article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the positive solution of the yamabe problem @xcite tells us that if @xmath1 is a compact smooth manifold ( without boundary ) , then every conformal class @xmath2 of riemannian metrics on @xmath1 contains a metric ( known as a yamabe metric ) of constant scalar curvature with the following special property .
its scalar curvature is the infimum of the scalar curvature @xmath3 , taken over all metrics in @xmath2 with constant scalar curvature and total volume @xmath4 .
the value of this scalar curvature is called the yamabe constant @xmath5 of @xmath2 .
equivalently , @xmath5 can be defined to be the minimum over metrics @xmath6 of the einstein - hilbert functional @xmath7 it is therefore natural to to ask if there is a `` best '' yamabe metric , and if so what its scalar curvature is . that motivates the following definition from @xcite .
the yamabe invariant of @xmath1 is defined by @xmath8 this supremum is not always attained , so the answer to the question about whether @xmath1 has a `` best '' metric of constant scalar curvature might be `` no . ''
the best that is known that there are singular metrics ( with singularities at a finite number of points ) which serve as the `` best '' approximation to an einstein metric on @xmath1 . nevertheless , @xmath9 is a diffeomophism invariant of @xmath1 .
it also turns out that @xmath10 if and only if @xmath1 admits a metric of positive scalar curvature , a much - studied condition ( @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) .
however , @xmath11 is possible even when @xmath1 admits no scalar - flat metric . in dimension
@xmath12 , gauss - bonnet quickly shows that @xmath13 . in dimension @xmath14
, @xmath9 can be positive , @xmath15 , or negative , and a lot is known about it from seiberg - witten theory ( @xcite and @xcite ) .
similarly , there is a conjectural connection between @xmath9 and `` geometrization '' when @xmath16 ( see for instance @xcite ) .
but even when @xmath16 , and especially when @xmath17 , it is not yet known if there are * any * manifolds with @xmath18 .
( the obvious candidates for such manifolds are hyperbolic manifolds , but for all we know they could have vanishing yamabe invariant . ) in fact , petean @xcite has proved that @xmath19 for any simply connected manifold of dimension at least @xmath20 . in this paper
we study the yamabe invariant for manifolds with finite fundamental groups .
our first main result is the following .
[ thm : thma ] let @xmath1 be a closed , connected , compact manifold with finite fundamental group @xmath21 , @xmath22 .
suppose all sylow subgroups of @xmath21 are abelian .
assume either that @xmath1 is spin and the order of @xmath21 is odd , or else that the universal cover of @xmath1 is non - spin
. then @xmath23 .
the proof of this result is somewhat involved .
first of all , we use surgery tools ( developed in the study of positive scalar curvature ) to reduce the assertion of theorem [ thm : thma ] to special situations . in particular
, we show that it is enough to study the case when @xmath21 is a finite abelian @xmath24-group . the central objects to understand here are the bordism groups @xmath25 and @xmath26 , and the proof amounts to the fact that all elements of these bordism groups may be represented by manifolds with nonnegative yamabe invariant .
a computation of these groups is quite hard , and the actual answer is known only for elementary abelian groups of odd order and few other cases ( see @xcite ) .
instead we use the knneth formula to build manifolds with nonnegative yamabe invariant to represent generators of these bordism groups .
there are two types of `` building blocks '' : tensor products ( which are realized by direct products of manifolds ) and torsion products ( which geometrically are just toda brackets ) .
we recall that toda bracket @xmath27 is defined when @xmath28 and @xmath29 .
then the manifold @xmath30 represents the toda bracket @xmath27 . as usual , to prove new geometric results we have to employ some _ new geometric techniques_. roughly , we show ( under some restrictions ) that if @xmath9 and @xmath31 are @xmath32 ( resp . , @xmath33 )
, then @xmath34 ( resp . , @xmath33 ) .
we prove this by analytical means using elementary differential geometry .
our second main result is the following .
[ thm : thmb ] let @xmath1 be a closed , connected , compact manifold with fundamental group @xmath21 of odd order .
suppose all sylow subgroups of @xmath21 are elementary abelian of rank @xmath35 .
assume that @xmath1 is non - spin and that @xmath36 .
then @xmath1 has a metric of positive scalar curvature . to put these results in context , it s worth recalling what is known about positive scalar curvature for manifolds with finite fundamental group . for such manifolds ( of dimension @xmath0 )
whose universal cover is non - spin , _ there are no known obstructions to positive scalar curvature_. for spin manifolds of dimension @xmath0 with finite fundamental group , the only known obstructions to positive scalar curvature come from the index theory of the dirac operator ( @xcite , @xcite ) , and it is known that `` stably '' these are the only obstructions @xcite .
in fact in @xcite , it was conjectured ( on the basis of extremely spotty evidence ) that the index theory of the dirac operator provides the only obstructions to positive scalar curvature on manifolds of dimension @xmath0 with finite fundamental group .
this conjecture has sometimes been called the gromov - lawson - rosenberg conjecture . however
, the `` stable '' theorem by itself does not actually answer the question of whether any _ particular _ manifold with vanishing dirac obstructions admits a metric of positive scalar curvature .
it _ is _ known @xcite that for spin manifolds of dimension @xmath0 with finite fundamental group with periodic cohomology , the dirac obstructions are the only obstructions to positive scalar curvature .
a similar theorem was proved by schultz @xcite , and independently by botvinnik and gilkey @xcite , for spin manifolds of dimension @xmath0 with fundamental group @xmath37 , @xmath24 an odd prime .
but very little was previously known about positive scalar curvature for manifolds with elementary abelian fundamental group of rank @xmath38 .
the proof of theorem [ thm : thmb ] is based on a reduction to the results of @xcite , again using toda brackets .
the outline of the paper is as follows .
section [ sec : redtools ] recalls the surgery and bordism theorems necessary for attacking the problems .
section [ sec : tools ] contains our basic geometric results on toda brackets .
section [ sec : todaresults ] puts together the topological and geometrical tools to prove theorem [ thm : thma ] and related results , and section [ sec : psc ] proves theorem [ thm : thmb ] and related results .
we would like to thank sergey novikov for his encouragement and support .
to warm up , we recall the following result of petean for simply connected manifolds : [ thm : peteansimplyconn ] if @xmath39 is a connected , simply connected closed manifold of dimension @xmath40 , then @xmath41 .
the proof of this fact is based on the following _ surgery theorem _ : [ thm : surgery]if @xmath1 is a closed manifold with connected components @xmath42 , and if another closed connected manifold @xmath43 can be obtained from @xmath1 by surgeries in codimension @xmath44 , and if @xmath45 for each @xmath46 , then @xmath47 .
this is really three theorems in one .
if @xmath48 for all @xmath46 , then @xmath1 admits a metric of positive scalar curvature , hence so does @xmath43 , by the surgery theorem of gromov - lawson and schoen - yau ( @xcite and @xcite some of the details are carefully redone in theorem 3.1 of @xcite ) , and so @xmath49 . if @xmath1 is disconnected and @xmath50 for some components and @xmath51 for other components , then we may first replace @xmath1 by the connected sum of its components , which has @xmath52 by iterated application of case ( b ) of @xcite , theorem 1 .
( see also @xcite . )
this reduces us to the case where @xmath1 is connected .
if @xmath1 is connected and @xmath53 , then the corollary to theorem 1 of @xcite says @xmath54 , so if @xmath9 is exactly @xmath15 , @xmath47 .
in this paper we will discuss what can be learned about the yamabe invariant for non - simply connected manifolds , using theorem [ thm : surgery ] .
many of the basic facts about manifolds of positive scalar curvature , which are proved using the surgery theorem of gromov - lawson and schoen - yau , have obvious counterparts for manifolds with nonnegative yamabe invariant , obtained by substituting theorem [ thm : surgery ] in the proof .
the proofs are almost identical to those in the positive scalar curvature case , so while we will give complete statements of the results , we will be brief when it comes to details of the proofs .
first we need to convert the surgery theorem , theorem [ thm : surgery ] , to a bordism theorem .
we repeat some definitions from @xcite and @xcite : let @xmath55 be a fibration . a _
@xmath56-structure _ on a manifold is defined to be a lifting of the ( classifying map of the ) stable normal bundle to a map into @xmath56 .
then one has bordism groups @xmath57 of manifolds with @xmath56-structures , defined in the usual way .
( for instance , if @xmath58 , mapping as usual to @xmath59 , then @xmath60 . )
we note that given a connected closed manifold @xmath1 , there is a choice of such a @xmath56 for which @xmath1 has a @xmath56-structure and the map @xmath61 is a 2-equivalence .
[ ex : cases ] the following special cases show that many of the classical bordism theories arise via this construction . 1 .
if @xmath1 is a spin manifold , choose @xmath62 , where @xmath63 , and let @xmath55 be the projection onto the second factor composed with the map @xmath64 induced by @xmath65 .
map @xmath1 to the first factor by means of the classifying map for the universal cover , and to the second factor by means of the spin structure .
the map @xmath61 is a @xmath12-equivalence since it induces an isomorphism on @xmath66 and @xmath67 .
the associated bordism theory is @xmath68 .
2 . if @xmath1 is oriented and the universal cover @xmath69 of @xmath1 is non - spin , choose @xmath70 , where @xmath63 , and let @xmath71 be the obvious map .
map @xmath1 to the first factor by means of the classifying map for the universal cover , and to the second factor by means of the orientation
. the map @xmath61 is a @xmath12-equivalence since it induces an isomorphism on @xmath66 and @xmath72 , with the map @xmath73 corresponding to @xmath74 . the associated bordism theory is @xmath75 .
if @xmath1 is not orientable and the universal cover of @xmath1 is non - spin , let @xmath63 , and let @xmath56 be defined by the homotopy pull - back diagram @xmath76 \ar[d ] & b\pi \ar[d]^{w_1}\\ bo \ar[r]^{w_1 } & \,{\mathbb r}{\mathbb p}^\infty,}\ ] ] where the maps labeled @xmath77 are defined by the first stiefel - whitney class .
note that @xmath59 has fundamental group @xmath78 and that @xmath79 induces an isomorphism on @xmath66 , so that @xmath56 has fundamental group @xmath21 .
the map @xmath55 can be taken to be a fibration with fiber @xmath80 , where @xmath81 is the fundamental group of the oriented double cover of @xmath1 .
then the maps of @xmath1 to @xmath82 by means of the classifying map for the universal cover and to @xmath59 by means of the classifying map for the stable normal bundle define a map from @xmath1 to @xmath56 which is a @xmath12-equivalence for the same reason as in the last example .
we will denote the associated bordism theory by @xmath83 ; it is a `` twisted version '' of unoriented bordism with coefficients in @xmath80 , and it obviously comes with a natural map to @xmath84 . in the special case where @xmath21 splits as @xmath85 , with @xmath86 ,
then @xmath56 becomes simply @xmath87 , and the associated bordism theory is @xmath88 . in general , @xmath83 is more complicated to describe , though the following proposition often tells as much as one needs to know about it .
[ prop : nonorient]in example [ ex : cases].3 , if @xmath89 is surjective , the natural map @xmath90 is injective , at least on classes in dimension @xmath91 represented by disjoint unions of connected manifolds @xmath42 with non - trivial @xmath77 and with each @xmath92 an isomorphism on @xmath66 . consider a class in @xmath93 represented by @xmath94 , @xmath95 , with @xmath96 the classifying map for the universal cover , and suppose the first stiefel - whitney class of @xmath42 is @xmath97 and is non - trivial .
assume @xmath98 bounds in @xmath99 .
that means we have a manifold @xmath100 with @xmath101 and with a map @xmath102 extending each @xmath103 .
we may assume @xmath104 is connected .
then @xmath105 ( the induced map on fundamental groups ) is a split surjection , with splitting map @xmath106 , where @xmath107 .
since the first stiefel - whitney class for @xmath104 must extend @xmath97 , it is trivial on @xmath108
. thus we may do surgery on embedded circles in the interior of @xmath104 and in the kernel of @xmath105 to reduce to the case where @xmath105 is also an isomorphism on @xmath66 .
( the assumption of dimension @xmath91 makes it possible to kill @xmath108 completely with such surgeries . ) then the first stiefel - whitney class of @xmath104 must be represented by @xmath109 , and we obtain a map @xmath110 extending @xmath111 , showing that @xmath112 bounds in @xmath93 .
the simply connected cases of the positive scalar curvature analogue of the following theorem were proved in @xcite ; the general case of the positive scalar curvature analogue , with this formulation , is in @xcite and @xcite .
[ thm : bordthm ] let @xmath39 be a connected @xmath56-manifold with @xmath113 , and assume that the map @xmath61 is a 2-equivalence .
then @xmath19 if and only if the @xmath56-bordism class of @xmath1 lies in the subgroup of @xmath57 generated by @xmath56-manifolds with nonnegative yamabe invariant .
let @xmath114 be a @xmath56-manifold @xmath56-bordant to @xmath1 .
the hypotheses combine ( via the method of proof of the @xmath115-cobordism theorem ) to show that @xmath1 can be obtained from @xmath114 by surgeries in codimension @xmath44
. then if each component of @xmath114 has nonnegative yamabe invariant , one can apply theorem [ thm : surgery ] to conclude that the same is true for @xmath1 .
this does it since addition in @xmath57 comes from connected sum and additive inverses correspond to reversal of orientation , etc .
, which does nt affect the yamabe invariant of the underlying manifold . fortunately for applications , one can do better than this . for simplicity ,
we restrict attention to the three cases discussed in examples [ ex : cases ] .
[ thm : homology ] let @xmath39 be a compact connected manifold with @xmath113 . 1 .
if , as in example [ ex : cases].1 , @xmath1 is spin with fundamental group @xmath21 , then @xmath19 if and only the class of @xmath111 in @xmath116 lies in the subgroup @xmath117 generated by classes of @xmath118 with @xmath43 a spin manifold with nonnegative yamabe invariant , and @xmath118 a map not necessarily an isomorphism on @xmath66 . here
@xmath119 is the homology theory corresponding to the connective real @xmath120-theory spectrum .
if , as in example [ ex : cases].2 , @xmath1 is oriented with fundamental group @xmath21 , and the universal cover of @xmath1 is * not * spin , then @xmath19 if and only the class of @xmath111 in @xmath121 lies in the subgroup @xmath122 generated by classes of @xmath118 with @xmath43 an oriented manifold with nonnegative yamabe invariant , and @xmath118 a map not necessarily an isomorphism on @xmath66 .
if , as in example [ ex : cases].3 , @xmath1 is non - orientable with fundamental group @xmath21 , and if the universal cover of @xmath1 is * not * spin , then @xmath19 if and only the class of @xmath111 in @xmath123 lies in the subgroup @xmath124 generated by classes of @xmath118 with @xmath43 a manifold with nonnegative yamabe invariant , and @xmath118 a map .
it was proved by jung and stolz ( see @xcite and @xcite ) that the kernel of the map @xmath125 in case 1 , and the kernel of the map @xmath126 in case 2 , are represented by manifolds with positive scalar curvature .
thus the result immediately follows from theorem [ thm : bordthm ] .
now consider case 3 .
the `` only if '' direction is obvious , so suppose we are given @xmath1 non - orientable with fundamental group @xmath21 and universal cover non - spin , and assume the class of @xmath111 in @xmath123 lies in the subgroup @xmath124 . by theorem [ thm : bordthm ] and example
[ ex : cases].3 , it suffices to show that the class of @xmath1 in @xmath93 lies in @xmath127 . by proposition [ prop : nonorient ] , @xmath93 is detected by its image in @xmath99 . since the unoriented bordism spectrum is an eilenberg - mac lane spectrum , @xmath128 and each class in @xmath99 is a sum of classes of the form @xmath129 , where @xmath130 represents a class in @xmath131 . here
the summand with @xmath132 corresponds to the image of @xmath111 in @xmath123 .
we claim every class in @xmath131 , @xmath133 , is represented by a manifold with nonnegative yamabe invariant .
indeed , multiplicative generators of @xmath131 can be taken to be real projective spaces and quadric hypersurfaces in products of real projective spaces ( @xcite , p. 97 ) .
all of these manifolds admit metrics of positive scalar curvature(cf .
the argument in the proof of @xcite , corollary c ) , except for a point in dimension @xmath132 .
so by theorem [ thm : surgery ] above and proposition [ prop : prod ] below , if the class of @xmath111 lies in @xmath124 , then the class of @xmath111 in @xmath99 is represented by a map @xmath118 , with @xmath43 a manifold with nonnegative yamabe invariant .
choose a bordism @xmath102 between @xmath111 and @xmath118 .
as in the proof of proposition [ prop : nonorient ] , we may assume ( by doing surgeries on the interior of @xmath104 ) that @xmath105 is an isomorphism on @xmath66 .
as in the proof of proposition [ prop : nonorient ] , this implies @xmath43 and @xmath1 represent the same element of @xmath93 , and we conclude using theorem [ thm : bordthm ] .
this is now enough machinery to deal with `` easy '' torsion - free fundamental groups : let @xmath39 be a closed connected @xmath134-manifold with a fundamental group @xmath21 which is either free abelian or of homological dimension @xmath135 .
this includes the fundamental groups of aspherical @xmath12-manifolds , @xmath136-manifolds , and @xmath14-manifolds .
assume either that @xmath1 is spin or that its universal cover is non - spin .
in the spin case , also assume that the atiyah - hirzebruch spectral sequence @xmath137 collapses .
this is automatic if @xmath21 is of homological dimension @xmath138 . then if @xmath40 , @xmath1 has nonnegative yamabe invariant . by theorem [ thm :
homology ] , it s enough to show that for each of these groups @xmath21 , @xmath122 exhausts @xmath121 and @xmath117 exhausts @xmath116 for @xmath40 .
the non - spin case is easy , since for @xmath21 free abelian and any @xmath134 , @xmath121 is generated additively by the classes of tori , which carry flat metrics and thus have yamabe invariant zero , whereas if @xmath21 has homological dimension @xmath135 , @xmath121 vanishes for @xmath40 .
so consider the spin case .
when @xmath21 is free abelian , the atiyah - hirzebruch spectral sequence collapses and @xmath139 thus this group is generated by the classes of @xmath140 , where the map @xmath141 factors through @xmath142 .
since , as pointed out in @xcite , @xmath119 is generated by the classes of manifolds of nonnegative yamabe invariant , we have the desired result .
the other cases are similar but easier .
most of this paper will now deal with the opposite extreme , the case where @xmath143 is finite . in this case , the following results reduce us to the case where the fundamental group is a @xmath24-group .
[ lem : coverings ] suppose @xmath39 is a closed connected manifold with @xmath19 , and suppose @xmath69 is a finite covering of @xmath1 .
then @xmath144 .
let @xmath145 be the number of sheets of the covering @xmath146 . by assumption , given @xmath147
, we can choose a conformal class @xmath2 on @xmath1 with @xmath148 .
that means there is a metric @xmath149 on @xmath1 with unit volume and constant scalar curvature @xmath150 .
lift the metric @xmath149 up to @xmath69 .
that gives a metric on @xmath69 with volume @xmath145 and scalar curvature @xmath151 .
rescaling , we get a metric on @xmath69 with unit volume and scalar curvature @xmath152 .
this being true for all @xmath147 , it follows that @xmath144 .
[ prop : invprop ] if @xmath66 and @xmath153 are groups and if @xmath154 is a group homomorphism , then @xmath155 sends @xmath156 to @xmath157 , @xmath158 to @xmath159 , and @xmath160 to @xmath161 . if @xmath66 is a subgroup of @xmath153 of finite index , then the transfer map on @xmath162 or @xmath163 sends @xmath157 to @xmath156 , @xmath159 to @xmath158 , and @xmath161 to @xmath160 .
the first statement is obvious from the definitions in theorem [ thm : homology ] .
the second statement follows from lemma [ lem : coverings ] , since the transfer is realized geometrically via coverings .
[ thm : ks ] let @xmath39 be a closed connected @xmath134-manifold with finite fundamental group @xmath21 . assume either that @xmath1 is spin or that its universal cover is non - spin .
for each prime @xmath24 , let @xmath164 be the inclusion of a sylow @xmath24-subgroup of @xmath21 , and let @xmath165 , @xmath166 , @xmath167 be the transfer maps
. then @xmath1 has nonnegative yamabe invariant if and only if @xmath168)$ ] lies in the subgroup @xmath169 in the oriented non - spin case , or in @xmath117 in the spin case , for each @xmath24 dividing the order of @xmath21 . in the non - orientable non - spin case , @xmath1 has nonnegative yamabe invariant if and only if @xmath170)$ ] lies in @xmath159 .
the proof is almost word - for - word as in @xcite , but we review the argument .
the `` only if '' statement is contained in proposition [ prop : invprop ] . as for the `` if '' statement , let @xmath171 , @xmath172 , or @xmath173 , and let @xmath56 be the subgroup @xmath174 , @xmath175 , or @xmath176 . similarly let @xmath177 , @xmath178 , or @xmath179 , and let @xmath180 be the subgroup @xmath181 , @xmath182 , or @xmath183 .
( we can work with reduced homology since @xmath184 and @xmath185 . )
note that @xmath186 is a finite group and @xmath56 is a subgroup ; we are trying to show that an element @xmath187 $ ] of @xmath186 lies in @xmath56 if @xmath168)\in b_p$ ] for all @xmath24 .
now @xmath188 is an endomorphism of @xmath186 which is an isomorphism on @xmath189 , since @xmath190 $ ] is a unit modulo @xmath24 . if @xmath168)\in b_p$ ] for all @xmath24 , then @xmath191)\in i_p(b_p)\subseteq b$ ] for all @xmath24 , by proposition [ prop : invprop ] .
so that means the image of @xmath187 $ ] in @xmath189 lies in @xmath192 for all @xmath24 , and thus @xmath187 $ ] lies in @xmath56 .
note , incidentally , that in the non - orientable , non - spin case , only the sylow 2-subgroup matters , since @xmath193 for @xmath194 .
[ thm : stablesplitting ] let @xmath21 be a finite group , and let @xmath195 be an idempotent in the stable homotopy category giving a stable splitting of @xmath82
. then @xmath196 maps @xmath122 , @xmath124 , and @xmath117 into themselves .
as pointed out in @xcite , the proof of the segal conjecture implies that the stable splittings of @xmath82 are essentially linear combinations of products of transfer maps and maps induced by group homomorphisms , so the result then follows from proposition [ prop : invprop ] .
in this section we present a number of analytic results that can be used to study the classes of manifolds with positive scalar curvature or with nonnegative yamabe invariant .
first we need a basic characterization of manifolds in the latter class .
[ prop : nnychar]let @xmath39 be a closed @xmath134-manifold .
then : i if @xmath1 does not admit a metric of positive scalar curvature , then @xmath197 here the infimum is taken over all riemannian metric @xmath149 on @xmath1 , and @xmath3 denotes the scalar curvature of @xmath149 .
ii suppose that for each @xmath147 , there exists a metric @xmath149 on @xmath1 with volume @xmath4 and @xmath198
. then @xmath19 .
the converse is true if @xmath199 or if @xmath200
. statement ( i ) is proposition 1 in @xcite . as for ( ii ) , suppose the condition is satisfied .
if @xmath10 , then we have nothing to prove , and if not , ( i ) shows that @xmath19 . in the converse direction , suppose @xmath19 .
if @xmath199 and if @xmath10 , then by a theorem of kazdan and warner @xcite , @xmath1 admits a metric @xmath149 with @xmath201 , and obviously we may rescale @xmath149 to have volume 1 without changing this condition .
if , on the other hand , @xmath11 , that means , by definition of the yamabe invariant ( recall equation ) , that for all @xmath147 , there exists a metric @xmath149 on @xmath1 with volume @xmath4 and @xmath202 constant and @xmath203 .
so again the condition of ( ii ) is satisfied .
another basic fact is the following : [ prop : prod ] suppose @xmath204 and @xmath205 are closed manifolds , @xmath206 , and also @xmath207
. then @xmath208 . if @xmath209 , then @xmath114 admits a metric of positive scalar curvature and so does @xmath210 , so @xmath211 . if @xmath212 , then by proposition [ prop : nnychar ] , given @xmath147 , there exists a metric @xmath213 on @xmath114 with volume @xmath4 and @xmath198 .
choose any metric @xmath214 on @xmath1 with volume @xmath4 .
if we give @xmath210 the product metric @xmath215 ( where @xmath216 means @xmath214 rescaled by multiplying all distances by @xmath217 ) , then this product metric has scalar curvature @xmath218 and volume @xmath219 .
so the integral in ( with @xmath210 in place of @xmath1 ) is @xmath220 for some constant @xmath221 ( the maximum of @xmath222 over @xmath1 ) independent of @xmath217 and @xmath223 .
so the idea is to take @xmath217 large , and then given @xmath217 , to take @xmath223 of order @xmath224 . in equation
, we see that the integral on the left - hand side is bounded by a constant times @xmath225 which goes to zero as @xmath226 . thus by proposition [ prop : nnychar ] ,
the result follows .
next , we discuss the extension of the minimal hypersurface technique of @xcite to the study of nonnegative yamabe invariant .
suppose @xmath39 is a closed manifold with metric @xmath149 , and suppose @xmath227 is a stable minimal hypersurface in @xmath1 . in @xcite , it was shown that if @xmath228 , then @xmath229)>0 $ ] , where @xmath230 denotes the induced metric on @xmath231 and @xmath232 $ ] is its conformal class . in particular , there is a metric in @xmath232 $ ] with positive scalar curvature , and this can be used to rule out positive scalar curvature metrics on many non - simply connected manifolds .
now it is not true that just because @xmath19 , then @xmath233 , since by proposition [ prop : prod ] , we can get a counterexample by taking @xmath234 and @xmath235 ( say with @xmath236 or @xmath14 ) .
however , the same estimates used in the proof of theorem in @xcite show that if @xmath237 , where @xmath120 is a constant , then because the second variation of the @xmath238-dimensional volume of @xmath231 is positive , one has @xmath239 for all functions @xmath240 not vanishing identically .
( here @xmath241 , the scalar curvature of @xmath231 , and @xmath242 are to be computed with respect to the induced metric @xmath230 . )
assume @xmath243 and consider the `` conformal laplacian '' @xmath244 of @xmath231 , where @xmath245 is the usual ( non - negative ) laplacian .
( recall that the dimension of @xmath231 is @xmath246 , not @xmath134 . )
then for @xmath247 as above we have @xmath248 note the use of equation at the last step .
this implies that if @xmath120 is close to @xmath15 , then the conformal laplacian @xmath249 is not too negative , and thus @xmath250 is not too negative , provided that the ( @xmath246)-dimensional volume of @xmath231 is not too large .
if @xmath251 , things are even easier : we instead take @xmath252 in equation and apply gauss - bonnet .
these arguments thus prove the following two results : [ thm : minhyp ] let @xmath39 be a closed manifold with metric @xmath149 , and suppose @xmath227 is a stable minimal hypersurface in @xmath1 .
also suppose that the metric @xmath149 is scalar - flat
. then @xmath233 .
immediate from the above estimates .
[ thm : circcrosssurf ] let @xmath253 be a closed oriented surface of genus @xmath254 , and let @xmath255
. then @xmath212 by proposition [ prop : prod ] .
it can not be strictly positive , by , theorem 8.1 , for example .
thus there is a sequence @xmath256 of metrics on @xmath257 with volume @xmath4 and constant scalar curvatures @xmath258 , with the scalar curvatures tending to @xmath15 as @xmath259 . on the other hand , for any such sequence of metrics , @xmath260 .
choose the metrics @xmath256 as in the statement of the theorem .
choose minimal submanifolds @xmath42 which are absolutely area - minimizing in the homology class @xmath261\in h_2(s^1\times m^2,{\mathbb z})$ ] for the metric @xmath256 . by inequality with @xmath262 , @xmath263 on the other hand , each @xmath42 must be a surface of genus @xmath264 , since it represents a non - trivial homology class in the infinite cyclic cover @xmath265 , while each mapping of a sphere into this space is null - homotopic and each mapping of a torus into this space factors through a circle ( since each abelian subgroup of @xmath143 is cyclic ) and is thus trivial in @xmath266 .
so by gauss - bonnet , @xmath267 . comparing this with equation
, we see the area of @xmath42 with respect to @xmath268 must tend to @xmath269 as @xmath270 , while the average value of @xmath271 must go to @xmath15 , and in particular , @xmath272 .
this in turn means @xmath273 , since otherwise we could choose representatives for the homology class @xmath187 $ ] in @xmath274 with bounded diameters , a contradiction .
the next two results are the most significant in this paper ; they will be used in the next section to deal with `` toda brackets , '' among the most intractable of bordism classes .
[ thm : psctoda]let @xmath275 and @xmath276 be closed manifolds , not necessarily connected , that admit metrics of positive scalar curvature .
suppose @xmath277 and @xmath278 for some compact manifolds with boundary , @xmath279 and @xmath280 .
form a new manifold @xmath281 of dimension @xmath282 , where @xmath283 and @xmath284 are the dimensions of @xmath275 and @xmath276 .
then @xmath1 admits a metric of positive scalar curvature .
we start by choosing metrics of positive scalar curvature , @xmath285 and @xmath286 , on @xmath275 and @xmath276 , respectively .
extend them to metrics @xmath287 and @xmath288 on @xmath279 and @xmath280 , which are product metrics in neighborhoods of the boundaries .
the trick is to write @xmath1 as a union of four pieces ( not two ) as follows : @xmath289 where the `` tubes '' @xmath290 and @xmath291 are ( as smooth manifolds ) @xmath292 and @xmath293 , respectively .
call the pieces here @xmath294 , @xmath295 , @xmath296 , and @xmath297 in that order .
since @xmath285 and @xmath286 have positive scalar curvature , we can choose ( very small ) constants @xmath298 and @xmath299 so that the metric @xmath300 on @xmath294 and the metric @xmath301 on @xmath297 have positive scalar curvature .
now all we have to do is choose the metric @xmath302 on @xmath290 to interpolate between @xmath303 and @xmath285 and the metric @xmath304 on @xmath291 to interpolate between @xmath305 and @xmath286 .
if the tubes @xmath290 and @xmath291 are `` very long , '' it is possible to do this so that @xmath290 and @xmath291 have positive scalar curvature , by the `` isotopy implies concordance '' theorem , @xcite , lemma 3 .
( in fact , in this case , one can write down an explicit warped product metric that does the trick . )
then all the metrics fit together to give a metric of positive scalar curvature on @xmath1 .
the next theorem is quite similar , but considerably more delicate .
[ thm : nnyitoda]let @xmath275 and @xmath276 be closed manifolds , not necessarily connected , each with nonnegative yamabe invariant . when @xmath42 is disconnected , we mean that each component is required to have nonnegative yamabe invariant .
suppose @xmath277 and @xmath278 for some compact manifolds with boundary , @xmath279 and @xmath280 .
form a new manifold @xmath281 of dimension @xmath282 , where @xmath283 and @xmath284 are the dimensions of @xmath275 and @xmath276 . then , excluding the case where @xmath306 , @xmath307 , and @xmath308 , it follows that @xmath19 .
we follow the same approach as in the proof of theorem [ thm : psctoda ] .
if @xmath309 and @xmath310 are both strictly positive , we re done by theorem [ thm : psctoda ] , so we may assume at least one of @xmath275 and @xmath276 has @xmath311 .
then we re excluding the case where the other manifold is a disjoint union of copies of @xmath312 or @xmath313 , so by proposition [ prop : nnychar ] , we may assume both manifolds have metrics of unit volume which are almost scalar - flat .
by proposition [ prop : nnychar ] , it is enough to show that @xmath1 has a metric for which the integral in is as small as one likes . we will estimate the integral separately over the four pieces of @xmath1 ( as in the last proof ) and add the results .
fix @xmath147 and choose metric @xmath285 and @xmath286 on @xmath275 and @xmath276 , respectively , each with volume @xmath4 and with small constant scalar curvatures , @xmath314 and @xmath315 , respectively , with @xmath316 .
extend @xmath285 and @xmath286 to metrics @xmath287 and @xmath288 on @xmath279 and @xmath280 , which are product metrics in neighborhoods of the boundaries .
then the scalar curvature of the metric @xmath317 on @xmath294 is @xmath318 and the scalar curvature of the metric @xmath319 on @xmath297 is @xmath320 .
( the constants @xmath321 and @xmath322 will be chosen later . ) furthermore , the volumes of these metrics are @xmath323 for @xmath294 and @xmath324 for @xmath297 . letting @xmath321 and @xmath322 go to @xmath15 ,
we see there are constants @xmath325 and @xmath326 with @xmath327 the right - hand sides of and can be rewritten as @xmath328 and @xmath329 respectively .
next we need to deal with the tubes @xmath290 and @xmath291 .
we give these warped product metrics of the form @xmath330 , @xmath331 , where @xmath332 is the standard metric on the line , and @xmath333 is the parameter along the length of the tube .
the function @xmath334 will be chosen to interpolate between @xmath15 and @xmath335 .
if we write @xmath336 , we need to choose @xmath103 as in the following picture , so that the graph has vanishing first and second derivatives at both ends : here @xmath337 , to be taken large , is the length of the tube . since @xmath338 and @xmath339
, the volume of @xmath340 will be bounded by @xmath337 , as will the volume of @xmath295 or @xmath296 , when we take the product with the metric @xmath286 on @xmath276 or @xmath285 on @xmath275 .
the scalar curvature of @xmath340 is given by equation ( 7.35 ) on p. 157 of @xcite , which gives : @xmath341 since @xmath342 , @xmath343 and @xmath344 .
now we can choose @xmath103 so that @xmath345 is bounded by a constant times @xmath346 and @xmath347 is bounded by a constant times @xmath348 .
thus the scalar curvature of @xmath340 is bounded in absolute value by @xmath349 for some constant @xmath350 .
thus the integrals over @xmath295 and @xmath296 give : @xmath351 now all we have to do is choose the parameters @xmath321 , @xmath223 , and @xmath337 to make all of , , , and simultaneously small .
we do this as follows .
first choose @xmath321 and @xmath322 very small . then choose @xmath337 large enough so that the terms @xmath352 are small .
then finally choose @xmath353 extremely small so that @xmath354 is also small .
that does it .
we re now ready for the first main results of this paper .
[ thm : nnyinonspinabel ] let @xmath39 be a closed , connected , @xmath134-manifold with abelian fundamental group , with non - spin universal cover , and with @xmath40
. then @xmath1 has nonnegative yamabe invariant .
consider the oriented case .
( the non - orientable case works exactly the same way , but with integral homology replaced by homology with coefficients in @xmath78 . ) by theorem [ thm : homology ] , it s enough to show that @xmath355 exhausts the image in @xmath356 of @xmath357 .
write @xmath143 as @xmath358 , with @xmath21 finite abelian .
since the homology of a free abelian group is torsion free , the knneth theorem gives @xmath359 and so the homology of @xmath360 is generated by classes of products of tori with homology classes of @xmath82 .
so by proposition [ prop : prod ] , we only have to show that @xmath361 exhausts the image in @xmath362 of @xmath357 .
in other words , we are reduced to the case of _ finite _ abelian groups . by theorem [ thm : ks ]
, we can further assume that @xmath143 is a finite abelian @xmath24-group for some prime @xmath24 ( and in the non - orientable case , we can further assume that @xmath363 ) .
we will come back to finite abelian @xmath24-groups after a short digression .
[ lem : cyclic]let @xmath21 be a cyclic group of prime power order @xmath364 . then each class in @xmath121
is represented by an oriented manifold with nonnegative yamabe invariant , and if @xmath365 , by an oriented manifold with positive scalar curvature .
if @xmath363 , then also each class in @xmath123 is represented by a manifold not necessarily orientable with nonnegative yamabe invariant , and if @xmath365 , by a manifold with positive scalar curvature .
note that @xmath366 is cyclic of order @xmath364 , with a generator represented by the lens space @xmath367 , and @xmath368 vanishes for @xmath369 .
since the lens space has positive scalar curvature except in the exceptional case @xmath370 , when it has a flat metric , the first statement is immediate .
now consider the case of @xmath78-homology and @xmath363 .
then @xmath123 is cyclic of order @xmath12 for all @xmath371 .
when @xmath372 , things are again easy , as we have generators @xmath373 for all the homology groups , with positive scalar curvature except in the exceptional case @xmath374 .
this leaves the case @xmath375 . in that case ,
@xmath376/(v^2)$ ] , where @xmath377 is the field of two elements , @xmath378 corresponds to the group homomorphism @xmath379 , and @xmath380 is the class of the central extension @xmath381 via the usual correspondence between @xmath382 and central extensions .
since the pull - back of this central extension to the unique two - element subgroup of @xmath21 is non - trivial , the inclusion @xmath383 induces an isomorphism @xmath384 and thus @xmath385 is the generator of @xmath386 . by duality , @xmath387 is an isomorphism , and so the generator of @xmath388 is represented by @xmath389 we still need to find geometric generators for @xmath390 .
since @xmath391 is generated by @xmath392 , and the cup product comes from restricting the exterior product @xmath393 to the diagonal copy of @xmath82 , we see by duality that the generator of @xmath390 is represented by @xmath394 , where @xmath395 is the diagonal map and @xmath396 denotes the transfer ( which corresponds to taking a covering ) . since transfer preserves the positive scalar curvature or the nonnegative yamabe invariant property by proposition [ prop : invprop ] , the proof is complete .
proof of theorem [ thm : nnyinonspinabel ] , continued recall that we have already reduced to the case where the fundamental group @xmath21 of @xmath1 is a finite abelian @xmath24-group , hence a finite product of cyclic @xmath24-groups . in the non - orientable case ,
since @xmath397 is by the knneth theorem generated by products of homology classes for cyclic groups , and since all these homology classes are represented by manifolds of nonnegative yamabe invariant by lemma [ lem : cyclic ] , we are done by proposition [ prop : prod ] . in the orientable case ,
things are more complicated because we have to deal with the tor terms in the knneth theorem , and also because the natural map ( the hurewicz homomorphism for @xmath398 ) @xmath399 may not be surjective , and may not be split onto its image .
thus the argument will require some care .
we prove the theorem by induction on the rank ( the number of cyclic factors in a product decomposition ) of @xmath21 .
if the rank is @xmath4 , @xmath21 is cyclic and we are done by lemma [ lem : cyclic ] .
so assume the result is true for @xmath24-groups of smaller rank , and write @xmath400 , where we may assume that @xmath364 is less than or equal to the order of every cyclic factor of @xmath401 , and thus less than or equal to the order of every cyclic factor of the homology of @xmath80 .
first assume that @xmath363 . this case is easier
because @xmath398 localized at @xmath12 is a direct sum of eilenberg - mac lane spectra ( see @xcite and @xcite , or @xcite for a review of the literature ) , and thus @xmath362 can be identified with a direct summand in @xmath75 , and similarly for @xmath401 . by inductive hypothesis ,
each cyclic factor ( say of order @xmath402 , @xmath403 ) in @xmath404 is generated by the class of a manifold @xmath405 , where @xmath43 has nonnegative yamabe invariant .
if @xmath406 is odd , then we get a corresponding tensor term in the knneth formula for the homology of @xmath82 , and it is represented by a product of @xmath43 with either @xmath407 or a lens space , and so it is represented by a manifold with nonnegative yamabe invariant . if @xmath408 is even , there is a contribution to @xmath121 of the form @xmath409 , coming from @xmath404 and @xmath410 , which we need to represent by a manifold of nonnegative yamabe invariant .
since tor is left exact and @xmath409 is cyclic of order @xmath364 , the map @xmath411 induced by the inclusion @xmath412 is an isomorphism , so without loss of generality , we may replace @xmath43 by something representing a multiple of its homology class , and assume @xmath413 .
choose @xmath414 , with @xmath415 either @xmath407 or a lens space , of dimension @xmath416 , generating @xmath410 .
we may assume the bordism classes of @xmath405 and @xmath417 both have order @xmath364 .
then their tor product in the homology of @xmath82 may be represented by the cobordism massey product @xmath418 ( see @xcite ) , or in other words , by a toda bracket construction as in theorem [ thm : nnyitoda ] .
more precisely , choose @xmath279 bounding @xmath419 over @xmath80 and @xmath280 bounding @xmath420 over @xmath421 , and glue together @xmath422 and @xmath423 along their common boundary . by theorem [ thm :
nnyitoda ] , the resulting manifold is represented by a manifold with nonnegative yamabe invariant .
( note that the exceptional case of that theorem never arises . )
this completes the inductive step .
now consider the case where @xmath24 is odd . in this case
, it s important to point out that the inductive hypothesis is simply that the _ image _ of @xmath424 is represented by manifolds with nonnegative yamabe invariant , as this map is not usually surjective .
however , in this case we have one additional tool in our arsenal , namely landweber s knneth theorem for oriented bordism @xcite . more precisely , we apply theorem a of @xcite , which applies since @xmath425 consists entirely of odd torsion and the atiyah - hirzebruch spectral sequence @xmath426 collapses for dimensional reasons .
( @xmath427 localized at @xmath24 is free over @xmath428 and concentrated in degrees divisible by @xmath14 , and @xmath429 is non - zero only in odd degrees . )
note also that the proof of landweber s theorem shows that @xmath430 has homological dimension @xmath4 over @xmath431 .
now observe that we have a commutative diagram with exact rows : @xmath432 \ar@{=}[d ] & { \omega}_*(b\pi')\otimes_{{\omega } _ * } { \omega}_*(b{\mathbb z}/p^k ) \ar[r ] \ar[d]^\alpha & { \omega}_*(b\pi ) \ar[r ] \ar[d]^\beta & { \operatorname{tor}}_{{\omega}_*}\bigl({\omega}_*(b\pi'),{\omega}_*(b{\mathbb z}/p^k)\bigr ) \ar[r ] \ar[d]^\gamma & 0 \ar@{=}[d ] \\ 0\ar[r ] & h_*(b\pi')\otimes_{{\mathbb z } } h_*(b{\mathbb z}/p^k ) \ar[r ] & h_*(b\pi ) \ar[r ] & { \operatorname{tor}}_{{\mathbb z}}\bigl(h_*(b\pi'),h_*(b{\mathbb z}/p^k)\bigr ) \ar[r ] & \,0 , } \ ] ] in which the vertical arrows are induced by the natural transformation @xmath433 .
note that the map @xmath434 is surjective , and denote the image of @xmath435 by @xmath436 ( for `` representable homology '' ) .
the image of @xmath437 is then obviously @xmath438 ; classes here are represented by products of manifolds of nonnegative yamabe invariant ( because of the inductive hypothesis ) , so these have nonnegative yamabe invariant .
the image of @xmath439 is by definition @xmath440 , whereas the image of @xmath441 is evidently contained in @xmath442 .
also @xmath441 factors through @xmath443 but this latter group is generated by cobordism massey products @xmath444 , where @xmath405 and @xmath414 . by inductive hypothesis , @xmath43 and @xmath415
can be chosen to have nonnegative yamabe invariant , hence so can this toda bracket by theorem [ thm : nnyitoda ] .
again , the exceptional case of that theorem never arises in our context .
so this shows that the image of @xmath441 is represented by manifolds of nonnegative yamabe invariant .
diagram now shows that @xmath440 is represented by manifolds with nonnegative yamabe invariant , which completes the inductive step for the case @xmath24 odd .
[ cor : abeliansylow ] let @xmath39 be a closed , connected , @xmath134-manifold with finite fundamental group @xmath21 , with non - spin universal cover , and with @xmath445
. also assume all sylow subgroups of @xmath21 are abelian .
then @xmath1 has nonnegative yamabe invariant .
this is immediate from theorem [ thm : nnyinonspinabel ] and theorem [ thm : ks ] . in the odd order case
, we can carry this over to the spin case as well : [ thm : oddspinabeliansylow ] let @xmath39 be a closed , connected , spin @xmath134-manifold with finite fundamental group @xmath21 of odd order , and with @xmath445 . also assume all sylow subgroups of @xmath21 are abelian .
then @xmath1 has nonnegative yamabe invariant . by petean s theorem @xcite ,
this is true when @xmath21 is trivial .
as before , it s enough to consider the case of an abelian @xmath24-group , @xmath24 odd . but for @xmath21 of odd order , the natural map @xmath446 is an isomorphism , since the map of spectra @xmath447 is an equivalence after localizing at @xmath24 ( see @xcite ) .
we prove the result by induction on the rank of @xmath21 .
thus write @xmath400 , and assume by inductive hypothesis that the theorem is true for @xmath401 .
since @xmath448 and @xmath449 are isomorphisms , we have by landweber s theorem @xcite an exact sequence @xmath450 by inductive hypothesis , each element of @xmath451 is represented by a map @xmath405 , with @xmath43 a spin @xmath115-manifold with nonnegative yamabe invariant , and similarly each element of @xmath452 is represented by a map @xmath414 , with @xmath415 a spin @xmath217-manifold with nonnegative yamabe invariant .
then @xmath453\otimes [ m''\to b{\mathbb z}/p^k]$ ] in the tensor term on the left side of is represented by @xmath454 , which has nonnegative yamabe invariant .
furthermore , the tor term @xmath455 , [ m''\to b{\mathbb z}/p^k])$ ] on the right of pulls back ( non - canonically ) to the class of a toda bracket @xmath456 , where @xmath457 is some element from the coefficient ring @xmath458 . since @xmath43 , @xmath457 , and @xmath415
all have nonnegative yamabe invariant ( we do nt even need to know anything about @xmath457 ! ) , it follows from theorem [ thm : nnyitoda ] that this toda bracket has nonnegative yamabe invariant , and this completes the inductive step .
it turns out that the method of proof of theorem [ thm : nnyinonspinabel ] , if we replace theorem [ thm : nnyitoda ] by theorem [ thm : psctoda ] , gives substantial results on the positive scalar curvature problem for manifolds with finite abelian fundamental group for which the universal cover is non - spin , since all of the homology generators constructed above have positive scalar curvature by theorem [ thm : psctoda ] , _ except _ for those involving toda brackets and products of one - dimensional homology classes .
we proceed to formalize this as follows : [ def : toral ] let @xmath21 be a finitely generated abelian group . call a class in @xmath121 or in @xmath123 _ toral _ if it is represented by a map @xmath459 .
note that any such map is determined up to homotopy by the associated map @xmath460 on fundamental groups , which we may assume without loss of generality to have image of rank @xmath461 , where @xmath462 is the _ rank _ of @xmath21 , that is , the minimal number of cyclic factors when we write @xmath21 as a direct sum of cyclic groups , so toral classes only exist in degrees @xmath461 .
let @xmath463 be the subgroup generated by the the toral classes , and call this the _
toral subgroup_. [ prop : toral ] let @xmath21 be an elementary abelian @xmath24-group of rank @xmath462 , that is , @xmath464 .
then for all @xmath371 , @xmath121 is also elementary abelian , of rank equal to @xmath465 the toral subgroup @xmath466 is a direct summand in @xmath121 , of rank the binomial coefficient @xmath467 .
note that this vanishes for @xmath468 .
the homology groups @xmath469 vanish for @xmath369 even and are @xmath470 for @xmath134 odd .
so by iterated applications of the knneth theorem , all integral homology groups of @xmath21 ( other than @xmath471 , which is @xmath472 ) , are elementary abelian @xmath24-groups .
consider the poincar series @xmath473 then @xmath474 the knneth theorem gives the recursion relation @xmath475 where the first term comes from the `` tensor terms '' and the second term comes from the `` tor terms . ''
putting together equations ( [ eq : p1 t ] ) and ( [ eq : prt ] ) yields by induction on @xmath462 the formula @xmath476 for @xmath371 , the coefficient of @xmath477 in this expression is @xmath478 which is the expression in the statement of the proposition . on the other hand ,
the toral subgroup is generated just by the products of distinct generators of @xmath479 , so in degree @xmath134 , we have @xmath467 possibilities . [
prop : toralnonorient ] let @xmath21 be an elementary abelian @xmath12-group of rank @xmath462 , that is , @xmath480 .
then for all @xmath481 , @xmath123 is also elementary abelian , of rank equal to @xmath482 the toral subgroup @xmath483 is a direct summand in @xmath123 , of rank the binomial coefficient @xmath467 .
note that this vanishes for @xmath468 .
this is easier than the previous case since there are no tor terms .
the poincar series for @xmath484 is @xmath485 , and so the poincar series for @xmath486 is @xmath487 which yields the desired formula .
[ def : atoral ] let @xmath21 be a finite abelian @xmath24-group , say @xmath488
. for purposes of this definition , @xmath489 means homology with @xmath472 coefficients if @xmath194 , and homology with @xmath78 coefficients if @xmath490 .
we define a splitting @xmath491 inductively as follows . if @xmath492 , @xmath493 if @xmath494 and @xmath495 for @xmath496 , so we let @xmath497 .
if @xmath498 , write @xmath400 , @xmath499 , where @xmath500 . choose a splitting of the knneth formula @xmath501 by induction , we have splittings @xmath502 and @xmath503 .
the atoral part of the tensor term on the left is defined to be @xmath504 then we define @xmath505 to be the direct sum of the atoral piece of the tensor term with the image of the @xmath506 term under the splitting of the exact sequence .
[ def : rh ] for any space @xmath507 , we denote by @xmath508 the image of the hurewicz map @xmath509 in @xmath510 , and call it the _
representable homology_. ( this already made an appearance in the proof of theorem [ thm : nnyinonspinabel ] . ) note that @xmath511 is a functor , though not a homology theory . by lemma [ lem : cyclic ] , @xmath512 when @xmath21 is a cyclic group .
[ prop : ranktwo ] let @xmath21 be an elementary abelian @xmath24-group of rank @xmath12 , where @xmath24 is an odd prime . then @xmath513 is generated as an abelian group by the images of @xmath514 , as @xmath515 runs over the cyclic subgroups of @xmath21 .
[ thm : rhstructure ] let @xmath21 be an elementary abelian @xmath24-group , where @xmath24 is an odd prime . then @xmath440 is generated as an abelian group by elements @xmath516 with @xmath517 a subgroup of @xmath21 with each @xmath518 a cyclic @xmath24-group .
we prove this by induction on the rank @xmath462 .
when @xmath492 , the statement is trivially true , and when @xmath519 , this is proposition [ prop : ranktwo ] .
now assume the result for smaller values of @xmath462 , and write @xmath520 , where @xmath401 has rank @xmath521 .
we again use diagram ( with @xmath372 ) .
the image of @xmath437 is taken care of by inductive hypothesis .
so consider the image of @xmath441 .
consider a representable class @xmath522 , where @xmath523 and @xmath524 . by inductive hypothesis
, we may assume @xmath333 is the image of a representable class @xmath525 , with @xmath517 a subgroup of @xmath401 with each @xmath518 a cyclic @xmath24-group .
it will suffice to show that @xmath526 is of the correct form back in @xmath527 .
represent each @xmath528 by a manifold @xmath529 which is either @xmath407 or a lens space , and also represent @xmath530 by a manifold @xmath531 which is either @xmath407 or a lens space .
then @xmath532 is represented by the homology toda bracket @xmath533\times \cdots \times [ m_j ] , p , [ l]\rangle$ ] . since
this toda bracket is representable , it must be that @xmath534\times \cdots \times [ m_j]$ ] and @xmath535 $ ] have order @xmath24 in bordism , so that the homology toda bracket lifts to a bordism toda bracket , giving a class in @xmath536 .
indeed , the representable part of the tor term in homology is the group generated by the images of the bordism toda brackets @xmath537 , p^s , [ l]\rangle$ ] , and such a bracket maps to zero in homology whenever @xmath538 .
now consider the bordism toda bracket @xmath533\times \cdots \times [ m_j ] , p , [ l]\rangle$ ] .
we know at least one of the @xmath539 $ ] , say @xmath540 $ ] , has order @xmath24
. then @xmath541\times \cdots \times [ m_j ] , p , [ l]\rangle = [ m_1]\times \cdots \times [ m_{j-1}]\times \langle [ m_j ] , p , [ l]\rangle\ ] ] ( @xcite , 2.1 ) .
now apply proposition [ prop : ranktwo ] to @xmath542 , p , [ l]\rangle$ ] , and this completes the inductive step .
[ thm : pscfiniteabelian ] let @xmath21 be an abelian @xmath24-group , and let @xmath40 .
suppose @xmath39 is a closed manifold with fundamental group @xmath21 and non - spin universal cover , and suppose @xmath543\in h_n{^{\text{\textup{atoral}}}}(b\pi)$ ] . if @xmath363 , also assume @xmath1 is not orientable .
if @xmath24 is odd , also assume @xmath21 is elementary abelian
. then @xmath1 has a metric with positive scalar curvature .
in particular , if @xmath544 , then * every * @xmath134-manifold with fundamental group @xmath21 and with non - spin universal cover has a metric of positive scalar curvature .
first consider the non - orientable case with @xmath363 . in this case , since there are no tor terms , @xmath545 where the @xmath546 are the cyclic factors of @xmath21 .
now each class in @xmath547 may by lemma [ lem : cyclic ] be represented by a manifold with nonnegative yamabe invariant , or with positive scalar curvature if the class is in degree @xmath264 .
so it immediately follows that @xmath548 is represented by manifolds of positive scalar curvature . that takes care of the non - orientable case .
now consider the case where @xmath24 is odd ( and @xmath1 is orientable ) .
we apply theorem [ thm : rhstructure ] .
this reduces us to the case of classes of the form @xmath525 , with each @xmath528 represented by either @xmath407 or a lens space .
the product manifold has positive scalar curvature unless all the @xmath528 s are @xmath4-dimensional , in which case we have a toral class .
is theorem [ thm : oddspinabeliansylow ] true without the odd order assumption ?
we presume so , but the proof would necessarily be much more complicated , since computing @xmath549 for a @xmath12-group is quite difficult .
j. c. alexander , cobordism massey products , trans .
amer . math .
soc . * 166 * ( 1972 ) , 197214 .
m. t. anderson , scalar curvature , metric degenerations and the static vacuum einstein equations on @xmath136-manifolds , i , geom .
9 * ( 1999 ) , no . 5 , 855967 .
b. botvinnik and p. gilkey , the eta invariant and the gromov - lawson conjecture for elementary abelian groups of odd order , topology appl .
* 80 * ( 1997 ) , no .
12 , 4353 .
b. botvinnik , p. gilkey , and s. stolz , the gromov - lawson - rosenberg conjecture for groups with periodic cohomology , j. differential geom .
* 46 * ( 1997 ) , no . 3 , 374405 .
m. gromov and h. b. lawson , jr .
, spin and scalar curvature in the presence of a fundamental group , i , ann . of math .
* 111 * ( 1980 ) , 209230 .
m. gromov and h. b. lawson , jr . ,
the classification of simply connected manifolds of positive scalar curvature , ann . of math.*111 * ( 1980 ) , 423434 . m. gromov and h. b. lawson , jr .
, positive scalar curvature and the dirac operator on complete riemannian manifolds , publ .
, no . 58
( 1983 ) , 83196 . d. c. johnson and s. w. wilson , the brown - peterson homology of elementary @xmath24-groups , amer .
j. math .
* 107 * ( 1985 ) , no . 2 , 427453 .
s. kawsik and r. schultz , positive scalar curvature and periodic fundamental groups , comment .
helv . * 65 * ( 1990 ) , no . 2 , 271286 . j. l. kazdan and f. warner , existence and conformal deformation of metrics with prescribed gaussian and scalar curvatures , ann . of math .
( 2 ) * 101 * ( 1975 ) , 317331 . o. kobayashi , scalar curvature of a metric with unit volume , math . ann .
* 279 * ( 1987 ) , no . 2 , 253265 .
p. landweber , knneth formulas for bordism theories , trans .
* 121 * ( 1966 ) , no . 1 , 242256 .
c. lebrun , kodaira dimension and the yamabe problem , comm .
* 7 * ( 1999 ) , no . 1 , 133156 . j. lohkamp , the space of negative scalar curvature metrics , invent .
* 110 * ( 1992 ) , 403407 .
i. madsen and r. j. milgram , _ the classifying spaces for surgery and cobordism of manifolds _ , ann . of math .
studies , vol .
92 , princeton univ . press , princeton , n.j . , 1979 . j. petean , computations of the yamabe invariant , math .
lett . * 5 * ( 1998 ) , no . 6 , 703709 . j. petean , the yamabe invariant of simply connected manifolds , j. reine und angew .
* 523 * ( 2000 ) , 225231 .
j. petean and g. yun , surgery and the yamabe invariant , geom . and funct
. anal . *
9 * ( 1999 ) , 11891199 .
j. rosenberg , the @xmath550-assembly map and positive scalar curvature , in _ algebraic topology pozna 1989 _ , lecture notes in math .
, vol . 1474 , springer , berlin , 1991 , pp .
. j. rosenberg , reflections on c. t. c. wall s work on cobordism , in _ surveys on surgery theory : volume 2 _ , s. cappell , a. ranicki , and j. rosenberg , eds . ,
annals of math .
studies , vol .
149 , princeton univ . press , princeton , nj , 2001 , pp . 4961 . j. rosenberg and s. stolz , a `` stable '' version of the gromov - lawson conjecture , in _ the ech centennial : proc . conf .
on homotopy theory _ , m. cenkl and h. miller , eds .
, contemp .
* 181 * ( 1995 ) , amer .
soc . , providence , ri , pp .
j. rosenberg and s. stolz , metrics of positive scalar curvature and connections with surgery , in _
surveys on surgery theory : volume 2 _ , s. cappell , a. ranicki , and j. rosenberg , eds .
, annals of math .
studies , vol . 149 , princeton univ . press , princeton , nj , 2001 , pp . 353386
. r. schoen , conformal deformation of a riemannian metric to constant scalar curvature , j. differential geom .
* 20 * ( 1984 ) , no . 2 , 479495 .
r. schoen and s .- t .
yau , on the structure of manifolds with positive scalar curvature .
manuscripta math .
* 28 * ( 1979 ) , no . 1 - 3 , 159183 . r. schultz , positive scalar curvature and odd order abelian fundamental groups , proc .
* 125 * ( 1997 ) , no . 3 , 907915 . s. stolz , simply connected manifolds of positive scalar curvature , ann .
of math .
( 2 ) * 136 * ( 1992 ) , no . 3 , 511540 . s. stolz , positive scalar curvature metrics existence and classification questions , in _ proceedings of the international congress of mathematicians _ , vol . 1 ( zrich , 1994 ) ,
birkhuser , basel , 1995 , pp .
625636 .
r. stong , _ notes on cobordism theory _
, mathematical notes , princeton univ . press , princeton , nj , 1968 . l. r. taylor , @xmath12-local cobordism theories , j. london math .
( 2 ) * 14 * ( 1976 ) , no . 2 , 303308 . * author addresses : * + boris botvinnik + department of mathematics + university of oregon + eugene , or 97403 - 1222 + ` [email protected] ` + jonathan rosenberg + department of mathematics + university of maryland + college park , md 20742 - 4015 + ` [email protected] ` | the yamabe invariant is an invariant of a closed smooth manifold defined using conformal geometry and the scalar curvature .
recently , petean showed that the yamabe invariant is non - negative for all closed simply connected manifolds of dimension @xmath0 .
we extend this to show that yamabe invariant is non - negative for all closed manifolds of dimension @xmath0 with fundamental group of odd order having all sylow subgroups abelian .
the main new geometric input is a way of studying the yamabe invariant on toda brackets .
a similar method of proof shows that all closed manifolds of dimension @xmath0 with fundamental group of odd order having all sylow subgroups elementary abelian , with non - spin universal cover , admit metrics of positive scalar curvature , once one restricts to the `` complement '' of manifolds whose homology classes are `` toral . ''
the exceptional toral homology classes only exist in dimensions not exceeding the `` rank '' of the fundamental group , so this proves important cases of the gromov - lawson - rosenberg conjecture once the dimension is sufficiently large . |
since the seminal paper by morris and thorne @xcite traversable lorentzian wormholes have attracted great interest . as solutions of the equations of gravity , these objects describe geometries which have two regions , either in the same universe or in different universes , connected by a throat @xcite . in static spacetimes , it is possible to define the throat as a minimal area surface satisfying a flare - out condition @xcite . within the framework of general relativity
, traversable wormholes must be threaded by exotic matter that violates the null energy condition @xcite .
the amount of exotic matter required at the throat can be made infinitesimally small by appropriately choosing the geometry of the wormhole @xcite . although it may imply that the throat should suffer large stresses @xcite .
discussions about the energy conditions in the context of wormholes can be find in refs .
@xcite . thin
shell wormholes are mathematical constructions made by cutting and pasting two manifolds to form a geodesically complete one , with a shell located at the joining surface , which corresponds to the throat where the exotic matter is present @xcite .
the junction - condition formalism is used for their study @xcite .
the stability analysis , under perturbations that preserve the symmetries , has been widely performed for a variety of thin shell wormholes .
poisson and visser carried out a linear analysis of the stability of wormholes constructed by joining two equal schwarzschild geometries @xcite .
barcel and visser studied wormholes built using branes with negative tensions @xcite .
ishak and lake analyzed transparent spherically symmetric thin shells and wormholes @xcite .
eiroa and romero extended the linearized stability study to reissner - nordstrm wormhole spacetimes @xcite .
lobo and crawford did the same analysis with wormhole geometries with a cosmological constant @xcite .
shell wormholes were studied within the context of dilaton gravity in refs .
cylindrical thin
shell wormholes have been also analyzed in refs .
for other recent interesting articles see refs .
@xcite .
the requirement of matter that violates the energy conditions is not only a feature of wormholes .
if general relativity is assumed as the gravity theory to describe the large scale of the universe , the accelerated expansion violates the strong energy condition ( see refs .
@xcite for observational data analysis ) .
therefore several models of exotic matter , which have been proposed in the cosmological context @xcite , have been also used for wormholes .
phantom energy has been considered in a variety of wormhole articles .
a chaplygin gas @xcite was used as the exotic matter in wormhole spacetimes @xcite . in particular
, a generalized chaplygin gas within a finite region around the throat , has been adopted by lobo @xcite as the exotic matter supporting wormholes , and by eiroa @xcite for thin
shell wormholes , in which the exotic matter is located at the joining surface . in the present paper , we construct and analyze the stability under perturbations preserving the symmetry of thin
shell wormholes supported by a generalized chaplygin gas in dilaton gravity .
that is , wormholes constructed by cutting and pasting geometries associated to a charged black holes with dilaton and maxwell fields . in sec .
[ sec : tswh ] , we review the basic aspects of construction and stability of thin shell wormholes with a generalized chaplygin gas . in sec . [
sec : dil - wh ] , we perform the study of thin shell wormholes in dilaton gravity . in sec .
[ sec : conclu ] , the results are summarized .
units such that @xmath0 are adopted throughout this work .
in this section , we review the formalism developed in ref .
@xcite for a wide class of spherically symmetric thin shell wormholes with a generalized chaplygin gas .
we start from the general metric @xmath1 where @xmath2 is the radial coordinate , @xmath3 and @xmath4 are the angular coordinates , and the functions @xmath5 and @xmath6 are positive from a given radius . from this geometry , we take two identical copies of a region with radius @xmath7 assuming @xmath8 in order to avoid the presence of horizons ( at @xmath9 ) and/or singularities ( at @xmath10 ) : @xmath11 , and we paste them at the hypersurface @xmath12 , to create a new manifold @xmath13 .
if @xmath14 ( condition of flare - out ) , this construction creates a geodesically complete manifold representing a wormhole with two regions connected by a throat of radius @xmath15 , where the surface of minimal area is located .
in such a way that the area of a surface at fixed radius is @xmath16 , with @xmath6 an increasing function for @xmath17 , being @xmath18 .
thus the area has a minimum at @xmath19 and the manifold @xmath20 represents a wormhole with throat @xmath21 . on this manifold
it is possible to define a new radial coordinate @xmath22 ( the signs correspond , respectively , to @xmath23 and @xmath24 ) that represents the proper radial distance to the throat , which is situated at @xmath25 . on the wormhole throat , which is a synchronous timelike hypersurface ,
we define coordinates @xmath26 , with @xmath27 the proper time on the shell . for the stability analysis , under perturbations preserving the symmetry ,
we let the radius of the throat be a function of proper time @xmath28 .
outside the throat , the geometry remains static .
the lack of gravitational waves is guarantee by the birkhoff theorem ( see refs .
@xcite for the conditions that should be satisfied in a spherically symmetric geometry ) . following the darmois - israel formalism @xcite ,
the second fundamental forms ( or extrinsic curvature ) associated with the two sides of the shell are given by @xmath29 where @xmath30 are the unit normals ( @xmath31 ) to @xmath32 in @xmath20 : @xmath33 using the orthonormal basis @xmath34^{-1/2}e_{\theta } , e_{\hat{\varphi}}=[h(a)\sin^2 \theta ] ^{-1/2 } e_{\varphi } \ }
$ ] , we obtain for the geometry ( [ e1 ] ) that @xmath35 and @xmath36 where a prime and the dot represent , respectively , the derivatives with respect to @xmath37 and @xmath28 . defining @xmath38\equiv k_{_{\hat{\imath}\hat{\jmath}}}^{+}-k_{_{\hat{\imath}\hat{\jmath}}}^{-}$ ] , @xmath39=[k_{\ ; \hat{\imath}}^{\hat{\imath}}]$ ] and introducing the surface stress - energy tensor @xmath40 , where @xmath41 is the surface energy density and @xmath42 are the transverse pressures , the einstein equations on the shell ( the well - known lanczos equations ) are @xmath43+kg_{\hat{\imath}\hat{\jmath}}=8\pi s_{\hat{\imath}\hat{\jmath } } , \label{e8}\ ] ] so we get @xmath44 and @xmath45 . \label{e10}\ ] ] the negative sign in eq .
( [ e9 ] ) plus the flare - out condition ( @xmath14 ) implies that @xmath46 , indicating the presence of exotic matter at the throat .
we adopt a generalized chaplygin gas as the exotic matter in the shell @xmath32 . for this gas
, the pressure has opposite sign to the energy density , resulting in a positive pressure .
then , the equation of state for the exotic matter at the throat can be written in the form @xmath47 where @xmath48 and @xmath49 are constants .
the original chaplygin gas @xcite with equation of state @xmath50 is recovered for @xmath51 , and it has the property that the squared sound speed is always positive , even in the case of exotic matter .
then , by replacing eqs .
( [ e9 ] ) and ( [ e10 ] ) in eq .
( [ e11 ] ) , we obtain the differential equation that should be satisfied by the throat radius of thin
shell wormholes threaded by exotic matter with the equation of state of a generalized chaplygin gas : @xmath52h(a)+[f(a)+\dot{a}^2]h'(a ) \}[h'(a)]^{\alpha}-2a[4\pi h(a)]^{\alpha + 1}[f(a)+\dot{a}^2]^{(1-\alpha)/2}=0 .
\label{e12}\ ] ] for the static solutions with throat radius @xmath53 , the surface energy density and pressure are given by @xmath54 and @xmath55 .
\label{e14}\ ] ] the throat radius of the static solutions , @xmath53 , should fulfill eq .
( [ e12 ] ) , that is @xmath56[h'(a_{0})]^{\alpha}-2a[4\pi h(a_{0})]^{\alpha + 1}[f(a_{0})]^{(1-\alpha)/2}=0 , \label{e15}\ ] ] with the condition @xmath57max@xmath58 ( if the metric has an event horizon or a singularity ) . following the standard potential approach , we study the stability of the static solution , under perturbations that preserve the symmetry . from eqs .
( [ e9 ] ) and ( [ e10 ] ) ( and recalling that the area of the wormhole throat is @xmath59 ) , the energy conservation equation is found : @xmath60 ^ 2 -2h(a)h''(a)\right\rbrace \frac{\dot{a}\sqrt{f(a)+ \dot{a}^2}}{2h(a)}. \label{p1}\ ] ] the first term in the left hand side is the internal energy change of the throat and the second the work done by the internal forces of the throat , while the right hand side represents a flux .
if @xmath61 ^ 2 -2h(a)h''(a)=0 $ ] , the flux term is zero and eq .
( [ p1 ] ) takes the form of a simple conservation equation .
this occurs when @xmath62 ( with @xmath63 and @xmath64 constants ) or @xmath65 ( this case is unphysical , since there is no throat @xcite ) .
it is straightforward to see that eq .
( [ p1 ] ) can be written in the form @xmath66 ^ 2 -2h(a)h''(a)\right\rbrace \frac{\sigma } { 2 h'(a)}=0 , \label{p3}\ ] ] where it was used that @xmath67 . since the pressure @xmath68 is a function of @xmath69 given by the equation of state , eq .
( [ p3 ] ) is a first order differential that can be recast in the form @xmath70 , for which a unique solution with a given initial condition always exists , provided that @xmath71 has continuous partial derivatives .
then , eq . ( [ p3 ] ) can be formally integrated to obtain @xmath72 . if we replace @xmath72 in eq .
( [ e9 ] ) , the dynamics of the wormhole throat is completely determined by a single equation : @xmath73 with @xmath74 ^{2}. \label{p5}\ ] ] replacing the surface energy density and the pressure corresponding to the static solutions ( eqs .
( [ e13 ] ) and ( [ e14 ] ) ) and using that @xmath75 , a taylor expansion to second order of the potential @xmath76 around the static solution yields @xcite : @xmath77 , \label{p11}\ ] ] with @xmath78 ^ 2}{2f(a_{0})}+\left [ \frac{(1-\alpha ) h'(a_{0})}{2h(a_{0})}+\frac{\alpha h''(a_{0})}{h'(a_{0})}\right ] f'(a_{0 } ) \nonumber \\ & & + ( \alpha + 1)\left [ \frac{h''(a_{0})}{h(a_{0 } ) } -\left ( \frac{h'(a_{0})}{h(a_{0 } ) } \right ) ^{2}\right ] f(a_{0 } ) .
\label{p12}\end{aligned}\ ] ] as usual , if the dynamics of the system is completely determined by an equation like eq .
( [ p4 ] ) , the stability under radial perturbations is guaranteed if and only if @xmath79 .
the four - dimensional modified einstein action in dilaton gravity includes the ( scalar ) dilaton field @xmath80 and the electromagnetic field @xmath81 , and in the einstein frame has the form @xcite @xmath82 where @xmath83 is the ricci scalar of the spacetime .
the parameter @xmath84 represents the coupling between the dilaton and the maxwell field and it will be restricted within the range @xmath85 . for @xmath86 , the action corresponds to the usual einstein
maxwell scalar theory . for @xmath87 ,
the action was obtained in the context of low energy string theory with a maxwell field , but with all other gauge fields and the antisymmetric field set to zero . the action given by eq .
( [ emd ] ) leads to the einstein equations with the dilaton and the maxwell fields as the sources @xcite : @xmath88 @xmath89 @xmath90 assuming @xmath91 , which means that the dilaton field does not depend on time , these field equations admit spherically symmetric black hole solutions @xcite in the form of eq .
( [ e1 ] ) , with metric functions , in schwarzschild coordinates , given by @xmath92 where the constants @xmath93 are related with the parameter @xmath84 , and the mass and charge of the black hole by @xmath94 and @xmath95 in the case of electric charge , the electromagnetic field tensor has non - null components @xmath96 , and the dilaton field is given by @xmath97 , where the asymptotic value of the dilaton @xmath98 was taken as zero . for magnetic charge ,
the metric is the same , with the electromagnetic field @xmath99 and the dilaton field obtained replacing @xmath100 by @xmath101 .
we can distinguish three cases : i ) @xmath86 , ii ) @xmath102 , and iii ) @xmath87 .
in the first case , the geometry corresponds to reissner - nordstrm metric , with @xmath103 and @xmath104 .
if @xmath105 , the constants @xmath106 and @xmath107 represent , respectively , the radii of the outer ( event ) and the inner horizons ; if @xmath108 both horizons have the same radius ; and if @xmath109 there is only a naked singularity at @xmath110 . in the second case
, the metric has an event horizon for @xmath111 and a singular surface for @xmath112 , and there are two possibilities : if @xmath113 , then @xmath114 which implies that @xmath115 ; if @xmath116 , we have @xmath117 and the event horizon is located inside the singularity radius ( so called naked singularity ) .
it is worthy of mention that there is a maximum value for the charge : @xmath118 , otherwise the geometry is not well defined . in the third case , one obtains @xmath119 and @xmath120 $ ] .
the event horizon at @xmath121 is outside or inside the singularity at @xmath122 , accordingly to @xmath123 or @xmath124 , respectively .
reissner - nordstrm wormholes ( @xmath86 ) with a generalized chaplygin gas were previously studied by one of the authors @xcite . in the present work ,
we focus on dilaton wormholes ( @xmath125 ) .
as it was done in a previous work @xcite , we construct thin
shell wormholes in dilaton gravity , assuming that the throat has a radius @xmath15 greater than @xmath106 and @xmath107 to eliminate the presence of horizons and singularities . for the geometry represented by the metric
( [ e1 ] ) , the static solutions with throat radius @xmath53 are obtained replacing @xmath126 and @xmath127 in eqs .
( [ e13 ] ) and ( [ e14 ] ) , thus we reobtain the surface energy density and pressure found in ref .
@xcite @xmath128 , \label{d4}\ ] ] @xmath129 .
\label{d5}\ ] ] but now the pressure and the surface energy density are related by the generalized chaplygin gas equation of state , instead of the linearized equation of state adopted in ref .
it is straightforward to see that the surface energy density and pressure corresponding to the reissner
nordstrm thin
shell wormholes studied in ref .
@xcite are recovered when @xmath86 . supported by a generalized chaplygin gas with @xmath130 :
the dashed curves represent the unstable static solutions with throat radius @xmath53 for given parameters @xmath106 , @xmath131 and @xmath132 ; there are not stable static solutions .
the darker ( lighter ) gray zones are unphysical , corresponding to a throat radius smaller than the horizon ( singularity ) radius of the original manifold .
the event horizon and the naked singularity match for a critical value of charge @xmath133 and the geometry is not well defined for @xmath134.,width=566 ] supported by a generalized chaplygin gas with @xmath135 : the dashed curves represent
the unstable static solutions with throat radius @xmath53 for given parameters @xmath106 , @xmath131 and @xmath132 ; the solid curves represent the stable static solutions .
the meaning of the gray zones and the charges @xmath136 and @xmath137 are explained in fig .
[ f1].,width=566 ] supported by a generalized chaplygin gas with @xmath138 : the dashed curves represent the unstable static solutions with throat radius @xmath53 for given parameters @xmath106 , @xmath131 and @xmath132 ; the solid curve represent the stable static solutions .
the meaning of the gray zones and the charges @xmath136 and @xmath137 are explained in fig .
note that the horizontal scale has changed.,width=566 ] supported by a generalized chaplygin gas with @xmath130 : the dashed curves represent the unstable static solutions with throat radius @xmath53 for given parameters @xmath106 , @xmath131 and @xmath132 ; there are not stable static solutions .
the meaning of the gray zones and the charges @xmath136 and @xmath137 are explained in fig .
[ f1].,width=566 ] supported by a generalized chaplygin gas with @xmath135 : the dashed curves represent the unstable static solutions with throat radius @xmath53 for given parameters @xmath106 , @xmath131 and @xmath132 ; there are not stable static solutions .
the meaning of the gray zones and the charges @xmath136 and @xmath137 are explained in fig .
[ f1].,width=566 ] supported by a generalized chaplygin gas with @xmath138 : the dashed curves represent the unstable static solutions with throat radius @xmath53 for given parameters @xmath106 , @xmath131 and @xmath132 ; the solid curve represent the stable static solutions ( see the magnified region ) .
the meaning of the gray zones and the charges @xmath136 and @xmath137 are explained in fig .
note that the horizontal scale has changed .
, width=566 ] when @xmath139 , we assume that the throat radius satisfy that @xmath140 for @xmath141 , since the event horizon contain the singularity , and @xmath142 for @xmath143 , because of the event horizon is inside of the singular surface ( thus is a naked singularity ) ; we have labeled the critical value of charge , for which the event horizon and the singularity have the same radial coordinate , as @xmath133 . as we pointed out above
, the charge can not take values greater than what we have called @xmath144 as the maximum possible value of charge for which the geometry is well defined . in the particular case of @xmath87
there is no maximum value of charge since the geometry is always well defined in our wormhole construction .
the throat radii of the static solutions are obtained by replacing eqs .
( [ d1a ] ) and ( [ d1b ] ) in eq .
( [ e15 ] ) , and solving it numerically ; these solutions are stable when satisfy that the second derivative of the potential , given by eq .
( [ p12 ] ) , is positive . for
a given set of parameters , the complexity of the expression of @xmath145 compel us to find the range of @xmath146 for which @xmath79 in a numerical form and display the results graphically . in figs .
[ f1][f3 ] we show the results obtained for values of dilaton parameter @xmath147 which is close to the reissner - nordstrm case ( i.e. @xmath86 ) , and @xmath130 , @xmath135 , @xmath138 , respectively . in figs .
[ f4][f6 ] we display the results found for @xmath148 , corresponding to an intermediate value of dilaton parameter , and @xmath130 , @xmath135 , @xmath138 , respectively .
the values of charge adopted were chosen to be representative in each case . in all plots ,
the solid curves correspond to stable solutions while the dashed lines to unstable ones . as shown in fig .
[ f1 ] ( where @xmath147 and @xmath130 ) , for small values of @xmath149 there are not stable solutions .
this is a very important difference with the reissner - nordstrm wormholes , studied in ref .
we can summarize the results as follows : * when @xmath141 there is one unstable solution with a throat radius @xmath150 which decreases with @xmath151 and approaches to the event horizon radius of the original manifold for large values of @xmath152 . *
if @xmath153 there is one unstable solution for small values of @xmath151 , there are two unstable solutions for intermediate values of @xmath151 , and there are no solutions for large values of @xmath151 , because the throat radius @xmath150 reaches the singularity of the original manifold at a finite value of @xmath151 . * when @xmath154 there is one unstable solution for small values of @xmath151 , there are three unstable solutions for intermediate values of @xmath151 , and there is one unstable solution for large values of @xmath151 , with the throat radius @xmath150 approaching asymptotically to the singularity of the original manifold . as shown in fig .
[ f2 ] ( where @xmath147 and @xmath135 ) , for intermediate values of @xmath149
there are stable solutions for properly chosen values of the parameters .
the main results are : * for @xmath141 there is one unstable solution with a throat radius @xmath150 that asymptotically approaches to the radius of the event horizon of the original manifold for large values of @xmath151 , faster than in the case of the small values of @xmath149 .
* for @xmath155 there are three solutions at small values of @xmath151 : the larger and the smaller ones are unstable and the middle one is stable , while for large values of @xmath151 there is only one unstable solutions with a throat radius @xmath150 that tends asymptotically to the singularity of the original manifold .
note that all cases have solution for any value of @xmath151 . as shown in fig .
[ f3 ] ( where @xmath147 and @xmath138 ) , the case of the ordinary chaplygin gas presents stable solutions .
we find that : * when @xmath156 and @xmath157 is not very close to @xmath158 , for small values of @xmath151 there is only an unstable solution .
the throat radius @xmath150 decreases with @xmath151 and it cuts the horizon radius of the original manifold at a finite value of @xmath151 , so there are no solutions for intermediate and large values of @xmath151 . * if @xmath159 and @xmath157 is very close to @xmath158 , for small values of @xmath151 there are two solutions with throat radius @xmath150 : the larger one is unstable and the smaller one is stable .
the throat radius cuts the horizon radius of the original manifold at a finite and very close to zero value of @xmath151 , so there are no solutions for intermediate and large values of @xmath151 . *
if @xmath160 , for small values of @xmath151 there are three solutions with throat radius @xmath150 : the larger and the smaller ones are unstable , and the intermediate one is stable . for intermediate and large values of @xmath151
there are no solutions since the throat radius reaches the singularity of the original manifold at very small values of @xmath161 .
* when @xmath162 , for small values of @xmath151 there are three solutions with throat radius @xmath150 : the larger and the smaller are unstable , the middle one is stable ; while for intermediate and large values of @xmath151 there is only one unstable solution that asymptotically approaches to the singularity of the original manifold . in fig .
[ f4 ] ( where @xmath148 and @xmath130 ) , the solutions are unstable for all values of charge .
the throat radius is a decreasing function of @xmath151 that presents asymptotic behavior for large values of @xmath151 .
sufficiently closer to the critical value of charge , there are three unstable solutions for a very small interval of @xmath151 . in fig .
[ f5 ] ( where @xmath148 and @xmath135 ) , for all values of charge the solutions are unstable .
the throat radius approaches asymptotically to the event horizon of the original manifold or to the singularity of the original manifold ( depending on the value of charge ) , in a faster way than the previous case , with a small value of @xmath149 .
sufficiently closer to the critical value of charge , there are three unstable solutions for a narrow interval of @xmath151 .
for both , @xmath130 and @xmath135 , there are solutions for any value of @xmath151 . in fig .
[ f6 ] ( where @xmath148 and @xmath138 ) there is one stable configuration , and the main results are the following : * when @xmath156 and @xmath157 is not very close to @xmath158 , for small values of @xmath151 there is only an unstable solution .
the throat radius @xmath150 decreases with @xmath151 and it cuts the horizon radius of the original manifold at a finite value of @xmath151 , so there are no solutions for intermediate and large values of @xmath151 . * if @xmath159 and @xmath157 is very close to @xmath158 , but not too much close , for small values of @xmath151 there are two solutions with throat radius @xmath150 : the larger one is unstable and the smaller one is stable .
the throat radius reaches the horizon radius of the original manifold for a finite and very close to zero value of @xmath151 , so there are no solutions for intermediate and large values of @xmath151 .
note that the horizontal scale was reduced in the plots in order to obtain a better visualization of the little region of the stability .
* when @xmath159 and @xmath157 is near enough to @xmath158 there are two unstable solutions for small values of @xmath151 .
the throat radius cuts the horizon radius of the original manifold at a finite and very close to zero value of @xmath151 , so there are no solutions for intermediate and large values of @xmath151 . *
if @xmath155 , there is only one unstable solution with a throat radius that tends asymptotically to the singularity of the original manifold .
when @xmath148 and @xmath163 , we have found that there are not stable solutions . only the case @xmath164 , that is the ordinary chaplygin gas , presents stable configurations . in all cases analyzed , when there are both stable and unstable static configurations , the throat radius @xmath150 of the stable solution is an increasing function of @xmath151 , while the unstable branch decreases with @xmath151 . as the value of the dilaton parameter increases , we find that the interval of @xmath150 for which @xmath79 reduces .
we observe that there are no roots of @xmath145 for @xmath165 when the dilaton parameter is such that @xmath166 . so that the second derivative remains negative for all possible values of @xmath152 , which means that the dilaton wormholes have no stable static solutions for @xmath165 if @xmath167 . in the specific case of @xmath51 , which corresponds to the ordinary chaplygin gas
, we obtain a very small stable region for some values of the charge close enough to @xmath168 ( but not too much close ) .
the interval of @xmath150 for which there are stable static solutions is smaller as far as the value of the dilaton coupling parameter increases . in particular , for the uppermost value @xmath87 , we encountered a very small interval , for a very narrow range of values of charge , where the wormholes have stable behavior ( the plots are not shown here for brevity ) .
in this work , we have studied the stability under perturbations preserving the spherical symmetry of thin shell wormholes in dilaton gravity supported by a generalized chaplygin gas . for the construction
we have performed the usual cut and paste procedure ; so the surface energy density and the pressure at the throat have been obtained following the junction formalism . for the stability analysis
we have implemented the standard potential approach .
we have found that for a fixed value of the dilaton coupling parameter @xmath84 , the possibility of obtaining a static stable solution increases as far as the exponent @xmath149 in the equation of state grows , i.e. as the generalized chaplygin gas gets closer to the ordinary one ( with @xmath164 ) .
the same result was previously found for reissner
nordstrm thin shell wormholes ( which correspond to @xmath86 ) in ref . @xcite . on the other hand , as the dilaton coupling @xmath84 increases ,
the values of the parameters for which stable static solutions exist lie in a smaller range , so the dilaton field reduces the presence of stable configurations .
this work has been supported by conicet and universidad de buenos aires . c. barcel and m. visser , int
d * 11 * , 1553 ( 2002 ) ; t.a .
roman , proceedings of the tenth marcel grossmann , rio de janeiro , brazil , 20 - 26 july 2003 , edited by m. novello , s.p .
bergliaffa , and r. ruffini ( world scientific publishing , singapore , 2006 ) [ arxiv : gr - qc/0409090 ] .
n. sen , ann .
( leipzig ) * 73 * , 365 ( 1924 ) ; k. lanczos , ann .
( leipzig ) * 74 * , 518 ( 1924 ) ; g. darmois , mmorial des sciences mathmatiques , fascicule xxv , chap . v ( gauthier - villars , paris , 1927 ) ; w. israel , nuovo cimento * 44b * , 1 ( 1966 ) ; * 48b * , 463(e ) ( 1967 ) ; for a review see p. musgrave and k. lake , class .
quantum grav .
* 13 * , 1885 ( 1996 ) .
lobo and p. crawford , class .
quantum grav .
* 22 * , 4869 ( 2005 ) ; j.p.s .
lemos and f. s. n. lobo , phys .
d * 78 * , 044030 ( 2008 ) ; k.a .
bronnikov and a.a .
starobinsky , mod .
a * 24 * , 1559 ( 2009 ) ; k.a . bronnikov and j.p.s .
lemos , phys .
d * 79 * , 104019 ( 2009 ) ; g.a.s .
dias and j.p.s .
lemos , phys .
d * 82 * , 084023 ( 2010 ) ; e.f .
eiroa and c. simeone , phys .
d * 83 * , 104009 ( 2011 ) .
a. kamenshchik , u. moschella and v. pasquier , phys . lett . *
b511 * , 265 ( 2001 ) ; n. bili , g.b . tupper and r.d .
viollier , phys . lett . *
b535 * , 17 ( 2002 ) ; m.c .
bento , o. bertolami and a.a .
sen , phys .
d * 66 * , 043507 ( 2002 ) ; for a review see v. gorini , a. kamenshchik , u. moschella and v. pasquier , gr - qc/0403062 .
f. rahaman , m. kalam and k. a. rahman , mod .
a * 23 * , 1199 ( 2008 ) ; v. gorini , u. moschella , a.yu .
kamenshchik , v. pasquier , and a.a .
starobinsky , phys .
d * 78 * , 064064 ( 2008 ) ; s. chakraborty and t. bandyopadhyay , int . j. mod
d * 18 * , 463 ( 2009 ) ; m. jamil , m.u .
farooq and m.a .
rashid , eur .
j. c * 59 * , 907 ( 2009 ) ; p.k.f .
kuhfittig , gen .
. grav . * 41 * , 1485 ( 2009 ) . | in this article , we construct spherical thin shell wormholes with charge in dilaton gravity .
the exotic matter required for the construction is provided by a generalized chaplygin gas .
we study the stability under perturbations preserving the symmetry .
we find that the increase of the coupling between the dilaton and the electromagnetic fields reduces the range of the parameters for which stable configurations are possible .
+ pacs number(s ) : 04.20.gz , 04.40.nr , 98.80.jk + keywords : lorentzian wormholes ; stability ; generalized chaplygin gas |
there have been many development in statistics in which geometry , represented by an extension from euclidean to more general spaces , has proved fundamental .
thus , reproducing kernel hilbert spaces is a part of gaussian process methods , sobolev spaces are used for , and besov spaces , for wavelets .
differential geometry , riemannian manifolds , curvature and geodesics are at the core of information geometry , shape analysis and manifold learning .
this paper is concerned with cat(0 ) , and , more generally , cat(k ) .
although related to riemannian manifolds ( @xmath0 denotes the riemannian curvature ) these metric spaces have a less rigid structure and are qualitatively different from the spaces mentioned above .
trees are among the first examples in which cat(0 ) properties have been used in statistics and bioinformatics @xcite : the set of trees ( the tree space ) becomes a geodesic metric space ( to be defined ) when for each tree , the edge lengths are allocated to entries in a vector of sufficient dimension to capture all the tree structures .
then , the metric is the euclidean geodesic distance on a simplicial complex called a fan " which is induced by the tree structure : the geodesics are the shortest piecewise linear paths in the fan .
such spaces have already been shown to be cat(0 ) @xcite by gromov s theory @xcite . for a random variable @xmath1 on a metric space
@xmath2 endowed with a metric @xmath3 the general _ intrinsic mean _ is defined by @xmath4 the empirical intrinsic mean based on data @xmath5 , sometimes called the frchet mean , is defined as @xmath6 where @xmath7 for euclidean space , @xmath8 , the sample mean . in general , @xmath9 is not necessarily convex and the means , @xmath10 , are not unique .
figure [ fig : hyperboloid ] shows that the curvature can affect the property of @xmath9 . in particular , for so - called cat(0 ) spaces , which ( trivially ) include euclidean spaces , the intrinsic means @xmath10 are unique . for data on
( a ) a hyperboloid ( curvature @xmath11 ) , ( b ) a plane ( @xmath12 ) and ( c ) a sphere ( @xmath13 ) .
the bluer represents the smaller value of @xmath9 . only for the sphere
, @xmath9 has multiple minima .
one hundred data points are sampled independently from a `` gaussian mixture , '' whose density is proportional to @xmath14 where @xmath15 and @xmath16.,title="fig:",width=166 ] + ( a ) hyperboloid for data on ( a ) a hyperboloid ( curvature @xmath11 ) , ( b ) a plane ( @xmath12 ) and ( c ) a sphere ( @xmath13 ) .
the bluer represents the smaller value of @xmath9 . only for the sphere
, @xmath9 has multiple minima .
one hundred data points are sampled independently from a `` gaussian mixture , '' whose density is proportional to @xmath14 where @xmath15 and @xmath16.,title="fig:",width=166 ] + ( b ) plane for data on ( a ) a hyperboloid ( curvature @xmath11 ) , ( b ) a plane ( @xmath12 ) and ( c ) a sphere ( @xmath13 ) . the bluer represents the smaller value of @xmath9 . only for the sphere
, @xmath9 has multiple minima .
one hundred data points are sampled independently from a `` gaussian mixture , '' whose density is proportional to @xmath14 where @xmath15 and @xmath16.,title="fig:",width=166 ] + ( c ) sphere even when the mean is not unique , the function @xmath9 can yield useful information , for example about clustering .
we can also define second - order quantities : @xmath17 and @xmath18 a key concept in the study of these issues is that the metrics are global geodesic metrics , that is metrics based on the shortest path between points measured by integration along a path with respect to a local metric .
the interplay between the global and the local will concern us to a considerable extent .
for some intuition , consider a circle with arc length as distance and the uniform measure .
the frchet intrinsic mean is the whole circle .
we can say , informally , that the non - uniqueness arises because the space is not cat(0 ) .
although the curvature is zero in the riemannian geometry sense , there are other considerations related to `` diameter '' , which prevents it from having the cat(0 ) property @xcite .
if we use the embedding euclidean distance rather than the arc distance , we have the _ extrinsic mean _ , but again , we obtain the whole perimeter . see @xcite for further discussion on the intrinsic and extrinsic means .
the empirical case is also problematic .
for example , two data points on a diameter , with arc length as distance , give intrinsic means on the perpendicular diameter .
simple geometry shows that the empirical extrinsic mean is the whole circle .
we also cover the more general cat(@xmath0 ) spaces giving some new results related to diameter " as well as conditions for the uniqueness of intrinsic means not requiring the spaces to be cat(0 ) .
the general form of @xmath9 depends , here , on three parameters , @xmath19 and it can be written in compact form : @xmath20 where the function @xmath21 and the construction of @xmath22 are given below .
once we have introduced this new class of metrics , variety of statistics can be generalised ; intrinsic mean , variance , clustering ( based on local minima of @xmath9 ) . for classification problems , we can select an appropriate metric by cross - validation .
theoretically , we have the opportunity to apply well studies areas of geometry compared with methods based on selection of a good `` loss function . '' in table [ table : generalised_statistics ] , we summarise such generalised statistics . ' '' '' + metrics & @xmath23 & @xmath24 ' '' '' + intrinsic mean & @xmath25 & @xmath26 ' '' '' + variance & @xmath27 & @xmath28 ' '' '' + clustering function & @xmath29 & @xmath30 ' '' '' + there are many ways to transform one metric into another , regardless of whether they are geodesic metrics .
a straightforward way is to use a concave function @xmath31 such that given a metric @xmath32 , the new metric is @xmath33 .
this is plausible if we use non - convex @xmath34 , which are useful , as will be explained , in clustering and classification .
such concave maps are often interpreted as loss functions , but we will consider them in terms of a change of metric which may lead to selection using geometric concepts .
this is particularly true for the construction of @xmath21 in this paper .
the basic definition and construction from a geodesic metric space to the special geodesics based on accumulation of density are given in the next section , together with the definition of a cat(0 ) space .
in section 3 , we first show that means and medians in simple one - dimensional statistics can be placed into our framework . because geodesics themselves are one - dimensional paths , this should provide some essential motivation .
the @xmath35-metric is obtained by a local dilation .
our computational shortcut is to use _ empirical _ graphs , whose vertices are data points .
we will need , therefore , to define empirical geodesics .
we start with a natural geodesic defined via a probability density function in which the distance along a path is the amount of density accumulated " along that path .
then , an empirical version is defined whenever a density is estimated .
we might have based the paper on kernel density estimates ; instead , we have adopted a very geometric construction based on objects such as the the delaunay complex and its skeleton graph . in section 4 , the @xmath36 metric is introduced .
it is based on a function derived from a geodesic metric via shrinking , pointwise , to an abstract origin ; that is to say an abstract cone is attached . the smaller the value of @xmath37 , the closer to the origin ( apex ) .
the geometry of the two dimensional case drives our understanding of the general case because of the one - dimensional nature of geodesics .
we prove that , for finite @xmath37 , the embedding cone is cat(@xmath0 ) with smaller @xmath0 than the original space .
we cover the more general cat(@xmath0 ) spaces , giving some new results related to diameter " , in section 5 , including conditions for the uniqueness of intrinsic means not requiring the spaces to be cat(0 ) .
section 6 provides a summary of the effect of changing @xmath38 and @xmath37 . after some discussion on the selection of @xmath38 and @xmath37 in section 7 , section 8 covers some examples .
the fundamental object in this paper is a geodesic metric space .
this is defined in two stages .
first , define a metric space @xmath39 with base space @xmath1 and metric @xmath40 .
sometimes , @xmath41 will be a euclidean space @xmath42 of dimension @xmath43 , containing the data points , but it may also be some special object such as a graph or manifold .
second , define the length of a ( rectilinear ) path between two points @xmath44 and the geodesic connecting @xmath45 and @xmath46 as the shortest such path .
this minimal length defines a metric @xmath47 , and the space endowed with the geodesic metric is called the geodesic metric space , @xmath48 .
the interplay between @xmath49 and @xmath50 will be critical for this paper , and , as mentioned , we will have a number of ways of constructing @xmath51
. for data points @xmath52 in @xmath1 , the empirical intrinsic ( frchet ) mean is @xmath53 there are occasions when it is useful to represent @xmath54 as a sub - manifold of a larger space ( such as euclidean space ) @xmath55 with its own metric @xmath56 .
we can then talk about the extrinsic mean : @xmath57 typically , the intrinsic mean is used as an alternative when the geodesic distance , @xmath51 is hard to compute .
the difficultly in considering the intrinsic mean in @xmath58 is that it may not lie in the original base space @xmath1 .
this leads to a third possibility , which is to project it back to @xmath1 , in some way , as an approximation to the intrinsic mean @xmath59 ( which may be hard to compute ) .
we will discuss this again in section 4 .
cat(0 ) spaces , which correspond to non - positive curvature riemannian spaces , are important here because their intrinsic means are unique .
the cat(0 ) property is as follows .
take any three points @xmath60 in a geodesic metric space @xmath1 and consider the geodesic triangle " of the points based on the geodesic segments connecting them .
construct a triangle in euclidean 2-space with vertices @xmath61 , called the comparison triangle , whose euclidean distances , @xmath62 , are the same as the corresponding geodesic distances just described : @xmath63 , etc . on the geodesic triangle
select a point @xmath45 on the geodesic edge between @xmath64 and @xmath65 and find the point @xmath46 on the edge @xmath66 of the euclidean triangle such that @xmath67 .
then the cat(0 ) condition is that for all @xmath68 and all choices of @xmath45 : @xmath69 for a cat(0 ) space ( i ) there is a unique geodesic between any two points , ( ii ) the space is contractible , in the topological sense , to a point and ( iii ) the intrinsic mean in terms of the geodesic distance is unique .
cat(@xmath0 ) spaces , a generalization of cat(0 ) spaces , are explained in section [ sec : cat_k ] .
let @xmath1 be a @xmath43-dimensional euclidean random variable absolutely continuous with respect to the lebesgue measure , with density @xmath70 .
let @xmath71\}$ ] be a parametrised integrable path between two points @xmath72 in @xmath73 , which is rectifiable with respect to the lebesgue measure .
let @xmath74 with appropriate modification in the non - differentiable case , be the local element of length along @xmath75 .
the weighted distance along @xmath75 is @xmath76 the geodesic distance is @xmath77 here we consider a random variable on euclidean space but this can be generalized for riemannian manifolds and even for singular spaces with a density with respect to a base measure naturally defined by the metric . from the geodesic distances on distributions we shall follow three main directions : 1 .
transform the geodesic metrics in various ways with parameters @xmath78 to obtain a wide class of metrics , 2 .
discover ( locally ) cat(0 ) and cat(@xmath0 ) spaces for certain ranges of the parameters , 3 .
apply empirical versions of the metrics based on an empirical graph whose nodes are the data points .
there is an important distinction between global transformations applied to the whole distance between points and local transformations applied to dilate the distance element .
the general @xmath35 metric is a dilation of the original distance @xmath43 and what we have referred to as a local metric .
it is obtained by transforming the density in ( [ eq : geodesic_dist ] ) .
thus for @xmath71\}$ ] between @xmath79 and @xmath80 , @xmath81 and @xmath82 changing @xmath38 essentially changes the local curvature . roughly speaking ,
when @xmath38 is more negative ( positive ) , the curvature is more negative ( positive ) . in the next subsection
, we look at the one - dimensional case .
although this case is elementary , good intuition is obtained by rewriting the standard version in terms of a geodesic metric .
assume that @xmath1 is a continuous univariate random variable with probability density function @xmath70 and cumulative distribution function ( cdf ) @xmath83 .
the mean @xmath84 achieves @xmath85 .
here we are using the euclidean distance : @xmath86 .
the median is defined by @xmath87 . on a geometric basis
, we can say that @xmath88 achieves @xmath89 , where we use a metric that measures the amount of probability between @xmath45 and @xmath90 : @xmath91 carrying out the calculations : @xmath92 which achieves a minimum of @xmath93 at @xmath94 , as expected .
now let us consider the sample version .
let @xmath95 be the order statistics , which we assume are distinct .
one of the first exercises in statistics is to show that @xmath96 minimises @xmath97 , with respect to @xmath98 .
for the median , first consider using the first of the two approaches with the empirical cdf @xmath99 .
we obtain various definitions depending on our definition of @xmath100 and @xmath101 , or just using convention . using the metric approach
the natural metric is to take @xmath102 with the standard definition of @xmath103 .
applied to distinct data points @xmath104 this is equal to @xmath105 .
for an arbitrary @xmath98 @xmath106 where @xmath107 .
then , @xmath108 is achieved at @xmath109 when @xmath110 is odd and at @xmath111 when @xmath110 is even .
another approach for the median would be to take a piecewise linear approximation to @xmath100 which is equivalent to having a density @xmath103 that is proportional to @xmath112 in the interval @xmath113 .
then , the metric is @xmath114 and @xmath115 is achieved at @xmath109 when @xmath110 is odd and at @xmath116 , when @xmath110 is even .
we can think of this last result in another way .
consider the points @xmath117 as points in @xmath118 $ ] , take the empirical mean of the points and transform them back with the cdf corresponding to @xmath103 , namely , the piecewise linear approximation .
the idea of weighting intervals should provide intuition when we extend the intervals to edges on a graph , because edges are one - dimensional .
there are a number of options to produce an empirical version of the @xmath22 metric , based on the data .
one such option would be to produce a smooth empirical density @xmath119 followed by numerical integration and optimization to compute the geodesics .
we prefer a much simpler method based on a graph constructed from the data .
all geodesic computation is then restricted to the graph .
we list some candidates ( see appendix [ graphs ] for a description of the second and third candidates listed below ) : 1 . the complete graph with vertices at the data points and all edges , 2 . the edge graph ( 1-skeleton ) of the delaunay simplicial complex with vertices at the data points , 3 .
the gabriel graph with vertices at the data points .
the discussion below applies to the complete graph or any connected sub - graph . for any such graph , define a version of the @xmath22 distance just for edges , @xmath120 where @xmath121 is the euclidean distance from @xmath122 to @xmath123 .
this can be explained by making a transformation @xmath124 we refer to this as _
edge regularization_. we then apply @xmath38 in the usual way to obtain @xmath125 the new length " of each edge @xmath126 is obtained by integrating this density " along the edge . in this sense
, @xmath127 also plays the role of density estimation .
although we need a regularization @xmath128 with respect to the dimension @xmath129 for density estimation @xcite , we manage the regularization by rescaling the parameter @xmath38 .
note that @xmath130 gives the unit length and @xmath131 restores the original length .
now we consider only the set of edges @xmath132 of the graph @xmath133 as a metric space with the metric defined by the geodesic : @xmath134 where the infimum is taken over all ( connected ) paths @xmath75 between @xmath135 and @xmath136 .
note that even if @xmath128 can estimate the local density well , it does not follow that the metric @xmath35 can be approximate by the metric @xmath137 since edge lengths @xmath127 on each path @xmath75 are not independent .
it is suggested that further theoretical work is necessary .
here we will admit @xmath137 as an approximation of @xmath35 .
if the graph is not a complete euclidean graph with weights equal to the euclidean lengths of the edges , some edges may not be in any edge geodesics between any pair of vertices .
for an edge - weighted graph @xmath138 with weights @xmath139 on the graph , @xmath140 , which is the union of all the edge geodesics between all pairs of vertices , is called the geodesic sub - graph ( or geodesic graph ) of @xmath138 .
we will see how the geodesic sub - graphs transform as the value of @xmath38 changes .
we make an important general position assumption that the set of values @xmath141 are distinct , that is there are no ties .
this is an additional general position assumption to that given for the delaunay cells ( see appendix [ graphs ] ) .
we order the values using only a single suffix for simplicity : @xmath142 where @xmath143 . for @xmath144 ,
this induces the @xmath145 values : @xmath146 now , consider the geodesics as @xmath147 .
recall that a circuit in a graph is a connected path that begins and ends in some vertex and an elementary circuit is a circuit that visits a vertex no more than once .
consider an edge @xmath148 that has the following property which we call @xmath149 : it is in an elementary circuit @xmath150 of the graph in which all other edges have smaller values of @xmath127 namely @xmath151 then , the path @xmath152 ( within the circuit ) from @xmath122 to @xmath123 not containing the edge @xmath153 has length smaller than @xmath154 when @xmath38 is sufficiently negative : @xmath155 from this argument , we see that for sufficiently large @xmath156 as @xmath38 approaches @xmath157 , every edge having property q is removed from the geodesic sub - graph , and we obtain a tree .
let us summarize this algorithm , which applies to a general edge - weighted graph with distinct edges .
we refer to this algorithm as the _ backwards _ algorithm .
it clearly gives a tree .
let @xmath158 and label the edges @xmath159 in increasing order of their weights .
starting with edge @xmath160 , remove @xmath160 if it is in a cycle otherwise continue to @xmath161 .
( general step ) continue downwards at each stage removing an edge if it is in a cycle of the remaining subgraph .
4 . stop if no more edges can be removed using step 3 .
there is a natural _ forwards _ algorithm that also yields a tree as follows . 1 .
let @xmath158 and label the edges @xmath159 in increasing order of their weights .
starting with @xmath162 , add an edge if adding it does not create a cycle .
3 . ( general step ) continue adding an edge at each step provided that the addition does not create a cycle . 4
. stop if no more edges can be added .
we have the following . given a connected edge - weighted graph @xmath133 , the backward and forward algorithms yield the same tree , which we call @xmath163 .
let @xmath164 and @xmath165 be the trees generated by the backward and forward algorithms respectively .
any edge of @xmath138 not in @xmath164 can not be in @xmath165 , because it is in a circuit of edges with lower weights than itself , by the forward construction .
this is sufficient because each tree has the same number of edges .
in fact , the tree can be constructed by the simple rule : remove all edges that are in a circuit with smaller " edges ; the order in which the edges are removed is irrelevant .
although there are several implementations of forward and backward minimal spanning tree algorithms , their importance here is to give intuition for the construction of geodesics as @xmath166 .
we now show that the tree @xmath163 is the minimal spanning tree in a strong sense .
[ tree1 ] let @xmath133 be a graph with @xmath167 and distinct edge weights @xmath168 .
there is a unique spanning tree @xmath163 whose ordered weights @xmath169 have the following minimal property . if @xmath170 are the ordered weights of any other spanning tree , then @xmath171 with strict inequality for at least one @xmath172 .
moreover @xmath163 is given by the forward ( or backward ) algorithm .
proof . the proof is obtained by contradiction .
let @xmath173 be the tree constructed by the forward algorithm .
let @xmath174 be another spanning tree with ordered weights @xmath170 and suppose that for some @xmath175 with @xmath176 , we have @xmath177 .
let the edges of @xmath178 and @xmath174 be @xmath179 and @xmath180 , respectively . since @xmath181 , we must have @xmath182 . in the sequence @xmath159 ,
let @xmath183 and @xmath184 ; then , we must have @xmath185 . by the nature of the forward algorithm ,
new vertices are used only when an edge is included by the algorithm : all unused edges form circuits and therefore use edges of the tree constructed up to that point .
thus , the subgraph of @xmath133 with edges @xmath186 has exactly @xmath187 vertices ; this value is attained on the addition of the last edge @xmath188 .
however , @xmath174 , which is a subgraph of @xmath133 , attains @xmath187 vertices at @xmath189 with @xmath185 which is a contradiction .
the ordering property in theorem [ tree1 ] is known from the theory of stochastic ordering : if the empirical cdfs of the distances @xmath190 and @xmath191 are @xmath140 and @xmath192 , respectively , then @xmath193 with strict inequality for at least one value ( in fact , over at least one non - empty interval ) .
thus , not only @xmath194 but also @xmath195 for any non - decreasing function @xmath31 . for sufficiently negative @xmath38 ,
the tree @xmath196 itself , that is the tree as a metric space with metric @xmath22 , is a cat(0 ) space .
we need to extend the metric somewhat so that it applies to the edges , in addition to the nodes . thus , for any two points @xmath197 on the tree ,
define @xmath198 where the integral is taken along the ( unique ) path @xmath199 on the tree and @xmath200 when line element @xmath201 is in edge @xmath202 in @xmath199 .
[ cat1 ] there is an @xmath203 such that for any @xmath204 , the geodesic sub - graph becomes the minimal spanning tree @xmath163 endowed with the @xmath22 metric and , therefore , becomes a cat(0 ) space .
we see that for sufficiently negative @xmath38 , _ every _ geodesic defined with the @xmath22 metric lies in the tree @xmath173 .
in fact , although we started with a general connected graph , any graph for which the edges can be mapped into a euclidean interval gives a cat(0 ) tree using this construction .
there are some well - known algorithms and considerable literature on minimal spanning trees .
remarkably , the minimal spanning tree for a complete euclidean graph is the same as the minimal spanning tree for the delaunay graph and is therefore a subgraph of the delaunay graph .
[ tree2 ] let @xmath138 be a complete edge - weighted graph whose weights are distinct and monotonically related to those of the complete euclidean graph @xmath205 of a point set @xmath1 in @xmath43 dimensions .
then , @xmath138 , @xmath205 , the gabriel graph and the delaunay graph of the point set @xmath1 all have the same and unique minimal spanning tree . versions of this result are known in two dimensions @xcite , but the authors had some difficulty in finding a concise proof in the literature .
we present a proof in appendix [ proof - mst ] .
thus , any algorithm for finding the minimal spanning tree of a graph will give the minimal spanning tree of the delaunay graph , without having to find the full delaunay complex for any euclidean space into which the points can be embedded .
further , as seen previously , the minimal spanning tree has the strong minimal property in the sense of theorem [ tree2 ] .
we may study the geometry as @xmath38 increases away from @xmath157 .
following theorem [ tree1 ] , we are interested in two cases : the euclidean delaunay graph and the complete euclidean graph . in both cases , we consider the @xmath22 metric .
[ thm : alpha - chain ] let @xmath206 be an edge - weighted graph with distinct weights @xmath207 and let @xmath208 be its geodesic subgraph ; then , @xmath209 proof .
this follows from the consideration of geodesics .
an edge @xmath153 in @xmath138 is not in @xmath208 if it is not a geodesic . in this case
, there is an alternative path @xmath75 from @xmath210 to @xmath175 such that @xmath211 .
however , this inequality is preserved if @xmath38 is decreased , so that @xmath212 is increased .
thus an edge absent from @xmath208 is absent from @xmath213 .
+ let @xmath214 be a geodesic disk of radius @xmath215 centred at @xmath216 .
define the maximum radius @xmath217 of the disk centred at @xmath45 as being cat(@xmath0 ) , that is @xmath218 see section [ sec : cat_k ] for the definition of cat(@xmath0 ) .
if @xmath1 is a metric graph , @xmath219 does not include a cycle whose length is less than @xmath220 consider a rescaling of @xmath1 such that the shortest ( longest ) edge length is 1 , and denote it as @xmath221 for @xmath222 ( @xmath223 ) .
if @xmath224 , @xmath225 because the @xmath38-chain is increasing for @xmath226 , each cycle in @xmath227 is removed one by one as @xmath38 decreases .
furthermore , each cycle length increases as @xmath38 decreases because , by the rescaling , every edge length is greater than 1 and it increases as @xmath38 decreases .
this gives the decreasing property of @xmath228 for @xmath222 .
we can prove the result for @xmath229 similarly . by the theorem
, @xmath227 becomes `` more cat(@xmath0 ) '' for a smaller @xmath226 . because rescaling of the graph does not affect the uniqueness of the intrinsic mean , @xmath208 tends to have a unique mean for a smaller @xmath226 .
figures [ fig : gaussian_alpha ] ( a)-(f ) are geodesic graphs with different values of @xmath38 for 50 samples of the standard 2-d normal distribution .
we give two cases in which we decrease @xmath38 : the delaunay graph in figure [ fig : gaussian_alpha ] and the complete graph in figure [ fig : gaussian_alpha_complete ] . by the time @xmath230 the cases are indistinguishable and have the same minimal spanning geodesic graph for large negative values of @xmath38 , as expected .
this is predictable from theorem [ tree2 ] and [ thm : alpha - chain ] and gives an important practical strategy : when the dimension is high and @xmath38 is small , use the complete graph rather than the delaunay graph because the former requires computational cost only proportional to @xmath43 , whereas the computational cost of the latter is @xmath231 @xcite @xcite . for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the delaunay graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( a ) @xmath233 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the delaunay graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( b ) @xmath131 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the delaunay graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( c ) @xmath234 + for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the delaunay graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( d ) @xmath235 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the delaunay graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( e ) @xmath236 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the delaunay graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( f ) @xmath237 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the complete graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( a ) @xmath233 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the complete graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( b ) @xmath131 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the complete graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( c ) @xmath234 + for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the complete graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( d ) @xmath235 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the complete graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( e ) @xmath236 for 50 samples from the standard 2-d normal distribution .
the initial graph ( @xmath131 ) is the complete graph .
the value @xmath232 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) , and the minimum is represented by a square.,title="fig:",width=170 ] + ( f ) @xmath237
the cat(0 ) property of euclidean space implies that we do not obtain multiple local minima of @xmath239 even for multi - modal distributions .
however , an appropriate concave transformation of the metric can modify the base data space making it less cat(0 ) .
we introduce the @xmath238 metric via a transformation @xmath240 as a candidate . for any geodesic metric space @xmath241 with metric @xmath242 and a parameter @xmath243
, we can define the metric @xmath244 where @xmath245 for a finite positive @xmath37.,width=264 ] since @xmath240 is a concave function on @xmath246 ( see figure [ fig : graph_of_g_beta ] ) , @xmath238 becomes a metric but not necessarily a geodesic metric .
we can express this conveniently as @xmath247 , where @xmath248 is the heaviside function , @xmath249 it is easiest to consider the case that @xmath242 is the euclidean distance on the real line . as @xmath250 for small values of @xmath43 ,
the metric behaves like @xmath251 , and as @xmath252 , @xmath43 behaves like @xmath253 rescaled to @xmath254 $ ] . for euclidean distances greater than @xmath37 , @xmath36 returns a constant distance of unity .
the metric has the effect of downsizing large distances to unity . because , as will soon be seen , @xmath238 can be recognized as a geodesic metric of a cone embedding @xmath1 , we refer to the mean @xmath255 as the _ @xmath37-extrinsic mean_. we consider a scheme in which the real line , or part of it , is mapped into the unit circle , where it can be represented by an angle @xmath256 . in the unit disk , @xmath257 ,
a point is represented by polar coordinates : @xmath258 .
we consider the top half of the unit disk , for which @xmath259 , namely @xmath260 we will now give a rule for travelling from a point @xmath261 in @xmath257 to the point p:@xmath262 . 1 .
if @xmath263 , travel in a straight line to p. this is the geodesic for the euclidean metric 2 . if @xmath264 , travel first in straight line to the origin @xmath265 and then in a straight line from @xmath265 to p. we now apply a similar rule to points @xmath266 and @xmath267 . 1
if @xmath268 , we take the euclidean distance @xmath269 2 . if @xmath270 , we take @xmath271 , indicating a route to the origin and out to the other point .
now , consider two points at @xmath256 and @xmath272 on the circumference of the circle ; then , in case ( d1 ) above , the distance between them is @xmath273 when @xmath274 , we obtain the value 2 . now , keeping @xmath275 fixed , let us rescale the distance inside the argument .
in terms of the two points just described , we place the first point at p so that @xmath276 and scale the second with a factor @xmath277 .
then , without changing the notation , and removing the factor of 2 , we have @xmath278 .
then , @xmath279 implies that we may use the euclidean distance for a wider interval .
case ( d2 ) above corresponds to the distance @xmath280 before rescaling and the new version @xmath281 achieves the value 1 .
this corresponds precisely to the @xmath36 metric of this section .
the extrinsic mean keeps the mean within the original data space but uses the @xmath238 metric .
we will see below that the @xmath238 metric involves , even in the general case , an extension of the original space by a single dimension .
this distinguishes it from the embeddings for extrinsic means used in the literature , which involve high - dimensional euclidean space , for example the case of the tree space .
controlling @xmath37 , as will be seen below , controls the value of @xmath0 when the embedding space is considered as a cat(@xmath0 ) space .
we have an indirect link between clustering and cat(@xmath0 ) spaces . as @xmath37 decreases while the embedding space becomes more cat(0 ) ( @xmath0 decreasing ) the original space becomes less cat(0 ) .
this demonstrates , we believe , the importance of the cat(@xmath0 ) property in geodesic - based clustering . in euclidean space ,
the standard euclidean distance dose not exhibit multiple local means because the space is trivially cat(0 ) .
however , by using the @xmath36-metric with a sufficiently small @xmath37 , the space can have multiple local means , as shown in figure [ fig:1dim_beta ] . against @xmath98 for different values of @xmath37.,title="fig:",width=207 ] + ( a ) density function against @xmath98 for different values of @xmath37.,title="fig:",width=207 ] + ( b ) @xmath282 + against @xmath98 for different values of @xmath37.,title="fig:",width=207 ] + ( c ) @xmath283 against @xmath98 for different values of @xmath37.,title="fig:",width=207 ] + ( d )
@xmath284 figure [ fig:1dim_beta_trajectory ] shows a plot of the local minima of @xmath285 against @xmath37 for the same samples as those in figure [ fig:1dim_beta ] . against @xmath37.,width=264 ]
the above construction is a special case of a general construction that applies to any geodesic metric space and hence to those in this paper .
let @xmath286 be a geodesic metric space with a metric @xmath287 .
a metric cone @xmath288 with @xmath289 is a cone @xmath290 \slash \mathcal{x } \times \{0\}$ ] with a metric @xmath291 for any @xmath292 .
the intuitive explanation is as follows .
see figure [ fig : metric - cone - explanation ] .
let @xmath293 be a subset @xmath294 with the geodesic metric on @xmath293 .
thus , @xmath293 and @xmath286 are the same as a set but endowing different metrics .
then @xmath293 is a rescaling of the metric on @xmath286 by @xmath37 . for any @xmath292 , their projections
@xmath295 give two points @xmath296 , respectively .
for a geodesic @xmath297 between @xmath45 and @xmath298 , consider a cone @xmath299\}$ ] spanned by @xmath300 .
this cone can be isometrically embedded into an `` extended unit circular sector '' , i.e. a covering @xmath301,\theta \in ( -\infty , \infty)\}/ \{(0,\theta)\mid \theta \in ( -\infty , \infty)\}$ ] of the unit disk corresponding to @xmath302 $ ] .
then @xmath303 and @xmath304 are also mapped into the extended unit circular sector ; the distance @xmath305 for @xmath283 corresponds to the case ( d2 ) of a disk if we set @xmath306 and @xmath307 .
this corresponds to the length of the blue line path in figure [ fig : metric - cone - explanation ] ( b1 ) and ( b2 ) .
for further details on metric cones , refer to @xcite .
.,title="fig:",height=102 ] .,title="fig:",height=94 ]
the following result indicates that the metric cone space preserves the cat(0 ) property of the original space and the smaller values of @xmath37 continue this process .
see section [ sec : versus ] for more details .
[ cone ] 1 .
if @xmath286 is a cat(0 ) space , the metric cone @xmath288 is also cat(0 ) for every @xmath289 .
if @xmath308 is cat(0 ) , @xmath309 is also cat(0 ) for @xmath310 .
if @xmath286 is cat(@xmath0 ) for @xmath311 , @xmath288 becomes cat(0 ) for @xmath312 . the proof is given in appendix [ proof - cone ] .
a cat(@xmath0 ) space for @xmath313 is a geodesic metric space satisfying the following cat(@xmath0 ) condition .
take any geodesic triangle @xmath314 whose perimeter is less than @xmath315 and select any point @xmath45 on a geodesic @xmath316 .
let @xmath317 be a comparison triangle , which has the same edge length as @xmath314 , on a surface with a constant curvature @xmath0 , i.e. a sphere with radius @xmath318 for @xmath319 , a plane for @xmath320 and a hyperbolic space for @xmath321 . set @xmath46 on @xmath322 satisfying @xmath323 .
then the cat(@xmath0 ) condition is that for all @xmath68 and all choices of @xmath45 , @xmath324 .
thus every cat(@xmath0 ) space is a cat(@xmath325 ) space for @xmath326 .
every metric graph is cat(@xmath0 ) for @xmath319 if and only if there is no loop shorter than @xmath327 because every metric tree is cat(0 ) and therefore cat(@xmath0 ) for @xmath319 .
let @xmath286 be a geodesic metric space and fix it throughout this section .
the diameter of a subset @xmath328 is defined as the length of the longest geodesic in @xmath329 .
we define classes @xmath330 , @xmath331 and @xmath332 as follows .
1 . @xmath330 : the class of subsets @xmath328 such that the geodesic distance function @xmath333 is strictly convex on @xmath329 for each @xmath334 .
here , `` convex '' means geodesic convex , i.e. a function @xmath239 on @xmath286 is convex iff for every geodesic @xmath335 on @xmath286 , @xmath336 is convex with respect to @xmath337 .
2 . @xmath331 for @xmath338 $ ] : the class of the subsets @xmath328 such that for any probability measure whose support is in @xmath329 and non - empty , the intrinsic @xmath339-mean exists uniquely .
we refer to @xmath340 as @xmath341 .
@xmath332 : the class of subsets @xmath328 such that for every pair @xmath342 , the geodesic between @xmath129 and @xmath343 is unique .
[ lem : k1 - 1 ] @xmath344 for any @xmath338 $ ] .
if @xmath345 , @xmath346 is a strictly convex function on @xmath329 for each @xmath347 ; hence , @xmath348 is strictly convex for any probability measure @xmath59 whose support is in @xmath329 and non - empty .
thus , @xmath349 .
next , assume that @xmath350 and @xmath351 ; then , there are at least two different geodesics , @xmath352 and @xmath353 , between @xmath45 and @xmath298 .
thus , there are two points @xmath46 and @xmath354 in @xmath355 such that there is no intersection of @xmath352 and @xmath353 between @xmath46 and @xmath354 . then , the mid points of @xmath46 and @xmath354 on each geodesic become intrinsic @xmath339-means of the measure with two equal point masses on @xmath46 and @xmath354 .
this implies that @xmath356 .
let @xmath357 and @xmath358 be the largest values ( including @xmath359 ) such that every subset whose diameter is less than the value belongs to @xmath360 and @xmath332 , respectively .
then , evidently from lemma [ lem : k1 - 1 ] , @xmath361 for @xmath362 .
note that if @xmath286 is cat(0 ) , @xmath363 .
in general , the following theorem holds .
[ diam1 ] [ thm : diameter ] * if @xmath286 is cat(@xmath0 ) , @xmath364 . *
if @xmath286 is cat(@xmath0 ) , @xmath365 .
* if @xmath286 is a surface with a constant curvature @xmath319 , @xmath366 . the proof is given in appendix [ proof - diameter ] . by theorem [ thm : diameter](1 ) , @xmath367
thus , a lower curvature @xmath0 gives a wider area where the intrinsic @xmath339-mean is unique . according to theorem [ thm : diameter](3 ) , this lower bound for @xmath368 is the best universal upper bound for any @xmath286 with cat(@xmath0 ) property . for @xmath369
, @xmath370 is bounded above by @xmath371 where @xmath372 is an increasing function of @xmath300 , as proved in appendix [ proof - gamma ] .
this upper bound shows that the parameter @xmath300 plays a role in controlling the uniqueness of the mean , but it does not in euclidean space , where the @xmath339-mean functions are always convex .
combining the three deformations by @xmath38 , @xmath37 and @xmath300 , a novel frchet mean is proposed : @xmath374 where @xmath375 we now review three main areas : ( i ) cat(0 ) and cat(k ) aspects , ( ii ) uniqueness of means and ( iii ) robustness against outlying data values . in this discussion
we will concentrate on cases that arise from the euclidean graph , considered as a basic metric space from which we can construct a geodesic metric space using@xmath22 , or as the geodesic space for the metric cone extension using the @xmath36 approach .
note that this is a restriction of a very general approach that would start with a density @xmath239 or a smooth empirical density @xmath103 ; first , construct the geodesic metric @xmath22 , and then , possibly apply the @xmath37-cone method to that geodesic metric space .
however we shall prefer the graph approach because it leads to tractable mathematics and much faster computation ; moreover , we are able to capture some of the structure of the data using the complete euclidean graph , the delaunay graph and the gabriel graph .
as mentioned above , we restrict ourselves to the case where the base space is the euclidean graph , i.e. @xmath376 .
it holds that as @xmath377 , all geodesics end up on the minimal spanning tree that is cat(0 ) .
as @xmath38 increases at some point , it ceases to be cat(0 ) but will be cat(k ) for some @xmath0 , which depends on @xmath38 and indeed increases with @xmath38 . it should be stressed that the theorems on @xmath37 are completely general .
that is they cover arbitrary metric cones based on a geodesic metric space .
if we start with the euclidean graph as our geodesic space , it may not be cat(0 ) , but it can be shown that it is a cat(k ) space for some @xmath0 and will eventually be cat(0 ) for sufficiently small @xmath37 .
while the space is cat(0 ) the intrinsic mean is unique and lies on the minimal spanning tree . for larger values of @xmath38
, there may not be a unique intrinsic mean lying on the geodesic graph , but theorem [ thm : diameter ] suggests that the mean may be unique for values of the diameter that are sufficiently small relative to @xmath378 . for @xmath379 , at @xmath131 and @xmath380 , we obtain the euclidean intrinsic mean and the median , respectively , and these , again , may not be unique . as noted above , the metric cone embedding the original data space becomes cat(0 ) for sufficiently small @xmath37 sufficiently small .
the following is a possible strategy : starting with the euclidean graph , apply the cone method decreasing @xmath37 until the intrinsic mean is unique ( this may happen at a larger value of @xmath37 than that required to be cat(0 ) ) .
then , project that unique mean back onto the original graph ; the projection is unique by construction provided that the intrinsic mean is not at the apex of the cone .
instead of judging this method now we will wait for some empirical studies . in the @xmath37 , instead of
the projection of the intrinsic mean in the cone back to the original geodesic space , we prefer the extrinsic mean in which the mean lies in the geodesic space but , but we use the @xmath37 cone metric @xmath238 .
as we decrease @xmath37 we tend to get more local minima .
this is the antithesis of obtaining a cat(0 ) space .
because the @xmath37 metric becomes more concave as @xmath37 decreases for very small @xmath37 , each data point will yield a single `` local mean '' .
for reasonable values of @xmath37 , each data cluster will have a local mean because the function @xmath240 is locally linear and intra - cluster distances are small .
however , when the intra - cluster distances are larger , the concavity dominates , and we obtain multiple local minima .
both larger values of @xmath38 and smaller values of @xmath37 make the function @xmath383 less convex and can lead to multiplicity of the local means .
figures [ fig : alpha - beta - difference ] shows the difference between the effects caused by positive @xmath38 and finite @xmath37 . compared with the original geodesic distance ( figure [ fig : alpha - beta - difference ] ( a ) ) , both ( b ) and ( c ) have multiple local minima of the function @xmath239 . in ( b ) , the three small groups of samples in the top right corner are dealt with one cluster and have a unique local minimum in their midst .
in ( c ) , each of the three small bunches has a local minimum .
however , ( c ) tends to have local minima on a ridge of each cluster . by tuning both @xmath38 and @xmath37 as in ( d ) , we can get local minima at the centre of each cluster . for 24 samples each from normal distributions with the s.d .
@xmath384 and 8 samples each from normal distributions with the s.d .
initial graph : delaunay .
the value for each @xmath98 is represented by the colour : red ( small ) , blue ( large ) , and minimum by a square.,title="fig:",width=207 ] + ( a ) @xmath386 for 24 samples each from normal distributions with the s.d . @xmath384 and 8 samples each from normal distributions with the s.d .
initial graph : delaunay .
the value for each @xmath98 is represented by the colour : red ( small ) , blue ( large ) , and minimum by a square.,title="fig:",width=207 ] + ( b ) @xmath387 + for 24 samples each from normal distributions with the s.d . @xmath384 and 8 samples each from normal distributions with the s.d .
initial graph : delaunay .
the value for each @xmath98 is represented by the colour : red ( small ) , blue ( large ) , and minimum by a square.,title="fig:",width=207 ] + ( c ) @xmath388 for 24 samples each from normal distributions with the s.d . @xmath384 and 8 samples each from normal distributions with the s.d .
initial graph : delaunay .
the value for each @xmath98 is represented by the colour : red ( small ) , blue ( large ) , and minimum by a square.,title="fig:",width=207 ] + ( d ) @xmath389 at the beginning of section [ onedim ] , we showed how the univariate median can be considered as an intrinsic mean with respect to a geodesic metric , in both the population case and the empirical case . because the univariate median is well known to be robust against outliers , we can expect means based on empirical geodesics to also be robust . in particular , the flattening of large distances when @xmath37 is small gives an indication as to where to look for robust means .
this flattening idea is not new .
huber @xcite introduced a function similar to our @xmath21 function , which is quadratic in a spherical neighbourhood of zero and flat outside .
if the generalised mean is computed as a frchet mean , we should take the square root of the huber loss function .
various generalisations , including a smoothed version of huber loss function , have been proposed ; see , for example , @xcite .
figure [ fig : robustness ] shows a geodesic graph and the value of @xmath390 for 19 samples from n(0,1 ) and one outlier at @xmath391 .
figure [ fig : robustness ] ( a ) uses ordinary geodesic distance and the sample closest to the outlier sample attains the minimum of @xmath239 .
all of @xmath233 for ( b ) , @xmath283 for ( c ) and @xmath392 for ( d ) attain the minimum of @xmath239 around the mean of the normal distribution and we can see their robustness .
it is interesting that @xmath38 , @xmath37 and @xmath300 can control the trade - off between uniqueness of the mean and robustness of the estimation . in euclidean space
this trade - off is typically studied via the convexities of the loss function , whereas we are making the perhaps novel suggestion that the curvature is an appropriate vehicle . for 19 samples from n(0,1 ) and one outlier at @xmath391 ( far above the figure region ) .
the value @xmath70 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) and the minimum is represented by a square.,title="fig:",width=207 ] ( a ) @xmath393 for 19 samples from n(0,1 ) and one outlier at @xmath391 ( far above the figure region ) .
the value @xmath70 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) and the minimum is represented by a square.,title="fig:",width=207 ] ( b ) @xmath394 + for 19 samples from n(0,1 ) and one outlier at @xmath391 ( far above the figure region ) .
the value @xmath70 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) and the minimum is represented by a square.,title="fig:",width=207 ] ( c ) @xmath395 for 19 samples from n(0,1 ) and one outlier at @xmath391 ( far above the figure region ) .
the value @xmath70 for each sample point @xmath45 is represented by the colours red ( small ) and blue ( large ) and the minimum is represented by a square.,title="fig:",width=207 ] ( d ) @xmath396
in this section , we suggest how to select @xmath38 and @xmath37 empirically from the data .
first , assume that we have euclidean data ( equivalnet to @xmath397 and recall the basic effect of decreasing @xmath38 from @xmath398 to @xmath157 . at @xmath399
, we make no change to the metric .
as @xmath38 decreases , we lose edges from the geodesic graph .
that is to say from time to time , an edge that is in a particular geodesic is discarded and every geodesic that passes through that edge then has to use an alternative route .
let us assume that at @xmath38 ( and under mild extra conditions ) , only a single edge @xmath400 is removed and let @xmath401 be its length .
let @xmath402 be the lengths of the edges on the new geodesic that will replace the removed edge .
in addition , let there be @xmath403 distinct geodesics that use @xmath400 .
it is straightforward to see that _ all _ geodesics that use @xmath400 will use the new arc for an interval @xmath404 , for sufficiently small @xmath405 .
the total change in geodesic length is @xmath406 and it is continuous at the current @xmath38 but the first derivative changes : @xmath407 is typically not zero . to see this ,
take the case where all the @xmath408 are equal .
then , the change in the first derivative is @xmath409 in graph theory , the number of geodesics using a particular edge , @xmath403 in our case , is sometimes called the _ edge betweenness_.
we might therefore refer to the term @xmath410 as the weighted betweenness .
this quantity measures changes in the configuration : if @xmath403 and @xmath411 are large then a long edge with large betweenness is removed , and it is replaced by @xmath0 shorter edges from the current geodesic graph . if @xmath412 is the betweenness of an edge @xmath202 , the total betweenness of a graph @xmath133 is the sum of all the individual edge betweennesses , @xmath413 and the weighted version is @xmath414 which except for a scalar factor is the @xmath415 variance given by @xmath416 , in this paper .
we shall in fact favour the use of @xmath417 ( @xmath418 , and with the above discussion in mind , we will see in examples 1 and 2 that plots of the second derivative of @xmath419 do indeed have pronounced peaks and there is some matching of the @xmath38-values at the peaks with the analogous differential of the aggregate betweenness .
section 4.1 and figure 4 are important for understanding the @xmath37 metric .
we can summarise the material in a way that will indicate how to estimate @xmath37 .
the first point is that @xmath37 provides a metric cone . in one dimension
, we wrap the real line around a circle and attach the origin .
then , the the metric cone is based on the euclidean metric _ inside the cone_. the enlarged space ( referred to as the embedding space ) is cat(0 ) with respect to this metric .
we claim that this construction is fundamental because even in larger spaces , the geodesics are one - dimensional .
every geodesic , in some sense , has its private geodesic , but with a common vertex .
moreover , by theorem [ cone ] , if the base space is cat(0 ) , the embedding space is cat(0 ) , and in both cases , we have a unique intrinsic mean and our statistics are well defined .
howeve , r if we compute the intrinsic mean restricted to the base space , e.g. euclidean space , then the uniqueness no longer holds .
as stated above , the space may not be cat(0 ) for small @xmath37 but may become more so for large @xmath37 .
we can use this to our advantage : for sufficiently large @xmath37 , we expect a single maximum @xmath420 but multiple maxima for smaller @xmath37 , as shown in figure [ fig:1dim_beta ] .
if we recall that the value of the function @xmath421 for a given @xmath37 is helpful in clustering , we can suggest a number of plots to show the local minima .
however , we can say more .
first , note that in one dimension , @xmath422 over @xmath423 $ ] is a smooth kernel with bandwidth @xmath37 .
thus , with @xmath424 , we see that @xmath425 is a smooth density .
this interpretation helps to intuitively choose @xmath37 : a `` typical value '' of @xmath426 , e.g. the average of @xmath427 , by analogy of selection of bandwidth for kernel functions .
another option is to use cycle lengths in the geodesic graph . as can be seen from theorem [ cat1 ] and the proof of theorem [ cone ] ,
if we set @xmath428 for a cycle @xmath300 and its length @xmath429 , then the metric cone generated by the cycle becomes cat(0 ) .
the use of @xmath37 is shown in examples 1 and 3 .
in this section , we apply the @xmath430 metric to real data . because the @xmath431 loss function is more familiar than deformation of metrics by @xmath38 and @xmath37 , we will set @xmath432 and focus on @xmath38 and @xmath37 throughout the section . for the @xmath35 metric ( for @xmath433 , @xmath432 )
, we briefly describe the computation . for each fixed @xmath38 and each pair of points
initially every @xmath434 is computed , giving a complete graph . on this graph
the @xmath122 to @xmath123 geodesic is computed for all @xmath435 . the geodesic graph , for this @xmath38 ,
is then computed as the union of all such geodesics .
the present version of the software computes the geodesic graph for a grid of around 100 points , depending on the range of @xmath38 . as mentioned
, we are interested , here , only in the range @xmath436 $ ] and typically consider the range @xmath437 $ ] , where @xmath438 is a small positive integer . for each @xmath38 , we compute @xmath439 and @xmath440 .
we apply the @xmath441 metric to the k - nearest neighbour ( k - nn ) method , one of the most popular classification methods .
we will see that if we can choose adequate values for @xmath38 and @xmath37 , the classification error can be reduced .
we use five data sets from he uci machine learning repository @xcite : ( i ) fisher s iris data set ( number of instances @xmath442 , number of attributes @xmath443 , number of clusters @xmath444 ) , ( ii ) wine data set ( @xmath445 , @xmath446 , @xmath444 ) , ( iii ) ionosphere data set ( @xmath447 , @xmath448 , @xmath449 ) , with only real attributes , ( iv ) breast cancer wisconsin ( diagnostic ) data set ( @xmath450 , @xmath451 , @xmath449 ) , and ( v ) yeast data set ( @xmath452 , @xmath453 , @xmath454 ) . the average @xmath455 norm of each data set is normalized to be one .
the euclidean complete graphs are used as the initial metric graphs ( @xmath131 ) , and classification is performed using the weighted k - nn method ( k=10 ) with a common weighting @xmath456 where @xmath43 is the distance to the neighbour data point but using @xmath441 for various values of @xmath457 and @xmath458 .
a half of the samples is selected at random as a training set and the rest half is used as a testing set to evaluate the classification result .
we repeat it 1000 times and estimate the error rate .
.classification by k - nn method [ cols= " < , > , > , > , > " , ] figure [ fig : sushi_dendrograms ] shows the trees ( dendrograms ) for men and women : the first two are the intrinsic means and the next two are centres selected by the minima of @xmath459 , for the selected values of @xmath38 and @xmath37 .
figure [ fig : sushi_f ] was constructed by first finding the two local minima of @xmath459 and then plotting the values of @xmath459 along a geodesic of maximal length passing through the two minima .
it transpires that there are only six trees on this geodesic ( including the two special points ) but we are able to interpolate between the points by extending the range of the argument @xmath98 to include points in the edge , considered as a subset of the metric graph .
.,title="fig:",width=207 ] + ( a ) men .,title="fig:",width=207 ] + ( b ) women .,title="fig:",width=207 ] + ( c ) cluster 0 .,title="fig:",width=207 ] + ( d ) cluster 1 along a geodesic passing its two local minima ( @xmath460 ) . the value at each sample point on the geodesic is marked as a circle .
the two red circles correspond to the local minima.,height=226 ]
the @xmath38 metric is a deformation of the starting geodesic , and as @xmath38 becomes more negative , the geodesic graph , namely the union of all the geodesics , becomes sparser , and in our graph representation , it becomes a tree .
the space is cat(k ) with smaller @xmath0 and finally becomes a tree , at which point the space becomes cat(0 ) .
it is quite difficult to see the tree computation because of the numerous short edges , but for moderate values of @xmath38 such as @xmath461 , the structure is tree - like .
abrupt changes in various statistics as @xmath38 changes reveal topological changes in the structure of the geodesic graph , a fact that can be used to tune @xmath38 .
the @xmath37 metric is non - geodesic " because although the function @xmath21 operates on a geodesic , that does not mean that the space is a geodesic space in the formal sense .
however the cone construction yields a geodesic metric space , which is cat(@xmath0 ) with a lower value of @xmath0 than the original space , and indeed may be cat(0 ) .
if the @xmath37 metric is projected back to the original space , that space can have a non - convex @xmath239-function with larger @xmath0 .
this is useful for finding clusters because of multiple minima of the @xmath239-function , which is itself similar to a kernel .
the means obtained by the @xmath37-metric may represent the first study of an extrinsic mean via embedding in non - euclidean spaces and the first application of metric cones to statistics and data analysis .
we believe that the curvature of the data space underlying this work demands further investigation whereby connections should be established with recent developments related to empirical geodesic graphs , for example in manifold learning
. one important direction should be the effect of the curvature of the space on the trade - off between the uniqueness of the frecht means and the robustness of estimation .
to this end , @xmath78 and @xmath300 can be seen as parameters that can be tuned to change the curvature and hence study the trade - off .
the delaunay complex is usually defined via the voronoi diagram .
let @xmath462 be distinct points in @xmath73 .
define for @xmath463 the @xmath464 hyperplane @xmath465 perpendicular to and bisecting the line from @xmath122 to @xmath123 .
again for @xmath466 let @xmath467 be the half space bounded by @xmath465 and containing @xmath122 .
then , the voronoi cell of @xmath122 is @xmath468 the elementary definition is that @xmath469 is all those no further from @xmath122 than any other @xmath123 .
the delaunay complex is a simplicial complex with vertices at @xmath122 and one - dimensional edges between @xmath122 and @xmath123 if and only if @xmath467 is a face of @xmath469 ( and @xmath470 ) .
the vertices @xmath122 and the edge set then define the delaunay graph : the graph of the complex .
note that in the text we shall assume a general position when no more than @xmath471 voronoi cells can have a common point , and this guarantees that the delaunay complex has maximum simplexes of dimension @xmath43 and @xmath471 vertices a well - known property of the voronoi construction , in the general position case , is that the circumscribed sphere of a delaunay simplex contains only the @xmath471 vertices of the simplex , and in fact this is a defining property of the maximal simplexes .
the gabriel graph is defined in an intuitively similar way : the edge from @xmath122 to @xmath123 is included if and only if the unique sphere whose diameter is that edge contains no third point @xmath472 .
it is straightforward to show that the gabriel graph is a subgraph of the delaunay graph .
see @xcite @xcite @xcite for theory , algorithms and applications of delaunay and gabriel graphs .
the delaunay complex has come to prominence in the areas of discrete tube theory and persistent homology @xcite .
\(1 ) denote the mapped points of @xmath68 and @xmath45 by the projection @xmath473 as @xmath474 and @xmath1 , respectively , as shown in figure [ fig : metric_cones ] ( left ) .
denote the origin of the metric cone as @xmath475 . if the sum of the lengths of the geodesics @xmath476 , @xmath477 and @xmath478 in @xmath293 exceeds @xmath479 , it is easy to see that the cone spanned by @xmath480 becomes cat(0 ) and @xmath314 satisfies the cat(0 ) property . therefore , assume that @xmath481 .
next , let @xmath317 be a comparison triangle of @xmath314 and let @xmath46 be a point on a geodesic @xmath485 such that @xmath486 .
thus , @xmath487 .
arrange the points @xmath488 , @xmath489 and @xmath490 in a three - dimensional euclidean space with origin @xmath491 such that the lengths of @xmath492 , @xmath493 and @xmath494 are equal to the lengths of @xmath495 , @xmath496 and @xmath497 , respectively .
denote the radial projection of @xmath488 , @xmath489 , @xmath490 and @xmath46 to a unit sphere as @xmath498 , @xmath499 , @xmath500 and @xmath501 , respectively , as shown in figure [ fig : metric_cones ] ( right ) . by the definition of a metric cone , @xmath502 and the geodesics @xmath503 , @xmath504 , @xmath505 and @xmath506 in the unit sphere are arcs satisfying @xmath507 , @xmath508 , @xmath509 and @xmath510 .
the unit sphere has a positive constant curvature and @xmath293 is cat(0 ) , @xmath512 .
however , since @xmath513 and @xmath502 , @xmath512 implies that @xmath514 by the property of a metric cone .
thus , @xmath314 has cat(0 ) property and ( 1 ) of the theorem is proved .
\(2 ) assume that @xmath515 and a metric cone @xmath309 is not cat(0 ) for proving the latter half of the theorem by contradiction .
then , there is a geodesic triangle @xmath516 in @xmath309 and a point @xmath136 on the geodesic @xmath517 such that the geodesic @xmath518 is longer than the corresponding geodesic of a comparison triangle . by defining @xmath519 and @xmath520 as above
, we can say that @xmath521 .
next , each of @xmath522 and @xmath523 corresponds to a point in @xmath524 and we can consider the corresponding points @xmath525 and @xmath526 in the other metric cone @xmath308 .
when restricted to @xmath524 , a geodesic @xmath527 is just a rescaling of @xmath528 and @xmath529 .
now , @xmath530 is a geodesic triangle on the unit sphere , but after rescaling by @xmath531 , we can get a geodesic triangle @xmath532 on a sphere of radius @xmath531 whose edges have the same length as @xmath533 . by a known result on spherical triangles with the same edge lengths on different spheres ,
a larger radius implies a `` thinner '' triangle and @xmath534 where @xmath535 is a point on the geodesic @xmath536 such that @xmath537 .
combining all the arguments gives @xmath538 select a non - degenerate geodesic triangle in @xmath308 by selecting arbitrary points @xmath539 and @xmath540 on the geodesics @xmath541 , @xmath542 and @xmath543 in @xmath308 , respectively , and let @xmath544 be the intersection point of @xmath545 and @xmath546 . then , by @xmath547 , we can say that @xmath548 .
this implies that @xmath308 is not cat(0 ) and ( 2 ) of the theorem is proved .
\(3 ) for @xmath320 , the statement holds by ( 1 ) . for @xmath319 and @xmath549 ,
it is sufficient to prove for @xmath550 by ( 2 ) .
let @xmath314 be a geodesic triangle in @xmath288 and let @xmath551 be a geodesic triangle in @xmath293 .
let @xmath474 be the projection of @xmath68 , respectively .
if the perimeter of @xmath551 is longer than or equal to @xmath552 , the cone spanned by the perimeter becomes cat(0 ) by the same argument as that for ( 1 ) .
therefore , @xmath314 is cat(0 ) and satisfies the cat(0 ) property . if the perimeter of @xmath551 is smaller than @xmath552 , since @xmath286 is cat(@xmath0 ) and @xmath293 is cat(1 ) , for any @xmath553 , @xmath554 is shorter than the corresponding great arc @xmath555 of a comparison triangle @xmath556 , which is a spherical triangle on the unit sphere . since a comparison triangle @xmath317 of @xmath314
can be embedded on the cone spanned by @xmath556 , @xmath557 is shorter than the corresponding line segment @xmath558 .
this means the @xmath314 satisfies the cat(0 ) property .
the proof is by contradiction .
we need to show that if we take an edge of the complete graph that is _ not _ a edge of the delaunay graph then it can be replaced by a shorter edge that is in the delaunay graph .
let @xmath126 be our test edge from vertex @xmath210 to vertex @xmath175 .
the edge can not be a boundary edge of the convex hull of the vertices , because it is well known that every such edge is in the delaunay graph .
thus , one of vertices @xmath210 or @xmath175 must be an interior point , say @xmath210 . consider the edge @xmath126 as it leaves @xmath210 . since @xmath210 is completely surrounded by delaunay cells that have @xmath210 as a vertex it must pass through one such cell that we call @xmath257 .
replace one vertex other than @xmath210 , say @xmath559 of @xmath257 , by the vertex @xmath175 .
now , we still have @xmath560 vertices for @xmath561 and we can construct their ( unique ) circumscribing sphere @xmath562 .
now , delaunay cells are uniquely determined as those sets of @xmath560 vertices whose circumscribing sphere does not contain any other points . from this
we see that @xmath559 can not be out side @xmath562 , because otherwise its circumscribing sphere would contain @xmath175 .
our claim is that any @xmath563-dimensional hyperplane @xmath248 that contains the edge @xmath126 has at least one edge of @xmath564 , say @xmath565 , that must be on the small side " of @xmath248 in relation to @xmath562 ; that is the side not containing the centre of @xmath562 . any such edge , and therefore , @xmath565 must be shorter than @xmath126 . then
, if @xmath126 were in the minimal spanning tree , we could replace it by @xmath565 , giving a contradiction .
the proof for the gabriel graph is more straightforward .
we need to show that the minimal spanning tree has the gabriel property .
take any edge @xmath126 of the tree if the diametric sphere contains the third vertex @xmath0 then we can replace @xmath126 by @xmath566 or @xmath567 , both of which have a shorter `` length '' than @xmath126 .
this gives a contradiction .
\(1 ) although this is a known result , for example @xcite @xcite , we show a short proof .
since a comparison triangle for the cat(@xmath0 ) property is on a sphere of radius @xmath318 , first consider the unit sphere @xmath568 and the geodesic distance @xmath43 on it .
take three points @xmath569 and think of the convexity of @xmath570 for @xmath571 . without losing generality ,
assume that @xmath572 is on the plane @xmath573 and @xmath316 is on the plane @xmath574 and let @xmath575 , @xmath576 and @xmath577 for @xmath578$],@xmath579 $ ] .
thus , @xmath580 and @xmath581 .
note that for @xmath582 , @xmath583 for @xmath582 is a convex of @xmath584 . for a @xmath585 , @xmath586 for @xmath571
, @xmath587 is a convex of @xmath584 iff @xmath588 since @xmath589 .
this means that if @xmath586 for @xmath571 , @xmath587 is a convex of @xmath584 .
if @xmath286 is cat(@xmath0 ) and has a diameter of at most @xmath590 , there is a comparison triangle @xmath317 on a sphere of radius @xmath318 such that its perimeter is at most @xmath591 and @xmath592 is a convex of @xmath593 for each @xmath594 because of the argument above after scaling by @xmath318 .
\(3 ) we show an example of the probability measure with a three - point support on @xmath568 such that the diameter is larger than @xmath595 but can be arbitrarily close to @xmath595 and the uniqueness of the intrinsic @xmath415-mean fails .
take @xmath598 , @xmath599 , @xmath600 and @xmath601 with @xmath602 , as in figure [ fig : cat_hemisphere1 ] .
let @xmath603 be the mid point of @xmath604 for @xmath605 .
put the point masses @xmath606 at @xmath149 and @xmath607 at @xmath608 and @xmath596 , and assume that there is a unique intrinsic median @xmath59 . by the symmetry , @xmath59 must be on the arc @xmath609 , and if we change the ratio @xmath610 , @xmath59 moves continuously on @xmath609 .
thus , we can set @xmath611 by tuning @xmath610 adequately .
however , @xmath415-dispersion from @xmath1 becomes @xmath612 , and @xmath415-dispersion from @xmath608 becomes @xmath613 .
this contradicts the assumption of @xmath1 being the unique @xmath415-intrinsic mean . since we can set @xmath256 and @xmath614 as arbitrarily small positive numbers , @xmath615
however , by ( 1 ) , @xmath616 ; thus , @xmath617 .
if @xmath286 is a surface with a constant curvature @xmath319 , @xmath618 where @xmath619 is the inverse function of @xmath620^{-1 } & \mbox{~for~ } 0 \leq \theta \leq \pi/6,\\ \left[\log_2\frac{\pi-2\theta}{\arccos(\sin^2\theta)}\right]^{-1 } & \mbox{~for~ } \pi/6 < \theta \leq \pi/2 \end{array } \right.\ ] ] for @xmath621 and @xmath622 for @xmath623 .
c1 : @xmath598 , @xmath625 , + @xmath626 where @xmath627 , as shown in figure [ fig : cat_hemisphere1 ] .
this satisfies @xmath628 .
+ c2 : @xmath598 , @xmath629 , and @xmath630 , as shown in figure [ fig : cat_hemisphere2 ] .
+ we put point masses @xmath98 at @xmath149 and @xmath607 at @xmath608 and @xmath596 . for @xmath631 ,
we consider c1 . as in the proof of theorem [ thm : diameter](3 ) , we can set @xmath632 .
let @xmath633 and @xmath634 denote @xmath339-dispersion from @xmath1 and @xmath608 , respectively .
then , @xmath635 therefore , @xmath636 is equivalent to @xmath637 bose , p. devroye , l. , lffler , m. , snoeyink , j. and verma , v. , almost all delaunay triangulations have stretch factor greater than @xmath644/2 , _ computational geometry _ , vol.44 , no.2 , pp . 121127 , 2011 .
matula , d. w. and robert r. s. , properties of gabriel graphs relevant to geographic variation research and the clustering of points in the plane , _ geographical analysis _ ,
vol.12 , no.3 pp . 205222 , 1980 . | a methodology is developed for data analysis based on empirically constructed geodesic metric spaces . for a probability distribution
, the length along a path between two points can be defined as the amount of probability mass accumulated along the path .
the geodesic , then , is the shortest such path and defines a geodesic metric .
such metrics are transformed in a number of ways to produce parametrised families of geodesic metric spaces , empirical versions of which allow computation of intrinsic means and associated measures of dispersion .
these reveal properties of the data , based on geometry , such as those that are difficult to see from the raw euclidean distances .
examples include clustering and classification . for certain parameter ranges ,
the spaces become cat(0 ) spaces and the intrinsic means are unique . in one case ,
a minimal spanning tree of a graph based on the data becomes cat(0 ) . in another , a so - called metric cone " construction allows extension to cat(k ) spaces .
it is shown how to empirically tune the parameters of the metrics , making it possible to apply them to a number of real cases . |
ideal hydrodynamics describes a wide variety of data from heavy ion collisions @xcite .
in particular ideal hydrodynamics successfully predicted the observed elliptic flow and its dependence on mass , centrality , beam energy , and transverse momentum .
nevertheless , the hydrodynamic approach failed in several respects .
first , above a transverse momentum @xmath0 the particle spectra deviate from hydrodynamics and approach a power law .
second , hbt radii are significantly too large compared to ideal hydrodynamics @xcite . considering the partial success of ideal hydrodynamics , viscous corrections may provide a natural explanation for these failures .
for an ideal bjorken expansion , the entropy per unit space - time rapidity @xmath1 is conserved . for a viscous bjorken expansion the entropy per unit rapidity increases as a function of proper time @xcite @xmath2 where @xmath3 is the shear viscosity . in this equation and below
we have neglected the bulk viscosity . for hydrodynamics to be valid , the entropy produced over the time scale of
the expansion ( to wit , @xmath4 ) must be small compared to the the total entropy ( @xmath5 ) .
this leads to the requirement that @xmath6 where we have defined the _ sound attenuation length _ as @xmath7 .
perturbative estimates of the shear viscosity in the plasma give @xmath8 .
below we take @xmath9 , assuming that non - perturbative effects shorten equilibration times .
viscosity modifies the thermal distribution function .
this modification influences the observed particle spectrum and hbt correlations .
the formal procedure for determining the viscous correction to the thermal distribution function is described in references @xcite .
however , the basic form of the viscous correction can be intuited without calculation .
first write @xmath10 , where @xmath11 is the equilibrium thermal distribution function and @xmath12 is the first viscous correction .
@xmath12 is linearly proportional to the spatial gradients in the system .
spatial gradients which have no time derivatives in the rest frame and are therefore formed with the differential operator @xmath13 . for a baryon free fluid ,
these gradients are @xmath14 , @xmath15 , and @xmath16 , where @xmath17 .
@xmath14 can be eliminated in favor of the other two spatial gradients using the condition that @xmath18 and the ideal equations of motion .
@xmath15 leads ultimately to a bulk viscosity and will be neglected in what follows .
finally , @xmath16 leads to a shear viscosity .
if @xmath12 is restricted to be a polynomial of degree less than two , then the functional form of the viscous correction is completely determined , @xmath19 for a boltzmann gas this is the form of the viscous correction adopted in this work . for bose and fermi gasses the ideal distribution function in eq .
[ correction ] is replaced with @xmath20 @xcite .
the coefficient @xmath21 is directly related to the sound attenuation length . indeed , using the distribution function in eq .
[ correction ] to determine the stress energy tensor , yields a relationship between the shear viscosity @xmath3 and the coefficient @xmath21 , @xmath22 for a boltzmann gas , eq .
[ tensor ] yields @xmath23 .
the thermal distribution function is now completely determined . in the next section
this correction is used to calculate corrections to the observables used in heavy ion collisions .
to quantify the effect of viscous corrections on spectra and hbt radii , i generalize the blast wave model . in the blast wave model used here , the matter undergoes a boost invariant bjorken expansion and decouples at a proper time of @xmath24 at a temperature of @xmath25 .
the matter is distributed uniformly up to a radius of @xmath26 , with velocity profile up to a maximum velocity @xmath27 .
is linearly rising : @xmath28 the blast wave model with these parameters closely models the output of a full hydrodynamic simulation @xcite and gives a reasonable fit of the data . to understand this figure qualitatively ,
consider a bjorken expansion of infinitely large nuclei .
the longitudinal pressure is reduced @xcite , @xmath31 . because the shear tensor is traceless , the _ transverse _ pressure is _ increased _ , @xmath32 .
thus , the matter distribution is pushed out to larger @xmath33 by the shear in the longitudinal direction .
more mathematically , the ratio of the corrected spectrum to the uncorrected spectrum is given by , @xmath34 for large @xmath33 we find , @xmath35 .
[ spectra ] reproduces the shape and dependence of the full viscous blast wave calculation shown in fig .
[ figspectra ] .
viscous corrections become of order one when the @xmath33 of the particle approaches @xmath36 gev .
this signals the breakdown of the hydrodynamic approach .
in fact , ideal hydrodynamics generally fails to reproduce the single particle spectra above @xmath33 of 1.5 gev .
viscosity provides a ready explanation for this breakdown . in non - central collisions elliptic flow
is calculated using the spectrum indicated in eq .
[ eqspectra ] . in non - central collisions ,
the matter is assumed to have a cylindrical distribution but the flow velocity has an elliptic component , @xmath37 . for non - central collisions
the parameters are : @xmath38 and @xmath39 . as illustrated in fig .
[ figv2 ] , viscosity reduces elliptic flow by a factor of three .
viscous corrections to @xmath40 and @xmath41 are large and negative . this may be understood qualitatively by again considering a simple bjorken expansion of infinite nuclei .
the longitudinal pressure is reduced , @xmath42 .
therefore the @xmath43 distribution ( @xmath44 ) is narrower .
however , by boost invariance the single particle distribution ( @xmath45 ) is a function of @xmath46 , which yields the relation , @xmath47 the @xmath48 distribution ( @xmath49 ) at mid rapidity is therefore narrower because the longitudinal pressure is reduced .
this implies that @xmath50 decreases due to the viscous longitudinal expansion . in summary
, viscosity provides a simple explanation for the very large radii predicted by ideal hydrodynamics .
+ * acknowledgments : * this work was supported by de - ac02 - 98ch10886 .
9 d. teaney , j. lauret , and e.v .
shuryak , phys .
86 * , 4783 ( 2001 ) ; d. teaney , j. lauret , and e.v .
shuryak , nucl - th/0110037 .
kolb , p.huovinen , u. heinz , h. heiselberg , phys . lett .
b * 500 * , 232 ( 2001 ) ; p. huovinen , p.f .
kolb , u. heinz , h. heiselberg , phys .
b * 503 * , 58 ( 2001 ) .
s. soff , s. a. bass , adrian dumitru , phys .
86 * , 3981 ( 2001 ) .
p. danielewicz , m. gyulassy , phys . rev .
d * 31 * , 53 - 62 ( 1985 ) . s. de groot , w. van leeuven , ch .
van veert , _ relativistic kinetic theory _ ( north - holland , 1980 ) .
peter arnold , guy d. moore , laurence g. yaffe , jhep * 0011 * , 001 ( 2000 ) . | i compute the first viscous correction to the thermal distribution function . with this correction ,
i calculate the effect of viscosity on spectra , elliptic flow , and hbt radii . indicating the breakdown of hydrodynamics , viscous corrections become of order one for @xmath0 .
viscous corrections to hbt radii are particularly large and reduce the outward and longitudinal radii .
this reduction is a direct consequence of the reduction in longitudinal pressure . |
dark matter halos contain baryons in the forms of stars and gas , but there is active debate about the level of agreement between the baryon fraction in halos and the cosmic mean , @xmath15 , as well as which baryonic phase dominates as a function of halo mass @xmath16 .
cosmological simulations suggest that the baryon content of massive galaxy clusters ( m@xmath17 m@xmath8 ) should be within 10% of the universal value ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ) . in galaxy clusters ,
most baryons reside in the hot , diffuse , x - ray emitting intra - cluster medium ( icm ) .
detailed x - ray measurements suggest that the hot gas fraction ( @xmath13 ) is considerably lower than the cosmic value in the inner regions of clusters ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) though recent measurements suggest that @xmath13 might approach the universal value at larger radii @xcite . depleted
gas mass fractions at small radii might indicate that the icm has been affected by non gravitational processes such as star formation or feedback from active galatic nuclei ( agn ) .
indeed , sources of heat such as these that are located near the centers of clusters might pump thermal energy into the icm causing it to expand towards the outskirts ( thus lowering @xmath13 in cluster cores ) .
significant sources of heat might even be capable of removing gas from lower - mass halos @xcite .
on the other hand , the fact that @xmath13 is lower than expected at small radii might also simply indicate that some gas has been transformed into other baryonic components such as galaxy stellar mass or intra - cluster light ( icl ) .
discriminating between these two scenarios is key towards understanding the efficiency of processes that regulate star formation and for tracing the thermo - dynamic history of icl .
galaxy groups are highly interesting in this respect since they have shallower potential wells than clusters and should therefore be more sensitive to non - gravitational processes ( _ e.g. _ , * ? ? ?
measurements of the baryon fraction at group scales , however , are disparate and subject to much debate ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in particular , one unresolved issue is whether or not the baryon content of galaxy groups is dominated by hot gas ( like clusters ) or if other baryonic phases such as galaxy stellar mass or icl also make a significant contribution . from x - ray studies
there is a fairly clear consensus that @xmath13 decreases towards lower halo masses ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the two other components , @xmath0 and @xmath18 , have proved more challenging to measure and are largely responsible for disagreements in the literature regarding the group scale baryon fraction .
the icl component is diffuse and faint ( with a typical surface brightness of 27 - 32 mag arcsec@xmath19 in the r - band ) and therefore extremely difficult to measure ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the @xmath0 component is subject to uncertainties regarding the assignment of galaxies to groups and clusters , but also to uncertainties inherent to stellar mass measurements themselves .
previous results on this topic can be broadly divided into two categories . in the first category ,
the decrease in @xmath13 towards lower halo masses is found to be exactly compensated by an increased contribution from @xmath0+@xmath18 in such a way that the baryon fraction is constant with halo mass and is found to be either slightly below @xcite or significantly below @xmath20 @xcite .
this would imply that there is a simple trade - off between @xmath13 and @xmath0+@xmath18 from group to cluster scales . in the second category , @xmath0+@xmath18 do not exactly compensate for the decrease of @xmath13 at group scales and thus the baryon fraction is found to be mildly decreasing towards lower halo masses ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
this could imply for example that the icm is more strongly affected by feedback mechanisms in groups than in clusters .
most of the work to date has not directly probed @xmath0 , but has focused instead on the relationship between halo mass and the total k - band luminosity of group and cluster galaxies ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although k - band luminosity is most sensitive to the low - mass stars that dominate the total mass of most stellar populations , it is not a direct tracer of this mass .
other work , for lack of multiband photometry , has estimated @xmath0 using simple mass - to - light ratio ( m / l ) estimates ( _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in practice
, however , mass - to - light ratio values can vary strongly with galaxy color , metallicity , and age .
no single mass - to - light ratio can capture the complexity of the full galaxy population @xcite .
no studies to date have actually accounted for the full morphological mix of galaxies in group environments and used spectral energy distribution ( sed ) fitting methods to derived @xmath0 . in this paper
we measure @xmath0 using stellar masses derived from full sed fitting to multi - band photometry ( including k - band ) from the cosmos survey .
we tackle the issue of measuring @xmath0 using two complementary approaches .
first , we use a halo occupation distribution ( hod ) model that has been calibrated to fit measurements of galaxy clustering , galaxy - galaxy lensing , and the galaxy smf as a function of redshift . using this method
we can derive @xmath0 down to much lower halo masses ( a few @xmath21 m@xmath22 ) than possible using group and cluster catalogs .
second , we use a group membership catalog to directly calculate @xmath0 for @xmath23 by summing together the stellar masses of group members .
interestingly , we find results that are significantly different than previous work . assuming a chabrier initial mass function ( imf ) , our results suggest that previously published estimates of @xmath0 on group scales are overestimated by a factor of about 2 to 5 .
the discrepancy is only partially reduced by assuming a salpeter imf .
we investigate the cause of this discrepancy and find that previous work has overestimated @xmath0 due to a combination of inaccurate stellar mass estimators and/or because they have assumed that all galaxies in groups and clusters are passive early type galaxies with a constant mass - to - light ratio .
the layout of this paper is as follows .
the data are briefly described in section [ cosmos_survey ] followed by the presentation of our method used to derive @xmath0 in section [ section_method ] .
our main results are presented in section [ results ] .
finally , we discuss the results and draw up our conclusions in section [ conclusions ] .
we assume a wmap5 @xmath24cdm cosmology with @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 km s@xmath31 mpc@xmath31 @xcite .
we assume a wmap5 cosmology to maintain consistency with the hod analysis of @xcite , which is an key ingredient for this paper .
all distances are expressed in physical mpc units .
halo mass is denoted @xmath16 , and we define @xmath32 where @xmath33 is the radius at which the mean interior density is equal to 500 times the critical density ( @xmath34 ) .
stellar mass , denoted @xmath35 , is derived using a chabrier initial mass function ( imf ) unless otherwise specified .
the integrated ( total ) stellar content is denoted @xmath0 and is defined as @xmath36 .
herein , @xmath0 is the sum of two components : a contribution from satellite galaxies denoted @xmath37 and a contribution from central galaxies denoted as @xmath38 ; it does not include the icl .
the function @xmath39 represents the hubble parameter evolution for a flat metric .
stellar mass scales as @xmath40 .
halo mass scales as @xmath41 .
all magnitudes are given on the ab system .
the cosmos survey @xcite , centered at 10:00:28.6 , + 02:12:21.0 ( j2000 ) , brings together a broad array of panchromatic observations with imaging data from x - ray to radio wavelengths and a large spectroscopic follow - up program ( zcosmos ) with the vlt @xcite . in the following sections we describe the key cosmos data sets and previously published analyses relevant to this paper . the goal of this paper is to estimate the total stellar content of dark matter halos .
stellar mass estimates are therefore an important component of this paper .
we adopt the same stellar masses as used by @xcite ( hereafter `` l11 '' ) and george et al .
( 2011 ) ( hereafter , `` ge11 '' ) .
these have been derived in a similar manner as @xcite but are based on updated redshift information ( v1.7 of the cosmos photo-@xmath42 catalog and the latest available spectroscopic redshifts as compiled by the cosmos team ) and use a slightly different cosmology ( @xmath30 km s@xmath31 mpc@xmath31 instead of @xmath43 km s@xmath31 mpc@xmath31 ) .
we refer to l11 and @xcite for details regarding the derivation of the stellar masses and only give a brief description here .
stellar masses are calculated using cosmos ground - based photometry ( filters @xmath44 ) and the depth in all bands reaches at least 25th magnitude ( ab ) with the @xmath45-band limited to @xmath46 .
stellar masses are estimated using the bayesian code described in @xcite assuming a chabrier initial mass function ( imf ) and the @xcite dust model .
the stellar mass completeness limits are shown in figure 2 of l11 and vary from @xmath47 m@xmath8 at @xmath48 to @xmath49 m@xmath8 at @xmath50 .
the entire cosmos region has been mapped through 54 overlapping _ xmm - newton _ pointings and additional _ chandra _ observations cover the central region ( 0.9 degrees@xmath51 ) to higher resolution @xcite .
these x - ray data have been used to construct a cosmos x - ray group catalog that contains 211 extended x - ray sources over 1.64 degrees@xmath51 , spanning the range @xmath52 , and with a rest - frame 0.12.4 kev luminosity range between @xmath53 and @xmath54 erg s@xmath31 .
the general data reduction process can be found in @xcite and details regarding improvements and modifications to the original catalog are given in @xcite , george et al .
( 2011 ) , and finoguenov in prep .
halo masses have been measured by @xcite for this sample by using weak lensing to calibrate the relationship between x - ray luminosity ( l@xmath55 ) and halo mass .
groups in this catalog have halo masses that span the range @xmath56 . to estimate @xmath0 from this group catalog ,
we need a method to distinguish galaxies that belong to groups from those in the field . for this
we use the group membership probability catalog from ge11 . in ge11 , the full probability distribution function ( pdf ) from the cosmos
photo-@xmath42 catalog @xcite has been used to derive a group membership probability for all galaxies with @xmath57 from the cosmos acs galaxy catalog @xcite .
the completeness and purity of the group membership assignment was studied in detail by ge11 using spectroscopic redshifts and mocks catalogs . within a radius of @xmath58
, the group membership catalog has an estimated purity of 84% and a completeness of 92% .
the completeness and purity of this catalog does not vary strongly with redshift at @xmath59 . in l11 a hod model
was used to probe the relationship between halo mass and galaxy stellar mass from @xmath2 to @xmath3 .
constraints were obtained by performing a joint fit ( in three redshift bins ) to galaxy - galaxy lensing , galaxy clustering , and the galaxy stellar mass function , with a 10 parameter hod model .
a detailed description of this model can be found in @xcite . in particular , as described in section 2.3 of that paper , this model can be used to derive @xmath0 as a function of @xmath16 .
hereafter , we refer to this approach as the `` hod method '' .
this method is discussed in more detail in section [ hod_method ] .
the aim of this paper is to derive @xmath0 as a function of @xmath60 . to achieve this goal , we use two different and complementary measurements .
first , we use a hod model that has been calibrated to fit measurements of galaxy clustering , galaxy - galaxy lensing , and the galaxy smf as a function of redshift .
second , we use a group membership catalog to directly calculate @xmath0 by summing together the stellar masses of group members .
there are advantages and disadvantages inherent to each method .
the hod method might be considered to be a more indirect probe of @xmath0 , however , it allows us to derive @xmath0 over a much wider halo mass range than with group and cluster catalogs .
for example , in cosmos , this technique allows us to probe @xmath0 down to the stellar mass completeness limit of the survey ( @xmath61 m@xmath8 at @xmath48 ) .
the group catalog method is more direct and less prone to modeling uncertainties ( for example , the assumption that satellite galaxies follow the same radial distributions as the dark matter ) .
however , this technique is limited by the halo mass range for which group and cluster catalogs can be constructed ( @xmath62 m@xmath8 ) and is subject to other uncertainties such as the accurate identification of the centers of groups and cluster of galaxies .
the group catalog method is also subject to uncertainties due to the completeness and purity of the membership assignment and to projection effects , that must be understood through mock catalogs . finally , we compare both methods with the results from the abundance matching technique described in @xcite , with the assumption of the cosmos stellar mass function .
this model has substantially fewer parameters but makes a specific assumption about the connection between galaxies and dark matter subhalos .
the relationship between halo mass and galaxy stellar mass can be described via a statistical model of the probability distribution @xmath63 that a halo of mass @xmath16 is host to n galaxies above some threshold in luminosity or stellar - mass .
this statistical model , commonly known as the halo occupation distribution ( hod ) , has been considerably successful at interpreting the clustering properties of galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although the hod is usually inferred observationally by modeling measurements of the two - point correlation function of galaxies , @xmath64 , it can also be used to model other observables such as galaxy - galaxy lensing and the stellar - mass function . in l11 the hod model was used to probe the relationship between halo mass and galaxy stellar mass from @xmath2 to @xmath3 .
constraints were obtained by performing a joint fit to galaxy - galaxy lensing , galaxy clustering , and the galaxy stellar mass function . in this paper
, we adopt the parameter fits from table 5 in l11 and we use these to study @xmath0 as a function of @xmath16 .
we calculate @xmath0 following the procedure outlined in section 2.3 of @xcite .
one potential concern with this method is that the calculation of @xmath65 might require extrapolating the hod model beyond the lower and upper stellar mass bounds for which the model has been calibrated .
as shown in l11 , @xmath0 is dominated by the stellar mass of central galaxies for @xmath66 m@xmath8 .
this concern is therefore only valid for halos larger than about @xmath67 m@xmath8 where satellite galaxies represent the dominant contribution to @xmath0 .
we now investigate the stellar mass range of satellite galaxies that build up the bulk of @xmath65 ( the satellite contribution to @xmath0 ) in halos above @xmath68 m@xmath8 .
the original hod analysis of l11 used @xmath69 halo masses ( defined with respect to 200 times the mean matter density @xmath70 ) . for simplicity , for the tests in this paragraph we will also use @xmath69 but the conclusions are also valid for @xmath1 masses .
let us consider the stellar fraction for satellites in a fixed stellar mass bin , @xmath71 , where @xmath72 represents a lower stellar mass limit and @xmath73 represents an upper stellar mass limit .
the expression for @xmath71 is given by : @xmath74 where @xmath75 represents the conditional stellar mass function for satellite galaxies ( _ e.g. _ , * ? ? ?
using this expression , we have tested how @xmath76 varies with the integral limits @xmath72 and @xmath73 .
we find that at fixed halo mass , the bulk of @xmath65 comes from galaxies with @xmath77 m@xmath8 .
for example , for @xmath78 m@xmath8 , 95@xmath79 and 90@xmath79 of @xmath65 is due to satellites with @xmath80 m@xmath8 and @xmath81 m@xmath8 respectively .
the hod model of l11 is calibrated down to @xmath82 m@xmath8 at @xmath48 and to @xmath83 m@xmath8 at @xmath50 .
therefore , the bulk of @xmath65 arises from satellites that are well within the tested limits of our model ( for a similar calculation , also see @xcite ) . as mentioned previously
, the l11 analysis assumes @xmath69 halo masses whereas in this paper we use @xmath1 halo masses so as to facilitate comparisons with previous work .
we therefore need to convert the l11 results from @xmath69 to @xmath1 . for this
we make two assumptions .
first , we assume that satellites galaxies are distributed according to the same profiles as their parent halos , namely , nfw profiles @xcite with @xmath84 @xcite .
second , we assume that the distribution of satellite galaxies is independent of stellar mass ( i.e , there is no `` satellite mass segregation '' ) . this assumption is supported by multiple observations @xcite but see also @xcite . under these two assumptions , @xmath65 is unchanged by any halo mass conversion ( because the same conversion factor applies to both the numerator and the denominator in @xmath65 ) .
the @xmath85 component does vary , however , because the halo mass ( denominator ) changes while the mass of the central galaxy ( numerator ) remains constant . for details on how to convert halo masses assuming a nfw profile , see appendix c in @xcite . for this conversion ,
we assume the mass - concentration relation of @xcite . in this section ,
we describe our procedure for calculating @xmath0 at group scales using the cosmos x - ray group catalog .
figure [ xgroup_bin ] shows the group sample as a function of redshift and @xmath86 .
blue boxes show the binning scheme that we adopt for this paper .
the redshift limits are selected to match those of l11 and are @xmath87 $ ] , @xmath88 $ ] , and @xmath89 $ ] . to ensure a robust group sample with a relatively clean member list , we include only groups with reliable optical associations having greater than 3 members , and we exclude close neighboring systems and groups near the edges of the field or masked regions where member assignment is difficult ( catalog flags xflag=1 or 2 , and poor = merger = mask=0 ; see ge11 for flag definitions ) .
this ensemble of quality cuts is synthesized by a global flag in ge11 called flag_include ( we select groups with flag_include=1 ) . in total , after these cuts , our catalog contains 129 groups at @xmath59 .
the determination of group centers is an important and non - trivial task , especially since these structures do not always have a visually obvious central galaxy .
mis - centering effects could lead to an underestimate of @xmath0 , especially if the mis - centering is large enough that the true central galaxy is excluded from the group selection for example .
george et al . in
prep show that weak lensing is an excellent tool that can be used to optimize centering algorithms by maximizing the stacked weak lensing signal at small radii ( below about 500 kpc ) .
for example , george et al . in prep also show that various naive centering algorithms such as the centroid of galaxy members , even if weighted by stellar mass or luminosity , is a poor tracer of the centers of dark matter halos .
we use the group centers from george et al . in prep that are found to maximize the stacked weak lensing signals at small radii , namely , the most massive group galaxy located within an nfw scale radius of the x - ray centroid , denoted mmgg@xmath90 .
we note , however , that george et al . in prep
have also identified a sub - set of groups ( of order 20% ) that host a more massive galaxy at a greater distance from the x - ray centroid than the nfw scale radius . for these groups ,
the weak lensing does not clearly indicate a preferred center among the two options .
we include these systems in our analysis but we have tested that our results are not affected by this choice .
group members are selected according to the following criteria : 1 .
@xmath91 where @xmath92 denotes the group membership probability derived in ge11 .
@xmath93 where @xmath94 is the projected distance to the group center ( defined hereafter by the position of the mmgg@xmath90 ) .
3 . galaxies must have a stellar mass above the completeness limit of the bin in consideration ( see table [ bin_schem ] ) .
stellar mass completeness limits for the group membership catalog are higher than those quoted in section [ stellar_masses ] , because the group membership catalog is limited to @xmath57 .
group members are selected only within a radius of @xmath14 for two reasons .
ge11 have shown that the purity of the group membership assignment drops at larger radii ( see their figure 5 ) . by only considering galaxies at @xmath93 ,
we reduce the errors due to purity corrections . in addition , most previously published work on this topic has used @xmath93 ; this choice thus facilitates comparisons with previous work ( see [ comparison ] ) .
once group members have been selected following the criteria outlined above , @xmath0 is derived for each group as the sum of the stellar masses of all group members associated with this group divided by the halo mass .
halo masses are derived from the x - ray luminosity following the weak lensing scaling relation of @xcite .
although in principle more direct than the hod method , measuring @xmath0 estimates from the group catalog has systematic errors that must be corrected for .
the two main effects on the measurement are : 1 . the completeness and purity of the membership selection . here
we include both contamination in the group membership selection due to neighboring galaxies but also the de - projection of a spherical nfw profile . indeed , the quantity that we are interested in is @xmath0 contained within a sphere of radius @xmath14 whereas our membership selection is akin to selecting galaxies in a cylinder in redshift space .
we therefore include a correction factor for this de - projection .
2 . the stellar mass completeness of the group membership catalog ( i.e. , we need to account for the contribution to @xmath0 from galaxies below the completeness limit ) . to derive these correction factors
, we use a suite of cosmos - like mock catalogs that have been described in ge11 .
mocks were created from a single simulation ( named `` consuelo '' ) 420 @xmath95 mpc on a side , resolved with 1400@xmath96 particles , and a particle mass of 1.87@xmath97 @xmath95 m@xmath8 .
this simulation can robustly resolve halos with masses above @xmath98 @xmath95 m@xmath8 which corresponds to a central galaxy stellar masses of @xmath99 @xmath95 m@xmath8 , well - matched to our completeness limit of @xmath100 at low redshift .
this simulation is part of the las damas suite ( mcbride et al . in prep ) .
halos are identified within the simulation using a friends - of - friends finder with a linking length of @xmath101 , and populated with mock galaxies according to the hod model of l11 .
see ge11 for further details regarding these mock catalogs .
we derive a single overall correction factor to @xmath0 due to the two effects described above .
to do so , we apply the group membership selection to the mock catalog and we compare the measured value of @xmath0 ( in the mocks ) to the true value of @xmath0 ( within a sphere of radius @xmath14 ) .
we derive correction factors separately for the contribution to @xmath0 from satellite and central galaxies .
these correction factors are denoted @xmath102 and @xmath103 respectively .
@xmath102 is defined as : @xmath104 where @xmath105 represents the `` true '' total mass from the mocks in a sphere of @xmath14 and down to the completeness limit of the mocks and @xmath106 represents our group member selection applied to the mocks .
@xmath103 represents the equivalent for central galaxies .
the correction factors are listed in table [ corr_fac ] for the three redshift bins and are at most 14% .
the completeness and purity correction component of @xmath102 is negative ( the measured value of @xmath65 is higher than it should be because of contamination ) whereas the stellar mass completeness component of @xmath102 is positive ( the measured value of @xmath65 must be augmented to account for satellites below our completeness limit ) .
the magnitude of the two corrections is fairly similar , but the opposite signs mean that these corrections tend to cancel each other out .
the redshift trends cause the net correction for @xmath102 to be negative at @xmath107 and positive at @xmath108 .
lllllhtb + [ -1.5ex ] bin i d & redshift & lower l@xmath55 limit & upper l@xmath55 limit & m@xmath109 completeness limit + & & @xmath110l@xmath55e(z)@xmath31/erg s@xmath31 ) & @xmath110l@xmath55e(z)@xmath31/erg s@xmath31 ) & @xmath111(m@xmath109/m@xmath8 ) + [ 1ex ] + a & @xmath87 $ ] & 41.96 & 42.43 & 9.3 + b & @xmath87 $ ] & 42.43 & 43.2 & 9.3 + c & @xmath88 $ ] & 42.43 & 43.2 & 9.9 + d & @xmath89 $ ] & 42.76 & 43.5 & 10.3 + lllll + [ -1.5ex ] bin
i d & redshift & @xmath102 & @xmath103 & + + a & @xmath87 $ ] & -0.029 & -0.019 + b & @xmath87 $ ] & -0.066 & -0.004 + c & @xmath88 $ ] & -0.047 & -0.017 + d & @xmath89 $ ] & 0.003 & -0.017 + due to the finite mass resolution of our mock catalogs , the contribution to @xmath0 from very low mass galaxies ( those that are not contained in our mock catalogs ) will not be included . using equation [ tot_msat_1 ] , we estimate that this contribution should only be of order 1 - 2@xmath79 , and we thus neglect it here .
we now calculate @xmath0 using the best fit hod parameters from l11 for each of the three redshift bins .
the results are shown in figure [ tot_sm_content ] .
the solid dark blue , blue , and turquoise lines show @xmath0 as a function of @xmath1 for @xmath112 , @xmath113 , and @xmath108 respectively . the contribution to @xmath0 from central galaxies , @xmath85 ,
is shown by the dotted red line and the contribution from satellites , @xmath65 , is shown by the dashed yellow line .
@xmath0 is dominated by central galaxies at @xmath114 m@xmath8 and by satellites at @xmath115 m@xmath8 .
note that @xmath0 in this figure does not include icl .
the redshift evolution of @xmath0 has been discussed at length in l11 .
briefly , the `` pivot halo mass '' is defined as the halo mass where @xmath0 reaches a maximum ( @xmath116 m@xmath8 ) . at fixed halo mass and below the pivot halo mass scale
, @xmath0 increases at later epochs .
this evolution can be explained by the fact that in this regime , central galaxy growth outpaces halo growth . at fixed halo mass and above the pivot halo mass scale
, there is tentative evidence that @xmath0 declines at later epochs . in l11
, we suggest that this trend might be linked to the smooth accretion of dark matter , which brings no new stellar mass , and amounts to as much as 40% of the growth of dark matter halos ( e.g. , * ? ? ?
* ) and/or the destruction of satellites and a growing icl component . the shaded grey area in figure [ tot_sm_content ] indicates the statistical error on @xmath0 .
the statistical error is small compared to the systematic uncertainty associated with the actual derivation of stellar masses .
we will return to this issue shortly in section [ sys_error ] .
we now calculate @xmath0 at group scales using the cosmos x - ray group catalog .
figure [ tot_sm_content_500c_matt ] shows our estimates for @xmath85 ( top row ) , @xmath65 ( middle row ) , and @xmath0 ( bottom row ) as derived from the group catalog ( the correction factors from table [ corr_fac ] have been applied ) .
each colored data point in figure [ tot_sm_content_500c_matt ] represents a measurement from one x - ray group .
the thick black data point represents the mean value in each halo mass bin ( see table [ bin_schem ] ) .
the horizontal error corresponds to the width of the bin while the vertical error represents @xmath117 where @xmath118 is the measured dispersion in the bin and @xmath119 is the number of groups in the bin .
the solid filled symbols in figure [ tot_sm_content_500c_matt ] indicate systems for which the weak lensing analysis of george et al . in prep suggests that the group centering might be unreliable ( mmgg@xmath90 is different than mmgg@xmath120 ) .
we do not find a strong correlation between these systems and either halo mass or @xmath0 , and find that the results are unchanged if these systems are removed from the analysis .
we find excellent agreement between the hod method and the group catalog method for @xmath85 ( compare the red squares to the red dotted line ) .
regarding @xmath65 , there is a small but consistent offset between the two methods ( compare the yellow diamonds to the yellow dashed line ) . indeed , in all three redshift bins , the @xmath65 values as measured from the group catalog are higher than those derived from the hod method by about 20% to 50% .
as mentioned in the previous section , the hod predictions for @xmath65 in figure [ tot_sm_content_500c_matt ] have been derived under the assumption that the distribution of satellite galaxies follows the same nfw as their parent halos .
we hypothesize that the offset for @xmath121 seen in figure [ tot_sm_content_500c_matt ] might indicate that either a ) the satellite mass distribution has different concentration than the dark matter or b ) the satellite mass distribution varies with stellar mass ( `` mass segregation '' ) . the observed level of offset , however , does not affect the main conclusions of this paper and so we leave a more detailed investigation of this effect for future work .
overall , the two methods show a remarkable level of agreement especially considering the fact that they are subject to very different types of systematic errors .
the two methods agree to within 30% for the estimated value for @xmath0 at group scales .
the statistical uncertainty on @xmath0 ( from both the hod method or the group catalog ) is much smaller than the systematic uncertainty associated with the determination of stellar masses .
for example , according to @xcite , the typical systematic error associated with photometric - based stellar mass estimates are of order @xmath123 dex .
this systematic uncertainty arises from the choice of a stellar population synthesis ( sps ) model , the choice of a dust attenuation model , and the assumed functional form of the star formation history .
in addition , there are also further uncertainties due to the choice of an imf .
further details regarding systematic errors in stellar mass measurements can be found in @xcite and @xcite . in this section ,
we attempt to quantify how the systematic errors described above might translate into an uncertainty on @xmath0 .
note that the aim here is only to provide a first attempt to discuss systematic errors on @xmath0 due to uncertainties in stellar mass estimates ( to our knowledge , this has not yet been quantified ) .
we will make certain assumptions that will undoubtedly need to be refined in future work .
we focus on the low redshift results here ( @xmath124 ) because the method used here relies on the availability of a large number of published smfs . to begin with , we note that the smf provides the strongest observational constraints on the hod model of l11 .
the other two observables used by l11 ( clustering and galaxy - galaxy lensing ) contain additional information , but this is less dominant than the information from the smf . in this paper , we will therefore make the simplifying assumption that the main source of systematic error in the determination of @xmath0 is due to uncertainties related to the smf . under this assumption ,
the task of determining systematic errors on @xmath122 now becomes one of determining systematic errors on the observed smf .
we consider two distinct sources of error for the smf : 1 .
a systematic error due to the choice of an imf . in this paper
, we specifically consider the salpeter and the chabrier imf @xcite .
to first order , the choice of a chabrier versus a salpeter imf , yields masses that are lower by about 0.25 dex . for simplicity
, we will assume here that we can simply convert between these two imfs using a difference of 0.25 dex and we will refer to this difference as the `` imf systematic error on the smf '' . note that another imf that is often considered is the kroupa imf @xcite .
the kroupa imf yields stellar masses that are larger than a chabrier imf by about 0.05 dex .
we neglect any possible variations of the imf with either galaxy type and/or redshift but note that this is another potential ( and currently poorly determined ) source of systematic error on @xmath0 .
all other sources of systematic error on the smf , for example , those due to the choice of a sps model , the choice of a dust attenuation model , and the assumed functional form of the star formation history
. we will refer to these together as the `` non - imf systematic errors on the smf '' . to estimate the non - imf systematic errors , we consider a compilation of various published smfs at @xmath125 .
figure [ smf_comp ] shows an ensemble of low - z ( @xmath125 ) mass functions from cosmos @xcite and from the sdss @xcite .
all mass functions in figure [ smf_comp ] have been converted to our assumed value of @xmath126 and to a chabrier imf .
observed variations between these different mass functions are due to a combination of systematic error , measurement error , sample variance , and redshift evolution .
we will adopt a conservative approach and assume that all the observed variation in figure [ smf_comp ] is due to non - imf systematic errors in the determination of stellar masses ( this is an assumption that will need to be improved on in future work ) .
we define a lower envelope and an upper envelope that is designed to encompass the observed range of smfs .
this is shown by the grey shaded region in figure [ smf_comp ] .
in what follows , we assume that this grey shaded region represents the non - imf systematic errors on the smf .
there is , however , one caveat with this method . our non - imf systematic error margin might be underestimated due to the fact that not all sps models are represented by the set of smfs shown in figure [ smf_comp ] .
for example , the sps models of @xcite are not represented in figure [ smf_comp ] . to date , however , there are no published @xmath125 smfs that have used the @xcite models .
we therefore simply note that this is a limitation with our current method that needs to be improved on in future work .
the next step is to understand how these non - imf errors on the smf might translate into a systematic error on @xmath0 . at present , the hod model of l11 can only be used in conjunction with clustering and galaxy - galaxy lensing measurement .
for this reason , we do not use the l11 hod method for this step .
instead , we will adopt an `` abundance matching '' methodology that assumes there is a monotonic correspondence between halo mass ( or circular velocity ) and galaxy stellar mass ( or luminosity ) ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
abundance matching techniques are generally designed to fit the smf alone and have fewer free parameters than the hod method .
we follow the abundance matching techniques of @xcite , using @xmath1 halo masses , using both the upper and the lower limits of the grey shaded region in figure [ smf_comp ] .
the abundance matching results are then reported in figure [ sys_err ] as the non - imf systematic error margin for @xmath0 ( solid light grey region ) .
the systematic errors margin for @xmath0 assuming a salpeter imf are estimated by shifting this region upwards by 0.25 dex in stellar mass ( hashed region in figure [ sys_err ] ) . in figure
[ sys_err ] , the cosmos results are slightly higher than our non - imf systematic error margin in figure [ sys_err ] at @xmath127 m@xmath8 .
this is due to the fact that the hod method of l11 and the abundance matching method of @xcite yield slightly different predictions for @xmath0 as function of halo mass ; this is discussed further in the following section .
@xcite have derived constraints on @xmath0 ( see their figure 10 ) by abundance matching to the sdss smf of @xcite .
figure [ sys_err ] shows these results , here translated to @xmath1 halo masses ( magenta dashed line ) and assuming a chabrier imf .
the @xcite results yield a lower amplitude for @xmath0 than the l11 cosmos results .
this is largely driven by differences between the cosmos mass function and the @xcite mass function ( shown in figure [ smf_comp ] ) . in order to investigate how much of these differences are in fact due to differences in the assumed stellar mass function ,
we have applied the abundance matching technique of @xcite to the cosmos @xmath128 smf .
this is shown as the blue dotted line in figure [ sys_err ] .
this can then be directly compared with the l11 hod results .
we find that the hod results in roughly @xmath129% higher values for all halo masses than the abundance matching technique shown here .
there are several possible sources for the 15% discrepancy between the hod and abundance matching methods .
one possibility is that the weak lensing and/or clustering is providing more information in the hod analyses than is available given the stellar mass function alone , and that this information pushes the values slightly higher .
another potential issue is that conversions have been made in each case to use @xmath1 halo masses ( hod results were calculated initially using @xmath69 and abundance matching was done using @xmath130 ) ; these conversions necessarily make assumptions about the radial profiles of satellites which may be inaccurate at the several percent level .
another possibility is that the abundance matching method might suffer from satellite incompleteness in the n - body simulations ( see e.g. wu et al . in prep ) , however , this is unlikely to be important below the group mass scale .
we leave tracking down the exact source of the difference to future work , and simply note that there is roughly a @xmath129% level uncertainty in the determination of the non - imf systematic uncertainty on @xmath0 due to the method used to derive @xmath0 from the smf .
overall , the difference between l11 and @xcite is consistent with our systematic error margin for @xmath0 , and these differences do not impact any of our conclusions significantly .
we now compare our @xmath0 measurements with two recently published results from @xcite ( hereafter `` gz07 '' ) and @xcite ( hereafter `` gi09 '' ) which have been derived using group and cluster catalogs .
figure [ tot_sm_content_500c ] shows our predictions for @xmath0 as a function of @xmath1 compared to the measurements of gz07 and gi09 . in both cases ,
our results suggest _ significantly lower values _ for @xmath0 than reported by either of these previous studies .
we now provide a more detailed comparison with both gz07 and gi09 and discuss possible explanations for the source of the discrepancy .
the left panel of figure [ tot_sm_content_500c ] shows the comparison of our low - z results with gz07 .
yellow diamonds show estimates from gz07 for the contribution to @xmath0 from satellite galaxies whereas dark blue squares show the total stellar fraction ( @xmath121+@xmath131+@xmath132 ) .
the yellow diamonds in figure [ tot_sm_content_500c ] are therefore directly comparable to our prediction for @xmath65 which is shown by the yellow dash - dash line .
gz07 find that @xmath65 increases towards lower halo masses whereas we predict the opposite trend .
it is clear from this figure that a large fraction of the observed discrepancy is due to the @xmath65 component ( and not so much to @xmath131 or @xmath133 ) .
we will now investigate the source of this difference in further detail . to estimate stellar masses , gz07
have used a cousins i - band mass - to - light ratio ( @xmath134 ) calibrated from the dynamical modeling of 2-d kinematic data from @xcite .
the @xcite sample is comprised of early type galaxies from the sauron sample ( bacon et al .
we note two important caveats with this approach : 1 .
the sauron galaxy sample is comprised solely of elliptical ( e ) and lenticular ( s0 ) galaxies which have larger @xmath135 values on average than intermediate and late type galaxies . if the early type fraction increases with halo mass as suggested by several studies ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , then this could explain both why gz07 find a steeper slope and a higher amplitude than we do for @xmath65 at group scales .
the @xmath135 ratios presented in @xcite are _ dynamical _ ( i.e total ) and are therefore sensitive to the dark matter fraction within the effective radius ( @xmath136 ) .
the dynamical mass - to - light ratios in @xcite might therefore be biased high by up to @xmath137 compared to the stellar mass - to - light ratios ( see figure 17 in @xcite for example ) .
we now investigate the mass - to - light ratios of our group members compared to the value used by gz07 . for this exercise ,
we consider group members from the membership catalog of ge11 in our two low redshift bins ( bins a and b in figure [ xgroup_bin ] ) .
we derive subaru @xmath138-band mass - to - light ratios from our stellar mass catalog using the absolute @xmath138-band ab magnitude that is provided by the cosmos photoz catalog . for this exercise
we adopt an ab solar luminosity of @xmath139 .
figure [ ml_comp ] shows the @xmath140 values for cosmos group members .
we use unextincted rest frame magnitudes from the cosmos photoz catalog to divide the histogram in figure [ ml_comp ] by galaxy color .
more specifically , we define galaxy color as @xmath141 where @xmath142 and @xmath143 are the unextincted rest - frame template magnitudes in the near - ultraviolet and the r bands defined in @xcite .
we adopt the same division as in that paper , namely : @xmath144 `` high activity '' or `` blue '' @xmath145 `` intermediate activity '' or `` intermediate '' @xmath146 `` quiescent '' or `` red '' we compute the difference in magnitudes between our subaru @xmath138 filter and a cousins i filter for a range of stellar population templates .
we find that a cousins @xmath134 corresponds to a subaru @xmath138 mass - to - light ratio of @xmath147 ( the exact value depends on galaxy color ) .
these values are represented in figure [ ml_comp ] by the hashed magenta region .
figure [ ml_comp ] shows that the stellar masses inferred by using a similar @xmath135 as gz07 are larger than our sed stellar masses by a factor of 2 to 10 ( assuming a chabrier imf ) .
the difference is particularly large for active and intermediate type galaxies .
the large spread observed in figure [ ml_comp ] highlights the fact that the group population is not well represented by any single @xmath138-band @xmath148 ratio value .
part of the observed spread can also be explained by the sensitivity of the @xmath138-band luminosity to recent star formation , underscoring the need for nir data to more accurately determine galaxy stellar mass .
the differences between our mass - to - light ratios and the value used by gz07 is reduced by assuming a salpeter imf ( figure [ ml_comp ] , right panel ) .
even with a salpeter imf , however , the gz07 m / l ratio is still well above the mean m / l ratio of our galaxy sample .
this is likely to be due to the fact that the @xcite mass - to - light ratios are dynamical and are thus sensitive to dark matter in addition to stellar mass .
another important fact worth highlighting in figure [ ml_comp ] is that a non negligible fraction of group members are star forming / intermediate type galaxies ( also see ge11 who find that the quenched fractions in groups is of order 40% to 60% at @xmath149 ) .
note however that figure [ ml_comp ] does not include the corrections described in section [ group_cat_method ] for de - projection or the completeness and purity of the membership selection .
these corrections were derived as adjustments to the total @xmath0 from mock catalogs that do not distinguish quiescent galaxies from active ones , so we can not directly apply them to the separate galaxy types in figure [ ml_comp ] . to address the possibility of color - dependent contamination of the group sample , we can test the purity and completeness of the membership selection using a subsample of galaxies with spectroscopic redshifts as done in ge11 .
the cosmos spectroscopic redshift sample , however , is dominated by zcosmos `` bright '' galaxies with @xmath150 and is therefore not representative of our group membership sample ( @xmath57 ) .
this will induce a degree of uncertainty that can only be reduced with deeper and more representative spectroscopic data and/or more representative mock catalogs .
having noted this caveat , we use this approach to compute the completeness and purity of the membership selection within @xmath14 and as a function of galaxy type . we find that the group populations of red , green , and blue galaxies are overestimated by 14% , 20% , and 30% respectively . after applying these correction factors to groups in bins a and b ( and for galaxies above our stellar mass completeness limit ) , we find that the mean red fraction at @xmath151 is roughly 50% .
so while we do see some added contamination of our group member sample among blue galaxies , they still represent a significant part of the population .
studies that aim to compute @xmath0 must take this varied group population into account .
it is clear that assuming that all group members are quiescent could lead to large biases in @xmath0 estimates .
finally , we now investigate the impact of using a single mass - to - light value similar to the one use by gz07 . for this exercice
we adopt a single value of @xmath152 and we neglect the small color dependance of the @xmath135 conversion from cousins i to subaru @xmath138 ( see figure [ ml_comp ] )
. using the cosmos group catalog , we re - compute @xmath65 using the absolute @xmath138-band ab magnitude provided by the cosmos photoz catalog and a mass - to - light ratio of @xmath152 to derive stellar masses . here
, we only consider @xmath65 . because our analysis does not include icl
, we can not compare with the gz07 @xmath131+@xmath133 component ( these were measured together as a single component by gz07 ) .
however , as noted in figure [ tot_sm_content_500c ] , most of the discrepancy between our results and gz07 occurs for @xmath65 .
figure [ mlratiotest ] shows that we are able to qualitatively reproduce the gz07 @xmath65 using our data and @xmath152 .
note however that the halo mass range and redshift range of our sample does not exactly match gz07 and so figure [ mlratiotest ] is not exactly a one - to - one comparison .
however , the fact that we obtain similar @xmath65 values as gz07 strongly suggests that the primary cause of the discrepancy is due the fact that @xmath152 largely overestimates the stellar masses of group members compared to our full sed fitting technique .
gi09 have used the cosmos data and an earlier version of our x - ray catalog to derive @xmath0 .
since we use the same data set and a similar group catalog , our results are directly comparable . the right panel of figure [ tot_sm_content_500c ] shows the comparison of our results with gi09 . since gi09 report the average over all groups at @xmath59 , their results are most comparable with our @xmath153 redshift bin . to estimate stellar masses , gi09 have used a galaxy - type dependent stellar mass - to-@xmath45 band luminosity ratio derived by using the analytical relation from @xcite assuming a salpeter imf .
however , as pointed out by @xcite , the @xcite @xmath154 relation was only calibrated for massive galaxies . in practice
the k - band mass - to - light ratio varies with galaxy age and color .
thus the @xcite @xmath154 relation overestimates the stellar masses of low mass galaxies . @xcite
have shown that the @xcite mass estimates contain galaxy - color dependent biases compared to sed fitting techniques ( see their figure 28 ) .
for example , @xcite find that the @xcite masses are biased high by more than 0.3 dex for star forming galaxies with @xmath155 .
thus , including these galaxies will lead to an overestimate of @xmath0 . since the gi09 results are based on cosmos data
, we can make a direct comparison between their mass estimates and ours .
we match our galaxy catalog with the photoz catalog that was used by gi09 .
we then recompute @xmath0 from our catalog using the same procedure as in section [ results_group ] but using the @xcite masses and the results are shown in section [ results_group ] .
for this exercise we have used the same redshift and halo mass bins as in section [ results_group ] ( see table [ bin_schem ] ) .
note however that we do not apply the purity / completeness and de - projection correction factors here . indeed , since the gi09 stellar masses are different than ours , we would need a new set of mock catalogs to perform this correction .
the correction values however are typically only of order 5 to 14% and are much smaller than difference between our results and gi09 ( a factor of @xmath1562.5 ) . neglecting these correction factors here
should therefore not change our main conclusions .
figure [ mlratiotest2 ] shows that we obtain roughly similar values for @xmath0 as gi09 when we use the @xcite masses . in detail , however , we do not exactly find the same mean values as a function of halo mass as gi09 but our methods are sufficiently different ( ge11 use a bayesian probabilistic method whereas gi09 use a background subtraction method ) that this is not too unexpected .
part of the observed difference in figure [ mlratiotest2 ] is due to the fact that gi09 use a salpeter imf whereas we use a chabrier imf ( this accounts for about 0.25 dex ) .
the remaining difference ( about 0.15 dex ) is probably due to a combination of differences in the techniques that we have used and to differences in stellar mass estimates .
in this paper we have derived the total stellar mass fraction , @xmath0 , as a function of host halo mass from @xmath2 to @xmath3 using hod methods , abundance matching methods , and direct estimates from group catalogs .
our stellar masses are derived from full sed fitting to multi - band photometry ( including k - band ) from the cosmos survey . assuming a chabrier imf , we find significantly lower estimates for @xmath0 at group scales than previous work . including ( non - imf ) systematic errors on stellar masses
, our analysis suggests that @xmath157 at @xmath7 m@xmath8 and @xmath158 at @xmath10 m@xmath8 .
we will make files available upon request including our @xmath0 estimates and the associated systematic uncertainties .
our main results are as follows : 1 . assuming a chabrier imf , we find that previously published estimates of @xmath0 on group scales could be overestimated by a factor of two to five .
this discrepancy is only partially reduced by assuming a salpeter imf .
we investigate the cause of this discrepancy and find that a large fraction of the observed difference can be explained by the use of over - simplistic mass - to - light ratio estimates
we show that galaxies in groups are a mixed population of quiescent , intermediate , and star forming galaxies . as a consequence , galaxies in groups
are not well represented by any single @xmath135 ratio value .
the assumption that all galaxies in groups are quiescent , for example , will lead to biases in @xmath0 estimates .
we quantify the systematic uncertainty on @xmath0 using abundance matching methods and show that the statistical uncertainty on @xmath0 ( @xmath4% ) is currently dwarfed by systematic uncertainties associated with stellar mass measurements ( @xmath5% excluding imf uncertainties ) .
we provide first estimates for the systematic error margin on @xmath0 due to uncertainties in stellar mass measurements .
while we have tried to provide a conservative estimate for these systematic errors , it is also clear that this is a non - trivial task that will require further investigation .
one aspect in particular that could be improved on in this respect would be to consider a larger set of smfs that cover a more representative range of sps models than we have used here .
4 . while direct measurements of @xmath0 using group and cluster catalogs are necessary and important , hod and abundance matching methods can probe @xmath0 over a much wider halo mass range than possible using group and cluster catalogs .
for example , in cosmos , hod methods allows us to probe @xmath0 down to the stellar mass completeness limit of the survey ( @xmath159 m@xmath8 and @xmath160 m@xmath8 at @xmath48 ) . in cosmos
, we show that the hod method and the group catalog method are in good overall agreement but exhibit some small but interesting differences regarding the contribution to @xmath0 from satellite galaxies ( @xmath65 ) .
5 . using identical data sets
, we find that the hod method of @xcite and the abundance matching method of @xcite agree to within 15% in terms of determining @xmath0 .
identifying the cause of remaining small differences between the two methods will require further investigation , but these small differences do not affect the main conclusions of this paper .
our finding that the total stellar fraction in groups is lower than previously estimated has interesting implications for the thermodynamic history of the intra - group gas . in simulations ,
the relative proportions of gas mass to stellar mass are directly related to the efficiency of cooling , star formation , and agn feedback ( _ e.g. _ , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
suppressed gas mass fractions at small radii might indicate that gas has been removed from the centers of groups and clusters by non gravitational processes .
for example , recent simulations from @xcite have suggested that low entropy gas might be ejected from the progenitors of present - day groups by agn feedback at high redshift ( primarily @xmath161 ) at the epoch when super - massive black holes are in quasar mode .
on the other hand , low gas mass fractions might also be exactly compensated for by @xmath0+@xmath18 in such a way that the total baryon fraction is constant with halo mass and with halo radius .
discriminating between these two scenarios is key towards understanding how much gas may have been ejected from these systems . in figure
[ compare_sun ] , we have compared our results with the mean gas mass fraction ( @xmath13 ) from @xcite , computed from a compilation of data from @xcite , @xcite , and @xcite .
this result demonstrates that @xmath0 is always lower than @xmath13 for halos above @xmath12 , for the assumption of either a salpeter or chabrier imf .
figure [ tot_f ] shows the sum of @xmath0 and @xmath13 . here
the solid grey region indicates the systematic error margin for @xmath0+@xmath13 assuming a chabrier imf ; the grey hashed region shows the systematic error margin for @xmath0+@xmath13 assuming a salpeter imf . even in the latter case
, we find a shortfall in @xmath0+@xmath13 compared to the cosmic mean , which increases towards lower halos masses .
it is important to note that figure [ tot_f ] does _ not _ include a baryonic contribution from icl .
while including icl in the baryon budget is obviously an important issue to be adressed , @xmath18 estimates suffer from large uncertainties related to the choice of a @xmath135 value .
for example , the icl estimates of @xcite assume @xmath134 ; figure [ ml_comp ] demonstrates this is on the high end compared to @xmath162 values predicted from sps models .
finally , we note that the main claim in this paper , that @xmath0 estimates are lower than previously estimated , is dominated by the @xmath65 component of @xmath0 . in summary ,
we have presented new measurements of the stellar fraction @xmath0 in groups and clusters from several complementary approaches , which we believe to be more robust than previous estimates . however , precise measurements of @xmath0 are limited by our systematic uncertainties associated with stellar mass measurements .
future improvements to the @xmath0 measurements presented in this paper will benefit most from efforts that lead to an improved understanding of galaxy stellar masses .
we thank ian mccarthy , andrey kravtsov , andrew wetzel , and stefania giodini for insightful discussions .
we are grateful to anthony gonzalez for useful comments on an early version of the manuscript and for providing data in electronic format .
we thank michele cappellari and claudia maraston for providing details on imf conversions and sps models .
we are grateful to jack bishop and al leon for inspiring conversations .
al acknowledges support from the chamberlain fellowship at lbnl and from the berkeley center for cosmological physics .
the hst cosmos treasury program was supported through nasa grant hst - go-09822 .
we wish to thank tony roman , denise taylor , and david soderblom for their assistance in planning and scheduling of the extensive cosmos observations .
rhw and psb received support from a nasa hst theory grant hst - ar-12159.01-a and from the u.s .
department of energy under contract number de - ac02 - 76sf00515 , and used computing resources at slac national accelerator laboratory .
we gratefully acknowledge the contributions of the entire cosmos collaboration consisting of more than 70 scientists . more information on the cosmos survey is available at * http://cosmos.astro.caltech.edu/*. it is a pleasure the acknowledge the excellent services provided by the nasa ipac / irsa staff ( anastasia laity , anastasia alexov , bruce berriman and john good ) in providing online archive and server capabilities for the cosmos data - sets . | measurements of the total amount of stars locked up in galaxies as a function of host halo mass contain key clues about the efficiency of processes that regulate star formation .
we derive the total stellar mass fraction @xmath0 as a function of halo mass @xmath1 from @xmath2 to @xmath3 using two complementary methods .
first , we derive @xmath0 using a statistical halo occupation distribution model jointly constrained by data from lensing , clustering , and the stellar mass function .
this method enables us to probe @xmath0 over a much wider halo mass range than with group or cluster catalogs .
second , we derive @xmath0 at group scales using a cosmos x - ray group catalog and we show that the two methods agree to within 30% . we quantify the systematic uncertainty on @xmath0 using abundance matching methods and we show that the statistical uncertainty on @xmath0 ( @xmath4% ) is dwarfed by systematic uncertainties associated with stellar mass measurements ( @xmath5% excluding imf uncertainties ) . assuming a chabrier imf
, we find @xmath6 at @xmath7 m@xmath8 and @xmath9 at @xmath10 m@xmath8 .
these values are significantly lower than previously published estimates .
we investigate the cause of this difference and find that previous work has overestimated @xmath0 due to a combination of inaccurate stellar mass estimators and/or because they have assumed that all galaxies in groups are early type galaxies with a constant mass - to - light ratio .
contrary to previous claims , our results suggest that the mean value of @xmath0 is always significantly lower than @xmath11 for halos above @xmath12 . combining our results with recently published gas mas fractions , we find a shortfall in @xmath0+@xmath13 at @xmath14 compared to the cosmic mean .
this shortfall varies with halo mass and becomes larger towards lower halos masses . |
in the early evolutionary stages of star formation process , sequential depletion of molecular species on grain mantles nurtures a peculiar low - temperature chemistry due to the removal of important gas - phase reactants , starting with sulfur - bearing species and followed by even volatile molecules such as co ( bergin & tafalla 2007 ) . besides @xmath9 , @xmath10
is thought to be least affected in this condensation process and results in an enrichment of its daughter products , @xmath11 and @xmath1 ( bergin et al .
the removal of the gas - phase @xmath12 also promotes ion - molecular reactions and induces a sharp increase in the abundance of deuterated molecules in dense cores ( millar , bennett , & herbst 1989 ) .
indeed , an enhancement of @xmath13 orders of magnitude in the @xmath3 ratio in star - forming cores ( crapsi et al . 2005 ; fontani et al . 2006 ; pillai et al .
2007 ) over the local interstellar value of @xmath14 ( oliveira et al .
2003 ) has been observed . in particular
, the deuterium fractionation of @xmath1 , @xmath15 , in low - mass star - forming cores shows an increasing trend with dynamical age in the prestellar phase ( crapsi et al . 2005 ) but a decreasing trend in the protostellar phase ( emprechtinger et al .
chemical models anticipate @xmath16 to be affected by a few factors such as the kinetic temperature and , for ion species , the electron abundance ( roueff et al . 2005 ) as well as the gas - phase co abundance ( aikawa et al . 2005 ) .
the abundance of gas - phase co is expected to decline in the prestellar phase through molecular depletion onto grain surfaces but to rise up in the protostellar phase through sublimation of ice mantles as the envelope warms up .
the correlation between deuterium fractionation and co depletion factor has been recognized in a compiled sample of prestellar and protostellar cores ( crapsi et al .
2005 ; emprechtinger et al .
2009 ) . in a subsample of taurus cores ,
a better correlation is found and leads to the speculation of external environment being influential to the evolution of a core ( crapsi et al .
on the other hand , there is no evidence of a consistent behavior of @xmath16 in the case of high - mass protostellar candidates although a clear but less dramatic enhancement has been observed in a number of massive protostellar objects ( fontani et al .
2006 ; pillai et al . 2007 ) .
the environs of high - mass protostars may not sustain low temperature long enough to build up deuterated species as abundant as those of low - mass objects . in this study
, we investigated a possible trend in the deuterium fractionation of @xmath1 with dynamical age using a sample of massive star - forming cores from one single irdc to reduce the environmental fluctuations among the selected cores .
infrared dark clouds ( irdcs ) were first discovered by infrared space observatory ( iso ) and the midcourse space experiments ( msx ) through silhouette against the bright , diffuse infrared background emission of the galactic plane ( egan et al .
1998 ; carey et al .
1998 , 2000 ; hennebelle et al . 2001 ; rathborne , jackson , & simon 2006 ) . because of their large mass ( @xmath17 ) , low temperature ( @xmath18 ) , and high density ( @xmath19 ) , irdcs
have been proposed to be in the earliest stage of massive star formation . at a distance of @xmath20 , irdc g28.34 + 0.06 ( hereafter g28 )
is associated with roughly @xmath21 in the infrared absorption region and contains several dense cores in different evolutionary stages ( carey et al .
2000 ; wang et al .
we have selected three dense cores in g28 to form an evolutionary sequence : starting with mm9 in an early stage of mass collection with one weak continuum source , followed by mm4 in a stage of mass fragmentation with at least five continuum sources , and mm1 in a later stage with embedded massive protostellar objects with a total luminosity of @xmath22 .
all the selected regions are associated with water masers , indicating star - forming activities ( wang et al .
2006 ) .
our study is enabled by the arizona radio observatory ( aro ) submillimeter telescope ( smt ) and the submillimeter array(sma ) .
the smt offers good sensitivity to detect weak line emission in extended structures while the sma can preferentially image compact structures in the regions of interest .
we observed the emissions of @xmath5 at @xmath23 and @xmath24 at @xmath25 towards the three selected regions in g28 with both the smt and sma .
single - dish observations of @xmath5 and @xmath24 toward g28-mm1 , mm4 , and mm9 were carried out on 2008 april 13 with the 10-meter smt on mount graham , arizona .
the observations were performed in the beam - switching mode .
the pointing centers followed the peak positions given by rathborne et al .
( 2006 ) : ( @xmath26 , @xmath27)(j2000 ) = ( @xmath28 ) for mm1 , ( @xmath26 , @xmath27)(j2000 ) = ( @xmath29 ) for mm4 , and ( @xmath26 , @xmath27)(j2000 ) = ( @xmath30 ) for mm9 . the primary beam is about 26 for @xmath1 and 32 for @xmath2 .
the spectral resolution is @xmath31 , corresponding to a velocity resolution of @xmath32 for @xmath1 and @xmath33 for @xmath2 .
the temperature scale @xmath34 was obtained using standard vane calibration , and the main beam temperature , @xmath35 , was derived through @xmath36 with a main beam efficiency @xmath37 .
the respective rms noise level is about @xmath38 and @xmath39 for the @xmath1 and @xmath2 data , respectively .
data reduction was done with the class package .
observations with the sma were carried out with seven antennas in the compact configuration on 2008 june 17 for @xmath1 and september 30 for @xmath2 .
the system temperature varied from 150 to 350 k during the @xmath1 observation and from 100 to 180 k during the @xmath2 observation .
the correlator was set to have a spectral resolution of @xmath40 , equivalent to @xmath41 and @xmath42 for @xmath1 and @xmath2 , respectively .
the projected baselines ranged from @xmath43 and @xmath44 , which are insensitive to structures larger than 8 ( wilner & welch 1994 ) . the full - width at half power ( fwhp ) of the primary beam is roughly 44 for @xmath1 and 53 for @xmath2 .
the observing cycle comprised scans of 1751 + 096 , mm1 , mm4 , mm9 , and 1911@xmath45201 , and was repeated every 27 minutes .
the phase centers of the first observation run were the same as those of the smt observations .
based on the results of the first observation run , we adjusted the phase center for the second observation run on the continuum peak in mm1 to ( @xmath26 , @xmath27)(j2000 ) = ( @xmath46 ) and in mm9 to ( @xmath26 , @xmath27)(j2000 ) = ( @xmath47 ) .
data inspection , bandpass and flux calibrations , as well as temporal gain derivation were done within the idl superset mir .
the flux scale was derived by observations of uranus and is estimated to be accurate within 15% .
imaging was performed using miriad package with natural weighting . to keep comparison easy , we present all the maps with center positions consistent with the phase centers of the second observation run . for every source in each observation
run , we used visibilities from both sidebands , each of a 2 ghz bandwidth , to generate a line - free continuum map .
channel maps were made with visibilities gridded into a velocity resolution of @xmath48 and resulted in rms noises , @xmath49 and @xmath50 , equivalent to @xmath51 and @xmath52 , with angular resolutions of @xmath53 and @xmath54 for @xmath1 and @xmath2 , respectively .
the @xmath0 transition of @xmath1 and @xmath2 were detected in all three sources with the smt ( fig .
[ fig_g28smt ] ) .
since both transitions contain numerous hyperfine components , we fit each spectrum of every source with a model comprised of thirty - eight hyperfine components with updated line frequencies and spontaneous emission rates ( pagani , daniel , & dubernet 2009 ) . for each individual source , all the hyperfine components of every @xmath55-level are assumed to be in thermal equilibrium at a single excitation temperature , @xmath56 , adopted from the ammonia observations with an angular resolution of 40 ( pillai et al .
the models are described by three more parameters : total column density , @xmath57 , systemic velocity , @xmath58 , full - width at half maximum ( fwhm ) as line width , @xmath59 .
model spectra are optimized with the minimization of the reduced @xmath60 value , @xmath61 , and the results are listed in table [ table_smtfits ] .
the @xmath1 spectrum towards mm1 appears to be doubly peaked , possibly affected by the presence of multiple sources ( zhang et al .
2009 ) and resulted in a significantly broader line width and a large @xmath61 .
in cold clouds , a correction of the cosmic background temperature , @xmath62 , is necessary when extracting the optical depth information of an observed spectrum with @xmath63,\ ] ] where @xmath64 is the main beam temperature of the spectra , and @xmath65 .
all the line emissions appear to have fairly small optical depths .
meanwhile , the optimized model also provides an estimate of optical depth by integrating optical depths of all the hyperfine components .
the maximum optical depth , @xmath66 , is less than 0.21 ( table [ table_smtfits ] ) .
the emission of all the observed lines is optically thin .
over all , the fitted line widths , @xmath59 , of all the transitions are much broader than their thermal line width , @xmath67 , suggesting a significant contribution from nonthermal motions .
our @xmath5 spectra also have larger line widths when compared to the @xmath68 spectra observed with the very large array ( vla ) and the effelsberg 100 m telescope ( pillai et al . 2007 ; wang et al . 2008 ) . since the two transitions have similar upper state energies , @xmath69 for @xmath1 and @xmath70 for @xmath11 , but fairly different critical densities , @xmath71 for @xmath1 ( daniel et al . 2005 ; pagani et al .
2009 ) and @xmath72 for @xmath11 ( evans 1999 ) , our @xmath1 observations tend to trace denser clumps possibly embedded in the inner region that is affected by star - forming activities as suggested by @xmath73 masers ( wang et al .
observations of lower @xmath55 transitions of @xmath1 such as the @xmath74 ( @xmath75 and @xmath76 ) line are needed to verify this interpretation .
the deuterium fractionation , @xmath16 , is calculated for every source ( table [ table_smtfits ] ) and shows a significant enhancement of roughly 3 orders of magnitude higher than the average @xmath3 ratio of @xmath14 in the local interstellar medium ( oliveira et al .
such an enhancement has been observed in a large sample of high - mass star - forming cores ( fontani et al .
2006 ) as well as low - mass prestellar and protostellar cores ( crapsi et al . 2005 ; roberts & millar 2007 ; emprechtinger et al .
in particular , we find a decreasing trend with evolutionary stage , from @xmath77 in the younger mm9 core to @xmath78 in the more evolved mm1 core , a change over a factor of roughly @xmath79 . although a decreasing trend over a factor of roughly @xmath80 has been identified in low - mass protostellar cores ( emprechtinger et al . 2009 ) , there was no evidence for such trend in a sample of massive star - forming cores associated with different molecular clouds ( fontani et al .
variations in thermal history , external environment , and initial chemical abundances across different molecular clouds may cause undesired confusion .
although our sample contains only three sources , the association in one single irdc helps to minimize the environmental variations among the sources and to reveal the gentle drop in deuterium fractionation as the dusty envelope warms up . using the sma , we further imaged the @xmath1 and @xmath2 emission to study more centrally concentrated structures .
the @xmath1 emission shows clear detections in mm1 and mm4 as well as a marginal detection in mm9 ( fig .
[ fig_g28n2hp ] ) .
since an interferometer is insensitive to structures larger than the scale corresponding to its shortest projected baseline , our sma observations serve as a spatial filter to probe the fractional flux contained in clumps smaller than 8 ( wilner & welch 1994 ) .
a comparison of the integrated intensity observed with the sma , @xmath81 , to that with the smt , @xmath82 , can be made by convolving the sma maps with the smt beam after masking out regions below @xmath83 , which are believed to be artifacts induced by the lack of short baselines .
for every source , we compare the @xmath1 integrated intensity observed with the sma to that with the smt ( fig .
[ fig_g28smt ] ) and estimate the fraction of the sma integrated intensity , @xmath84 , which indicates the state of a mass concentration process ( table [ table_mms ] ) .
the small values of @xmath84 suggest that most of the @xmath1 emission is in structures larger than 8 .
an increasing trend from mm9 to mm1 is found and agrees with the presumed evolutionary stage of the central sources . on the other hand
, @xmath2 is not detected in all regions at a level of @xmath85 , translating to an upper limit of @xmath86 at @xmath87 .
compared to the smt integrated intensity in mm9 , this detection limit sets a maximum fraction of 19% for emission coming from compact structures .
given the high critical density of @xmath71 , the large fraction of the @xmath88 emission missed by the sma observations strongly suggests the presence of cold and dense gas in scales larger than 8 ( @xmath8 ) .
similar to @xmath1 , most of the @xmath2 emission is in extended structures .
continuum emissions at @xmath89 and @xmath90 are detected in all sources with similar morphology .
for the first time , a compact continuum source is detected in mm9 ( fig .
[ fig_g28n2hp ] ) . since the visibility coverages of the two observation runs were quite different , we made continuum maps with visibilities of projected baselines within @xmath91 , the range that all sources have in common at the two frequencies . if the dust emission is optically thin , the continuum flux density @xmath92 , where @xmath93 ( shepherd & watson 2002 ) is the dust opacity at the observing frequency @xmath94 , and @xmath95 is the planck function at a dust temperature @xmath96 .
the opacity spectral index , @xmath97 , can be measured by comparing flux densities at two observing frequencies . in cold clouds , the condition @xmath98 makes the rayleigh - jeans approximation inappropriate .
we estimate the opacity spectral index in the selected regions with @xmath99}{\log ( \nu_2/\nu_1 ) } - 3,\ ] ] and the core mass with @xmath100 , where @xmath101 is the distance of the source . assuming thermal equilibrium between gas and dust over similar spatial scales in the calculations of @xmath102 and @xmath103
, we adopt the gas temperature derived from the vla ammonia observations with a comparable angular resolution of @xmath104 ( zhang et al .
2009 ) to be the dust temperature , @xmath96 .
the results are listed in table [ table_mms ] .
because of the limited frequency span of @xmath105 , we note that a systematic uncertainty of 15% in flux density measurements will produce a fairly large uncertainty of @xmath106 in our calculations .
the dust opacity spectral index shows an increasing trend from mm9 to mm1 as the central sources evolve although the multiplicity in mm4 may complicate the interpretation of its averaged @xmath102 .
nevertheless , a smaller value of @xmath102 is observed in mm9 with respect to mm1 . at millimeter wavelengths ,
the value of @xmath102 is a good probe for the size distribution of dust grains ( miyake & nakagawa 1993 ) , and a smaller @xmath102 suggests that mm9 is surrounded by larger dust grains as a result of grain growth in high - density environment . on the other hand ,
the larger @xmath102 in the more evolved region mm1 may be attributed to possible changes in the size distribution and chemical composition of the surrounding dust grains , which have been exposed to strong radiation fields generated by the associated massive ysos . in cold , dense environment
, co tends to freeze out onto dust grains while @xmath1 and @xmath11 suffer less from depletion ( bergin et al .
2002 ; tafalla et al .
2004 ) . since co is the major destroyer of molecular ions , its removal from the gas phase results in a subsequent enrichment of @xmath1 ( aikawa et al .
2005 ) . in later evolutionary stages
when the internal heating of ysos becomes important , the temperature will rise up and lead to sublimation of ice mantle , which returns volatile species such as co back to the gas phase .
the warmer temperature together with the reappearance of gas - phase co can alter the competition among chemical reaction routes and destroy @xmath1 that has been produced during the cold early phase ( roueff et al .
2005 ; aikawa et al .
2005 ) . in mm1
, a deficiency of @xmath1 is observed in the location of the dust continuum emission as well as the @xmath109 emission ( fig .
[ fig_g28mm1 ] ) in a previous study ( zhang et al .
observationally , a centrally heated temperature structure from @xmath110 to @xmath111 is derived in mm1 over spatial scales of roughly @xmath112 to @xmath113 ( pillai et al .
2006 ; zhang et al .
2009 ) . theoretically ,
a significant drop in @xmath16 is predicted across this temperature range ( roueff et al .
2005 ) , and an instant release of co is expected due to a significant drop of the co sublimation timescale from @xmath114 at @xmath115 to @xmath116 at @xmath117 ( collings et al .
this hypothesized release of co in the warm region can be traced in mm1 with the @xmath6 emission and dust continuum emission .
a contrast in chemical composition can occur across the boundary between the cold , outer part and the warm , inner part that has been altered by the newly released gas - phase reactants .
this chemical contrast has recently been observed in afgl 5142 by comparing emissions of @xmath1 and @xmath11 ( busquet et al .
our main findings are summarized as follows : 1 . a moderate decreasing trend over a factor of @xmath79 in deuterium fractionation of @xmath1 with evolutionary stage .
such trend extends the use of the @xmath4 ratio as an evolutionary tracer to high - mass protostellar candidates . 2 .
an @xmath1 void resulting in part from a warmer temperature of the core and an instant release of co from sublimated grain mantles in the warm region of mm1 .
3 . a large fraction of the @xmath88 flux missed by the interferometer suggesting the presence of cold and dense gas over a rather large area ( @xmath118 ) in g28 .
lccc @xmath119 & 16.0 & 16.6 & 13.2 + @xmath57 ( @xmath120 ) & @xmath121 & @xmath122 & @xmath123 + @xmath58 ( @xmath124 ) & @xmath125 & @xmath126 & @xmath127 + @xmath59 ( @xmath124 ) & @xmath128 & @xmath129 & @xmath130 + @xmath61 & 122.9 & 9.2 & 1.5 + @xmath66 & 0.21 & 0.16 & 0.08 + @xmath57 ( @xmath131 ) & @xmath132 & @xmath133 & @xmath134 + @xmath58 ( @xmath124 ) & @xmath135 & @xmath136 & @xmath137 + @xmath59 ( @xmath124 ) & @xmath138 & @xmath139 & @xmath140 + @xmath61 & 1.8 & 0.9 & 1.4 + @xmath66 & 0.007 & 0.003 & 0.005 + @xmath16 & @xmath141 & @xmath142 & @xmath143 lccc @xmath81 ( @xmath144 ) & 1.12 & 0.33 & 0.05 + @xmath82 ( @xmath144 ) & 12.93 & 7.17 & 2.23 + @xmath84 ( % ) & 8.6 & 4.6 & 2.4 + @xmath56 & 30.0 & 13.0 & 13.7 + @xmath145 ( jy ) & 0.59 & 0.15 & 0.03 + @xmath146 ( jy ) & 1.24 & 0.28 & 0.04 + @xmath102 & 2.3 & 2.0 & 0.9 + @xmath103 ( @xmath147 ) & 277 & 209 & 36 | we have observed the @xmath0 transition of @xmath1 and @xmath2 to investigate the trend of deuterium fractionation with evolutionary stage in three selected regions in the infrared dark cloud ( irdc ) g28.34 + 0.06 with the submillimeter telescope ( smt ) and the submillimeter array ( sma ) .
a comprehensible enhancement of roughly 3 orders of magnitude in deuterium fractionation over the local interstellar @xmath3 ratio is observed in all sources . in particular , our sample of massive star - forming cores in g28.34 + 0.06 shows a moderate decreasing trend over a factor of 3 in the @xmath4 ratio with evolutionary stage , a behavior resembling what previously found in low - mass protostellar cores .
this suggests a possible extension for the use of the @xmath4 ratio as an evolutionary tracer to high - mass protostellar candidates . in the most evolved core , mm1
, the @xmath5 emission appears to avoid the warm region traced by dust continuum emission and emission of @xmath6 sublimated from grain mantles , indicating an instant release of gas - phase co. the majority of the @xmath1 and @xmath2 emission is associated with extended structures larger than @xmath7 ( @xmath8 ) . |
genetic regulatory networks had been modeled using discrete and continuous mathematical models .
an important contribution to the simulation science is the theory of sequential dynamical systems ( sds ) @xcite . in these paper , the authors developed a new theory about the sequential aspect of the entities in a dynamical systems .
in particular laubenbacher and pareigis created an elegant mathematical background of the sds , and with it solve several aspects of the theory and applications .
probabilistic boolean networks ( pbn ) had been recently introduced @xcite , to model regulatory gene networks .
the pbn are a generalization of the widely used boolean network model ( bn ) proposed by kauffman ( 1969 ) @xcite .
while the pbn eliminate one of the main limitations of the bn model , namely its inherent determinism , they do not provide the framework for considering the sequential behavior of the genes in the network , behavior observed by the biologists . here , we introduce the probabilistic structure in sds , using for each vertices of the support graph a set of local functions , and more than one schedule in the sequence of the local function selected to form the update functions , obtaining a new concept : probabilistic sequential dynamical system ( pss ) . the notion of simulation of a pss
is introduced in section [ ehom ] using the concept of homomorphism of pss ; and we prove that the category of * sds * is a full subcategory of the of the category * pss*. several examples of homomorphisms , subsystems and simulations are given .
on the other hand , deterministic sequential dynamical systems have been studied for the last few years .
the introduction of a probabilistic structure on sequential dynamical systems is an important and interesting problem .
our approach take into account , a number of issues that have already been recognized and solved for sequential dynamical systems .
this section is an introduction with the definitions and results of sequential dynamical system introduced by laubenbacher and pareigis . here
we use sds over a finite field . in this paper , we denote the finite field @xmath0 by @xmath1 , where @xmath2 is a prime number . a sequential dynamical system ( ) over a finite field @xmath3 consists of 1 . a finite graph @xmath4 with @xmath5 vertices , and a set of edges @xmath6 .
2 . a family of local functions @xmath7 , that is @xmath8 where @xmath9 depends only of those variables which are connected to @xmath10 in @xmath11 .
3 . a permutation @xmath12 in the set of vertices
@xmath13 , called an update schedule ( i.e. a bijective map @xmath14 .
the global update function of the is @xmath15 .
the function @xmath16 defines the dynamical behavior of the and determines a finite directed graph with vertex set @xmath17 and directed edges @xmath18 , for all @xmath19 , called the state space of @xmath20 .
the definition of homomorphism between two uses the fact that the vertices @xmath21 of an and the states @xmath22 together with their evaluation map @xmath23 form a contravariant setup , so that morphisms between such structures should be defined contravariantly , i.e. by a pair of certain maps @xmath24 or by a pair @xmath25 , with the graph @xmath26 having @xmath27 vertices . here
we use a notation slightly different the one using in @xcite .
let @xmath28 and @xmath29 be two .
let @xmath30 be a digraph morphism , and @xmath31 be a family of maps in the category of * set*. the map @xmath32 is an adjoint map , because is defined as follows : consider the pairing @xmath33 and similarly @xmath34 the induced adjoint map is @xmath35 .
then @xmath36 and @xmath37 induce the adjoint map @xmath38 defined as follows : 1 .
@xmath39 then @xmath32 is an homomorphism of if for a set of orders @xmath40 associated to @xmath41 in the connected components of @xmath42 , the map @xmath32 holds the following condition and the commutative diagram .
@xmath43 where @xmath44 and @xmath45 .
if @xmath46 , then @xmath47 , and the commutative diagram is the following . 1 . @xmath48 for examples and properties see@xcite .
it is clear that the above diagrams implies the following one 1 .
the following definition give us the possibility to have several update functions acting in a sequential manner with assigned probabilities .
all these , permit us to study the dynamic of these systems using markov chains and other probability tools .
we will use the acronym pss ( or ) for plural as well as singular instances .
[ pss ] a probabilistic sequential dynamical system ( pss ) @xmath50 over @xmath1 consists of : 1 . a finite graph @xmath4 with @xmath51 vertices ; 2 .
a set of local functions @xmath52 for each vertex @xmath10 of @xmath11 .
( i. e. a bijection map latexmath:[$\sim : v_\gamma\rightarrow \{f_i 3 . a family of @xmath27 permutations @xmath54 in the set of vertices @xmath55 .
4 . and a set @xmath56 , of selection probabilities . we select one function in each set @xmath57 , that is one for each vertices @xmath10 of @xmath11 , and a permutation @xmath58 , with the order in which the vertices @xmath10 are selected ( similarly sds ) , so there are @xmath59 possibly different update functions @xmath60 , where @xmath61 .
the probabilities are assigned to the update functions , so there exists a subset @xmath62 of update functions such that @xmath63 , @xmath64 .
the state space of @xmath65 is a digraph whose vertices are the elements of @xmath66 and there are an arrow going from @xmath67 to @xmath68 if there exists a functions @xmath16 , such that @xmath69 . for each one of the selected functions in @xmath70 we have an sds inside the pss , so
the state space of the function @xmath16 is a subdigraph of the state space of the pss , so , the state space of @xmath65 is a superposition of all inside .
when we take the whole set of update functions generated by the data , we will say that we have the _ full _ pss .
we denote by @xmath71 the complement of set @xmath70 in the set of all update functions , and we will call the pss @xmath72 building by the same data but taking @xmath73 as a set of update functions , one complement of the pss @xmath74 . all the complements have the same set of function but they can use different set of probabilities . the probability of the arrow going from @xmath67 to @xmath68 for example , is the sum of the probabilities of all functions @xmath16 , such that @xmath69 .
the pss represents a generalization of the : a is a pss for which every set of local functions has one element , and there is only one permutation in the family of permutations .
the update functions of the pss have probabilities and the state space of the pss ( or high level digraph ) is described by a transition matrix , and the dynamic is described by a markov chain .
let @xmath75 be the following pss : \(1 ) the graph : @xmath76 \(2 ) if @xmath77 , then the sets of local functions are the following : @xmath78 \(3 ) the schedules or permutations are @xmath79 we obtain the following table of functions , and we select all of them for @xmath80 because the probabilities given by @xmath81 . @xmath82 the update functions are the following : @xmath83 \(4 ) the probabilities that we assign are : @xmath84 . in order to study the state space
, it is convenient to determine the transition matrix of the system .
the transition matrix of @xmath80 is @xmath85 , @xmath86 we notice that there are several pss that we can construct with the same initial data of functions and permutations , but with different set of probabilities , that is , sub - pss of @xmath80 see [ hompss ] .
for example if @xmath87 and @xmath88 , then @xmath89
one objective in this section is to show that : homomorphism of pss is a natural extension of homomorphism of .
we remark , in the introductory section [ sds ] , the definition of . in the definition of a homomorphism of pss
we establish conditions to connect the support graphs , the state space and the assigned probabilities .
let @xmath90 and @xmath91 be two pss over a finite field @xmath1 .
a homomorphism from @xmath92 to @xmath93 is a pair of functions @xmath94 where : 1 .
@xmath95 is a graph morphism , and @xmath96 is a family of maps in the category of * set*. they induce the adjoint function @xmath32 , see ( [ defh ] ) ( i ) .
2 . the induced adjoint map @xmath97 is a map such that for all update function @xmath16 in @xmath98 there exists an update function @xmath99 such that @xmath94 is an -homomorphism from @xmath100 to @xmath101 .
that is , the diagrams ( d1 ) , and ( d2 ) commute for all @xmath16 and the selected @xmath102 .
the @xmath103-condition ) for a fixed real number @xmath104 , the map @xmath32 satisfies the following : @xmath105 for all @xmath16 in @xmath98 , and its selected @xmath102 in @xmath106 by condition ( [ hompss].2 ) , and @xmath107 , we will say that an homomorphism @xmath94 from @xmath92 to @xmath93 is an * isomorphism * if @xmath108 and @xmath32 are bijective functions , and @xmath109 for all @xmath110 , in @xmath22 , and all @xmath16 . we denote it by @xmath111 .
_ in the definition of homomorphism the condition ( [ hompss].3 ) is a consequence of the condition ( [ hompss].1 ) , and ( [ hompss].2 ) .
_ suppose @xmath112 satisfies ( [ hompss].1 ) , ( [ hompss].2 ) , and @xmath113 , then for some @xmath114 we have one of the following cases * @xmath115 and @xmath116 , it is impossible by ( [ hompss].2 ) .
* @xmath117 , and @xmath118 , it is impossible because there exists at least other element @xmath119 such that @xmath120 , and so on @xmath121 . because the sum of probabilities of all arrow going up of @xmath122 is equal @xmath123 , then @xmath124 .
therefore the third condition ( [ hompss].3 ) holds , and always @xmath103 exists . _
if @xmath125 is a pss - homomorphism , and @xmath126 denote the transition matrix of @xmath127 , and the entry @xmath114 of @xmath128 is @xmath129 then : @xmath130 for all possible @xmath110 and @xmath131 in @xmath22 . _
the @xmath103-condition ( [ hompss].3 ) asserts that for a fixed real number @xmath104 , the map @xmath32 satisfies the following : @xmath132 for all @xmath16 in @xmath98 , and its selected @xmath102 in @xmath106 by condition ( [ hompss].2 ) , and @xmath107 .
if we have a function @xmath133 going from @xmath110 to @xmath134 in @xmath66 , then there exists a function @xmath135 going from @xmath122 to @xmath136 , so @xmath137 .
then for @xmath138 , and by the chapman - kolmogorov equation @xcite , we have the following : @xmath139 @xmath140 @xmath141 @xmath142 by condition ( 2 ) in definition of homomorphism .
then we proved that @xmath143 . using mathematical induction over @xmath27
, we can conclude that * @xmath144 let @xmath145 be the set of function in @xmath106 associated to a function in @xmath98 .
if @xmath146 , and we denote by @xmath147 and @xmath148 , then condition ( [ hompss].3 ) implies that @xmath149 , where @xmath150 is the maximum number of functions @xmath151 going from one state to another in the state space @xmath22 .
denoting @xmath152we know that @xmath153 .
so , if @xmath126 denote the transition matrix of @xmath127 , and the entry @xmath114 of @xmath154 is @xmath155 then the condition ( [ hompss].3 ) implies that : * @xmath156 for all possible @xmath110 and @xmath131 in @xmath22 . using equations ( [ mt].1 ) , and ( [ mt].2 ) , and denoting @xmath157^{m } $ ] , we conclude that * @xmath158 for all possible @xmath110 and @xmath131 in @xmath22 .
so , for all real number @xmath159 there exists @xmath160 , such that , @xmath161 for all natural number @xmath162 , and for all possible @xmath163 .
in fact , using notation of equation ( [ mt].2 ) , we have @xmath164 , and this implies @xmath165 .
similarly @xmath166 , so , selecting @xmath27 such that @xmath167 , we obtain @xmath168 where @xmath150 is the maximum number of functions going from one state to another in the state space of the power @xmath169 of the functions .
therefore @xmath130 for all possible @xmath110 and @xmath131 in @xmath22 , and the theorem holds if @xmath108 and @xmath32 are bijective functions , and the condition ( [ hompss].3 ) holds , but the probabilities are not equal , we will say that @xmath92 , and @xmath93 are @xmath103-equivalent , and we write @xmath170 .
so , @xmath92 , and @xmath93 are @xmath103-equivalent if there exist @xmath171 , and @xmath172 , such that for all @xmath173 and @xmath174 , we have @xmath175 .
consider @xmath176 defined as follows : if @xmath177 is the set of functions in @xmath106 selected in ( [ hompss].2 ) for the map @xmath32 then the new probability of @xmath178 is @xmath179 the set of functions @xmath145 in @xmath180 , together with the new probabilities defined above , form a new pss , that we will call image of @xmath181 .
so , the graph @xmath26 , the update functions @xmath145 determine de local functions associate to each vertex of @xmath26 , similarly the permutations using by these functions , and finally the new probabilities assigned .
we will say an injective monomorphism is a pss - homomorphism such that @xmath108 is surjective and the set of functions @xmath182 , for all @xmath10 are injective functions , and so @xmath32 , @xcite .
therefore , we will say that a pss @xmath183 is sub probabilistic sequential system of @xmath184 if there exists an injective monomorphism from @xmath183 to @xmath184 .
( [ defh].1 ) let @xmath185 be a pss .
the pair of functions @xmath186 is the _ identity homomorphism _ , and it is an example of an isomorphism .
( [ defh].2 ) an homomorphism @xmath187 of pss is an _ injective monomorphism _ if @xmath108 is surjective and @xmath188 is injective , for example see ( [ ehom].3 ) .
similarly we will say that an homomorphism is a _ surjective epimorphism _ if @xmath108 is injective and @xmath188 is surjective , for a complete description of the properties of this class of monomorphism and epimorphism see section 7 in @xcite .
we consider that the pss @xmath180 is simulated by @xmath20 if there exist a injective monomorphisms @xmath189 or a surjective epimorphism @xmath190 .
in this section we give several examples of pss - homomorphism , and simulation . in the second example
we show how the condition ( [ hompss].2 ) is verified under the supposition that a function @xmath108 is defined .
so , we have two examples in ( [ ehom].2 ) , one with @xmath108 the natural inclusion , and the second with @xmath108 the only possibility of a surjective map . in the last example we have a complete description of two pss , where we have only one permutation and two or less functions for each vertices in the graph .
in particular this homomorphism is an injective monomorphism , so is an example of simulation too . *
( [ ehom].1 ) * for the pss , in the examples [ exam1 ] , and [ ma ] we now define the natural inclusion @xmath191 . it is clear that the inclusion satisfies the two first condition to be a homomorphism , and the third one is a simple consequence of the theorem [ c3 ] . in fact
@xmath192 and @xmath193 . here
@xmath194 is simulated by @xmath195 * ( [ ehom].2 ) * consider the two graphs below @xmath196 suppose that the functions associated to the vertices are the families @xmath197 , for @xmath198 and @xmath199 for @xmath42 .
the permutations are @xmath200 , @xmath201 and @xmath202 , @xmath203 , so , @xmath204 , and @xmath205 .
then , we have constructed two pss , each one with two permutations and only one function associated to each vertex in the graph ; denoted by : @xmath206 _ case ( a ) _ we assume that there exists a homomorphism @xmath207 from @xmath65 to @xmath208 , with the graph morphism @xmath209 is given by @xmath210 .
suppose the functions @xmath211 are giving , and the adjoint function @xmath212 is defined too .
if @xmath171 is an homomorphism , which satisfies the definition ( [ hompss ] ) , then the following diagrams commute : @xmath213 @xmath214 _ case ( b ) _ consider now the map @xmath215 , defined by @xmath216 , @xmath217 , @xmath218 , and @xmath219 . if there exists an homomorphism @xmath220 that satisfies ( [ hompss].2 , then @xmath221 @xmath222 * ( [ ehom].3 ) * we now construct a pss - homomorphism from @xmath223 to @xmath224 , with the property that @xmath108 is surjective and the functions @xmath225 are injective , that we call a injective monomorphism .
the pss @xmath226 has a support graph @xmath11 with three vertices , and the pss @xmath227 has a support graph @xmath42 with four vertices @xmath228 the homomorphism @xmath229 , has the contravariant graph morphism @xmath230 , defined by the arrows of graphs , as follows @xmath216 , @xmath231 and @xmath232 .
the family of functions @xmath233 , @xmath234 ; @xmath235 the adjoint function is @xmath236 the first condition in the definition [ hompss ] holds . the pss @xmath237 with data of functions @xmath238 and @xmath239 one permutation or schedule @xmath240 and probabilities @xmath241 , so @xmath242 ; where the two update functions are @xmath243 @xmath244 the pss @xmath245 is a pss , with the following data : the families of functions : @xmath246 ; @xmath247 , @xmath248 ; and @xmath249 .
@xmath250 one schedule @xmath251 , the eight possible update functions , and its probabilities @xmath252 whose determine @xmath253 probabilities , are the following : @xmath254 we claim @xmath255 is a homomorphism .
we will prove that the following diagrams commute .
@xmath256 in fact , @xmath257 on the other hand , @xmath258 we verify the composition of functions as follows @xmath259 @xmath260 , + @xmath261 , + @xmath262 + similarly we check the condition for @xmath263 and @xmath264 .
the third condition holds , because the initial @xmath265 , because @xmath266 .
in fact : @xmath267 and @xmath268 to describe the connection between the state spaces , see figure 2 .
in this section , we prove that the pss with the _ homomorphisms _ form a category , that we denote by * pss*. in theorem [ fsds ] , we prove that the category of sequential dynamical systems * sds * is a full subcategory of * pss*. the composite @xmath273 of two graph morphisms is obviously again a graph morphism .
the composite @xmath274 is again a digraph morphism which satisfies the conditions ( [ hompss].1 ) , and ( [ hompss].2 ) .
in fact , using the proposition and definition 2.7 in @xcite , we can check these condition and conclude that the third condition holds , too by theorem [ ma ] .
so , @xmath275 is again a pss - homomorphism .
let @xmath65 , and @xmath276 be two pss over the finite field @xmath1 . for all @xmath103-homomorphisms @xmath277 and @xmath278 , then there exists a morphism @xmath279 such that the following diagram commutes @xmath280 let @xmath282 be a pss over the finite field @xmath1 .
let @xmath283 be the set of all update functions that we can construct with the local functions and the permutation in @xmath184 .
let us consider , the free abelian group generated by @xmath284 , that we denote by @xmath285 , then @xmath286 .
we can notice that we are working over a finite field with characteristic the prime number @xmath2 , so @xmath287 for all @xmath288 .
we will take the quotients of these groups by @xmath2 , that is @xmath289 , and these groups are finite , @xcite . denoting @xmath290 , for an abelian group @xmath291 ,
we rewrite the above relation by @xmath292 and , there exists a covariant functor from the category * pss * to the category of small abelian groups with morphism of such , @xmath293 , defined as follows .
@xmath294 1 .
the object function is defined by @xmath295 .
the arrow function which assigns to each homomorphism @xmath296 in the category * pss * an homomorphism of abelian groups @xmath297 which is defined in a natural way , because @xmath298 , where @xmath299 , and @xmath300 , then @xmath301 @xmath85 is a functor , in fact , the functor @xmath85 gives the possibility to work with pss using group theory , for example , because @xmath304 , and @xmath305 we assign probabilities to the set @xmath71 in some way , and we consider that all possible different assignations are @xmath103-isomorphic . a complement of the pss @xmath65 is the pss @xmath306 , and all of the complement are @xmath103-equivalent , so we can select a distribution of probabilities for the complement having in account particular applications .
we use the definition of complement in order to define a decomposition of a pss in two sub pss , only looking the set of functions .
one of the mean problem in modeling dynamical systems is the computational aspect of the number of functions and the calculation of steady states in the state space , in particular the reduction of number of functions is one of the most important problem to solve for determine which part of the network _
state space _ could be simplify . | in this paper we introduce the idea of probability in the definition of sequential dynamical systems , thus obtaining a new concept , probabilistic sequential system .
the introduction of a probabilistic structure on sequential dynamical systems is an important and interesting problem .
the notion of homomorphism of our new model , is a natural extension of homomorphism of sequential dynamical systems introduced and developed by laubenbacher and paregeis in several papers . our model ,
give the possibility to describe the dynamic of the systems using markov chains and all the advantage of stochastic theory .
the notion of simulation is introduced using the concept of homomorphisms , as usual .
several examples of homomorphisms , subsystems and simulations are given .
[ multiblock footnote omitted ] |
electronic noise generated by the thermal excitation of charge carriers has been observed and explained by johnson and nyquist more than 80 years ago @xcite and discussed in great detail in the literature since then .
more recently , it has been found that this so - called johnson - nyquist noise is the main source of dephasing in mesoscopic systems at low temperatures of a few kelvins , where phonons are frozen out .
dephasing puts an ir cut - off for interference phenomena , such as quantum corrections to the classical conductivity.@xcite the current interest in this topic arises from studies of dephasing in mesoscopic systems which consist of connected quasi-1d disordered wires , fig . [ figgraph ] , including connected rings and grids .
@xcite it has been found ( both experimentally @xcite and theoretically @xcite ) that dephasing depends not only on the dimensionality , but also on the geometry of the system .
the noise correlation function is well - understood for macroscopically homogeneous systems such as infinite wires or isolated rings , but has so - far not been studied in multiply - connected networks with leads attached at arbitrary points .
the goal of this paper is to give a transparent and systematic description of the thermal noise properties for such systems . in particular
, we will derive an expression for the fluctuations of the scalar electric potentials for arbitrary geometries , eq . , and a general expression for the corresponding dephasing rate , eq . .
throughout , we assume that a description of the noise in terms of scalar potentials is sufficient , i.e. we neglect the fluctuations of the transverse component of the electromagnetic field ( for a detailed discussion of the latter , see ref . [ ] ) .
let us start by reviewing simplified arguments to derive the noise correlation function : johnson and nyquist concluded that thermal noise in electrical conductors is approximately white , meaning that the power spectral density is nearly constant throughout the whole frequency spectrum .
if in addition the fluctuations are uncorrelated for different points in space , a correlation function for the random thermal currents in the classical limit is independent of frequency @xmath0 and momentum @xmath1 .
the power spectrum of the current density reads @xmath2 here , @xmath3 is the drude conductivity of the disordered system , @xmath4 and @xmath5 are the diffusion constant and the density of states , respectively . and
@xmath6 of multiply - connected networks attached to leads ( denoted by the usual ground symbol ) at arbitrary points . ] naively applying ohm s law , @xmath7 , to eq .
and using the relation between the electric field and the scalar potential , @xmath8 , we find @xmath9 the correlation function , eq . , corresponds to the coupling of a given electron to the bath of the surrounding electrons .
@xcite thus @xmath10 describes the process of successive emission and re - absorption of a photon , which is described effectively by the scalar potential @xmath11 .
the factor @xmath12 coincides with the solution of a diffusion equation in an infinite system , which reflects the fact that the currents , eq . , are uncorrelated in space .
these simple arguments are based on the homogeneity of the system and have assumed a local relation between potential and current , whereas transport properties in disordered metals are substantially non - local .
@xcite in this paper , we derive an analogy of eq . for disordered systems with arbitrary geometry and dimensionality ; this will in particular apply to networks of disordered wires . a detailed calculation , which takes into account all properties of the mesoscopic samples , has to be done in the real - space representation .
starting points are the usual linear response formalism and the fluctuation - dissipation theorem ( fdt ) .
@xcite although most ingredients of the following discussion will be familiar to experts , we hope that the manner in which they have been assembled here will be found not only to be pedagogically useful , but also helpful for further theoretical studies .
the paper is organized as follows : in section [ heuristic ] , we propose a heuristic description of the potential fluctuations . in section [ sectnoisecorr ] , we review a microscopic approach to the noise correlation function , based on a relation of the fluctuations of the scalar potentials to the fluctuations of the density , using the random phase approximation ( rpa ) . in section [ sectdensityresponse ]
we evaluate the density response function @xmath13 for disordered systems by using a real - space representation for arbitrary geometries .
we apply this result to the noise correlation function in section [ sectfinalcorrfunc ] . in section [ sectnetworks ] ,
we show how the noise correlation function can be calculated for networks of disordered wires .
finally , in section [ dephasing ] we discuss the relation to the fundamental problem of dephasing by electronic interactions .
a description of fluctuations in metals within the linear response formalism naturally starts with an analysis of the density fluctuations in the model of non - interacting electrons described by the standard free - electron hamiltonian @xmath14 .
this system is perturbed by an external scalar potential @xmath15 coupled to the density operator @xmath16 : @xmath17 the response of the ( induced ) charge density , @xmath18 \nonumber\\ & = - \int { \ ! { \rm d}{{\bm{y}}}\ , } \chi({{\bm{x}}},{{\bm{y}}},{\omega } ) \ , v({{\bm{y}}},{\omega } ) \ , , \label{linresponsev}\end{aligned}\ ] ] is governed by the ( retarded ) density response function : @xmath19 \right\rangle } \ , .\end{aligned}\ ] ] here @xmath20 and @xmath21 denote quantum / statistical averaging with respect to the perturbed and unperturbed hamiltonian , respectively .
the fdt relates the equilibrium density fluctuations to the imaginary ( dissipative ) part of the response function , @xmath22 } \
, , \end{aligned}\ ] ] where @xmath23 in writing eqs .
( [ nnchi])([eqfdtf ] ) we have exploited detailed balance and time - reversal symmetry .
the latter implies @xmath24.@xcite the question which we are going to address in this paper is how to characterize the fluctuations of the electric potential @xmath11 . for this purpose
we consider the `` dual '' case , where some external density @xmath25 is the perturbation that couples to the `` potential operator '' @xmath26:@xcite @xmath27 the linear response of @xmath26 to the perturbation can be written as @xmath28 defining the response function @xmath29 . in analogy to eq .
, the response function also characterizes the equilibrium fluctuations of the potential:@xcite @xmath30 } \ , .\ ] ] calculating the response function @xmath31 is a complicated task because it requires precise knowledge of the potential operator @xmath32 .
instead , we can identify the potential @xmath33 in eq . with the response @xmath34 in eq .
to relate @xmath29 to @xmath13 : in the limit of strong screening in good conductors ( called the unitary limit @xcite ) , electroneutrality is satisfied locally .
therefore , the induced charge exactly compensates the external charge : @xmath35 . now inserting eq . into eq .
( or vice versa ) , we obtain @xmath36 if @xmath13 is known , eqs .
( [ vvichi ] ) and allow one to calculate the correlation function of the scalar potential .
let us recall the well - known case of macroscopically homogeneous diffusive systems .
the expression for the disordered averaged response function @xmath37 reads@xcite @xmath38 where we used eq . .
inserting eq . into eq . , we find @xmath39 which reduces to eq . in the limit @xmath40 .
in eqs . ( [ defhint1 ] ) and , we introduced the operators @xmath41 and @xmath26 assuming that either @xmath15 or @xmath25 are external perturbations .
in fact , the fluctuations originate inside of the system and the starting point of a microscopic description is the part of the hamiltonian , which describes electron interactions , @xmath42 where @xmath43 is the bare coulomb interaction . in the mean - field approximation , eq .
gives rise to a correction , called hartree contribution , to the electron energy : @xmath44 where @xmath45 .
the coulomb interactions are dynamically screened , which can be accounted for in the framework of the rpa , provided that the electron density is high , @xmath46 see fig .
[ figurpa ] .
note that @xmath47 [ see the definition in eq .
] is equal to the bubble diagrams of fig .
[ figurpa ] , see e.g. ref .
in appendix [ appendixscreening ] , we recall how to obtain eq . within a self consistent treatment of the screening problem . using the rpa in eq . and comparing the result with equation eq . , we observe that the potential fluctuations are due to electronic interactions and that the operator of the scalar potential is given by @xmath48 eq .
allows us to relate the correlation function of the potentials to the correlation function of the density fluctuations : @xmath49 by inserting eqs .
( [ eqinitialinteq ] ) and into eq . , re - ordering the terms in the rpa series and using the fact that @xmath50 is real , we find ( see fig . [ proofurpa ] ) @xmath51 } \ , .\ ] ] we emphasize that the derivation of eq
. has not used any other assumption than the rpa .
thus , eqs .
( [ eqinitialinteq ] ) and are a microscopic ( and more rigorous ) counterpart of the phenomenological eqs .
( [ vvichi ] ) and . ) . ]
in disordered metals the motion of the electrons is diffusive , provided that @xmath52 where @xmath53 is the fermi wave - vector , @xmath54 the mean free path and @xmath55 the system size .
it can be accounted for by substituting the disorder - averaged density response function , @xmath37 , into the phenomenological eqs .
( [ vvichi ] ) and or the microscopic eqs .
( [ eqinitialinteq ] ) and .
the function @xmath37 has been calculated for macroscopically homogeneous systems by vollhardt and wlfle .
@xcite in the following , we will show how to generalize their calculation to inhomogeneous systems .
a useful starting point is a coordinate representation of the density response function , eq . , in terms of the advanced and retarded green s functions , @xmath56 , @xmath57 \\
\nonumber & \times \overline{g^r({{\bm{x}}},{{\bm{y}}},\epsilon+{\omega})g^a({{\bm{y}}},{{\bm{x}}},\epsilon ) } \\
\nonumber & + f(\epsilon ) \overline{g^r({{\bm{x}}},{{\bm{y}}},\epsilon+{\omega})g^r({{\bm{y}}},{{\bm{x}}},\epsilon ) } \\
\nonumber & - f(\epsilon+{\omega } ) \overline{g^a({{\bm{x}}},{{\bm{y}}},\epsilon+{\omega})g^a({{\bm{y}}},{{\bm{x}}},\epsilon ) } \big ) \ , , \end{aligned}\ ] ] see fig .
[ figchi0v](a ) . here , @xmath58 is the fermi distribution function , and @xmath59 denotes disorder averaging .
the combinations @xmath60 and @xmath61 give short - range contributions , since the average of the products decouple , e.g. @xmath62 , and the disorder averaged green s functions @xmath63 and @xmath64 decay on the scale @xmath65 .
we will consider contributions to the thermal noise which are governed by distances larger than @xmath54 , cf .
therefore , details of the behavior on short scales are not important for our purposes and we replace the short - range contributions , @xmath60 and @xmath61 , by a delta function . the long range contributions @xmath66 can be calculated by standard methods,@xcite @xmath67 where @xmath68 is the impurity vertex function , @xmath69 ( the factor @xmath70 , where @xmath71 is the transport time , originates from the impurity line ) , see fig . [ figchi0v](b ) . the short - ranged product , @xmath72 , can be expanded as @xmath73 $ ] , which is obtained by transforming the product to momentum space and expanding in the transferred momentum @xmath1 and frequency @xmath0 , realizing that terms of order @xmath1 vanish due to symmetry . as a result ,
. reduces to a diffusion equation : @xmath74 where @xmath75 is the diffusion constant for a @xmath76 dimensional system .
thus , the vertex function is proportional to the diffusion propagator , @xmath77 . collecting the short- and long - range contributions and taking the limit @xmath78
, we obtain from eq .
@xmath79 eq . is valid for arbitrary geometries since it is based only on the diffusive approximation and does not require macroscopic homogeneity .
let us simplify eq . for a disordered conductor .
using eq . and @xmath80 eq .
can be written as @xmath81 where we introduced the thomas - fermi screening wave - vector @xmath82 , which corresponds to the inverse screening length in three dimensional ( 3d ) bulk systems . the kernel of eq .
is a solution to the diffusion equation which can be expanded in terms of eigenfunctions of the laplace operator .
consequently , the kernel is always separable and eq .
has a unique solution ( see e.g. ref .
[ ] for details on how the solution can be found ) . using the semi - group property of the diffusion propagators , @xmath83 \ , , \ ] ]
one can check that @xmath84 satisfies eq . .
in practice , the 3d thomas - fermi screening length @xmath85 is a microscopic scale , thus the typical value of the first term of the rhs . of eq .
, @xmath86 , is larger than @xmath87 for good conductors ( this is the so - called unitary limit , for details see ref . [ ] ) : @xmath88 in this limit , using the diffusion eq . , we obtain from eq . : @xmath89 \ , .\ ] ] we remind that @xmath90 is always real . as a result , eqs .
( [ eqfdtvv ] ) and yield @xmath91 where @xmath92 is given by eq . .
the real - space demonstration of eqs .
( [ urpadiff ] ) and for macroscopically inhomogeneous systems , are among the main results of the paper .
it is worth emphasizing the frequency - space factorization of the correlator , which plays an important role in the theory of dephasing , cf .
section [ dephasing ] .
the relation of eq . to the correlation function of the currents , eq . , is discussed in appendix [ appendix : currentcorrelation ] , and allows to put the presentation of the introduction on firm ground .
note that eq .
allows one to neglect the term @xmath93 in eq . and , thus , to reduce eq .
to the form of the phenomenological integral equation , with @xmath94 taking the place of @xmath29 .
[ the same replacement leads from eq . to eq . .
] in other words , the electric potential of the fluctuating charge densities itself is negligible when screening is strong enough ( i.e. good conductors in the unitary limit ) , justifying a _ posteriori _ our assumptions in the phenomenological section [ heuristic ] . the fact that the correlation function of the potential is proportional to the solution of the diffusion equation at zero frequency , cf .
eq . , may be understood as a non - local version of the johnson - nyquist theorem , since @xmath90 can be related to the classical dc - resistance @xmath95 between the points @xmath96 and @xmath6 ( see ref . [ ] ) : @xmath97-p({{\bm{x}}},{{\bm{y}}},0 ) \right\ } \ , .\ ] ] for example , in an infinitely long quasi-1d wire of cross - section @xmath98 , the solution of the diffusion equation is @xmath99 where @xmath100 is the component of @xmath96 along the wire .
hence we recover a resistance proportional to the distance between the points , @xmath101 .
let us now illustrate the calculation of the noise correlation function , eq . , for a network of disordered wires .
the main ingredient to eq .
is the solution of the diffusion equation at zero frequency .
wires allow a quasi-1d description of diffusion , where transverse directions can be integrated out since @xmath102 is assumed to be constant on the scale of the width of the wire . as a result ,
we replace @xmath103 , where @xmath98 is the cross - section of the wires and @xmath104 solves the 1d diffusion equation in the network , @xmath100 and @xmath105 being coordinates along the wires . recently
, effective methods have been developed to solve the resulting diffusion equation for arbitrary networks.@xcite we will review these methods in this section and evaluate the noise correlation function for a simple example .
we start by introducing some basic notations : a network is a set of vertices , labeled by an index @xmath106 , connected via wires of arbitrary length , say @xmath107 for the wire connecting vertices @xmath106 and @xmath108 .
let us define a vertex matrix @xmath109 as @xmath110 where @xmath111 if the vertices @xmath106 and @xmath108 are connected and @xmath112 otherwise .
the solution of the diffusion equation at zero frequency between arbitrary vertices @xmath113 of the network is given by the entries of the inverse matrix @xmath114 divided by the diffusion constant : @xmath115 this allows us to calculate the noise correlation function between arbitrary points of a network by inserting vertices and inverting @xmath114 . as an aside , note that arbitrary boundary conditions can be included in this scheme easily ( see refs .
[ ] for details ) .
let us consider the network shown in fig .
[ figgraphring ] , representing a ring connected to absorbing leads .
and the length of the connecting arms is @xmath116 .
vertices are labeled by numbers @xmath117 .
vertices `` @xmath118 '' and
`` @xmath119 '' denote the points where the ring is connected to the arms .
vertex `` @xmath120 '' is always placed in the upper arc , defining the running coordinate @xmath100 .
vertex `` @xmath121 '' determines the @xmath105-coordinate and is placed either in the adjacent arc ( panel a ) or in the left arc ( panel b ) or in the same arc ( panel c ) . ] for simplicity , we assumed that the ring is symmetric : the two arcs are of the same length @xmath122 and the connecting arms of length @xmath116 .
we evaluate the noise correlation function for two points in this network by inserting two vertices , called `` @xmath120 '' and `` @xmath121 '' .
vertex `` @xmath120 '' is always placed in the upper arc , encoding the running coordinate @xmath100 in the length of the connected wires .
vertex `` @xmath121 '' determines the @xmath105-coordinate and is placed either in the lower arc or in the left connecting arm or in the upper arc . in the first case ,
[ figgraphring](a ) , the vertex matrix , eq . , is given by @xmath123 the diffusion propagator is then given by @xmath124 , and we obtain from eq .
the correlation function as a function of the running coordinates @xmath100 , @xmath105 @xmath125 $ ] : @xmath126 when vertex `` @xmath121 '' is placed in the connecting arm , @xmath127 $ ] and @xmath128 $ ] [ fig . [ figgraphring](b ) ] , we get @xmath129 finally , when vertex `` @xmath121 '' is placed in the same arc of the ring as vertex `` @xmath120 '' [ see fig .
[ figgraphring](c ) ] , following the same logic we obtain , with @xmath130 , @xmath131 all other configurations can be found by symmetry arguments . , where @xmath10 is given by eqs .
( [ vvringarcadjacent])([vvringarcsame ] ) , for a fixed coordinate @xmath100 in the upper arm of the ring ( indicated by the dot ) , as a function of @xmath105 traversing the network .
@xmath132 is linear in @xmath105 and its derivative has a discontinuity at @xmath133 . ]
we plot @xmath134 for @xmath105 traversing the whole network in fig .
[ fluctring ] .
note that the resulting function is linear in @xmath105 and its derivative has a discontinuity at @xmath133 ( cf . ref .
the precise characterization of potential fluctuations is very important in studying phase coherent properties of disordered metals at low temperatures . to be specific ,
let us discuss a particular coherent property : the weak localization correction to the conductivity .
let us recall that the weak localization ( wl ) correction @xmath135 is a small contribution to the averaged conductivity arising from quantum interference of reversed diffusive electronic trajectories.@xcite at low temperatures , dephasing is dominated by electron interactions , that can be accounted for through a contribution to the phase accumulated by two time - reversed interfering trajectories in a fluctuating electric field @xcite : @xmath136 = \int_0^t { \ !
{ \rm d}\tau \ , } \,\big[v({{\bm{x}}}(\tau),\tau)-v({{\bm{x}}}(\tau),t-\tau)\big ] \ , .\ ] ] when averaged over the gaussian fluctuations of the electric field , @xmath137 yields a phase difference which cuts off the contributions of long electronic trajectories . introducing the trajectory - dependent dephasing rate @xmath138=\frac1{2t}\langle \phi[{{\bm{x}}}(\tau)]^2\rangle_{v}$ ] , the weak localization correction takes the form:@xcite @xmath139 } \rangle_{\{{{\bm{x}}}(\tau)\ } } \ , , \ ] ] where @xmath140 is the average with respect to closed diffusive trajectories of duration @xmath141 starting from @xmath96 ( not to be confused with the thermal average @xmath142 over the electric potential @xmath11 ) .
the phase fluctuations can then be related to the potential fluctuations : @xmath143 ^ 2\rangle = \int_0^t { \ ! { \rm d}\tau \ , } { \ ! { \rm d}\tau \ , } ' \int_{-\infty}^\infty\frac { { \ ! { \rm d}{\omega}\ , } } { 2\pi } \left [ e^{-i{\omega}(\tau-\tau ' ) } - e^{-i{\omega}(\tau+\tau'-t ) } \right ]
\\ \label{phasefluct } & \qquad \qquad \qquad \times \langle |v|^2 \rangle_\varphi({{\bm{x}}}(\tau),{{\bm{x}}}(\tau'),{\omega } ) \ , .\end{aligned}\ ] ] here we have introduced a new noise correlator , @xmath144 obtained from eq . by replacing @xmath92 with a modified function @xmath145 ( given below ) , on the origin
of which we now comment .
equation is well - known in the theory of dephasing : its version symmetrized with respect to frequency arises naturally when comparing the diagrammatic calculation of the dephasing time @xcite with the influence functional approach describing electrons moving in a random gaussian field @xmath11 .
@xcite diagrammatically , the symmetrized eq
. represents the keldysh component of the screened electron interaction propagator , the only substantial difference being that the diagrammatically calculated correlation function involved in the dephasing process acquires so - called `` pauli - factors '' that account for the fact that the fermi sea limits the phase space available for inelastic transitions .
@xcite these factors lead to the following replacement of the function @xmath92 in eq . and here also in eq .
: @xmath146 this restricts the energy transfer to @xmath147 , @xcite but does not affect the factorization of the correlator . inserting eq . into eq
. leads to @xmath148 ^ 2\rangle = \frac{2t}{\nu } \int_0^t { \ ! { \rm d}\tau \ , } { \ ! { \rm d}\tau \ , } ' \ , p({{\bm{x}}}(\tau),{{\bm{x}}}(\tau'),0 ) \int_{-\infty}^\infty\frac { { \ ! { \rm d}{\omega}\ , } } { 2\pi } \\ \label{eq:3 } & \qquad\qquad \times \left [ e^{-i{\omega}(\tau-\tau ' ) } - e^{-i{\omega}(\tau+\tau'-t ) } \right ] \frac{{\omega}}{2t}f_\varphi({\omega } ) \ , .\end{aligned}\ ] ] the fact that the frequency dependent function @xmath149 is symmetric allows us to add to @xmath150 the term @xmath151 $ ] , which does not contribute to the integral .
therefore we finally end up with the following expression for the dephasing rate , @xmath152 & = e^2 t \int_0^t\frac { { \ !
{ \rm d}\tau \ , } } { t}\int_0^t { \ ! { \rm d}\tau \ , } ' \ , \left [ \delta_t(\tau+\tau'-t ) - \delta_t(\tau-\tau ' ) \right ] \\\label{eq : functionaldecoh } & \qquad \times { \mathcal{r}}\left({{\bm{x}}}(\tau),{{\bm{x}}}(\tau')\right ) \
, , \end{aligned}\ ] ] written in terms of the resistance @xmath153 , defined in eq . .
the function @xmath154 , a broadened delta function of width @xmath155 and height @xmath156 , is the fourier transform of @xmath149 , which is given by @xmath157 eq .
, which is one of the main results of our paper , generalizes the results obtained in refs .
[ ] for an infinite wire and an isolated ring to arbitrary geometry . in the classical noise limit , @xmath158 ,
@xmath159 may be replaced by a @xmath160 function : the second term of eq .
vanishes and we recover the results of refs . [ ] .
let us now illustrate eq . by calculating the dephasing time for the well - understood case of one and two - dimensional isolated simply - connected samples .
the dephasing time can be extracted from the condition @xmath161 where @xmath162 is given by the functional eq .
, averaged over the typical closed random walks @xmath163 of duration @xmath141 in the system
. the problem is governed by the interplay of three time - scales : the thouless time @xmath164 , depending on the system size @xmath55 , the thermal time @xmath165 ( related to the thermal length @xmath166 ) , and the dephasing time @xmath167 .
\(i ) diffusive regime , @xmath168 ( @xmath169 ) : this is the regime considered in refs . [ ] , where the width of the broadened delta functions in eq . , @xmath170 , is the shortest time - scale .
thus , when averaging over paths @xmath163 , the characteristic length - scale @xmath171 entering the resistance @xmath172 can be determined as follows : for the first @xmath173 term this length is governed by free diffusion , since @xmath174 , hence @xmath175 . for the second term
the characteristic length is set by the width of the delta function , @xmath176 . in 1d , where @xmath177 , the first term dominates and we immediately obtain from eq .
@xmath178 . in 2d ,
the diffuson at zero frequency is logarithmic as well as the resistance , @xmath179 , where @xmath76 is the width of the sample , which can be understood from the fact that the resistance of a plane connected at two corners scales logarithmically with the system size .
. gives @xmath180 , and for the dephasing time : @xmath181 .
\(ii ) ergodic regime , @xmath182 ( @xmath183 ) : the width of @xmath159 in eq .
is still the shortest time - scale but , in contrast to ( i ) , the typical trajectories @xmath163 explore the whole system , setting the length - scale of diffusion to the system size @xmath55 , cf . refs .
[ ] . in full analogy to the diffusive regime , but replacing @xmath184 by @xmath55 , we find for 1d , @xmath185 , and for 2d , @xmath186.@xcite these examples show that for non - trivial geometries dephasing due to electron interactions can not be accounted for through a unique dephasing rate depending only on dimensionality , but must be described by a functional of the trajectories @xmath163 since the qualitative behavior of @xmath167 follows from the geometry dependent typical distance @xmath187 . for sufficiently low temperatures , on the other hand , eq .
is capable to describe the crossover to a 0d regime , where , apart from a dependence on the total system size , geometry becomes unimportant : \(iii ) 0d regime , @xmath188 ( @xmath189 ) : here , the width of the delta functions in eq . , @xmath170 , is larger than @xmath190 .
hence , the trajectories reach the ergodic limit @xmath191 before the electric potential has significantly changed : dephasing is strongly reduced .
let us denote the maximal resistance reached at the ergodic limit as @xmath192 and replace the resistance in eq .
by @xmath193 , without changing the result , since @xmath192 is constant and its contribution vanishes after integrating over @xmath194 and @xmath195 .
the difference @xmath196 is nonzero only during time differences @xmath197 , before reaching ergodicity .
thus , the leading contribution comes from the second @xmath198 term in eq . , which is constant at its maximum @xmath156 during such short time - scales .
we find @xmath199 $ ] and since the @xmath192 term dominates , we obtain a dephasing time @xmath200 , independent of geometry and with the characteristic @xmath201 behavior . @xcite
in this paper we have considered fluctuations of the scalar electric potentials in macroscopically inhomogeneous metals . we have shown how to relate the density fluctuations to the potential fluctuations , emphasizing the role of electronic interactions , provided a real space derivation of the density response function and illustrated these general ideas for the case of networks of metallic wires .
finally we have obtained a trajectory - dependent functional , eq . , which describes dephasing by electron interactions for arbitrary geometries and accounts for the quantum noise contribution .
when applied to networks , eq . can describe the full crossover from the 0d to the 1d and the 2d regime .
we acknowledge illuminating discussions with f. marquardt , and support from the dfg through sfb tr-12 ( o. ye . ) , de 730/8 - 1 ( m. t. ) and the cluster of excellence , nanosystems initiative munich .
[ sectappendix ]
we recall here how to obtain eq . using a self - consistent treatment of screening in real space.@xcite starting points
are the following three equations : ( i ) the excess charge density is decomposed into external and induced contributions @xmath202 ( ii ) the induced charge is related to the potential @xmath33 by the density response function , cf .
@xmath203 ( iii ) the poisson equation @xmath204 self - consistency lies in the fact that the response involves the screened potential @xmath33 and not the bare `` external '' potential related to @xmath205 .
the screened effective interaction between electrons @xmath206 is obtained by placing an external charge at @xmath6 , so that the external density is @xmath207 , and associating the resulting screened potential @xmath33 in eq . with @xmath208 .
we obtain @xmath209 convolution with the coulomb interaction gives eq . .
we discuss here the relation between the density and the current density correlations .
the response of the ( induced ) current density is characterized by the conductivity tensor @xmath210 , @xmath211 which is related to eq . by current conservation : @xmath212 the thermal fluctuations of the current density can be obtained from @xmath213 } $ ] , in analogy to the discussion in section [ heuristic ] , assuming time - reversal symmetry , @xmath214 .
let us now examine the case of disordered metals .
the classical contribution to the averaged non - local dc - conductivity has been derived in ref .
their result can be generalized straightforwardly to non - zero frequencies , @xmath215 \ , , \ ] ] which obeys the condition with eq .
substituted for @xmath13 . for the current correlations ,
we find @xcite @xmath216 } \big\ } \ , .\end{aligned}\ ] ] since the diffuson , @xmath217 , decays exponentially on a length scale @xmath218 , this expression shows that current correlations can be considered as purely local over the scale @xmath219 , i.e @xmath220 . in the limit of classical noise ,
@xmath221 , we recover precisely eq . .
b. l. altshuler , a. g. aronov , and d. e. khmelnitsky , _ effects of electron - electron collisions with small energy transfer on quantum localization _ , j. phys .
c * 15 * , 7367 ( 1982 ) .
m. ferrier , l. angers , a. c. h. rowe , s. guron , h. bouchiat , c. texier , g. montambaux , and d. mailly , _ direct measurement of the phase coherence length in a gaas / gaalas square network , _ phys .
lett . * 93*(24 ) , 246804 ( 2004 ) .
f. schopfer , f. mallet , d. mailly , c. texier , g. montambaux , c. buerle , and l. saminadayar , _ dimensional crossover in quantum networks : from macroscopic to mesoscopic physics _ ,
lett . * 98 * , 026807 ( 2007 ) .
m. ferrier , a. c. h. rowe , s. gueron , h. bouchiat , c. texier , and g. montambaux , _ geometrical dependence of decoherence by electronic interactions in a gaas / gaalas square network _ , phys .
lett . * 100 * , 146802 ( 2008 ) .
m. treiber , o. m. yevtushenko , f. marquardt , j. von delft , and i. v. lerner , _ dimensional crossover of the dephasing time in disordered mesoscopic rings : from diffusive through ergodic to 0d behavior _ , in _ perspectives of mesoscopic physics - dedicated to yoseph imry s 70th birthday _ , edited by a. aharony , o. entin - wohlman ( world scientific , singapore , 2010 ) , chap . 20 .
in general , if the electromagnetic field is not quantized , as is the case here , the potential is not an operator ; the assumption of a potential operator arises from electronic interactions that relate the internal electronic density operator to the electric potential .
note that the fdt expresses thermal correlations @xmath223 } $ ] , dominated by short range contributions ; this should not be confused with the mesoscopic correlations @xmath224 studied in refs .
[ ] , dominated by long range contributions . | we study johnson - nyquist noise in macroscopically inhomogeneous disordered metals and give a microscopic derivation of the correlation function of the scalar electric potentials in real space . starting from the interacting hamiltonian for electrons in a metal and the random phase approximation
, we find a relation between the correlation function of the electric potentials and the density fluctuations which is valid for arbitrary geometry and dimensionality .
we show that the potential fluctuations are proportional to the solution of the diffusion equation , taken at zero frequency . as an example , we consider networks of quasi-1d disordered wires and give an explicit expression for the correlation function in a ring attached via arms to absorbing leads .
we use this result in order to develop a theory of dephasing by electronic noise in multiply - connected systems . |
systems operating far from thermodynamic equilibrium are usually subjected to the action of noise which may profoundly influence their performance . over the past two decades the extensive theoretical and experimental studies in physics ,
information theory and biology have documented various phenomena of noise - induced order , noise - facilitated kinetics or noise - improved signal transmission and detection .
almost all research in this field assumes that the noise process involved can be characterized by finite variance . yet , in many situations the external non - thermal noise can be described by distribution of impulses following the heavy - tail stable law statistics of infinite variance .
notably , infinite variance does not exclude finite spread or dispersion of the distribution around its modal value .
in fact , for the family of @xmath0-stable noises the interquantile distance can be used as a proper measure of the distribution width around the median .
dynamical description of the system and its surroundings is typically carried out within a stochastic picture based on langevin equations @xcite .
the basic equation of this type reads @xmath1 where @xmath2 is the deterministic ( and possibly time - dependent ) `` force '' acting on the system and @xmath3 stands for the `` noise '' contribution describing interaction between the system and its complex surrounding .
if the noise can be considered as white and gaussian , the above equation gives rise to the classical langevin approach used in the analysis of brownian motion .
the whiteness of the noise ( lack of temporal correlations ) corresponds to the existence of time - scale separation between the dynamics of a relevant variable of interest @xmath4 and the typical time scale of the noise .
hence white noise can be considered as a standard stochastic process that describes in the simplest fashion the effects of `` fast '' surroundings @xcite .
although in various phenomena , the noise can be indeed interpreted as white ( i.e. with stationary , independent increments ) , the assumption about its gaussianity can be easily violated .
the examples range from the description of the dynamics in plasmas @xcite , diffusion in energy space , self - diffusion in micelle systems , exciton and charge transport in polymers under conformational motion and incoherent atomic radiation trapping to the spectral analysis of paleoclimatic @xcite or economic data @xcite , motion in optimal search strategies among randomly distributed target sites @xcite , fluorophore diffusion as studied in photo - bleaching experiments , interstellar scintillations @xcite , ratcheting devices @xcite and many others @xcite .
the present work overviews properties of lvy flights in external potentials with a focus on astonishing aspects of noise - induced phenomena @xcite like resonant activation ( ra ) , stochastic resonance ( sr ) and dynamic hysteresis @xcite . in particular , the persistency of the sr occurrence is examined within a continuous and a two - state description of the generic system @xcite composed of a test particle moving in the double well - potential and subject to the action of deterministic , periodic perturbations and @xmath0-stable lvy type noises . in the same system the appearance of dynamic hysteresis is documented .
moreover , an archetypal escape problem over a fluctuating barrier is analyzed revealing the ra phenomenon in the presence of non - equilibrium lvy type bath .
the model system is described by the following ( overdamped ) langevin equation driven by a lvy stable , white noise @xmath3 @xmath5 to examine occurrence of stochastic resonance and dynamic hysteresis , the generic double - well potential with the periodic perturbation ( see left panel of fig .
[ fig : model ] ) has been chosen @xmath6 with @xmath7 .
with @xmath7 for @xmath8 ( the left panel ) and the generic linear potential slope dichotomously switching between two configurations characterized by differend heights @xmath9 ( right panel).,width=264 ] in turn , for inspection of resonant activation , the action of the potential @xmath10 has been approximated by a linear slope switching dichotomously between two distinct configurations @xmath11 ( see right panel of fig . [
fig : model ] ) .
the detection of ra has been carried for a relatively `` high '' barrier @xmath12 , which under the normal diffusion condition guarantees the proper separation of time scales ( the actual times of escape events and time of diffusive motion within the potential well @xcite ) of the process .
following former studies on ra @xcite , the lower barrier height has been chosen to take one of two values @xmath13 or @xmath14 .
a particle has been assumed to start its motion at the reflecting boundary @xmath15 and continued as long as the position fulfilled @xmath16 .
it should be stressed that with generally non - gaussian white noise the knowledge of the boundary location alone can not specify in full the corresponding boundary conditions for reflection or absorption , respectively @xcite .
the trajectories driven by non - gaussian white noise display irregular , discontinuous jumps . as a consequence , the location of the boundary itself is not hit by the majority of discontinuous trajectories .
this implies that regimes beyond the location of the boundaries must be properly accounted for when setting up the boundary conditions .
in particular , multiple recrossings of the boundary location from excursions beyond the specified state space have to be excluded .
as it has been demonstrated elsewhere @xcite , incorporation of lvy flights into kinetic description requires use of nonlocal boundary conditions which , for the case of absorbing boundary at , say , @xmath17 calls for extension of the absorbing regime to the semiline beyond that point , @xmath18 . the barrier fluctuations causing alternating switching between the high ( @xmath19 ) and low ( @xmath20 ) barrier configurations have been approximated by the markovian dichotomous noise @xcite with the exponential autocorrelation function @xmath21 [ \eta(t')-\langle \eta(t')\rangle ] \rangle = \frac{1}{4}(h_+-h_-)^2\exp(-2\gamma |t - t'|)$ ] .
here @xmath22 represents the transitions rates between @xmath23 and @xmath24 states , and both lvy and dichotomous noises have been assumed to be statistically independent .
the sample trajectories @xmath4 have been obtained by a direct integration of the langevin equation ( [ eq : langevin ] ) using the standard techniques of integration of stochastic differential equations with respect to the lvy stable pdfs @xcite @xmath25 here @xmath26 and @xmath27 is distributed according to the stable probability density function @xmath28 whose corresponding fourier transform for @xmath29 is @xmath30 , \label{eq : charakt}\ ] ] while for @xmath31 this expression reads @xmath32 .
\label{eq : charakter}\ ] ] the parameter @xmath33 $ ] denotes the stability index , yielding the asymptotic long tail power law for the @xmath34-distribution , which for @xmath35 is of the @xmath36 type .
the parameter @xmath37 ( @xmath38 ) characterizes the scale whereas @xmath39 ( @xmath40 $ ] ) defines an asymmetry ( skewness ) of the distribution @xcite .
exemplary probability density functions for symmetric ( @xmath41 ) and skewed ( @xmath42 ) cases have been displayed in fig .
[ fig : densities ] . with @xmath41 ( from the bottom to the top ) ( left panel ) and asymmetric stable densities for @xmath43 with @xmath44 ( from left to right ) ( right panel).,width=302 ]
the response of the ensemble average coordinate @xmath45 to external periodic perturbations ( cf .
( [ eq : langevin ] ) ) gives rise to the phenomenon of `` dynamic hysteresis '' . in order to register the behavior ,
we have performed an analysis of trajectories @xmath4 based on a two - state approximation . for this purpose , we have defined an occupation probability of being in the left ( with the respect to the barrier position @xmath46 ) or right state according to the definition of the occupation probabilities : @xmath47=\int\limits^{x_b}_{-\infty}p(x , t)dx=1-p_{\mathrm{right}}(t).\ ] ] note that the relative occupation of either one of those states can be altered not only by the modulation of the potential , but also by an appropriate tuning of the additive noise parameters . when plotted as a function of the periodic driving @xmath48 ,
the occupation probability @xmath49 exhibits a characteristic hysteresis loop .
quite obviously , the nonzero skewness parameter @xmath39 introduces asymmetry to the dynamic hysteresis loops , cf .
[ fig : sh_a19 ] . as can be inferred from fig .
[ fig : sh_a19 ] , the positive @xmath39 `` biases '' the motion , causing the trajectories to stay more likely to the right of the barrier top @xmath15 . with fixed @xmath50 .
the time step of the integration @xmath51 .
results were averaged over @xmath52 realizations .
initial conditions @xmath53 were sampled from the interval @xmath54 $ ] .
the particle is moving in the modulated double - well potential ( [ eq : potential ] ) with @xmath55.,width=264 ] the overall behavior of the dynamic hysteresis loops @xmath56 under the @xmath0-stable noises could be deduced from the inspection of the trajectories of the process presented in fig .
[ fig : strajectory ] . for the symmetric stable noise
, i.e. @xmath41 , the process spends , on average , the same amount of time in the left / right states and consequently , occupations of both states are equal . for non - zero and increasing @xmath57 a larger asymmetry in the distribution of residence times in the right / left states is registered .
the asymmetry of the occupation probability in either one of the states is reflected in the shape of the dynamic hysteresis loop , which with an increasing @xmath57 becomes distorted into the direction determined by the sign of the skewness parameter .
the probability of finding the process in the right / left state depends strongly on the stability index @xmath0 , cf .
[ fig : sh_alpha ] . for symmetric noises ( @xmath41 )
the area of the hysteresis loop decreases with decreasing stability index @xmath0 , see fig .
[ fig : sh_alpha ] .
this observation is a direct consequence of a heavy - tailed nature of the noise term in eq .
( [ eq : generallangevin ] ) . with decreasing @xmath0 , larger excursions of the particle are possible and
these occasional jumps of trajectories may be of the order of , or even larger than the distance separating the two minima of the potential @xmath10 .
a combined interplay of both noise parameters @xmath58 may result in a permanent locking of the process in one of its states , see right panel of fig .
[ fig : sh_alpha ] . in order to quantify the sr phenomenon @xcite we have used the standard measures of the signal - to - noise ratio snr and the spectral power amplification @xmath59 @xcite .
the snr is defined as a ratio of a spectral content of the signal in the forced system to the spectral content of the noise @xmath3 : @xmath60/s_n(\omega).\ ] ] here @xmath61 represents the power carried by the signal , while @xmath62 estimates the background noise level . in turn , the spectral power amplification ( @xmath59 ) is given by the ratio of the power of the driven oscillations to that of the driving signal at the driving frequency @xmath63 .
closer examination of these quantifiers reveals that @xmath59 and snr behave in a typical way both in the continuous model , cf .
[ fig : genericpower ] , and in its two - state analogue ( results not shown ) .
the spectral amplification @xmath59 has a characteristic bell - shape form indicating detection of stochastic resonance within the interval of suitably chosen noise intensity @xmath64 .
for example , if the stability index is set to @xmath50 and trajectories are simulated with extremely weak noise intensities @xmath64 , the periodic signal is not well separated from the noisy background and consequently snr stays negative .
however , snr becomes positive with increasing noise intensity what indicates emerging separation of the signal from the noisy background .
[ fig : genericpower ] results for @xmath59 and snr analysis ( in a continuous sr model ) with various stability index @xmath0 are presented .
note , that for a given @xmath0 the power spectra for @xmath65 are the same .
therefore , stochastic resonance quantifiers derived from the power spectra are equivalent for @xmath65 with the same value of @xmath0 . with @xmath41 ( the left panel ) and @xmath66 ( the right panel ) .
parameters of the simulation like in fig .
[ fig : sh_a19].,width=302 ] -stable process for the two - state and the continuous model constructed by use of eq .
( [ eq : langevin ] ) .
the time step of the integration @xmath67 , the scale parameter @xmath68 .
the frequency of the cycling voltage @xmath69 .
noise parameters as indicated in figures.,width=264 ] switching from the continuous model to the two - state approximation ( cf .
fig.5 ) does not change the qualitative behavior of the sr quantifiers . with large @xmath0 ( @xmath70 ) ] however , the slope of the decaying part of the signal - to - noise ratio is more flat for the two - state description than for the continuous model ( results not shown ) .
our numerical analysis implies that @xmath59 is more sensitive to the variation of noise parameters than snr , see fig .
[ fig : genericpower ] . at decreasing values of the stability index @xmath0 ,
the maximum of @xmath71 drops and shifts towards higher values of the noise intensity .
obviously , the decrease of @xmath0 weakens the stochastic resonance and reduces system performance .
the diminishment of spectral amplification for @xmath35 indicates that the input signal is worse reproduced in the recorded output .
this behavior is easier detectable for symmetric noises ( cf .
left panel of fig .
[ fig : genericpower ] ) than for asymmetric ones ( see the right panel ) and can be readily explained by the analysis of exemplary trajectories , see fig .
[ fig : strajectory ] . sharp spikes clearly visible in the right corner panel of fig .
[ fig : strajectory ] are due to the heavy - tailed nature of the lvy stable distribution and they are more pronounced the smaller the index @xmath0 becomes .
their presence indicates that for a sufficiently small @xmath0 the trajectory becomes discontinuous and , on average , switches between left and right wells of the potential @xmath10 are realized by sudden long - jump escape events fairly independent of the periodic driving .
although the snr is less sensitive to variations in noise parameters ( cf . fig .
[ fig : genericpower ] ) , the shape of this function flattens for decreasing values of @xmath0 thus hampering detection of the resonant value of the noise optimal intensity for which the sr phenomenon is most likely perceived .
( the left panel ) , @xmath72 ( the right panel ) and various @xmath0 .
the simulation parameters like in fig .
[ fig : strajectory ] . lines are drawn to guide the eye.,width=302 ] as already discussed in the preceding sections , the stochastic kinetics driven by additive non - gaussian stable white noises is very different from the gaussian case . for @xmath35 , a test particle moving in the linear potential can change its position via extremely long ,
jump like excursions .
this in turn requires the use of nonlocal boundary conditions in evaluation of the mean first passage time ( mfpt ) @xcite . in this paragraph
the above issue is taken care of when generating first passage times ( fpts ) by monte carlo simulations .
simulated trajectories are representative for a motion of a test particle over the interval @xmath73 $ ] ( see right panel of fig .
[ fig : model ] ) influenced by independent dichotomous switching of the potential slope and subjected to additive white lvy noise .
a particle starts its motion at @xmath15 where the reflecting boundary is located .
the absorbing boundary is located at @xmath17 meaning , that the whole semi - axis @xmath74 is assumed absorbing , yielding zero pdf @xmath75 for all @xmath18 . from the ensembles of collected first passage times
we have evaluated the mean values of the distributions and our findings for mfpt are presented in fig .
[ fig : skewedm88 ] .
integration of eq .
( [ eq : langevin ] ) was performed for a series of descending time steps of integration @xmath76 to ensure that the results are self consistent .
the left panel of fig .
[ fig : skewedm88 ] displays behavior of derived mfpts as a function of the stability index @xmath0 and the rate of the barrier modulation @xmath22 . in the right panel sample cross - sections of the surfaces mfpt(@xmath77 )
are drawn . for @xmath31 , or adequately @xmath78 ,
the procedure of simulating skewed ( @xmath42 ) stable random variables becomes unstable .
this can be well explained by examining the form of the additive noise - term characteristic function @xmath79 - see eqs .
( [ eq : charakt ] ) and ( [ eq : charakter ] ) .
the exponential functions are no longer continuous functions of the parameters and exhibit discontinuities when @xmath80 .
therefore , for the clarity of presentation , these parameter sets have been excluded from consideration .
for @xmath81 with @xmath82 ( the left panel ) and sample cross - sections @xmath83 for various @xmath0 : ` @xmath84 ' : @xmath85 ; ` @xmath86 ' : @xmath87 ; ` @xmath88 ' : @xmath89 ; ` @xmath90 ' : @xmath91 and ` @xmath92 ' : @xmath93 ( the right panel ) .
the results were calculated by direct integration of eq .
( [ eq : langevin ] ) with the time step @xmath67 and averaged over @xmath52 realizations .
horizontal lines represent asymptotic values of @xmath94 and @xmath95 .
they have been evaluated by use of the monte carlo method with @xmath96 and averaged over @xmath97 realizations .
error bars represent standard deviation of the mean.,width=302 ] the depicted results indicate appearance of the ra phenomenon which is best visible when the potential barrier is switching between two barrier configurations characterized by @xmath81 .
however , the dependence of ra on noise parameters is highly nontrivial .
the overall tendency in kinetics is similar to cases described in former paragraphs : heavier tails ( @xmath35 ) in distribution of noise increments @xmath3 result in stronger discontinuities of trajectories causing the ra effect to become inaudible and fading gradually .
notably , for totally skewed additive noise @xmath98 which acts in favor of the motion to the right , the ra seems to disappear for @xmath89 .
it reappears again for smaller values of @xmath0 , when the driving lvy white noise becomes a one sided lvy process with strictly positive increments which tend to push trajectories to the right . to further examine the character and distribution of escape events from @xmath15 , we have analyzed survival probabilities @xmath99 constructed from generated trajectories at fixed values of frequencies @xmath22 . for better comparison with the gaussian ra scenario , the summary of results is displayed in fig .
[ fig : survm88 ] with sets of data relating to @xmath100 and @xmath89 . the left and right panels of fig .
[ fig : survm88 ] present behavior of mfpt as the function of the barrier modulation parameter @xmath22 , the survival probability @xmath101 and exemplary cross - sections of the survival probability surface .
the insets depict behavior of the survival probability @xmath101 at short time scales ( small @xmath102 ) .
( left panel ) and @xmath89 ( right panel ) for @xmath81 .
mfpts curves ( upper panel ) , survival probability surfaces @xmath101 ( middle panel ) and sample cross - sections of @xmath101 surface ( lower panel ) for small ( @xmath84 ) , resonant ( @xmath86 ) and large ( @xmath103 ) @xmath22-values are depicted .
simulations parameters as in fig .
[ fig : skewedm88 ] .
note the log scale on z - axis ( middle panel ) and y - axis ( lower panel).,width=302 ] at low rates of the barrier switching process ( small @xmath22 ) , two distinct time scales of the barrier crossing events can be observed .
the fast time scale corresponds to passages over the barrier in its lower state , while the large time scale is pertinent to the slower process , i.e. the passages over the barrier in its higher energetic state .
this effect is well pronounced after the ra phenomenon sets up .
the presence of the two time scales for small @xmath22 and only one time scale for large @xmath22 explains the asymptotic behavior of mfpts .
namely , for low values of the switching rate @xmath22 , mfpt is an average value of mfpts over both configuration of the barrier . with increasing frequency @xmath22 the distinguishable time scales coalesce and disappear
this is due to the fact that the barrier changes its height multiple times during the particle s motion so that resulting mfpt describes kinetics over the average potential barrier .
we have examined the influence of various types of stochastic @xmath0-stable drivings on dynamic properties of a generic two - state model system .
although our primary interest was to understand the effects of lvy - type drivings with the stability index @xmath104 , the numerical analysis as applied in these studies remains valid for any set of parameters characterizing @xmath0-stable excitations .
the response to external or / and parametric perturbations has been examined by analyzing qualitative changes in system s dynamic behavior as expressed in the onset of resonant activation , stochastic resonance and dynamic hysteresis . due to the inherent symmetry of the potential ,
the sr quantifiers constructed at a given value of the stability index @xmath0 have been the same for @xmath65 .
on the other hand , the asymmetry of the driving noise has been shown to influence strongly the population of states and therefore has affected mostly the appearance and performance of the dynamic hysteresis .
the system efficiency has been described by standard sr measures , i.e. signal - to - noise ratio ( snr ) and spectral power amplification ( @xmath59 ) .
those quantifiers have been shown to behave in a typical way .
in other words , by tuning the noise intensity @xmath64 , the maximum in the snr(@xmath64 ) and @xmath71 could be detected , thus proving that stochastic resonance is a robust phenomenon which may be observed also in systems subjected to the action of impulsive jump - noise stochastic processes .
decrease in stability index @xmath0 results in larger jump - like excursions of the process @xmath4 and consequently causes weakening of sr .
out of two sr measures , the spectral power amplification has turned out to be more sensitive to variations of @xmath0 than snr . a more pronounced drop in system efficiency
has been observed for symmetric noises with @xmath41 .
the nonequilibrated , non - thermal lvy white noise affects also a paradigm scenario of escape kinetics . by numerically implementing the set up boundary conditions for the problem
, we have investigated the statistics of escape times over the dichotomously fluctuating barrier .
the manifestation of the ra phenomenon has been analyzed within a certain frequency @xmath22 regime pointing that by a continuous readjustment of the external noise paramaters the resonant activation can be either suppressed or re - induced .
lvy white noise with @xmath105 extends a standard brownian noise to a vast family of impulsive jump - like stochastic processes .
our studies document that dynamical systems driven by such sources can also benefit and display a noise - enhanced order .
the research has been supported by the marie curie tok cocos grant ( 6th eu framework program under contract no .
mtkd - ct-2004 - 517186 ) and european science foundation ( esf ) via ` stochastic dynamics : fundamentals and applications ' ( stochdyn ) program .
additionally , bd acknowledges the support from the foundation for polish science . | a standard approach to analysis of noise - induced effects in stochastic dynamics assumes a gaussian character of the noise term describing interaction of the analyzed system with its complex surroundings .
an additional assumption about the existence of timescale separation between the dynamics of the measured observable and the typical timescale of the noise allows external fluctuations to be modeled as temporally uncorrelated and therefore white .
however , in many natural phenomena the assumptions concerning the abovementioned properties of `` gaussianity '' and `` whiteness '' of the noise can be violated . in this context ,
in contrast to the spatiotemporal coupling characterizing general forms of non - markovian or semi - markovian lvy walks , so called lvy flights correspond to the class of markov processes which still can be interpreted as white , but distributed according to a more general , infinitely divisible , stable and non - gaussian law .
lvy noise - driven non - equilibrium systems are known to manifest interesting physical properties and have been addressed in various scenarios of physical transport exhibiting a superdiffusive behavior . here
we present a brief overview of our recent investigations aimed to understand features of stochastic dynamics under the influence of lvy white noise perturbations .
we find that the archetypal phenomena of noise - induced ordering are robust and can be detected also in systems driven by non - gaussian , heavy - tailed fluctuations with infinite variance . |
in this paper , we consider the family of pattern - replacement equivalence relations referred to as the `` indices and values adjacent '' case in @xcite .
each such equivalence is determined by a partition @xmath0 of a subset of @xmath1 for some @xmath2 . before introducing our results ,
we provide background definitions .
a * word * is a series of * letters * , where each * letter * has an integer value .
* example : * the word @xmath10 has letters @xmath11 , @xmath12 , @xmath13 , @xmath14 , and @xmath15 . a * permutation * of size @xmath16 is an @xmath16-letter word using each letter from @xmath11 to @xmath16 exactly once .
* example : * @xmath17 is a permutation of size @xmath18 .
@xmath8 denotes the set of permutations of size @xmath16 .
* example : * @xmath19
. the word @xmath20 * forms the permutation * @xmath21 if for some integer @xmath22 , @xmath23 as shorthand , we may just say that @xmath24 * forms * @xmath25 .
* example : * @xmath26 forms the permutation @xmath27 .
an even simpler example is that @xmath27 forms the permutation @xmath27 . a * replacement partition * is a partition of a subset of @xmath1 for some @xmath2 .
each replacement partition @xmath0 determines an equivalence relation which we refer to as the * @xmath0-equivalence*. a * hit * is a contiguous subword of a permutation that forms @xmath25 for some @xmath28 .
a * @xmath0-hit * is a contiguous subword of a permutation that forms a permutation in a partition @xmath0 of @xmath1 .
* example : * the subword @xmath26 of the permutation @xmath29 forms the permutation @xmath27 . given a pattern - replacement partition @xmath0 , we define the @xmath0-equivalence on @xmath8 in the following manner .
given a @xmath0-hit @xmath30 in a permutation @xmath24 that forms the permutation @xmath31 , we are allowed to rearrange the letters within @xmath30 in any way such that the resulting word forms some @xmath32 in the same part of @xmath0 as @xmath33 .
a * class * , or * equivalence class * , containing @xmath24 is the set of permutations that can be reached from @xmath24 by repeated rearrangements of the type just described .
if two permutations @xmath34 and @xmath35 are in the same equivalence class , we say that @xmath34 is * equivalent * to @xmath35 , or that @xmath36 .
let @xmath0 be a replacement partition .
let @xmath30 be a @xmath0-hit in a permutation @xmath24 such that @xmath30 forms a permutation in the @xmath37th part of @xmath0 . rearranging the letters in @xmath30 in any way such that @xmath30 still forms a permutation in the @xmath37th part of @xmath0 is referred to as a * @xmath0-rearrangement*. * example : * let @xmath38 .
then applying a @xmath39-rearrangement to the hit @xmath40 allows us to rearrange the hit either as @xmath41 or as @xmath42 .
this results in the permutations @xmath43 and @xmath44 respectively .
the * @xmath0-equivalence * is the reflexive - transitive closure of @xmath0-rearrangement . in other words ,
the permutations that can be reached from a permutation @xmath24 by means of repeated @xmath0-rearrangements are considered * equivalent * to @xmath24 under the @xmath0-equivalence .
the set of permutations equivalent to @xmath24 comprises an * equivalence class*. * example : * in figure [ fig1 ] we show a visual depiction of the equivalence class containing @xmath45 under the @xmath4-equivalence .
each of the above permutations are considered * equivalent * under the relation .
two permutations are connected by a line segment if one can be reached from the other by the rearrangement of a single @xmath46-hit .
for example , @xmath47 connects to @xmath45 because @xmath48 forms the permutation @xmath49 and thus can be rearranged to form the permutation @xmath49 , becoming @xmath50 .
[ fig1 ] under the @xmath4-equivalence . ]
we will refer to the type of equivalence relations studied in this project as * pattern - replacement * equivalence relations . in the literature
, they are referred to either as doubly adjacent pattern - replacement equivalence relations or as the `` indices and values adjacent '' case . in 2010 ,
linton , propp , roby , and west posed a number of open problems in the area of pattern - replacement equivalences @xcite .
five , in particular , have remained unsolved until now , the enumeration of equivalence classes under the @xmath51-equivalence , under the @xmath4-equivalence , under the @xmath5 equivalence , and under the @xmath52-equivalence . in section [ secapplications ] ,
we find formulas for three of the five equivalences and systems of representatives for the equivalence classes of the other two .
we generalize our results to hold for all replacement partitions of @xmath7 , as well as for an infinite family of other replacement partitions . in order to do this ,
we first introduce the notion of @xmath53-confluence in section [ secconfluence ] ; when @xmath53-confluence is satisfied by a partition @xmath0 , one can easily find a set of representative permutations for the equivalence classes in @xmath8 under the @xmath0-equivalence .
we construct a set of tools allowing us to often quickly prove that a pattern - replacement equivalence satisfies this confluence .
finally , we use the prior results in the section to present a formula for the number equivalence classes in @xmath8 under pattern - replacement equivalences satisfying certain conditions . in section [ secsc ]
, we use our results to completely characterize the equivalence classes in @xmath8 under the @xmath1-equivalence .
our results connect interestingly to stanley s results on the @xmath54-equivalence @xcite . for this entire paper ,
fix @xmath16 and @xmath2 be positive integers such that @xmath55 .
we will be considering equivalence classes of @xmath8 under doubly adjacent pattern - replacement equivalences involving patterns of size @xmath2 . before continuing , we provide some useful background on the notion of a binary relation being * confluent*. let @xmath56 be a binary relation on a finite set @xmath57 and @xmath58 be the reflexive - transitive closure of @xmath56 .
we say that @xmath56 is * confluent * if * there is no @xmath59 such that @xmath60 and @xmath61 ; * for each connected component @xmath62 of @xmath56 considered as a directed graph ( with @xmath57 as the set of nodes ) , there is a unique minimal node @xmath63 .
( i.e. , there is a unique node @xmath63 such that for all @xmath64 with @xmath65 , we have that @xmath66 . ) the following lemma , known as the diamond lemma , is well known , and is easy to prove .
let @xmath56 be a binary relation on a finite set @xmath57 such that @xmath67 is false for @xmath68 .
assume the following is true . 1
. there does not exist an infinite ( possibly repeating ) sequence @xmath69 such that for all @xmath37 , @xmath70 , and @xmath71 .
in other words , @xmath56 * terminates*. 2 . for all @xmath72 such that @xmath73 and @xmath74 , there exists @xmath75 such that @xmath76 and @xmath77 where @xmath78 denotes the reflexive - transitive closure of @xmath56 .
then , @xmath56 is confluent .
in this section , we introduce the notion of a replacement partition @xmath0 being @xmath53-confluent . when a partition is @xmath53-confluent , it is easy to characterize a set of permutations in @xmath8 exactly one of which is in a given equivalence class in @xmath8 under the @xmath0-equivalence .
such a permutation is referred to as a @xmath53-root permutation .
we provide three results which together allow one to often determine quickly that a pattern - replacement equivalence satisfies @xmath53-confluence ( proposition [ propsimpleconf ] , theorem [ thmcombineconf ] , theorem [ thmomni ] ) .
we then present a formula for the number of equivalence classes in @xmath8 under the @xmath0-equivalence when @xmath0 is @xmath53-confluent and satisfies certain conditions ( theorem [ thmconfcount ] ) .
let @xmath24 be a word , each letter of which has a distinct value .
the * tail size * of @xmath24 is the smallest positive integer @xmath22 such that the first @xmath22 letters of @xmath24 contain the @xmath22 smallest letter values in @xmath24 .
* example : * the tail size of @xmath79 is @xmath15 because @xmath80 contains @xmath11 , @xmath13 , @xmath12 , and @xmath15 .
let @xmath24 be a word , each letter of which has a distinct value .
* @xmath24 is * right leaning * if the tail size of @xmath24 is less than @xmath16 .
* @xmath24 is * left leaning * if the tail size of a written backwards version of @xmath24 is less than @xmath16 .
* @xmath24 is * omni leaning * if it is neither left leaning nor right leaning .
* example : * @xmath79 is right leaning because its tail size is @xmath15 .
consequently , writing it backwards to get @xmath81 gives us a left leaning permutation .
* example : * one can check that @xmath82 is omni leaning .
the sets of right leaning , left leaning , and omni leaning permutations in @xmath8 respectively are denoted by @xmath83 , @xmath84 , and @xmath85 .
note that the three sets do not intersect .
let @xmath86 , and @xmath30 be a hit in @xmath24 .
if @xmath30 comprises all of @xmath24 we say @xmath24 is * left polarized*. otherwise , let @xmath34 be the average value of the letters in @xmath30 , and let @xmath87 be @xmath24 except with the entire hit @xmath30 replaced by a single letter whose value is @xmath34 .
let @xmath35 be the letter in @xmath87 closest to @xmath34 in value out of the ( either one or two ) letters immediately adjacent to @xmath35 in @xmath87 .
if there is a tie , than pick the one on @xmath35 s left , although one can check it does not matter which we pick . if @xmath34 and @xmath35 are in increasing order in @xmath87 , then @xmath30 is * right polarized*. if @xmath34 and @xmath35 are in decreasing order in @xmath87 , then @xmath30 is * left polarized*. * example : * the hit @xmath88 in the permutation @xmath89 is left polarized . here ,
@xmath90 and thus @xmath91 .
the two letters immediately adjacent to @xmath34 in @xmath87 are @xmath11 and @xmath13 .
since @xmath13 is closer to @xmath18 than is @xmath11 , @xmath92 . in @xmath87 , @xmath34 and @xmath35
are in decreasing order , meaning @xmath30 is left polarized .
a hit @xmath30 is * backward * in a permutation @xmath24 if @xmath30 either * is both right polarized and left leaning , * or is both left polarized and right leaning , * or is omni leaning . otherwise , @xmath30 is * forward*. let @xmath0 be a partition of a subset @xmath93 of @xmath1 .
let @xmath94 .
then @xmath62 is a * straightening set * of @xmath0 if * for each part of @xmath0 containing at least one left leaning permutation , @xmath62 contains exactly one left leaning permutation from that part ; * for each part @xmath62 of @xmath0 containing at least one right leaning permutation , @xmath62 contains exactly one right leaning permutation from that part ; * and for each part of @xmath0 containing only omni leaning permutation , @xmath62 contains exactly one permutation from that part .
* example : * @xmath95 is a straightening set of @xmath96
. let @xmath62 be a straightening set of a partition @xmath0 .
a @xmath0-hit @xmath30 in @xmath97 is * @xmath53-straightened * if either * @xmath30 is forward and is in @xmath62 , * or @xmath30 is backward and is in @xmath62 , and there is no @xmath0-rearrangement that when applied to @xmath30 results in @xmath30 being forward . in turn ,
* @xmath53-straightening * a @xmath0-hit @xmath30 is defined as rearranging a @xmath0-hit to be @xmath53-straightened by means of a @xmath0-rearrangement .
* example : * let @xmath98 and @xmath99 .
consider the permutation @xmath100 .
the hit @xmath101 is right leaning but left polarized and is thus backwards .
consequently , @xmath53-straightening it rearranges it to be the forward hit @xmath102 , bringing us to the permutation @xmath103 . on the other hand
, @xmath104-straightening the hit @xmath102 in this new permutation does nothing at all .
let @xmath0 be a partition of a subset of @xmath1 and @xmath62 be a straightening set of @xmath0 .
let the binary relation @xmath56 on @xmath8 be such that for @xmath105 , @xmath106 exactly when @xmath87 can be reached from @xmath24 by @xmath53-straightening a hit in @xmath24 that was not already @xmath53-straightened .
we refer to @xmath56 as the * @xmath53-straightening operator*. let @xmath0 be a partition of a subset of @xmath1 and @xmath62 be a straightening set of @xmath0 .
the partition @xmath0 is said to be * @xmath62-confluent * if the @xmath53-straightening operator is confluent .
let @xmath0 be @xmath62-confluent .
a permutation @xmath24 is said to be a * @xmath53-root permutation * if it is the unique permutation in its equivalence class under the @xmath0-equivalence such that it can be reached from any other equivalent permutation by means of repeated @xmath53-straightenings .
* example : * it turns out that @xmath4 is @xmath107-confluent .
figure [ fig2 ] demonstrates this for a particular equivalence class in @xmath108 .
an arrow between two permutations means one can be reached from the other under the @xmath4-equivalence by a @xmath109-straightening .
observe how all paths lead to the permutation @xmath45 which is consequently the @xmath109-root permutation in the class .
[ fig2 ] is @xmath110-equivalent for equivalence class containing @xmath45 . ]
[ propsimpleconf ]
let @xmath0 be a partition of @xmath111 with parts @xmath112 .
let @xmath62 be the straightening set of @xmath0 containing the lexicographically smallest permutation from each part of @xmath0 .
suppose that for any two hits @xmath30 and @xmath113 in any permutation @xmath114 such that @xmath113 forms a permutation in @xmath115 , @xmath53-straightening @xmath30 leaves the letters in @xmath113 still in a contiguous subword forming a permutation in @xmath115 .
then @xmath0 is @xmath62-confluent .
let @xmath56 be the @xmath53-straightening operator .
since @xmath116 implies that @xmath35 is lexicographically smaller than @xmath34 , @xmath56 terminates . given @xmath114 ,
let @xmath117 be the permutation @xmath87 reached by @xmath53-straightening the @xmath0-hit beginning with @xmath118 in @xmath24 if such a @xmath0-hit exists and @xmath24 otherwise .
note that @xmath119 due to the restrictions we have imposed . thus by the diamond lemma , @xmath0 is @xmath62-confluent .
we state the following lemmas without proof for the sake of brevity .
[ lemoverlappinghit ] let @xmath114 and @xmath30 and @xmath113 be overlapping hits of size @xmath2 in @xmath24 .
then , @xmath30 and @xmath113 are both forward and either are both left leaning or are both right leaning .
[ lemletterorder ] let @xmath24 and @xmath87 be equivalent under the @xmath0-equivalence where @xmath0 partitions a subset of @xmath1 .
suppose @xmath120 for @xmath121 .
then , @xmath122 .
[ lemlocalorderchange ] let @xmath123 be such that @xmath87 can be reached from @xmath24 by a @xmath0-rearrangement ( where @xmath0 partitions a subset of @xmath1 ) .
let @xmath37 and @xmath124 be such that @xmath118 is not in @xmath30 .
then , @xmath125 .
now we introduce a slightly less obvious lemma .
[ lemignorebackwardhits ] let @xmath126 be a @xmath127-confluent partition
. then 1 .
and @xmath128-straightening any forward hit in a permutation results in a permutation with the same number of @xmath126-hits that are both backward and non-@xmath128-straightened .
2 . @xmath128-straightening a non-@xmath128-straightened backward hit in a permutation results in a permutation with one fewer @xmath126-hits that are both backward and non-@xmath128-straightened ; observe that ( a ) a backward @xmath126-hit can not overlap any other @xmath126-hit ( by lemma [ lemoverlappinghit ] ) , and ( b ) @xmath128-straightening one @xmath126-hit can not change whether another @xmath126-hit is right polarized or left polarized ( by lemma [ lemlocalorderchange ] ) . by ( a ) and ( b ) ,
@xmath128-straightening a forward @xmath126-hit in a permutation does not change the set of backward @xmath126-hits in the permutation . also by ( a ) and ( b )
, @xmath128-straightening a backwards non-@xmath128-straightened hit can not rearrange another @xmath128-straightened backward hit to become non-@xmath128-straightened .
let @xmath129 be partitions of @xmath130 respectively such that @xmath131 and @xmath132 .
we say that @xmath0 is a _ disjoint union _ of @xmath133 and @xmath134 if each part of @xmath0 is a disjoint union of some ( possibly empty ) part from @xmath133 and some ( possibly empty ) part from @xmath134 .
* example : * @xmath135 is a disjoint union of @xmath136 and @xmath137
. [ thmcombineconf ] let @xmath0 be a @xmath62-confluent partition of @xmath111 and @xmath138 be a @xmath139-confluent partition of @xmath140 .
let @xmath141 and @xmath126 be a disjoint union of @xmath0 and @xmath138 .
then @xmath126 is @xmath127-confluent .
let @xmath56 acting on @xmath8 be the @xmath128-straightening operator and @xmath58 be the reflexive - transitive closure of @xmath56 . since @xmath127 is clearly a straightening set of @xmath126 , it is sufficient to show that @xmath56 is confluent . to do this , we will use the diamond lemma .
we begin by proving that @xmath56 terminates .
suppose there exists an infinite chain @xmath142 . for each @xmath37
, let @xmath143 be a @xmath126-hit @xmath30 , the @xmath128-straightening of which brings one from @xmath144 to @xmath145 . without loss of generality ,
an infinite number of @xmath37 satisfy that @xmath143 is right leaning in @xmath144 ( since otherwise , an infinite number of @xmath37 would satisfy that @xmath143 is left leaning ) .
let @xmath22 be the number of backward @xmath126-hits in @xmath146 .
by lemma [ lemignorebackwardhits ] there are at most @xmath22 values of @xmath37 such that @xmath143 is backward in @xmath144 .
let @xmath124 be the greatest such @xmath37 .
observe that for @xmath147 , @xmath143 is forward in @xmath144 . by lemma [ lemoverlappinghit ] , the action of @xmath128-straightening a right leaning @xmath126-hit commutes with the action of @xmath128-straightening a left leaning @xmath126-hit .
recall that there exists an infinite number of @xmath37 such that @xmath143 is right leaning in @xmath144 , and thus also an infinite number of such @xmath37 where @xmath148 .
it follows that for a given value of @xmath149 , there exists an infinite chain @xmath150 such that @xmath151 is reached from @xmath152 by means of the @xmath128-straightening of a forward right leaning @xmath126-hit .
however , since such hits are also @xmath0-hits and @xmath0 is @xmath62-confluent , the pigeonhole principle implies this can not be true for @xmath153 , a contradiction
. we will now consider condition ( 2 ) of the diamond lemma .
suppose @xmath154 are such that @xmath73 , @xmath155 , @xmath34 reaches @xmath35 by means of the @xmath128-straightening of the @xmath126-hit @xmath30 , and @xmath34 reaches @xmath2 by means of the @xmath128-straightening of the @xmath126-hit @xmath113 .
if @xmath30 and @xmath113 do not overlap in @xmath34 , then @xmath128-straightening @xmath30 and then @xmath128-straightening @xmath113 in @xmath34 yields the same permutation as does @xmath128-straightening @xmath113 and then @xmath128-straightening @xmath30 .
suppose instead that @xmath30 and @xmath113 overlap .
by lemma [ lemoverlappinghit ] , we can assume without loss of generality that both are forward and right leaning . by the assumption that @xmath0 is @xmath62-confluent
, it follows that there exists @xmath156 such that @xmath157 and @xmath77 .
the result follows from the diamond lemma .
we now show that omni leaning permutations can be nicely inserted into arbitrary parts of partitions without screwing up the confluence properties of the partition .
[ thmomni ] let @xmath0 be a @xmath62-confluent partition of a subset @xmath93 of @xmath1 .
let @xmath158 such that @xmath159 is not in @xmath93 . 1 .
let @xmath160 be @xmath0 except with @xmath159 added to one of the parts of @xmath0 .
then @xmath160 is @xmath62-confluent .
2 . let @xmath160 be @xmath0 except with @xmath159 added as a part of size @xmath11 .
then @xmath160 is @xmath161-confluent .
we only bother to prove ( 1 ) , as ( 2 ) is straightforward .
suppose @xmath160 is @xmath0 except with @xmath159 added to one of the parts of @xmath0 . let @xmath56 acting on @xmath8 be the @xmath162-straightening operator .
we will show that @xmath56 satisfies the conditions of the diamond lemma .
any @xmath160-hit forming @xmath159 is backwards .
thus , by lemma [ lemignorebackwardhits ] , if there exists an infinite chain @xmath163 , then there is some @xmath22 such that @xmath164 and no @xmath160-hits forming @xmath159 are used to go from @xmath144 to @xmath145 in the chain for @xmath165 .
but since @xmath0 is @xmath62-confluent , no such chain can exist .
thus @xmath56 terminates . observe that a @xmath160-hit forming @xmath159 can not overlap any other @xmath160-hit ( lemma [ lemoverlappinghit ] ) . since @xmath0 is @xmath62-confluent
, it follows that condition ( 2 ) of the diamond lemma is satisfied by @xmath56 .
thus @xmath56 is confluent .
[ thmconfcount ] let @xmath0 be a @xmath62-confluent partition of @xmath1 .
suppose that for @xmath0-hits @xmath30 and @xmath113 in @xmath114 to overlap implies that at least one of @xmath30 and @xmath113 is @xmath166-straightened .
pick @xmath167 and let @xmath168 where @xmath169 is the number of parts in @xmath0 .
then the number of equivalence classes in @xmath8 under the @xmath0-equivalence , which we will denote as @xmath170 , is @xmath171 let @xmath172 be the number of @xmath86 such that @xmath24 contains a @xmath0-hit @xmath30 such that the first letter of @xmath30 is in position @xmath37 in @xmath24 and @xmath30 is not @xmath53-straightened in @xmath24 . by the inclusion - exclusion principle , @xmath173 , |s|=j } | \bigcap_{i \in s } t_i |.\ ] ] note that for any @xmath0-hit @xmath30 in @xmath114 such that @xmath30 is not @xmath53-straightened , there are exactly @xmath174 possibilities for @xmath30 given the position of its first letter , the value of its smallest letter , and whether @xmath30 is right polarized or left polarized .
let us calculate the number of ways to construct @xmath124 non - overlapping @xmath0-hits and place them in a permutation so that none of them are @xmath53-straightened .
let @xmath175 be the set containing exactly the @xmath124 letters that will each be the smallest letter in one of our @xmath124 @xmath0-hits .
we claim there are @xmath176 possibilities for @xmath175
. since no two non-@xmath53-straightened @xmath0-hits can overlap , the possibilities are exactly the sets of @xmath124 integers from @xmath11 to @xmath177 such that no two of them are within @xmath178 of each other in value .
these sets , in turn , are in bijection with binary words containing @xmath179 letters , @xmath124 of which are ones . indeed , given such a word @xmath24
, we may substitute each @xmath11 with the letter @xmath11 followed by @xmath178 copies of the letter @xmath13 , bringing us to @xmath87 .
we then choose to stick the letter @xmath37 in @xmath175 exactly when the @xmath37th letter of @xmath87 is @xmath11 .
the inverse bijection is easy to see .
given @xmath175 , the relative positions of the first letters of each of the @xmath124 @xmath0-hits along with the letters not in any of the @xmath124 @xmath0-hits can be chosen in @xmath180 ways .
this , in turn , determines which hits are right polarized and left polarized .
given this choice , the arrangement of each individual hit can be chosen from @xmath174 possibilities .
hence @xmath181 , |s|=j } | \bigcap_{i \in s } t_i | = { n - j(c-1 ) \choose j } ( n - j(c-1 ) ) ! ( n!-|p|)^j.\ ] ] inserting this into the formula found by the inclusion - exclusion principle and slightly rearranging yields the desired result .
in this section , we apply our results from section [ secconfluence ] to all replacement partitions of @xmath7 ( theorem [ thms_3case ] ) , on the way solving three open problems of linton , propp , roby , and west @xcite .
we then apply our results to a particularly interesting infinite family of pattern - replacement equivalences ( theorem [ thmeasyinf ] ) .
[ thms_3case ] ( a ) let @xmath0 be a partition of @xmath7 .
then there exists a straightening set @xmath62 of @xmath0 such that @xmath0 is @xmath62-confluent .
( b ) furthermore , if @xmath182 , @xmath183 , and @xmath184 are not all in the same part of @xmath0 and @xmath49 , @xmath185 , and @xmath186 are not all in the same part of @xmath0 , then the number of equivalence classes in @xmath8 under the @xmath0-equivalence can be established with theorem [ thmconfcount ] . by proposition [ propsimpleconf ] , @xmath187 is @xmath188-confluent , @xmath189 is @xmath3-confluent ) , @xmath190 is @xmath191-confluent , and @xmath192 is @xmath193-confluent .
( these are all the partitions of @xmath194 ; note that @xmath195 is empty . ) note that only in the final of these cases does the partition not satisfy the requirements for theorem [ thmconfcount ] to be applied . by symmetry
, similar statements can be said for each partition of @xmath196 .
since every partition of @xmath7 is the disjoint union of a partition of @xmath196 and a partition of @xmath194 , ( a ) follows from theorem [ thmcombineconf ] .
since two hits of size three of size @xmath12 in a permutation @xmath24 can only overlap if either both are in @xmath196 or both are in @xmath194 ( lemma [ lemoverlappinghit ] ) , ( b ) follows as well .
note that theorem [ thms_3case ] resolves three open problems , the enumeration of equivalence classes under each of the @xmath3-equivalence , the @xmath4-equivalence , and the @xmath197-equivalence .
it also makes progress on two additional open problems , the enumerations of equivalence classes under the @xmath198-equivalence and @xmath52-equivalence ; in each case , theorem [ thms_3case ] allows one to characterize a set of permutations ( @xmath53-root permutations ) , exactly one of which appears in each equivalence class in @xmath8 .
[ thmeasyinf ] let @xmath199 and @xmath200 such that @xmath201 for all @xmath202 .
let @xmath0 be the partition @xmath203 .
the number of equivalence classes in @xmath8 under the @xmath0-equivalence is @xmath171 it is trivial that @xmath204 is @xmath205-confluent and that @xmath206 is @xmath207-confluent . by theorem [ thmcombineconf ] , @xmath208 ( which we will refer to as @xmath0 ) is @xmath209-confluent . observe that if two @xmath0-hits overlap , they either form @xmath144 and @xmath210 for some @xmath202 , or they form @xmath152 and @xmath211 for some @xmath202 ( by lemma [ lemoverlappinghit ] ) . in either case , both @xmath0-hits are forward and thus @xmath212-straightened .
thus , pretending that all the permutations in @xmath1 not in @xmath0 are each in a part of size one of @xmath0 , we can apply theorem [ thmconfcount ] to obtain the desired formula .
observe that theorem [ thmeasyinf ] counts equivalence classes in @xmath8 under the @xmath0 equivalence for at least @xmath213 choices of @xmath214 for a given @xmath2 .
this is easy to see by allowing only half of the omni leaning permutations in @xmath1 to be assigned to the @xmath144 and allowing only the other half to be assigned to the @xmath152 ( so the choices for @xmath144 come from exactly half of the permutations in @xmath1 , and similarly for @xmath152 ) .
it should be surprising that equivalence relations with different class structures often have the same number of classes in @xmath8 .
for example , by theorem [ thms_3case ] , for all partitions @xmath0 of @xmath7 of the form @xmath215 the @xmath0-equivalence yields the same enumeration . yet it is easy to see that the structure of the equivalence classes created under such equivalence differs greatly between equivalence relations .
in fact , theorem [ thmeasyinf ] provides an infinite number of examples of this phenomenon .
one infinite family of equivalence relations in particular , is worth studying in greater depth .
the @xmath1 equivalence has strictly fewer equivalence classes than any equivalence defined using another replacement partition of @xmath1 .
consequently , enumerating the equivalence classes under the @xmath1-equivalence is an important tasks . in this section ,
we make progress on this by completely characterizing the equivalence classes . in the case of the @xmath54-equivalence ( @xmath216 ) ,
stanley @xcite enumerated and characterized the equivalence classes in @xmath8 and found representative elements for each one .
surprisingly , in our work , we will find a set of representative elements ( the @xmath217-root permutations ) that in the case of the @xmath54-equivalence differ from the already found one .
we will refer to the permutation @xmath218 as @xmath219 and to the permutation @xmath220 as @xmath221 .
* example : * when @xmath222 , the @xmath1-equivalence is the @xmath223-equivalence , @xmath224 , and @xmath225 .
[ lemstayinr ] for any two @xmath1-hits @xmath30 and @xmath113 in a permutation @xmath114 such that @xmath113 forms a permutation in @xmath111 , @xmath226-straightening @xmath30 leaves the letters in @xmath113 still forming a permutation in @xmath111 .
this is because rearranging the letters of a contiguous subword of a right leaning @xmath1-hit to be in increasing order will still yield a right leaning @xmath1-hit .
[ lemhitmaintaining ] for any two @xmath1-hits @xmath30 and @xmath113 in a permutation @xmath114 , @xmath227-straightening @xmath30 leaves the letters in @xmath113 still forming a permutation in @xmath1 by lemma [ lemoverlappinghit ] , we only need to consider the case where @xmath30 and @xmath113 are either both left leaning or both right leaning . without loss of generality , we can assume they are both right leaning ( and since they overlap , are both in @xmath111 ) .
the lemma thus follows from lemma [ lemstayinr ] .
[ thmscrep ] the @xmath1-equivalence is @xmath228-confluent .
we provide two proofs of this . by lemma [ lemstayinr ] and proposition [ propsimpleconf ] ,
the @xmath111-equivalence is @xmath229-confluent . by symmetry ,
the @xmath140-equivalence is @xmath230-confluent .
it follows from theorem [ thmcombineconf ] that the @xmath231-equivalence is @xmath232-confluent .
it follows from theorem [ thmomni ] that the @xmath1-equivalence is @xmath233-confluent .
the following alternative proof is an interesting application of the diamond lemma .
let @xmath56 be the @xmath234-straightening operator .
observe that @xmath227-straightening a @xmath1-hit in a permutation @xmath24 does not take any other @xmath1-hits that were @xmath235-straightened and rearrange them to not be so .
( observation ( 1 ) ) as a consequence of observation ( 1 ) , @xmath236-straightening a @xmath1-hit in a permutation @xmath24 results in a permutation @xmath87 with strictly more @xmath236-straightened @xmath1-hits .
since the number of such hits a permutation can have is bounded , @xmath56 terminates . as an additional consequence of observation ( 1 ) and lemma [ lemhitmaintaining ] , the actions of @xmath234-straightening the @xmath1-hit beginning in position @xmath37 commutes with the action of @xmath234-straightening the @xmath1-hit in position @xmath237 for
any given @xmath37 and @xmath237 .
thus @xmath56 satisfies condition ( 2 ) of the diamond lemma . by the diamond lemma , @xmath56 is confluent .
[ thmsumclasses ] for a given equivalence class of @xmath8 under the @xmath1-equivalence , let @xmath24 be the unique @xmath238-root permutation in the class .
( such a @xmath24 exists by theorem [ thmscrep ] . )
then @xmath24 is of the form @xmath239 such that 1 .
each @xmath240 is an increasing or decreasing word of consecutive integers .
2 . for all @xmath241
, @xmath25 is of the form @xmath242 such that @xmath243 .
pick @xmath244 such that @xmath245 , such that each @xmath240 is an increasing or decreasing word of consecutive integers , and such that @xmath22 is minimal .
( a little thinking shows that there is only one such pick . )
observe that in @xmath24 , there is no @xmath1-hit @xmath30 containing letters from both @xmath240 and @xmath246 for any @xmath37 .
if there were , then the @xmath1-hit would have to be @xmath247-straightened ( as are all @xmath1-hits in @xmath24 ) .
but if this were the case , then @xmath240 and @xmath246 would concatenate to form an increasing or decreasing word of consecutive integers , a contradiction because we picked @xmath244 to minimize @xmath22 . by lemma [ lemhitmaintaining ] and theorem [ thmscrep ]
, it follows that for @xmath241 , there are no @xmath1-hits containing both letters from @xmath240 and from @xmath246 for any @xmath37 . if there were , then since repeated @xmath227-straightenings can bring us from @xmath25 to @xmath24 , @xmath24 would have to contain a @xmath1-hit containing both letters from @xmath240 and @xmath246 , a contradiction . given @xmath114 ,
we define @xmath248 to be the unique @xmath240 defined in the proof of theorem [ thmsumclasses ] .
observe that stanley found a result similar to theorem [ thmsumclasses ] in the case of the @xmath249-equivalence @xcite . by the previous theorem ,
all that remains in characterizing the equivalence classes in @xmath8 under the @xmath1-equivalence is to characterize the equivalence class containing @xmath250 and the one containing @xmath251 for all @xmath33 ; by symmetry , it is sufficient to characterize the equivalence class containing the identity permutation in @xmath8 , @xmath252 .
let @xmath24 be a word of size @xmath16 , each letter of which has a distinct value .
recall that the * tail size * of @xmath24 is the smallest positive integer @xmath22 such that the first @xmath22 letters of @xmath24 contain the @xmath22 smallest letter values in @xmath24 .
we say that if @xmath24 is empty , it has no * irreducible blocks*. otherwise , let @xmath22 be the tail size of @xmath24 , let @xmath34 be a word made of the first @xmath22 letters of @xmath24 , and let @xmath35 be a word made of the final @xmath253 letters of @xmath24 .
we say that the * irreducible blocks * of @xmath24 are @xmath254 and @xmath255 for @xmath256 satisfying that @xmath257 exists .
a permutation is @xmath2-toothed if the following is true
. 1 . for each @xmath258 , @xmath259 ( where @xmath260 is the number of letters in @xmath258 ) ; 2 .
there is some sequence of consecutive integers @xmath261 such that @xmath262 * example : * figure [ fig3 ] is a visual depiction of the irreducible block decomposition of the permutation @xmath263 .
note that @xmath264 , @xmath265 , @xmath266 , and @xmath267 .
consequently , @xmath24 is @xmath12-toothed , @xmath15-toothed , and @xmath268-toothed .
however , @xmath24 is not @xmath13-toothed since @xmath269 . [ fig3 ] .
] [ thmtooth ] let @xmath114 .
then @xmath270 exactly if @xmath24 is @xmath2-toothed . ] .
observe that every @xmath2-toothed permutation in @xmath8 except for the identity contains at least one @xmath1-hit not forming @xmath219 .
also note that @xmath271-straightening a @xmath1-hit in a @xmath2-toothed permutation in @xmath8 is exactly the same as rearranging the @xmath1-hit to form @xmath219 . finally , observe that rearranging a @xmath1-hit in a @xmath2-toothed permutation to form @xmath219 yields another @xmath2-toothed permutation .
thus it follows from theorem [ thmscrep ] that every @xmath2-toothed permutation in @xmath8 is equivalent to @xmath272 under the @xmath1-equivalence .
it is straightforward to check that every permutation in @xmath8 that is equivalent to @xmath272 under the @xmath1-equivalence must also be @xmath2-toothed .
let @xmath273 be the set of @xmath2-toothed permutations in @xmath8 .
let @xmath274 be the set @xmath275 .
it follows from theorem [ thmtooth ] and theorem [ thmsumclasses ] that the size of any equivalence class under the @xmath1-equivalence is a product of elements of @xmath274 .
consequently , it is interesting to study @xmath276 .
we do so for @xmath222 . for @xmath277 , @xmath278 is the value of the coefficient of @xmath279 in the generating function @xmath280 let @xmath281 be the number of permutations in @xmath8 comprising only irreducible blocks of size @xmath282
. if the final irreducible block is of size @xmath11 , the final letter of such a permutation is @xmath16 .
if the final irreducible block is of size @xmath13 , the final two letters are @xmath283 . if the final irreducible block is of size @xmath12 , the final three letters are @xmath284 , or @xmath285 , or @xmath286 .
thus @xmath287 for @xmath288 , @xmath289 , and @xmath290 .
thus if @xmath291 is the generating function satisfying @xmath292f = f_n$ ] , then @xmath293 . rearranging yields
@xmath294 now consider the permutations in @xmath8 comprising only of irreducible blocks of size @xmath282 but containing no consecutive irreducible blocks whose sizes add to @xmath12 .
for @xmath277 , it is easy to see that such a permutation can only exist when @xmath16 is even and that it is of the form @xmath295 ( comprising only of irreducible blocks of size two ) .
thus for @xmath277 , @xmath278 is the value of the coefficient of @xmath279 in the generating function @xmath296
we present three directions of future work . 1 .
is it interesting to extend the notion of confluence introduced in this paper to the case where instead of @xmath53-straightening @xmath0-hits , one @xmath53-straightens contiguous subwords of some fixed size .
when this size of @xmath2 , this reduces to the notion of @xmath53-straightening we have used in this paper .
2 . is there a formula counting the number of classes in @xmath8 under the @xmath1-equivalence ?
such a formula would provide a lower bound for the number of equivalence classes under an arbitrary doubly adjacent pattern - replacement equivalence .
3 . is there a formula for @xmath276 ? such a formula is of interest since the size of any equivalence class under the @xmath1-equivalence is a product of elements of @xmath297 .
william kuszmaul and ziling zhou , _ equivalence classes in @xmath8 for three families of pattern - replacement relations .
_ mit primes , 2013
. http://web.mit.edu/primes/materials/2012/kuszmaul-zhou.pdf .
under review by _ electronic journal of combinatorics_. a. pierrot , d. rossin , j. west , _ adjacent transformations in permutations . _ fpsac 2011 proceedings , discrete math .
proc . , 2011 .
http://www.dmtcs.org/dmtcs-ojs/index.php/proceedings/article/view/dmao0167/3638 .
s. linton , j. propp , t. roby , j. west , _ equivalence relations of permutations generated by constrained transpositions .
_ dmtcs proceedings , north america , july 2010 .
http://www.dmtcs.org/dmtcs-ojs/index.php/proceedings/article/view/dman0168 , also arxiv:1111.3920v1 . | in this paper , we consider the family of pattern - replacement equivalence relations referred to as the `` indices and values adjacent '' case .
each such equivalence is determined by a partition @xmath0 of a subset of @xmath1 for some @xmath2 . in 2010 ,
linton , propp , roby , and west posed a number of open problems in the area of pattern - replacement equivalences .
five , in particular , have remained unsolved until now , the enumeration of equivalence classes under the @xmath3-equivalence , under the @xmath4-equivalence , under the @xmath5 equivalence , and under the @xmath6-equivalence .
we find formulas for three of the five equivalences and systems of representatives for the equivalence classes of the other two .
we generalize our results to hold for all replacement partitions of @xmath7 , as well as for an infinite family of other replacement partitions .
in addition , we characterize the equivalence classes in @xmath8 under the @xmath1-equivalence , finding a generalization of stanley s results on the @xmath9-equivalence . to do this ,
we introduce a notion of confluence that often allows one to find a representative element in each equivalence class under a given equivalence relation . using an inclusion - exclusion argument
, we are able to use this to count the equivalence classes under equivalence relations satisfying certain conditions . |
ebola virus is currently affecting several african countries , mainly guinea , sierra leone , and liberia .
ebola was first discovered in 1976 in the democratic republic of the congo near the ebola river , where the disease takes its name @xcite . since then
, ebola outbreaks have appeared sporadically in africa .
the virus , previously known as ebola haemorrhagic fever , is the deadliest pathogens for humans .
the early signs and symptoms of the virus include a sudden onset of fever and intense weakness and headache . over time , symptoms become increasingly severe and include diarrhoea , raised rash , internal and external bleed in ( from nose , mouth , eyes and anus ) . as the virus spreads through the body , it damages the immune system and organs @xcite .
ebola virus is transmitted to an initial human by contact with an infected animal s body fluid . on the other hand
, human - to - human transmission can take place with direct contact ( through broken skin or mucous membranes in , for example , the eyes , nose , or mouth ) with blood or body fluids of a person who is sick with or has died from ebola .
it is also transmitted indirectly via exposure to objects or environment contaminated with infected secretions @xcite .
mathematical models are a powerful tool for investigating human infectious diseases , such as ebola , contributing to the understanding of the dynamics of the disease , providing useful predictions about the potential transmission of the disease and the effectiveness of possible control measures , which can provide valuable information for public health policy makers @xcite .
epidemic models date back to the early twentieth century , to the 1927 work by kermack and mckendrick , whose model was used for modelling the plague and cholera epidemics .
in fact , such epidemic models have provided the foundation for the best vaccination practices for influenza @xcite and small pox @xcite .
currently , the simplest and most commonly implemented model in epidemiology is the sir model .
the sir model consists of three compartments : susceptible individuals @xmath0 , infectious individuals @xmath1 , and recovered individuals @xmath2 @xcite . when analysing a new outbreak , the researchers usually start with the sir model to fit the available outbreak data and obtaining estimates for the parameters of the model @xcite .
this has been the case for the modelling of the spreading mechanism of the ebola virus currently affecting several african countries @xcite . for more complex mathematical models , with more than three state variables ,
see @xcite . in our previous works @xcite , we used parameters identified from the recent data of the world health organization ( who ) to describe the behaviour of the virus . here
we focus on the mathematical analysis of the early detection of the ebola virus . in section [ sec2 ]
, we briefly recall the analysis study of the sir model that we presented in our previous study of the description of the behaviour of ebola virus @xcite . in section [ sec3 ] ,
we add to the basic model of section [ sec2 ] the demographic effects , in order to provide a description of the virus propagation closer to the reality .
this gives answer to an open question posed in remark 1 of @xcite and at the end of @xcite .
our aim in studying the model with vital dynamics is to provide useful predictions about the potential transmission of the virus .
we also consider an induced death rate for the infected individuals .
after numerical simulations , in section [ sec4 ] we control the propagation of the virus in order to minimize the number of infected individuals and the cost of vaccination .
we end with section [ sec : conc ] of conclusions .
in this section , we present and briefly discuss the properties of the system of equations corresponding to the basic sir ( susceptible infectious recovery ) model , which has recently been used in @xcite to describe the early detection of ebola virus in west africa . in the formulation of the basic sir model ,
we assume that the population size is constant and any person who has completely recovered from the virus acquired permanent immunity .
moreover , we assume that the virus has a negligibly short incubation period , so that an individual who contracts the virus becomes infective immediately afterwards .
these assumptions enables us to divide the host population into three categories , * @xmath3 for susceptible : denotes individuals who are susceptible to catch the virus , and so might become infectious if exposed ; * @xmath4 for infectious : denotes infectious individuals who are able to spread the virus through contact with the susceptible category ; * @xmath5 for recovered : denotes individuals who have immunity to the infection , and consequently do not affect the transmission dynamics in any way when in contact with other individuals .
the model is described mathematically by the following system of non - linear differential equations : @xmath6 \dfrac{di(t)}{dt } = \beta s(t)i(t)- \mu i(t),\\[0.2 cm ] \dfrac{dr(t)}{dt } = \mu i(t ) , \end{cases}\ ] ] where @xmath7 is the infection rate and @xmath8 is the recovered rate .
the initial conditions are given by @xmath9 we can see that @xmath10 = 0 $ ] , that is , the population size @xmath11 is constant during the period under study : @xmath12 for any @xmath13 , which is far from the reality .
in the basic sir model , we ignore the demographic effects on the population . in this section , we study a variant of the basic model by considering vital dynamics , that is , by adding the birth and death rates into the model
. moreover , we increase the death rate of the infectious class by considering an induced death rate associated to the infected individuals .
such model is new in the ebola context @xcite .
if we expand the sir model by including the demographic effects , assuming a constant rate of births @xmath14 , an equal rate of deaths @xmath15 per unit of time , and an induced death rate @xmath16 , then the mathematical model is described by the following system of differential equations : @xmath17 \dfrac{di(t)}{dt } = \beta s(t)i(t ) - \mu i(t ) - \left(\gamma + \gamma_{i } \right ) i(t),\\[0.2 cm ] \dfrac{dr(t)}{dt } = \mu i(t ) - \gamma r(t ) .
\end{cases}\ ] ] figure [ fig : sir_bd2 ] shows the compartment diagram of the sir model with vital dynamics , that is , with demographic ( birth and death ) effects , and an induced death rate . and @xmath15 and an induced death rate @xmath16 .
] firstly , we start by analysing the equations of the model that serve as the basis for the propagation dynamics of the ebola virus with death and birth rates . as we shall see ( cf .
theorem [ thm2 ] ) , the dynamics are determined by the basic reproduction number @xmath18 an equilibrium point @xmath19 of satisfies , by definition , @xmath20 @xmath21
@xmath22 there are two biologically meaningful equilibrium points : if @xmath23 , then there is no disease for the population , and the equilibrium point is called a disease - free equilibrium ; otherwise , if @xmath24 , then the equilibrium point is called endemic . by adding equations and , we obtain that @xmath25 then , @xmath26 from and , we get @xmath27 therefore , or @xmath23 or @xmath28 if @xmath23 , then from we obtain that @xmath29 .
it follows from that @xmath30 .
we just proved that there is a virus free equilibrium @xmath31 given by @xmath32 if holds , then there is another equilibrium with @xmath33 by substituting into , we obtain that @xmath34 and , using in , we get @xmath35.\ ] ] we just obtained the endemic equilibrium @xmath36 given by @xmath37 where the expressions of @xmath38 , @xmath39 and @xmath40 are given by .
next result summarizes what we have obtained so far .
[ thm2 ] let @xmath41 be the basic reproduction number defined by .
if @xmath42 , then the disease free equilibrium @xmath43 of the virus is obtained , which corresponds to the case when the virus dies out ( no epidemic ) .
if @xmath44 , then the equilibrium @xmath45 of the virus is obtained , in agreement with expressions , and the virus is able to invade the population ( endemic equilibrium ) .
we now present a simulation of the model , taking into account the real outbreak of ebola virus occurred in guinea in 2015 and by using the world health organization ( who ) data .
precisely , the epidemic data used in our study is borrowed from the who web site @xcite .
the birth rate @xmath46 and death rate @xmath47 of the model are obtained from the specific statistical study of the demographic of guinea in 2015 @xcite .
the parameters @xmath48 , @xmath49 , @xmath15 and @xmath16 are obtained by identification by using the real data of who . to estimate the parameters of the model
, we adapted the initialisation of @xmath1 with the reported data of who by fitting the real data of confirmed cases of infectious in guinea . the result of fitting
is shown in figure [ i_sml_data_model_induced ] .
the comparison between the curve of infectious obtained by our simulation and the reported data of confirmed cases by who shows that the mathematical model fits well the real data by using @xmath50 as the transmission rate , @xmath51 as the infectious rate , @xmath52 as the recovered rate , and @xmath53 as the induced death rate . by comparing the value of @xmath16 with the death rate @xmath15
, we remark that @xmath54 .
the initial susceptible , infected and recovered populations , are given by @xmath55 respectively .
the choice of @xmath56 is in agreement with the data shown in figure [ i_sml_data_model_induced ] of the who data . by using the value of guinea s population , which is estimated at @xmath57 in @xmath58 , and the number of confirmed infectious cases
( obtained from who ) , the initialization of @xmath56 corresponds to the number of infected divided by the total number of population .
then , in reality , the initial number of infected , in the period between january 2015 and march 2015 , is given by @xmath59 , that is , the number of confirmed infectious cases represents @xmath60 of the total of population .
figure [ i_sml_data_model_induced ] presents the curve of infectious individuals @xmath4 simulated by using and obtained from the who real data . ] the evolution of susceptible , infected and recovered groups over time , is shown in figure [ fig : bd : sir2 ] .
+ we also studied numerically the equilibria , by solving numerically the sir model with the same parameters and the same initialization .
figure [ bd_fig1:sir2 ] shows the evolution of the susceptible individuals @xmath3 over time .
we see that the oscillations in the numbers of the three compartments damp out over time , eventually reaching an equilibrium . in our mathematical analysis of the model
, we found that the equilibrium @xmath38 is computed theoretically by .
when we calculate the value of this theoretical result , we find @xmath61 , which is equal to the @xmath38 computed by the numerical resolution of the model . figure [ bd_fig2:sir2 ] shows the evolution of the infected individuals @xmath4 over time .
we know that the equilibrium @xmath39 is given by .
when we calculate the value of this theoretical result , we find @xmath62 , which agrees with the value computed by the numerical resolution of the model . figure [ bd_fig3:sir2 ] shows the evolution of the recovered individuals @xmath5 over time .
the equilibrium @xmath40 is in this case computed theoretically by , which coincides the value @xmath63 computed numerically .
+ the fact that the reached equilibrium @xmath64 computed theoretically coincides with the value found by the numerical simulation of the model , is a validation of our study of the sir model with vital dynamics and induced death rate , which describes well the currently detection of ebola virus in guinea .
nowadays there are several trial vaccinations against ebola .
one was already applied in guinea and seems highly effective @xcite . in this section
, we present a strategy for the control of the virus , by introducing into the model a control @xmath65 representing the vaccination rate at time @xmath66 .
the control @xmath65 is the fraction of susceptible individuals being vaccinated per unit of time , taking values on the interval @xmath67 $ ] .
then , the mathematical model with control is given by the following system of non - linear differential equations : @xmath68 \dfrac{di(t)}{dt } = \beta s(t)i(t ) - \mu i(t ) - ( \gamma + \gamma_i ) i(t),\\[0.2 cm ] \dfrac{dr(t)}{dt } = \mu i(t ) -
\gamma r(t ) +
u(t ) s(t ) . \end{cases}\ ] ]
the goal of the strategy is to reduce the infected individuals and the cost of vaccination .
precisely , the optimal control problem consists of minimizing the objective functional @xmath69 dt,\ ] ] where @xmath70 is the control variable , @xmath71 $ ] , which represents the vaccination rate at time @xmath66 , and the parameters @xmath72 and @xmath73 denote , respectively , the weight on cost and the duration of the vaccination program . in the quadratic term of , @xmath72 is a positive weight parameter associated with the control @xmath65 and the square of the control variable reflects the severity of the side effects of the vaccination .
one has @xmath74 , where @xmath75 \right\}\ ] ] is the admissible control set with @xmath76 .
the existence of an optimal solution follows from theorem 2.1 of @xcite ( see also the study of the spread of influenza a ( h1n1 ) @xcite by using the sir model ) . in our study of the control of the virus , we use the parameters defined in section [ sec4_subsec2 ] . for the numerical simulations of the optimal control problem
, we have used the solver @xcite , which is based on a multiple shooting method , including automatic differentiation and based ultimately on the semidirect multiple shooting algorithm of bock and pitt @xcite .
the solver comes as a self - contained public domain software environment , written in , for automatic control and dynamic optimization . because is a nonlinear control system , functional is quadratic in the control and linear in the phase variable @xmath1 , it is not clear that the numerically found solution through to our optimal control problem gives the global minimum to the functional , or only a local one .
indeed , with the optimal control problem is approximated by a finite dimensional optimization problem , which is then solved by techniques from mathematical programming .
this only gives a candidate for local minimizer .
because of this , we have also used a dynamic programming approach and checked our results by solving it with , which is a / toolbox for optimal control developed since 2014 in the framework of the inria
saclay initiative for an open source optimal control toolbox , being supported by the team `` commands '' ( see http://bocop.org ) . while implements a local optimization method , the package implements a global optimization method .
similarly to the dynamic programming approach , the optimal control problem is solved by in two steps : first the hamilton jacobi bellman equation satisfied by the value function of the problem is solved ; then the optimal trajectory is simulated from any chosen initial condition @xcite .
the obtained results through are coincident with those obtained through and , because of that , we claim to have found the global minimum to the problem .
+ figure [ fig6 ] shows the significant difference in the number of susceptible , recovered , and infected individuals with and without control . in figure
[ cntrlbd_s ] , we see that the number of susceptible @xmath0 , in case of optimal control , decreases faster during the vaccination campaign .
it reaches 34.78% at the end of the campaign against 44.12% in the absence of optimal control .
figure [ cntrlbd_r ] shows that the number of recovered individuals increases rapidly .
the number @xmath77 increases more rapidly in case of control than without control . in figure
[ cntrlbd_i ] , the time - dependent curve of infected individuals shows that the peak of the curve of infected individuals is less important in case of control .
in fact , the maximum value on the infected curve @xmath1 under optimal control is 16.3% , against 47.69% without any control ( see figure [ cntrlbd_i ] ) .
the other important effect of control , which we can see in the same curve , is the period of infection , which is less important in case of control of the virus .
the value of the period of infection is @xmath78 days in case of optimal control , versus more than @xmath79 days without vaccination .
this shows the efficiency of vaccination in controlling ebola . for problem with initial conditions , @xmath80 $ ] , @xmath81 days , and @xmath82.[cntrlbd_u ] ] , in case of optimal control _ versus _ without control .[cntrlbd_n_induced ] ] in conclusion , one can say that figure [ fig6 ] shows the effectiveness of optimal vaccination in controlling ebola .
figure [ cntrlbd_u ] gives a representation of the optimal control @xmath65 ; while figure [ cntrlbd_n_induced ] shows the evolution of the number of total population @xmath83 over time .
we see that the total number of population is bigger in case of vaccination ( less people dying ) .
mathematical modelling of the detection of a virulent virus such ebola is a powerful tool to understand the dynamics of the propagation of the virus in a population .
the main aim is to provide useful predictions about the potential transmission of the virus .
the important step after modelling is to study the properties of the system of equations that describes the propagation of the virus . in this work , we analysed a sir model with vital dynamics for the early detection of ebola virus in west africa , by adding demographic effects and an induced death rate , in order to discuss when the model makes sense mathematically and to study the information provided by the model .
we simulated the model in the case of a basic reproduction number @xmath44 , which describes the current situation of ebola virus in guinea .
we studied the equilibria .
the system of equations of the model was solved numerically and the numerical simulations confirmed the theoretical analysis of the equilibria for the model .
finally , we controlled the propagation of the virus by minimizing the number of infected individuals and the cost of vaccination and showing the importance of optimal control .
this research was partially supported by the portuguese foundation for science and technology ( fct ) through project uid / mat/04106/2013 of the center for research and development in mathematics and applications ( cidma ) and within project toccata , ref .
ptdc / eei - aut/2933/2014 .
d. ariens , m. diehl , h. j. ferreau , b. houska , f. logist , r. quirynen and m. vukov , _ acado toolkit user s manual _ , optimization in engineering center ( optec ) and department of electrical engineering , ku leuven , 2014 .
m. barry , f. a. traor , f. b. sako , d. o. kpamy , e. i. bah , m. poncin , s. keita , m. cisse and a. tour , _ ebola outbreak in conakry , guinea : epidemiological , clinical , and outcome features _ ,
mdecine et maladies infectieuses * 44 * ( 2014 ) , no . 1112 , 491494 .
l. borio et al .
[ working group on civilian biodefense ; corporate author ] , _ hemorrhagic fever viruses as biological weapons : medical and public health management _ , journal of the american medical association * 287 * ( 2002 ) , no .
18 , 23912405 .
s. f. dowell , r. mukunu , t. g. ksiazek , a. s. khan , p. e. rollin and c. j. peters , _ transmission of ebola hemorrhagic fever : a study of risk factors in family members _ , kikwit , democratic republic of the congo , 1995 .
commission de lutte contre les epidmies kikwit .
j. infect .
dis . * 179 * ( 1999 ) , suppl . 1 , s87s91 . h. r. joshi , s. lenhart , m. y. li and l.wang , _ optimal control methods applied to disease models _ , in _ mathematical studies on human disease dynamics _ , 187207 , contemp . math . , 410 , amer . math .
soc . , providence , ri , 2006 .
j. a. lewnard , m. l. ndeffo mbah , j. a. alfaro - murillo , f. l. altice , l. bawo , t. g. nyenswah and a. p. galvani , _ dynamics and control of ebola virus transmission in montserrado , liberia : a mathematical modelling analysis _ , the lancet infectious diseases * 14 * ( 2014 ) , no . 12 , 11891195 .
a. rachah and d. f. m. torres , _ mathematical modelling , simulation and optimal control of the 2014 ebola outbreak in west africa _ , discrete dyn .
* 2015 * ( 2015 ) , art .
i d 842792 , 9 pp . | we present a mathematical analysis of the early detection of ebola virus . the propagation of the virus is analysed by using a susceptible , infected , recovered ( sir ) model . in order to provide useful predictions about the potential transmission of the virus ,
we analyse and simulate the sir model with vital dynamics , by adding demographic effects and an induced death rate .
then , we compute the equilibria of the model .
the numerical simulations confirm the theoretical analysis .
our study describes the 2015 detection of ebola virus in guinea , the parameters of the model being identified from the world health organization data .
finally , we consider an optimal control problem of the propagation of the ebola virus , minimizing the number of infected individuals while taking into account the cost of vaccination . |
the near - infrared wavelength range offers valuable probes of the dynamics , composition , and excitation mechanisms present in the circumnuclear gas of active galactic nuclei ( agn ) . among the near - infrared emission lines which have been observed to be strongest in seyfert galaxies
are the recombination lines of hydrogen , pa@xmath1 ( @xmath2=1.2818@xmath3 m ) and br@xmath4 ( @xmath2=2.1655@xmath3 m ) , collisionally excited lines of singly ionized iron , e.g. , [ ] ( @xmath2=1.2567@xmath3 m ) , and rotational - vibrational lines of molecular hydrogen , e.g. , h@xmath5 @xmath6=10s(1 ) ( @xmath2=2.1212@xmath3 m ) .
additionally , there are very high excitation coronal lines , e.g. , [ ] ( @xmath2=1.2525@xmath3 m ) , which have been seen in a number of seyfert galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
to date , most published infrared spectra of seyfert 2 galaxies use single large beams centered on the nucleus ( e.g. , * ? ? ? * ) .
thus , very little spatial information on the strong near - infrared lines is available .
however , some studies have shown that near - infrared line emission is extended over a spatial scale easily probed with ground - based spectroscopy .
for example , near - infrared [ ] and h@xmath5 emission has been seen extended over several arcseconds ( @xmath71 kpc ) in the nucleus of the seyfert 2 galaxy ngc 1068 @xcite , the seyfert 1.5 galaxy ngc 7469 @xcite , and the seyfert 1.5 galaxy ngc 4151 @xcite . in each case , there were overall spatial and kinematic similarities between the infrared and optical line emission , but differences were also apparent on spatial and velocity scales within the nlr . clearly , there is important diagnostic information about the circumnuclear regions of seyfert 2 galaxies to be gained from spatially resolved near - infrared spectroscopy .
this paper presents longslit near - infrared spectra of the seyfert 2 galaxies mk 1066 , ngc 2110 , ngc 4388 , and mk 3 , in wavelength ranges which include the lines pa@xmath1 , br@xmath4 , [ ] ( @xmath2=1.2567@xmath3 m ) , h@xmath5 @xmath6=10 s(1 ) , and [ ] ( @xmath2=1.2525@xmath3 m ) .
these four objects were chosen from a larger project to search for spatially extended infrared line emission in @xmath820 seyfert 1 and seyfert 2 galaxies . because desirable targets in this larger project were those likely to show extended emission , seyfert galaxies visible from palomar observatory ( @xmath9 )
were selected with redshift cz@xmath105400 km s@xmath11 , corresponding to a spatial scale of 350 pc/@xmath12 .
( throughout this paper , for the determination of distances and hence projected scales , we assume @xmath13 km s@xmath11 mpc@xmath11 . ) galaxies with known extended optical ionized gas were preferentially observed , followed in priority by galaxies with spatially extended radio emission on scales of @xmath14 .
the final sample of galaxies observed includes 11 seyfert 1.x ( 1 , 1.5 , and 1.9 ) galaxies and 12 seyfert 2 galaxies .
this paper reports the results from this program for the four seyfert 2 galaxies that show the best well - detected extended circumnuclear emission line regions in the near - infrared . throughout this paper , for the determination of distances and hence projected scales , we assume @xmath13 km s@xmath11 mpc@xmath11 .
the near - infrared spectra discussed in this paper were obtained at the hale 200@xmath12 telescope with the palomar near - infrared spectrometer @xcite .
observations at resolution @xmath0 were made in two spectral ranges for each galaxy .
one spectral range in the j - band atmospheric window covers 0.07@xmath3 m and includes the [ ] , [ ] , and pa@xmath1 lines in the rest frame of each galaxy .
the other spectral range in the k - band atmospheric window covers 0.13@xmath3 m and includes the h@xmath5 @xmath6=10 s(1 ) and br@xmath4 lines .
table [ sey2obslog ] is a log of the observations , including the resolution ( in km s@xmath11 ) , on - source integration time , and position angle of the slit for each observation . since we are interested in exploring the near infrared emission line flux ratios and dynamics as a function of distance from the seyfert nucleus , we chose slit position angles to coincide with the position angle of known ionization cones or other anisotropy previously mapped through narrow band optical emission line imaging . in the case of two of the galaxies ( mk 1066 and ngc 4388 ) ,
there are additional observations along a secondary slit through the nucleus , perpendicular to the primary position angle .
the spectrometer is a reimaging system which disperses the light from the telescope and feeds the spectra to an externally mounted infrared camera .
the spatial scale of the 256@xmath15256 nicmos hgcdte array in the camera is 0.165@xmath12 pixel@xmath11 .
the width of the adjustable slit was approximately 0.6@xmath12 , yielding the resolutions listed in table [ sey2obslog ] . in order to accurately center
the nucleus in the slit , a fold mirror inside the spectrometer , but in front of the grating , allows infrared imaging with a @xmath16 field of view . between subsequent observations of each galaxy ,
the nucleus was moved back and forth along the slit by 20@xmath12 , to allow for background subtraction while maximizing the amount of on - source integration time .
each individual exposure was 600s for the j - band ( 1.2@xmath3 m ) wavelength range , and 300s for the k - band ( 2.2@xmath3 m ) wavelength range .
along with the observation of each galaxy , we observed a g - dwarf star at a similar airmass ( within 0.1 of that of the galaxy ) , and where possible in the same area of the sky , in order to correct for atmospheric transmission . a chopping secondary mirror on the telescope uniformly spread the light of this star along the length of the slit .
in this way , the spectrum of the atmospheric calibrator is also used as a flat - field .
the data were reduced following the procedure described in @xcite .
bad pixels were removed by linear interpolation , and the spatial and spectral curvature of the field of view was measured and removed using a point source spectrum and the spectrally unresolved oh night sky lines .
spectral data frames were divided by the appropriate g - star spectrum , in order to remove atmospheric absorption lines and to correct for efficiency variations across the field of view . in order to correct for the blackbody slope of the g - star continuum , and to remove absorption features intrinsic to the star , each g - star spectrum
was first divided by a solar spectrum following the procedure of @xcite .
mk 1066 is an inclined sb0 + galaxy with a systemic velocity of 3625 km s@xmath11 @xcite .
this corresponds to a calculated distance of 48 mpc , which yields a spatial scale of 234 pc arcsec@xmath11 .
hubble space telescope ( hst ) imaging of the nuclear regions shows a narrow `` jetlike '' feature in a narrow band image which includes [ oiii ] and h@xmath1 .
this jet is oriented at 315@xmath17 and extends @xmath18 northwest of the nucleus . in h@xmath19 and [ nii ] ,
the jet is observed on both sides of the nucleus @xcite .
the nuclear radio source is linear , and extended over 2.6@xmath12 .
the 5 ghz radio emission is oriented at the same position angle as the optical line emission @xcite .
we obtained infrared spectra of mk 1066 along two position angles , 135/315@xmath17 and 45/225@xmath17 .
the 135@xmath17 angle of the slit was chosen to align with the position angle of the optical ionization cones @xcite .
these spectra show strong line emission , concentrated on the nucleus .
four emission lines are visible in figures [ mk1066grey ] , [ mk1066spec135 ] , and [ mk1066spec45 ] : pa@xmath1 and [ ] in the j - band spectra , and br@xmath4 and h@xmath5 @xmath6=10 s(1 ) in the k - band spectra .
no [ ] ( @xmath2=1.2525@xmath3 m ) is visible on the nucleus ( to an upper limit on the equivalent width of @xmath20 @xmath3 m ) .
in addition , there is data along a 45@xmath17 slit chosen to be perpendicular to the primary slit .
the infrared lines are visible over a spatial extent of @xmath21 ( 1.2 kpc ) along the 135@xmath17 slit , in comparison to the fwhm of the nuclear continuum which was @xmath22 at 1.25@xmath3 m and @xmath23 at 2.15@xmath3 m in our spectra ( both of which are greater than the 0.50.7@xmath12 fwhm of the atmospheric seeing ) . in contrast
, the lines are less spatially extended along the 45@xmath17 slit , and are visible only over an extent of @xmath24 .
the br@xmath4 line is strongest on the nucleus , but the pa@xmath1 , [ ] , and h@xmath5 lines peak about @xmath25 northwest of the nucleus ( see figure [ mk1066lv ] ) .
figure [ mk1066grey ] shows two greyscale plots of the reduced 1.251.31@xmath3 m spectra along both position angles .
the horizontal axis is observed wavelength , and the vertical axis is position along the slit in arcseconds from the nucleus .
the pa@xmath1 and [ ] lines are spatially resolved along both position angles in figure [ mk1066grey ] .
the tilt in the lines indicates that there is a spatial velocity gradient in the [ ] and pa@xmath1 lines along the 135@xmath17 slit ; this will be discussed in section [ sec : mk1066velocitystruc ] . along the 45@xmath17 slit ,
no velocity gradient is immediately apparent .
figures [ mk1066spec135 ] and [ mk1066spec45 ] are one - dimensional projections in various spatial bins extracted from the two - dimensional data .
figure [ mk1066spec135 ] shows the spectra along the 135@xmath17 slit , and figure [ mk1066spec45 ] along the 45@xmath17 slit .
the size of each spatial bin used in the extractions is 4 pixels , or 0.67@xmath12 .
this matches the point spread function ( psf ) at the time of the observations ( [email protected]@xmath12 ) .
the pa@xmath1 profile near the nucleus shows enhanced emission blueward of the line center .
these blueward features on pa@xmath1 are visible not only in mk 1066 , but also in several other seyfert galaxies @xcite . were this pa@xmath1 emission
, it would be at a velocity of @xmath26 km s@xmath11relative to the centroid of the pa@xmath1 emission , which is higher than the velocities observed in other lines in mk 1066 .
these blue features on pa@xmath1 are probably primarily due to a blend of two lines at 1.2785@xmath3 m and 1.2791@xmath3 m , which have been resolved from pa@xmath1 in high resolution spectroscopy of the planetary nebula bd@xmath27 @xcite .
there may also be a small contribution from [ ] ( @xmath28 m ) . for t@xmath29 k and n@xmath29 @xmath30 ,
the strength of the [ ] ( @xmath28 m ) line is expected to be @xmath31 of the strength of the [ ] ( @xmath2=1.2567@xmath3 m ) line @xcite . in mk 1066 , [ ] ( @xmath28 m ) could account for at most half of the emission on the blue side of pa@xmath1 .
table [ mk1066eqw ] in appendix [ sec : tabappndx ] lists the equivalent widths and flux ratios of the lines observed in mk 1066 . because [ ] and pa@xmath1 were observed simultaneously and
are close in wavelength , we may obtain accurate [ ] /pa@xmath1 flux ratios which are relatively insensitive to uncertainties in flux calibration or differential reddening .
similarly , we may obtain reliable h@xmath5/br@xmath4flux ratios from our k - band spectra .
figure [ mk1066fluxrat ] shows these ratios as a function of position along the slit , using the total integrated fluxes for each species in each spatial bin to derive the line flux ratios . a striking feature of figures [ mk1066spec45 ] and [ mk1066fluxrat ] is the rapid rise of the h@xmath5/br@xmath4 flux ratio with distance from the nucleus along the 45@xmath17 slit , perpendicular to the position angle of the optical ionization cones .
the h@xmath5 line is detected 2@xmath12 away from the nucleus in both directions , while the br@xmath4 line is detected only within 0.67@xmath12 of the nucleus .
the h@xmath5/br@xmath4 flux ratio 1.33@xmath12 from the nucleus is @xmath323 , in contrast to the nuclear value of @xmath33 . by comparison , along the 135@xmath17 slit
, the h@xmath5/br@xmath4 flux ratio always remains between about 0.4 and 0.9 , with an indication of a gradient across the nucleus .
the [ ] /pa@xmath1 flux ratio varies with position along the 45@xmath17 slit , although not as rapidly as does the h@xmath5/br@xmath4 flux ratio .
there is a clear local minmum in the [ ] /pa@xmath1 ratio at the nucleus along both slits .
[ sec : mk1066wings ] figure [ mk1066lv ] is a set of position versus velocity plots along the 135@xmath17 slit for each of the four lines identified in figures [ mk1066spec135 ] and [ mk1066spec45 ] . in each case
, a smoothed continuum has been subtracted from the spectrum .
velocity is plotted relative to 3625 km s@xmath11 , the systemic velocity of mk 1066 @xcite , and position is distance in arcseconds on the sky from the continuum peak .
a velocity gradient is visible in each of the lines , in the sense that the lines are blueshifted southeast of the nucleus and redshifted northwest of the nucleus .
the slope of the velocity gradient does not appear to be the same for all of the lines .
specifically , the h@xmath5 line appears to have a steeper gradient ( @xmath8300 km s@xmath11 arcsec@xmath11 , or @xmath81300 km s@xmath11 kpc@xmath11 ) in comparison to the other three lines ( whose gradients are @xmath8200 km s@xmath11 arcsec@xmath11 , or @xmath8900 km s@xmath11 kpc@xmath11 ) .
in addition , br@xmath4 appears to have a slightly steeper gradient ( by @xmath850 km s@xmath11 arcsec@xmath11 ) than pa@xmath1 southeast of the nucleus .
the gradients of the [ ] and pa@xmath1 lines are similar .
another feature apparent in figure [ mk1066lv ] is that all of the lines , except for br@xmath4 , peak northwest of the nucleus .
this is discussed in section [ mk1066extinction ] .
table [ mk1066kinematics ] in appendix [ sec : tabappndx ] lists the velocity centroids and full widths at half maxima of the complete line profiles for the lines observed in mk 1066 along both slits in each spatial bin .
many of the lines show a complicated asymmetric profile .
these lines were decomposed into the smallest number of gaussian components necessary to statistically describe the line profile .
figure [ mk1066vvspos ] shows plots of the velocity of the strongest narrow gaussian component of each line as a function of position along the slit .
usually , this is equivalent to a plot of the mode ( peak ) of the line profile as a function of position , because most of the line profiles in mk 1066 are dominated by a strong , central narrow component .
for reference , a line connecting the br@xmath4 velocity peaks is drawn on each plot , and the h@xmath19 velocity curve @xcite is shown on the pa@xmath1 plot .
the same relative velocity gradients between the different species mentioned for figure [ mk1066lv ] are visible in figure [ mk1066vvspos ] .
it is apparent the difference between the h@xmath5 and br@xmath4 velocity slopes is greater than the difference between the h@xmath5 and pa@xmath1 velocity slopes .
two arcseconds southeast of the nucleus , all of the velocity profiles except for that of br@xmath4 ( which is undetected ) show a local velocity reversal of 5075 km s@xmath11 .
in addition to the primary narrow emission line components whose centroid velocities are plotted in figure [ mk1066vvspos ] , there are weaker components which are displaced in velocity relative to the center of the lines . southeast of the nucleus , where the peaks of the line profiles are blueshifted relative to the systemic velocity , all of the lines show a noticeable enhanced red wing ( although it is very weak in h@xmath5 ) . on the nucleus and to the northwest where the peaks of the line profiles are redshifted , the pa@xmath1 and [ ] lines show an enhanced blue wing .
while this blue wing is almost certainly a feature of the [ ] line profile itself , the blue wing of the pa@xmath1 line suffers contamination from other species as discussed above
. for illustration , figure [ mk1066closeups ] shows the profiles of the [ ] , pa@xmath1 , and br@xmath4 lines on the nucleus and 1.33@xmath12 southeast and northwest of the nucleus .
southeast of the nucleus , the red wing is stronger relative to the rest of the line in [ ] and br@xmath4 than it is in pa@xmath1 .
similarly , northwest of the nucleus , the blue wing is most prominent in [ ] .
it is not visible in the br@xmath4 line , and in pa@xmath1 it may be contaminated by the aforementioned and [ ] lines .
@xcite present longslit optical spectroscopy of mk 1066 along a 134@xmath17 slit .
the h@xmath19 , h@xmath1 , [ nii ] , [ sii ] , and [ oi ] lines all have a velocity gradient extending from + 100 km s@xmath11 2 - 3@xmath12 northwest of the nucleus to -90 km s@xmath11 2 - 3@xmath12 southeast of the nucleus .
this is similar to the velocity structure seen in the infrared lines ( see figure [ mk1066vvspos ] ) , although the peaks of the pa@xmath1 and particularly the br@xmath4 lines are slightly blueshifted relative to the optical lines to the southeast of the nucleus .
in addition to the primary kinematic component seen in optical emission lines , @xcite identify a second kinematical component , most clearly seen in [ oiii ] .
the [ oiii ] emission is slightly blueshifted relative to the other lines northwest of the nucleus and redshifted by as much as 200 km s@xmath11relative to the other lines southeast of the nucleus .
this structure is qualitatively similar to the enhanced wings seen in the infrared lines discussed in section [ sec : mk1066wings ] .
specifically , the infrared lines show an enhancement blueward of the peak on the nucleus and to the northwest , and they show an enhancement redward of the peak to the southeast .
it is likely that some of the wings seen on the infrared emission lines are associated with the kinematic component @xcite identify in [ oiii ] . since pa@xmath1 and br@xmath4 are both hydrogen recombination lines ( n=53 and n=74 respectively ) , they should arise in the same physical locations .
their intrinsic case b flux ratio is well known ( assuming radiative transfer effects are negligible ) . because of reddening , in dusty regions the observed pa@xmath1 line will be suppressed relative to the observed br@xmath4 line .
if the dust is intermixed with the line emitting clouds , and there are strong velocity gradients , the two lines can appear to have different velocity structures . in mk 1066 , both
the infrared and optical data indicate that extinction effects are stronger to the southeast than they are to the northwest .
a difference in both the spatial and velocity distributions of pa@xmath1 and br@xmath4 is visible in figure [ mk1066lv ] .
the br@xmath4 line is peaked close to the continuum peak , and is stronger relative to pa@xmath1 toward the southeast than it is toward the northwest .
the simplest explanation of this difference is a larger extinction in the southeast .
spatially variable extinction may also be responsible for the slightly different pa@xmath1 and br@xmath4 velocity gradients seen in figure [ mk1066lv ] .
to the southeast , where dust extinction appears greater , the discrepancy between the velocity gradients for the pa@xmath1 and br@xmath4 lines is larger .
extinction could also explain the difference between the velocities of the infrared lines and the optical h@xmath19 line southeast of the nucleus @xcite .
the h@xmath19-infrared slope discrepancy is in the same sense as the pa@xmath1br@xmath4 slope discrepancy , i.e. , lines at shorter wavelengths have shallower gradients ; see figure [ mk1066vvspos ] .
this same extinction effect is seen in the red wings of the pa@xmath1 and br@xmath4 lines southeast of the nucleus as discussed in section [ sec : mk1066wings ] ; the wing is relatively more prominent in br@xmath4 than in pa@xmath1 suggesting that the extinction to the region where the wings originate is greater than the extinction to the region where the bulk of the line emission originates .
if the wings represent outflowing gas , then the redshifted receding gas , which may be partially obscured by dust in the disk of the galaxy , would show higher extinction .
there is also evidence from optical wavelengths that the outflowing gas in the southeast is heavily extinguished .
the optical line emission is seen emerging from the nucleus in ionization cones at position angles of 135@xmath17/315@xmath17 .
the [ oiii]+h@xmath1 narrowband image @xcite is much stronger to the northwest , suggesting suppressed emission to the southeast .
the 5 ghz radio emission @xcite indicates that the strength of the outflow is slightly asymmetric , but the difference in the strength of the radio emission northwest and southeast of the nucleus is not as great as the difference in the strength of the [ oiii ] emission which @xcite associate with the outflow .
extinction can not explain the discrepancies between the h@xmath5 and br@xmath4lines , since the differential extinction between the two is negligible .
the h@xmath5 line shows a different velocity gradient than br@xmath4(figure [ mk1066lv ] ) , and the peak of the h@xmath5 emission is shifted to the northwest relative to the peak of the br@xmath4 emission .
the differences in the structure of these lines may indicate that the emission seen from each species originates from different clouds along the same line of sight .
additionally , there may be a systematic variation in the physical conditions along the cone axis of the galaxy .
the infrared data in combination with the data in @xcite indicate that there are two dominant kinematic components in near the nucleus of mk 1066 .
one component shows up in infrared h@xmath5 , in the peaks of the infrared lines , and in optical h@xmath19 , [ ] , and other species .
this component shows a clear velocity gradient along the 135@xmath17 slit , is blueshifted to the southeast , and is redshifted to the northwest .
the second , higher - excitation component shows up in the wings on the infrared lines and in optical [ ] , and has the opposite velocity trend from the first component .
@xcite identify the first , low - excitation component with a rotating disk .
the velocity gradient seen in the infrared line emission for this component along the 135@xmath17 slit is consistent with a rotation .
the much smaller ( though non - zero ) velocity gradient along the 45@xmath17 slit indicates that the major axis of the rotating disk must be closer to 135@xmath17 than 45@xmath17 ; a fit of a uniformly rotating disk to the peaks of the infrared line emission along both slits suggests that the major axis of the disk is at @xmath34 .
following @xcite , we identify the second component with high - excitation outflowing gas .
this second component is not seen off of the nucleus in the infrared spectra along the 45@xmath17 slit , and the optical [ ] emission is clearly extended along an angle of @xmath35 .
consequently , the data indicate that the outflow is _ not _ perpendicular to the rotating disk ; rather , the two physical systems are at low relative inclination when projected on to the plane the sky .
figure [ mk1066art ] is a cartoon drawing of the system based on the optical and infrared data . in this picture
, there is an inclined disk with a major axis of 120@xmath17 projected on the sky .
this orientation explains both the greater spatial extent and the larger velocity gradient in the infrared lines along the 135@xmath17 slit .
the disk is dusty , explaining the extinction effects discussed in section [ mk1066extinction ] .
there is a biconical outflow associated with the higher - excitation gas , as drawn . to the northwest
, this outflow is relatively unobscured , as it is in front of the plane of the galaxy .
the blue wings seen on the infrared lines to the northwest are associated with this outflow . to the southeast ,
the outflow is obscured by the plane of the galaxy , suppressing the [ oiii ] emission observed in that direction .
extinction resulting from this obscuration enhances the br@xmath4/pa@xmath1 ratio on the red side of the line profile southeast of the nucleus ( see figure [ mk1066closeups ] ) .
the fact that various features of the [ ] emission line profiles correspond to those identified with the active nucleus in other species , such as the wings seen in [ oiii ] @xcite , suggests that the [ ] emission is dominated by processes directly associated with the seyfert nucleus .
the radio emission of mk 1066 @xcite appears to represent two oppositely directed radio jets aligned with the outflow seen in [ oiii ] . in the wings on the infrared line profiles we associate with this outflow , [ ] emission is enhanced relative to pa@xmath1 ( see figure [ mk1066closeups ] ) .
shocks resulting from the interaction of outflowing radio plasma with the ambient gas are a natural mechanism for the production of [ ] emission .
fast shocks can both destroy dust grains , increasing the gas phase abundance of iron , and create large partially ionized regions , two conditions which enhance [ ] emission .
the most striking feature of the h@xmath5 emission is that its strength does not drop off as quickly with distance as the other infrared lines from the nucleus perpendicular to the outflow axis . while br@xmath4 is only detected out to 150pc from the nucleus at a position angle of 45@xmath17 , h@xmath5 is detected to @xmath32300pc .
the presence of h@xmath5 emisison along both slits shows that warm molecular gas is ubiquitous in the near - nuclear disk of mk 1066 . since the h@xmath5
is not confined to the cone axis , it is evidently not directly dependent on ionization from the active nucleus or the radio jets for its excitation .
there is probably little h@xmath5 emission associated with the high excitation outflow , since the wings seen in the [ ] and pa@xmath1 lines associated with this outflow were not detected in the h@xmath5 line profiles .
heating by x - rays from the seyfert nucleus likely can not explain the extent of the h@xmath5 emission perpendicular to the ionization cone axis .
slow ( v@xmath36 km s@xmath11 ) shocks originating in the nucleus may heat molecular clouds in the disk of the galaxy immediately surounding the nucleus , but the relatively flatter distribution of h@xmath5 relative to the other species along the 45@xmath17 slit argues for an extended source of heating for the molecular gas in the regions within @xmath37 pc of the nucleus of mk 1066 . 1 . the infrared pa@xmath1 , br@xmath4 , and [ ] lines are extended over 5@xmath12 ( 1.2
kpc projected on the plane of the sky ) along a 135@xmath17 position angle coincident with the direction of the optically identified ionization cones .
along a perpendicular angle , their intensity drops off faster with distance from the nucleus ; 1.3@xmath12 ( 0.3
kpc projected ) away from the nucleus at 45@xmath17 , the intensity of the lines relative to the nucleus is 20% of their value along a position angle of 135@xmath17 .
the infrared lines have strongly peaked narrow components .
the peaks of the lines have a velocity gradient similar to the optical h@xmath19 , h@xmath1 , [ nii ] , [ sii ] , and [ oi ] lines along a position angle of 135@xmath17 , which @xcite identify as part of a rotating , liner - like disk .
3 . there are secondary components in the form of enhanced wings in the infrared lines whose kinematics differ from the strong , narrow components in a manner similar to the difference between the kinematics of [ oiii ] emission and the other optical lines . the [ oiii ] emission line and the near infrared line wings are likely associated with a nuclear outflow which produces blueshifted emission to the northwest of the nucleus .
the relative spatial distributions of br@xmath4 and pa@xmath1 indicate that there is greater extinction to the southeast of the nucleus than there is to the northwest .
this may explain differences seen in the velocity centers of hydrogen recombination lines to the southeast , which is most likely due to obscuration of the gas by an inclined disk in mk 1066 .
the h@xmath5 line is more extended than the other lines along an angle perpendicular to the ionization cone . along the axis of the cone
, it shows a steeper velocity gradient than pa@xmath1 , br@xmath4 or [ ] .
we propose that there is a measurable contribution to the [ ] emission from shocks associated with the nuclear outflow .
it may have a direct link to the radio jets seen at 3.6 cm .
the h@xmath5 emission , on the other hand , arises primarily from the inner ( 500pc ) disk of mk 1066 , most of which is presumably shielded from direct illumination of the nucleus .
this warm molecular gas may be heated by extended local sources such as star formation .
ngc 2110 is an s0 galaxy with a systemic velocity of 2342 km s@xmath11@xcite , corresponding to a calculated distance of 31 mpc , implying a spatial scale of 150 pc arcsec@xmath11 .
hst narrowband imaging in optical h@xmath19+[nii ] and [ oiii ] emission shows a narrow jetlike feature near the nucleus extending to the north at a position angle of @xmath8340@xmath17 , and a weaker opposing feature extending to the south at @xmath8160@xmath17 @xcite . the feature curves into an `` s '' shape 4@xmath12 away from the nucleus in each direction .
at 1.49 ghz , ngc 2110 shows a triple radio source extended over 4@xmath12 with a position angle of @xmath80@xmath17 @xcite , slightly different from the position angle of the optical line emission .
the curvature of the `` s '' of the optical line emission is towards the direction of the radio emission . in x - rays ,
ngc 2110 has a 2 - 10 kev luminosity of 1.4@xmath38 erg s@xmath11 with a comparatively high absorbing column density of @xmath39 @xmath40 @xcite .
this x - ray luminosity is closer to that typical for a moderate luminosity seyfert 1 galaxy than for a seyfert 2 galaxy ( @xmath41 erg s@xmath11 ) @xcite .
rosat high resolution imager ( hri ) data show a spatially extended x - ray source , with secondary x - ray emission 4@xmath12 north of the nucleus which may be associated with the soft x - ray excess seen by @xcite .
our longslit spectra of ngc 2110 along the ionization cones at a position angle of 160@xmath17 ( figure [ ngc2110spec ] ) show strong emission lines in [ ] , pa@xmath1 , h@xmath5 @xmath6=10 s(1 ) , and br@xmath4 .
no [ ] ( @xmath2=1.2525@xmath3 m ) is detected on the nucleus ( with an upper limit on the equivalent width of @xmath42 m ) .
figure [ ngc2110spec ] shows the summed spectra in spatially adjacent 1@xmath12 bins along the slit .
these lines are visible on the nucleus , and are seen to be extended by @xmath43 ( 900 pc ) . in comparison ,
the fwhm of the continuum is 2.4@xmath12 at 1.2@xmath3 m and 1.5@xmath12 at 2.2@xmath3 m , while the seeing at the time of the observations was @xmath23 .
the most striking feature of the ngc 2110 j - band spectrum is the strength of the [ ] line . on the nucleus ,
the [ ] /pa@xmath1 flux ratio is [email protected] , which is greater than that observed in any other seyfert galaxy in this survey @xcite .
table [ ngc2110eqw ] in appendix [ sec : tabappndx ] lists the equivalent widths and flux ratios of the lines observed in ngc 2110 .
the [ ] /pa@xmath1 and h@xmath5/br@xmath4 line ratios are very large on and near the nucleus .
these flux ratios are plotted as a function of position along the slit in figure [ ngc2110fluxrat ] . near the nucleus , the h@xmath5/br@xmath4 flux ratio is close to 3 , locally rising at a position @xmath45 north of the nucleus , and dropping off in the furthest spatial bins .
the [ ] /pa@xmath1 flux ratio peaks on the nucleus , falling rapidly with radius away from the nucleus . for most seyfert galaxies , values of [ ]
/pa@xmath1 typically range between 0.5 and 2.0 ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , while the h@xmath5 @xmath6=10 s(1)/br@xmath4 flux ratio typically lies between 0.5 and 1.5 ( e.g. , * ? ? ?
the flux ratios observed in ngc 2110 are significantly higher than these , with the [ ] line in particular being unusually strong in comparison to the pa@xmath1 line . indeed ,
the value of 8.1 for the [ ] /pa@xmath1 flux ratio on the nucleus is higher than that observed in any published seyfert galaxy spectrum ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . this high [ ] /pa@xmath1 ratio is due to both a strong [ ] line and a weak pa@xmath1 line in comparison to other seyfert galaxies . the equivalent width of the [ ] line on the nucleus , 25@xmath46 m , is larger by a factor of 2 - 4 than that typically seen in seyfert galaxies , although it is not the largest known equivalent width of [ ] in a seyfert 2 galaxy ( e.g. , ngc 449 ; * ? ? ?
* ) . in figure [ ngc2110lv ]
, we show a set of position versus velocity plots for each of the four near - infrared lines seen in ngc 2110 . a smoothed continuum has been subtracted from the data in each case .
velocity is plotted relative to the systemic velocity of ngc 2110 .
all of the lines in figure [ ngc2110lv ] show a gradient with velocity increasing from north to south . unlike mk 1066 ,
the velocity gradients appear to be quite similar for the four infrared lines seen in ngc 2110 .
figure [ ngc2110vvspos ] is a plot of the velocity centroid of each line profile as a function of position for each of the four lines .
the centroids for pa@xmath1 have excluded the flux due to the contamination from and [ ] on the blue side of pa@xmath1 ( as discussed in section [ sec : mk1066 ] ) .
table [ ngc2110kinematics ] in appendix [ sec : tabappndx ] lists these velocity centroids , as well a the full widths at half maximum of each of the line profiles .
the velocity structure in figure [ ngc2110vvspos ] is dominated by the gradient from north to south .
[ ] appears different from the other species in figure [ ngc2110lv ] because it is much wider on the nucleus than it is off of the nucleus .
the shape of the contours convolves the spatial variation of the line width and the gradient in the line s velocity .
the slope of the gradient in all species is steeper to the south than to the north ( figure [ ngc2110vvspos ] ) , reaching velocities of about @xmath47100 to @xmath47150 km s@xmath11 at 3@xmath12 . south of the nucleus and @xmath48250 to @xmath48300 km s@xmath11 at 3@xmath12 south of the nucleus .
a second feature apparent in figure [ ngc2110vvspos ] is that the [ ] line is systematically at a lower velocity than the other lines .
this effect is strongest south of the nucleus , although here the strong blue wings in [ ] pull the centroids to lower velocities . in each spatial bin
the velocity peak of the [ ] line is shifted to the blue with respect to the other lines by about 50100 km s@xmath11 , which is several times the uncertainty in the wavelength calibration . in many cases , the lines are asymmetric , although the asymmetries are mostly small enough that the velocity centroids are very close to the velocities of the peaks of the lines . on the nucleus and 1@xmath12 south of the nucleus , pa@xmath1 , [ ] , and h@xmath5
show a blue wing in places , while north of the nucleus , br@xmath4 , [ ] , and h@xmath5 show red wings .
the [ ] line is significantly broader than the other lines , particularly close to the nucleus ( see figure [ ngc2110lv ] ) . on the nucleus ,
the fwhm of the [ ] line is 500 km s@xmath11 , as compared to 340 km s@xmath11 for the pa@xmath1 line .
in addition , the [ ] line has wings with fwhm @xmath49 km s@xmath11 which are not visible in the pa@xmath1 line .
the [ ] line is narrower at larger radii , having a fwhm of @xmath50 km s@xmath11 , 3@xmath12 away from the nucleus . at 3@xmath12 south of the nucleus , the [ ]
line remains @xmath51 km s@xmath11 broader than the pa@xmath1 line , while north of the nucleus the line widths are comparable .
@xcite ( hereafter cited as s99 ) performed spatially extended infrared spectroscopy of ngc 2110 with resolutions of @xmath52 and @xmath53 .
they have data along two slit angles : one at 170@xmath17 , similar to our slit angle of 160@xmath17 , and one at a perpendicular angle of 80@xmath17 .
the features of their spectra along the 170@xmath17 slit are also seen in this work , including the notable increase in the [ ] /pa@xmath1 ratio on the nucleus .
any quantitative differences between the results of the two studies are easily explained by differences in slit widths and seeings .
one exception is the slope of the continuum in the 2.15@xmath3 m spectra , where the slope seen in this work is opposite that seen in s99 .
this is probably also a result of differences in the sampled spatial region , but may be partially a result of differences in data reduction , indicating the difficulty in absolute calibration of the continuum in infrared spectra . since this work is primarily concerned with the spatial and velocity distribution of emission line radiation , it is important to consider how these differences affect calculated line flux ratios . for the differences seen between the two nuclear continuum spectra
, the h@xmath5/br@xmath4 flux ratio should not be misestimated by more than @xmath815% , comparable to our quoted error on the nuclear flux ratio .
s99 saw a nuclear h@xmath5/br@xmath4 flux ratio of [email protected] , consistent with our value of [email protected] ; this indicates that the flux ratios observed in each case are in fact robust . along the perpendicular 80@xmath17 slit ,
s99 have data in a wavelength range including br@xmath4 and h@xmath5 ( @xmath2=2.1212@xmath3 m ) .
is well detected to a greater radius along this 80@xmath17 slit , which is perpendicular to the ionization cones .
although s99 have weak or no detection of br@xmath4 off of the nucleus in either direction , the s / n of the off - nuclear h@xmath5detection is better along the 80@xmath17 slit than it is along the 170@xmath17 slit .
this , coupled with our observations of the h@xmath5/br@xmath4 ratio along our 160@xmath17 slit , suggests that also as with mk 3 , the h@xmath5/br@xmath4 ratio off of the nucleus is larger in the slit perpendicular to the ionization cone than it is along the cone axis .
s99 report the detection of a broad pa@xmath1 line in ngc 2110 .
this line has a width of 1200 km s@xmath11 , and is highly asymmetric to the blue .
although the same spectral feature is present in the data here ( see the nuclear spectrum in figure [ ngc2110spec ] ) , we interpret the enhanced blue wing as contamination from and [ ] ( see section [ sec : mk1066 ] ) , rather than as evidence for high velocity blueshifted pa@xmath1 emission . the [ ]
( @xmath28 m ) line which contributes to this contamination is generally only @xmath31 the strength of the [ ] ( @xmath2=1.2567@xmath3 m ) line @xcite .
however , given that [ ] ( @xmath2=1.2567@xmath3 m ) is @xmath54 times as strong as pa@xmath1 on the nucleus of ngc 2110 , it is likely that the [ ] ( @xmath28 m ) line contributes more significantly to the emission blueward of pa@xmath1 than is the case in most seyfert 2 galaxies .
the velocity structure seen in h@xmath19 , h@xmath1 , and [ oiii ] @xcite is quite similar to the velocity structure seen in the infrared lines plotted in figures [ ngc2110lv ] and [ ngc2110vvspos ] .
@xcite and @xcite describe this velocity field as a rotating disk with a projected major axis of 161@xmath17 .
( this paper only presents infrared data for ngc 2110 along one slit position angle .
although the rotation identified by @xcite is evident in this data , the lack of data along two spatial dimensions precludes any independent conclusions about the major axis of the rotation . )
near the nucleus , the [ oiii ] line has a higher root mean square ( rms ) width ( equivalent to 1@xmath55 for a gaussian line profile ) than the h@xmath1 line by 100 - 200 km s@xmath11 @xcite , although the higher [ oiii ] signal - to - noise allows weak wings to be visible in the former which are not visible in the latter .
south of the nucleus , [ oiii ] shows a tendency toward enhanced blue wings , similar to what is observed in [ ] , and to a lesser extent in h@xmath5 .
these enhanced wings show some spatial similarities with the 2 cm emission @xcite , which shows three peaks aligned along a roughly north - south axis .
the strongest peak is on the nucleus .
the blue wings are strongest in [ ] , h@xmath5 , and [ oiii ] 2@xmath12 south of the nucleus , which is the approximate position of one of the secondary radio peaks .
the other off - nuclear radio peak is 2@xmath12 north of the nucleus , where we see a red shoulder in [ ] and h@xmath5 .
this suggests that , as with mk 1066 , there may be two kinematical components : a rotating disk , and a second velocity field , possibly an outflow associated with the radio jets ( see below ) .
note that even though the _ projected _ position angle of these outflowing jets appears similar to the major axis of the rotation identified by @xcite , the two physical systems need not be aligned in space .
the fact that the [ ] /pa@xmath1 ratio is very large on the nucleus and rapidly drops with radius suggests that the [ ] emission is associated with the seyfert nucleus itself .
there is indirect evidence that the physical process responsible for the [ ] emission in ngc 2110 is x - ray heating and ionization .
ngc 2110 has a 210 kev x - ray luminosity which is an order of magnitude greater than that typical for a seyfert 2 galaxy . on the nucleus ,
the [ ] /pa@xmath1 ratio is also nearly an order of magnitude greater than the typical value for seyfert 2 galaxies . out to radii 34@xmath12 north of the nucleus , the [ ] /pa@xmath1 ratio remains atypically large . about @xmath56 north of the nucleus , at the position of the secondary 0.12.4 kev x - ray emission peak identified by @xcite , there is a lower limit on the [ ] /pa@xmath1 flux ratio of @xmath57 ( due to an undetected pa@xmath1 line ) . south of the nucleus , opposite the direction of the extended x - ray emission , the [ ] /pa@xmath1 ratio drops to unity .
although the [ ] excitation may be dominated by x - ray ionization , there is likely to be an additional contribution due to fast shocks resulting from the interaction of outflowing radio plasma with surrounding gas .
this would explain the enhanced red and blue wings respectively north and south of the nucleus which are spatially coincident with the radio peaks seen by @xcite at 1.49 ghz .
nuclear x - ray heating of molecular clouds may also be the primary mechanism responsible for the h@xmath5 emission , although the enhanced h@xmath5 emission is not as dramatic as is the case with [ ] in this galaxy .
the h@xmath5/br@xmath4 ratio in ngc 2110 is large for seyfert galaxies , though the ratio is not strongly peaked on the nucleus as is the case for the [ ] /pa@xmath1ratio .
the h@xmath5/br@xmath4 flux ratio is nearly constant over the 6@xmath12 ( 900pc ) where the lines are detected .
in contrast to mk 1066 , the presence of the wings in the h@xmath5 line profiles indicates that there may be warm molecular gas in the outflowing clouds . 1 . in ngc 2110 , the [ ]
/pa@xmath1 flux ratio is unusually large , reaching a maximum value of 8.1 on the nucleus and dropping off rapidly with radius .
the h@xmath5/br@xmath4 ratio is also large , being @xmath58 within 2@xmath12 of the nucleus .
2 . in general ,
the infrared lines show a smooth velocity gradient similar to the velocity gradient seen in optical emission by @xcite .
3 . the centroid of the infrared [ ] line is systematically shifted with respect to the other lines by about 50 - 100 km s@xmath11 to the blue . additionally , the [ ] line is @xmath59 km s@xmath11 broader than the other infrared lines near the nucleus .
4 . there are enhanced wings on the infrared lines . to the south , where line centroids are redshifted with respect to the systemic velocity ,
there are enhanced blue wings . to the north , where line centroids are blueshifted with respect to the systemic velocity
, there are enhanced red wings .
these wings are strongest in [ ] and h@xmath5 at positions coincient with the secondary radio peaks of @xcite .
we suggest that the [ ] excitation is dominated by nuclear x - ray photoionization .
however , the enhanced off - nuclear wings which are spatially coincident with the nonthermal 2 cm radio peaks indicate that there is some emission associated with outflowing gas .
fast shocks may contribute to the [ ] emission at these locations . while the h@xmath5 emission is likely to be enhanced by the strong x - ray flux in ngc 2110 ,
the distribution of the warm molecular gas is not as centrally peaked as is the [ ] emission .
this may indicate a large molecular reservoir surrounding the nucleus .
in addition , the wings seen in the h@xmath5 line profile suggest that at least some of the h@xmath5 emitting gas takes part in the outflow driven by radio jets .
ngc 4388 is a highly inclined spiral galaxy at a systemic velocity of 2525 km s@xmath11 @xcite .
although its hubble type is uncertain @xcite , @xcite classify it as sb(s)b pec .
it is generally considered to be a member of the virgo cluster @xcite , implying that the distance to the galaxy is probably @xmath814 mpc @xcite .
this yields a spatial scale of 68 pc arcsec@xmath11 .
optical narrowband imaging @xcite shows an h@xmath19+[nii ] extent of @xmath60 east and west of the nucleus , aligned with the major axis of the galaxy at position angle @xmath61 .
this emission originates in the spiral arms of the galaxy , and is probably due to hii region complexes .
close to the nucleus on a spatial scale of @xmath21 , the [ oiii ] emission is extended about the nucleus in a halo - like distribution that has a major axis at a position angle of @xmath62 .
high resolution radio maps at 4.86 ghz show emission which is elongated southwest of the nucleus at a similar position angle of 200@xmath17 @xcite . at 4.86 ghz ,
ngc 4388 is double peaked , with a primary peak on the nucleus and a secondary peak 2.5@xmath12 southwest of the nucleus .
there is a weak tertiary peak about 1@xmath12 northeast of the nucleus .
ngc 4388 has also been detected in x - rays , with a 2 - 10 kev luminosity of @xmath63 erg s@xmath11 @xcite .
the soft ( 0.1 - 2.4 kev ) x - ray emission of ngc 4388 is extended over @xmath64 @xcite .
because h@xmath19 emission is extended over @xmath60 along a position angle of 120@xmath17 @xcite , we obtained a `` sky '' frame distant ( 200@xmath12 east ) from the center of the galaxy . subtracting this sky frame from spectra of
the galaxy verified that no infrared emission lines were detected at a radius greater than @xmath65 in this direction to a 3@xmath55 flux limit of @xmath66 erg @xmath67s@xmath11 in a @xmath68 spatial beam .
figures [ ngc4388spec30 ] and [ ngc4388spec120 ] are one - dimensional projections in various spatial bins of the data along the slits oriented at position angles of 30@xmath17 and 120@xmath17 respectively .
the size of each spatial bin used in the extractions is 6 pixels ( 1@xmath12 ) , slightly larger than the psf at the time of the observations , which was 0.70.8@xmath12 for the spectra at both position angles .
the nuclear spectra for the two slit position angles show qualitative differences , which are likely attributed to vagaries of spatial sampling with the rectangular synthetic aperture .
the 30@xmath17 slit was chosen to align with the major axis of the distribution of [ oiii ] emission near the nucleus @xcite .
the 120@xmath17 slit was chosen to be perpendicular to the [ oiii ] axis . in ngc 4388 ,
the [ ] , pa@xmath1 , br@xmath4 , and h@xmath5 lines are detected with a spatial extent up to 8@xmath12 ( 540 pc ) along the 30@xmath17 slit .
along the perpendicular 120@xmath17 slit , the lines diminish quickly with distance from the nucleus , but are still visible out to a radius of 23@xmath12 .
in addition to these lines , there is a narrow line seen on the nucleus blueward of [ ] , which we identify as the fine structure line [ ] at a rest wavelength of 1.25249@xmath3 m @xcite . @xcite
identify this line as [ ] in the circinus galaxy .
however , @xcite and @xcite identify this line as ( @xmath69 m ) , with possibly some contamination from [ ] in ngc 1068 and ngc 4151 respectively .
the arguments in favor of this line being primarily [ ] are threefold .
first , the line is not strong enough to account for the observed flux in the objects where this line is seen .
recombination cascade calculations @xcite indicate that the strength of ( @xmath69 m ) should be @xmath70% of the strength of the ( @xmath71 m ) line .
the latter line , blended with ( @xmath72 m ) and [ ] ( @xmath28 m ) , comprises emission which is frequently seen as a weak blue wing on the pa@xmath1 line profile .
however , when seen ( see also @xcite ) , the line blueward of [ ] is of comparable strength or stronger than the blend blueward of pa@xmath1 , arguing against its identification as .
secondly , the observed wavelength centroid is more consistent with [ ] than with .
if the line is identified as [ ] , it is blueshifted by @xmath73 km s@xmath11 relative to pa@xmath1 .
if the line is , its blueshift would be 174 km s@xmath11 .
since the ionization potential of he is only 24.5 ev , we would expect he to show a similar spatial and velocity structure to the h@xmath74 lines .
s@xmath75 , on the other hand , has a much larger ionization potential of 328 ev @xcite , and may arise from very different physical regions than does most of the pa@xmath1 and br@xmath4emission .
finally , the line we identify as [ ] is only seen in a small number of galaxies for which infrared spectra are available @xcite .
a line would be expected to be seen with a similar strength relative to pa@xmath1 in most seyfert galaxies , as the two lines should arise from similar physical regions .
the strength of a high ionization coronal line may vary greatly between objects , however , due to variations in the ionization conditions , as well as differences in extinction to the regions responsible for the h@xmath74 emission and the coronal emission .
we therefore conclude that the line observed at a rest wavelength of 1.252@xmath3 m is [ ] .
the [ ] line is very strong on the nucleus .
its equivalent width of 4.5@xmath76 m is nearly half that of [ ] .
it is not as spatially extended as the other infrared lines along the position angle of either slit .
it is marginally resolved spatially along the 30@xmath17 slit , showing a stronger flux ( [email protected] of the nuclear [ ] flux ) in the @xmath77 spatial bins than would be expected for atmospheric seeing of 0.8@xmath12 ( 0.08 of the nuclear flux ) .
it is not resolved along the 120@xmath17 slit .
table [ ngc4388eqw ] in appendix [ sec : tabappndx ] lists the equivalent widths and flux ratios of the infrared lines detected in ngc 4388 along both slits .
figure [ ngc4388fluxrats ] shows the [ ] /pa@xmath1 and h@xmath5/br@xmath4 flux ratios along both slits . on the nucleus ,
the [ ] /pa@xmath1 flux ratio is about 0.4 , a relatively low value for a seyfert 2 galaxy ; however , the [ ] /pa@xmath1 flux ratio rises off the nucleus along both slits . the [ ]
/pa@xmath1 ratio is largest where the [ ] velocity centers are most discrepant from the pa@xmath1 velocity centers , specifically to the southwest along the 30@xmath17 slit and to the southeast along the 120@xmath17 slit ( see section [ sec:4388velocitystruc ] ) . along the 30@xmath17 slit , which is coincident with the major axis of the optical [ oiii ] emission @xcite
, the h@xmath5/br@xmath4 ratio is @xmath71 on the nucleus , decreasing slowly with radius . along the 120@xmath17 slit , however , the flux of the h@xmath5 line does not diminish nearly as quickly as does the flux from the br@xmath4 line . at a distance 2@xmath12 southeast and northwest of the nucleus ,
the br@xmath4 line is @xmath78 of its nuclear strength , while the h@xmath5 line is @xmath79 of its nuclear strength ( see figure [ ngc4388spec120 ] ) .
this behavior is qualitatively similar to that seen in mk 1066 .
figure [ ngc4388lv30 ] is set of continuum - subtracted position / velocity contour plots for the emission lines observed along the 30@xmath17 slit for ngc 4388 .
there is an enhanced blue wing of pa@xmath1 on the nucleus , that is probably dominated by a combination of ( @xmath2=1.2786@xmath3 m ) and [ ] ( @xmath2=1.2788@xmath3 m ) , as discussed in section [ sec : mk1066 ] .
table [ ngc4388kinematics ] in appendix [ sec : tabappndx ] lists the centroid and width of each line in each spatial bin along both slits .
figures [ ngc4388vvspos30 ] and [ ngc4388vvspos120 ] show the velocity centroids of the line profiles as a function of position along the 30@xmath17 and 120@xmath17 slits respectively .
northeast of the nucleus along the 30@xmath17 slit , the pa@xmath1 , br@xmath4 , and [ ] lines agree in velocity centroid , and show only a small increase in velocity with distance from the nucleus . the h@xmath5 line is slightly redshifted relative to the other lines . southwest of the nucleus , however , the lines show very different velocity structures .
the pa@xmath1 , br@xmath4 , and [ ] lines turn towards increasing velocity with radius to the southwest , while the velocity gradient of h@xmath5 flattens .
the [ ] line curves back more sharply than the pa@xmath1 line , until its peak is about 100 km s@xmath11 redward of pa@xmath1 3@xmath12 south of the nucleus .
figure [ ngc4388closeups ] shows profiles of the pa@xmath1 and [ ] lines southwest of the nucleus , and indicates that the difference between [ ] and pa@xmath1 is not the result of errors in velocity measurements , but that the centroid and profile of [ ] is genuinely redshifted relative to pa@xmath1 .
br@xmath4 also differs from pa@xmath1 southwest of the nucleus , increasing in velocity closer to the nucleus than does pa@xmath1 . along the 120@xmath17 slit ,
the infrared line fluxes drop off faster with radius than they do along the 30@xmath17 slit . along this slit
, there is no evidence for significant velocity changes in the pa@xmath1 line .
where it was detected , the velocity peak of the br@xmath4 line is marginally consistent with that of the pa@xmath1 line , as is the [ ] line everywhere except for 1@xmath12 southeast of the nucleus , where [ ] is blueshifted by @xmath80 km s@xmath11 relative to pa@xmath1 .
the h@xmath5 line is always blueshifted relative to the br@xmath4 line by 50100 km s@xmath11 .
in addition to different variations in velocity centroid with position , the lines have differing profiles . along the 30@xmath17 slit , not only is [ ] generally wider than pa@xmath1 , and at a different velocity centroid southwest of the nucleus , but the asymmetries in the lines are different .
for instance , 23@xmath12 southwest of the nucleus the [ ] /pa@xmath1 ratio is preferentially enhanced at redshifted velocities ( figure [ ngc4388closeups ] ) .
these asymmetries suggest that systems of clouds with different flux ratios at different velocities are superimposed along the line of sight . along the 120@xmath17 slit , the [ ] line
has an enhanced red shoulder to the northwest of the nucleus .
again , the pa@xmath1 line is redshifted relative to the [ ] line , so this red shoulder may correspond to a component which is associated with the pa@xmath1 peak .
in their nuclear spectrum which was marginally resolved spatially along a 13@xmath17 slit , @xcite observed pa@xmath1 and [ ] with similar profiles .
this is in contrast to our data , where [ ] is clearly wider than pa@xmath1 near the nucleus along a 30@xmath17 slit .
narrowband optical imaging and spectroscopy ( e.g. * ? ? ?
* ; * ? ? ?
* ) , radio imaging @xcite , as well as this study and that of @xcite indicate that ngc 4388 is a complicated system with much spatial structure at all wavelengths .
consequently , the differences in observed line profiles almost certainly arise from differences in the spatial region sampled due to slit sizes and angles as well as seeing conditions .
similarly , @xcite observed a [ ] /pa@xmath1 flux ratio of @xmath80.7 , in comparison to our nuclear ratio of [email protected] .
since we observe a rapidly changing flux ratio with position , the difference between the two measurements of the nuclear ratio may also be easily ascribed to the different spatial region sampled .
@xcite report a faint broad ( full width at `` zero intensity '' of 6000 km s@xmath11 ) h@xmath19 line on the nucleus , while @xcite ( 1988 ; 1996 ) report the detection of a broad line in scattered h@xmath19 flux within 5@xmath12 of the nucleus .
figures [ ngc4388spec30 ] and [ ngc4388spec120 ] show no convincing evidence of a broad pa@xmath1 or br@xmath4 line . the upper limit on the flux of a broad ( fwhm@xmath81 km s@xmath11 ) pa@xmath1 line on the nucleus is @xmath82 erg @xmath83 s@xmath11 , corresponding to an upper limit in its equivalent width of @xmath84 m .
@xcite report on optical spectroscopy in slits oriented at a number of position angles , including two angles , 0@xmath17 and 60@xmath17 , which bracket our 30@xmath17 slit . to the north and east , within 5@xmath12 of the nucleus , the optical lines show a shallow gradient to the red with all line velocities within 100 km s@xmath11 of the systemic velocity , consistent with the infrared emission lines . optically , this gradient continues to the south and west , with lines more distant from the nucleus at lower velocity , in contrast to the increasing velocity seen in infrared br@xmath4 and [ ] emission .
based on the offset between the centroid of the ir continuum image of ngc 4388 obtained with the spectrometer in imaging mode and a star from the hubble guide star catalog 3@xmath85 away , the peak of the infrared continuum emission corresponds to the brighter , northeastern radio peak of two peaks separated by @xmath86 in the 4.86 ghz map of @xcite . as with mk 1066 ( section [ sec : mk1066 ] )
, there is a spatial correspondence in ngc 4388 between enhanced [ ] emission and extended radio emission . to the southwest close to the second peak of radio emission @xcite , the [ ] /pa@xmath1 ratio increases to values @xmath87 ( see figure [ ngc4388fluxrats ] ) , whereas within 1@xmath12 of the nucleus the [ ] /pa@xmath1 ratio is @xmath88 . to the northeast ,
the [ ] /pa@xmath1 ratio is also higher than it is on the nucleus , but is still not as large as the values to the southwest .
additionally , to the southwest the [ ] line profiles are most disparate from the pa@xmath1lines profiles , where the peak of the [ ] emission is redshifted relative to the peak of the pa@xmath1 emission .
the velocity field of ngc 4388 is known to be complicated .
there is a rotating disk with major axis at position angle @xmath89 and inclination @xmath90 @xcite .
however , in addition to this , there is blueshifted gas out of the plane of the galaxy to the north and south , as well as other localized sites within 20@xmath12 where the velocities of the gas are discrepant from pure rotation @xcite . in the infrared spectra
, we are probably observing at least two kinematic components within 4@xmath12 of the nucleus of ngc 4388 .
one kinematic component is the rotating disk , which is responsible for the velocities of the lines northeast of the nucleus , and for the velocity of the h@xmath5 line southwest of the nucleus ( figure [ ngc4388vvspos30 ] ) .
these velocities match most closely the velocities of the optical lines which @xcite identify with the rotation curve of the galaxy .
in addition to this rotating component , there may be a receding outflow to the southwest , associated with the radio emission @xcite .
the shoulders associated with this outflow pull the pa@xmath1 , br@xmath4 , and [ ] line centroids to higher velocities than would be the case from pure rotation .
the velocity difference is greater for [ ] than it is for pa@xmath1 because of the higher [ ] /pa@xmath1 ratio in the outflow than in the disk .
this receding outflow has a higher [ ] /pa@xmath1 ratio than the gas in the disk of the galaxy .
@xcite observed red components in the [ oiii ] line profiles along a 23@xmath17 slit ; to the southwest , the mean [ oiii ] velocities are redshifted from the peak [ oiii ] velocities by 5080 km s@xmath11 .
the br@xmath4 line southwest of the nucleus probably includes contributions from both the disk and outflow components .
the velocities of br@xmath4 here are redshifted relative to pa@xmath1 , and relative to optical emission lines that correspond to the rotation of the disk . additionally , the br@xmath4/pa@xmath1 ratio is higher to the southwest than it is to the northeast ( although it is highest on the nucleus ) .
this suggests that the emission associated with the outflow is viewed through substantial extinction in the disk of the galaxy .
the [ ] emission also increases off the nucleus along the 120@xmath17 slit which is perpendicular to the putative outflow axis .
this increased emission does not correspond to notable differences in emission at other wavelengths .
it is likely that this strong [ ] emission is associated with the disk of the galaxy , as it shares a similar ( although not identical ) velocity gradient with the other lines ( figure [ ngc4388vvspos120 ] ) .
the h@xmath5 emission appears to be originating primarily in the disk of the galaxy , as its velocity as plotted in figures [ ngc4388vvspos30 ] and [ ngc4388vvspos120 ] follows closely the pa@xmath1 and optical line velocities identified with the rotating disk @xcite .
the h@xmath5/br@xmath4 ratio diminishes away from the nucleus along the direction of the radio emission , suggesting that the h@xmath5 emission is not associated with the radio jets or the outflow .
similar to mk 1066 , the h@xmath5/br@xmath4 ratio shows a striking increase along the slit perpendicular to the outflow axis . along this direction
, the h@xmath5emission appears to share a similar spatial structure to the [ ] emission .
the ratios of both lines to hydrogen recombination lines ( figure [ ngc4388fluxrats ] ) and the velocity curve of both lines ( figure [ ngc4388vvspos120 ] ) along a position angle of 120@xmath17 share the same shape ( although the h@xmath5 emission increases faster relative to br@xmath4 than the [ ] emission does relative to pa@xmath1 ) . the mechanism of excitation of the [ ] and
h@xmath5 lines in the disk of ngc 4388 at radii of @xmath32150pc are probably related . 1
. strong pa@xmath1 , br@xmath4 , [ ] , and h@xmath5 emission is seen from ngc 4388 .
these lines are observed to be extended by @xmath91 along the direction of greatest extent of optical [ oiii ] line emission .
infrared line emission is seen extended by @xmath21 perpendicular to this direction .
we identify a line at the rest wavelength of 1.252@xmath3 m with the infrared coronal line [ ] .
this line is only marginally spatially resolved at a position angle of 30@xmath17 . on the nucleus ,
its flux is 1/6 the flux of pa@xmath1 .
the [ ] /pa@xmath1 flux ratio is lowest on the nucleus with a value of 0.4 .
it rises to values @xmath92 southwest of the nucleus where the velocity centers of the pa@xmath1 and [ ] lines are most different , spatially coincident with a secondary peak of radio emission .
the h@xmath5/br@xmath4 flux ratio is highest ( @xmath322 - 3 ) away from the nucleus along the 120@xmath17 slit , perpendicular to the axis of the extended [ oiii ] emission . along this direction ,
the h@xmath5/br@xmath4 and [ ] /pa@xmath1 ratios both increase away from the nucleus . along the 30@xmath17 slit ,
the two line ratios show opposing trends , with the h@xmath5/br@xmath4 ratio increasing and the [ ] /pa@xmath1 ratio decreasing with distance from the nucleus .
4 . northeast of the nucleus , the infrared lines in ngc 4388 show structure which is similar to the optical line measured by @xcite . southwest of the nucleus
, the br@xmath4 and [ ] lines are redshifted relative to pa@xmath1by 50150 km s@xmath11 .
5 . in some cases where different lines have different velocity centroids , there is substructure suggestive that the same velocity components may be present in every line with different strengths .
specifically , southwest of the nucleus where the [ ] emission is redshifted relative to the pa@xmath1emission , there are enhanced blue wings in [ ] and enhanced red wings in pa@xmath1 . 6 .
a possible explanation for the velocity structures seen in the infrared lines is that there are at least two kinematical components .
most velocities are governed primarily by a rotating disk .
there may additionally be an obscured redshifted outflow southwest of the nucleus .
the coincidence of strong , redshifted [ ] emission to the southwest may be evidence of shock - excited [ ] emitting gas associated with this outflow .
mk 3 is an s0 galaxy at a systemic velocity of 4046 km s@xmath11 @xcite .
this gives a calculated distance of 54 mpc , which in turn yields a spatial scale of 262 pc arcsec@xmath11 .
narrowband optical imaging with the hst in several lines include [ oiii ] and h@xmath19 shows that the nlr of the galaxy is s - shaped @xcite . on small spatial scales of @xmath93 ,
the nlr is at a position angle of 86@xmath17 . on larger spatial scales of a few arcseconds ,
the apparent position angle of the nlr is 113@xmath17 @xcite .
radio observations at 14.9 ghz show a linear structure oriented at a position angle of 86@xmath17 @xcite . there are two radio peaks , the stronger one 1.1@xmath12 west of the nucleus , the weaker one 0.4@xmath12 east of the nucleus .
higher resolution maps of the source at 5 ghz resolve the linear structure into a number of components . in particular , the eastern lobe is resolved into three components aligned along the axis of the source , together with a third component coincident with the core of the galaxy .
mk 3 is also quite bright in x - rays , with an x - ray luminosity of @xmath94 erg s@xmath11 ( with h@xmath95 km s@xmath11 mpc@xmath11 ) , and a very high absorbing column of @xmath96 @xmath83 @xcite .
the 0.12 kev x - ray source is unresolved with the rosat hri on a scale of @xmath56 @xcite .
the reduced 1.29@xmath3 m and 2.21@xmath3 m spectra of mk 3 along a 113@xmath17 slit are plotted in figure [ mk3spec ] , in spatial bins along the slit separated by 1@xmath12 .
the position angle of the 113@xmath17 slit was chosen to align with the major axis of the [ oiii ] distribution on scales of @xmath93 @xcite .
the [ ] and pa@xmath1 emission lines are visible in the 1.2@xmath3 m spectra , and the h@xmath5 and br@xmath4 emission lines are visible in the 2.2@xmath3 m spectra .
note that for these observations , the br@xmath4 line was centered in the k - band spectrum so that the h@xmath5 line is near the blue edge of the spectrum .
this may lead to inaccuracies in placing the continuum on the blue side of the h@xmath5 , resulting in underestimates of the flux of the h@xmath5 line and any blue wings in the profile of the h@xmath5line .
the [ ] , pa@xmath1 , h@xmath5 , and br@xmath4 lines are extended along the slit by @xmath97 ( 1.8 kpc ) .
the equivalent widths of the lines on the nucleus , given in table [ mk3eqw ] in appendix [ sec : tabappndx ] , are larger by a factor of @xmath92 than those seen in most seyfert 2 nuclei @xcite .
most of the extended line emission is characteristically narrow ( fwhm@xmath98 km s@xmath11 ) , yet on the nucleus the infrared lines are quite broad ( see figure [ mk3nucprof ] ) .
the fwhm of the nuclear lines are br@xmath4@xmath8700 km s@xmath11 , pa@xmath1 @xmath81000 km s@xmath11 , and [ ] @xmath81150 km s@xmath11 .
these lines are also markedly asymmetric , with strong blue wings .
the peak of the pa@xmath1 line on the nucleus is at a velocity @xmath99 km s@xmath11 greater than the centroid of the full line provile .
although [ ] is not obvious , it could easily be hidden within the broad blue wings of the [ ] line profile while the observed width of the pa@xmath1 line on the nucleus is comparable to the fwhm of broad pa@xmath1 lines seen in some seyfert 1 galaxies ( e.g. , * ? ? ?
* ) , the observed line profiles in mk 3 probably do not indicate the detection of a canonical broad line region , since [ ] and pa@xmath1 lines are both broad and have similar profiles . because the critical density of [ ] is only n@xmath100 , the broad [ ] emission
can not be produced in a high density ( @xmath101 ) broad line region .
some or all of the high velocity pa@xmath1 line flux is therefore probably coming from the nlr .
table [ mk3eqw ] in appendix [ sec : tabappndx ] lists the equivalent widths and flux ratios of the lines observed in mk 3 .
figure [ mk3fluxrat ] shows the flux ratios of [ ] /pa@xmath1 and h@xmath5/br@xmath4 as a function of position along the slit .
the [ ] /pa@xmath1ratio changes by a factor of nearly 4 along the slit .
this change is a monotonic increase from east to west , although it is poorly determined 3@xmath12 west of the nucleus due to the low signal to noise ratio in the pa@xmath1 line .
the h@xmath5/br@xmath4 flux ratio , close to @xmath102 in all but one spatial bin , is relatively low for seyfert galaxies .
the h@xmath5/br@xmath4 flux ratio is anomalously large 2@xmath12 west of the nucleus .
the strong , double - peaked h@xmath5 emission responsible for this enhanced flux ratio is visible in figure [ mk3spec ] .
the higher velocity peak is unresolved , and is at a velocity of 450 km s@xmath11 relative to the systemic velocity of mk 3 .
this peak is not well detected in every individual image that goes into the summed spectrum of mk 3 , and so we suspect that this peak is an artifact associated with poor background subtraction .
because h@xmath5 at the redshift of mk 3 falls right at the wavelength of an atmospheric oh line , the spatial bins in which the detection of h@xmath5 is of low significance may be containmated by poor subtraction of this line .
if this is the case , then the h@xmath5/br@xmath4ratio would be @xmath103 , which is still higher than in the other spatial bins , although the caveat about background subtraction applies to the full h@xmath5 profile in this spatial bin .
[ sec : mk3velsubstruc ] figure [ mk3lv ] shows position versus velocity plots for each of the four emission lines seen in mk 3 .
a smoothed continuum has been subtracted from the data in each case .
table [ mk3kinematics ] in appendix [ sec : tabappndx ] lists the velocity centroids and total full widths at half maximum of the lines in each of the 1@xmath12 spatial bins for which spectra are plotted in figure [ mk3spec ] . to the east ,
the pa@xmath1 line is narrow and stronger than any narrow pa@xmath1emission to the west .
a very shallow velocity gradient at the limit of our resolution is marginally visible in this narrow component .
the pa@xmath1 line is redshifted by @xmath73 km s@xmath11 @xmath104 east of the nucleus .
the br@xmath4 line is not as well detected as the pa@xmath1 line , but appears to share a similar velocity structure , although the shallow velocity gradient is not apparent in br@xmath4 . the br@xmath4 line is slightly blueshifted relative to the pa@xmath1 line east of the nucleus .
the pa@xmath1and br@xmath4 line peaks appear to share a shallow velocity decrease from east to west .
the total blueshift of the lines is @xmath51 km s@xmath11west of the nucleus .
the velocity structure of the [ ] line to the east is different from that of the hydrogen recombination lines . to the east of the nucleus and on the nucleus , the peak of the [ ] line is blueshifted by @xmath8100200 km s@xmath11relative to the pa@xmath1 and br@xmath4 lines . west of the nucleus , the velocity of the [ ] line is closer to that of the hydrogen recombination lines .
the h@xmath5 line is not confidently detected east of the nucleus .
west of nucleus by 2@xmath12 , the h@xmath5 line shows the aforementioned peculiar double - peaked shape ( see figure [ mk3spec ] ) . on the nucleus ,
the [ ] , pa@xmath1 , and br@xmath4 lines all have qualitatively similar profiles .
all three lines show strong enhanced emission to the blue of the peak .
however , as is visible in figure [ mk3nucprof ] , the profiles differ in detail . in br@xmath4 ,
the blue wing is not as strong relative to the peak as it is for pa@xmath1 and [ ] .
the peak of the [ ] line profile is shifted to the blue of the pa@xmath1 line profile by 80 km s@xmath11 , and the fwhm of [ ] is about 100 km s@xmath11 larger than the fwhm of pa@xmath1 , apperently due to enhanced emission on the blue wing of [ ] ( see figure [ mk3nucprof ] ) .
while the profiles to the east of the nucleus are narrow , figure [ mk3lv ] clearly shows that 12@xmath12 west of the nucleus the lines are broader , with the fwhm of pa@xmath1 line equal to 500600 km s@xmath11 .
figure [ mk3closeups ] shows the profiles of the br@xmath4 , pa@xmath1 , and [ ] lines 1@xmath12 west of the nucleus .
two features are apparent from this plot .
first , the red wing of the br@xmath4 line is stronger relative to the line peak than the red wing of pa@xmath1 .
this may indicate that there is receeding gas obscured behind dust , as the longer wavelength br@xmath4 is less affected by extinction than pa@xmath1 .
second , the [ ] /pa@xmath1 ratio is largest with a value of @xmath105 at a redshifted velocity near @xmath106 km s@xmath11 . the blue hump on the pa@xmath1 and [ ] line profiles has an [ ] /pa@xmath1 ratio of about 1 , close to the nuclear flux ratio ( see section [ sec : mk3fluxrat ] ) .
the velocity gradient seen in the off - nuclear narrow pa@xmath1 line is generally consistent with the rotation curve seen in optical h@xmath1 and [ oiii ] emission by @xcite .
however , the optical slit position angle is 86@xmath17 , while the infrared data presented here was obtained through a slit at a position angle of 113@xmath17 . east of the nucleus , the [ oiii ] line peaks at 120 km s@xmath11 , while west of the nucleus it peaks at -120 km s@xmath11 , similar to the limiting off - nuclear velocities of @xmath107 km s@xmath11 in the infrared lines .
@xcite also saw an increase in the fwhm of the optical h@xmath1 and [ oiii ] lines near the nucleus , similar to that seen in the pa@xmath1 , br@xmath4 , and [ ] lines .
this broader emission is blueshifted relative to the peak of the lines , again consistent with that seen in the infrared lines .
in addition , a red wing in the [ oiii ] and h@xmath1 profiles is similar to the enhanced red emission seen in the br@xmath4 and [ ] lines in figure [ mk3closeups ] , which was also strongest 1@xmath12 west of the nucleus .
a gaussian decomposition of the optical lines placed the relative velocity of the redshifted component at @xmath108 km s@xmath11 ; our gaussian decomposition of the infrared lines places the velocity of the reddened component of the infrared lines at somewhat higher velocity , closer to @xmath37 km s@xmath11 , although this is uncertain given the complex blended line profiles ( see figure [ mk3closeups ] ) .
the spatial location of this red wing is coincident with the stronger of two radio peaks seen at 2 cm @xcite .
the relatively high extinction seen for this reddened component in the infrared br@xmath4 and pa@xmath1 lines is consistent with the picture of this radio peak as a lobe coincidental with a receding outflow of ionized gas which is obscured behind the plane of the galaxy .
the weaker radio lobe is 0.4@xmath12 east of the nucleus along the same position angle .
@xcite do not identify any component in the optical spectra with the eastern radio lobe .
however , the enhanced blue wings seen on the nucleus in both the infrared and [ oiii ] emission lines may be part of an approaching outflow associated with the eastern radio lobe . as with mk 1066 and
ngc 4388 , and to a lesser extent ngc 2110 ,
the [ ] emission in mk 3 shows spatial correlations with the radio emission which suggests that processes involving the radio jet s interaction with circumnuclear gas are significantly enhancing the [ ] emission . west of the nucleus at the position of the western radio peak , the [ ] line shows enhanced redshifted emission , which is also seen in the pa@xmath1 and br@xmath4lines as well as in the optical lines @xcite . at the position of this radio peak west of the nucleus , the [ ] /pa@xmath1 ratio is higher than it is to the east of the nucleus where the radio emission is much weaker .
this enhanced redshifted emission 1@xmath12 west of the nucleus also has a larger br@xmath4/pa@xmath1 flux ratio , suggesting higher extinction to the region which is the source of the enhanced [ ] emission .
this is consistent with the picture of the region as a receding outflow which is partially obscured by the galactic disk .
the eastern radio peak is too close ( 0.4@xmath12 ) to the nucleus to distinguish infrared line emission spatially coincident with it from emission directly from the nucleus .
however , if the two knots of radio emission represent the radio lobes associated with a biconical outflow , the enhanced blue wings in the infrared and optical lines on the nucleus may in fact be correlated with the second radio lobe . higher resolution 5 ghz maps of mk 3
@xcite resolve the eastern lobe into a number of individual linearly spaced components .
the infrared lines near the nucleus show a complicated , evidently multi - component structure in velocity ( see figure [ mk3spec ] ) .
the multi - component spatial nature of the radio emission and the multiple velocity components in the infrared line emission suggest the identification of the broad blue wings on the infrared lines near the nucleus with the eastern radio lobe .
the [ ] emission on the nucleus is blueshifted relative to the pa@xmath1 emission , suggesting that the [ ] /pa@xmath1 flux ratio may be enhanced at the position of the eastern radio peaks .
given this spatial correspondence , it is likely that the [ ] emission is enhanced by fast shocks resulting from the interaction of the radio jet and the circumnuclear gas . except for the anomalously high value 2@xmath12 west of the nucleus , the h@xmath5/br@xmath4 ratio is lower in mk 3 than it is in most seyfert galaxies . west of the nucleus
, the narrow pa@xmath1 and perhaps br@xmath4 emission may be associated with the rotation of a galactic disk , based on similarities with optical lines @xcite . at these positions ,
the h@xmath5 emission is suppressed relative to the br@xmath4 emission , indicating that mk 3 continues relatively little warm molecular gas . 1 . at a position angle of 113@xmath17 , the infrared line emission is observed to be extended by @xmath97 .
2 . there is a fwhm@xmath81000 km s@xmath11 blue wing visible on the nucleus of mk 3 , most most prominent in [ ] and pa@xmath1 but also seen in br@xmath4 .
there is a red wing on the line profiles , arising from a region of relatively high extinction , @xmath93 west of the nucleus .
this red wing corresponds to a redshifted component seen in optical emission @xcite .
both wings are spatially consistent with the position of two radio peaks @xcite .
3 . the pa@xmath1 and br@xmath4 narrow line peaks may follow a barely resolved smooth velocity gradient which lies at @xmath109 km s@xmath11 east of the nucleus and @xmath110 km s@xmath11 west of the nucleus .
east of the nucleus , the br@xmath4 line is marginally blueshifted relative to pa@xmath1 .
the [ ] emission is blueshifted relative to the pa@xmath1 emission both east of the nucleus , and significantly so on the nucleus .
the [ ] /pa@xmath1 flux ratio monotonically increases from @xmath111 to @xmath112 over 5@xmath12 from east to west across the nucleus .
the h@xmath5/br@xmath4 flux ratio is close to @xmath102 everywhere except where it is anamalously high , possibly as a result of an artifact , 2@xmath12 west of the nucleus .
we suggest that emission associated with an outflow of ionized gas contributes significantly to the infrared line emission in mk 3 .
the [ ] /pa@xmath1 ratio is enhanced in components identified with this outflow and coincident with the 2 cm radio lobes , suggesting a possible shock contribution to the [ ] emission .
the low h@xmath5/br@xmath4 ratio indicates that in contrast to other sefyert galaxies , the circumnuclear disk is relatively depleted in warm molecular gas .
one obvious conclusion from the data presented here is that the [ ] /pa@xmath1ratio in seyfert galaxies varies significantly with position on scales of @xmath113 . of
the four galaxies observed , each showed a different dependence of [ ] /pa@xmath1 on position , whether it is monotonically increasing across the nucleus ( mk 3 ) , highest at the nucleus ( ngc 2110 ) , lowest at the nucleus ( ngc 4388 ) , or close to constant with point - to - point variations ( mk 1066 ) .
these data show that flux ratios integrated over large nuclear beams may be of limited value for drawing conclusions about the excitation processes of the infrared lines .
the flux ratios are observed to vary significantly on scales of @xmath114 pc , and to show significant velocity structure .
circumnuclear starbursts and processes associated with the central agn directly have both been implicated in the generation of [ ] emission lines seen in seyfert galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . our data in general support the active nucleus as the power source of [ ] in the seyfert galaxies .
despite the differences in the details of the velocity structure of the infrared lines , the velocity gradients of [ ] tend to match those of pa@xmath1 and br@xmath4 .
the [ ] line can generally be kinematically associated with other spectral features linked with the nuclear activity ( e.g. , [ oiii ] ) , or sites of enhanced radio emission .
the one exception is the extended [ ] emission along position angle 120@xmath17 in ngc 4388 .
there are two primary proposed mechanisms for [ ] emission associated with the agn itself .
because the ionization potential of fe@xmath74 is only 16.2 ev , the [ ] line will be strong in partially ionized regions .
if clouds illuminated by the seyfert nucleus have a high enough column density , only the side of the cloud facing the nuclear ionizing source will be fully ionized .
because the ionization cross section for high - energy x - rays is relatively small , x - rays can penetrate deeply into these dense clouds , creating large warm partially ionized regions suitable for [ ] emission @xcite .
fast shocks can also produce partially ionized regions heated to 6000 k @xcite .
models indicate that these shocks may also destroy dust grains @xcite .
because much of the iron in the local interstellar medium is locked up in dust grains , the destruction of those grains would enhance the gas phase fe abundance , thereby increasing the infrared [ ] emission .
@xcite report a slightly stronger correlation between infrared [ ] and 6 cm radio emission than between pa@xmath1 and 6 cm emission .
based on the difference in the correlations , @xcite conclude that @xmath115% of the [ ] emission in seyfert galaxies is due to the action of fast shocks , and that the bulk of it is due to x - ray photoionization .
the data presented here for mk 1066 , ngc 2110 , ngc 4388 , and mk 3 , combined with the calculations of @xcite , provide spatial and kinematic evidence that indicate that x - ray photoionization is responsible for most of the [ ] emission , but that the [ ] emission is probably enhanced in some places by a contribution from shocks . in most cases the [ ] emission tends to be spatially associated with the pa@xmath1 emission , and the primary component of [ ] often shares the same velocity structure as the pa@xmath1 line .
thus , most of the [ ] emission probably comes from gas physically and kinematically associated with the nlr , as traced by the pa@xmath1 emission .
except for the nucleus of ngc 2110 , the flux ratios observed at various positions in the galaxies are within the range of those seen in the spectra of @xcite .
while the bulk of the [ ] appears to be photoionized , there are clear signatures of local enhancements of the [ ] emission from shocks associated with the expanding radio plasma . in all four of the galaxies ,
spectral features were identified in the [ ] emission line profile which occured at or near the position of peaks in the non - thermal radio emission . in many cases
, these features showed relatively high [ ] /pa@xmath1 ratios . in mk 3 and
ngc 4388 , the [ ] emission was enhanced relative to the pa@xmath1 emission , or showed the largest kinematical differences from the pa@xmath1 emission , at positions spatially coincident with radio peaks .
this suggests that processes associated with radio jets are contributing measurably to the [ ] emission .
the most probable explanation of this is fast shocks resulting from the interaction of outflowing radio plasma with the circumnuclear gas .
molecular hydrogen emission can result from ultraviolet fluorescence of hydrogen molecules in photodissociation regions illuminated by photons from ob stars or by a nuclear uv continuum source .
alternatively , the h@xmath5emission may be excited thermally , through either slow ( v@xmath36 km s@xmath11 ) shocks @xcite , or through x - ray heating of dense molecular clouds @xcite .
observations of the flux ratio between h@xmath5 ( @xmath6=10 s(1 ) ) ( @xmath2=2.1212@xmath3 m ) and h@xmath5 ( @xmath6=21 s(1 ) ) ( 2.2471@xmath3 m ) rule out ultraviolet fluorescence as a significant mechanism in most seyfert galaxies @xcite .
in all four of the galaxies discussed here , we have identified features in [ ] and hydrogen recombination line profiles which we associate with outflows or other interactions with radio lobes and jets . except for ngc 2110 ,
these features are weaker or invisible in the h@xmath5 line profile ; the kinematics of the h@xmath5 are usually better explained by galactic rotation . in addition , in those two galaxies where we have data perpendicular to the outflow axis ( mk 1066 and ngc 4388 ) , and possibly as well in ngc 2110 ( s99 ) , the h@xmath5 is significantly more extended than the atomic hydrogen recombination lines , with the h@xmath5/br@xmath4 flux ratio increasing by a factor of @xmath84 or more at 200300 pc away from the nucleus . although x - ray heating of molecular clouds that cross into the ionization cones can give rise to the extended h@xmath5 emission along the cone axes , it is unlikely that the nucleus is responsible for the enhanced h@xmath5 emission along the axis perpendicular to the cones . according to the standard model of seyfert galaxies , the radiation from the central source
is obscured except along the axis of the ionization cones ; thus , nuclear x - rays can not penetrate significantly into the gas perpendicular to the cone axis to produce the extended h@xmath5 emission .
while these x - rays would heat the inner face of the obscuring molecular torus , the size of that torus is believed to be on the order of a few parsecs , so this can not explain the extended h@xmath5 emission that we observe .
also , if molecular hydrogen emission does emerge from photodissociation regions heated by x - rays from a central source , we would expect strong br@xmath4 emission illuminated by ultraviolet radiation from the same central source .
shocks driven directly or indirectly by the nucleus are another possible explanation for the excitation of the h@xmath5 emission .
slow shocks moving through high density molecular clouds in the disk of the galaxy could excite molecular hydrogen emission . where the shocks encounter lower density clouds
, they should be faster , and produce substantial [ ] emission .
the [ ] /pa@xmath1 ratio does rise with distance from the nucleus , at least initially , perpendicular to the primary direction of extended high - excitation gas in both mk 1066 and ngc 4388 . alternatively , secondary shocks in molecular clouds circumscribing radio outflows might give rise to enhanced h@xmath5 emission .
this process might produce the differing [ ] and h@xmath5 spatial morphologies observed in ngc 1068 by @xcite .
one problem with this scenario is that , in contrast to the axis of ioniziation cones where outflowing radio plasma may drive shocks into the surrounding gas , there is no obvious mechansim to drive shocks to distances of @xmath81 kpc perpendicular to the cone axis .
the relatively large extent of the h@xmath5 emission perpendicular to the ionization cones in mk 1066 and ngc 4388 suggest that there are local source of heating which is giving rise to much of the h@xmath5 emission . in both galaxies
, there may be contribution to the h@xmath5 emission from young stars heating molecular clouds .
while this may contribute to the h@xmath5 emission , it is unlikely to be the only source , as the h@xmath5/br@xmath4 ratios observed in the extended h@xmath5 gas tend to be higher than the values @xmath88 typically observed in starburst galaxies @xcite .
ngc 4388 has a bar oriented at position angle @xmath61 @xcite .
dynamical processes associated with this bar may be creating shocks which give rise to the observed h@xmath5 emission . in mk 1066 , it is also possible that the bulk motions of clouds which do not move completely with the rotation of the disk are colliding with other clouds , creating slow shocks which excite the cold molecular gas . in mk 3 , the lower h@xmath5/br@xmath4 ratio combined with the narrower h@xmath5 line profile suggest that local star formation activity may be able to explain most of the h@xmath5 emission , with no need for a contribution from outflowing gas associated with the nucleus . by contrast ,
in ngc 2110 , where wings associated with the outflow are seen in h@xmath5 , nuclear x - ray heating may be a more significant contributor .
1 . we have observed strong infrared pa@xmath1 , br@xmath4 , [ ] , and h@xmath5 line emission which is spatially extended on scales of a few hundred parsecs in mk 1066 , ngc 2110 , ngc 4388 , and mk 3 . 2 .
we have detected the near infrared coronal line [ ] ( @xmath2=1.2525@xmath3 m ) in the nucleus of ngc 4388 . in contrast to the other infrared lines , this line is only marginally spatially extended .
there is no apparent [ ] emission from mk 1066 , ngc 2110 , or mk 3 .
this is similar to what is seen optically , that high ionization species are visible in some but not all seyfert galaxies . 3 .
while often showing similar velocity trends over the central few arcseconds , the [ ] , h@xmath5 , and h lines have profiles which differ in detail from each other over the full spatial range probed by these observations .
the peaks of lines are shifted relative to each other at some spatial positions in some galaxies , and some species show enhanced blue and red wings relative to other species .
these differences are usually due to the effects of extinction , an enhancement to the [ ] emission associated with other indicators of the action of an outflow , or a possible local source of heating of h@xmath5 .
the [ ] /pa@xmath1 and h@xmath5/br@xmath4 flux ratios can vary dramatically on scales of 10@xmath116 pc near the nuclei of seyfert 2 galaxies .
the details of the spatial structure of the [ ] /pa@xmath1 ratio frequently support the notion that partially ionized regions created by x - ray photoionization from the nucleus , or shocks from the interaction of outflowing gas with ambient gas , or both , are enhancing the circumnuclear [ ] emission .
the spatial variation of the h@xmath5/br@xmath4 ratio is again frequently ascribable to local heating of h@xmath5 .
5 . the [ ] emission appears to be associated with the nuclear activity , not requiring the invocation of a circumnuclear starburst in any of these four galaxies .
the [ ] can be kinematically and spatially linked to processes dominated by the active nucleus
. in all of these galaxies , there are correlations between features in the [ ] emission and both radio jets and optically identified outflows , suggesting that some [ ] emission is produced by fast shocks resulting from interactions between radio jets and circumnuclear gas .
the h@xmath5 emission , although peaked on the nucleus , is relatively strong compared to hydrogen recombination and [ ] emission perpendicular to the axis of ionization cones or the major axis of high - excitation gas in mk 1066 and ngc 4388 .
the spatial distribution and line profile of h@xmath5 generally differs from that of [ ] . a local source such as bar - driven shocks or circumnuclear star formation
may be partially responsible for the h@xmath5emission in these galaxies .
the authors would like to thank the staff and night assistants at palomar observatory for assistance in completing these observations .
additionaly , they would like to thank james larkin for discussion and assistance in the earlier phases of the project .
in this appendix are tables with data from the infrared spectra of mk 1066 , ngc 2110 , ngc 4388 , and mk 3 . tables [ mk1066eqw ] through [ mk3eqw ] list the equivalent widths for the integrated line profile for each species observed in each spatial bin .
they additionally list relevant flux ratios .
tables [ mk1066kinematics ] through [ mk3kinematics ] give the centroids and line widths ( fwhm ) for each species in each spatial bin . | we present near - infrared spectra with resolutions of @xmath0 in the emission lines pa@xmath1 , [ ] ( @xmath2=1.2567@xmath3 m ) , br@xmath4 , and h@xmath5 @xmath6=10s(1 ) of the nuclei and circumnuclear regions of the four seyfert 2 galaxies mk 1066 , ngc 2110 , ngc 4388 , and mk 3 .
all of these galaxies show strong near - infrared line emission that is detected at radii several times the spatial resolution , corresponding to projected physical scales of 0.07 to 0.7 kpc .
velocity gradients are detected in these nuclei , as are spatial variations in line profiles and flux ratios .
we compare the spatial and velocity distribution of the line emission to previously observed optical line and radio emission .
the evidence indicates that the [ ] emission is associated with the seyfert activity in the galaxies .
our data are consistent with x - ray heating being responsible for most of the [ ] emission , although differences in [ ] and pa@xmath1 line profiles associated with radio emission suggests that the [ ] emission is enhanced by fast shocks associated with radio outflows .
the h@xmath5emission is not as strongly associated with outflows or ionization cones as is the emission in other lines , but rather appears to be primarily associated with the disk of the galaxy . |
the discovery of roughly two dozen tdes with x - ray and optical / uv telescopes ( e.g. , ) has afforded us the opportunity to study accretion over a wide range of rates and is also promising for the discovery and characterization of otherwise quiescent supermassive black holes ( smbhs ) @xcite .
standard treatments predict a temperature that is @xmath5 k near the peak of the burst and decreases with the luminosity @xcite .
however , although x - ray detected tdes reach this temperature , the tdes discovered via optical / uv observations have lower temperatures @xmath0 k , and those temperatures remain nearly steady even as the luminosity of the sources drop by @xmath6 orders of magnitude ( e.g. , * ? ? ?
* ; * ? ? ?
this has been explained as a consequence of either an optically thick shroud of gas at many times the radius of the disk @xcite or a wind from the disk @xcite . why do the processes which reduce the temperature in optical / uv detected tdes fail to operate for the x - ray detected tdes that have temperatures of @xmath7 k ?
the ranges in length and time scales required for full hydrodynamic simulations of debris circularization in tidal disruptions of a main - sequence star by a smbh from a marginally bound orbit mean that such simulations are currently computationally infeasible .
the main reason is that such simulations need to be able to follow the debris orbit with radius @xmath8 , and at the same time need to resolve the thickness of the stream which , if is self - gravitating , is much less than @xmath9 @xcite .
the addition of realistic cooling mechanisms and general relativistic hydrodynamics would make simulations even more time - consuming .
however , studies of stars disrupted by smbhs from initially bound orbits @xcite and white dwarfs disrupted by intermediate mass black holes @xcite have revealed that debris circularization is more difficult than previously thought . the nozzle shock and
the instabilities at pericenter are not strong enough to circularize the debris quickly @xcite .
instead , tidal stream intersection is the most effective way to produce shocks and dissipate the debris orbital energy @xcite .
this intersection is a result of the orbital apsidal precession of the debris on the same plane around the smbh .
however , if apsidal precession is small and intersection happens far from the pericenter of the initial orbit , then the collision is mild and circularization can be delayed substantially ( see , e.g. , @xcite and the newtonian simulation of @xcite ) . in this case
, most of the matter would take much longer to accrete onto the black hole than it would in the standard picture . as a consequence ,
the luminosity and the disk temperature would both be less than those seen in x - ray detected tdes .
here we propose that the high temperatures of some tdes can be explained in a picture in which the debris disk is small due to relatively large apsidal precession .
in particular , we argue that if the disk is small enough that the initial temperature exceeds @xmath10 k , then ( unlike when @xmath11 k ) the opacity is only weakly sensitive to the temperature , which eliminates the strong dependence of wind rates on temperatures that features in one explanation for the lower temperatures seen in optical / uv tdes @xcite . in section 2
we perform calculations of ballistic motion to explore how debris stream intersection and the consequent disk size depend on the black hole mass @xmath12 and the stellar orbit penetration parameter @xmath2 . in section 3
we calculate the disk temperature , where we show that high-@xmath2 encounters , particularly with black holes of @xmath13 , are the best candidates for tdes with temperatures @xmath14 k. further discussions and some observational considerations are given in section 4 .
as the nozzle shock is weak , the trajectory of the debris is nearly ballistic until the tidal stream self - intersects .
although some self - intersection would occur even in newtonian gravity because there is a spread in debris binding energy , the dynamics of the intersection are dominated in our case by general relativistic pericenter precession . if we assume for simplicity a non - rotating ( schwarzschild ) black hole ( rotational corrections are mild unless the pericenter is very close to the hole ) , then the precession angle @xmath15 over a single orbit is , to first order , @xmath16 ( @xcite ) , where @xmath17 is the semi - major axis of the orbit , and @xmath18 is the orbital eccentricity . here and
henceforth we use geometrized units in which @xmath19 .
if the orbit is close to the hole , the exact precession rate is larger than the rate given by this expression , but the differences are mild for @xmath20 . in reality
the ellipse precesses continuously , but for the high - eccentricity orbits we consider almost all the precession occurs at pericenter .
we therefore treat the debris orbit as a closed ellipse with an instantaneous shift of @xmath15 at pericenter passage , as illustrated in figure [ intersection ] .
the head of the stream essentially follows the most bound debris orbit , and after passing pericenter it intersects the trailing part of the incoming stream ( the shifted orbit is represented by the blue ellipse in figure [ intersection ] ) .
the incoming stream lies between the most bound debris orbit ( the black ellipse ) and the marginally bound debris orbit ( the black dotted curve ) .
if the initial trajectory of the star is nearly parabolic , intersection typically happens in a time after disruption of @xmath21 @xcite , where @xmath22 is the orbital period of the most bound orbit .
we now study the dynamics of the most bound orbit for the disruption of a star of mass @xmath23 and a radius @xmath24 , where @xmath25 is the radius of the sun , by a smbh with mass @xmath26 .
a star on an initially parabolic orbit has zero specific binding energy : @xmath27 where @xmath28 is the tidal radius and @xmath29 is the orbital speed at @xmath30 .
if we neglect the rotation of the star and the redistribution of energy during the compression / rebound processes , the most bound orbit has energy : @xmath31 because @xmath32 .
the semi - major axis and eccentricity of this orbit are @xmath33 and @xmath34 here @xmath35 is the pericenter distance , and we have defined the penetration factor @xmath36 .
the orbital period of the most bound debris is : @xmath37 ( note that we neglect the weak increase in @xmath22 with increasing @xmath2 found by @xcite ) .
therefore , the most bound orbit in a main - sequence star
smbh disruption is highly eccentric and its orbital period ranges between @xmath38 month ( @xmath39 ) and @xmath38 year ( @xmath40 ) .
as discussed previously , the narrowness of the tidal stream means that the intersection radius of the tidal streams is given by apsidal precession .
treating the most bound orbit as an ellipse with an instantaneous pericenter shift @xmath15 , a geometrical calculation shows that the original ellipse and the shifted ellipse intersect at a radius @xmath41 figure [ radiusangle](a ) shows that this stream intersection radius changes sensitively with @xmath12 and @xmath2 ( hereafter we use a solar - type star in calculations ) . for more massive holes ,
the tidal radius relative to the size of the hole decreases , so stronger apsidal precession reduces the intersection radius in units of the black hole gravitational radius @xmath42 . for higher @xmath2 , @xmath35 is smaller , which again implies stronger precession and closer intersection .
the intersection angle @xmath43 of the outgoing most bound orbit with the incoming stream is given by @xmath44 @xmath43 is plotted as a function of @xmath12 and @xmath2 in figure [ radiusangle](b ) . streams usually collide with @xmath45 if @xmath46 .
the larger the intersection angle , the more effectively the collision reduces the orbital energy of the debris .
some fraction of orbital energy is lost during stream collision due to shocks and instabilities .
the exact fraction depends on the collision angle and velocity / density contrast of the streams @xcite .
here we adopt an inelastic collision model in which the outgoing and incoming streams have similar mass ( cf . * ? ? ? * ) . from momentum conservation ,
the post - collision speed of the stream at the intersection point is @xmath47 where the speed @xmath48 of the streams just before collision is given by @xmath49 the specific energy loss in the collision is @xmath50 .
the debris forms an elliptical disk after collision , and the semi - major axis of this elliptical disk is @xmath51 for deep encounters , @xmath52 and thus @xmath53 we plot this characteristic size of debris disk in figure [ disk ] .
the disk size shrinks rapidly as @xmath12 or @xmath2 increases .
this is because the energy loss is large when the intersection is close and the collision speed is fast .
when @xmath2 is large , the disk radius is comparable to the classical circularization radius @xmath54 , for which the debris materials are fully circularized .
the evolution of the post - collision elliptical disk depends on further stream stream and stream disk interactions .
the circularization timescale has traditionally been assumed to be several dynamical timescales of the most bound orbit : @xmath55 with @xmath56 @xcite .
however , @xcite found that for a white dwarf disrupted by an intermediate - mass black hole the disk is only partially circularized after @xmath57 if the stream collision is mild . though it is unclear how this result extrapolates to a higher mass ratio tde , it is plausible that the circularization timescale will be shorter for higher @xmath2 where the collision is stronger . for our calculation we assume at @xmath58 the debris materials are largely circularized and the disk becomes stable , which corresponds to the peak of the disk luminosity , but our results are not qualitatively changed unless @xmath59 is at least an order of magnitude larger than we assume . from the classical @xmath60-disk model , the inflow timescale of the disk is @xmath61 where @xmath60 is the viscosity parameter of the disk , @xmath62 is the disk aspect ratio at a radius @xmath63 , and @xmath64 is the orbital period at the outer edge of a disk of size @xmath65 .
therefore , in deep plunging tdes , because @xmath66 , @xmath67 , and the inflow timescale is likely to be shorter than the circularization / fallback timescale . in this situation
the onset of accretion is therefore rapid ; it can easily occur before the disk circularizes fully . for a tde around a black hole with @xmath68few@xmath69 ,
the mass fallback rate is greater than the eddington accretion rate for the first weeks to years . when the circularization is efficient due to the strong stream collisions as discussed above , the gas supply rate to the disk is also super - eddington .
modern simulations of super - eddington accretion flow ( e.g. , * ? ? ?
* ) show that such disks are geometrically and optically thick , which together with strong outflows completely obscure the inner region of the disk .
studies of the outflow structure and rate of super - eddington accretion are far from complete , so for simplicity we assume that @xmath70 of the fallback mass flows out in a wind , launched from near the innermost stable circular orbit ( isco ) at the orbital speed there ( for motivation see @xcite ) . at the circularization time
, the wind from the power - law decay part of the fallback ( starting at @xmath71 ) has reached out to @xmath72 . taking the time - averaged @xmath73 ,
the optical depth of the wind beyond @xmath65 is : @xmath74 where @xmath75 is the electron scattering opacity , and @xmath76 at the schwarzschild isco .
using @xmath77 from @xcite , and given @xmath78 , we find that @xmath79 except in the case of the most extreme super - eddington accretion , for which @xmath80 .
thus the wind beyond the disk does not significantly change the observed temperature .
we therefore compute the effective temperature @xmath81 of the system using a photospheric size @xmath82 : @xmath83 here @xmath84 is the luminosity of the disk , @xmath85 is the stefan - boltzmann constant , and we cap the luminosity at eddington ; this is consistent with some simulations @xcite , although other simulations find that the luminosity could be somewhat higher @xcite , which would imply higher temperatures . the accretion efficiency @xmath86 is always taken to be @xmath87 , which is comparable to the specific binding energy at the isco of a moderately spinning black hole .
this calculation of the photosphere size is applicable to super - eddington accretion .
if the fallback rate at @xmath88 is sub - eddington , we use the fallback rate to calculate the peak accretion power , but @xmath89 is insensitive to @xmath90 .
we plot @xmath81 in figure [ temperature ] . @xmath81 increases with increasing @xmath2 . to reach the @xmath91 k seen in x - ray selected tdes
, it is necessary that @xmath92 and @xmath93 .
note that disk radiative transfer effects can alter the spectrum so that the best - fit planck temperature can be up to @xmath94 times larger than the effective temperature ( see , table 1 of @xcite ) .
therefore , very large @xmath2 may not be required to explain tdes with @xmath95 k.
we consider the tidal disruption of a main - sequence star by a non - spinning smbh , and in particular the tidal stream intersection and circularization driven by apsidal precession of the debris orbit .
we show that because apsidal precession is greater for more massive black holes and deeper orbital penetration , such encounters also lead to stream intersection closer to the hole .
this strengthens shocks , which enhances energy losses and decreases the circularization timescale .
the resulting super - eddington accretion disk is small and has a short inflow timescale , so the initial disk temperature is high . for a fixed black hole mass ,
stars disrupted in closer orbits produce higher temperature events .
however , for fixed @xmath2 this disk temperature does not decrease monotonically with increasing black hole mass ; it appears that for x - ray detected tdes it is necessary that @xmath92 and @xmath96 .
this is consistent with the mass distribution of the small number of x - ray observed tde flares reported in @xcite .
such high @xmath2 encounters are most likely to be produced when the orbital phase space of the stars is not depleted by disruptions . in this case , the probability of an event with @xmath97 scales as @xmath98 . if , in contrast , the inner galactic zone is depleted , diffusion processes tend to produce low-@xmath2 encounters ( e.g. , @xcite ) .
therefore we expect that @xmath99 of the fully disruptive events reach a temperature high enough to produce substantial x - rays .
reprocessing from , e.g. , dust could reduce this fraction .
nonetheless , this ratio is consistent with the ratio of the observed x - ray tde rate @xcite and optical / uv tde rate @xcite , though both estimates bear uncertainties due to the small sample sizes and modeling assumptions .
we note that not many tdes have been observed at their peaks with x - ray telescopes , so a more complete search on x - ray tdes will greatly improve our understanding .
there are other models proposed to explain x - ray tdes .
for example , inferred temperature differences among tdes may be partially due to viewing angle dependent obscuration by the disk @xcite or wind @xcite , leading in the extreme case to a jetted tde when viewed down the jet @xcite .
if the disk is geometrically thin , unobscured , and the accretion rate can stay relatively constant with radius , then the disk temperature can be dominated by the temperature near disk inner regions @xcite .
non - thermal processes such as electron scattering and comptonization can drive part of the tde disk to have a high temperature @xcite .
thermal or non - thermal x - ray emission could also arise from either the disk corona or jet , where a similar ambiguity exists for the origin of sgra * radio and x - ray emission . around a spinning black hole ,
the nodal precession of the debris stream can delay substantially the time of first stream intersection @xcite , and this likely also delays the onset of accretion . as debris streams may only partially collide
, circularization can take longer and the peak luminosity will be lower than it would be around a non - spinning hole
. however , if the initial stellar orbit is aligned with the black hole spin , the stream intersection will be very similar to that discussed in this paper .
the extra precession produced by black hole spin on a retrograde orbit can draw the intersection closer to the hole .
this produces an even smaller disk and thus promotes faster onset of accretion and increases disk temperature .
ld and jcm acknowledge nasa / nsf / tcan ( nnx14ab46 g ) , nsf / xsede / tacc ( tgphy120005 ) , and nasa / pleiades ( smd-14 - 5451 ) .
mcm was supported in part by nsf ( ast-1333514 ) .
we are grateful for conversations with t. alexander , r. blandford , j. cannizzo , b. cenko , r. cheng , s. gezari , j. guillochon , s. komossa , j. krolik , m. macleod , e. ramirez - ruiz , h. shiokawa , n. stone and s. van velzen .
we also thank the anonymous referee for helpful comments .
( equation ( 1 ) ) .
panels ( a ) , ( b ) , and ( c ) represent the scenarios for a smbh and a solar type star .
panel ( d ) represents an imbh - white dwarf disruption scenario as in @xcite .
stream - stream intersection occurs where the black solid line and blue dashed line intersect . in a deep plunge ( panel ( b ) ) or when the black hole is more massive ( panel ( c ) ) the stream intersects closer to the hole.,width=576 ] : 1 - black solid , 2 - blue dotted , 4 - green dashed , 8 - orange dot - dashed , and 16 - magenta long - dashed .
panel ( a ) shows that the interaction radius decreases with larger @xmath2 .
panel ( b ) shows that the collision angle in most cases is between 40 and 160@xmath100 .
the energy loss is more efficient when the collision angle is large.,width=576 ] | one of the puzzles associated with tidal disruption event candidates ( tdes ) is that there is a dichotomy between the color temperatures of @xmath0 k for tdes discovered with optical and uv telescopes , and the color temperatures of @xmath1 k for tdes discovered with x - ray satellites . here
we propose that high - temperature tdes are produced when the tidal debris of a disrupted star self - intersects relatively close to the supermassive black hole , in contrast to the more distant self - intersection that leads to lower color temperatures . in particular , we note from simple ballistic considerations that greater apsidal precession in an orbit is the key to closer self - intersection .
thus larger values of @xmath2 , the ratio of the tidal radius to the pericenter distance of the initial orbit , are more likely to lead to higher temperatures of more compact disks which are super - eddington and geometrically and optically thick . for a given star and @xmath2
, apsidal precession also increases for larger black hole masses , but larger black hole masses imply a lower temperature at the eddington luminosity .
thus the expected dependence of the temperature on the mass of the black hole is non - monotonic .
we find that in order to produce a soft x - ray temperature tde , a deep plunging stellar orbit with @xmath3 is needed and a black hole mass of @xmath4 is favored .
although observations of tdes are comparatively scarce and are likely dominated by selection effects , it is encouraging that both expectations are consistent with current data . |
in a cosmological model with closed , nontrivial space topology , a distant astronomical source is predicted to produce several images , corresponding to the diffferent paths its radiation can take to reach the observer on earth .
see lachize - rey & luminet @xcite for a review .
theoretically these multiconnected spatial sections are manifolds @xmath0 of the form @xmath1 where @xmath2 is the universal covering space of @xmath3 and @xmath4 is a discrete group of isometries acting freely on @xmath5 .
this action tesselates @xmath2 into cells which are copies of a fundamental polyhedron ( fp ) .
the fp has an even number of faces , which are pairwise congruent ; @xmath0 may be represented by the fp , with members of these pairs identified ( more colloquially , they are `` glued together '' ; more technically , see massey @xcite ) . in practice one
looks at the fp as the physical space where the observer and the sources exist , and at its copies as the apparent space of repeated cosmic images . hence the tendency to consider those images within the fp as ` real , ' those in other cells as ` ghosts . ' but it also appears reasonable to call real the nearest image of an object , the other ones being the ghosts .
it is the purpose of this paper to show that these two criteria for classifying multiple images are not always consistent with each other .
one of the possible universe models with einstein - de sitter metric @xmath6 has as spatial section the orientable manifold e2 in fig .
17 of @xcite .
a klein bottle surface @xmath7 is imbedded in e2 , so we may use this nonorientable surface as our simplified model of cosmic space .
it turns out that the fp ( here fundamental _ polygon _ ) is not unique ; it depends on an arbitrary _ basepoint _ in its definition : the fp of a manifold @xmath8 with basepoint @xmath9 is the set of points @xmath10 _
_ distance__@xmath11 _ _
distance__@xmath12 let @xmath0 @xmath13 be the klein bottle obtained from the square @xmath14 of side length @xmath15 in fig .
1 , with coordinates @xmath16 and basepoint @xmath17 this fundamental polygon will be called fp0 .
group @xmath4 is generated by the motions @xmath18 and @xmath2 is the euclidean plane .
the identified pairs of sides are @xmath19 @xmath20 .
now we choose another basepoint , @xmath21 and find the corresponding fp as indicated in fig .
1 . the six images of @xmath22 in the figure ( @xmath23 are sufficient to determine the irregular hexagon @xmath24 as fp1
. the points on line @xmath25 are equidistant from @xmath22 and @xmath26 , and thus belong to the border ; and similarly for the other sides .
the coordinates of the vertices are @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 and @xmath32 , and the identifications are @xmath33 , @xmath34 , @xmath35
in fig . 2 , let the observer s position be @xmath22 , and two images of a cosmic source be located at @xmath36 and @xmath37 the nearest image is @xmath38 , and if the observer is using fp1 , then by both criteria at the end of sec .
i @xmath38 would be the real image and @xmath39 a ghost .
but if she or he is using fp0 , which is much easier to handle mathematically despite the fact that the observer is _ not _ at its center , then by first criterion the real image is @xmath39 and @xmath38 is a ghost , with the opposite holding by the second rule . therefore , in order that the terminology ` real ' vs. ` ghost ' ( or ` source ' vs. ` ghost ' ) could be used consistently , one would have to always work with an fp where the basepoint is at the observer s position .
a further advantage of thus having the nearest images labeled as sources would be in the comparison with astrophysical data , which are usually richer for nearby objects .
however , the calculations are much simpler in the more symmetrical fp like fp0 above , or in those fp s which mathematicians say have `` maximum injectivity radius '' - see weeks @xcite , for example .
this is what was done by the author in @xcite , where the fp is a regular icosahedron with basepoint at its center , far from the observer s position . | universes with multiconnected spatial sections predict multiple images of cosmic sources .
a confusing terminology exists in the naming of these images as real ones vs. ghosts . here
an attempt is made to clarify the situation . |
consider a sequence @xmath5 of i.i.d . random variables with support on @xmath3 $ ] and having distribution function @xmath6 .
for any fixed @xmath7 , the distributions of @xmath8 and @xmath9 have been well studied ; in fact it is shown in elementary texts that @xmath10 and @xmath11 .
but what if we have a situation where the number @xmath0 of @xmath12 s is random , and we are instead considering the extrema @xmath13 and @xmath14 of a random number of i.i.d .
random variables ?
now the _ sum _
@xmath15 of a random number of i.i.d .
variables , defined as @xmath16 satisfies , according to wald s lemma @xcite , the equation @xmath17 provided that @xmath0 is independent of the sequence @xmath18 and assuming that the means of @xmath2 and @xmath0 exist . the purpose of this paper is to show that the distributions in ( 1 ) and ( 2 ) can be studied in many canonical cases , even if @xmath0 and @xmath19 are correlated . the main deviation from the papers @xcite , @xcite and @xcite , where similar questions are studied , is that the variable @xmath2 is concentrated on the interval @xmath3 $ ] unlike the above references , where @xmath2 has lifetime - like distributions on @xmath20 . even then , we find that many new and interesting distributions arise , none of them to be found , e.g. , in @xcite or @xcite . in another deviation from the theory of extremes of random sequences ( see , e.g. , @xcite ) , we find that the tail behavior of the extreme distributions is not relevant due to the fact that the distributions have compact support .
we next cite three examples where our methods might be useful .
first , we might be interested in the strongest earthquake in a given region in a given year .
the number of earthquakes in a year , @xmath0 , is usually modeled using a poisson distribution , and , ignoring aftershocks and similarly correlated events , the intensities of the earthquakes can be considered to be i.i.d .
random variables in @xmath21 $ ] whose distribution can be modeled using , e.g. , the data set maintained by caltech at @xcite .
second , many small world " phenomena have recently been modeled by power law distributions , also sometimes termed discrete pareto or zipf distributions .
see , for example , the body of work by chung and her co - authors @xcite , @xcite , and the references therein , where vertex degrees @xmath22 in internet - like graphs " @xmath23 ( e.g. , the vertices of @xmath23 are individual webpages , and there is an edge between @xmath24 and @xmath25 if one of the webpages has a link to the other ) are shown to be modeled by @xmath26^{-1}}{n^k}\ ] ] for some constant @xmath27 , where @xmath28 is the riemann zeta function @xmath29 thus if the vertices @xmath30 in a large internet graph have some bounded i.i.d .
property @xmath12 , then the maximum and minimum values of @xmath12 for the neighbors of a randomly chosen vertex can be modeled using the methods of this paper .
third , we note that @xmath0 and the @xmath12 may be correlated , as in the csug example ( studied systematically in section 3 ) where @xmath31 $ ] and @xmath32 follows the geometric distribution @xmath33 .
this is an example of a situation where we might be modeling the maximum load that a device might have carried before it breaks down due to an excessive weight or current .
it is also feasible in this case that the parameter @xmath4 might be unknown .
here is our general set - up : suppose @xmath34 are i.i.d .
random variables following a continuous distribution on @xmath3 $ ] with probability density and distribution functions given by @xmath35 and @xmath36 respectively .
@xmath0 is a random variable following a discrete distribution on @xmath37 with probability mass function given by @xmath38 .
let @xmath39 and @xmath40 be given by ( 1 ) and ( 2 ) respectively .
then the p.d.f.s @xmath41 of @xmath39 and @xmath40 are derived as follows : since@xmath42^n,\ ] ] we see that@xmath43^{n-1}f(y),\ ] ] and consequently , the marginal p.d.f . of @xmath39 is @xmath44^{n-1}p(n ) .
\end{aligned}\ ] ] in a similar fashion , the p.d.f . of @xmath40 can be shown to be @xmath45^{n-1}p(n);\ ] ] what is remarkable is that the sums in ( 3 ) and ( 4 ) will be shown to assume simple tractable forms in a variety of cases .
our paper is organized as follows . in section 2 , we study the case of @xmath46 $ ] and @xmath47 .
we call this the standard uniform geometric model .
the csug ( correlated standard uniform model ) is studied in section 3 .
section 4 is devoted to a summary of a variety of other models .
since @xmath48 , and @xmath49 for some @xmath50 , we have from ( 3 ) that the p.d.f . of @xmath39 in the sug model is given by @xmath51 ^ 2}.\end{aligned}\ ] ] , width=449,height=316 ] similarly , ( 4 ) gives that @xmath52 ^ 2}\nonumber\\ & = & \frac{\theta}{[\theta+(1-\theta)z]^2}.\end{aligned}\ ] ] , width=449,height=316 ] if the random variable @xmath39 has the sug maximum distribution " ( 5 ) and @xmath53 , then @xmath54 @xmath55 ^ 2}dy\\ & = & \int_1^\theta \lr\frac{1-u}{1-\theta}\rr^k \times \frac{\theta}{u^2}\times \lr-\frac{1}{1-\theta}\rr du\\ & = & \frac{\theta}{(1-\theta)^{k+1}}\int_\theta^1 \frac{(1-u)^k}{u^2}du\\ & = & \frac{\theta}{(1-\theta)^{k+1}}\int_\theta^1 \frac{\sum_{j=0}^k { k \choose j}(-u)^j}{u^2}du\\ & = & \frac{\theta}{(1-\theta)^{k+1}}\sum_{j=0}^{k}{k \choose j}\int_\theta^1(-u)^{j-2}du , \end{aligned}\ ] ] as claimed .
the random variable @xmath39 has mean and variance given , respectively , by @xmath56 using proposition 2.1 , we can directly compute the mean and variance by setting @xmath57 , and using the fact that @xmath58 ^ 2 $ ] for any random variable @xmath59 .
( this proof could equally well have been based on calculating the moments of @xmath60 and then recovering the values of @xmath61 and @xmath62 .
the same is true of other proofs in the paper . ) if the random variable @xmath40 has the sug minimum distribution " and @xmath53 , then @xmath63 @xmath64 ^ 2}dz\\ & = & \int_\theta^1 \lr\frac{u-\theta}{1-\theta}\rr^k \times \frac{\theta}{u^2}\times \frac{1}{1-\theta}du\\ & = & \frac{\theta}{(1-\theta)^{k+1}}\int_\theta^1 \frac{(u-\theta)^k}{u^2}du\\ & = & \frac{\theta}{(1-\theta)^{k+1}}\int_\theta^1 \frac{\sum_{j=0}^k { k \choose j}u^{k - j}(-\theta)^j}{u^2}du\\ & = & \frac{\theta}{(1-\theta)^{k+1}}\sum_{j=0}^{k}{k \choose j}(-\theta)^j\int_\theta^1 u^{k - j-2}du , \end{aligned}\ ] ] as asserted .
the random variable @xmath40 has mean and variance given , respectively , by @xmath65 using proposition 2.3 , it is easily to compute the mean and variance by setting @xmath66 .
the m.g.f.s of @xmath67 are easy to calculate too .
notice that the logarithmic terms above arise due to the contributions of the @xmath68 and @xmath69 terms , and it is precisely these logarithmic terms that make , e.g. , method of moments estimates for @xmath4 to be intractable in a closed ( i.e. , non - numerical ) form .
similar difficulties arise when analyzing the likelihood function and likelihood ratios .
the correlated standard uniform geometric ( csug ) model is related to the sug model , as the name suggests , but @xmath2 and @xmath0 are correlated as indicated in section 1 .
the csug problems arise in two cases .
one case is that we conduct standard uniform trials until a variable @xmath12 exceeds @xmath70 , where @xmath4 is the parameter of the correlated geometric variable , and the maximum of @xmath71 is what we seek .
the maximum is between 0 and @xmath70 .
the other case is where standard uniform trials are conducted until @xmath12 is less than @xmath4 , and we are looking for the minimum of @xmath71 . the minimum is between @xmath4 and 1 .
specifically , let @xmath72 be a sequence of standard uniform variables and define @xmath73 or @xmath74 in either case @xmath0 has probability mass function given by @xmath75 note that this is simply a geometric random variable conditional on the success having occurred at trial 2 or later .
clearly @xmath0 is dependent on the @xmath2 sequence . under the csug model , the p.d.f . of @xmath39 ,
defined by ( 1 ) , is given by @xmath76 , width=449,height=316 ] the conditional c.d.f . of @xmath39 given that @xmath77 is given by @xmath78 taking the derivative , we see that the conditional density function is given by @xmath79 consequently , the p.d.f . of @xmath39 in the csug model is given by @xmath80 this completes the proof . the p.d.f . of @xmath40 under the csug model
is given by @xmath81 , width=449,height=316 ] the conditional cumulative distribution function of @xmath40 given that @xmath77 is given by @xmath82 thus , the conditional density function is given by @xmath83 which yields the p.d.f .
of @xmath40 under the csug model as @xmath84 which finishes the proof . if the random variable @xmath39 has the csug maximum distribution " and @xmath53 , then @xmath85 @xmath86 as claimed .
the random variable @xmath39 has mean and variance given , respectively , by @xmath87and @xmath88 using proposition 3.3 , we can directly compute the mean and variance by setting @xmath89 .
notice that the variance of @xmath39 is smaller than that of @xmath39 under the sug model , with an identical numerator term . also , the expected value is smaller under the csug model than in the sug case . if the random variable @xmath40 has the csug minimum distribution " and @xmath53 , then @xmath90 routine , as before .
the random variable @xmath40 has mean and variance given , respectively , by @xmath91and@xmath92 a special case of proposition 3.5 ; note that as in the sug model , @xmath93 .
the intermingling of polynomial and logarithmic terms makes method of moments estimation difficult in closed form , as in the sug case .
however , if @xmath4 is unknown , the maximum likelihood estimate of @xmath4 can be found in a satisfying form , both in the cgug maximum and csug minimum cases .
suppose that @xmath94 form a random sample from the csug maximum distribution with unknown @xmath4 .
since the pdf of each observation has the following form : @xmath95 the likelihood function is given by @xmath96 the mle of @xmath4 is a value of @xmath4 , where @xmath97 , which maximizes @xmath98 .
let @xmath99 .
since @xmath100 , it follows that @xmath101 is a increasing function , which means the mle is the largest possible value of @xmath4 such that @xmath102 .
thus , this value should be @xmath103 , i.e. , @xmath104 .
suppose next that @xmath105 form a random sample from the csug minimum distribution .
since the pdf of each observation has the following form : @xmath106 it follows that the likelihood function is given by @xmath107 as above , it now follows that @xmath108 .
the general scheme given by ( 3 ) and ( 4 ) is quite powerful . as another example ,
suppose ( using the example from section 1 ) that @xmath109 and @xmath46 $ ] .
then it is easy to show that @xmath110 and that @xmath111 .
( the expected value of @xmath39 can also be calculated by using the identity @xmath112 . in this section ,
we collect some more results of this type , without proof : uniform - poisson model here we let @xmath46 $ ] and @xmath113 , so that @xmath0 follows a left - truncated poisson distribution . under the uniform - poisson model ,
@xmath114 @xmath115 @xmath116 @xmath117 in some sense , the primary motivation of this paper was to produce extreme value distributions that did not fall into the beta family ( such as @xmath118 for the maximum of @xmath7 i.i.d . @xmath119 $ ] variables ) .
a wide variety of non - beta - based distributions may be found in @xcite . can we add extreme value distributions to that collection ? in
what follows , we use both the beta families @xmath120 and @xmath121 , the arcsine distribution , and a beyond beta " distribution , the topp - leone distribution , as input variables " to make further progress in this direction .
geometric - beta(2,2 ) model . here
@xmath122 and @xmath47 . in this case
we get @xmath123 ^ 2}\ ] ] and @xmath124 ^ 2}.\ ] ] poisson - beta(2,2 ) model . here
@xmath122 and @xmath125 , the poisson(@xmath4 ) distribution left - truncated at 0 . in this case we get @xmath126 and @xmath127 geometric - arcsine model . here
@xmath128 and @xmath47 . in this case
we get @xmath129^{-1/2}}{[1-(1-\theta)\frac{2}{\pi}\arcsin\sqrt{y}]^2}\ ] ] and @xmath130^{-1/2}}{[1-(1-\theta)(1-\frac{2}{\pi})\arcsin\sqrt{z}]^2}.\ ] ] poisson - arcsine model . here
@xmath128 and @xmath125 .
here we have @xmath131^{-1/2}e^{-\theta(1-\frac{2}{\pi}\arcsin\sqrt{y})}}{1-e^{-\theta}}\ ] ] and @xmath132^{-1/2}e^{-\frac{2\theta \arcsin\sqrt{z}}{\pi}}}{1-e^{-\theta}}.\ ] ] geometric - topp - leone model . here
@xmath133 and @xmath47 : @xmath134 ^ 2}\ ] ] and @xmath135\}^2}.\ ] ] poisson - topp - leone model . @xmath133 and @xmath125 : @xmath136}}{1-e^{-\theta}}\ ] ] and @xmath137}}{1-e^{-\theta}}.\ ] ]
the research of ag was supported by nsf grants 1004624 and 1040928 .
99 b. aiello , f. chung , and l. lu . a random graph model for power law graphs . "
_ experimental mathematics _ * 10 * , 5366 , 2001 . f. chung , l. lu , and v. vu .
eigenvalues of random power law graphs . "
_ annals of combinatorics _ * 7 * , 2133 , 2003 .
r. durrett , _ probability : theory and examples . _ wadsworth and brooks / cole , pacific grove , 1991 .
earthquake data set , available at s. kotz and j. van dorp . _ beyond beta : other continuous families of distributions with bounded support and applications .
_ world scientific publishing co. , singapore , 2004 .
m. leadbetter , g. lindgren , and h. rootzn .
_ extremes and related properties of random sequences and processes . _ springer verlag , new york , 1983 .
f. louzada , e. bereta , and m. franco .
on the distribution of the minimum or maximum of a random number of i.i.d .
lifetime random variables . " _ applied mathematics _ * 3 * , 350353 , 2012 . f. louzada , m. roman and v. cancho .
the complementary exponential geometric distribution : model , properties , and a comparison with its counterpart . "
_ computational statistics and data analysis _ * 55 * , 25162524 , 2011 .
a. morais and w. barreto - souza .
a compound class of weibull and power series distributions . "
_ computational statistics and data analysis _ * 55 * , 14101425 , 2011 .
mit course lecture notes ; civil and environmental engineering department , probability and statistics in engineering , spring 2005 . | we study a new family of random variables , that each arise as the distribution of the maximum or minimum of a random number @xmath0 of i.i.d .
random variables @xmath1 , each distributed as a variable @xmath2 with support on @xmath3 $ ] .
the general scheme is first outlined , and several special cases are studied in detail . wherever appropriate , we find estimates of the parameter @xmath4 in the one - parameter family in question .
v [ section ] [ section ] [ thm]claim [ thm]definition [ thm]lemma [ thm]corollary [ thm]remark [ thm]proposition [ thm]example [ thm]equation [ section ] [ thm]question [ 1]#1 keywords : maximum and minimum ; random number of i.i.d .
variables ; statistical inference . |
as is well known , s - wave superconductivity is possible even in highly disordered systems .
examples are superconducting intermetalic alloys @xmath4such as @xmath5 , @xmath6@xcite@xmath7 , heavily doped cubic perovskites @xmath4such as @xmath8 with @xmath9@xmath7@xcite and chevrel phases @xmath4 such as @xmath10@xmath7.@xcite what lies beyond these initially surprising facts is anderson s theorem@xcite according to which pairing of time - reversed states leads to a finite gap @xmath11 in the density of states .
namely , the one particle states involved in the pairing need not be eigenstates of any translation operators and hence both @xmath12 and @xmath13 are only weakly influenced by the disorder .
the key assumptions required for anderson s theorem are @xmath14 non- magnetic scattering only @xmath4time reversal symmetry@xmath7 @xmath15 the self - consistent order parameter @xmath12 does not fluctuate from configuration to configuration .
previously gyrffy _
et al._@xcite examined the conditions for @xmath15 to hold , and found that spatial fluctuations in @xmath12 could be neglected provided that coherence length @xmath16 is much greater than the lattice spacing , @xmath17 .
anderson s theorem also follows from the classic theory of abrikosov and gorkov.@xcite their argument is based on perturbation theory and the proof requires that the real part of the self energy @xmath18 varies slowly near the fermi energy @xmath19.@xcite it turns out that this is equivalent to the condition @xmath20 in eq . [ eq : coherent length ] .
the purpose of this paper is to examine disordered s - wave superconductors for which the spatial fluctuations of @xmath12 can not be neglected .
the results of an such inquiry will be important for narrow band or short coherence length superconductors where @xmath21 or @xmath22 , such as some of those mentioned above . in the work of gyrffy _
et al._@xcite the order parameter fluctuations due to impurities were studied using a perturbative technique .
zhitomirsky and walker@xcite also calculated corrections to @xmath13 beyond anderson s theorem due to order parameter fluctuations evaluated perturbatively . in the present paper
the goal is to treat the spatially varying @xmath12 and charge density fully self - consistently within the coherent potential approximation ( cpa ) , and hence allow for the case of arbitrarily strong @xmath12 or charge fluctuations . in this regime
a number of interesting new issues arise .
for example ghosal _
et al._@xcite showed that strong disorder scattering leads to strong spatial variations in @xmath12 , with the formation of superconducting islands where @xmath12 is large and other regions where @xmath12 is small .
moreover , they found that the spectral gap persists even when @xmath12 is very small in large regions of the sample .
by contrast it was argued by opperman@xcite and ziegler@xcite that order parameter fluctuations lead to a finite density of states within the gap .
similarly , in the calculations of the non - self - consistent density of states by annett and goldenfeld@xcite an impurity _ band tail _ develops due to spatial fluctuations in @xmath12 , and eventually this leads to gap - less s - wave superconductivity .
obviously the fluctuations in @xmath12 can arise either due to randomness in the single particle site energy , @xmath23 , at each atomic site , as in an alloy , or due to randomness in the attractive interaction potential , @xmath24 , as considered by litak and gyrffy.@xcite here we consider only the former case of alloy type disorder . in this paper
we calculate the spatially random , self - consistent order parameter and charge due to ab binary alloy type disorder .
the on site energy is @xmath25 on a fraction c of lattice sites and @xmath26 on the fraction @xmath27 .
the green functions are calculated using the coherent potential approximation@xmath28@xcite .
this has been shown to be exact in the limit of infinite dimensions@xcite and to reproduce the results of the self - consistent born approximation for weak scattering @xmath29 where w is the bandwith @xmath7 and the self - consistent t - matrix approximation in the limit of low impurity concentration ( @xmath30 ) .
we calculate the self - consistent a or b site order parameter @xmath31 and charge density @xmath32 .
interestingly , we find that particle - hole symmetry leads to @xmath33 , and we show that anderson s theorem applies exactly in this case , even though @xmath22 or @xmath34 .
a similar result was obtained previously for weak ( @xmath35 ) @xcite and strong @xmath36@xcite interactions , but only for the weak disorder limit . by contrast
the result here is exact for both weak and strong scattering .
evidently , this latter result implies that @xmath37 is not a necessary condition for anderson s theorem .
furthermore in the extreme disorder limit @xmath38 ) we have a disorder induced band - splitting in the normal state .
remarkably , if the chemical potential lies in the band gap there is a superconductor to insulator quantum phase transition at @xmath2 .
if the chemical potential lies inside one of the split bands , the normal state is metalic but there are impurity states inside the superconducting gap .
below , in section ii , we describe the model , and our cpa formalism and report our numerical results for the cases of weak @xmath39 and intermediate @xmath40 scattering . in section iii
we show that the gap fluctuation vanishes in the case of particle - hole symmetry , whilst in section iv we argue that within cpa anderson s theorem is exact in this case .
section v contain our results for the strong scattering @xmath41 case where we find a superconductor - insulator transition .
we use an attractive u single band hubbard model defined by the hamiltonian : @xmath42 where @xmath43 , @xmath44 are , respectively , the creation and annihilation operators of electrons with spin @xmath45 on the lattice site @xmath46 , @xmath47 is the local occupation number operator , @xmath48 is the chemical potential , and @xmath49 is the hopping integral from site @xmath46 to site @xmath50 .
@xmath24 is the attractive pairing interaction at site i. in all of the numerical calculations shown below the interaction potential is @xmath51 . for other interaction strengths we found similar results .
@xmath23 is the site diagonal random disorder potential which takes on values @xmath25 with probablity c and @xmath26 with probablity @xmath27 . after applying the hartree - fock - gorkov ( hfg ) approximation to eq .
[ eq : s - wavehamiltonian ] our task becomes a study of the gorkov equation : @xmath52 where @xmath53 is the fourier transform , with respect to the complex - time variable @xmath54 , of the greens function @xmath55.the self - consistency conditions for the local order parameter and charge density are @xmath56 and @xmath57 where @xmath58 is a positive infinitesimal .
the task at hand is to solve the above equations for each configuration @xmath59 of the site energy and pairing interaction , to obtain the local order parameter and charge @xmath60 and hence to calculate the configurationally averaged greens function @xmath61 .
whilst most of the salient features of disordered superconductors are well described by the abrikosov and gorkov(ag)@xcite theory , recently the cpa has also been brought to bear on the problem.@xcite within the cpa one replaces the on site random potential with a site independent self - energy matrix @xmath62 which is determined by the condition that an a or b impurity , corresponding to @xmath25 and @xmath26 respectively does not scatter on the average .
thus the average green function is given by @xmath63 in the previous application of the cpa to the above model@xcite the condition which determined @xmath64 was implemented under the assumption that the pairing potential @xmath65 does not fluctuate with the site energies .
therefore the self - consistency conditions ( eqs .
[ eq : local order parameter ] , [ eq : s wave charge density ] ) were satisfied only on average . however , this presumption is not a necessary part of the cpa and , as will be seen presently , is unduely restrictive . here , we consider the more general case where on an a or b site , with site energy @xmath25 or @xmath26 , the pairing potential is allowed to be @xmath66 or @xmath67 respectively and the two local gaps , @xmath66 and @xmath67 , are determined by the condition that the corresponding local gap equations ( eq .
[ eq : local order parameter ] ) are separately satisfied .
thus if the probability that a site is occupied by an a atom is c and that for a b atom is @xmath27 the generalized cpa condition is @xmath68 where the single site t - matrices , @xmath69 and @xmath70 , are given by @xmath71 in terms of the local , single site , scattering potential matrix @xmath72 for @xmath73 .
note that in the above expression not only @xmath23 and @xmath65 are allowed to fluctuate but the local @xmath74 also take on different values on a and b sites .
evidently , such variations arise from charge fluctuations . in the present hfg approximation
this is described by @xmath75 clearly , once the above cpa problem has been solved for a set of @xmath76 and @xmath77 self - consistency requires that they are recalculated using the relations @xmath78 @xmath79 where @xmath80 is the greens function matrix averaged over all configurations with an a or b atom , respectively , on the site i. in our calculation we assume , the non - magnetic case , that @xmath81 , and hence @xmath82 moreover , the patially averaged green functions @xmath80 are approximated by the green function for a or b impurity in the cpa effective lattice described by the self - energy @xmath64 .
they are given by @xmath83 in short , equations @xmath84 fully specify a self - consistent procedure , which when carried to convergence , constitutes the complete cpa for disordered superconductors , in the hfg approximation , for the model hamiltonian in eq .
[ eq : s - wavehamiltonian ] .
note that since the cpa is the mean field theory of disorder and the hfg approximation is that for superconductivity , the above theory should be regarded as the mean field theory which treats disorder and electron interaction simultaneously and on an equal footing . by treating the self - consistency only on average , earlier
works@xcite did not include the effects of fluctuations in @xmath12 and @xmath85 .
previously the full cpa was only implemented to determine the influence of @xmath65 fluctuation on @xmath13.@xcite in the remainder of this paper we investigate the full consequences of treating interaction and disorder together on the basis of eqs .
@xmath84 .
our results for the gap and charge fluctuations as functions of the average band filling @xmath86 are shown in figs.@xmath87 and @xmath88 . to simplify matters we used a @xmath89 square lattice with lattice constant @xmath90 and band energy @xmath91
where t is the nearest neighbour hopping amplitude .
in these calculations we have taken the energy difference @xmath92 to be a significant fraction of the bandwidth @xmath93 @xmath94 and hence we are in the fairly strong scattering regime @xmath4 for @xmath95 .
as fig.@xmath87 shows @xmath96 except at the point @xmath97 , namely a half filled band . clearly from fig.@xmath88 @xmath98 at any filling except @xmath99 or @xmath88
thus , unlike in previous calculations@xcite the @xmath12 and @xmath85 fluctuations are central features of our results .
for emphasis we show in fig.@xmath100 the standard deviations of the order parameter @xmath101 and charge density @xmath102 as predicted by our cpa calculations .
remarkably , at half filling the fluctuations in the pairing potential go to zero , whilst the charge density fluctuations are at their strongest .
to investigate the origin of this interesting phenomena we studied the case where @xmath103 but the band is still half filled at @xmath97 .
for this case @xmath0 and @xmath104 are shown in fig.@xmath105 as a function of temperature .
evidently for all @xmath106 , @xmath107 and hence @xmath108 . in what follows we unravel the root cause of this behavior .
recall that in figs.@xmath109 the order parameter fluctuations vanish , @xmath33 , for the case @xmath97 and @xmath110 , namely equal concentrations of a and b atoms .
this special case is one where the fermi energy is at the center of the band and the density of states is symmetric , and hence particle - hole symmetry occurs . for this case and only this case @xmath111 , @xmath112 and @xmath113 .
furthermore , since the self - energy always obey the same symmetries as the green functions @xmath114 for particle - hole symmetry there is also the property that @xmath115 namely , @xmath116 and consequently @xmath117 noting that the cpa respects these symmetries we can rewrite the cpa condition in eq .
[ eq : cpa condition ] , as @xmath118 combining this with particle - hole symmetry as described by eqs .
[ eq : sigma symmetry ] , [ eq : sig - p - h ] , [ eq : relsigma ] , [ eq : realgreen ] implies that @xmath119 and hence , since the second bracket is non - zero , that @xmath120 therefore particle - hole symmetry implies the absence of the order parameter fluctuations .
in non - magnetic disordered local s - wave superconductors the traditional argument leading to anderson s theorem assumes that the fluctuations of the order parameter are negligible , @xmath121 .
anderson s theorem shows that in this case there are no bound states between quasi - particles and impurity sites and therefore the quasi - particle energy gap is absolute .
namely , there are no impurity states inside of the gap.@xcite furthermore @xmath13 is found simply by replacing , the clean system density of states @xmath122 , with its disordered system average @xmath123 in the gap equation . in an alternative route to the same result abrikosov and gorkov@xcite use perturbation theory which implies that the self - energy can be approximated by @xmath124 where @xmath54 is a wave vector and frequency independent quasi - particle life time .
this is justified in the case of the non - self - consistent born approximation by the assumption that the relevant energy scale is @xmath125 @xmath126 is debeye frequncy@xmath7 and hence only states near the fermi surface are relevant .
more implicitly it is assumed that near the fermi surface the density of states is a constant and hence there is effectively particle - hole symmetry .
however in general we can not assume that @xmath127 is of this form .
in the case of narrow band superconductors , there is no debeye cut off and so one can not assume that only states near the fermi level are significant .
in particular @xmath128 will not be a constant , and @xmath129 need not be zero . nevertheless ,
if for some reason particle - hole symmetry is obeyed , anderson s theorem will obtain in full cpa , self - consistent born , and t - matrix approximations . in figs.@xmath130 and @xmath131
we illustrate using our explicit calculation the energy dependence of @xmath132 , @xmath133 , @xmath134 and @xmath135 for the cases of particle - hole symetric or nonsymmetric cases .
note that function @xmath136 is defined by @xmath137 and @xmath136 is the analytical continuation of @xmath129 to the real axis . figs.@xmath130 and @xmath131 show that for particle - hole symmetry , @xmath97 and @xmath110 , @xmath136 is equal to zero but in other cases @xmath138 .
we shall now analyze the consequences of the particle - hole symmetry @xmath139 .
the green function for cpa , self - consistent born or t - matrix approximations can be written in the form @xmath140 where the renormalized parameters @xmath141 , @xmath142 and @xmath143 and @xmath144 are given by @xmath145 note that these renormalized parameters are the same as in the original paper of abrikosov and gorkov@xcite , where @xmath64 was computed in the non - self - consistent born approximation .
they found that @xmath146 near the fermi surface is independents of @xmath147 , and therefore eventually it can be absorbed into the chemical potential .
the same procedure was followed martin et al
.. @xcite however we proceed without these simplifications and retain the full energy dependence of @xmath132 and @xmath133 .
now , particle - hole symmetry implies that @xmath148 , and therefore @xmath149 . using the above relations the gap equation becomes @xmath150 where @xmath151 and , @xmath152 is the renormalized band energy .
surprisingly in the case of particle - hole symmetry the quantity @xmath153 in eq .
[ eq : anderson ] becomes equal to the disorder average normal state density of states , namely , @xmath154 and hence , without further assumptions , anderson s theorem obtains . to be
quick and explicit we note that this last step follows from the property of a delta function that @xmath155 , and hence @xmath156 in short , we again have anderson s theorem that @xmath157 is given by the usual gap equation , but with the disorder average normal density of states .
however , unlike other proofs , ours does not neglect the energy dependence of @xmath132 or @xmath133 .
one of the main virtues of the cpa in the normal state is the fact that it describes band splitting correctly.@xcite namely , for @xmath158 less than the bandwidth , @xmath159 in our case , it predicts an effective band somewhere in between the bands of pure a or pure b metal , whilst for @xmath158 bigger than the half bandwidth cpa predicts two , smeared , but nevertheless well defined , bands seperated by a gap .
the two bands are at energies where there would have been an a or b band in one of the pure systems . in this later case the wave function corresponding to the a band
is large mainly on the a sites and that in the b band is significant only on the b sites .
in what follows we shall investigate superconductivity in this split band regime of cpa . to investigate the consequences for superconductivity in the above split band regime we have solved eq.6 - 15 using a strong scattering potential @xmath160 . from these solutions two interesting points emerged .
firstly we found that for the particle - hole symmetric case , reported in fig.@xmath161 , @xmath33 and hence even in this strong scattering state there are no fluctuations in @xmath12 .
nevertheless , scattering has a large effect on the superconductivity through the configurationally average density of states @xmath123 in the eq .
[ eq : average density of state ] .
namely as shown in fig.@xmath161 @xmath162 for the critical strength of scattering @xmath163 . the other interesting phenomenon is a superconductor to insulator transition .
recently scalettar , trivedi and huscroft@xcite discovered a superconductor - insulator transition in the disordered attractive u hubbard model , using monte corlo simulations .
they found that for strong disorder the superconducting gap is replaced by an insulating gap , both for weak and strong interaction u. clearly the superconductor to insulator transition in our cpa calculaton has a similar origin , although our model of disorder is different .
in general the band filling @xmath86 is given by @xmath164 where @xmath165 and @xmath77 are the partial averaged occupation numbers on a or b sites respectively .
we shall discuss three different cases of eq .
[ eq : filling ] : case @xmath14 @xmath166 , case @xmath15 @xmath167 and case @xmath168 @xmath169 . in the first case
the a band is empty and the filling of the b band@xmath170 is @xmath171 for the second case the b band is completely , doubly , occupied and the a band completly empty.in the third case the a band is partially occupied with @xmath172 and the b band is fully occupied .
the second case is a special case of eq .
[ eq : afilling ] .
fig.@xmath173 shows the a , b and average normal density of states for these three different concentrations . in fig.@xmath174
the a sites are approximately empty @xmath175 but there is more than one electron on the b sites .
the graph plotted was for the case @xmath97 and @xmath176 , and therefore from eq .
[ eq : bfilling ] we have @xmath177 .
fig.@xmath174 is a band metal with hopping of electrons from the b sites to the b sites . in fig.@xmath178
the a sites are almost completely empty , @xmath179 , and the b sites are doubly occupied , @xmath180 .
therefore there is a gap at the fermi energy and the system is a band insulator . in fig.@xmath181
the b band is fully occupied while the a band is partially filled . for this case
we had @xmath97 and @xmath182 and so by eq .
[ eq : afilling ] , the band filling of the a sites is @xmath183 .
similarly fig.@xmath181 is a metal band with hopping from the a sites to the a sites . in fig.@xmath178
there is no hopping and we can regard this state as a band insulator . as illustrated in fig.@xmath184 , in the split band regime the superconductor gap closes,@xmath185 as @xmath186 but is replaced by an insulating gap for @xmath187 . evidently , this can also happen in the non particle - hole symmetric case as shown in fig.@xmath188 .
the general condition for the superconductor - insulator transition is saturation of the b sites with @xmath88 electrons of opposite spin while the a sites are empty .
this happens when @xmath189 in this case the fermi level lies in the gap and all the b states are filled and the a states are empty . for completeness
we show the suppression of @xmath104 by disorder for particle - hole symmetric and non particle - hole symmetric cases in fig.@xmath190 , obeying the condition from eq .
[ eq : transition ] .
now numerically we shall test our prediction , in eqs .
[ eq : afilling ] , [ eq : bfilling ] , for the two non - particle - hole symmetric cases:@xmath191 and @xmath192 , as discussed at the beginning of this section . in this cases
there is no superconductor - insulator transition . for a greater understanding of the details of the band splitting mechanism
we plot the density of states , the order parameter and the charge density on a and b sites . in the first case
the concentration of a sites is less than b sites , but in the second case the concentration of a sites is more than of b sites .
as expected , in fig.@xmath193 we see that there are two split bands , a normal empty band on the a sites and a superconducting band on b sites .
conversely in fig.@xmath194 the b band is a normal doubly occupied band and the a band is the superconducting band .
the effect of disorder on the average order parameter and the a and b sites local order parameters @xmath195 and charge densities @xmath196 are shown in figs.@xmath197 and fig.@xmath198 .
one can see that , for weak scattering the local order parameters and charge densities of a and b sites are approximatly the same but with increasing disorder the difference between them will increase , with one of @xmath66 or @xmath67 going to zero@xmath4depending on the relative concentrations@xmath7 . in fig.@xmath199 and fig.@xmath200 the concentration of a sites is less than that of b sites therefore @xmath201 , @xmath202 while @xmath203 , @xmath204 .
in contrast to this , in fig.@xmath205 and fig.@xmath206 the concentration of a sites is more than that of b sites , consequently @xmath207 , @xmath208 and @xmath209 , @xmath210 .
schematically fig.@xmath211 shows the physical mechanism of the band splitting in terms of a and b lattice sites .
clearly when all b sites become doubley occupied and all the a sites become empty there is no hopping and the system becomes insulating .
therefore as a function of the disorder , @xmath212 , or band filling @xmath86 there is a @xmath2 superconductor - insulator quantum phase transition .
this transition corroborates the suggestion of scalettar , trivedi and huscroft@xcite that disorder in superconductors can lead to a superconductor - insulator transition .
however it shuld be noted that they used a uniform random distribution of site energies , @xmath23 , wheras we used a binary alloy model .
in this paper , for the first time , we applied the full cpa for disorder superconductors . in our calculations
the self - consistency equations were solved fully within cpa so that self - consitency was properly satisfied on each type of cpa impurity sites , and not just on the average .
it is in this form that the cpa - hfg is the only ` controled ' mean field theory of disordered superconductors.@xcite we plotted the local order parameter and charge density of a or b atoms of a binary alloy for both weak and strong scattering limits .
we found that only for one special case are the order parameters of a and b sites equal , and consequently the fluctuations of @xmath12 are zero .
for this point we have shown analytically that the condition for anderson s theorem is fullfilled not only for weak scattering but also for strong scattering in a particle - hole symmetric band .
by contrast neither , the density of states nor @xmath13 is exactly constant , although in the gap equation , the quasi - particle energy gap is absolute . in the gap equation the normal clean system density of states in the gap equation
is replaced by the normal disorder average density of states , and therefore the only changes of the superconducting density of states and @xmath13 come from the latter . in narrow band binary alloy s - wave superconductors , we showed that strong disorder leads to two different interesting phenomena : @xmath14 band splitting with a quantum superconductor - insulator phase transition at @xmath2 , @xmath15 band splitting without a phase transition . in this last case in terms of concentration and average band filling , one band is normal @xmath4doubly occupied or empty@xmath7 and superconductivity is only present in the another band that is partially occupied .
99 see a.m. finkelstein , physica b , * 197 * , 636(1994 ) and references therein .
matthias , e.a wood , e. corenzwit and v.b bala , j.phys .
solids , * 1 * , 188 , ( 1956 ) .
b. batlogg , physica b , * 126 * , 275 ( 1984 ) .
woolam , r.b .
somoano , p. oconnor , phys .
lett . * 32 * , 712 ( 1974 ) .
anderson , j.phys .
solids * 11 * , 26 ( 1959 ) .
b.l.gyrffy , g. litak and k.i .
wysokinski , in _
`` fluctuation phenomena in high temperature superconductors '' _ , nato advanced study institute , series 3 high technology - vol .
32 , edited by m. ausloos and a.a .
varlamov ( kluwer academic , dordrecht , 1997 ) p. 385 .
a.a . abrikosov and l.p .
gorkov , zh .
fiz ( jetp)*35 * , 1158 ( 1958 ) .
a.a . abrikosov and l.p .
gorkov , zh .
fiz ( jetp)*36 * , 319 ( 1959 ) . g. rickayzen , _ `` green s function and condensed matter '' _ , ( academic press 1980 ) .
j b ketterson and s n song , _ `` superconductivity '' _ , ( cambridge university press 1999 ) .
zhitomirsky and m.b .
walker , phy .
lett . * 80 * , 5413 ( 1998 ) .
a. ghosal , m. randeria and n. trivedi , phy .
* 81 * 3940 ( 1995 ) .
r. opperman , z. phys .
b * 63 * , 33 ( 1985 ) . k. ziegler , commun . math . phys . * 120 * , 177 ( 1988 ) . j.f .
annett and n. goldenfeld j. low temp . phys . ,
* 89*,nos.1/2 , 197 ( 1988 ) . g. litak , b.l .
gyrffy , phys .
b * 62 * ( 2000 ) . g. litak and k.i .
wysokinski , r. micans and s. robaszkiewicz , physica c * 199 * , 191 ( 1992 ) .
g. litak , a.m. martin , b.l .
gyrffy , j.f .
annett and k.i .
wysokinski , physica c * 309 * , 257 ( 1998 ) .
a.m. martin , g. litak , b.l .
gyrffy , j.f .
annett and k.i .
wysokinski , phys .
b * 60 * , 7523 ( 1999 ) .
r. moradian , j.f . annett and b.l .
gyrffy , phys .
b * 62 * , 3508 ( 2000 ) .
r.j eliott , j.a .
krumhansl and p.l .
leath , rev .
, * 46 * , 465 ( 1974 ) .
g. litak , b.l .
gyrffy , and k.i .
wysokinski , physica c * 308 * , 132 ( 1998 ) .
g. litak , b.l .
gyrffy , and k.i .
wysokinski , molecular physics reports ( poland ) * 15/16 * , 87 ( 1996 ) .
h. lustfeld , j. low temp . phys .
* 12 * , 595 ( 1973 ) r.t .
scalettar , n. trivedi , and c. huscroft , phys .
b * 59 * , 4364 ( 1999 ) . | we have studied the effects of strong impurity scattering on disordered superconductors beyond the low impurity concentration limit . by applying the full cpa to a superconductiong a - b binary alloy
, we calculated the fluctuations of the local order parameters @xmath0 and charge densities , @xmath1 for weak and strong on site disorder .
we find that for narrow band alloy s - wave superconductors the conditions for anderson s theorem are satisfied in general only for the case of particle - hole symmetry . in this case
it is satisfied regardless whether we are in the weak or strong scattering regimes .
interestingly , we find that strong scattering leads to band splitting and in this regime for any band filling we have a critical concentration where a superconductor - insulator quantum phase transition occurs at @xmath2 .
# 1@xmath3#1 |
the standard model ( sm ) of particle physics is a renormalizable quantum field theory encoding our knowledge of the fundamental particles and their ( electroweak and strong ) interactions . despite its tremendous success to describe many phenomena with high accuracy for over 40 years including the recent experimental confirmation of the existence of its last missing piece , the higgs boson fundamental questions remain open ( such as the small higgs boson mass compared to the planck scale , dark matter , matter - antimatter asymmetry , neutrino masses , ... ) which may likely _ not _ be fully answered through the study of proton - proton collisions at the large hadron collider ( lhc ) .
notwithstanding their lower center - of - mass energies , high - energy @xmath0 colliders feature several advantages in new physics studies compared to hadronic machines , such as direct model - independent searches of new particles coupling to z*/@xmath13 * with masses up to @xmath14 , and clean experimental environment with initial and final states very precisely known theoretically , ( i.e. well understood backgrounds without `` blind spots '' of p - p searches ) .
combined with high - luminosities , an @xmath0 collider can thus provide access to studies with @xmath15 precision at the permille level , allowing indirect constraints to be set on new physics up to very - high energy scales @xmath16 ( 1 tev)/@xmath17 . plans exist to build future circular ( fcc - ee@xcite , cepc@xcite ) and/or linear ( ilc@xcite , clic@xcite ) @xmath0 colliders ( fig .
[ fig : ee_lumi ] ) .
colliders under consideration . _
_ ] the advantages of circular machines are ( i ) their much higher luminosity below @xmath18 400 gev ( thanks to much larger collision rates , adding continuous top - up injection to compensate for luminosity burnoff ) , ( ii ) the possibility to have several interaction points ( ips ) , and ( iii ) a precise measurement of the beam energy e@xmath19 through resonant transverse depolarization@xcite .
linear colliders , on the other hand , feature ( i ) much larger @xmath1 reach ( circular colliders are not competitive above @xmath18 400 gev due to synchroton radiation scaling as e@xmath20/r ) , and ( ii ) easier longitudinal beam polarization . at the fcc ( with a radius r = 80100 km ) , @xmath0 collisions present clear advantages with respect to lep ( e.g. @xmath2110@xmath22 more bunches , and @xmath23 @xmath240.1 mev compared to @xmath242 mev ) and to ilc ( crab - waist optics scheme , up to 4 ips ) yielding luminosities @xmath21(10@xmath2210 ) larger in the @xmath1 = 90350 gev range@xcite .
table [ tab : runs ] lists the target fcc - ee luminosities , and the total number of events collected at each @xmath1 obtained for the following cross sections ( including initial state radiation , and smearings due to beam - energy spreads ) : @xmath25 43 nb , @xmath26 0.29 fb , @xmath27 4 pb , @xmath28 200 fb , @xmath29 0.5 pb , and @xmath30 30 fb . with these target luminosities , the completion of the fcc - ee core physics program ( described in the next sections ) requires 10 years of running .
.__target luminosities , events / year , and years needed to complete the w , z , h and top programs at fcc - ee .
[ @xmath31 = 10@xmath32 @xmath33s@xmath2 corresponds to @xmath34 = 1 ab@xmath2/yr for 1 yr = 10@xmath35 s ] . _ _ [ cols="<,^,^,^,^,^,^",options="header " , ] figure [ fig : fccee_sm ] shows limits in the w - mass vs. top - mass plane ( left ) , and the energy reaches of a subset of dimension-6 operators of an effective field theory of the sm parametrizing possible new physics ( right)@xcite . such measurements impose unrivaled constraints on new weakly - coupled physics . whereas electroweak precision tests ( epwt ) at lep bound any bsm physics to be above @xmath36 7 tev , fcc - ee would reach up to @xmath37 100 tev .
@xmath38 plane at fcc - ee and other colliders@xcite .
right : energy reaches for dim-6 operators sensitive to ewpt , obtained from precision measurements at fcc - ee and ilc@xcite.__,title="fig:",height=170 ] @xmath38 plane at fcc - ee and other colliders@xcite .
right : energy reaches for dim-6 operators sensitive to ewpt , obtained from precision measurements at fcc - ee and ilc@xcite.__,title="fig:",height=170 ]
in the range of c.m . energies covered by the fcc - ee , higgs production peaks at @xmath18 240 gev dominated by higgsstrahlung ( @xmath39 ) , with some sensitivity also to vector - boson - fusion ( @xmath40 ) and the top yukawa coupling ( @xmath41 with a virtual higgs exchanged among the top quarks ) at @xmath1 = 350 gev . the target total number of higgs produced at the fcc - ee ( 4 ips combined , all years ) amounts to 2.1 million at 240 gev , 75000 in @xmath42 at 350 gev , and 19000 in s - channel @xmath43 at @xmath1 = 125 gev ( table [ tab : runs ] ) . with such large data samples ,
unique higgs physics topics are accessible to study : * high - precision model - independent determination of the higgs couplings , total width , and exotic and invisible decays ( table [ tab : higgs])@xcite . *
higgs self - coupling through loop corrections in hz production@xcite . *
first - generation fermion couplings : ( u , d , s ) through exclusive decays h@xmath44 ( @xmath45)@xcite , and electron yukawa through resonant @xmath46 at @xmath47@xcite .
the recoil mass method in @xmath48 is unique to lepton colliders and allows for a high - precision tagging of higgs events irrespective of their decay mode ( fig .
[ fig : hz ] ) .
it provides , in particular , a high - precision ( @xmath240.05% ) measurement of @xmath49 and , therefore , of @xmath50 . from the measured value of @xmath51 and different
known decays fractions @xmath52 , one can then obtain the total higgs boson width with @xmath53 uncertainty combining different final states .
the @xmath54 final state and the distribution of the mass recoiling against the lepton pair can also be used to directly measure the invisible decay width of the higgs boson in events where its decay products escape undetected .
the higgs boson invisible branching fraction can be measured with an absolute precision of 0.2% .
if not observed , a 95% c.l .
upper limit of 0.5% can be set on this branching ratio@xcite .
in addition , loop corrections to the higgsstrahlung cross sections at different center - of - mass energies are sensitive to the higgs self - coupling@xcite .
the effect is tiny but visible at fcc - ee thanks to the extreme precision achievable on the @xmath55 coupling .
indirect and model - dependent limits on the trilinear @xmath56 can be set with a @xmath57 uncertainty , comparable to that expected at hl - lhc .
+ the large higgs data samples available also open up to study exotic ( e.g. flavour - violating higgs ) and very rare sm decays .
first- and second - generation couplings to fermions are accessible via exclusive decays @xmath58 , for @xmath45 , with sensitivity to the u , d , s quark yukawas@xcite .
the @xmath59 channel appears the most promising with @xmath60 events expected .
the low mass of the electron translates into a tiny @xmath61 branching ratio @xmath62 which precludes any experimental observation of this decay mode and , thereby a determination of the electron yukawa coupling .
the resonant s - channel production , despite its small cross section@xcite , is not completely hopeless and preliminary studies indicate that one could observe it at the 5@xmath63-level accumulating 75 ab@xmath2 at fcc - ee running at @xmath1 = 125 gev with a c.m
. energy spread commensurate with the higgs boson width itself ( @xmath644 mev ) which requires beam monochromatization@xcite .
+ summarizing in terms of new physics constraints , deviations @xmath65 of the higgs boson couplings to gauge bosons and fermions with respect to the sm predictions can be translated into bsm scale limits through : @xmath66 . the expected 0.15% precision for the most precise coupling , @xmath67 , would thus set competitive bounds , @xmath36 7 tev , on any new physics coupled to the scalar sector of the sm .
the precision electroweak and higgs boson studies , summarized previously , not only impose competitive constraints on new physics at multi - tev scales but can also be interpreted in terms of limits in benchmark susy models ( cmssm and nuhm1)@xcite .
other studies exist that have analyzed the impact of fcc - ee on other key bsm extensions such as e.g. direct searches of dark matter ( dm)@xcite and right - handed neutrinos@xcite through z and h bosons rare decays .
figure [ fig : fccee_bsm ] ( left ) shows the limits in the plane ( branching ratio , dm mass ) for the decays @xmath68 . measurements of the invisible z and h widths are the best collider options to test dm lighter than @xmath69 that couples via sm mediators .
figure [ fig : fccee_bsm ] ( right ) shows the unrivaled limits that can be set in sterile neutrinos searches via decays @xmath70 with @xmath71 as a function of their mass and mixing to light neutrinos@xcite . vs. @xmath72 plane ( left)@xcite , and ( ii ) sterile neutrinos as a function of their mass and mixing to light neutrinos ( inverted hierarchy ) for 10@xmath3 z decays ( right)@xcite .
_ , title="fig : " ] vs. @xmath72 plane ( left)@xcite , and ( ii ) sterile neutrinos as a function of their mass and mixing to light neutrinos ( inverted hierarchy ) for 10@xmath3 z decays ( right)@xcite .
_ , title="fig : " ]
electron - positron collisions at @xmath18 90350 gev at the fcc - ee provide unique means to address many of the fundamental open problems in particle physics via high - precision studies of the w , z , higgs , and top - quark with permil - level uncertainties , thanks to the huge luminosities @xmath73 ab@xmath2 and the exquisite beam energy calibration . such measurements set indirect constraints on bsm physics up to scales @xmath74 tev for new particles coupling to the scalar and electroweak sm sectors , respectively .
rare higgs and z bosons decays are sensitive to dark matter and sterile neutrinos with masses up to @xmath75 60 gev .
m. bicer _ et al . _
[ tlep design study working group ] , jhep * 1401 * , 164 ( 2014 ) ; [ arxiv:1308.6176 [ hep - ex ] ] .
m. ahmad _ et al .
_ [ cepc - sppc study group ] , ihep - cepc - dr-2015 - 01 , ihep - th-2015 - 01 , hep - ep-2015 - 01 ( 2015 ) . t. behnke _ et al . _
[ ilc collaboration ] , arxiv:1306.6327 [ physics.acc-ph ] ( 2013 ) .
l. linssen , a. miyamoto , m. stanitzki and h. weerts [ clic collaboration ] , doi:10.5170/cern-2012 - 003 ; [ arxiv:1202.5940 [ physics.ins-det ] ] ( 2012 ) .
m. koratzinos , a. blondel , e. gianfelice - wendt and f. zimmermann , arxiv:1506.00933 [ physics.acc-ph ] ( 2015 ) .
f. zimmermann , 1st fcc annual workshop , washington ( usa ) , march 2015 ; https://indico.cern.ch/event/340703/session/109/contribution/190/attachments/668692/919126/fcc_ee_machine_overview_final.pdf [ ] r. tenchini , arxiv:1412.2928 [ hep - ex ] ( 2014 ) .
janot , arxiv:1512.05544 [ hep - ph ] , jhep to appear ( 2016 ) .
d. denterria _ et al .
_ , arxiv:1512.05194 [ hep - ph ] ( 2015 ) .
p. janot , jhep * 1504 * , 182 ( 2015 ) ; [ arxiv:1503.01325 [ hep - ph ] ] . j. ellis and t. you , arxiv:1510.04561 [ hep - ph ] ( 2015 ) .
m. mccullough , phys .
d * 90 * , 015001 ( 2014 ) [ arxiv:1312.3322 [ hep - ph ] ] .
a. l. kagan , g. perez , f. petriello , y. soreq , s. stoynev and j. zupan , phys .
lett . * 114 * , 101802 ( 2015 ) ; [ arxiv:1406.1722 [ hep - ph ] ] .
d. denterria , g.wojcik , r.aleksan , 8th fcc - ee phys .
workshop , lpnhe - paris , 2014 , + https://indico.cern.ch/event/337673/session/6/contribution/20/attachments/661128/908747/dde_ee_higgs_schannel_fcc_paris_oct14.pdf [ ] ; in preparation ( 2016 ) .
s. jadach and r. a. kycia , arxiv:1509.02406 [ hep - ph ] , phys .
b to appear ( 2016 ) .
a. strumia , 9th fcc - ee physics workshop , sns - pisa , feb .
2015 , + https://agenda.infn.it/getfile.py/access?contribid=19&sessionid=7&resid=0&materialid=slides&confid=8830 [ ] ( 2015 ) | the physics case for electron - positron beams at the future circular collider ( fcc - ee ) is succinctly summarized .
the fcc - ee core program involves @xmath0 collisions at @xmath1 = 90 , 160 , 240 , and 350 gev with multi - ab@xmath2 integrated luminosities , yielding about 10@xmath3 z bosons , 10@xmath4 w@xmath5w@xmath6 pairs , 10@xmath7 higgs bosons and 4@xmath810@xmath9 @xmath10 pairs per year .
the huge luminosities combined with @xmath11 knowledge of the c.m .
energy will allow for standard model studies at unrivaled precision .
indirect constraints on new physics can thereby be placed up to scales @xmath12 7 and 100 tev for particles coupling respectively to the higgs and electroweak bosons .
= 11.6pt |
the discovery of iron based high - temperature superconductors [ ] high - temperature superconductors stimulated quite intensive research work [ ] .
recently another class of high - t@xmath8 superconductors isostructural to 122-family of iron pnictides was discovered iron chalcogenides k@xmath9fe@xmath3se@xmath3 [ ] , cs@xmath9fe@xmath3se@xmath3 [ ] and ( tl , k)fe@xmath9se@xmath3 [ ] .
values of superconducting critical temperatures t@xmath8 are comparable for both families of pnictides and chalcogenides and are about 30 - 50k [ ] .
further interest to these chalcogenides was stimulated by experimental observation of rather nontrivial antiferromagnetic ordering with very high neel temperature of about 550k and fe vacancies ordering in the same range of temperatures in k@xmath10fe@xmath11se@xmath3 ( the so called 245 phase ) [ ] . despite intensive experimental work
there is still no consensus on the composition of the phase , responsible for high-@xmath12 superconductivity in these systems .
the most common point of view is that kfe@xmath3se@xmath3 ( 122 phase ) is the parent compound for superconductivity , while 245 phase is insulating [ ] . some other phases in this system were also reported [ ] .
below we shall concentrate on electronic structure calculations for the parent 122 phase with different levels of hole doping .
crystallographically pnictides afe@xmath3as@xmath3 , and chalcogenides fe(se , te ) , afe@xmath3se@xmath3 are quite similar with main structural motiff layers of fe(as , se)@xmath13 tetrahedra . recently reported theoretical electronic band structures of fe(se , te ) [ ] and afe@xmath3as@xmath3 [ ] are found to be nearly identical to each other especially for those bands crossing fermi level .
afe@xmath3as@xmath3 and afe@xmath3se@xmath3 compounds are simply isostructural .
however lda ( local density approximation ) electronic structure of afe@xmath3se@xmath3 differs quite remarkably from that of afe@xmath3as@xmath3 as was directly shown in refs .
[ ] . from the very beginning of studies of iron based superconductors
it was recognized , that the account of electronic correlations on fe sites is rather essential for correct description of the physics of pnictide materials [ ] . to this end
lda+dmft hybrid computational scheme [ ] was employed .
the main conclusion was that correlations lead to simple narrowing ( compression ) of lda bandwidth by the factor of the order of 2 or 3 .
this observation agrees rather well with variety of angular resolved photoemission spectroscopy ( arpes ) data on afe@xmath3as@xmath3 compounds .
fermi surface maps obtained from arpes experiments for afe@xmath3as@xmath3 are quite similar to those obtained from simple lda calculations : there are 2 or 3 hole cylinders around @xmath14-pointin the brillouin zone and 2 electron fermi surface sheets around ( @xmath15 ) point . up to
now there are only few lda+dmft papers devoted to fe chalcogenides [ ] .
recently we performed the investigation of electronic structure of hole doped iron chalcogenide k@xmath4fe@xmath5se@xmath3 in the normal phase [ ] , inspired by available arpes data , for this system [ ] , especially those obtained in ref .
[ ] , using both the standard lda+dmft and the novel lda@xmath0+dmft computational approach [ ] .
the results of our calculations agree rather well with the general picture of arpes , obtained in ref .
[ ] , with lda@xmath0+dmft showing slightly better agreement .
we showed that this iron chalcogenide is actually more strongly correlated in the sense of bandwidth renormalization ( energy scale compression by the factor of about 5 in the energy interval @xmath61.5 ev ) around the fermi level ) , than the typical iron pnictides ( with compression factor of about 2 or 3 [ ] ) , though the coulomb interaction strength is almost the same in both families .
moreover the k@xmath4fe@xmath5se@xmath3 system demonstrates absence of well defined quasiparticle bands on the fermi level in contract to pnictides .
in this paper we continue our lda@xmath0+dmft study of system .
we investigate the evolution of correlation effects upon hole doping , performing lda@xmath0 calculations at four doping levels : hypothetical stoichiometric composition with 29 valence electrons per unit cell through intermediate values of valence electrons 28 and 27.2 downto experimentally obtained composition k@xmath4fe@xmath5se@xmath3 with 26.52 electrons per unit cell .
we shall demonstrate that the actual doping dependence of correlation effects on the electronic structure in iron chalcogenides is apparently more complicated than in iron pnictides and does not reduce to the simple picture of universal bandwidth renormalization ( compression ) .
the paper is organized in a following manner . in sec .
[ cd ] we discuss crystallographic structure , methodological and computational details of lda@xmath0+dmft .
comparative study of lda@xmath0 bands and lda@xmath0+dmft spectral function maps within wide and narrow energy intervals around the fermi level , together with orbitally resolved densities of states are presented in sec .
finally we summarise our results in sec .
[ concl ] . + dmft calculated spectral function maps with renormalized by compression factor of 1.3 lda@xmath16 bands ( grey lines ) for k@xmath1fe@xmath2se@xmath3 for different hole doping levels : n@xmath17=29 , 28 , 27.2 , 26.52 ( from bottom to top ) along high symmetry direction of the first brillouin zone .
fermi level is zero.,title="fig : " ] + dmft calculated spectral function maps with renormalized by compression factor of 1.3 lda@xmath16 bands ( grey lines ) for k@xmath1fe@xmath2se@xmath3 for different hole doping levels : n@xmath17=29 , 28 , 27.2 , 26.52 ( from bottom to top ) along high symmetry direction of the first brillouin zone .
fermi level is zero.,title="fig : " ] + dmft calculated spectral function maps with renormalized by compression factor of 1.3 lda@xmath16 bands ( grey lines ) for k@xmath1fe@xmath2se@xmath3 for different hole doping levels : n@xmath17=29 , 28 , 27.2 , 26.52 ( from bottom to top ) along high symmetry direction of the first brillouin zone .
fermi level is zero.,title="fig : " ] + dmft calculated spectral function maps with renormalized by compression factor of 1.3 lda@xmath16 bands ( grey lines ) for k@xmath1fe@xmath2se@xmath3 for different hole doping levels : n@xmath17=29 , 28 , 27.2 , 26.52 ( from bottom to top ) along high symmetry direction of the first brillouin zone .
fermi level is zero.,title="fig : " ]
the k@xmath9fe@xmath3se@xmath3 system is isostructural to ba122 pnictide ( for the last one see ref . [ ] ) with ideal body centered tetragonal space group i4/mmm .
the k@xmath9fe@xmath3se@xmath3 has @xmath18=3.9136 and @xmath19=14.0367 with k ions occupying @xmath20 , fe
@xmath21 and se @xmath22 positions with @xmath23=0.3539 [ ] . this crystal structure was used in band structure calculations for k@xmath4fe@xmath5se@xmath3 within the linearized muffin - tin orbitals method ( lmto ) [ ] using default settings [ ] . to take into account local coulomb correlations
we apply here lda@xmath0+dmft [ ] approach which is a modification of the well known lda+dmft method [ ] .
lda+dmft hamiltonian is usually written as @xmath24 the general problem with lda+dmft is that some portion of local electron - electron interaction for presumed correlated @xmath25 shells is already included in the standard lda ( @xmath26 ) . to avoid its double counting due to hubbard interaction @xmath27 one has to subtract from @xmath26 the so called double counting correction term @xmath28 .
( explicit expressions for @xmath26 and @xmath27 can be found elsewhere [ ] . )
the lda@xmath0+dmft approach is the new attempt to solve the double counting problem , which is due to the absence of universal expression for @xmath28 , since there is no explicit microscopic ( or diagrammatic ) link between the model ( hubbard like ) hamiltonian approach and the standard lda . to be short the main idea of
the lda@xmath0+dmft is _ to exclude explicitly _ from the charge density the contribution of the presumably strongly correlated @xmath25 shells on the initial step of lda calculations .
then this redefined charge density ( for some chosen orbital basis @xmath29 ) @xmath30 is used to calculate the local exchange - correlation energy @xmath31 in lda and perform the self - consistent lda band structure calculations for correlated bands at the lda stage of lda+dmft .
further on the local @xmath32 electron correlations are taken into account within the dmft .
all states not counted as strongly correlated are then treated with the full power of dft / lda and _ full _ @xmath33 in @xmath34 .
once this lda@xmath0 calculations with redefined charge density were done , what is left for the correlated states out of interaction on the lda stage would be just the hartree contribution which can be written in the fully localized limit form ( fll ) , and which is the most consistent definition of double counting term here ( other forms can be also used [ ] ) : @xmath35 where @xmath36 is the total number of electrons on strongly interacting orbitals and number of electrons per spin , @xmath37 .
effective five orbital impurity problem for k@xmath1fe@xmath2se@xmath3 within dmft was solved by hirsh - fye quantum monte - carlo algorithm [ ] , at temperature 280k .
lda@xmath0+dmft densities of states and spectral functions were obtained as discussed in ref .
coulomb parameters were taken to be @xmath38=3.75 ev and @xmath39=0.6 ev respectively [ ] , which are very close to calculated ones [ ] . to define dmft lattice problem we used the full ( i.e. without any downfolding or projecting ) lda hamiltonian , which included all fe-3d , se-4p and k-4s states .
in fig . 1 there we present the comparison of lda@xmath16+dmft calculated spectral function maps in the wide energy window @xmath62 ev along high symmetry directions in the first brillouin zone with the renormalized lda@xmath16 bands ( grey lines ) for k@xmath1fe@xmath2se@xmath3 at different hole doping levels n@xmath17 .
renormalization ( bandwidth compression ) factor of lda@xmath16 bands here is only 1.3 , so that bandwidth renormalization due to correlations is rather weak .
the lower panel of fig .
1 lda@xmath0+dmft shows data for stoichiometric kfe@xmath40se@xmath3 compound with total number of valence electrons n@xmath17=29 . for this composition
all quasiparticle bands are rather well defined for this wide energy range around the fermi level . for hole doped cases with
n@xmath17=28 , 27.2 and 26.52 ( 2nd , 3rd and 4th panels from bottom correspondingly ) we see that fe-3d bands obtained from lda@xmath0+dmft become less pronounced with hole doping
. overall rigid shift of lda@xmath16 bands from stoichiometric case down to the most hole doped one is about 0.3 ev . in fig .
2 we show lda@xmath0+dmft spectral function maps in the vicinity ( -0.5 0.2 ev ) of the fermi level with dominant orbital character of quasiparticle bands denoted explicitly by black squares for @xmath41 , black circles for @xmath42 , white circles for @xmath43 and white squares for @xmath44 .
we can see thet the orbital characters and forms of quasiparticle bands in this energy interval ( -0.1 0.1 ev ) change with the increase of hole doping , although characters of quasiparticle bands located outside this region remain the same .
main orbital character of bands crossing the fermi level is @xmath41 and @xmath42 . also from fig . 2
one can conclude that close enough to the fermi level ( for all hole dopings ) there are uniformly no well defined quasiparticle bands ( though some low intensity maxima of spectral density can still be seen ) .
this fact tells that the k@xmath1fe@xmath2se@xmath3 for all hole dopings is more correlated than 122 pnictide system .
[ cols="^,^,^,^,^",options="header " , ] [ tab1 ] this `` pseudogap '' like behaviour can be explicitly observed in fig .
3 for all fe-3d orbitals , where orbitally resolved bare lda@xmath16 and lda@xmath0+dmft densities of states ( dos ) for all four hole doping levels are presented . also by watching these doses
one can actually see , that upon hole doping correlation effects become stronger .
this fact manifests itself in a different way for orbitals of various symmetry .
first of all , for all fe-3d orbitals we observe narrowing of doses . for @xmath43 ( third panel from top ) and @xmath44 ( upper panel )
this narrowing is most evident at @xmath61 ev and @xmath60.5 correspondingly .
for @xmath41 ( second panel from top ) and @xmath42 ( bottom panel ) increase of narrowing with doping is mostly concentrated within the energy interval @xmath60.4 ev . to get more insight into the lda@xmath0+dmft self - energy effects on bare lda@xmath0 bands we have determined energy scale renormalizations and energy shifts for variety of separate dispersions of bare lda@xmath0 band structure depicted in fig . 4 , which rather accurately fit bare bands to those plotted in the fig . 2 . also in fig .
4 we show the standard lda bands ( dashed lines ) , just to emphasize that lda@xmath16 dispersions are quite close to lda ones ( see also refs . [ ] ) . the obtained energy scale renormalization ( bandwidth compression ) factors and energy shifts ( shown in brackets ) are collected in the table 1 for all hole doping levels .
these results show more complicated picture of bare lda@xmath16 dispersion transformations , than those obtained in ref . [ ] , where we have proposed that all fe-3d band dispersions for the k@xmath4fe@xmath5se@xmath3 should be just compressed by a factor of 5 to get reasonable agreement with experiment .
in contrast to our previous work ref .
[ ] here we fit lda@xmath0 bare bands exactly to maxima positions of the spectral function ( see stars on fig . 2 ) .
thus we get more detailed picture which in general agrees with our early conclusions of ref .
here we actually see , that different parts of bands in the first brillouin zone acquire different ( bandwidth or mass ) renormalization factors , which change with doping . for example the @xmath45 band ( see fig .
4 ) almost does not have doping dependence at all ( see table 1 ) .
renormalization for the part of @xmath41 band ( 2 ) only slightly depend on doping .
however , the @xmath41 band ( 1 ) becomes monotonically more correlated ( renormalization factor grows up to 4.2 ) while hole doping increase .
correlation renormalization of the @xmath42 orbital demonstrates nonmonotonic behaviour and is more pronounced for intermediate dopings .
the renormalization factor of the @xmath43 band near the @xmath14 point abruptly increase up to 4.7 at n@xmath17=26.52 . here
one should note that all four doping levels were treated with the same coulomb interaction parameters . despite this fact upon doping several orbitals of k@xmath1fe@xmath2se@xmath3 become more correlated ( narrowed ) .
this phenomenon is apparently related to the change of the correlated orbitals occupancy .
the data collected in the table 1 might be helpful for interpretation of aprpes spectra in simple terms of renormalized bare lda band dispersions .
in this paper we have performed detailed lda@xmath0+dmft study of correlation effects in k@xmath1fe@xmath2se@xmath3 system at four hole doping levels : from hypothetical stoichiometric composition with 29 valence electrons per unit cell through intermediate values of valence electrons 28 and 27.2 towards composition k@xmath4fe@xmath5se@xmath3 with 26.52 electrons per unit cell , for which there are available arpes data on electronic dispersions [ ] . within dmft correlation effects
are concentrated in the self - energy which provides two types of modifications of the bare spectra at each energy : broadening ( lifetime effects ) by imaginary part of self - energy and energy shift due to its real part .
we have shown , that in rather wide energy window @xmath62 ev ( in terms of lda energy scale ) for all dopings the lifetime effects are relatively weak and renormalization ( compression ) of quasiparticle bandwidths remains nearly the same and of the order of @xmath71.3 .
however near the fermi level self - energy effects become more pronounced and complicated . in particular , the renormalization ( bandwidth compression ) factor grows from 1.3 for stoichiometric composition upto nearly 5 for k@xmath4fe@xmath5se@xmath3 ( for @xmath41 and @xmath43 bands ) .
though all calculations were done with the same value of coulomb ( hubbard ) interaction , this increase of renormalization upon doping tells us that correlations also increase with doping . also following recent tendency of experimental arpes papers
where energy shifts and and renormalization factors are determined separately for different bare lda bands [ ] to fit arpes data , we have provided quasiparticle mass renormalizations and energy shifts , which rather accurately transform bare fe-3d lda@xmath0 bands of various symmetries into lda@xmath0+dmft quasiparticle bands in different regions of the first brillouin zone . in fact , the detailed picture of band renormalizations close to the fermi level in k@xmath1fe@xmath2se@xmath3 is pretty complicated .
for example , the @xmath45 bands near the @xmath14-point and the @xmath41(2 ) bands in the middle of @xmath14-x and p - n directions almost do not feel doping changes . on the opposite , the @xmath41(1 )
bands in all directions and the @xmath43 band near the @xmath14-point become monotonically more correlated ( renormalized ) with doping increase .
finally , correlation renormalization of the @xmath42 band demonstrates non monotonic behaviour with doping , becoming more correlated for intermediate dopings .
however , the general conclusion of ref .
[ ] remains valid k@xmath1fe@xmath2se@xmath3 systems demonstrates more pronounced correlation effects in contrast to 122 iron pnictides .
this is clearly reflected in the absence of well defined quasiparticle bands in the vicinity of the fermi level , which demonstrates a kind of `` pseudogap '' behavior ( `` dark '' region around the fermi level in fig .
the interesting problem for the future studies is possible manifestation of these effects in optical conductivity , as well as their role in the formation of superconducting state .
we thank a.i . poteryaev for providing us qmc code and many helpful discussions . this work is partly supported by rfbr grant 11 - 02 - 00147 and
was performed within the framework of programs of fundamental research of the russian academy of sciences ( ras ) `` quantum mesoscopic and disordered structures '' ( 12-@xmath46 - 2 - 1002 ) and of the physics division of ras `` strongly correlated electrons in solids and structures '' ( 012-t-2 - 1001 ) .
nsp acknowledges the support of the dynasty foundation and international center of fundamental physics in moscow .
dmft(qmc ) calculations were performed on supercomputer `` uran '' in the institute of mathematics and mechanics of ub ras . + dmft spectral density maps along high symmetry directions of the first brillouin zone .
maxima of intensity are shown by stars .
dominant orbital character of quasiparticle bands is marked as follows : black squares - @xmath41 , black circles - @xmath42 , white circles - @xmath43 , white squares - @xmath44 .
fermi level is zero . ] ( dashed black line ) and lda@xmath16+dmft density of states ( dos ) for k@xmath1fe@xmath2se@xmath3 for different dopings n@xmath17 .
thick black line corresponds to n@xmath17=26.52 , thick grey line
n@xmath17=27.20 , thin black line
n@xmath17=28.00 , grey dot - dashed line n@xmath17=29.00 .
fermi level is zero . ]
( full line ) and lda ( dashed line ) band dispersions for stoichiometric kfe@xmath3se@xmath3 system with orbital characters explicitly shown .
numbers in brackets after orbital symmetry symbol enumerate corresponding parts of the bands in the first brillouin zone ( see table 1 ) . ]
k. held , i.a .
nekrasov , n. blmer , v.i .
anisimov , d. vollhardt .
b * 15 * , 2611 ( 2001 ) ; g. kotliar , s.y .
savrasov , k. haule , v.s .
oudovenko , o. parcollet , c.a .
phys . * 78 * , 865 ( 2006 ) .
y. zhang , l. x. yang , m. xu @xmath47 , nature materials * 10 * , 273 ( 2011 ) , arxiv : 1012.5980 ; x .- p .
wang , t. qian , p. richard @xmath47 , europhys .
lett . * 93 * , 57001 ( 2011 ) , arxiv : 1101.4923 ; d. g mou , sh .
liu , x. jia @xmath47 , phys .
lett . * 106 * , 107001 ( 2011 ) ; l. zhao , d. mou , sh .
liu @xmath47 , phys .
b * 83 * , 140508(r ) ( 2011 ) . | we present detailed lda@xmath0+dmft investigation of doping dependence of correlation effects in novel k@xmath1fe@xmath2se@xmath3 superconductor .
calculations were performed at four different hole doping levels , starting from hypothetical stoichiometric composition with total number of electrons equal to 29 per unit cell through 28 and 27.2 electrons towards the case of 26.52 , which corresponds to chemical composition k@xmath4fe@xmath5se@xmath3 studied in recent arpes experiments . in general case
the increase of hole doping leads to quasiparticle bands in wide energy window @xmath62 ev around the fermi level becoming more broadened by lifetime effects , while correlation induced compression of fe-3d lda@xmath0 bandwidths stays almost the same and of the order of @xmath71.3 for all hole concentrations .
however close to the fermi level situation is more complicated . here in the energy interval from -1.0 ev to 0.4 ev the bare fe-3d lda@xmath0 bands
are compressed by significantly larger renormalization factors up to 5 with hole doping increase , while the value of coulomb interaction remains the same .
this fact manifests the increase of correlation effects with hole doping in k@xmath1fe@xmath2se@xmath3 system .
moreover in contrast to typical pnictides k@xmath1fe@xmath2se@xmath3 does not have well defined quasiparticle bands on the fermi leves but pseudogap like `` dark '' region instead .
we also find that with the growth of hole doping fe-3d orbitals of various symmetries are affected by correlations in a different way in different parts of brillouin zone . to illustrate this we determine quasiparticle mass renormalization factors and energy shifts , which transform the bare fe-3d lda@xmath0 bands of various symmetries into lda@xmath0+dmft quasiparticle bands .
these renormalization factors effectively mimic more complicated energy dependent self - energy effects and can be used to analyze the available arpes data . |
exclusive reactions of pion production are an important source of information on the nature of the strong interaction . they allow to study the dynamics of the processes as well as the pion - pion interaction . up to now a number of experiments which study production of pions by pions were fulfilled on unpolarized proton targets @xcite in the intermediate energy region .
their results showed the predominant role of the one pion exchange ( ope ) and noticeable effects of absorption .
the whole picture was satisfactorily described by models of the one pion exchange with absorption ( opea ) @xcite and regge - model with moving branchings @xcite .
nevertheless the absence of information on the spin dependence of the reaction at the intermediate energies ( except the preliminary results of this experiment @xcite ) did not allow to go further in the understanding of the reaction mechanism , particularly it was not possible to estimate contribution of the axial - vector exchange .
here we should mention two experiments with transversely polarized targets performed at high energies .
authors of the first experiment @xcite used a polarized proton target to study the reaction @xmath6 at the incident pion beam momentum 17.2 gev / c . in the other @xcite the reaction @xmath7
was explored at 5.98 and 11.85 gev / c on a polarized deutron target . in both experiments
large asymmetries ( up to 50% ) in the @xmath0-meson production were observed .
this was explained as a result of the essential contribution of the axial - vector exchange ( a@xmath8-meson ) , that caused doubts about the results of partial wave analyses , which neglected this mechanism .
untrivial results of these experiments put forward the question of the energy dependence of the asymmetries .
the latter would allow to estimate intercept of axial - vector regge trajectory .
the angular distributions of the products of the reaction ( [ eq : pip ] ) at the transversely polarized target allow to fulfill a nearly full model independent amplitude analysis of the reaction @xcite .
so in the region of the dipion masses below 1 gev where contribution of waves with @xmath9 is negligible , one can directly reconstruct modules of all the transversity amplitudes and most of the phases .
besides all , this makes possible to separate the intensities of the dipions produced in the @xmath10- and @xmath11-wave states without additional assumptions and thus study the mass dependence of @xmath10-wave which is important for the scalar meson spectroscopy . up to now the main source of information about light quark meson spectroscopy
is the partial wave analyses of the pion - pion scattering .
old analyses were based on the data obtained on unpolarized targets . in order to study
the pion - pion interaction authors of these works separated the diagram of the one pion exchange and then made a transition from the scattering on the virtual pion to the scattering on the real one @xcite .
the other method is to perform a partial wave analysis basing on the amplitudes of the reaction .
but the model independent amplitude reconstruction is impossible without measurements on polarized targets . in the absence of such data
the assumption was used that in the helicity system one could disregard amplitudes without the nucleon helicity flip @xcite .
in addition the partial wave analyses used strong but not well proven assumptions of `` spin coherence '' , that is the absence of @xmath5 exchanges , and `` phase coherence '' , that is the equality of phases of @xmath11-waves with zero and unit helicities . on the other hand
the analysis of the situation using roy s equations @xcite , incorporating conditions of unitarity , analyticity and crossing symmetry , showed that solutions of partial wave analysis are self content and generally correct @xcite , though individual analyses had noticeable difference in the @xmath10-wave phases .
the results of partial wave analyses had an ambiguity in the behavior of the @xmath12 phase ( @xmath10-wave with zero isospin ) at dipion masses above 700 mev , so called `` up - down '' ambiguity .
the `` up '' solution revealed more quick change of the @xmath12 phase under the peak of @xmath0-meson and was considered as resonance opposite to the `` down '' solution .
the direct measurements of the mass dependence of the pion - pion interaction intensity with two @xmath13-mesons in the final state did not lead to a clear conclusion about the existence of the relatively narrow scalar resonance @xcite .
the experiments at high energies with polarized targets @xcite showed large polarization effects that contradicted to the assumptions on which partial wave analyses were based .
the model - independent amplitude analysis of the data from @xcite performed by m. svec in @xcite gave 4-fold ambiguity in the solution , but all the variants corresponded to the resonant behaviour of the @xmath10-wave in the @xmath0-meson region .
the important additional argument in favour of the existence of the resonance , obtained in this analysis , is the weak mass dependence of the relative phase between @xmath10-wave and @xmath11-wave with zero helicity .
the simultaneous amplitude analysis of the reaction ( [ eq : pip ] ) together with the reaction @xmath14 @xcite gave two branches of the solution .
one corresponded to a relatively narrow resonance ( @xmath15 mev and @xmath16 mev ) and the other to a wide one ( @xmath17 mev and @xmath18 mev ) .
the recent partial wave analysis of the data @xcite , which took into account the axial - vector exchange , gave also two branches of the solution for the phase @xmath12 @xcite .
so the current amount of the experimental data over peripheral meson production is not sufficient to solve the `` up - down '' ambiguity , which preserves the open problem in the meson spectroscopy .
nevertheless this data points out to the existence of the scalar - isoscalar resonance in the mass region below 1 gev ( wide sigma - meson @xmath19 and/or narrow @xmath20 ) .
the history of the search of the scalar - isoscalar meson is long and contradictory .
the @xmath21-meson was for the first time proposed in the work @xcite .
up to 1974 this resonance was present in the pdg tables as @xmath21 , @xmath22 or @xmath23 . in 1996
it again appeared in pdg tables @xcite as @xmath19 .
thus the modern state of the problem of light scalar - isoscalar meson requires additional experimental information and to provide such information was one of the goals of this work .
we present here a new experimental data on the reaction ( [ eq : pip ] ) on the polarized proton and liquid hydrogen targets in the region of small momentum transfered at @xmath24 gev / c .
the results are compared with the ones at high energies ( @xmath25 gev / c @xcite ) .
the results are analyzed with the help of the regge phenomenology , which gives the natural connection between the high and intermediate energy regions .
here we only want to review the basic formalism which could be found elsewhere in more details @xcite . at fixed beam momentum
the reaction ( [ eq : pip ] ) is described by 5 kinematic variables .
we use two energy variables ( the squared momentum transferred @xmath26 instead of @xmath26 ] and the dipion invariant mass @xmath28 ) and three angular variables .
one angle @xmath29 describes the reaction plane and is defined as the angle between the normal to the reaction plane and the target polarization @xmath30 .
the other two angles describe the dipion decay in its rest frame and are defined as angles @xmath31 ( gotfried - jackson angle ) and @xmath32 ( treiman - yang angle ) of the negative pion in the jackson coordinate system ( helicity system of the t - channel , axis @xmath33 along the beam and @xmath34 perpendicular to the reaction plane ) .
the dynamics of the reaction ( [ eq : pip ] ) could be described by the set of the helicity amplitudes @xmath35 , where @xmath36 and @xmath37 are the spin and the helicity of the dimeson , while @xmath38 and @xmath39 are the neutron and proton helicities correspondingly . at @xmath40
gev the dipion production with spins @xmath41 ( @xmath10-wave ) and @xmath42 ( @xmath11-wave ) is dominant .
thus the reaction ( [ eq : pip ] ) in the energy region under consideration is described by 8 complex amplitudes : two for dipion production in @xmath10-state ( with and without neutron helicity flip ) and 6 for dipion production in @xmath11-state ( with dipion helicities + 1,0,-1 with or without neutron helicity flip ) .
an experiment with a transversely polarized target yield 15 elements of spin density matrix ( sdme ) @xmath43 , defining the angular distributions of the reaction products .
nine of them reflect the interference of the helicity flip and non - flip amplitudes and can be measured only in polarized target experiment . the normalized process intensity as function of the sdme and the angular variables
could be written as follows : @xmath44 where @xmath45 with two additional relations : @xmath46 the first of them expresses the normalization and the second defines @xmath47 conventional polarized target asymmetry .
the data on transversely polarized target is best analyzed in terms of the nucleon transversity amplitudes with definite t - channel exchange naturality : @xmath10 , @xmath48 , @xmath49 , @xmath50 and @xmath51 , @xmath52 , @xmath53 , @xmath54 .
these amplitudes are linear combinations of helicity amplitudes and correspond to definite recoil nucleon transversity ( `` down '' and `` up '' respectively ) .
the amplitudes @xmath10 , @xmath51 and @xmath48 , @xmath52 describe production of the @xmath10-wave and @xmath11-wave dipions with zero helicity , respectively .
amplitudes @xmath10 , @xmath51 , @xmath48 , @xmath52 , @xmath49 , @xmath53 are dominated by unnatural exchange , while amplitudes @xmath50 , @xmath54 - by natural one .
@xmath55 these amplitudes are connected to sdme by the following equations : @xmath56 it is also useful to introduce partial - wave intensities @xmath57 and partial - wave polarizations
@xmath58 : @xmath59 where @xmath60 .
the absolute values of the transversity amplitudes could be reconstructed from the experiment on a polarized target .
this is not so for the helicity amplitudes .
the matrix elements measured allow to reconstruct the absolute values of all the amplitudes involved and all relative phases between them , except the relative phase between the two groups of amplitudes with different recoil nucleon transversity .
the relations between the sdme and the amplitudes could be put in the form of two independent similar systems of equations @xcite , one for amplitudes
@xmath10 , @xmath48 , @xmath49 and @xmath54 and the other for amplitudes with the opposite recoiled nucleon transversity .
each of the systems could be reduced to a cubic equation in respect to @xmath61 or @xmath62 , correspondingly , which has two positive solutions .
this provides two - fold ambiguity in the resulting amplitudes and four - fold ambiguity in the partial wave intensities .
the apparatus used in this work is the further development of the experimental setup * spin * @xcite designed for the investigation of polarization effects in two- and three - particle reactions with two charged particles in the final state .
it is a two - arm magnetic spectrometer with wire chambers and a transversely polarized proton target .
spectrometric capabilities of the setup are provided by the magnet of the polarized target equipped with spark chambers placed in its field .
the apparatus is located at the itep accelerator in the beam of particles with maximum momentum 2.1 gev / c .
the typical beam intensity is @xmath63 pions per spill , the spill duration is about 1 s and the frequency is 2025 spills per minute .
the beam angular divergences are @xmath64 and @xmath65 mrad in the horizontal and vertical planes respectively .
its dimensions on the target are 810 mm in both projections .
the momentum spread of the beam is @xmath66% .
the central beam momentum was determined with the precision better than 0.5% by bending of the incident particles in the well known magnetic field of the polarizing magnet and by the time - of - flight difference between negative pions and anti - protons of the beam @xcite .
the setup * spin * is shown in fig .
[ fig : spin ] .
the basic elements are : * polarized proton target ( * ppt * ) , placed in the center of a wide aperture magnet ( * m * ) .
the magnet ( * m * ) is c - shaped iron magnet with approximately axially symmetrical horizontal field .
the field in the center is 2.55 t with non - uniformity in the region of the target less than @xmath67 .
the bending power from the center is 0.45 t@xmath68 m .
the protons of the target are polarized by the dynamic nuclear orientation method .
the target material is propanediol c@xmath69h@xmath70o@xmath71 doped by hbma - cr@xmath72 complexes .
target dimensions are @xmath73 mm@xmath74 ( width@xmath68height@xmath68length ) .
the working temperature of the target 0.5 k is achieved with a -evaporation type cryostat .
the polarization is measured by the nmr method with the precision of 5% .
the average polarization during the data taking was @xmath75% .
* liquid hydrogen target , which can be placed in the same cryostat as the polarized target .
this allows to make measurements on both types of targets without additional readjustments of the setup just by replacing the target .
the liquid hydrogen is produced by the cooling of the gaseous hydrogen by the liquid helium at the atmospheric pressure .
the helium stream in the heat exchanger was automatically adjusted so as to keep the hydrogen pressure in the closed volume of the target constant .
this allows to maintain the volume of the liquid hydrogen with the precision better than 2% .
* wire spark chambers with magnetostrictive readout ( * mwsc123 * ) , which allow to measure trajectories of the incoming pion ( * mwsc18 * ) and two outgoing charged particles : ( * mwsc916 * ) in the lower arm and ( * mwsc1723 * ) in the upper arm . the chambers ( * mwsc912 * ) of the lower arm placed into the magnetic field provide the spectrometric capabilities of the apparatus .
* system of scintillation and cherenkov counters ( * s110 * and * 1,2 * ) for triggering .
the threshold cherenkov counters provide suppression of protons and are used in the trigger , which was formed according to the equation : @xmath76 where veto counter * s10 * is placed on the continuation of the beam line after the magnet .
the adjustment of the apparatus was performed in two subsidiary runs . in one of them
the special copper wire target was used to produce events with known vertex coordinates and in the other the beam which momentum was controlled with time - of - flight technique was bended by the magnet into different arms of the setup .
data processing was performed in two stages . at the first stage
the kinematic parameters of the individual events were reconstructed . for every event the interaction point and particles momenta as well as
their error matrixes were reconstructed taking into account multiple scattering in the materials of the setup @xcite .
the final kinematic parameters were found using the hypothesis that the missing particle is a neutron .
the distribution of the events over missing mass squared is shown in fig . [
fig : mmass ] ( solid line ) .
further event selection was based on the @xmath77 of the hypothesis that the missing particle is a neutron .
we assumed that in the background events one or more missed pions are produced and then the missing mass in such events should be greater the mass of a neutron .
we made processing with different rejections over @xmath77 and missing mass and found no significant difference .
finally we rejected events whose missing mass was above the mass of nucleon and @xmath78 . in fig .
[ fig : mmass ] `` good '' events are shown by hatched area .
the event distributions over momentum transferred @xmath79 , dipion invariant mass @xmath28 , gotfried - jackson angle @xmath31 and treiman - yang angle @xmath32 on the liquid hydrogen target are shown in fig .
[ fig : events ] .
these figures describe the kinematic region covered by the setup . 99 l.d .
jacobs et al . , _ phys .
rev . _ * d6 * ( 1972 ) 1291 .
b. haber et al . , _ phys .
_ * d10 * ( 1974 ) 1387 . a.b .
wicklund et al . , _ phys .
rev . _ * d17 * ( 1978 ) 1197 . g.a .
leksin _ uspehi fizicheskih nauk _ * 102 * -3 ( 1970 ) 337 ; + a.v .
arefyev et al . , _ yaderaya fizika _ * 27 * ( 1964 ) 161 .
charlesworth et al . , _ nucl .
_ * b65 * ( 1973 ) 253 .
reynolds et al . , _ phys .
_ * 184 * ( 1969 ) 1424 .
bacon et al . , _ phys .
* 157 * ( 1967 ) 1263 .
w. ochs and f. wagner , _ phys .
_ * 44b * ( 1973 ) 271
kaydalov and b.m .
karnakov , _ yadernaya fizika _ * 11 * ( 1970 ) 216 ; + b.m .
guzhavin , a.b .
kaydalov and a.f .
nilov , _ yadernaya fizika _ * 15 * ( 1972 ) 162 .
alekseev et al . , _ @xmath80 newsletter _
* 6 * ( 1992 ) 35 ; + i.g .
alekseev et al . , _ yad . fiz .
_ * 55 * ( 1992 ) 1912 .
h. becker et al . , _ nucl .
phys . _ * b150 * ( 1979 ) 301 ; + h. becker et al . , _ nucl .
phys . _ * b151 * ( 1979 ) 46 .
a. de lesquen et al . , _ phys .
rev . _ * d32 * ( 1985 ) 21 ; + a. de lesquen et al . ,
_ * d39 * ( 1989 ) 21 ; + a. de lesquen et al .
rev . _ * d42 * ( 1990 ) 934 ; + a. de lesquen et al .
rev . _ * d45 * ( 1992 ) 55 ; + a. de lesquen et al .
* d46 * ( 1992 ) 949 .
g. lutz and k. rybitcki , max planc institute , munich , internal report no .
mpi - pae / exp .
ei.75 , 1978 ( unpublished ) . k.n . muhin and o.o .
patarakin , _ uspehi fizicheskih nauk _ * 133 * ( 1981 ) 377 .
williams , _ phys . rev .
_ * d1 * ( 1970 ) 1312 .
roy , _ phys .
* b36 * ( 1971 ) 353 .
patarakin , v.n .
tihonov and k.n .
muhin , _ nucl .
_ * a598 * ( 1996 ) 335 .
cason et al . , _ phys .
rev . _ * d28 * ( 1983 ) 1586 ; and references 12 - 19 therein c. amsler et al .
_ * b342 * ( 1995 ) 433 ; + c. amsler et al . , _ phys .
_ * b355 * ( 1995 ) 425 .
m. svec , _ phys . rev . _ * d53 * ( 1996 ) 2343 ; + m. svec , _ phys .
_ * d55 * ( 1997 ) 5727 ; + m. svec , * hep - ph*/9707495 .
donohue and y. leroyer , _ nucl .
* b158 * ( 1979 ) 123 .
r. kaminski , l. lesniak and k. rybicky , _ zeit . f. phys .
_ * c74 * ( 1997 ) 79 .
m. gell - mann and m. levy , _
nuovo cim .
_ * 16 * ( 1960 ) 705 .
review of particle physics , _ phys .
rev . _ * d54 * ( 1996 ) 1 .
e. byckling and k. kajantie , _ particle kinematics _ , john wiley and sons , london 1973 .
k. gotfried and j.d .
jackson , _
nuovo cim . _ * 34 * ( 1964 ) 735 ; + j.d .
jackson , _
nuovo cim . _
* 34 * ( 1964 ) 1644 .
kimel and j.f .
owens , _ nucl .
_ * b122 * ( 1977 ) 464 .
b. hyams et al . , _ nucl .
* b64 * ( 1973 ) 134 ; + g. grayer et al . , _ nucl .
phys . _ * b75 * ( 1974 ) 189 ; + i.g .
alekseev et al . , _ nucl .
phys . _ * b348 * ( 1991 ) 257 .
alekseev et al . , preprint itep 184 - 88 , moscow 1988 .
h. eichinger and m. regler , _ review of track fitting methods in counter experiments , _ book cern 81 - 06 , geneva 1981 .
solmitz , _ annu .
* 14 * ( 1964 ) 375 ; + j. bystricky et al . , _
nuovo cim . _ * a1 * ( 1971 ) 601 .
r. brun , _ geant user guide and reference manual _ , cern dd/78/2 , 1978 .
sobol , _ monte - carlo numerical methods _ , moscow , nauka , 1973 .
m. babou et al . , _ nucl .
methods _ * 160 * ( 1979 ) 1 .
nilov et al . , _ yadernaya fizika _ * 22 * ( 1975 ) 116 .
j. pit and m. roos , _ nucl .
* b6 * ( 1968 ) 325 .
blatt and v.f .
weiskopf , _ theoretical nuclear physics _ , new - york , wiley&sons , 1952
. m. svec , * hep - ph*/9707495
. m. svec , * hep - ph*/9608339
brabson et al . , proc . of hadron95 , manchester 1995 , world
scientific , 1996 , p. 494 .
d. alde et al . , _ z. phys .
_ * c66 * ( 1995 ) 375 ; + a.a .
kondashov et al . , proc .
of the 28@xmath81 international conference on high energy physics , warsaw , 1996 , p. 474 .
j. ellis and j. lanik , _ phys .
lett . _ * b150 * ( 1985 ) 249 . m.s .
shifman et al . , _ nucl .
phys . _ * b147 * ( 1979 ) 385 , 448 .
g. bali et al . , _ phys .
* b309 * ( 1993 ) 378 .
f. butler et al . , _ phys .
* 70 * ( 1993 ) 2849 ; + m. chen et al .
( suppl . )
* 34 * ( 1994 ) 357 .
k. ishikawa et al . , _ phys .
lett . _ * 116b * ( 1982 ) 429 ; + b. berg and a. billoire , _ nucl .
phys . _ * b221 * ( 1983 ) 109 ; + w.h .
hamber and m. urs mceiller , _ phys .
rev . _ * d29 * ( 1984 ) 928 .
close et al . , _ phys .
lett . _ * b319 * ( 1993 ) 291 . | we present the results of the experimental study of the reaction of dipion production by the beam of the negative pions with momentum 1.78 gev / c on the polarized proton and liquid hydrogen targets .
the experiment covers the region of dipion masses near the mass of @xmath0-meson and small momenta transferred @xmath1 ( gev / c)@xmath2 .
the whole set of spin density matrix elements was reconstructed and model - independent and model - dependent amplitude analyses of the reaction were performed .
the experiment allows to exclude the ambiguity of the amplitude analysis existing at high energies .
the results contain evidences in favour of existing of narrow @xmath3 @xmath4 .
the data also allows to estimate intercept of @xmath5-meson regge trajectory .
the experiment was performed at the itep proton synchrotron , moscow .
pacs number(s ) : 13.75.gx , 13.88.+e . rotate , , , , , , , , , , , amplitude analysis , polarized proton target , rho - meson , scalar mesons , spin density matrix elements , |
at this conference we heard of the spectacular progress that has occurred in observational cosmology in recent years .
we also learned about the upcoming missions that are poised to provide new data to further constrain or even rule out leading theoretical models .
these advances have been and continue to be the engines that drive contemporary cosmology .
they have brought to forefront the astonishing success of the friedmann , lematre , robertson , walker ( flrw ) models , and perturbations thereof .
indeed , it appears that the rich data that we now have , and are likely to accumulate in the near future , would be adequately described by these simple applications of general relativity and quantum field theory on the resulting cosmological backgrounds .
however these theories are conceptually incomplete .
they assume that the universe begins with a big bang at which matter densities and space - time curvature become infinite . with inflationary scenarios
there were initial hopes that perhaps the big - bang singularity could be avoided because the inflaton fails to satisfy the strong energy condition often used in the singularity theorems of general relativity . however , borde , guth and vilenkin @xcite soon established that this hope was misplaced .
inflation can be eternal in the future but , if we go back in time using einstein s equations , one again finds that the space - time ends and physics simply comes to a halt at the big bang .
but all our experience with fundamental physics suggests that this can not be the situation in the real world .
this must be a prediction of a theory that has been pushed well beyond the domain of its validity . to know what _ really _ happened near the putative big bang
, we must work with a genuine unification of general relativity and quantum physics , an unification which
does not pre - suppose that space - time is a smooth continuum , and which can encompass the rich non - linear structure of strong gravity that lies beyond the scope of perturbation theory .
but the burden on this desired unification is heavy . in the context of cosmology
there is a long list of fundamental questions that must be satisfactorily addressed by such a theory .
here is an illustrative list encompassing some of the contemporary issues .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ @xmath0 if general relativity is transcended , how close to the putative big bang does a smooth space - time of einstein s make sense ?
in particular , can one show from first principles that this approximation is valid at the onset of inflation ?
+ @xmath0 is the big - bang singularity naturally resolved by the quantum version of einstein s equations ? or , is some external input such as a new principle or a boundary condition at the big bang essential ?
an outstanding example of such an external input is the hartle - hawking proposal @xcite .
+ @xmath0 is the quantum evolution across the ` singularity ' deterministic ?
one needs a fully non - perturbative framework to answer this question in the affirmative . in the pre - big - bang @xcite and ekpyrotic / cyclic @xcite scenarios , for example , so far the answer is in the negative because these theories pre - suppose the space - time continuum of general relativity and this approximation fails at the big - bang .
+ @xmath0 if the singularity is resolved , what is on the ` other side ' ?
is there just a ` quantum foam ' , far removed from any classical space - time , or , is there another large , classical universe ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ such questions are fundamental and must be faced squarely because the resolution of classical singularities can profoundly shift the paradigm underlying contemporary cosmology .
this in itself makes it imperative that we understand the quantum nature of the big bang .
but there could also be another rich pay - off : the new paradigm may enable us to address open issues that are observationally significant .
for example , if there is a classical pre - big - bang branch to the universe , the horizon problem would disappear and the observed large scale homogeneity could be simply a consequence of the fact that even the most distant parts of the universe would have been in causal contact in the past . if there is a pre - big - bang branch , we would not have to specify the initial conditions for perturbations on the singularity , where the applicability of current theories is least reliable .
there would be more natural ways of specifying these conditions which , in turn , may well lead to small , potentially observable deviations from current predictions .
finally , our experience with general relativity itself suggests that , once the physics of the planck regime is well understood , we may be handed with novel predictions of central importance to the next generation of astrophysicists and cosmologists . loop quantum gravity ( lqg ) @xcite is well suited to embark on this mission because it does not pre - suppose a classical space - time it is background independent . at a fundamental level , everything ,
_ including geometry , _ is described in the paradigm of quantum physics @xcite .
classical space - times emerge only on coarse graining of semi - classical quantum states . finally , since the approach is fully non - perturbative , it is well suited for the strong field regime near the putative big bang . in this chapter
i will discuss loop quantum cosmology ( lqc ) , the application of the principles of lqg to cosmology @xcite .
initial ideas appeared in @xcite and were developed in detail for a variety of cosmological models in @xcite .
_ in all cases , the big - bang and big - crunch singularities are resolved in a direct physical sense . _
the resulting planck scale physics has been explored using analytical and numerical solutions to the quantum einstein equations as well as effective equations which capture the leading quantum corrections .
singularity theorems are avoided _ not _ because one uses matter violating energy conditions .
indeed , in the models that have been studied in most detail , _ all energy conditions hold_. the theorems are inapplicable because quantum geometry effects modify einstein equations themselves .
physically , quantum geometry gives rise to a new repulsive force .
this force is _ utterly _ negligible under normal circumstances .
it is only when the matter density becomes about 1% of the planck density or curvature approaches @xmath1 that the repulsive force and hence the deviation from classical general relativity becomes significant .
but then the repulsive force rises _ very _ quickly , overwhelms classical attraction and causes a quantum bounce .
the density and curvature start falling and once they are below the scales just mentioned , the force again becomes negligible and classical general relativity again becomes an excellent approximation . immediately to the future of the bounce there is a robust phase of super - inflation which is not encountered in general relativity @xcite .
but it is short lived and in absence of a suitable inflaton potential it does not yield a sufficient number of e - foldings .
however , in presence of suitable potentials such as @xmath2 super - inflation _ funnels the phase space trajectories to initial conditions which virtually guarantee a slow roll inflation with 60 or more e - foldings _ @xcite .
this is in striking contrast to what happens in general relativity where it has been argued @xcite that the probability of @xmath3 e - folding decreases as @xmath4 .
the super - inflationary phase is also likely to have other phenomenological consequences such as production of gravitational waves that are being analyzed .
the article is organized as follows .
section [ s2 ] lays out the conceptual setting and section [ s3 ] provides a bird s eye view of lqc through illustrative results .
we will see that not only has lqc answered many of the long standing questions but it has also opened new vistas .
section [ s4 ] summarizes the origin of the novel predictions and places them in a broader perspective .
@xmath5@xmath6 plane , where @xmath7 denotes the volume and @xmath8 the scalar field . following the convention in general relativity ,
the ( emergent ) time variable @xmath8 is plotted along the y - axis .
@xmath5 k=0 trajectories .
@xmath6 a k=1 trajectory . in the k=0 case
, @xmath9 is the physical volume of a fiducial cell / box ; @xmath10 where @xmath11 is the scale factor .
the final physical results are of course insensitive to the choice of this cell.,title="fig:",width=307 ] plane , where @xmath7 denotes the volume and @xmath8 the scalar field . following the convention in general relativity , the ( emergent ) time variable @xmath8 is plotted along the y - axis .
@xmath5 k=0 trajectories .
@xmath6 a k=1 trajectory . in the k=0 case
, @xmath9 is the physical volume of a fiducial cell / box ; @xmath10 where @xmath11 is the scale factor .
the final physical results are of course insensitive to the choice of this cell.,title="fig:",width=307 ] to set the stage , let me use the simplest cosmological models : flrw space - times with a massless scalar field .
these models are instructive because in classical general relativity all their solutions have a big - bang ( and/or big - crunch ) singularity .
therefore , a quantum resolution of these singularities is non - trivial . furthermore , it is not difficult to incorporate potentials , additional matter fields and anisotropies .
figure [ class ] illustrates classical dynamics for k=0 and k=1 models without a cosmological constant .
@xmath8 is the massless scalar field while @xmath10 denotes the _ physical _ volume of the universe in the k=1 case and of a fixed fiducial cell in the k=0 case . if k=0 there are two classes of trajectories . in one
the universe begins with a big - bang and expands continuously , while in the other , it contracts continuously into a big crunch . in the k=1 case ,
the universe begins with a big bang , expands to a maximum volume and then undergoes a recollapse to a big crunch singularity .
now , in quantum gravity , one does not have a single space - time in the background but rather a probability amplitude for various space - time geometries .
therefore , unlike in the classical theory , one can not readily use , e.g. , the proper time along a family of preferred observers as a clock .
however , along each dynamical trajectory , the massless scalar field @xmath8 is monotonic .
therefore , it serves as a good clock or ` emergent time ' with respect to which the physical degrees of freedom the matter density or the volume , anisotropies , curvature scalars and other matter fields , if any evolve .
note incidentally that , in the k=1 case , because of the classical recollapse , volume ( or , the scale factor ) is double - valued along any dynamical trajectory .
therefore it can not serve as a global clock variable , while the scalar field does fulfill this role .
if one has a full quantum theory , one can proceed as follows .
choose a classical solution , i.e. a dynamical trajectory in the @xmath12 plane .
the momentum @xmath13 conjugate to @xmath8 is a constant of motion .
let us suppose its value on our trajectory is @xmath14 .
next , choose a point @xmath15 on the trajectory where the matter density and space - time curvature are low .
this point describes the state of the flrw universe at a late time when general relativity is expected to be valid . at the ` time ' @xmath16 construct a wave packet which is sharply peaked at @xmath17 and @xmath18 and evolve it backward and forward in ( the scalar field ) time .
we are then led to two questions : \i ) _ the infrared issue : _ does the wave packet remain peaked on the classical trajectory in the low curvature regime ? or ,
do quantum geometry effects accumulate over the cosmological time scales , causing noticeable deviations from classical general relativity ?
in particular , in the k=1 case , is there a recollapse and if so for large universes does the value @xmath19 of maximum volume agree with that predicted by general relativity @xcite ?
\ii ) _ the ultraviolet issue : _ what is the behavior of the quantum state when the curvature grows and enters the planck regime ? is the big - bang singularity resolved without any extra input ? or , do we need to supplement dynamics with a new principle , such as the hartle - hawking ` no boundary proposal ' @xcite ?
what about the big - crunch ? by their very construction ,
perturbative and effective descriptions have no problem with the first requirement .
however , physically their implications can not be trusted at the planck scale and mathematically they generally fail to provide a deterministic evolution across the putative singularity . since the non - perturbative approaches often start from deeper ideas , it is conceivable that they could lead to new structures at the planck scale which modify the classical dynamics and resolve the big - bang singularity . but once unleashed , do these new quantum effects naturally ` turn - off ' sufficiently fast , away from the planck regime ?
the universe has had some _ 14 billion years _ to evolve since the putative big bang and even minutest quantum corrections could accumulate over this huge time period leading to observable departures from dynamics predicted by general relativity .
thus , the challenge to quantum gravity theories is to first create huge quantum effects that are capable of overwhelming the extreme gravitational attraction produced by matter densities of some @xmath20 near the big bang , and then switching them off with extreme rapidity as the matter density falls below this planck scale .
this is a huge burden !
the question then is : how do various approaches fare with respect to these questions ?
the older quantum cosmology the wheeler - dewitt ( wdw ) theory passes the infra - red test with flying colors .
but unfortunately the state follows the classical trajectory into the big bang ( and in the k=1 case also the big crunch ) singularity .
the singularity is not resolved because _ expectation values of density and curvature continue to diverge _ in epochs when their classical counterparts do @xcite . for a number of years
, the failure of the theory to naturally resolve the big bang singularity was taken to mean that quantum cosmology can not , by itself , shed significant light on the quantum nature of the big bang . indeed , for systems with a finite number of degrees of freedom we have the von neumann uniqueness theorem which guarantees that quantum kinematics is unique .
the only freedom we have is in factor ordering and this was deemed insufficient to alter the status - quo provided by the theory .
the situation changed dramatically in lqg .
in contrast to the theory , a well established , rigorous kinematical framework _ is _ available in full lqg @xcite .
furthermore , this framework is uniquely singled out by the requirement of diffeomorphism invariance ( or background independence ) @xcite .
if one mimics it in symmetry reduced models , one finds that a key assumption of the von - neumann theorem is violated . as a result ,
_ one is led to new quantum mechanics _
@xcite ! this quantum theory is inequivalent to the theory already at the kinematic level .
quantum dynamics built in this new arena agrees with the theory in ` tame ' situations but differs dramatically in the planck regime , leading to a natural resolution of the big bang and the big crunch singularities .
this section is divided into three parts . in
the first i consider the flrw models without a cosmological constant in some detail , in the second i summarize the more general situation , and in the third i discuss the phenomenon of super inflation and its surprising implications for inflationary scenarios .
the main lqc results can be summarized as follows @xcite .
@xmath5@xmath6 the k=0 case . in the backward evolution ,
i.e. , as @xmath8 decreases , the wave function follows our post big - bang branch at low densities and curvatures but undergoes a quantum bounce at matter density @xmath21 and joins on to the classical trajectory that was contracting to the future . @xmath6 the k=1 case .
the quantum bounce occurs again at @xmath21 . since the big bang and the big crunch are replaced by quantum bounces and the classical re - collapse survives , the evolution undergoes cycles . in these simulations , @xmath22 , @xmath23 , and @xmath24.,title="fig:",width=307 ] the k=0 case . in the backward evolution ,
i.e. , as @xmath8 decreases , the wave function follows our post big - bang branch at low densities and curvatures but undergoes a quantum bounce at matter density @xmath21 and joins on to the classical trajectory that was contracting to the future . @xmath6 the k=1 case .
the quantum bounce occurs again at @xmath21 . since the big bang and the big crunch
are replaced by quantum bounces and the classical re - collapse survives , the evolution undergoes cycles . in these simulations , @xmath22 , @xmath23 , and @xmath24.,title="fig:",width=307 ]
let us first consider the k=0 model without a cosmological constant .
following the strategy outlined in section [ s2 ] , let us fix a point at a late time on the trajectory corresponding to an expanding classical universe , construct a gaussian wave function which is sharply peaked at that point , and evolve it using the lqc hamiltonian constraint .
one then finds the following @xcite .
the wave packet remains sharply peaked on the classical trajectory so long as the matter density @xmath25 remains below 1% of the planck density @xmath26 .
thus , as in the theory , the lqc evolution meets the infra - red challenge successfully .
let us evolve the quantum state back in time , toward the singularity
. in the classical solution scalar curvature and the matter energy density keep increasing and eventually diverge at the big bang .
the situation is very different in lqc . as mentioned in section [ s1 ] ,
once the density and curvature enter the planck scale _ quantum geometry effects become dominant creating an effective repulsive force which rises very quickly , overwhelms the classical gravitational attraction , and causes a bounce thereby resolving the big bang singularity .
_ ( see fig [ fig : lqc1 ] . )
numerical simulations show that the density acquires its maximum value @xmath27 at the bounce point .
although in the planck regime the peak of the wave function deviates very substantially from the general relativistic trajectory of figure [ class ] , it follows an effective trajectory with very small fluctuations .
this effective trajectory was derived using techniques from geometric quantum mechanics @xcite .
the effective equations it satisfies incorporate the leading corrections from quantum geometry which modify the left hand side of einstein s equations .
however , to facilitate comparison with the standard form of einstein s equations , one moves this correction to the right side through an algebraic manipulation .
then , one finds that the friedmann equation [ fe ] ( ) ^2 = . of classical general relativity
is replaced by [ lqc - fe ] ( ) ^2 = ( 1 - ) . here , as usual ` dot ' refers to the derivative with respect to _ proper time _ and @xmath28 , where @xmath29 is the barbero - immirzi parameter of lqg ( whose value @xmath30 is determined by the black hole entropy calculations in lqg ) .
by plugging in numbers one finds @xmath31 .
thus , @xmath32 , found in numerical simulations .
furthermore , one can show analytically @xcite that the eigenvalues of the density operator on the _ physical _ quantum states are bounded above by @xmath33 , also given by @xmath34 .
thus , _ there is an excellent match between the quantum theory which provides _
@xcite , _ the effective equations which provide _
@xmath35 @xcite _ and numerical simulations which provide @xmath36 _ @xcite .
in classical general relativity the right side , @xmath37 , of the friedmann equation is positive , whence @xmath38 can not vanish ; the universe either expands forever from the big bang or contracts forever ending in the big crunch . in the lqc effective equation ,
on the other hand , @xmath38 vanishes when @xmath39 at which a quantum bounce occurs : to the past of this event , the universe contracts while to the future , it expands .
this is possible because the lqc correction @xmath40 _ naturally _ comes with a negative sign .
this is non - trivial . in the standard brane world scenario , for example
, friedmann equation also receives a @xmath40 correction but it comes with a positive sign ( unless one artificially makes the brane tension negative ) whence the singularity is not resolved . consider the standard inflationary scenario for the @xmath2 potential with phenomenologically determined values of @xmath41 .
then the standard initial conditions at the onset of inflation are such that the quantum correction @xmath40 is of the order @xmath42 , and hence completely negligible .
thus , lqc calculations provide _ an a priori justification _ for using classical general relativity during inflation .
+ in the closed , k=1 model , the situation is similar but there are two additional noteworthy features @xcite . although they are not important from phenomenological considerations , they reveal surprising properties of the domain of applicability of classical general relativity . to start with , classical general relativity is again an excellent approximation to the lqc evolution till matter density @xmath25 becomes about 1% of the planck density @xmath26 but , as the density increases further , the two evolutions start diverging rapidly . again
, quantum geometry effects become dominant creating an effective repulsive force which rises very quickly , overwhelms the classical gravitational attraction , and causes a bounce thereby resolving both the big bang and the big crunch singularities .
surprisingly these considerations apply _ even to universes whose maximum radius @xmath43 is only _ @xmath44 . for these universes ,
general relativity is a very good approximation in the range @xmath45 !
the matter density acquires its minimum value @xmath46 at the recollapse .
the classical prediction @xmath47 is correct to one part in @xmath48 .
it is rather astonishing that general relativity is so accurate even in situations where one would have expected quantum corrections to dominate .
on the other hand there are also situations in which one would have at first expected general relativity valid , where quantum corrections dominate !
recall that the volume of the closed universe acquires its minimum value @xmath49 at the quantum bounce .
@xmath49 scales linearly with @xmath50 .
consequently , @xmath49 can be _ much _ larger than the planck size .
consider for example a quantum state describing a universe which attains a maximum radius of a megaparsec .
then the quantum bounce occurs when the volume reaches the value @xmath51 , _ some @xmath52 times the planck volume .
_ a more realistic universe would have a much larger maximum radius and hence @xmath50 and then @xmath49 would be much lager !
although it is at first surprising that quantum effects can dominate when the universe is so large , a moment s reflection shows that deviations from the classical behavior are triggered when the density or curvature reaches the planck scale .
the size of the volume by itself is not the relevant factor for quantum gravity effects .
+ thus , in the simplest k=0 and k=1 flrw models , classical singularities are replaced by quantum bounces and lqc provides a rather detailed picture of the physics in the planck regime .
the situation is essentially the same in the open , k=-1 model @xcite .
( although the detailed analysis in @xcite has certain drawbacks , they can be overcome using techniques introduced in @xcite . )
furthermore , the singularity resolution does not cause infra - red problems : there is close agreement with classical general relativity away from the planck scale .
the ultraviolet - infrared tension is avoided because , although quantum geometry effects are truly enormous in the planck regime , _ they die extremely quickly .
_ in this sub - section i will discuss the fate of singularities and planck scale physics in models with increasing generality .
_ beyond the big bang and big crunch singularities : _ in section [ s3.1 ] i focused on the big bang and big crunch singularities where the scale factor goes to zero and the matter density and curvature diverges at a finite proper time . but for general matter , new types of singularities can arise even with k=0 and @xmath53 .
for example , in the ` big - rip ' singularity , the scale factor _ diverges _ in a finite proper time along with energy density and pressure .
another possibility is the ` sudden death ' singularity where the curvature and/or its derivatives diverge at a _
finite _ values of the scale factor ( and proper time ) . in the spatially homogeneous , isotropic context ,
if matter has equation of state @xmath54 , there is a good control on the type of singularities that can arise @xcite .
from a physical perspective , we can divide them into two broad categories : strong singularities ( for example , the big bang and the big crunch ) where observers would be crushed and tame , weak ones ( such as those resulting from shell - crossing ) through which they would be able to propagate .
all possible occurrences have been recently analyzed using effective equations of lqc @xcite .
the result is that all strong singularities are resolved but not necessarily the weak ones .
since physics does not stop at the weak singularities , they are widely regarded as ` harmless ' .
an interesting situation arises in the context of a phantom model in which general relativity predicts the occurrence of a big - rip singularity .
if one uses the effective equations of lqc , the energy density remains bounded but the pressure and the rate of change of the hubble parameter diverge @xcite . at first this
might seem as a problem for lqc .
however , a closer examination @xcite shows that this is a weak singularity beyond which geodesics can be extended .
thus , in this case the quantum geometry effects tame a strong singularity and render it harmless .
@xmath5@xmath6 on the y - axis
. @xmath5 negative cosmological constant . in lqc ,
the big bang and big crunch singularities are resolved and the universe also undergoes a classical re - collapse .
therefore , the qualitative behavior is similar to that in k=1 , @xmath53 model of fig .
@xmath6 inflation with @xmath2 potential .
lqc reproduces the general relativistic inflation captured by the long , horizontal part of the curve that slopes down .
however , if we evolve backward in time using lqc , one obtains the near vertical segment on the left showing that the big - bang is replaced by a quantum bounce . near the bounce
, the potential plays a negligible role in the lqc dynamics.,title="fig:",width=307 ] on the y - axis .
@xmath5 negative cosmological constant . in lqc ,
the big bang and big crunch singularities are resolved and the universe also undergoes a classical re - collapse .
therefore , the qualitative behavior is similar to that in k=1 , @xmath53 model of fig .
@xmath6 inflation with @xmath2 potential .
lqc reproduces the general relativistic inflation captured by the long , horizontal part of the curve that slopes down .
however , if we evolve backward in time using lqc , one obtains the near vertical segment on the left showing that the big - bang is replaced by a quantum bounce . near the bounce
, the potential plays a negligible role in the lqc dynamics.,title="fig:",width=307 ] _ negative cosmological constant : _ the k=0 , @xmath55 model is discussed in @xcite . in this case , in classical general relativity
the situation is analogous to that in the k=1 model , discussed above .
the universe starts out with a big bang , expands to a maximum value of the scale factor , at which the positive energy density of matter is exactly balanced by the negative energy density in the cosmological constant .
if the universe were to expand further , the total energy density would become negative .
but since the friedmann equation ( [ fe ] ) of classical general relativity does not allow the total energy density to become negative , the universe recollapses . in lqc
the model has been analyzed in detail both by solving the quantum einstein s equations numerically and by analyzing the solutions of effective equations .
again , for states which are sharply peaked gaussians at late times , the effective equations provide an accurate account of full quantum dynamics , including the planck regime .
the quantum corrected friedmann equation again has the form ( [ lqc - fe ] ) where @xmath25 now includes also the energy density in the cosmological constant ; the big bang and the big crunch are replaced by quantum bounces .
the qualitative picture is thus similar to that in the k=1 , @xmath53 models .
the evolution is ( approximately ) cyclic .
the total density is maximum ( @xmath56 ) at the quantum bounce , the universe then expands , and the total density decreases .
when it vanishes , there is a classical re - collapse , the universe starts contracting , and the total density increases till it reaches the maximum value , @xmath35 when there is again a quantum bounce .
the value of @xmath35 is the same as in the @xmath53 case .
( see fig 3.a ) _ positive cosmological constant : _ in classical general relativity , the situation in the k=0 , @xmath57 case is similar to that in the k=0 , @xmath53 case , discussed in section [ s3.1 ] .
but the classical trajectory which starts out at the big - bang now expands to an infinite volume ( so that the energy density @xmath58 in the scalar field goes to zero ) at a _
value @xmath59 of the emergent time @xmath8 .
this opens up an interesting possibility in lqc . because the evolution in the emergent time @xmath8 is unitary in lqc
, one can continue it beyond the point at which the density goes to zero .
states which are semi - classical in the low @xmath58 regime again follow effective trajectories which now naturally extend beyond @xmath59 . since @xmath58 remains bounded , it is convenient to draw these trajectories in the @xmath58-@xmath8 plane .
they agree with the classical trajectories in the low @xmath58 regime and the extension is just the analytical continuation of classical trajectories
. it would be very interesting to work out physical ramifications of this unforeseen feature of lqc dynamics .
finally , in this case , lqc predicts that the cosmological constant has a maximum possible value . as one would expect the bound is given by the planck scale .
therefore it is not of phenomenological interest .
however , it is conceptually interesting that quantum geometry effects imply that @xmath60 can not even in principle be arbitrarily large . _
inflationary scenarios : _ the lqc framework has also been extended to incorporate an inflaton .
quantum einstein s equations have been numerically solved for the @xmath2 potential @xcite .
lqc admits wave functions that are sharply peaked at a standard inflating trajectory .
this is just as one would expect because in the standard scenario inflation occurs well away from the planck regime where general relativity is an excellent approximation to lqc .
when such a wave function is evolved _
backward _ in time , one again finds that , rather than following the classical trajectory into the big bang , the wave function bounces ( see fig .
3.b ) ) . again
, effective equations provide an excellent approximation to the full quantum equations .
numerical simulations as well as the effective equations show that the potential has negligible effect near the big bounce .
all the principal features of the lqc evolution reported above are recovered , including the value of @xmath35 .
_ anisotropies and gravitational waves : _ inclusion of anisotropies leads to bianchi models . in these models , the weyl curvature is not zero because , physically , they admit gravitational waves . detailed lqc analysis has been carried out in bianchi i and ii models at the analytical level @xcite ( and the method extends to all class a bianchi models ; bianchi ix , for example , is discussed in @xcite ) .
it establishes that the big bang singularity is resolved .
effective equations again show that the matter density @xmath61 is bounded above .
however , as one might expect , the bound now is lower than @xmath35 because there is also some energy in the gravitational waves that is not captured in matter density . in the isotropic models ,
space - time curvature is completely encoded in the ricci scalar . in the anisotropic case
, we also have non - trivial weyl curvature .
consequently , there is not just one bounce .
rather , any time a curvature invariant ( more precisely , shear ) enters the planck regime , quantum geometry effects intervene and dilute it , thereby avoiding the singularity that would have occurred had one used general relativity .
thus , the planck regime is much richer .
however , outside this regime , general relativity is again an excellent approximation to lqc . finally , as in the isotropic case
, matter can be chosen to satisfy all energy conditions .
the singularity theorems are transcended because einstein s equations are modified just in the right fashion by lqc . _ beyond homogeneity : _ in mathematical general relativity , the so called gowdy cosmological models have drawn much attention because they admit _ local , inhomogeneous degrees of freedom _ and yet are sufficiently simple to be tractable .
these models have been studied in detail using lqc @xcite .
one can construct a quantum theory of gowdy models using conventional field theoretical methods which do not take into account quantum geometry of lqc .
one then finds that the singularity is not resolved . on the other hand , using lqc to treat just the homogeneous degrees of freedom already suffices to resolve singularities , even if the inhomogeneous modes are treated using more conventional techniques rooted in fock quantization .
the general expectation is that the ultra - violet behavior would further improve if all modes are treated using loop quantum gravity .
but already this ` hybrid ' scheme introduced by the madrid group has given some confidence that _ the singularity resolution is not tied to the homogeneous models .
_ _ general space - like singularities : _ in general relativity there is a conjecture due to belinskii , lifshitz and khalatnikov ( bkl ) which says that as one approaches space - like singularities in general relativity , ` spatial derivatives of basic fields become sub - dominant relative to the time derivatives ' and dynamics at any spatial point is well approximated by that of homogeneous models @xcite .
bianchi i dynamics plays a dominant role and from time to time there are transitions from one bianchi i solution to another , mediated by a bianchi ii solution . by now
there is considerable support for this conjecture both from rigorous mathematical and numerical investigations @xcite .
this , in turn , provides support for the hope that the lessons on the quantum nature of singularities we learned from the bianchi i and bianchi ii models may be valid much more generally .
in particular , it suggests that the quantum geometry effects of lqg may well resolve _ generic _ space - like , strong curvature singularities of classical general relativity .
these are precisely the singularities of direct interest to cosmology .
there is now a formulation of the bkl conjecture in the framework that underlies lqg @xcite .
together with the quantization procedure introduced in the bianchi ii models @xcite which , as remarked above , is directly applicable to a wide class of homogeneous models this form of the bkl conjecture should enable one to obtain concrete results for general space - like singularities .
so far , i have focused on singularity resolution and the new physics in the planck regime .
it turns out that lqc has unforeseen implications also to the inflationary scenarios and phenomena that occur away from the planck regime .
because of space limitation , i will outline only one of these and mention a few other avenues that are being pursued both by cosmology and lqc communities . in this sub - section , i will list some basic predictions of lqc which are rather surprising .
these new features arise in the planck regime but are important for setting the initial conditions for slow roll inflation at much lower energies . for simplicity and definiteness , in the detailed discussion
i use the k=0 , @xmath53 isotropic cosmologies and assume that the kinetic energy of the inflaton is positive .
as we saw , lqc is generally formulated using volume @xmath7 of a fiducial cell ( rather than the scale factor @xmath11 ) , and its conjugate momentum @xmath62 @xcite . on solutions to einstein s equations ,
@xmath63 where , as before , @xmath64 is the barbero - immirzi parameter of lqc and @xmath65 is the hubble parameter .
however , lqc modifies the einstein dynamics and on solutions to the effective equations we have [ h ] h = 2b 2b where @xmath66 is the ` area - gap ' , the smallest non - zero eigenvalue of the area operator . in lqc @xmath62
ranges over @xmath67 and general relativity is recovered in the limit @xmath68 .
as i already mentioned , quantum geometry effects modify the geometric , left side of einstein s equations .
in particular , the lqc friedmann equation first emerges as [ lqc - fe2 ] = ( + v ( ) ) .to compare with the standard friedmann equation ( [ fe ] ) , it is often convenient to rewrite it in the form ( [ lqc - fe ] ) using ( [ h ] ) . by inspection
it is clear from eqs ( [ lqc - fe ] ) , ( [ h ] ) and ( [ lqc - fe2 ] ) that , away from the planck regime i.e .
, when @xmath69 or , @xmath70 we recover classical general relativity .
modifications in the planck regime are drastic . recall that in general relativity , the hubble parameter @xmath71 is large throughout the planck regime and diverges at the singularity .
by contrast , in lqc @xmath71 _ vanishes _ at the bounce ( because @xmath72 there ) and eq.([h ] ) implies that it has a _
finite _ , maximum value , @xmath73 .
second , eq .
( [ lqc - fe2 ] ) implies that the density @xmath25 is bounded by @xmath74 .
third , if the potential @xmath75 is bounded below , say @xmath76 , then it follows from ( [ lqc - fe2 ] ) that @xmath77 is bounded by @xmath78 .
fourth , if the potential grows unboundedly for large @xmath79 , then @xmath79 is also bounded .
for example , for @xmath80 , we have @xmath81 .
finally , the derivative with respect to proper time of @xmath62 and the hubble parameter are now given by [ dot ] = - 4 g ^2 = -4 g ( 2b ) ^2 as a consequence , @xmath62 _ decreases monotonically in every solution _ , starting from @xmath82 and ending with @xmath83 ( the bounce occurs at the mid - point , @xmath84 ) .
when the potential is bounded below , @xmath85 is also bounded by @xmath86 .
thus , a large number of physical quantities which are unbounded in general relativity can not exceed certain finite , maximum values in lqc .
in particular , these results imply that if the potential is bounded below , the matter density and curvature can not diverge anywhere on lqc space - times .
one can also show that a solution where our fixed fiducial cell has a finite volume initially can not evolve to a configuration where the volume becomes zero .
thus , _ the lqc solutions are everywhere regular irrespective of whether one focuses on matter density , curvature or the scale factor .
_ finally , every solution undergoes precisely one bounce at @xmath87 where the hubble parameter vanishes because of eq .
( [ h ] ) and the density reaches its maximum value @xmath35 because of eq .
( [ lqc - fe ] ) .
let me now turn to _ super - inflation _
( see @xcite and especially @xcite ) . at the bounce point
, we have @xmath88 whence ( [ h ] ) implies that @xmath89 and it follows from ( [ dot ] ) that @xmath90 is positive .
the universe expands to the future of the bounce point and contracts in its past .
eq ( [ dot ] ) shows that @xmath90 continues to be positive till @xmath62 has decreased to @xmath91 at which point it vanishes ( and then it becomes and remains negative as @xmath9 increases monotonically to infinity ) .
thus , _ every lqc solution has a super - inflationary phase from _
@xmath87 to @xmath92 .
eqs ( [ h ] ) and ( [ dot ] ) imply that at the beginning of this phase , the hubble parameter @xmath71 vanishes and @xmath90 is very large , while at the end of this phase @xmath71 assumes its maximum value and @xmath90 vanishes .
note that the occurrence of this phase is universal ; it exists even if the inflaton potential is zero ! _
this is a robust feature of lqc that has no analog in general relativity . _
it is natural to ask if this quantum geometry driven super - inflation could be a substitute for the standard inflation .
if so , we would not have to invoke _ any _ potential !
unfortunately the answer is in the negative .
detailed considerations show that in absence of a potential , this phase is very short lived in proper time and therefore does not yield enough e - foldings to be a viable substitute for the standard inflation .
however , it was recently realized that in presence of the ` standard ' potentials used in inflation , this phase has an unforeseen consequence : it naturally funnels dynamical trajectories to initial conditions that _ lead to a long slow roll inflation with at least 68 e - foldings _ @xcite .
inflationary scenarios have had an extraordinary success especially in accounting for structure formation .
this success brings added urgency to a long standing question : does a sufficiently long , slow roll inflation require fine tuning of initial conditions or does it occur generically in a given theoretical paradigm ? ( see e.g. @xcite ) .
such a slow roll requires that initially the inflaton must be correspondingly high - up in the potential .
how did it get there ?
is it essential to invoke some rare quantum fluctuations to account for the required initial conditions because the a priori probability for their occurrence is low ? or , is a sufficiently long , slow roll inflation robust in the sense that it is realized in ` almost all ' dynamical trajectories of the given theory ? to make these questions precise , one needs a stream - lined framework to calculate probabilities of various occurrences _ within a given theory_. a mathematically natural framework to carry out this analysis was introduced over two decades ago ( see , e.g. , @xcite ) .
it invokes laplace s principle of indifference @xcite to calculate the _ a priori _ probabilities for various occurrences .
more precisely , the idea is to use ( a flat probability distribution @xmath93 and ) the canonical liouville measure @xmath94 on the space @xmath95 of solutions @xmath96 of the theory under consideration to calculate the _ relative volumes _ in @xmath95 occupied by solutions with desired properties @xcite . in our case , then , the a priori probability is given by the _
liouville volume occupied by the sub - space of solutions in which a sufficiently long , slow roll inflation occurs .
further physical input can provide a sharper probability distribution @xmath97 and a more reliable likelihood than the ` bare ' a priori probability
. however , a priori probabilities can be directly useful if they are very low or very high . in these cases
, it would be an especially heavy burden on the fundamental theory to come up with the physical input that significantly alters them .
the task , then , is to calculate the _ a priori _ probability that there is inflation with at least 68 e - foldings . as in general relativity ,
this can be done by introducing the natural liouville measure on the space @xmath95 of solutions and calculating the fractional liouville volume of @xmath95 occupied by solutions with adequate number of e - foldings . in general relativity , it has been argued @xcite that the resulting a priori probability for obtaining 68 or more e - foldings is suppressed by a factor larger than @xmath98 .
what is the situation in lqc ?
now the equations of motion are [ ddot ] = v + 3h + v _ , = 0 where @xmath99 , and @xmath100 are , in addition , subject to the lqc - modified friedmann equation ( [ lqc - fe ] ) .
it is simplest to write the natural liouville measure on the space @xmath95 of solutions in terms of values @xmath101 of the scalar field and the volume at the bounce : _ = ^ _ v _ where @xmath102 is the fraction of the total density that is in the potential energy at the bounce .
furthermore , in lqc @xmath79 takes values in the _ bounded _
interval @xmath103 .
the question now is : with respect to this measure , what is the fractional volume in @xmath95 occupied by solutions with at least 68 e - folds of slow roll inflation ? recall that the hubble parameter @xmath71 takes its maximum value at the end of super - inflation and the friction term in the equation of motion for @xmath8 is given by @xmath104 .
the phenomenologically preferred value of @xmath41 is of the order @xmath105 .
therefore , at the end of super - inflation , the friction term is necessarily large and initial conditions are naturally set for a long slow roll . to calculate the probabilities ,
let us use the phenomenological value of @xmath41 .
one can perform detailed analytical calculations using usual approximations and check their validity via high precision numerical simulations .
one finds that 68 or more e - foldings are guaranteed if the fraction @xmath106 is larger than @xmath107 @xcite .
this in turn implies that the a priori probability of a slow roll with at least 68 e - foldings is greater than 0.99 .
this conclusion is quite robust : one can change the inflaton mass by a couple of orders of magnitude or add a quartic term to the potential ( with phenomenological bounds on the coupling constant ) .
the probability remains greater than 0.99 .
thus , even when one uses the same methods that were used for general relativity in @xcite , the conclusion in lqc is _ opposite _ of that in @xcite .
this comes about because , unlike in general relativity , @xmath108 are bounded in lqc ; all solutions are singularity free and undergo a quantum bounce ; and the bounce is followed by a period of super - inflation which funnels dynamical trajectories to initial conditions that are well suited for a long slow roll .
+ there are several other phenomenological implications of super inflation , and more generally , of the lqc quantum corrections to einstein s equations , particularly on scalar and tensor perturbations ( see e.g. @xcite ) .
i discussed some of these in my talk .
unfortunately , space limitation does not allow me to discuss these developments here .
singularities of general relativity are perhaps the most promising gates to physics beyond einstein .
they provide a fertile conceptual and technical ground in our search of a new paradigm in cosmology as well as fundamental physics .
consider some of the deepest conceptual questions we face today : the issue of the beginning and the end end , the arrow of time , and the puzzle of black hole information loss .
their resolutions hinge on the true nature of space - time singularities . in my view
, considerable amount of contemporary confusion about such questions arises from our explicit or implicit insistence that singularities of general relativity are true boundaries of space - time ; that we can trust causal structure all the way to these singularities ; that notions such as event horizons are absolute even though changes in the metric in a planck scale neighborhood of the singularity can move event horizons dramatically or even make them disappear altogether @xcite .
lqg is well suited to address the issue of the fate of classical singularities in quantum gravity because it is fully non - perturbative and does not pre - suppose that we have a smooth geometry all the way up to the big - bang and big crunch @xcite .
therefore lqc has been used to address many long standing cosmological questions in detail @xcite .
the scalar field serves as emergent time .
strong curvature singularities of classical general relativity are either resolved @xcite or converted to harmless weak singularities by the quantum geometry effects @xcite . in particular , the big bang and the big crunch
are naturally replaced by quantum bounces . on the ` other side ' of the bounce there
is again a large universe .
general relativity is an excellent approximation to quantum dynamics once the matter density falls below one percent of the planck density .
thus , lqc successfully meets both the ` ultra - violet ' and ` infra - red ' challenges .
furthermore results obtained in a number of models using distinct methods re - enforce one another .
one is therefore led to take at least the qualitative findings seriously : _
big bang is not the beginning nor the big crunch the end .
_ quantum space - times could be vastly larger than what general relativity had us believe ! how can the quantum space - times of lqc manage to be significantly larger than those in general relativity when those in the theory are not ?
main departures from the theory occur due to _ quantum geometry effects _ of lqg .
there is no fine tuning of initial conditions , nor a boundary condition at the singularity , postulated from outside .
furthermore , matter can satisfy all the standard energy conditions .
why then does the lqc singularity resolution not contradict the standard singularity theorems of penrose , hawking and others ?
these theorems are inapplicable because _ the left hand side _ of the classical einstein s equations is modified by the quantum geometry corrections of lqc .
what about the more recent singularity theorems that borde , guth and vilenkin @xcite proved in the context of inflation ?
they are not tied to einstein s equations .
but , motivated by the eternal inflationary scenario , they assume that the expansion remains positive if we recede in the past along any geodesic . because of the pre - big - bang contracting phase , this assumption is violated in the lqc effective theory .
a qualitative picture that emerges is that the non - perturbative quantum geometry corrections create a _ ` repulsive ' _ force . while this force is negligible under normal conditions
, it dominates when curvature approaches the planck scale and can halt the collapse that would classically have led to a singularity . in this respect , there is a curious similarity with the situation in the stellar collapse where a new repulsive force comes into play when the core approaches a critical density , halting further collapse and leading to stable white dwarfs and neutron stars .
this force , with its origin in the fermi - dirac statistics , is _ associated with the quantum nature of matter_. however , if the total mass of the star is larger than , say , @xmath109 solar masses , classical gravity overwhelms this force .
the new repulsive force of lqc is _ associated with the quantum nature of geometry_. it comes into play only near planck densities but is then so strong that it can counter the classical , gravitational attraction , irrespective of how large the collapsing mass is .
it is this force that prevents the formation of singularities .
at first one might think that , since quantum gravity effects concern only a tiny region , whatever they may be , their influence on the global properties of space - time should be negligible whence they would have almost no bearing on the issue of the beginning and the end . however , as we saw , once the singularity is resolved , vast new regions appear on the ` other side . '
new possibilities open up that were totally unforeseen in general relativity @xcite .
first , matter density , curvature scalars , and the hubble parameter _
are all bounded_. second , even in absence of a potential , there is a robust super - inflationary phase immediately after the quantum bounce @xcite . as we discussed in section [ s3.3 ]
this novel phase has interesting consequences .
in particular , it naturally leads to the phenomenon of _ hubble funneling _ : on every solution the hubble parameter is driven to its largest possible value at the end of super - inflation . as a consequence , if there is a potential such as @xmath2 , initial conditions are naturally set for a long slow roll inflation @xcite .
one does not have to appeal to rare quantum fluctuations to explain how the inflaton managed to get sufficiently high in the potential to seed a slow roll inflation leading to 68 or more e - foldings . finally , _
these conclusions are insensitive to what happened in the pre - big - bang branch . _
another major direction of current work concerns perturbations that seed structure formation . in the new paradigm provided by lqc
, one is led to re - examine the issue of initial conditions for these perturbations .
we no longer have to specify them at a singularity .
it is most natural to set them in the infinite past where matter density and curvatures tend to zero , whence it makes direct physical sense to speak of vacuum fluctuations .
the lqc paradigm leads us to re - asses the old questions from entirely new angles ( see , e.g. , @xcite ) .
this opens up a wealth of interesting challenges and lqc provides novel ideas and tools to meet them .
* acknowledgments : * i would like to thank alex corichi , tomasz pawlowski , param singh , david sloan , victor taveras , kevin vandersloot and edward wilson - ewing for collaboration and discussions . this work was supported in part by the nsf grant phy0854743 , the alexander von humboldt foundation , the george a. and margaret m. downsbrough endowment and the eberly research funds of penn state .
m. gasperini and g. veneziano , the pre - big bang scenario in string cosmology , phys . rept . * 373 * , 1 ( 2003 ) j. khoury , b. a. ovrut , p. j. steinhardt , n. turok , the ekpyrotic universe : colliding branes and the origin of the hot big bang , phys.rev .
d64 ( 2001 ) 123522 j. khoury , b. .
ovrut , n. seiberg , p. j. steinhardt , n. turok , from big crunch to big bang , phys.rev . *
d65 * 086007 ( 2002 ) a. ashtekar and j. lewandowski , background independent quantum gravity : a status report , class .
* 21 * , r53-r152 ( 2004 ) a. ashtekar , j. lewandowski , d. marolf , j. mouro and t. thiemann , quantization of diffeomorphism invariant theories of connections with local degrees of freedom .
phys . * 36 * , 64566493 ( 1995 ) a. ashtekar , t. pawlowski , p. singh and k. vandersloot , loop quantum cosmology of k=1 frw models , phys . rev . *
d75 * , 0240035 ( 2006 ) ; + l. szulc , w. kaminski , j. lewandowski , closed frw model in loop quantum cosmology , class .
* 24 * , 2621 - 2635 ( 2006 ) m. martin - benito , l. j. garay and g. a. mena marugan , hybrid quantum gowdy cosmology : combining loop and fock quantizations , _ _ phys . rev.__**d78 * * , 083516 ( 2008 ) + d. brizuela , g. mena - marugan and t. pawlowski , big bounce and inhomogeneities , ` arxiv:0902.0697 ` + g. mena - marugan and m. martin - benito , hybrid quantum cosmology : combining loop and fock quantization , intl .
a24 * , 2820 - 2838 ( 2009 ) v. a. belinskii , i. m. khalatnikov , and e. m. lifshitz , oscillatory approach to a singular point in relativistic cosmology , adv . phys .
* 19 * , 525 ( 1970 ) ; + a general solution of the einstein equations with a time singularity , adv
. phys . * 31 * , 639 ( 1982 ) e. j. copeland , d. j. mulryne , n. j. nunes , m. shaeri , super - inflation in loop quantum cosmology , phys . rev .
* d77 * , 023510 ( 2008 ) ; + the gravitational wave background from super - inflation in loop quantum cosmology , phys .
* d79 * , 023508 ( 2009 ) | this short review is addressed to cosmologists .
general relativity predicts that space - time comes to an end and physics comes to a halt at the big - bang .
recent developments in loop quantum cosmology have shown that these predictions can not be trusted .
quantum geometry effects can resolve singularities , thereby opening new vistas .
examples are : the big bang is replaced by a quantum bounce ; the ` horizon problem ' disappears ; immediately after the big bounce , there is a super - inflationary phase with its own phenomenological ramifications ; and , in presence of a standard inflaton potential , initial conditions are naturally set for a long , slow roll inflation independently of what happens in the pre - big bang branch . |
correlation is a quantum - mechanical patent with non - perturbative nature , which gives rise to diverse quantum phenomena .
starting from the basic , the notion of exchange correlation states that two _ independent _ fermions ( bosons ) experience an effective exchange repulsive ( attractive ) force when their wave functions highly overlap . in the case of interacting particles ,
correlation often leads to versatile orderings , for example ( anti-)ferromagneticsm , superconductorvity , and so on .
however , there remains insufficient understanding of what role the quantum correlation plays in the paramagnetic phase away from quantum criticality .
this question began to attract attentions after the behaviours of many transition - metal oxides were found outside the box of the fermi liquid theory .
they were soon classified as _ strongly - correlated _ materials .
the interest , with correlation still unknown , reaches its peak after the discovery of the @xmath0-based transition - metal oxides ( cuprates ) , so - called the high-@xmath1 superconductors @xcite .
the parent compounds of the high-@xmath1 superconductors is insulating antiferromagnets ( afm ) @xcite .
after chemical doping with charge carriers , the antiferromagnetic ordering disappears , followed by an enigmatic paramagnetic phase before the superconductivity emerges . the paramagnetic phase is now known as the pseudogap phase where a gap opens in the electronic spectrum without exhibiting any signature of conventional phase transition at @xmath2 , higher than the superconducting transition temperature @xmath1 @xcite .
similar phenomenon is also seen in heavy - fermion superconductors @xcite and iron pnictides @xcite , which share similar phase diagram to cuprates .
this ambiguous transition is often regarded as a crossover .
after almost three decades from its discovery , the central debates still focuses on the formation of the pseudogap and its relevancy to the superconductivity in the lower temperature range .
intensive studies have been done both experimentally and theoretically to identify whether or not the pseudogap phase belongs to a broken - symmetry state .
recently , experimental data in cuprates are accumulated to indicate that the pseudogap phase breaks time - reversal symmetry and preserves the translational symmetry @xcite . however , in most of those data , the time - reversal symmetry begins to fluctuate precursory to the pseudogap transition .
so far , the evidence is still vague that the pseudogap formation belongs to any symmetry - breaking scenario .
on the other hand , it is generally believed that the strong electron - electron repulsive interaction baptises the strong correlation .
this naive belief was first challenged by comanac et al .
@xcite . they found that a large hubbard @xmath3 value is not needed but the antiferromagnetic correlation is crucial to fit the experimental data of optical conductivity .
a recent numerical calculation also indicates that hubbard @xmath3 value decreases as the system size increases @xcite .
meanwhile , it has been advocated by laughlin that the coulomb interactions in the cuprates is simply the same as they are in elemental si or na metal @xcite .
however , to completely overrule the wrong belief , a mechanism responsible for pseudogap formation is needed in a framework of the _ weak - coupling _ theory . in this article
, we construct a weak - coupling theory for the pseudogap formation , where the gauge interaction weakly coupled to electrons acquires a mass leading to a gap - like structure in the electronic spectrum .
the non - perturbative mass acquisition mechanism identifies the quantum correlation , where the remanent antiferromagnetic fluctuation becomes the longitudinal mode of the gauge field .
moreover , the pseudogap transition is identified as a bkt-_like _ transition , and the transition temperature is computed .
most importantly , we provide a scheme for the pseudogap formation _ without _ breaking time - reversal and translational symmetry .
finally , a generalisation to the iron pnictides and heavy - fermion systems is briefly discussed .
in cuprates , mobile charge carriers are introduced in the mott insulator by chemical doping .
as the electrons become more and more mobile , the electron scattering process becomes more and more important .
the complete description of the scattering process should include the current - current ( cc ) interaction in the one - band hubbard model @xmath4 where @xmath5 is the electron creation ( annihilation ) operator , @xmath6 , and @xmath7 is the current operator @xmath8 when the scattering process occurs in the lattice level , the vertex may represent a copper site . even though the cc interaction can be small in most strongly - correlated materials , it is a relevant ingredient for the description of the non - perturbative effect , which the pseudogap formation will be demonstrated to be one later .
most importantly , the introduction of the cc interaction allows us to formulate a weak - coupling theory for high-@xmath1 materials , which was originally thought as a problem similar to qcd . taking the extended hubbard model to the continuous limit
, the theory can be written in the path integral formalism @xmath9 where @xmath10 , @xmath11 , and @xmath12 and @xmath13 are the charge and the current density respectively .
the theory described by eq .
( [ lagrangian1 ] ) is , however , not renormalizable since the dimension of the coupling constants @xmath14 and @xmath15 is @xmath16=[u_1]=l$ ] .
this situation is similar to fermi s @xmath17-decay theory , where a non - renormalizable _ four - fermion _ vertex is introduced to produce a finite matrix element between a neutron , a proton , an electron , and a neutrino @xcite .
the non - renormalizability hinders theorists from controlling the intrinsic infinity in the theory .
later , the @xmath17-decay problem is completely solved by glashow , weinberg , and salam @xcite , who constructed the standard model for the electroweak interaction , where @xmath18 and @xmath19 gauge bosons were introduced accounting for the weak interaction . here , we play the same trick . we introduce a fictitious gauge field @xmath20 accounting for the effective interaction between electrons . @xmath21[(\frac{\vec{\nabla}}{i}-g\vec{a})\psi_\sigma]\nonumber \\&-&ga_0\psi^\dag_\sigma\psi_\sigma\!-\!\frac{1}{4}f_{\mu\nu}f_{\mu\nu}\!-\!\frac{1}{2}m_1 ^ 2\vec{a}\cdot\vec{a}\!+\!\frac{1}{2}m_0 ^ 2a^2_0,\label{effective}\end{aligned}\ ] ] where @xmath22 is the field strength , @xmath23 is the coupling constant , @xmath24 and @xmath25 are the mass parameters for the @xmath26 and @xmath27 component respectively , and the metric @xmath28 is adopted .
since the theory is non - relativistic , there are two mass parameters for different interaction strengths in the charge and the current channels . integrating over @xmath26 and @xmath27 , eq .
( [ effective ] ) becomes @xmath29 where @xmath30 , @xmath31 is a small parameter , and terms with higher order of @xmath32 are ignored . using
the grassmann variable , terms with the same spin at the same spatial location automatically vanish . comparing eq .
( [ lagrangian1 ] ) and eq .
( [ lagrangian2 ] ) , @xmath14 and @xmath15 can be obtained @xmath33 in the low - energy and long - wavelength limit .
now , the coupling constant @xmath23 has a dimension @xmath34 $ ] .
the theory becomes renormalizable .
moreover , @xmath35 represents the repulsive interaction , and @xmath36 represents the attractive interaction .
it is consistent with our physical intuition of electromagnetism .
namely , the charge interaction between two electrons is repulsive due to the coulomb interaction .
the current interaction is attractive since two conducting wires with the same direction of electric current attract to each other due to the lorentz force .
similar situation of the current interaction between spinons was considered by lee _ _ e__t al . before @xcite .
different from other mechanism by exchanging vector bosons , for example phonons or magnons , the exchange of gauge boson produces both repulsive and attractive interactions .
mass of a gauge boson determines the interaction length scale since @xmath37 has the dimension of length .
however , the finiteness of mass makes eq .
( [ effective ] ) not gauge - invariant . as we know , the parent compound of cuprates is a 3d antiferromagnet ( afm ) @xcite .
in addition , the spin anisotropy favours the directions of spins to point in the cooper - oxygen plane @xcite . by the chemical doping
, 3d afm vanishes and antiferromagnetic ( af ) correlation becomes short - ranged and highly anisotropic in space @xcite . due to the spin anisotropy , the remnant af fluctuation can be described by a phase field @xmath38 .
now , as two electrons interact with each other by exchanging a gauge boson , in the path of the exchange , the gauge boson couples to the af fluctuation , which can be described by the following lagrangian @xmath39 where @xmath40 and @xmath41 are the covariant derivative , and @xmath42 is the gauge coupling for the phase field . the gauge transformation is defined by @xmath43 then , the total lagrangian is manifestly gauge invariant @xmath44[(\frac{\vec{\nabla}}{i}-g\vec{a})\psi_\sigma]]\nonumber \\&-&ga_0\psi^\dag_\sigma\psi_\sigma-\frac{1}{4}f_{\mu\nu}f_{\mu\nu}\nonumber \\ & + & \frac{1}{2}m_0 ^ 2(d_0 \phi)^\dag(d_0 \phi)-\frac{1}{2}m_1 ^ 2(d_i \phi)^\dag(d_i \phi).\label{fulllagrangian}\end{aligned}\ ] ] the mass acquisition of the gauge boson described here is known as the stckelberg mechanism @xcite , which is a common method in the mass generation in the context of the string theory @xcite .
the lagrangian in eq .
( [ fulllagrangian ] ) contains the correct description for the pseudogap transition . in the high temperature region ,
the gauge field is massless .
since the phase field is fluctuating , the vacuum expectation value @xmath45 .
the phase field is like a needle in a clock .
the needle has a finite length , but different directions cancel one another in the high temperature region . as the temperature goes down , the field tends to be static .
the vacuum expectation value of the field is frozen to @xmath46 @xcite .
now , we can simply gauge away the phase field by choosing @xmath47 . in this case ,
@xmath48 and @xmath49 .
the last two terms in eq .
( [ fulllagrangian ] ) soon become the last two terms in eq .
( [ effective ] ) , and the gauge bosons acquire a mass .
similar to the anderson - higgs mechanism , the gauge boson combines with the goldstone mode of the phase field and becomes massive .
in fact , the stckelberg mechanism was sometimes regarded as a special higgs mechanism with an infinite higgs mass .
the phase transition described above belongs to the 2d xy universality class @xcite .
therefore , a berezinski - kosterlitz - thouless ( bkt ) transition at finite temperature is expected @xcite .
the transition temperature is @xmath50 for the simplest case when @xmath51 .
the bkt transition is a phase transition of infinite order . in the high temperature phase ,
it is a disordered phase with an exponential correlation , and in the low temperature phase , it is a quasi - ordered phase with a power - law correlation . in this case
, the pseudogap is a bkt-_like _ transition . in the low temperature phase
, the af fluctuation becomes the longitudinal mode of the gauge boson . in the unitary gauge ,
the propagator of the @xmath52 field is simply zero .
namely , it does not appear in any physical process , nor can the power - law af correlation be observed .
let us now consider the symmetry property in the pseudogap state . since @xmath26 is even and @xmath27 is odd under time - reversal transformation , the phase field @xmath53 under time - reversal transformation . in the high temperature phase , @xmath54 ,
so it is time - reversal symmetric . in the low temperature phase , the phase field
can always be gauged away and becomes @xmath55 .
both time - reversal symmetry and translational symmetry is preserved in the pseudogap phase . across the pseudogap transition ,
the band structure is also modified . using the standard feynman diagram technique , we consider the diagrams up to @xmath56 as shown in ( fig .
2 ) for the computation of the electron self - energy .
we note that the dimensionless vertex @xmath57 in ( fig .
2a ) has the value @xmath58 in support of the current perturbative scheme . given the full green s function of electron @xmath59 , @xmath60 where @xmath61 and @xmath62 are the amplitudes for the diagrams in ( fig .
2a ) and ( fig .
2b ) respectively . after some algebra , @xmath61 = @xmath63 and re@xmath64 .
namely , the diagram in ( fig .
2a ) contributes to a finite energy gap and the one in ( fig .
2b ) does not .
we can do the same calculation for the hole in the valence band .
then , we obtain a finite energy gap @xmath65 which vanishes when @xmath66 .
therefore , at the pseudogap transition , the gauge field combines with the af fluctuation and acquires a mass , opening a finite energy gap in the electronic spectrum , which is the origin of the pseudogap formation .
the current theory is a zero - temperature calculation , and the pseudogap is isotropic in momentum space . in the real materials
, the pseudogap structure can be versatile due to the competition with the superconducting state with exotic pairing symmetry .
recently , with considerably - improved sample quality , a _ nodeless _ pseudogap structure was confirmed experimentally in the highly - underdoped la@xmath67sr@xmath68cuo@xmath69 ( lsco ) systems @xcite . while the pseudogap is nodeless in the low and zero temperature , the _ fermi - arc _ feature is restored and robust as either temperature or doping level increases @xcite .
the interesting temperature and doping dependence of the pseudogap structure might be due to the momentum - dependent hierarchy energy scales @xcite .
the pseudogap in cuprate is seemingly intrinsically nodeless , since even in la@xmath67ba@xmath68cuo@xmath69 ( @xmath70 ) , the failed high-@xmath1 superconductor , a gap is opened in the nodal direction when the system is approaching to zero temperature @xcite . while the robustness of the fermi arc at finite temperature remains intriguing , further studies with the consideration of the lattice symmetry and the @xmath71-wave superconducting instability should clarify this issue .
estimating the parameters in the theory , the fundamental properties in cuprates can be reproduced . taking @xmath72 ev , @xmath73 ev , and @xmath74 ev
, we obtain the hubbard term @xmath75 ev , the pseudogap magnitude @xmath76 @xmath77ev @xcite , and the pseudogap transition temperature @xmath78 k. the values of @xmath24 and @xmath25 imply the length scale @xmath79 , which explains why the on - site interaction is most relevant . as observed in the angle - resolved photoemission , the magnitude of the pseudogap and the pseudogap transition temperature has a roughly linear relation with doping @xcite . assuming that the hubbard-@xmath3 value is correlated with the magnitude of the pseudogap
, we can obtain the doping dependence of the @xmath80 , @xmath81 and @xmath82 phenomenologically as shown in ( fig .
while the interaction length scale is independent of doping , the electron - electron interaction decreases monotonically as the doping increases . at the critical doping @xmath83
, @xmath81 is zero and the system becomes fermi liquid .
interestingly , we found that the amplitude of the af fluctuation , @xmath84 , also vanishes at @xmath85 with @xmath86 provided that @xmath87 .
a quantum phase transition is implied as the onset of the af fluctuation .
since it is the end point of the finite - temperature pseudogap transition , the quantum phase transition belongs to the 2d xy universality class . in other words ,
the quantum phase transition is a continuous phase transition of infinite order .
( @xmath88 is assumed .
labelled in blue in ( a ) ) , the gauge coupling @xmath81 ( labelled in red in ( a ) ) , and the amplitude of the af fluctuation @xmath84 ( labelled in blue in ( b ) ) .
@xmath85 represents the critical doping where the quantum phase transition occurs .
the gradient green area represents the antiferromagnetic phase .
the gradient grey area is the pseudogap phase .
, title="fig : " ] ( @xmath88 is assumed .
labelled in blue in ( a ) ) , the gauge coupling @xmath81 ( labelled in red in ( a ) ) , and the amplitude of the af fluctuation @xmath84 ( labelled in blue in ( b ) ) .
@xmath85 represents the critical doping where the quantum phase transition occurs .
the gradient green area represents the antiferromagnetic phase .
the gradient grey area is the pseudogap phase .
, title="fig : " ]
introducing the massive gauge interaction , our approach offers a detour from the hubbard model toward a weak - coupling theory in the continuous limit .
the current theory can also describe the interaction of finite range , using different mass scales , which can be applicable to iron pnictides and heavy - fermion superconductors @xcite .
nevertheless , it sheds light on the pseudogap formation in those systems .
similar to cuprates , the existence of the remnant af phase fluctuation could be universal , after the afm is frustrated by chemical doping or pressure . different from cuprates , iron pnictides and heavy - fermion superconductors
are multi - band systems .
considering the extended hubbard model of the minimal active bands , a natural generalization is to introduce a non - abelian gauge interaction .
similar to the @xmath18 and @xmath89 bosons in the weak interaction , non - abelian gauge interactions in the multi - band systems , coupled to the @xmath90 af fluctuation , allow different mass ( or zero mass ) depending on the magnitudes of the pseudogap ( or no gap ) . for further development ,
more experimental constraints are required , for example , whether or not there are multiple pseudogap transitions at different temperatures and there are different magnitudes of the pseduogap ?
finally , we remark that the current theory does not imply that the real photon acquires a mass in the pseudogap phase .
the real photon couples to the electron but does not couple to the af fluctuation since the phase field already becomes the longitudinal mode of the internal gauge boson .
different from other gauge theories in the strong - coupling limit @xcite , we construct a weak - coupling gauge theory for the pseudogap formation in cuprates without breaking time - reversal and translational symmetry .
it is not electrons but the _ interaction _ between electrons that couples to the af fluctuation .
the non - perturbative mechanism of the mass acquisition identifies the strong correlation in cuprates .
the pseudogap formation is , therefore , a signature of the strong correlation .
the strong electron - electron interaction and the strong correlation should _ not _ be on the equal footing .
the relevancy of the pseudogap phase and quantum phase transition to the exotic superconductivity should be the next radical research direction to pursue .
i deeply appreciate the enlightening discussions with pei - ming ho , who patiently taught me the stckelberg physics .
i am grateful for the stimulating discussions with jiunn - wei chen . for many years , with highest gratitude , my high-@xmath1 knowledge had been benefited from the educational discussions with robert laughlin , dung - hai lee , naoto nagaosa , and shoucheng zhang .
this work was supported by the national science council of taiwan under nsc 100 - 2112-m-002 - 015-my3 , ntu grant no .
101r7831 , and center for theoretical sciences in ntu .
35ifxundefined [ 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 * * , ( ) | pseudogap formation is an ubiquitous phenomena in strongly - correlated superconductors , for example cuprates , heavy - fermion superconductors , and iron pnictides .
as the system is cooled , an energy gap opens in the excitation spectrum before entering the superconducting phase .
the origin of formation and the relevancy to the superconductivity remains unclear , which is the most challenging problem in condensed matter physics . here
, using the cuprate as a model , we demonstrate that the formation of pseudogap is due to a massive gauge interaction between electrons , where the mass of the gauge boson , determining the interaction length scale , is the consequence of the remnant antiferromagnetic fluctuation inherited from the parent compounds . extracting from experimental data ,
we predict that there is a quantum phase transition belonging to the 2d xy universality class at the critical doping where pseudogap transition vanishes .
pseudogap , weak - coupling theory , strongly - correlated electrons , cuprates |
to express macroscopic observables in terms of the microscopic parameters of composite particles , one has to reduce an infinite number of mechanical degrees of freedom to a few hydrodynamical degrees of freedom@xcite .
two typical techniques are projection and coarse - graining . in the projection method , one can either project the density matrix to a reduced density matrix for the relevant degrees of freedom ( schrodinger picture)@xcite or project a sum of microscopic mechanical variables to a macroscopic slow variable ( heisenberg picture)@xcite .
while in the coarse - graining method , excepting the obligatory ensemble average , one still has to carry out at least one of three average procedures : ( 1 ) coarse - graining in time@xcite ; ( 2 ) coarse - graining in space@xcite ; or ( 3 ) coarse - graining in the eigenvalue spectrum of energy and other collective variables@xcite . in the coarse - graining techniques , the spatial coarse grained average plays an essential role in formulating all macroscopic problems .
macroscopically , a non - equilibrium process in a fluid is described by some local hydrodynamic ( thermodynamic , macroscopic ) variables : temperature , stress , concentration and macroscopic velocity field etc@xcite .
the fundamental assumption behind this macroscopic description is that there exists a local equilibrium at each coarse - grained spatial ` point ' and at every coarse - grained temporal ` moment'@xcite .
similarly in the kinetic theories , at least three of the arguments of the single - particle distribution function , time , position and translational momentum are defined in a coarse - grained sense ( far more coarse - grained than the requirement of uncertainty principle)@xcite . as an alternative to the distribution function in phase space , onsager introduced a distribution function for the coarse grained moments of the conserved quantities ( displacements)@xcite . to obtain hydrodynamic equations and transport coefficients , m. s. green noticed that the resolution of a macroscopic measurement is finite , so that one must eliminate the high wave number components of a collective variable which are are unmeasurable .
the observed value of a hydrodynamic variable is given by averaging the truncated@xcite collective variable over the distribution function of markoff process@xcite .
it is well - known that temporal coarse - graining ( tcg ) is not necessary for deriving the macroscopic maxwell equations ( mame ) from the microscopic ones@xcite .
a safe lower limit of length @xmath0 for a macroscopic description of reflection and refraction of visible light can be taken as @xmath1@xcite .
the time scale @xmath2 associated with @xmath0 is in the range of atomic or molecular motions , a tcg at a time scale @xmath2 is meaningless@xcite .
moreover , if one averages the motion of the particles over a time scale longer than @xmath2 , the scattering phenomena will be smeared@xcite . in the mame ,
the sources of fields are the total charge density and the total current density@xcite . at the same time , the induced charge density and the induced current density are also the responses of the system to the external fields .
since no tcg is taken in deriving mame , one should not require tcg in computing current density ( conductivity)@xcite . the energy dissipation of system is taken care of by the thermal contact with a thermal bath@xcite .
it seems that tcg is not needed for the irreversible processes caused by the mechanical distrubances@xcite .
an open question is whether tcg is necessary to describe an irreversible process induced by an internal ( thermal ) disturbance@xcite .
several schemes have been designed to replace an internal disturbance with a fictitious mechanical disturbance@xcite .
but these methods assume _ a priori _ that the navier - stokes equation is valid@xcite .
in other theories , tcg is often taken along with spatial coarse - graining and ensemble average@xcite . in
particularly , tcg is viewed as a critical step producing irreversibility in the kinetic approach ( master or boltzmann equation)@xcite .
although a velocity gradient is often viewed as an _ internal _ disturbance , it can be realized in a ` mechanical ' manner . in a fluid , two obvious ways to produce a velocity gradient are ( i ) moving a boundary plate which confines the fluid ; and ( ii ) stirring the fluid with a rod in the middle . in both situations ,
a velocity gradient is produced by the interaction between the fluid and an external object at the solid - fluid interface . a non - equilibrium
` hamiltonian ' has been phenomenologically introduced to effectively compute viscosity@xcite .
in addition , for a mechanical disturbance@xcite , one may define microscopic response in a pure state .
the observed macroscopic response ( consequently transport coefficient ) is given by spatial coarse graining the microscopic response and averaging over possible initial conditions@xcite .
it is worthwhile to explore whether we could describe viscosity in a more ` mechanical ' manner : ( i ) not invoking temporal coarse - graining ; and ( ii ) not assuming local equilibrium as the starting point . in this paper , we show that shear viscosity can be strictly calculated from a time - dependent schrodinger equation , and time coarse - graining can be avoided .
one may repeat the same procedure for bulk viscosity _
mutatis mutandis_. in the discussion , we will only use the schrodinger picture . because a hydrodynamic quantity is a bilinear form of the many - body wave function , all the conclusions are valid for any set of identical particles : bosons or fermions .
only when we take concrete approximations for the many - body wave functions , do we need to know whether the particles are fermions or bosons . in sec .
[ cons ] , we first express the velocity gradient in a fluid as a constraint on the wave function of the fluid , cf .
eq.([bc ] ) . to prescribe an internal disturbance like velocity gradient
, one needs the wave function of the system .
the wave function must be determined self - consistently with the given velocity gradient . with velocity gradient as a constraint , the evolution equation for wave function
is then derived using the lagrange multiplier method@xcite , cf .
eqs.([eu1],[eu3],[eu4 ] ) . from the appearance ,
it is a time - dependent schrodinger equation .
two additional terms [ eqs.([eu3],[eu4 ] ) ] appear in the ` hamiltonian ' . both of them contain the occupation probabilities of admissible initial states , which reflects the fact that a velocity gradient is an internal disturbance . on the other hand a mechanical perturbation is completely specified by the time dependence of external field@xcite . in sec .
[ shou ] , we first discuss the entropy production rate of the system + bath , cf .
eqs.([entr],[tre ] ) .
secondly , we check the mass conservation law , cf . eqs.([hmc0],[hmc ] ) .
thirdly , the microscopic response eq.([flu ] ) to the velocity gradient ( the momentum flux in a pure state ) is obtained from the momentum conservation law in a pure state .
the dissipation caused by a mechanical perturbation depends on the occupation probabilities of admissible initial states . because a velocity gradient is an internal disturbance , the dissipation caused by a velocity gradient depends on the _ squares _ of the occupation probabilities of admissible initial states
this feature is clearly seen in the time rates of change of energy , mass and momentum .
representing velocity gradient as a constraint on the wave function [ eq.([bc ] ) ] depends critically on the assumption that a suitable spatial coarse - graining is adequate@xcite to describe internal friction . in sec .
[ put ] we show that the spatial coarse - graining average automatically contains a coarse - graining in time [ eq.([3 ] ) ] and a coarse - graining in eigenvalue spectrum of collective variables [ sec .
[ egcr ] ] . in sec .
[ dirac ] , we apply the method of variation constants to obtain the lagrange multipliers which appear in the solution of the time - dependent schrodinger equation .
the lagrange multipliers characterize the interaction on the system exerted by the boundary plates .
the phenomenological non - equilibrium ` hamiltonian'@xcite can be obtained from @xmath3 , the interaction of system with plates linear in velocity gradient .
the shear viscosity is read from the macroscopic momentum flux ( [ mrf ] ) which is deduced from the wave function of system at some moment . in sec .
[ tcd ] , we explain that applying dirac perturbation theory to transport process is valid .
various time scales in the momentum transport process are discussed .
we show that the length scale of spatial coarse - graining is determined by an intrinsic time scale . in sec .
[ fph ] , we show that the macroscopic stress tensor derived from the microscopic response method is the same as that derived from averaging the momentum flux operator over the density matrix .
we developed a cumulant expansion for the density operator .
when we replace the operator in the exponent with its expectation value , we reproduce the non - equilibrium density matrix obtained from other approaches .
a mechanical disturbance on a system can be expressed by the coupling between the mechanical degrees of freedom of the system and some specified external parameters which may depend on time and position .
the disturbance caused by a velocity gradient is more complicated .
we will see that to describe a fluid with a velocity gradient , the occupation probabilities of admissible initial states enter the ` hamiltonian ' of the fluid . in sec.[msm ] we briefly summarize the microscopic response method for a mechanical perturbation@xcite .
later , we will show that one can still define a momentum flux in a pure state ( microscopic response to velocity gradient ) , viscosity can be calculated from a modified time - dependent schrodinger equation . consider a @xmath4particle system
@xmath5 described by hamiltonian @xmath6 , we use indexes @xmath7 to label the eigenvalues @xmath8 and eigenstates @xmath9 of @xmath6:@xmath10 if the system @xmath5 is in thermal contact with a thermal reservoir @xmath11 at temperature @xmath12 , and is in equilibrium with it@xcite , the probability that the system is in state @xmath13 is @xmath14 because the system is macroscopic , comparing the energy of system @xmath5 , the energy exchange rate between the system and the bath is slow@xcite . in a time period
much longer than a macroscopic measurement , the system can be viewed as isolated , and described by a wave function . for a macroscopic fluid ,
the detailed dynamics of the surrounding bath is not important .
if a mechanical disturbance described by hamiltonian @xmath15 is exerted on @xmath5 , the state @xmath16 of @xmath5 will evolve according to @xmath17 the above description requires that@xcite @xmath5 is in good thermal contact with @xmath11 such that the energy generated by the disturbance @xmath15 inside @xmath5 can be rapidly transferred into @xmath11 , and the system is maintained at the temperature @xmath12 of bath @xmath11 .
if we assume that the system is initially in state @xmath18 , the state @xmath19 of system at time @xmath20 is determined by eq.([tse ] ) with initial condition @xmath21 .
for convenience we adiabatically introduced @xmath15 . by means of the microscopic conservation law in a pure state , we can define@xcite the corresponding microscopic response ( flux ) in state @xmath19
. one can always spatially resolve a quantity @xmath22 in state @xmath23 into its fourier components : @xmath24 because any instrument has finite spatial resolution , one can only detect those @xmath25 with @xmath26 , where @xmath27 is some instrument limited cut - off wave number@xcite . to express a macroscopic measurement
, one has to make spatial coarse - grained average@xcite:@xmath28 where @xmath29 is any function satisfying : ( 1 ) @xmath30 ; and ( 2 ) the fourier components of @xmath29 tends rapidly to zero for @xmath31 .
the integral in eq.([cg ] ) is over the sample . since the system is in thermal contact with a reservoir , i.e. a member of a canonical ensemble
, one does not know which state the system is initially in .
the measured quantity is an ensemble average of @xmath32:@xmath33 where @xmath34 is the probability that the system is in state @xmath21 in the remote past .
since the microscopic response @xmath22 is a bilinear form of @xmath23 , it can be computed from eq.([tse ] ) rather than from the liouville equation for the density matrix . since the spatial coarse - graining ( [ cg ] ) and the average over initial conditions ( [ ea ] ) do not involve the temporal evolution of the system , the macroscopic response @xmath35 and consequently transport coefficient are also determined by the time - dependent schrodinger equation@xcite . the averaging procedure ( [ cg],[ea ] ) is simpler than , and equivalent to kubo s linear response theory@xcite .
the procedure ( [ cg],[ea ] ) aims at the steady states @xmath36 of @xmath5 under a monochromatic driving .
the general linear casual response to a disturbance with arbitrary time dependence will be discussed in sec.[caus ] .
consider a fluid composed of @xmath37 identical particles with mass @xmath38 .
we use indexes @xmath39 , @xmath40 , @xmath41 to label the particles . denote the interaction energy between two particles at @xmath42 and @xmath43 as @xmath44 . then the hamiltonian of system is @xmath45 put the system in an external field , denote the interaction potential energy for a particle at @xmath42 in the external field as @xmath46 .
the interaction hamiltonian with external field is @xmath47 in the remote past @xmath48 , suppose that the system is in state @xmath18 .
driving the system with a velocity gradient and @xmath15 , @xmath18 will evolve into a state @xmath49 at time @xmath20 .
eq.([canp ] ) indicates that we do not have the full knowledge of the system at the initial moment .
the evolution of the system can then be understood to be probabilistic@xcite .
the interaction @xmath50 among molecules renders the macroscopic motion involving a velocity gradient irreversible@xcite . using the definition of mass density operator @xcite @xmath51 ,
the mass density for state @xmath19 is a function of time and position:@xmath52 where @xmath53 the arguments of @xmath23 are @xmath54 . the momentum density operator is defined as@xcite@xmath55 , \label{dc}\ ] ] where @xmath56 are indexes for three cartesian components .
the momentum density in state @xmath23 is @xmath57 according to eqs.([cg],[ea ] ) , the measured macroscopic mass density is@xmath58 where the arguments of @xmath23 are @xmath59 .
the measured macroscopic momentum density is@xmath60 the macroscopic velocity field should be defined as@xcite:@xmath61 it satisfies the requirement that the macroscopic velocity is the macroscopic momentum of unit mass of fluid@xcite .
direction ( right ) with speed @xmath62 .
the bottom oefg , front ( right ) bank aefd and back ( left ) bank bogc are static .
the width of channel is @xmath63 , the depth of channel is @xmath64.,width=302 ] we first describe the simplest experimental setup for measuring shear viscosity in fig.[fig1 ] .
let the fluid flow in a rectangular channel along the @xmath65 direction .
we choose a coordinate system such that the two banks bogc and aefd of the channel are @xmath66 and @xmath67 , the bottom surface oefg of channel is @xmath68 , the top surface abcd of the fluid is bordered by a plate @xmath69 with velocity @xmath62 .
the points on the bottom plate oefg are denoted as @xmath70 with @xmath71 and @xmath72 , the boundary condition on bottom is @xmath73 , @xmath56 .
the points on the top plate abcd are denoted as @xmath74 with @xmath71 and @xmath72 , the boundary condition on the top plate is @xmath75 .
the points on the back ( left ) bank bogc of channel are denoted as @xmath76 with @xmath71 and @xmath77 , the boundary condition on the back bank bogc is @xmath78 , @xmath56 .
the points on the front ( right ) bank aefd of channel are denoted as @xmath79 with @xmath71 and @xmath77 , the boundary condition on the front bank aefd is @xmath80 , @xmath56 .
these no - slip conditions for a viscous fluid are the boundary conditions for the navier - stokes equations@xcite .
we imposed a velocity gradient @xmath81 in the fluid . to enforce a velocity gradient as a constraint on the wave function , it is helpful to allow an arbitrary velocity field on each boundary surface .
we use a two dimensional mesh to characterize the positions of the points on a boundary surface . for a point on the @xmath82 plate ( @xmath83 ) , we use two indexes @xmath84 and @xmath85 to describe its position , where @xmath86 and @xmath87 .
thus the position vector of a point on the @xmath82 plate with indexes @xmath84 and @xmath85 is denoted as @xmath88 , there are @xmath89 points on the @xmath82 boundary surface .
a general velocity field on the @xmath82 boundary surface may be specified @xmath90 , @xmath56 . in terms of eq.([mfv ] ) , the boundary condition on the @xmath82 boundary surface is expressed as@xmath91 where @xmath92 is a know function of @xmath93 . making use of eqs.([md],[mod],[ond ] ) , eq.([nsli ] ) becomes @xmath94@xmath95=0.\ ] ] eq.([bc ] ) represents @xmath96 constraints on the wave function of system .
the evolution equation of @xmath19 has to be modified with constraints ( [ bc ] ) . to consider the effect of constraints ( [ bc ] ) on the state @xmath23 of the many - particle system
, we first notice that the schrodinger equation@xmath97 is the euler equation @xmath98 of the action@xcite:@xmath99 where @xmath100 , @xmath101 , @xmath102 is the volume element in configurational space , the arguments of @xmath23 are @xmath103 . the well - known lagrangian density for eq.([se ] ) is@xcite:@xmath104@xmath105\psi_{a}^{\ast}\psi_{a}.\ ] ]
we apply the lagrange multiplier method@xcite to find the evolution equation of @xmath19 driven by a velocity gradient specified by eq.([bc ] ) .
first of all we symmetrize eq.([bc ] ) respect to all particles .
secondly , we multiply the expression with @xmath106 and integrate it over the time interval@xcite @xmath107 $ ] : @xmath108@xmath109@xmath110=0 , \label{cst}\ ] ] where the @xmath111 function is produced when we change @xmath112 in eq.([bc ] ) to @xmath113 in eq.([cst ] ) .
the macroscopic velocity at every point on each plate is a constraint . the overall lagrangian density including the constraint ( [ cst ] ) is@xmath114@xmath115@xmath116\ ] ] where the lagrange multiplier @xmath106 is a real function of time@xcite with dimension length@xmath117time@xmath118 .
roughly speaking , @xmath106 is a product of the @xmath119 component of velocity at point @xmath88 of the @xmath82 plate and the disturbed volume of fluid by a molecule at @xmath88 .
we will see this in more detail in sec.[cz ] .
the equation of motion for @xmath19 can be obtained from the euler equation:@xmath120 it is@xmath121 where@xmath122@xmath123 and@xmath124@xmath125 one may notice that eq.([eu1 ] ) contains only @xmath23 and its partial derivatives .
other @xmath126 do not appear . in this sense , the velocity gradient is a special mechanical perturbation . on the other hand ,
the solution @xmath19 of eq.([eu1 ] ) involves @xmath127 parameters @xmath128 , that should also satisfy @xmath127 constraints ( [ bc ] ) which mix all @xmath129 .
thus velocity gradient is also an internal disturbance@xcite . in @xmath130 and @xmath3 ,
the integrals behind the summation signs are the macroscopic momentum density and negative momentum density operator respectively .
thus the subtracted momentum density automatically happens@xcite in @xmath131 , which has to be introduced in some previous theories@xcite .
because @xmath132 , one can not find a system with wave function @xmath133 such that @xmath134 , so that the process described by eq.([eu1 ] ) is irreversible@xcite .
in classical mechanics , constraints can also be dealt with the lagrange multiplier method@xcite .
a constraint produces a reaction force on each particle in the system .
the constraints ( [ bc ] ) enforce the interaction on the system from four the plates . defining the macroscopic velocity field of a fluid by eq.([mfv ] ) has a profound consequence : the resultant constraint ( [ bc ] ) on wave function is bilinear .
the evolution equation ( [ eu1 ] ) for @xmath19 is linear and does not include other states . the evolution of the state of system driven by a velocity gradient is _ almost _ determined in the framework of pure mechanics .
if we had defined the macroscopic velocity field of fluid as @xmath135 with@xmath136 we would have two serious difficulties : ( i ) this definition breaks the basic requirement that the _ macroscopic _ mass flux density @xmath137 must always be the _
macroscopic _ momentum of unit volume of fluid@xcite ; ( ii ) the resulted temporal evolution equation for @xmath19 is nonlinear , and @xmath19 couples with other @xmath138 . certainly , assuming the fluid is incompressible , the labyrinthine evolution equations could be linearized and decoupled . for a mechanical perturbation , external fields are some specified functions of time .
the interaction hamiltonian @xmath15 is expressed by the external fields , the coordinate and momentum operators of all particles@xcite .
the states @xmath36 of system are not involved in @xmath15 . on the other hand ,
the macroscopic velocity field of a fluid is defined in eq.([mfv ] ) through macroscopic mass density ( [ md ] ) and macroscopic mass flux ( [ ond ] ) .
the later two quantities involve the states @xmath36 of system , spatial coarse - graining and average over various initial conditions . before specifying the velocity gradient , which involves the states @xmath139 of system
, we can not determine the evolution of the state of system driven by a velocity gradient .
from the charge conservation law in a pure state , one can _ directly _ use the time - dependent schrodinger equation ( [ tse ] ) to derive@xcite the current density @xmath140 in state @xmath23 ( the microscopic response ) from the temporal change of charge density @xmath141 in state @xmath19 .
one does not need the quantal operator of current density@xcite , although the results are the same from the two points of view . in contrast , we can not use mass conservation in a pure state to derive mass flux @xmath142 in state @xmath19 .
we have to invoke the definition of momentum density ( [ dc ] ) . in this sense
, the velocity gradient is an internal disturbance@xcite .
the states @xmath36 of system driven by velocity gradient have to be determined by the equation of motion ( [ eu1 ] ) and the constraints ( [ bc ] ) self - consistently .
in this section , we assume that @xmath106 and @xmath129 are known . the time rate of change of energy , mass density and momentum density are derived .
these relations are formal but exact .
we will find @xmath106 and @xmath129 in sec.[dirac ] . the temporal or spatial change in _ local _ energy _ density _ is closely related to the temperature gradient . since @xmath5 is a member of the canonical ensemble at temperature @xmath12 , we can not talk about the time rate of change of _ local _ energy _
density_. for an isolated system @xmath5 , the total energy is conserved:@xmath143 consider a more general state of isolated system @xmath5 with initial value @xmath144 , where @xmath145 are some constants .
the time evolution of @xmath146 is @xmath147 . from the orthogonality of eigenfunctions of @xmath6 ,
it is easy to check that the average energy of @xmath148 is conserved : @xmath149 and@xmath150 now assume that the system @xmath5 is in good contact with a thermal reservoir @xmath11 in temperature @xmath12 . thus , even if we apply a velocity gradient and an external field on the system , @xmath5 is maintained at the same temperature @xmath12 . because @xmath5 is macroscopic , in a time period much longer than a macroscopic measurement
, @xmath5 can be viewed as quasi - closed@xcite .
therefore we can use a wave function @xmath19 to describe the system .
the average energy of @xmath5 in state @xmath151 is defined as @xmath152 because @xmath19 satisfies eq.([eu1 ] ) , one can get the time rate of change @xmath153 of the energy @xmath154 in pure state @xmath19 .
the average energy @xmath155 of @xmath5 at time @xmath20 is @xmath156 .
the time rate of energy exchange @xmath157 between @xmath5 and @xmath11 is determined by @xmath158@xmath159@xmath160@xmath161@xmath162@xmath161@xmath163@xmath164@xmath165@xmath166 where the arguments of @xmath23 are@xmath167 .
the last term is the power of the internal force near the boundary surfaces . in classical mechanics , for a group of particles , the internal forces do work on the system@xcite . only for a rigid body ,
is the work done by internal forces zero .
it is interesting to notice that only for the regions close to the boundary plates [ @xmath168 decays very rapidly when @xmath169 , the internal forces do work . in eq.([tre ] ) , we can see an obvious distinction between a velocity gradient and a mechanical perturbation .
the first term contains @xmath170 , is the power due to the external field .
the remaining terms contains @xmath171^{2}$ ] , the square of the occupation probability of admissible initial states .
they are the contributions to power due to the friction at boundary surfaces .
the first @xmath170 comes from prescribing the macroscopic boundary condition as constraints on the wave function of a pure state .
the interaction with plates are expressed by @xmath131 in the ` hamiltonian ' of system , cf .
eqs.([eu3],[eu4 ] ) . since @xmath153 is determined by @xmath172 , from eq.([eu ] ) , @xmath153 includes a @xmath170 . to obtain @xmath157 from @xmath153 , the average about possible initial states ( [ ea ] ) introduced the second @xmath170 .
the temperature of @xmath5 is maintained at @xmath12 by the bath @xmath11 .
all the energy generated in the system is transferred to @xmath11 .
the entropy @xmath173 of the bath @xmath11 increases with time according to @xmath174 substituting eq.([tre ] ) into eq.([entr ] ) and noticing that @xmath175 has dimension of velocity , each term is in the form of temperature@xmath118 @xmath176velocity@xmath177momentum@xmath176velocity gradient , except the first and last term .
this is consistent with the macroscopic law of entropy production rate@xcite . with the help of eq.([eu1 ] )
, one may calculate the time rate of change@xcite of @xmath141 . further applying the averaging procedure ( [ cg],[ea ] ) to the expression for @xmath178 ,
one obtains the macroscopic law of mass conservation:@xmath179 where@xmath180@xmath181^{2}\rho^{\psi_{a}}(\mathbf{r}^{\prime},t).\ ] ] in the r.h.s of eq.([hmc ] ) , the first term @xmath182 is the usual macroscopic mass flux , and is given by eq.([ond ] ) .
the second term is an additional mass flux caused by the given velocity gradient .
two truncation functions@xcite @xmath183 and @xmath184 indicate that the correction is nonzero only near the boundary surfaces .
again , @xmath171^{2}$ ] reflects the fact that a velocity gradient is an internal disturbance . taking the time derivative of eq.([mod ] ) , one obtains the time rate of change of the momentum density@xcite @xmath185 in state @xmath19 . in terms of eq.([eu1 ] ) , one has:@xmath186@xmath187@xmath188,\ ] ] where @xmath189 @xmath190 @xmath56 , and the momentum flux @xmath191 in state @xmath23 is@xmath192@xmath193\]]@xmath194 in eq.([flu ] ) , the arguments of @xmath23 are @xmath195 . to obtain the divergence form for the second term in eq.([flu ] ) , we used the fact that the range of interaction force is much shorter than@xcite @xmath196 .
although velocity gradient is an internal disturbance , by viewing the velocity gradient as a constraint on the wave function , one can still define a microscopic response : the momentum flux @xmath197 in state @xmath19 . applying eqs.([cg],[ea ] ) ,
one obtains the measured macroscopic momentum flux:@xmath198 after @xmath19 is obtained , the shear viscosity can be read out from eqs.([flu],[mdu ] ) . in the rhs of eq.([cond ] ) , the first term is the momentum density produced by the external field per unit time , the second term is the momentum density produced by the constraint ( [ bc ] ) per unit time . because the moving plate drives the fluid , @xmath5 gains momentum and energy .
thus there are source terms in the equations for the time rate of change , cf .
( [ cond ] ) and ( [ tre ] ) .
since no mass is produced by a velocity gradient , there is not a source term in the time rate of change of mass density , cf .
eq.([hmc0 ] ) .
there is a subtle difference between eqs.([cond],[flu ] ) and the corresponding results in previous theories@xcite . in all previous theories , conservation laws and fluxes ( responses )
are derived for an isolated fluid system without imposing velocity gradient .
the system satisfies @xmath199 or newton s equation for an isolated system@xcite .
the operator of momentum flux is@xcite@xmath200@xmath201@xmath202\delta(\mathbf{r}_{i}-\mathbf{r } ) .
\label{dfu}\ ] ] to obtain the viscosity , one must average operator ( [ dfu ] ) over a non - equilibrium density matrix ( or distribution function ) @xmath203 which accommodates the effects of velocity gradient@xcite .
this reminds us of some older theories for the response to an electromagnetic field . in those theories ,
the current operator is derived from the schrodinger equation for an isolated system ( without external electromagnetic field ) .
the resulting current density missed the vector potential term and broke gauge invariance@xcite . in eq.([flu ] ) , @xmath19 is the solution of eq.([eu1 ] ) rather than that of eq.([fs ] ) .
thus the influence of velocity gradient has been included in @xmath204 .
if we know @xmath151 , eq.([flu ] ) directly gives the microscopic response @xmath191 to a velocity gradient in state @xmath19 .
we will compare the results derived from two method in sec.[fph ] . at this point
, we only note that eq.([flu ] ) reduces to the ordinary conservation law in the bulk , and is also correct near to the boundary surfaces . for the effects which are first order in velocity gradient , eq.([cond],[flu ] ) gives the same results as those based on eqs.([fs],[dfu ] ) .
let us check whether the average procedure ( [ cg],[ea ] ) is adequate and enough to describe the dissipation produced by a velocity gradient . in a liquid ,
suppose that the force between two neighboring molecules is @xmath205 , where @xmath206 is the average distance between two molecules , @xmath50 is the interaction energy between two molecules at a distance @xmath206 .
the time needed to relax a change @xmath207 in momentum is @xmath208 .
the time needed for a molecule to move a distance @xmath206 is @xmath209 . near the equilibrium state ,
the potential energy is balanced by kinetic energy @xmath210 , and the two microscopic relaxation times are the same order @xmath211 and will be denoted as @xmath212 . we show that the spatial coarse - grained average contains temporal coarse - graining automatically . applying the spatial coarse - graining average ( [ cg ] ) to the microscopic law of mass conservation in a pure state @xmath23 , the time rate of change of a conserved quantity must be the same order as the divergence of the corresponding flux @xmath213 , i.e.@xmath214 where @xmath215 is the number density of fluid .
the relaxation time @xmath216 for a fluid drop is @xmath217 . because the cut - off wave vector is @xmath27 , the time needed for a molecule to pass a distance @xmath196 is @xmath218 , the same order as @xmath216 .
the temporal coarse - graining is automatically realized in the conservation law of mass by the spatial coarse - graining ( [ cg ] ) .
applying spatial coarse - graining to the microscopic law of momentum conservation in a pure state @xmath23 , one has @xmath219 , i.e.@xmath220 the relaxation time of momentum in a fluid drop is the same as @xmath216 .
the temporal coarse - graining is automatically realized in the conservation law of momentum by the spatial coarse - graining ( [ cg ] ) .
we do not have energy relaxation .
we derive the coarse - grained time scale @xmath216 from another point of view
. there are @xmath221 molecules in a coarse - grained fluid drop .
because the range of molecular interaction is only extended to @xmath206 , a fluid drop interacts with its surrounding only through its surface .
thus the force on a fluid drop is @xmath222 .
the acceleration of the fluid drop is @xmath223 , the time needed for the fluid drop to move a distance @xmath224 is @xmath225 .
two characteristic times for the coarse - grained fluid drop are the same order @xmath226 , will be denoted as @xmath227 . because @xmath228 , we can see @xmath229 , the coarse - grained time scale is much longer than the microscopic time scale @xmath230 .
the time coarse graining induced by the spatial coarse graining is fine enough to describe macroscopic motion .
the shear viscosity is order of @xmath231 . in a macroscopic measurement ,
the acceleration on a fluid drop produced by the velocity gradient is @xmath232^{-1}$ ] .
the macroscopic relaxation time is order of @xmath233t_{c}\thicksim t_{c}l_{y}/d$ ] , where @xmath64 is a macroscopic length .
therefore @xmath234 . in summary
, we have@xmath235 for the thermal motion of molecule , we could make a similar estimation : the velocity and kinetic energy of a molecule should use @xmath236 and @xmath237 respectively .
relation ( [ 3 ] ) holds . in conclusion , the spatial coarse - graining ( [ cg ] ) included a temporal coarse - graining , which is adequate and enough to describe the dissipation produced by a velocity gradient . in sec.[msm ] , the quantal mean value of any operator in a pure state is spatially coarse - grained with eq.([cg ] ) .
many authors choose to coarse - grain the eigenvalue spectrum of collective variables .
we will illustrate that the spatial coarse - graining in the microscopic response includes coarse - graining in the eigenvalue spectrum of collective variables . because the cut - off wave vector is @xmath27 , there are @xmath238 molecules in a coarse - grained fluid drop . a collective variable for a liquid drop is a symmetric sum of single - particle variables@xcite .
the typical interval @xmath239 between two eigenvalues of a collective variable is @xmath240 , where @xmath241 is the typical interval between two eigenvalues of the corresponding single - particle variable@xcite . excepting macroscopic quantum phenomenon ( superconductivity and superfluidity ) , even in a microscopic relaxation time @xmath242 , the action for a drop is @xmath243 .
therefore the collective variables are classical variables .
since the relative error for a collective variable is@xcite @xmath244 , the measurement error for the energy of a fluid drop is @xmath245 .
the energy interval for a single molecules is @xmath246 .
a macroscopic energy measurement with spatial resolution @xmath196 must involve @xmath247 eigenvalues and corresponding eigenvectors of the energy operator of a liquid drop .
in linear response theory , to obtain the observed macroscopic response to a time - dependent mechanical perturbation , dirac s method of the variation of constants is frequently used to solve the liouville equation@xcite or schrodinger equation@xcite .
we first use dirac s perturbation theory to get lagrange multipliers , and then discuss its applicable conditions . consider a state @xmath19 with initial condition in the remote past @xmath21 . in first order of @xmath248 , the normalized wave function is@xmath249 for both @xmath250 and @xmath251 , the first order expansion coefficient is given by:@xmath252_{ja}(t ) .
\label{1c}\ ] ] the second term in eq.([1c ] ) is the contribution from the moving plate and three static surfaces:@xmath253_{cb}(t)=\frac{1}{2}p^{\phi_{b}}\int_{-\infty}^{t}dt^{\prime}e^{i\omega_{cb}t^{\prime}}\int d\mathbf{r}f(\mathbf{r}_{\mu_{\sigma}\nu_{\sigma}}^{(\sigma)}-\mathbf{r})\]]@xmath254@xmath255@xmath256 where @xmath257 .
denote @xmath206 as the average distance between two neighboring molecules , @xmath258 the frequency of a monochromatic driving velocity @xmath259 , @xmath260_{ja}\thicksim\lbrack(\omega_{ja}+\omega ) d^{4}]^{-1}$ ] characterizes the transition caused by a point @xmath88 at the @xmath82 plate , and has the dimension of time@xmath261length@xmath262 . the first term in eq.([1c ] )
is the contribution from the external field:@xmath263\phi_{a}. \label{fc}\ ] ] the first term in the rhs of eq.([sup ] ) is the free evolution term , @xmath264 is caused by the requirement that @xmath265 should be correct to first order in @xmath266 . in the kinetic theories ,
one concerns various transition probabilities .
we only need the @xmath251 terms in eq.([sup ] ) .
before we apply the external field and move the top plate , the system @xmath5 is in equilibrium at the bath temperature @xmath12 : the system is in various stationary states \{@xmath13 } according to eq.([canp ] ) . in an ordinary fluid
, there is no macroscopic mass flux at any boundary surface @xmath267:@xmath268@xmath269 where @xmath56 , @xmath83 , the arguments of @xmath13 are ( @xmath270 ) .
superfluidity is excluded by the requirement ( [ sta ] ) . to conserve space in the subsequent calculations , we rewrite eq.([sup ] ) as @xmath271 where @xmath272 for @xmath250 ; @xmath273 for @xmath251 .
the summation includes both @xmath250 and @xmath251 .
substitute eq.([sup1 ] ) into boundary conditions ( [ bc ] ) , and notice eq.([sta ] ) for four boundary surfaces , then we have @xmath274@xmath275@xmath276\]]@xmath277@xmath278\},\ ] ] where @xmath279 is the constant mass density of liquid in absent of external field and boundary conditions ( [ bc ] ) . making use of eq.([1c ] ) , eq.([cst1 ] ) is reduced to a group of linear equations for lagrange multipliers:@xmath280(t)\lambda_{\mu_{\sigma^{\prime}}^{\prime } \nu_{\sigma^{\prime}}^{\prime}}^{\gamma^{\prime}(\sigma^{\prime})}(t)=b_{\mu_{\sigma}\nu_{\sigma}}^{\gamma(\sigma)}(t),\label{li}\ ] ] where@xmath281@xmath282@xmath283\]]@xmath284\},\ ] ] and@xmath285(t)=\int d\mathbf{r}f(\mathbf{r}_{\mu_{\sigma}\nu_{\sigma}}^{(\sigma ) } -\mathbf{r})\label{l2}\]]@xmath286_{db}(t)e^{-it\omega_{db}}\]]@xmath287\]]@xmath288_{db}^{\ast}(t)e^{it\omega_{db}}\]]@xmath289\}.\ ] ] the dimension of @xmath290 is momentum density . the first term in eq.([l1 ] ) is the momentum density produced by a point @xmath291 at the @xmath82 plate @xmath292 .
the second term is the momentum density produced by the interference between the external field and the constraint of velocity gradient @xmath293d^{-3}$ ] , where @xmath294 is the frequency of the external field .
@xmath295 is the net momentum density at @xmath291 on the @xmath82 plate .
the dimension of @xmath296 is momentum density@xmath261time@xmath261length@xmath262 . @xmath297\thicksim\lbrack(\omega
_ { db}+\omega)d^{4}]^{-1}(\hbar / d)d^{-3}$ ] describes the effect of a point @xmath298 at the @xmath299th plate to the point @xmath88 on the @xmath82 plate .
eq.([li ] ) describes the momentum density at point @xmath300 produced by all points on the four boundary plates .
the lagrange multipliers @xmath301 are determined by the coupled linear equations ( [ li ] ) , and they are in general nonlinear functions of the velocity gradient . the lowest order approximation to @xmath301 is first order in @xmath302 : @xmath303[(\omega _ { db}+\omega)d]d^{3}. \label{mda}\ ] ] from eq.([tre ] ) , we can see that the lowest order dissipation rate is second order in velocity gradient , which is consistent with the result derived from navier - stokes equation and the second law of thermodynamics@xcite .
in addition , from eqs.([mda],[eu3],[eu4 ] ) , @xmath130 is second order and @xmath3 is first order in velocity gradient .
this is similar to the coupling between electromagnetic field and material : there are both first order and second order terms in the vector potential .
the small parameter for the perturbation caused by @xmath15 is @xmath304 , for the perturbation caused by @xmath3 is @xmath305 , for the perturbation caused by @xmath130 is @xmath306^{2}$ ] .
if we substitute the estimation ( [ mda ] ) into eq.([eu4 ] ) , one has@xmath307 the quantity in \ { } is the dissipative part of the phenomenological non - equilibrium hamiltonian for the @xmath308 molecule@xcite . here
@xmath309 corresponds to the strain rate ( @xmath258 the frequency of a monochromatic driving ) , @xmath206 the displacement of an atom , @xmath310 the momentum of an atom at point @xmath88 .
of course , eq.([oneq ] ) has not changed from the boundary surfaces to the bulk@xcite . from @xmath311 , eqs.([z1],[fc],[1c ] )
tells us all the expansion coefficients in eq.([sup1 ] ) i.e. the state @xmath19 .
next , we notice that no momentum flux exists in an equilibrium state:@xmath312 where@xmath313@xmath314\]]@xmath315 is the momentum flux in pure state @xmath13 . by means of eq.([sup1 ] ) , the momentum flux ( [ flu ] ) in a pure state is reduced to @xmath316@xmath317@xmath318@xmath319@xmath320 applying the average procedure ( [ cg],[ea ] ) to eq.([fl ] ) , the observed macroscopic momentum flux ( [ mdu ] ) becomes @xmath321@xmath322@xmath323@xmath324@xmath320 eq.([mrf ] ) is first order in @xmath266 .
the order @xmath81 contribution comes from @xmath3 . to recover usual viscosity ( the proportional constant between momentum flux and velocity gradient @xmath81 ) , in @xmath325
, we only need the first term of eq.([z1 ] ) .
applying the method of variation of constants ( mvc ) to a macroscopic system worries many scientists@xcite .
the kinetic approach insists that the long time behavior of a large system should be described by a time coarse - grained master equation@xcite .
we will show that with some caution , the mvc is applicable to calculate shear viscosity .
we first estimate the time period during which the system can be described by a wave function .
the thermal reservoir plays the role of damping for @xmath5 .
eq.([eu1 ] ) should be revised as@xmath326 where @xmath327 is the net power exerted on @xmath5 .
usually the last term is not written out , but is implicitly taken into account@xcite .
@xmath5 receives the power generated by the external field and velocity gradient , in the same time @xmath5 delivers extra energy to @xmath11 to maintain the equilibrium with @xmath11 .
if at the initial moment @xmath48 , the system is exactly in a pure state @xmath328 , @xmath19 is uniquely determined by eq.([os ] ) and is finite for any @xmath329
. then eq.([cond ] ) and the corresponding macroscopic equation obtained by the averaging procedure ( [ cg],[ea ] ) works for any @xmath329 .
however , we only know @xmath170 , an incomplete specification of the initial state of @xmath5@xcite , so that a wave function description for @xmath5 can not last very long .
let us use@xmath330\phi_{a}|\label{coh}\]]@xmath331\phi_{b}|\ ] ] to define a coherence time @xmath332 for state @xmath23 .
the expression inside the modulus sign on the lhs is the remaining probability amplitude of state @xmath18 at time @xmath20 if @xmath21 , the expression inside the modulus sign on the rhs is the probability amplitude of state @xmath18 at time @xmath20 if @xmath333 .
when @xmath334 , it is not possible to describe @xmath5 with a wave function .
the energy exchange between @xmath5 and @xmath11 makes the system a member of the canonical ensemble at temperature @xmath12 , and the system distributes itself into various stationary states @xmath9 according to eq.([canp ] ) .
the incomplete specification of initial state ( the indeterminacy of initial state ) makes the system dephase in a time period @xmath335 .
the phase randomization continues ceaselessly , so that one no longer needs the solution @xmath19 of eq.([eu1 ] ) for @xmath336 .
expressions ( [ z1],[fc ] ) indicate that dirac s perturbation theory can be used only for a time interval much shorter than @xmath337 we can see that @xmath338 is the same order as @xmath335 . in other words ,
the method of variation of constants fails long before the system loses its wave function description .
the multi - scale method and fano s ansatz@xcite can be used to obtain @xmath19 for a longer period : we will discuss them elsewhere . as long as the system can be described by a wave function ( @xmath339 ) , we can apply the method of variation of constants to calculate the time evolution of wave function .
we show that the indeterminacy of the interaction time will not lead to any serious error in the results obtained by the mvc .
from@xmath340 we see that the characteristic interaction time is @xmath341 , where @xmath258 is the frequency of external conditions . for some interaction time
@xmath20 , @xmath342 because any external disturbance ( including velocity gradient ) is only exerted on a few degrees of freedom , the external distance can quickly relax to other degrees of freedom@xcite .
there is an indeterminacy @xmath343 in the interaction time , which is the order of microscopic relaxation time @xmath344 .
the error caused by this indeterminacy is@xmath345 the relative error is @xmath346 in addition , the phase randomization induced by the energy exchange between bath and system leads to a destructive interference in any observable which is bilinear in wave function @xmath19 , i.e. , the error will not accumulate with time .
therefore the indeterminacy in the interaction time does not cause serious error in any observable .
this conclusion will not change for two degenerate levels @xmath347 in a static field @xmath348 and for two levels that are in resonance with external condition @xmath349 . in both situations ,
the characteristic interaction time is @xmath350^{-1}$ ] , the relative error of transition amplitude is @xmath351 . for details ,
see ref.@xcite .
we show that the mvc is compatible with spatial coarse - graining ( [ cg ] ) .
the effect of a monochromatic external field is simple@xcite .
an electromagnetic field or a gravitational field , affects every molecule at the same time .
the whole system reaches steady state after a microscopic relaxation time @xmath352 .
the cutoff wave vector @xmath27 is determined by the precision of macroscopic measurement , is not related to any time scale .
but for a velocity gradient , the length scale of spatial coarse - graining relates to an intrinsic time scale .
if a velocity gradient with time dependence @xmath353 is applied to @xmath5 by a plate or a rod , after a time period several times of the characteristic time @xmath354^{-1}\}$ ] , the molecules within a distance @xmath355 to the plate or rod reach steady state . @xmath356 and momentum flux @xmath357 contain terms with @xmath353 factor and terms with @xmath358 factor .
the coefficient of @xmath353 gives us the real part of the viscosity @xmath359 , the coefficient of @xmath358 gives us the imaginary part of the viscosity @xmath360 .
for the aim of computing viscosity , we choose cutoff wave vector @xmath27 through @xmath361 . for the steady state under a monochromatic velocity gradient , the macroscopic response of @xmath5 at time @xmath20 corresponds to the microscopic response at the same time @xmath20 . the averaging procedure ( [ cg],[ea ] )
reflects this fact . for a time - dependent _ mechanical _ perturbation ,
the macroscopic response at time @xmath20 depends on all the past history @xmath362 of the external conditions .
therefore the macroscopic response is a convolution of all monochromatic transport coefficients and the external disturbance@xcite . for a velocity gradient with arbitrary time - dependence ,
the macroscopic response is more complicated . for an internal disturbance with arbitrary time dependence
, we make a temporal fourier resolution : @xmath363 because @xmath266 in eq.([eu1 ] ) is complex , we will not require @xmath364^{\ast}$ ] .
the first order transition amplitude is a sum of every monochromatic transition amplitude:@xmath365 .
\label{sr}\ ] ] according to eqs.([fl],[mdu ] ) , the macroscopic momentum flux is a frequency integral over all monochromatic momentum flux .
one must notice that this result originated from _ first _ order perturbation theory . unlike a classical oscillator , in eq.([eu1 ] )
the external conditions are multiplied by @xmath19 .
although eq.([eu1 ] ) is linear in @xmath23 , the response to @xmath266 is not the sum of the responses to @xmath366 in general .
the density matrix of the system is often defined as@xcite@xmath367 where @xmath368 is the number of systems in an ensemble , @xmath369 is the probability amplitude that the @xmath370 system is in state @xmath371 at time @xmath20:@xmath372 the density matrix @xmath373 can also be defined by the initial value problem@xcite:@xmath374,\text { } \widetilde{\rho}(t=-\infty)=\rho^{eq}.\label{liu}\ ] ] here @xmath375 is the equilibrium density operator , and @xmath376tr@xmath377 . by means
of@xmath378 it is easy to check that if @xmath266 is hermitian , @xmath373 is the same as @xmath379 in eq.([wd ] ) .
integrating by parts , @xmath3 becomes @xmath380@xmath381 where @xmath382 ( @xmath56 ) is the momentum operator for the @xmath308 particle .
we see that @xmath248 is hermitian .
thus we can use either eq.([wd ] ) or eq.([liu ] ) to define density matrix .
the former is more convenient for us since we already obtained probability amplitudes in sec.[cz ] .
the perturbation expansion ( [ sup1 ] ) for a time - dependent wave function @xmath19 leads to a similar expansion for the density matrix@xmath383 where the superscripts indicate the order in @xmath266 . because ( i ) the system is enclosed by a bath with temperature @xmath12 ; and ( ii ) a state with initial value @xmath18 is described by eq.([sup1 ] ) .
the zero order density matrix is@xcite @xmath384 .
it is easy to check that @xmath385 , and the first order correction for off - diagonal element is@xmath386 in the basis \{@xmath18 } , the matrix element of the momentum flux operator ( [ dfu ] ) is@xmath387_{bc}=\langle\phi_{b}|\widehat{\pi}_{\alpha\beta}(\mathbf{r})|\phi_{c}\rangle=\frac{\hbar^{2}}{2m}n\int d\tau_{1 } \label{ymf}\]]@xmath388\]]@xmath389 according to standard statistical mechanics@xcite , the observed macroscopic momentum flux @xmath390 is@xmath391_{bc}\rho_{cb}^{(1)}(t ) .
\label{usa}\ ] ] substitute eq.([ymf ] ) into eq.([usa ] ) , we obtain eq.([mrf ] ) which was obtained by the microscopic response method .
the entropy @xmath392 of system @xmath5 is defined by@xcite@xmath393 where @xmath394 is the eigenfunction of @xmath6 .
the time dependence of entropy is included in the density operator @xmath395 .
the time rate of change of @xmath396 is written as@xmath397 let @xmath398 , then @xmath399 .
it is easy to check that@xmath400,\text { \ } n=1,2,3,\cdots . \label{na}\ ] ] making use of@xmath401 eq.([na ] ) leads to a well - known result@xcite @xmath402 . \label{ze}\ ] ] then , we combine liouville equation ( [ liu ] ) and eq.([ze ] ) , to obtain@xmath403 . \label{seq}\ ] ] by means of eq.([seq ] ) , eq.([sb ] ) becomes @xmath404|n\rangle=0 .
\label{zss}\ ] ] if a system is in good thermal contact with a reservoir such that the heat generated by the external field and velocity gradient can be instantaneously transferred to the surrounding bath , its temporal evolution can be described eq.([eu1 ] ) or eq.([liu ] ) .
the entropy @xmath392 of system @xmath5 is a constant .
the irreversibility ( entropy production ) of the momentum transport process is reflected in eqs.([entr],[tre ] ) .
eq.([zss ] ) does not depend on the initial condition for @xmath379 .
the key reasons are that ( 1 ) the system is described by a ` hamiltonian ' .
the density matrix satisfies a memoryless liouville equation ( [ liu ] ) ; ( 2 ) tr@xmath405tr@xmath406 is correct for two arbitrary operators @xmath407 and @xmath408 . for the same reasons , if the entropy of a microscopic canonical ensemble is defined by eq.([shs ] ) , one has also@xcite @xmath409 . removing irrelevant degrees of freedom can avoid this difficulty@xcite .
eqs.([liu],[ze],[seq ] ) have a common form@xmath410 , \label{ce}\ ] ] where @xmath411 , @xmath412 and @xmath413 .
one can solve eq.([ce ] ) with perturbation method@xcite@xmath414 where the superscript indicates the order in @xmath266 . before we introduce external field and velocity gradient ,
the system is in equilibrium .
the initial value of @xmath415 is its equilibrium value @xmath416 : @xmath417 and @xmath418 , @xmath419 . since @xmath416 is function of @xmath6 , the zero order solution of eq.([ce ] ) is@xcite@xmath420 the first order solution is@xcite@xmath421e^{i(t - t^{\prime})h_{0}/\hbar}. \label{1s}\ ] ] it is easy to check that:@xmath422 for any eigenstate @xmath371 of @xmath6 . for @xmath423 , @xmath424(t=-\infty)=\ln\rho^{eq}$ ] , so that @xmath425^{(0)}(t)=-\beta h_{0}-\ln z. \label{xz0}\ ] ] the first order solution is@xmath425^{(1)}(t)=\frac{i\beta}{\hbar}\int_{-\infty}^{t}dt^{\prime } e^{-i(t - t^{\prime})h_{0}/\hbar}[h_{1}(t^{\prime}),h_{0}]e^{i(t - t^{\prime } ) h_{0}/\hbar}. \label{1he}\ ] ] we recognize that @xmath426/\hbar$ ] is the dissipated power caused by the external field and velocity gradient . excepting a local non - uniformity of temperature which does not appear in the canonical ensemble , @xmath424^{(1)}(t)$ ] is mori s @[email protected]
obtain the term linear in velocity gradient @xmath81 , we only need @xmath428/\hbar$ ] . by means of eqs.([hs],[eu4 ] ) , the term linear in @xmath81 is@xmath429^{(1)}(t)|a\rangle =- p^{\phi_{a}}\beta\int_{-\infty}^{t}dt^{\prime}\sum_{\gamma\sigma\mu_{\sigma}\nu_{\sigma}}\lambda_{\mu _ { \sigma}\nu_{\sigma}}^{\gamma(\sigma)}\]]@xmath430@xmath431\cdot\nabla_{\mathbf{r}_{j}}\frac{\partial}{\partial r_{j\gamma}}\]]@xmath432\frac{\partial}{\partial r_{j\gamma}}+\frac{\hbar^{2}}{2m}\sum_{j}[\nabla_{\mathbf{r}_{j}}\frac{\partial\delta(\mathbf{r}_{j}-\mathbf{r})}{\partial r_{j\gamma}}]\cdot\nabla_{\mathbf{r}_{j}}\]]@xmath433\}e^{i(t - t^{\prime } ) h_{0}/\hbar}|a\rangle . \label{xz1}\ ] ] eq.([xz1 ] ) is @xmath434 multiplied by the dissipated energy from @xmath48 to @xmath20 .
eqs.([xz0],[xz1 ] ) forms a cumulant expansion for the density matrix @xmath435}$]:@xmath436^{(1)}(t)+\cdots\}.\label{czk}\ ] ] the operator in \ { } on rhs of eq.([xz1 ] ) is the divergence of the stress tensor ( momentum flux ) .
if we replace the momentum flux operator with its expectation value , and change the source of velocity gradient from boundary surfaces to bulk , eq.([czk ] ) is reduced to the non - equilibrium density matrix obtained by many different methods@xcite .
we need not invoke onsager s regression assumption@xcite and nonconservative force from bath@xcite .
according to the microscopic response method@xcite , the observed macroscopic momentum density eq.([ond ] ) in a many - body system can be obtained from the microscopic momentum density in a pure state by spatial coarse - graining and averaging over all possible initial conditions . if we adopt a no - slip boundary condition eq.([nsli ] ) , we can view a velocity gradient as a constraint on the many - body wave function of the system eq.([bc ] ) . the evolution equation for wave function eq.([eu1 ] ) is then derived from the lagrange multiplier method .
the lagrange multipliers have been obtained by the method of variation of constants , eqs.([li],[l1],[l2 ] ) .
they express the interactions on the system exerted by the moving boundary plate and three static plates . from the evolution equation eq.([eu1 ] ) , we calculated the time rate of change of mass density , momentum density and total energy in a pure state : eqs.([hmc0],[cond],[tre ] ) .
the dissipation caused by a velocity gradient contains @xmath171^{2}$ ] , in contrast to the dissipation caused by a mechanical disturbance@xcite which contains @xmath170 .
this is an obviously statistical character , velocity gradient is an internal disturbance .
although velocity gradient is an internal disturbance , by means of the conservation law of momentum in a pure state , we could still define momentum flux in a pure state eq.([flu ] ) .
the observed macroscopic momentum flux eq.([mdu ] ) is obtained by spatial coarse - graining the momentum flux in a pure state and averaging over all possible initial conditions .
the shear viscosity can be read out from eq.([mrf ] ) .
comparing to the traditional theories of viscosity , the present _ ansatz _ does not need temporal coarse graining . taking a spatial coarse - grained average over the microscopic response
automatically contains temporal coarse - graining and coarse - graining in the eigenvalue spectrum of collective variables .
we compared the hamiltonian derived from lagrange multiplier method with the phenomenological non - equilibrium hamiltonian .
the non - equilibrium density matrix implied by the present method can be reduced to those derived by other methods .
the present _ ansatz _ can be generalized to all internal disturbances .
we can view concentration gradient , temperature gradient and velocity gradient as constraints on the many - body wave function of system , and have an unified theory for diffusion , thermal conductivity and viscosity .
this work is in progress . | by viewing a velocity gradient in a fluid as an internal disturbance and treating it as a constraint on the wave function of a system , a linear evolution equation for the wave function is obtained from the lagrange multiplier method .
it allows us to define the microscopic response to a velocity gradient in a pure state .
taking a spatial coarse - graining average over this microscopic response and averaging it over admissible initial states , we achieve the observed macroscopic response and transport coefficient . in this scheme , temporal coarse - graining is not needed .
the dissipation caused by a velocity gradient depends on the square of initial occupation probability , whereas the dissipation caused by a mechanical perturbation depends on the initial occupation probability itself .
we apply the method of variation of constants to solve the time - dependent schrodinger equation with constraints .
the various time scales appearing in the momentum transport are estimated .
the relation between the present work and previous theories is discussed . |
a simple closed curve embedded into the euclidean @xmath7-space is called a _ knot_. a knot @xmath1 can be embedded many different ways in space . a _ stick knot _ is a knot which consists of finite line segments , called _
sticks_. one natural question concerning stick knots may be the _ stick number _ @xmath0 of a knot @xmath1 which is defined to be the minimal number of sticks necessary to construct this stick knot .
several upper and lower bounds on stick number for various classes of knots and links were founded .
the most general result is negami s inequality @xcite : @xmath8 for any nontrivial knot or link @xmath1 other than the hopf link , where @xmath2 is the minimal crossing number of @xmath1 .
calvo @xcite improved the lower bound to @xmath9 . recently
huh and oh @xcite utilized the arc index @xmath10 to determine more precise upper bound , showing that @xmath11 for any nontrivial knot @xmath1 ( especially @xmath12 for a non - alternating prime knot ) .
they mainly use the fact that @xmath13 for any nontrivial knot @xmath1 in @xcite and converted each arc presentation of @xmath1 into a stick knot by using @xmath14 sticks .
precise stick number for specific knots with small crossing number are known due to work by randell @xcite and meissen @xcite .
adams et al .
@xcite found precise stick numbers for an infinite class of knots , namely the @xmath15 torus knots and all their compositions . in independent work ,
jin @xcite determined stick numbers for a broader class of knots , showing that the stick number for any @xmath16 torus knot with @xmath17 is @xmath18 by using the superbridge index .
there is another way to obtain an upper bound of the stick number for particular classes of knots and links .
mccabe @xcite proved that @xmath19 for any @xmath5-bridge knot or link @xmath1 other than the unlink and the hopf link by realizing the stick knot visually with @xmath20 sticks . this upper bound is sharp for small knots and links of at most five crossings . in @xcite furstenberg et al .
reduced mccabe s upper bound by @xmath21 for a few classes of @xmath5-bridge knots or links , and proposed as an open question whether or not @xmath22 for all @xmath5-bridge knots and links with crossing number at least @xmath23 .
and meissen @xcite showed that this inequality holds for all knots and links with crossing number @xmath24 . in this paper
we give the answer .
[ thm : main ] @xmath25 for all @xmath5-bridge knots and links @xmath1 with @xmath26 .
we describe the standard projection of a @xmath5-bridge knot or link in terms of the conway notation which will be useful for the stick construction .
conway @xcite introduced the concept of tangles , portions of a knot contained in a topological sphere which intersect the knot exactly four times .
an _ integral tangle _ is made from two strands that wrap around each other , and identified by the number of half - twists ( i.e. crossings ) within it as in figure [ fig : conway](a ) .
more precisely the integer inside the circle is positive if it indicates the number of right - handed half - twists and negative if left - handed .
these integral tangles are connected together as in figure [ fig : conway](b ) to form a @xmath5-bridge knot or link which is represented by a _
conway notation _
note that if all @xmath28 are positive integers , then the positive and negative signs of integers in the figure appear alternately , so it gives a non - nugatory alternating projection of a @xmath5-bridge knot or link . as summarized in (
* section 2 ) , these are well known facts that any @xmath5-bridge knot or link can be represented by conway notation @xmath27 with positive integers @xmath28 and odd number @xmath29 due to work by burde and zieschang @xcite , and this non - nugatory alternating projection displays the minimal number of crossings due to kauffman @xcite , murasugi @xcite , and thistlethwaite @xcite . to simplify the cases of the main proof
we may assume that @xmath30 . for ,
when @xmath31 and @xmath32 , we use the conway notation @xmath33 instead of the original one , both of which indicate the same knot or link . and when @xmath34 , we can use its mirror image @xmath35 which has eventually the same crossing number and stick number as the original one . in the rest of this section ,
we illustrate how to construct integral @xmath36-tangles with @xmath37 by using @xmath38 sticks .
now observe the @xmath39-tangle which consists of right - handed @xmath39 half - twists as in figure [ fig : tangle](a ) . in this construction , this tangle with @xmath39 even has one vertical stick , called a _ core stick _ , and the remaining @xmath39 sticks wrapping around the core stick .
and the tangle with @xmath39 odd has two sticks which are almost parallel and very close to each other at their one endpoints so that they look like two slightly perturbed broken pieces of the core stick , and @xmath40 sticks wrapping around these two sticks as in the figure .
both tangles have right - handed @xmath39 half - twists respectively , and so @xmath39 crossings .
the right figure shows the drawing of the tangle with @xmath39 even in the view point from the top .
the @xmath41-tangle is simply the mirror image of the positive one .
we use symbols with circles and numbers as drawn in figure [ fig : tangle](b ) to indicate these tangles .
let @xmath1 be any @xmath5-bridge knot or link with @xmath26 . recall that @xmath1 has a conway notation @xmath27 where @xmath29 is odd and all @xmath28 are positive integers with @xmath30 . here
. first we construct @xmath1 of a conway notation @xmath43 when @xmath44 by using @xmath45 sticks .
@xmath1 can be obtained from the @xmath46-tangle by closing off the four ends in the standard way .
as drawn in figure [ fig : p ] , we utilize the @xmath47-tangle consisting of @xmath48 sticks and add the remaining @xmath49 sticks to make @xmath50 more left - handed half - twists as needed .
figure [ fig:6 ] shows a specific realization of the link @xmath51 which is the union of two unknotted circles both of which consist of four sticks .
the coordinates of vertices are @xmath52 and @xmath53 .
following the tangle construction in the previous section we can replace the @xmath54-tangle by the @xmath47-tangle .
hence this specific example verifies that our scheme is also realizable for any case @xmath55 .
now we construct @xmath1 of a conway notation @xmath56 for positive integers @xmath57 with @xmath58 by using @xmath59 sticks .
assume that @xmath60 and @xmath61 for the general case .
as in the first link in figure [ fig : pqr ] , we rotate the @xmath46-tangle part around a horizontal line for untwisting one half - twist of the @xmath62-tangle . since all three integers @xmath63 , and @xmath64 are at least @xmath5 , we can construct these @xmath46-tangle , @xmath65-tangle and @xmath66-tangle by using @xmath59 sticks in total as described in the last paragraph of section [ sec : conway ] .
now connect them together to build a knot or link @xmath56 as drawn in the figure so that two pairs of sticks are joined into two sticks and two new sticks are added for closing off .
thus the total number of sticks are unchanged as desired .
easy calculation tells us that there are five more non - general cases as follows ; @xmath67 with @xmath60 and @xmath58 , @xmath68 with @xmath60 and @xmath69 , @xmath70 with @xmath61 and @xmath58 , @xmath71 with @xmath69 , and @xmath72 with @xmath73 . note that the case @xmath68 with @xmath74 and @xmath58 has the same result as @xmath68 with @xmath60 and @xmath69 .
similar constructions to the general case can be applied for these five cases .
the results of all the cases are drawn in figure [ fig : pqr ] .
one can see that each has exactly @xmath59 sticks . to guarantee that all these six cases can be realized in @xmath7-space
as intended , we provide the vertex - coordinates of the simplest case of each , such as @xmath75 , @xmath76 , @xmath77 , @xmath78 , @xmath79 and @xmath80 in the following table .
note that each of the cases @xmath75 , @xmath77 and @xmath78 consists of two circles , so they need two sets of vertices indicating two circles respectively .
+ + + finally we construct @xmath1 of a conway notation @xmath27 for the cases of @xmath81 by using @xmath82 sticks , that is @xmath83 sticks .
first we consider the general cases that all @xmath28 are greater than @xmath21 .
see figure [ fig : general ] .
we construct all @xmath84-tangles individually by using @xmath85 sticks in total .
now connect them together to build the knot or link as drawn in the figure .
in this scheme , all @xmath28-tangles with @xmath86 even share just one core stick , and each pair of @xmath84-tangle and @xmath87-tangle for @xmath88 share exactly one stick . on the other hand we have to add some extra sticks to connect each pair of @xmath89-tangle and @xmath90-tangle for @xmath91 , and two more sticks to connect the left - most core stick to @xmath92-tangle and @xmath93-tangle .
thus the total number of sticks is @xmath94 that is @xmath82 .
now we illustrate how to build up a stick presentation of a special case where all @xmath28 s are @xmath5 in three stages as in figure [ fig : realization ] .
note that one can easily apply this construction to general cases where some of @xmath28 s are greater than @xmath5 . at the first stage of this construction ,
figure [ fig : realization](a ) shows how to build the main frame of this stick presentation .
starting from the origin point @xmath95 in @xmath96 , take a long vertical stick along @xmath97-axis , and more sticks wrapping around this vertical stick from top to bottom so that these non - vertical sticks rotate clockwise in the view point from the top .
let @xmath98 be the terminal point of this part .
the intermediate point @xmath99 in the figure is placed on the @xmath100-plane so that its @xmath101-coordinate is positive . in figure
[ fig : realization](b ) we see how a connected series of sticks runs back and forth between the sticks constructed already .
the two end points of this connected series of sticks are denoted by @xmath102 and @xmath103 . finally as in figure
[ fig : realization](c ) we connect @xmath102 and @xmath103 by two sticks which share one end point @xmath104 .
obviously the vertex @xmath104 must lie far below the @xmath100-plane .
also we connect @xmath105 and @xmath98 by two sticks sharing one end point @xmath106 which lies farther below than the the vertex @xmath104 with respect to the @xmath107-coordinate .
now we handle the cases when @xmath28 s are @xmath21 .
we distinguish six cases .
the first three cases are illustrated in figure [ fig:1 ] and the other cases are in figure [ fig:11 ] .
first consider the case @xmath108 for some @xmath86 among odd integers @xmath109 .
replace the subgraph of the main graph between two horizontal lines by the graph indicating @xmath108 drawn on the right top in figure [ fig:1 ] .
next consider the case @xmath108 for some @xmath86 among even integers @xmath110 .
replace the related subgraph by the graph drawn on the right middle . in this case
the main frame will be bent slightly . for the case
@xmath111 , replace the related subgraph by the graph on the right bottom .
note that the stick @xmath112 does not pass the other sticks because the vertex @xmath104 lies far below the @xmath100-plane .
now consider the next three cases where both or either @xmath113 or @xmath114 is @xmath21 .
figure [ fig:11](a ) illustrates how to realize our scheme for the case @xmath115 . at the top of the main graph
we remove the dotted line segments and connect the two pairs of end points properly .
note again that the stick @xmath112 does not pass the other sticks .
the last two cases when only @xmath116 and only @xmath117 can be constructed following the procedure in figure [ fig:11](b ) and ( c ) respectively . abgw colin c. adams , bevin m. brennan , deborah l. greilsheimer and alexander k. woo , _ stick numbers and composition of knots and links _ , j. knot theory ramif .
* 6 * ( 1997 )
149161 . y. bae and c. park , _ an upper bound of arc index of links _ , math
* 129 * ( 2000 ) 491500 .
g. burde and h. zieschang , _ knots _ , walter de gruyter & co. ( 1985 )
. j. calvo , _ characterizing polygons in @xmath96 _ , in physical knots , contemporary mathematics * 304 * ( 2002 ) 3753 .
j. conway , _ an enumeration of knots and links , and some of their algebraic properties _
, in computational problems in abstract algebra , pergamon press , new york ( 1970 ) 329358 .
e. furstenberg , j. li and j. schneider , _ stick knots _ , chaos , solitons & fractals * 9 * ( 1998 ) 561568 . y. huh and s. oh , _ an upper bound on stick numbers of knots _ , to appear in j. knot theory ramif . g. t. jin , _ polygon indices and superbridge indices of torus knots and links _ , j. knot theory ramif .
* 6 * ( 1997 ) 281289 .
l. kauffman , _ state models and the jones polynomial _
, topology * 26 * ( 1987 ) 395407 .
l. mccabe , _ an upper bound on edge numbers of 2-bridge knots and links _ , j. knot theory ramif .
* 7 * ( 1998 ) 797805 . m. meissen , homepage at http://www.bethelks.edu/meissen . k. murasugi , _ jones polynomials and classical conjectures in knot theory _ ,
topology * 26 * ( 1987 ) 187194 .
s. negami , _ ramsey theorems for knots , links , and spatial graphs _ , trans .
* 324 * ( 1991 ) 527541 .
r. randell , _ an elementary invariant of knots _ , j. knot theory ramif . * 3 * ( 1994 ) 279286 .
m. thistlethwaite , _ a spanning tree expansion of the jones polynomial _ , topology * 26 * ( 1987 ) 297309 . | negami found an upper bound on the stick number @xmath0 of a nontrivial knot @xmath1 in terms of the minimal crossing number @xmath2 of the knot which is @xmath3 .
furthermore mccabe proved @xmath4 for a @xmath5-bridge knot or link , except in the case of the unlink and the hopf link . in this paper
we construct any @xmath5-bridge knot or link @xmath1 of at least six crossings by using only @xmath6 straight sticks .
this gives a new upper bound on stick numbers of @xmath5-bridge knots and links in terms of crossing numbers . |
two important issues of lepton number violation are intimately related to each other .
one is the possible existence of neutrino majorana masses , as evidenced by the ongoing excitement generated by the recent report of atmospheric neutrino oscillations @xcite , as well as previous other indications of solar @xcite and accelerator @xcite neutrino oscillations .
the other is one of the very challenging question in cosmology to generate the baryon asymmetry of the universe starting from a symmetric universe @xcite .
since the electroweak anomalous processes breaks both the baryon and the lepton numbers , still conserving the @xmath0 quantum number , the baryon asymmetry of the universe is no longer independent of the lepton number violation of the universe @xcite .
if there is very fast lepton number violation before the electroweak phase transition , then that can erase the @xmath0 asymmetry of the universe @xcite and hence the baryon asymmetry of the universe . on the other hand , if any lepton asymmetry is generated at some high temperature , that can get converted to a baryon asymmetry of the universe before and during the electroweak phase transition @xcite .
lepton number violation is required to give a majorana mass to the neutrinos .
depending on the scale at which this lepton number is violated , this interaction may or may not satisfy the out - of - equilibrium condition .
if this interaction is faster than the expansion rate of the universe , it can erase all lepton asymmetry of the universe before the electroweak phase transition . in those models
one then require additional inputs to explain the baryon asymmetry of the universe . on the other hand , in models of leptogenesis the lepton number violating interaction required to give majorana masses to the neutrino also satisfy the out - of - equilibrium condition .
if there is enough cp violation in the leptonic sector @xcite , then this can generate a @xmath0 asymmetry of the universe . the anomalous baryon number violation in the presence of the sphalerons
will then convert this @xmath0 asymmetry to a baryon asymmetry of the universe before the electroweak phase transition @xcite .
it is also possible to generate a baryon asymmetry of the universe during the electroweak phase transition @xcite , but the condition that this asymmetry will not be erased after the electroweak phase transition gives a strong bound on the mass of the higgs @xcite , which makes these less likely . as a result leptogenesis appears to be the most attractive scenario for generating a baryon asymmetry of the universe at present . in this article
we shall thus summarise the possibility of leptogenesis in different models of neutrino masses .
we shall first review the original idea of baryogenesis in the context of grand unified theories and show why @xmath0 is always conserved in the baryon asymmetry thus generated .
then we show the relationship between the baryon and lepton number in the presence of the sphaleron processes and how @xmath1 is washed out .
this also implies constraints on the lepton number violation and hence on the neutrino masses . at the end
we discuss different classes of models of neutrino masses which can naturally accomodate leptogenesis and then summarise .
the subject of baryogenesis originated when sakharov @xcite pointed out that for the generation of a baryon asymmetry of the universe we need three conditions * _ baryon number violation _
, * _ @xmath2 and @xmath3 violation _ , and * _ departure from thermal equilibrium_. it was then realised that grand unified theories ( guts ) satisfies all these criterion @xcite .
the quark - lepton unification implies baryon number violation in guts .
since fermions belong to chiral representation , @xmath2 is maximally violated .
departure from thermal equilibrium was also naturally satisfied in these models since the scale of unification is sufficiently high , and the universe was expanding very fast in that epoch .
so , any reasonble gut coupling would imply departure from equilibrium .
violation of @xmath3 was then the only crucial point , which had to be incorporated in these theories .
however , it was not difficult to consider some of the couplings to be complex so that there exist tree level and one loop diagrams which could interfere to give us enough baryon asymmetry in the decays of the heavy gauge and higgs bosons @xcite .
this was considered to be one of the major successes of guts that it can explain the baryon asymmetry of the universe .
after several years it was realised that the chiral nature of the weak interaction also breaks the global baryon and lepton numbers in the standard model @xcite .
although both @xmath4 and @xmath5 are broken , a combination @xmath0 remains invariant since the baryon and lepton number anomalies happens to be the same in the standard model .
since these classical global @xmath1 number symmetry is broken by quantum effects due to the presence of the anomaly , these processes were found to be very weak at the zero temperature .
but at finite temperature these @xmath1 number violating interactions were found to be very strong in the presence of some static topological field configuration - sphalerons @xcite .
in fact , during the period @xmath6 these interactions are so strong that in no time the particles and anti - particles attain their equilibrium distributions . as a result ,
since @xmath7 is conserved and hence the masses of the particles and anti - particles are same , the number density of baryons becomes same as that of the anti - baryons and that will wash out any primordial @xmath1 asymmetry of the universe .
we shall now discuss why gut baryogenesis always generated only @xmath1 asymmetry @xcite , which would be erased by the sphaleron transitions . in specific gut scenarios such as @xmath8 and @xmath9 ,
@xmath10 is either a global or a local symmetry respectively .
hence the asymmetry generated by the above mechanism is @xmath10 conserving @xcite .
when the scalar or vector bosons decay only into fermions , any attempt to generate a @xmath11 asymmetry leads to its large suppression in all these models . we shall now prove that if the decay products are sm fermions only , this is in fact a generic property of any baryon asymmetry generated by the above described mechanism .
this follows from an operator analysis analogous to the one used to show that the minimal scenarios of proton decay conserve @xmath0 @xcite . for definitness
we consider scalars @xmath12 and @xmath13 , but obviously the result generalizes also to vectors .
baryogenesis is possible in guts because there exist new gauge and higgs bosons , whose decays violate baryon number . when these heavy particles ( say @xmath12 ) decay into two quarks and into a quark and an antilepton , the baryon and lepton numbers are broken @xcite . for @xmath3 violation
this mechanism requires two heavy gauge or higgs bosons , @xmath12 and @xmath13 , each of which should have two decay modes , @xmath14 so that there exist one - loop vertex corrections to these decays .
the required @xmath3 violation occurs due to the interference between tree and loop diagrams . as required by the out - of - equilibrium condition
, masses of these particles must satisfy @xmath15 where , @xmath16 is the decay rate of the heavy particle @xmath12 ; @xmath17 is the hubble constant ; @xmath18 is the effective number of massless degrees of freedom ; and @xmath19 is the planck scale .
let us start from the lagrangian giving the decays of @xmath12 and @xmath13 , @xmath20 where @xmath21 denote any sm fermion . to obtain a nonzero @xmath3 violation from the interference between tree and vertex diagrams , we require @xmath12 and @xmath13 to be distinct from each other and to have different decay modes .
one can then write down all possible combinations of @xmath22 , @xmath4 , @xmath2 , and @xmath23 , with @xmath12 and @xmath13 , and find out the decay modes of @xmath12 and @xmath13 .
since the out - of - equilibrium condition and the nonvanishing of the absorptive part of the loop integral require these scalars @xmath12 and @xmath13 to be much heavier than the fermions , we can integrate them out and write down the diagrams in terms of the four - fermion effective operators of the sm , as shown in fig . 1 .
one can in principle also have the self - energy - type diagrams with the fermions in the loop for generating the @xmath3 asymmetry . in this case , after integrating out the heavy scalars , the effective diagrams in terms of the four - fermion operators are exactly the same as in the vertex - correction case , so the conclusions will not be changed .
3.75in.1 in -.2 in this simple but crucial step allows us to use existing knowledge on sm four - fermion operators for baryon number violation which have been studied extensively in the literature @xcite .
it was found that all these operators conserve @xmath0 to the lowest order .
any @xmath0 violating operator will be suppressed by @xmath24 compared to the @xmath1 violating operators . in models with an intermediate symmetry breaking scale or with new higgs scalars at some intermediate scales , this suppression factor may be softened a little , but still strong enough to rule out any possibility of generating enough @xmath0 asymmetry of the universe . on the other hand , any four - fermion operator which violates only lepton
number requires all the fermions to be the same ; hence it can not generate the required @xmath3 asymmetry .
therefore a @xmath0 asymmetry , needed to survive the sphaleron processes , is impossible to generate with the sm four - fermion operators .
anomaly breaks any classical symmetry of the lagrangian at the quantum level .
so , all local gauge theories should be free of anomalies
. however , there may be anomalies corresponding to any global current , which means that such global symmetries of the classical lagrangian are broken through quantum effects . in the standard model
the baryon and lepton number global symmetries are anomalous @xcite @xmath25\ ] ] which will break the @xmath1 symmetry . however , the anomaly corresponding to the baryon and lepton numbers are same and as a result there is no anomaly for the @xmath0 charge .
because of the anomaly @xcite , @xmath1 is broken during the electroweak phase transition , but their rate is very small at zero temperature , since they are suppressed by quantum tunnelling probability , @xmath26 $ ] . at finite temperature , this @xmath1 number violation becomes very fast in the presence of a non - trivial static topological soliton configuration , called the sphalerons @xcite , and the quantum tunnelling suppression factor is now replaced by the boltzmann factor @xmath27 $ ] where the potential or the free energy @xmath28 is related to the mass of the sphaleron field . as a result , at temperatures between @xmath29
the sphaleron mediated baryon and lepton number violating processes are in equilibrium .
for the simplest scenario of @xmath30 , the sphaleron induced processes are @xmath31 , given by , @xmath32 .
\label{sph}\ ] ] it can be shown that any @xmath0 asymmetry before the electroweak phase transition will get converted to a baryon and lepton asymmetry of the universe , which can be seen from an analysis of the chemical potential @xcite . above the electroweak scale , all the particles
could be assumed to be ultrarelativistic .
the particle asymmetry , _
i.e. _ the difference between the number of particles ( @xmath33 ) and the number of antiparticles ( @xmath34 ) can be given in terms of the chemical potential of the particle species @xmath35 ( for antiparticles the chemical potential is @xmath36 ) as @xmath37 where @xmath38 for bosons and @xmath39 for fermions .
in the standard model there are quarks and leptons @xmath40 and @xmath41 ; where , @xmath42 corresponds to three generations .
in addition ,
the scalar sector consists of the usual higgs doublet @xmath43 , which breaks the electroweak gauge symmetry @xmath44 down to @xmath45 . in table 1
, we presented the relevant interactions and the corresponding relations between the chemical potentials . in the third column
we give the chemical potential which we eliminate using the given relation .
we start with chemical potentials of all the quarks ( @xmath46 ) ; leptons ( @xmath47 , where @xmath48 ) ; gauge bosons ( @xmath49 for @xmath50 , and 0 for all others ) ; and the higgs scalars ( @xmath51 ) .
.relations among the chemical potentials [ cols="^,^,^",options="header " , ] the chemical potentials of the neutrinos always enter as a sum and for that reason we can consider it as one parameter .
we can then express all the chemical potentials in terms of the following independent chemical potentials only , @xmath52 . we can further eliminate one of these four potentials by making use of the relation given by the sphaleron processes , @xmath53 .
we then express the baryon number , lepton numbers and the electric charge and the hypercharge number densities in terms of these independent chemical potentials , @xmath54 where @xmath55 is the number of higgs doublets @xmath43 . at temperatures above the electroweak phase transition , @xmath56 ,
both @xmath57 and @xmath58 must vanish , while below the critical temperature @xmath57 should vanish , but since @xmath59 is now broken we can consider @xmath60 and @xmath61 .
these conditions and the sphaleron induced @xmath62 conserving , @xmath63 violating condition will allow us to write down the baryon asymmetry in terms of the @xmath62 number density as , @xmath64 thus the baryon asymmetry of the universe after the electroweak phase transition will depend only on the primordial @xmath0 asymmetry of the universe , while all the primordial @xmath1 asymmetry will be washed out . before proceeding further , we shall briefly discuss what do we mean when we say that some interaction is fast and that will erase some asymmetry @xcite . in equilibrium
the number density of particles with non - zero charge @xmath65 would be same as the antiparticle number density since the expectation value of the conserved charge vanishes .
a mathematical formulation of this statement reads that the expectation value of any conserved charge @xmath65 is given by , @xmath66}{{\rm tr } \left [ e^{-\beta h}\right]}\ ] ] and since any conserved charge @xmath65 is odd while @xmath17 is even under @xmath67 transformation this expectation value vanishes .
so for the generation of the baryon asymmetry of the universe we have to circumvent this theorem either by including nonzero chemical potential , or go away from equilibrium or violate @xmath67 . in
most of the popular models @xmath67 conservation is assumed and one starts with vanishing chemical potential for all the fields which ensures that the entropy is maximum in chemical equilibrium . then to generate the baryon asymmetry of the universe one needs to satisfy the out - of - equilibrium condition @xcite .
the requirement for the out - of - equilibrium condition may also be stated in a different way @xcite .
if we assume that the chemical potential associated with @xmath4 is zero and @xmath67 is conserved , then in thermal equilibrium the phase space density of baryons and antibaryons , given by @xmath68^{-1}$ ] are identical and hence there can not be any baryon asymmetry .
whether a system is is equilibrium or not can be understood by solving the boltzmann equations . but a crude way to put the out - of - equilibrium condition is to say that the universe expands faster than some interaction rate . for example , if some b - violating interaction is slower than the expansion rate of the universe , this interaction may not bring the distribution of baryons and antibaryons of the universe in equilibrium .
in other words , before the chemical potentials of the two states gets equal , they move apart from each other .
thus we may state the out - of - equilibrium condition as @xmath69 where , @xmath70 is the interaction rate under discussion , @xmath18 is the effective number of degrees of freedom available at that temperature @xmath71 , and @xmath19 is the planck scale .
in the standard model there is no lepton number violation .
however , one can consider a higher dimensional effective operator which violates @xmath0 , given by @xmath72 there is no origin of such interactions within the standard model .
so one expects that some new interaction at some high energy will give us this effective interaction at low energy .
the scale of the new interaction @xmath73 , which is also the scale of lepton number ( and also @xmath0 number ) violation , will determine if this interaction is fast enough to erase all primordial @xmath0 asymmetry . since during the same time @xmath1 asymmetry is also washed out by the sphaleron transitions , there will not be any residual baryon asymmetry of the universe after the electroweak phase transition . as a result
, the survival of the baryon asymmetry of the universe will then require this interaction to be slower than the expansion rate of the universe , @xmath74 which gives a bound @xcite on the lepton number violating scale to be , @xmath75 .
when the higgs doublets @xmath43 acquires a @xmath76 , the higher dimensional operator will induce a majorana mass of the left - handed neutrinos .
this bound on the heavy scale @xmath73 will then imply a bound on the mass of the left - handed neutrinos , @xmath77 3.5in.6 in -2 in it is also possible to give a bound on the neutrino mass in a more general way @xcite . unless the neutrinos are dirac particles @xcite , during the electroweak phase transition there will be interactions of the type , @xmath78 which violate lepton number .
these interactions are mediated by a virtual left - handed neutrino exchange as shown in figure 2 . here
@xmath79 and @xmath80 are the generation indices .
depending on the physical mass of the left - handed majorana neutrinos these processes can wash out any baryon asymmetry between the time when the higgs acquires a @xmath76 and the @xmath81 freeze out , _
i.e. _ , between the energy scales 250 gev and 80 gev .
the condition that these processes will be slower than the expansion rate of the universe , @xmath82 gives a bound on the majorana mass of the left - handed neutrinos to be , @xmath83 this bound is on each and every element of the mass matrix and not on the physical states .
there are other lepton number violating interactions like the scattering processes @xmath84 ( mediated by a virtual left - handed neutrino ) and decays of @xmath85 and the higgs @xmath43 , which also give similar bounds on the left - handed neutrino mass
. in some specific models one may give stronger bounds on the mass of the neutrinos @xcite . in models with right
handed neutrinos ( @xmath86 ) , the neutrino masses comes from the see - saw mechanism @xcite .
the lagrangian for the lepton sector containing the mass terms of the singlet right handed neutrinos @xmath87 and the yukawa couplings of these fields with the light leptons is given by , @xmath88 where @xmath43 is the usual higgs doublet of the standard model ; @xmath89 are the light leptons , @xmath90 are the complex yukawa couplings and @xmath91 is the generation index . without loss of generality we work in a basis in which the majorana mass matrix of the right handed neutrinos is real and diagonal with eigenvalues @xmath92 . once the higgs doublet @xmath43 acquires a @xmath76 , the masses of the neutrinos in the basis @xmath93 $ ] is given by , @xmath94 where , @xmath95 and @xmath96 are @xmath97 matrices . in the limit when all eigenvalues of @xmath73 are much heavier than those of @xmath55 , and the matrix @xmath73 is not singular , this matrix may be block diagonalised .
it then gives three heavy right handed majorana neutrinos with masses @xmath98 and the majorana mass matrix of the left - handed neutrinos will be given by , @xmath99 in this scenario the see - saw masses of the left - handed neutrinos explain naturally why they are so light .
the decay of @xmath100 into a lepton and an antilepton , @xmath101 breaks lepton number .
since the lightest of the right handed neutrinos ( say @xmath102 ) will decay at the end , this interaction ( @xmath102 decay ) should be slow enough so as not to erase the baryon asymmetry of the universe , which now implies @xmath103 which can then give a very strong bound @xcite on the mass of the lightgest of the left - handed neutrinos to be @xmath104 in models @xcite , where the left - handed neutrino mass is not related to any heavy neutrinos through see - saw mechanism , the abovementioned bounds may not be valid .
in addition , there are several specific cases even within the framework of see - saw models ( like the singular see - saw mechanism where @xmath105 ) , where these bounds are not applicable .
these bounds are also not valid if some global u(1 ) symmetry is exactly conserved up to an electroweak anomaly @xcite .
furthermore , in some very specific models where a baryon asymmetry of the universe is generated after the electroweak phase transition @xcite , or there are some extra baryon number carrying singlets which decays after the electroweak phase transition @xcite , it is possible to avoid all the bounds from constraints of survival of the baryon asymmetry of the universe .
we shall now discuss similar bounds on the supersymmetric r - parity violating and zee type radiative models .
although the earlier bounds on the see - saw mechanism is not applicable when the decays of the right handed neutrinos generate a lepton asymmetry of the universe , since leptogenesis is not possible in these r - parity breaking models or the zee - type models , in these models one needs additional inputs to generate a baryon asymmetry of the universe . in the r - parity violating models ,
the unavoidable lepton number violation at the supersymmetry breaking scale will erase any primordial @xmath4 or @xmath5 or @xmath106 asymmetry @xcite .
this is so unless @xmath107 is conserved @xcite even after the electroweak phase transition .
this has been pointed out earlier from a general dimensional analysis , but none of the existing models of neutrino masses through r - parity violation could accomodate this symmetry since that can not allow required neutrino mixing matrix .
possible solutions to this problem could be to break r- parity spontaneously after the electroweak symmetry breaking @xcite , or to generate a baryon asymmetry of the universe in r - parity breaking scenarios @xcite , or generate a baryon asymmetry of the universe during the electroweak phase transition @xcite .
but these models are incapable of accomodating the interesting feature of leptogenesis , namely generating a baryon asymmetry of the universe from the interaction which gives a neutrino mass .
similarly the zee - type models @xcite considered so far can not account for the observed baryon asymmetry of the universe , if they have to explain the present neutrino mass spectrum .
although the radiative models have the advantage that they can reproduce the required maximal mixing naturally , they erase any primordial lepton asymmetry of the universe and hence the baryon asymmetry of the universe .
this severe constraint on the zee - type models are also valid in both supersymmetric and non - supersymmetric scenarios .
it is not impossible to find an alternative where a baryon asymmetry of the universe is generated after it has been washed out by this interaction , but that will not be related to the neutrino mass . in the mssm
, r - parity of a particle is defined as @xmath108 where @xmath4 is its baryon number , @xmath5 its lepton number , and @xmath109 its spin angular momentum . hence the sm particles have @xmath110 and their supersymmetric partners have @xmath111 . using the common notation where all chiral superfields are considered left - handed , the three families of leptons and quarks are given by @xmath112 @xmath113 where @xmath79 is the family index , and the two higgs doublets are given by @xmath114 where the @xmath115 content of each superfield is also indicated .
if r - parity is conserved , the superpotential is restricted to have only the terms @xmath116 if r - parity is violated but not baryon number , then the superpotential contains the additional terms @xmath117 resulting in nonzero neutrino masses either from mixing with the neutralino mass matrix @xcite or in one - loop order @xcite .
if lepton - number violating interactions such as @xmath118 are in equilibrium in the early universe , any pre - existing lepton asymmetry would be erased . to make sure that this does not happen , the following condition has to be satisfied : @xmath119 assuming that the supersymmetry breaking scale @xmath120 is @xmath121 gev , we find @xmath122 which is very much below the typical minimum value of @xmath123 needed for radiative neutrino masses @xcite .
a similar bound was presented from dimensional arguments @xcite .
larger values of @xmath124 are allowed if there is a conserved @xmath125 symmetry @xcite
. however , there would be other severe phenomenological restrictions in that case @xcite .
this bound can not be evaded even if one uses the bilinear term for neutrino masses instead , because the induced mixing would introduce trilinear couplings which violate lepton number and an effective @xmath124 is unavoidable .
this means that although r - parity violation may exist , it will have very little consequences .
in particular , it will not contribute significantly to neutrino masses . in models of radiative neutrino masses
@xcite , in addition to the suppression due to the @xmath126 factor of each loop , there is often another source of suppression due to the yukawa couplings involved . in the original zee model ,
the sm is extended to include a charged scalar @xmath127 and a second higgs doublet .
the relevant terms of the interaction lagrangian are given by @xmath128 where two higgs doublets are needed or else there would be no @xmath129 coupling .
lepton number is violated in the above by two units , hence we expect the realization of an effective dimension - five operator @xmath130 for naturally small majorana neutrino masses @xcite .
this occurs here in one loop and the elements of the @xmath97 neutrino mass matrix are given by @xmath131 where @xmath132 and @xmath133 are the charged - lepton masses which come from @xmath134 but not @xmath135 .
the function @xmath136 is given by @xmath137 since the @xmath138 terms in eq .
( 11 ) are likely to be dominant , this model has two nearly mass - degenerate neutrinos which mix maximally .
this is very suitable for explaining the atmospheric neutrino data @xcite , but only in conjunction with the lsnd data @xcite .
let @xmath139 tev , @xmath140 gev , @xmath141 gev , @xmath142 , and @xmath143 to satisfy eq .
( 9 ) , then the @xmath138 terms generate a neutrino mass of 0.0013 ev , which is very much below the necessary 1 ev or so indicated by the lsnd data .
we note that eq .
( 8) constrains the combination @xmath144 , whereas @xmath145 goes like @xmath146 .
hence neutrino masses would only decrease if we increase @xmath147 .
as long as there is a suppression from @xmath138 ( which comes of course from the yukawa coupling @xmath148 ) , the conflict with leptogenesis is a real problem .
in the standard model neutrinos are massless . to make them massive , there exist four generic mechanisms , namely , the see - saw mechanism @xcite , the triplet higgs mechanism @xcite , the radiative mass generation @xcite and through r - parity violation @xcite .
all these models require lepton number violation , which can erase the primordial baryon asymmetry of the universe .
so , the most promising scenario will be to see if this lepton number violation could be used to generate a baryon asymmetry of the universe .
this is done in models of leptogenesis @xcite .
the see - saw mechanism and the triplet higgs mechanism for neutrino masses are the two mechanisms , which can accomodate leptogenesis in the minimal models , which we shall summarise next .
although one can extend the zee - type models also or have a complicated scenario of r - parity breaking models for generating a lepton asymmetry of the universe @xcite , we shall not discuss them in this article . to give a small majorana mass to the left - handed neutrino through a see - saw mechanism , right - handed neutrinos were introduced ( @xmath86 ) . in these models
neutrino masses come from the see - saw mechanism @xcite .
the lagrangian for the lepton sector containing the mass terms of the singlet right handed neutrinos @xmath87 and the yukawa couplings of these fields with the light leptons is given by eqn ( [ n ] ) . without loss of generality we work in a basis in which the majorana mass matrix of the right handed neutrinos is real and diagonal with eigenvalues @xmath92 , and assume @xmath149 .
because of the majorana mass term , the decay of @xmath100 into a lepton and an antilepton , breaks lepton number .
there are two sources of cp violation in this scenario : * vertex type one loop diagrams which interferes with the tree level diagram given by figure 3 .
this is similar to the @xmath3 violation coming from the penguin diagram in @xmath150decays .
* self energy type one loop diagrams could interfere with the tree level diagrams to produce cp violation as shown in figure 4 .
this is similar to the @xmath3 violation in @xmath151 oscillation , entering in the mass matrix of the heavy majorana neutrinos .
2.25in- .3 in 2.25in- .3 in in the first paper on leptogenesis @xcite , the vertex type diagram was only mentioned .
subsequently , it has been extensively studied @xcite and the amount of @xmath3 asymmetry is calculated to be , @xmath152 } { \sum_\alpha main contribution to the asymmetry comes from the lightest right handed neutrino ( @xmath102 ) decay , when the other heavy neutrinos have already decayed away . initially the self energy diagram was considered for @xmath3 violation as an additional contribution @xcite .
it was then pointed out @xcite that this @xmath3 violation enters in the mass matrix as in the @xmath151 oscillation .
before they decay , the right handed neutrinos were considered to oscillate to an anti - neutrino and since the rate of @xmath153 , an asymmetry in the right handed neutrino was obtained before they decay @xcite . as a result , when the two heavy right handed neutrinos are almost degenerate , _
i.e. _ , the mass difference is comparable to their width , there may be a resonance effect which can enhance the @xmath3 asymmetry by few orders of magnitude @xcite .
this effect was then confirmed by other calculations @xcite , one of which @xcite uses a field - theoretic resummation approach @xcite used earlier to treat unstable intermediate states . for large mass difference
the amount of cp asmmetry from the self energy contribution becomes equal to the vertex correction , which has to be added to get the final asymmetry .
although the @xmath3 asymmetry was found to be non - vanishing , in thermal equilibrium unitarity and @xmath7 would mean that there is no asymmetry in the final decay product .
however , when the out - of - equilibrium condition of the heavy neutrinos decay is considered properly , one could get an asymmetry as expected .
consider the decays of @xmath154 and @xmath155 . if they were generated in the early universe , in a short time scale @xmath155 could decay and recombine , but @xmath154 may not be able to decay or recombine . as a result in the decay product
there will be an asymmetry in @xmath156 and @xmath157 if there is @xmath3 violation . in the lepton number violating two body scattering processes
@xmath3 violation in the real intermediate state plays the most crucial role , which comes since the decay take place away from thermal equilibrium .
in the case of right handed neutrino decay , the asymmetry is generated when the lightest one ( say @xmath102 ) decay . before its decay ,
the pre - existing lepton asymmetry is washed out by its lepton number violating interactions .
so the out - of - equilibrium condition now implies that the lightest right - handed neutrino should satisfy the out - of - equilibrium condition when it decays , which is given by , @xmath158 which gives a bound on the mass of the lightest right - handed neutrino to be @xmath159 finally the lepton asymmetry and hence a @xmath0 asymmetry generated at this scale gets converted to a baryon asymmetry of the universe in the presence of sphaleron induced processes . to give a majorana mass to the neutrino
, one can either introduce a right handed neutrino as in the see - saw mechanism , or else one can introduce two complex @xmath59 triplet higgs @xcite scalars ( @xmath160 ) .
the @xmath76s of the triplet higgses can give small majorana masses to the neutrinos through the interaction @xmath161 + h.c.\ ] ] if the triplet higgs acquires a @xmath76 and break lepton number spontaneously , then there will be majorons in the problem which is ruled out by precision z width measurement at lep .
however , in a variant of this model @xcite lepton number is broken explicitly through an interaction of the triplet with the higgs doublet @xmath162 let @xmath163 and @xmath164 , then the conditions for the minimum of the potential relates the @xmath76 of the two scalars by @xmath165 , where @xmath73 is the mass of the triplet higgs scalar and the neutrino mass matrix becomes @xmath166 . in this case
the lepton number violation comes from the decays of the triplet higgs @xmath167 , @xmath168 the coexistence of the above two types of final states indicates the nonconservation of lepton number . on the other hand , any lepton asymmetry generated by @xmath169 would be neutralized by the decays of @xmath170 , unless cp conservation is also violated and the decays are out of thermal equilibrium in the early universe . in this case
there are no vertex corrections which can introduce cp violation .
the only source of cp violation is the self energy diagrams of figure 5 .
2.5in- .5 in if there is only one @xmath171 , then the relative phase between any @xmath172 and @xmath35 can be chosen real .
hence a lepton asymmetry can not be generated . with two @xmath171 s ,
even if there is only one lepton family , one relative phase must remain . as for the possible relative phases among the @xmath172 s
, they can not generate a lepton asymmetry because they all refer to final states of the same lepton number . in the presence of the one loop diagram
, the mass matrix @xmath173 and @xmath174 becomes different .
this implies that the rate of @xmath175 no longer remains to be same as @xmath176 .
since by @xmath7 theorem @xmath177 , what it means is that now @xmath178 \neq \gamma[\xi_b \to \xi_a ] .$ ] this is a different kind of cp violation compared to the cp violation in models with right handed neutrinos . if we consider that the @xmath179 is heavier than @xmath180 , then after @xmath179 decayed out the decay of @xmath180 will generate an lepton asymmetry given by , @xmath181 } \over { 8 \pi^2 ( m_1 ^ 2 - m_2 ^ 2 ) } } \left [ { { m_1 } \over \gamma_1 } \right].\ ] ] in this model the out - of - equilibrium condition is satisfied when the masses of the triplet higgs scalars are heavier than @xmath182 gev .
the lepton asymmetry thus generated after the higgs triplets decayed away would be the same as the @xmath0 asymemtry before the electroweak phase transition . during the electroweak phase transition
, the presence of sphaleron fields would relate this @xmath0 asymmetry to the baryon asymmetry of the universe .
the final baryon asymmetry thus generated can then be given by the approximate relation @xmath183 .
this allows us to obtain a neutrino mass of order ev or less , as well as the observed baryon asymmetry of the universe @xmath184 as desired . in general
, it is not possible to discriminate the see - saw mechanism from the triplet higgs mechanism for neutrino masses .
however , in some specific supersymmetric inflationary models , where the reheat temperature is lower than @xmath185 gev , the see - saw mechanism is preferred . on the other hand the leptogenesis scenario in the triplet higgs mechanism has several nice features , like the absence of the vertex diagrams or its detectability in the near future in the accelerators @xcite .
however , in the left - right symmetric models both the scenarios are present and can contribute to the neutrino masses as well as to leptogenesis @xcite .
the majorana masses of the neutrinos implies lepton number violation .
one very important consequence of this lepton number violation in the early universe is that it can erase any primordial baryon asymmetry of the universe in the presence of the sphaleron field before the electroweak phase transition .
this gives bound on the mass on the neutrinos .
while a general analysis can give somewhat weak bound , in some specific models these bounds could be very important .
for example , in zee - type radiative models or the r - parity breaking supersymmetric models this is very restrictive .
if one attempts to explain the atmospheric neutrino problem , then these models would wash out all primordial baryon asymmetry of the universe .
this implies that most models of neutrino masses based on these two scenarios are incomplete and more inputs are required to explain the present baryon asymmetry of the universe in these models . on the other hand , in the see - saw mechanism and the triplet higgs mechanism , the lepton number violation that gives masses to the neutrinos
also generate a lepton asymmetry of the universe , which then get converted to a baryon asymmetry of the universe in the presence of the sphaleron field .
r. davis , prog .
. phys . * 32 * , 13 ( 1994 ) ; y. fukuda _ etal .
lett . * 77 * , 1683 ( 1996 ) ; * 81 * , 1158 ( 1998 ) ; p. anselmann _ etal . _ ,
b357 * , 237 ( 1995 ) ; * b361 * , 235 ( 1996 ) ; j. n. abdurashitov _
et al . _ ,
b328 * , 234 ( 1994 ) .
shaposhnikov , nucl . phys .
* b 287 * ( 1987 ) 757 ; nucl . phys . *
b 299 * ( 1988 ) 797 ; a.d .
dolgov , _ phys .
* 222 * ( 1992 ) 309 ; v.a .
rubakov and m.e .
shaposhnikov , _ usp .
* 166 * ( 1996 ) 493 ; _ phys .
* 39 * ( 1996 ) 461 ; m.e .
shaposnikov , nucl . phys .
* b 287 * ( 1987 ) 757 ; n. turok and j.zadrozny , phys .
* 65 * ( 1990 ) 2331 ; a. kazarian , s. kuzmin and m.e .
shaposnikov , phys . lett .
* b 276 * ( 1992 ) 131 ; l. mclerran , m.e .
shaposnikov , n. turok and m. voloshin , phys .
* b 256*(1991 ) 451 ; for a recent review see a. riotto , hep - ph/9807454 and a. riotto and m. trodden , hep - ph/9901362 .
bochkarev , s.v .
kuzmin and m.e .
shaposnikov , phys . lett .
* b 244 * ( 1990 ) 275 ; m. dine , p. huet and r. singleton jr , nucl . phys .
* b 375 * ( 1992 ) 625 ; m.e .
shaposnikov , phys . lett . * b 316 * ( 1993 ) 112 .
mohapatra , _ unification and supersymmetry _
( springer - verlag ) ( 1992 ) ; a. masiero , in _ grand unification with and without supersymmetry and cosmological implications _ ,
international school for advanced studies lecture series no . 2 , ( world scientific ) ( 1984 ) 1 ; a. zee , ( ed . ) _ unity of forces in the universe _ * 1 * ( 1980 ) ( world scientific ) ; p. langacker , phys . rep . * 72 * ( 1981 ) 185 .
m. gell - mann , p. ramond , and r. slansky , in _
supergravity _ , edited by p. van nieuwenhuizen and d. z. freedman ( north - holland , amsterdam , 1979 ) , p. 315 ; t. yanagida , in _ proceedings of the workshop on the unified theory and the baryon number in the universe _ , edited by o. sawada and a. sugamoto ( kek report no
. 79 - 18 , tsukuba , japan , 1979 ) , p. 95 ; r. n. mohapatra and g. senjanovic , phys .
lett . * 44 * , 1316 ( 1980 ) .
s. nandi and u. sarkar , phys .
* 56 * ( 1986 ) 564 ; a.s .
joshipura and u. sarkar , phys .
* 57 * ( 1986 ) 33 ; a. masiero , d.v .
nanopoulos and a.i .
sanda , phys .
* 57 * ( 1986 ) 663 ; e. farhi and l. susskind , phys
d 20 * ( 1979 ) 3404 ; s. dimopoulos , nucl . phys .
* b 168 * ( 1980 ) 69 ; l. wolfenstein , nucl . phys . * b 175 * ( 1980 ) 93 ; s. nussinov , phys . lett .
* b 165 * ( 1985 ) 55 .
a. zee , phys . lett .
* b 93 * ( 1980 ) 389 ; a. yu .
smirnov and m. tanimoto , phys .
d55 * ( 1997 ) 1665 ; k. s. babu and e. ma , mod .
lett . * a4 * , 1975 ( 1989 ) ; n. gaur , a. ghosal , e. ma , and p. roy , phys .
d58 * , 071301 ( 1998 ) ; e. ma , hep - ph/9807386 ( phys . lett .
* b * , in press ) ; c. jarlskog , m. matsuda , s. skadhauge , and m. tanimoto , hep - ph/9812282 .
hempfling , nucl .
* b 478 * ( 1996 ) 3 ; h. niles and n. polonsky , nucl .
* b484 * ( 1997 ) 33 ; a. yu . smirnov and f. vissani ,
* b 460 * ( 1996 ) 37 ; s. rakshit , g. bhattacharyya and a. raychaudhuri , hep - ph/9811500 ; b. mukhopadhyaya , s. roy and f. vissani , phys .
* b 443 * ( 1998 ) 191 ; a.s .
joshipura and s. vempati , hep - ph/9808232 ; hep - ph/9903435 ; e.j .
chun , s.k .
kang , c.w .
kim and u.w .
lee , hep - ph/9807327 ; e.j .
chun and j.s .
lee , hep - ph/9811201 ; for a recent discussion , see for example m. drees , s. pakvasa , x. tata , and t. ter veldhuis , phys .
* d57 * ( 1998 ) r5340 .
e. ma and u. sarkar , phys .
* 80 * ( 1998 ) 5716 ; u. sarkar , phys . rev . * d 59 * ( 1999 ) 031301 ; e. ma , m. raidal and u. sarkar , hep - ph/9808484 ; e. ma , s. sarkar and u. sarkar , hep - ph/9812276 ; d. delepine and u. sarkar , hep - ph/9811479 .
gelmini and m. roncadelli , phys . lett .
* b 99 * ( 1981 ) 411 ; c. wetterich , nucl . phys .
* b 187 * ( 1981 ) 343 ; g. lazarides , q. shafi and c. wetterich , nucl .
* b 181 * ( 1981 ) 287 ; r.n .
mohapatra and g. senjanovic , phys .
d 23 * ( 1981 ) 165 ; r. holman , g. lazarides and q. shafi , phys . rev . *
d 27*(1983 ) 995 ; g. lazarides and q. shafi , report no hep - ph/9803397 .
e. ma , phys .
b433 * , 74 ( 1998 ) ; e. ma and u. sarkar , phys . lett .
* b439 * ( 1998 ) 95 ; e. ma , d.p .
roy and u. sarkar , phys . lett . * b 444 * ( 1998 ) 391 ; e. ma and d. p. roy , phys . rev . * d58 * , 095005 ( 1998 ) . g.d .
coughlan , g.g .
ross , r. holman , p. ramond , m. ruiz - altaba and j.w .
valle , phys .
* 158b * , ( 1985 ) 401 ; g.d .
coughlan , g.g .
ross , r. holman , p. ramond , m. ruiz - altaba and j.w .
valle , phys .
160b * ( 1985 ) 249 ; a. masiero and j.w .
valle , phys . lett .
* b251 * ( 1990 ) 273 ; t. multamaki and i. vilja , phys . lett . * b433 * ( 1998 ) 67 .
a. masiero and a. riotto , phys . lett .
* b 289 * ( 1992 ) 73 ; b. campbell , s. davidson and k.a .
olive , phys .
lett , * b 303 * ( 1993 ) 63 ; nucl . phys . *
b 399 * ( 1993 ) 111 ; r. adhikari and u. sarkar , phys . rev . *
d 55 * ( 1997 ) 3836 ; r. adhikari and u. sarkar , phys .
* b 427 * ( 1998 ) 59 .
m. fukugita and t. yanagida , phys . lett .
* b 174 * ( 1986 ) 45 ; p. langacker , r. peccei and t. yanagida , mod .
a 1 * , 541 ( 1986 ) ; m.a .
luty , phys . rev .
* d 45 * ( 1992 ) 445 ; h. murayama , h. suzuki , t. yanagida and j. yokoyama , phys .
* 70 * ( 1993 ) 1912 ; r.n .
mohapatra and x. zhang , phys
. rev . * d 45 * ( 1992 ) 5331 ; a. acker , h. kikuchi , e. ma and u. sarkar , phys
d 48 * ( 1993 ) 5006 ; t. gherghetta and g. jungman , phys.rev . *
d 48 * ( 1993 ) 1546 ; k. enqvist and i. vilja , phys .
* b 299 * ( 1993 ) 281 ; p.j . odonnell and u. sarkar , phys
d 49 * ( 1994 ) 2118 ; a. ganguly , j.c .
parikh and u. sarkar , phys . lett .
* b 385 * ( 1996 ) 175 ; w. buchmller and m. plmacher , phys .
* b 389 * ( 1996 ) 73 ; phys .
* 431 * ( 1998 ) 354 ; l. covi , e. roulet and f. vissani , phys .
* b 384 * ( 1996 ) 169 ; m. plmacher , zeit .
* c 74 * ( 1997)549 ; j. faridani , s. lola , p.j .
odonnell and u. sarkar , eur .
j. * c 7 * ( 1999 ) 543 .
l. covi , e. roulet and f. vissani , phys . lett .
* b 384 * ( 1996 ) 169 ; l. covi and e. roulet , phys .
* b 399 * ( 1997 ) 113 ; l. covi , e. roulet and f. vissani , phys . lett . * b 424 * ( 1998 ) 101 ; w. buchmller and m. plmacher , hep - ph/9710460 ( revised ) ; m. flanz and e.a .
paschos , hep - ph/9805427 .
j. papavassiliou and a. pilaftsis , phys .
* 75 * ( 1995 ) 3060 ; phys .
d 53 * ( 1996 ) 2128 ; phys . rev . * d 54 * ( 1996 ) 5315 ; a. pilaftsis , phys .
* 77 * ( 1996 ) 4996 ; nucl . phys .
* b 504 * ( 1997 ) 61 ; zeit . phys . * c 47 * ( 1990 ) 95 ; a. pilaftsis and m. nowakowski , phys . lett . * b 245 * ( 1990 ) 185 ; _ mod .
_ * a 6 * ( 1991 ) 1933 .
j. c. pati and a. salam , phys
* d10 * ( 1974 ) 275 ; r. n. mohapatra and j. c. pati , phys . rev . * d11 * ( 1975 ) 566 ; r. n. mohapatra and g. senjanovic , phys . rev . * d12 * ( 1975 ) 1502 ; r. e. marshak and r. n. mohapatra , phys .
* 44 * ( 1980 ) 1316 . | majorana masses of the neutrino implies lepton number violation and is intimately related to the lepton asymmetry of the universe , which gets related to the baryon asymmetry of the universe in the presence of the sphalerons during the electroweak phase transition . assuming that the baryon asymmetry of the universe is generated before the electroweak phase transition , it is possible to dicriminate different classes of models of neutrino masses . while see - saw mechanism and the triplet higgs mechanism are preferred , the zee - type radiative models and the r - parity breaking models requires additional inputs to generate baryon asymmetry of the universe during the electroweak phase transition . plus
1pt
@=12 -0.5 in 0.0 in 0.0 in 8.5 in 6.5 in [ 1 ] june 1999 * models of neutrino masses and baryogenesis + * |
zero - metallicity population iii stars ( hereafter p3 ) are thought to have preceded the normal metal - enriched stellar populations , but because they would be located at high @xmath0 they are inaccessible to direct observations by current telescopes .
if massive , they are expected to have left a significant level of diffuse radiation shifted today into ir , and it was suggested that the cosmic infrared background ( cib ) contains a significant contribution from p3 in near - ir , both its mean level and anisotropies ( see review by kashlinsky , 2005 and references therein ) . this has recently received strong support from measurements of cib anisotropies in deep spitzer / irac images @xcite .
if p3 are responsible for even a fraction of the claimed nir cib they would provide a high comoving density of photons all the way to the p3 era . in this _ letter _ we analyze effects of such photons on spectra of high-@xmath0 high - energy gamma - ray bursts ( grbs ) and blazars that would be observed with the upcoming glast lat instrument to 300 gev .
we show that the entire claimed nir cib excess ( nirbe ) can be explained if only @xmath24% of the baryons have gone through p3 stars .
this would result in @xmath3 photons/@xmath4 whose present day energy is between 1 and 4 . such photons would provide a large optical depth due to photon - photon absorption for grbs ( and other sources ) at energies that will be probed with glast . detecting this spectral damping in forthcoming grb observations will provide an important test of the p3 era parameters .
ay wavelengths @xmath5 the total flux produced by the observed galaxies matches the levels of the cib within its uncertainties , but in the near - ir ( nir ) the claimed levels of the cib are substantially higher than the net fluxes produced by galaxies out to flux limits where this contribution saturates ( see review by kashlinsky 2005 for details ) .
[ cib_dc ] shows the cib _ excess _ levels ( filled circles ) over the net flux from galaxies observed in deep surveys ( open symbols ) ; the caption discusses the details .
the excess is significant at @xmath6 , the range we term nir , and its bolometric flux is @xcite : @xmath7 at @xmath8 we evaluated the upper limits shown in fig .
[ cib_dc ] described in the caption .
the wavelengths @xmath910 contribute little , so we adopt the value of @xmath10 for what follows . it was suggested that the nirbe is produced by massive p3 stars at high @xmath0 ( @xmath11 ) @xcite
. significant energy release by p3 is suggested from the recent measurement of cib anisotropies in deep exposure spitzer data @xcite . because p3 stars , if massive , would radiate at the eddington limit , where @xmath12 , the total flux produced by them is largely model - independent @xcite .
we reproduce briefly the argument from kashlinsky et al ( 2004 ) : each star would produce flux @xmath13 , where @xmath14 is the luminosity distance . because for massive stars @xmath12 , the total comoving luminosity density from p3 is @xmath15 , where @xmath16 is their luminosity function and @xmath17 is the mass fraction of baryons locked in p3 at any given time . in the flat universe , the volume per unit solid angle subtended by cosmic time @xmath18 is @xmath19
. finally , these stars would radiate at efficiency @xmath20 ( @xmath21 for hydrogen burning ) .
this then leads to the closed expression for the total bolometric flux from these objects : @xmath22 here @xmath23 is the mean mass fraction of baryons locked in p3 stars and @xmath24 is a suitably averaged redshift over their era .
the total flux is a product of the maximal luminosity produced by any gravity - bound object , @xmath25 , distributed over the surface of the hubble radius , @xmath26=@xmath27 , and the fairly understood dimensionless parameters . from wmap observations we adopt @xmath28=0.0224 @xcite and , since the massive stars are fully convective ,
their efficiency is close to that of hydrogen burning @xcite , @xmath20=0.007 . requiring that p3 stars are responsible for the flux given by eq .
[ excess ] leads to : @xmath29 assuming @xmath3010 this is somewhat less than the @xmath31 value suggested by madau & silk ( 2005 ) and considerably less than the @xmath32 value of dwek et al ( 2005 ) . within the nirbe uncertainty , only @xmath92% of the baryons had to go through p3 .
this is not unreasonable considering that primordial clouds are not subject to many of the effects inhibiting star formation at the present epochs , such as magnetic fields , turbulent heating etc .
the only criterion for p3 formation seems to be that primordial clouds turning - around out of the primordial density field have the virial temperature , @xmath33 , that can enable efficient formation of and cooling by molecular hydrogen @xcite . assuming spherical collapse of gaussian fluctuations and the @xmath34cdm model from wmap observations @xcite the fraction of collapsed haloes at @xmath0=10 with @xmath35(400,2000)k is @xmath36 in good agreement with eq .
[ pop3fraction ] as can be derived from fig . 2 in kashlinsky et al ( 2004 ) .
. [ pop3fraction ] was evaluated from 1 to 4 , but with significant cib excess flux outside that range , @xmath23 would increase .
however , at wavelengths @xmath37 the high-@xmath0 emissions would be below the lyman break and would be reprocessed to @xmath38 ( santos et al 2002 ) . if @xmath39 , which is unlikely in light of wmap polarization data ( kogut et al 2003 ) , the rest frame lyman break may be redshifted to @xmath401 , but the possible extra cib excess from @xmath411 will be compensated by @xmath23 in eq .
[ pop3fraction ] decreasing with @xmath42 . at longer wavelengths , the cib excess given by eq .
[ excess ] can at most increase @xmath23 by @xmath43 .
the above estimate is subject to two caveats : first , it assumes the nir cib at the levels given by eq .
[ excess ] , which were derived assuming a specific set of zodiacal light models
. the latter may carry large systematic uncertainties , which are not included in the formal uncertainties in eq .
[ pop3fraction ] .
our estimate of @xmath23 is proportional to the cib excess summarized in fig .
[ cib_dc ] and eq .
[ excess ] and would change accordingly if these are superseded by future measurements . secondly
, this estimate assumes that p3 were massive stars leading to the effective efficiency , @xmath20 , very close to that of the hydrogen burning , @xmath20=0.007 . if p3 were less massive than @xmath44 , the effective @xmath20 of their energy release would decrease by a factor of a few @xcite requiring significantly larger values of @xmath23 for a given @xmath45 .
independently of the cib considerations , if p3 were less massive than @xmath46 , their fraction must be very small in order not to overproduce the metallicities of the poorest population ii stars @xcite , leading to cib fluxes from p3 significantly lower than eq .
[ excess ] .
if p3 at early epochs produced even a fraction of the claimed nirbe , they would supply abundant photons at high @xmath0 .
the present - day value of @xmath47=1 mjy / sr corresponds to the comoving number density of photons per logarithmic energy interval , @xmath48 , of @xmath49=0.6 @xmath50 and if these photons come from high @xmath0 then @xmath51 . these photons also had higher energies in the past , @xmath52(0.1 - 0.3)@xmath53ev , providing an abundance of absorbers for sufficiently energetic photons at high redshifts via @xmath54 @xcite .
stecker & de jager ( 1993 ) have pioneered applications of the @xmath55-@xmath55 absorption to constraining the present - day cib from high - energy spectra of low-@xmath0 blazars .
madau & phinney ( 1996 ) and salamon & stecker ( 1998 ) have considered effects by evolving normal galaxy populations on potential future intermediate @xmath0 ( @xmath56 ) blazars . however , as shown in sec .
2 , p3 stars are likely to have provided a far more abundant source of photons at high @xmath0 to interact with high - energy gamma - ray photons .
grbs are the obvious objects whose high - energy emissions would be damped by the absorption from the nirbe photons .
this effect was difficult to detect with egret because of its low sensitivity .
observations by egret have detected only 6 grb s with one of them being a record energy 18 gev photon @xcite .
a successor to the gamma - ray observatory , nasa s glast is to be launched in 2007 .
its large - area telescope ( lat ) will provide significantly improved sensitivity needed for detecting high-@xmath0 grbs out to @xmath57=300 gev with better than 10% energy resolution above 100 mev , and its large field - of - view should detect @xmath2 100 - 300 grbs / yr . it is expected that @xmath250 of these would have enough high - energy photons to measure spectral indices at @xmath580.1 gev with uncertainty better than 0.1 .
we show in this section that the photons produced by the p3 stars should leave a detectable signature by damping the high - energy part of the spectra of high-@xmath0 grbs .
we denote with @xmath59 and @xmath60 the present day energies of the cib and grb photons respectively and the primes refer to rest - frame energies , e.g. @xmath61 .
the photon - photon absorption , being electromagnetic in nature , has cross - section @xmath2 that for the thompson scattering , @xmath62 ; it is @xmath63 $ ] , where @xmath64 and @xmath65 .
the cross - section has a sharp cutoff as @xmath66 , peaks at @xmath67 at @xmath68 , and is @xmath69 for @xmath70 .
the mean free path of grb photons in the presence of cib would be @xmath71 mpc . the right vertical axis in fig .
[ cib_dc ] shows the product of @xmath72 and the comoving photon number density for given @xmath47 .
the figure also marks the regions defined by the photon - photon absorption threshold .
discussion in sec.2 makes it plausible that , if the nirbe originates from p3 stars , it must have a sharp truncation corresponding to the redshifted lyman break , or @xmath73 .
thus we assume that , at least at high @xmath0 , there are few photons at wavelengths that are shifted today to @xmath411 . at longer wavelengths
there is no observational evidence for cib excess over that from ordinary " galaxies containing population i and ii , but at the same time only spectra of the sources of very high energy @xmath55-rays at very high @xmath0 would be affected by that range of the cib .
thus it appears that there exists a narrow wavelength window of @xmath74 in which the cib photons can interact with high-@xmath60 , high-@xmath0 grbs and probe the emissions from the , so far putative , p3 era .
we assume a flat universe dominated by the cosmological constant and that the nirbe photons originated from p3 at redshifts higher than that of the grbs , so that @xmath75 . in this case , the optical depth due to photon - photon absorbtion is : @xmath76 where @xmath77 is the present - day photon density corresponding to the observed cib excess .
[ tau ] shows the resultant optical depth and contributions to it from different @xmath0 .
we adopted @xmath78=1.24 ev ( 1 ) and the form of @xmath77 corresponding to the solid line in fig .
[ cib_dc ] .
the onset of @xmath791 occurs rapidly at @xmath80 gev .
the grb spectra at these energies should either be strongly damped or there had to be only negligible energy releases from the p3 era .
given the high values of @xmath81 this would still hold even if p3 era produced only a small part of @xmath45 , but assumes that the latter has a lyman cutoff redshifted to @xmath21today . if the p3 era extended to lower @xmath0 , the lyman cutoff would occur at the observer wavelength @xmath82 and grb spectra would be damped at proportionately lower @xmath60 .
thus , _ the location of this cutoff may also serve as an indicator of the grb s redshift_. with the advertised lat energy resolution of @xmath83 at @xmath841 gev , one could determine grb redshifts from the damping by the p3 photons to better than @xmath85 .
how robust is this result ?
first , photons due to optical counts of galaxies at much later times do not affect much the grb spectra because @xmath86 at low @xmath0 and they contribute @xmath87 a few as fig . [ cib_dc ] shows .
furthermore , unlike the p3 photons for which @xmath88 , those contributed by galaxies would produce a still smaller contribution at earlier times compared to p3 .
second , even if the nirbe is not entirely cosmological , fig .
[ tau ] shows that p3 emissions should still lead to a very large optical depth , which scales as @xmath89
. thus p3 would likely be the dominant contributors to the optical depth to grbs at high energies .
finally , the magnitude of @xmath81 at a given bolometric cib flux should not be sensitive to the cib excess spectrum because the ( already narrow ) range of 1 - 4 available to dampen grbs at high @xmath0 is further decreased by @xmath53 .
[ tau ] shows that the optical depth from p3 is very high , @xmath90 which , when combined with the sudden onset of the optically thick regime , would lead to an identifiable damping by the p3 photons .
the damping will affect progressively lower energy part of the rest - frame grb spectra as one moves to higher @xmath0 . at these energies ,
the grb emission is likely produced by the inverse compton and the synchrotron self compton components and is expected to be high ( or even dominate ) for the typical lorentz factors involved ( piran 2004 and references therein ) ; e.g. dingus et al ( 1998 ) construct an average spectrum from four egret burst and find the differential photon index of @xmath91 out to @xmath210gev .
it may be sufficient to use glast observations of fairly low @xmath0 grbs ( say , @xmath922 - 4 determined spectroscopically in afterglow observations ) to establish the existence of the cib from p3 epochs ; if positive then the higher @xmath0 grbs can be used to further verify that it comes from the p3 epochs and/or calibrate the @xmath0 determinations from the damping .
although spectra of more and more distant blazars are now measured with new instruments , such as hess , these blazars are still too close for unambiguous detection of the p3 emissions .
the farthest blazars with known spectra are at @xmath930.13 - 0.18 and the spectra extend to @xmath943 tev ( dwek et al 2005 , aharonian et al 2005 ) .
the right panel in fig .
[ tau ] shows the optical depth of a blazar at @xmath0=0.18 produced by the @xmath55-@xmath55 absorption 1 ) due to observed galaxies , 2 ) due to nirbe from dirbe and irts , and 3 ) omitting the irts - based point at 1.65 and assuming the lower end of the dirbe - based results .
even if the entire nirbe is correct and originates from p3 , the _ additional _ damping from it is small and is more pronounced only at @xmath581 tev , where measurements and interpretation are difficult .
this is because p3 contribute to the background over a short range of wavelengths longward of @xmath95 .
if p3 were to contribute only a fraction of the nirbe , their emissions will not contribute appreciably to the observed spectra of @xmath96 blazars , but would be seen in the high-@xmath0 grbs . if glast collects a large sample of blazars and other agns at @xmath97 , fig .
2 shows that they can also be used to probe the emissions from the p3 era . | we discuss the contribution of population iii stars to the near - ir ( nir ) cosmic infrared background ( cib ) and its effect on spectra of high-@xmath0 high - energy gamma - ray bursts ( grbs ) and other sources .
it is shown that if population iii were massive stars , the claimed nir cib excess will be reproduced if only @xmath1 of all baryons went through these stars .
regardless of the precise amount of the nir cib from them , they likely left enough photons to provide a large optical depth for high - energy photons from distant grbs .
observations of such grbs are expected following the planned launch of nasa s glast mission .
detecting such damping in the spectra of high-@xmath0 grbs will then provide important information on the emissions from the population iii epoch and location of this cutoff may serve as an indicator of the grb s redshift .
we also point out the difficulties of unambiguously detecting the cib part originating from population iii in spectra of low @xmath0 blazars .
# 1 2sr 2m4sr |
nature harvests solar energy with a remarkably high _ quantum efficiency _ , the percentage of charge carriers created by photons .
recent spectroscopic experiments @xcite and theoretical models @xcite , provide evidence that efficient light harvesting in nature occurs by a quantum mechanism involving sustained electronic coherence @xcite and entanglement @xcite between chromophores . while the chromophores are chlorophyll molecules containing large networks of conjugated carbon bonds that surround a charged magnesium ion , they have largely been represented in theoretical studies @xcite by one - electron models that neglect the effects of electron correlation and entanglement within chromophores .
two advanced methods in electronic structure , density - matrix renormalization group @xcite and two - electron reduced - density - matrix theory @xcite , have recently shown that networks of conjugated bonds as in acene chains @xcite , acene sheets @xcite , and chlorophyll are associated with polyradical character that can not be adequately described without a strongly correlated many - electron quantum model .
[ fig : lip ] -electron model by lipkin , meshkov , and glick .
illustration by k. naftchi - ardebili , the university of chicago , 2011 . used with permission.,width=240,height=144 ] in this paper
we examine the efficiency of light harvesting where we represent each chromophore by a correlated @xmath1-electron model to treat strong electron correlation .
figure 1 illustrates the replacement of one - electron models for each of the 7 chromophores in the fenna - matthews - olson ( fmo ) complex of green - sulfur bacteria by @xmath1-electron models of increasing complexity .
here we employ the @xmath1-electron lipkin - meshkov - glick ( lmg ) model @xcite used extensively in electronic structure @xcite and quantum information @xcite .
we find that strong electron correlation , considered as a _
unique variable in this model _ , dramatically enhances the efficiency of the energy transfer to the reaction center by more than 100% . the results from the multi - electron / exciton models , which unlike one - electron / exciton models allow us to control the amount of strong electron correlation in the chromophores , are consistent with the notion that strong electron correlation is likely employed by nature to enhance its energy - transfer efficiency as much as other factors such as ( i ) environmental noise @xcite and ( ii ) entanglement between chromophores @xcite , which have been extensively studied in the recent literature .
in the multi - electron / exciton model the correlation of electrons within a molecular subunit like a chromophore is intrinsically connected with the entanglement of electrons between molecular subunits .
furthermore , the results suggest a general design principle for man - made materials in which electron correlations and entanglements both within and between subunits are simultaneously tuned for achieving enhanced quantum efficiencies .
the results give guidelines and inspiration for the engineering of new materials if their properties can be tailored to match the parameters of the model . in combination with other recent advances , including the study of functional subsystems of the fmo complex @xcite , this interplay of electron entanglements on different length scales may enable us to develop materials with quantum efficiencies approaching those found in natural processes from photosynthesis to bioluminescence .
the fenna - matthews - olson ( fmo ) complex of green - sulfur bacteria contains three identical subunits , each with a network of seven chromophores .
very recent crystallographic and quantum chemistry studies @xcite indicate that there is likely an eighth chromophore in the fmo of _ prosthecochloris aestuarii _ , which due to sample preparation is not present in the ultrafast spectroscopic studies @xcite .
theoretical models of an fmo complex s subunit typically represent each of the seven chromophore by a _ one - electron model _ in which the electron has access to two energy levels separated by the excitation energy of the chromophore
. interactions @xmath2 between pairs of chromophores are modeled by the exchange of single - electron excitations ( or _ excitons _ ) between them : @xmath3 where @xmath4 the quantum numbers @xmath5 and @xmath6 denote the seven sites of the chromophores while the quantum number @xmath7 , equal to + 1 or -1 , indicates one of the two energy levels within each chromophore .
the second - quantized operator @xmath8 ( @xmath9 ) creates ( annihilates ) an electron on chromophore @xmath5 in energy level @xmath7 .
the 7 parameters @xmath10 are the excitation energies of the chromophore , and the 21 parameters @xmath11 are the coupling energies between all pairs of chromophores .
typical values for the excitation and coupling energies are given in the 7x7 hamiltonian of ref .
@xcite , derived from the work of adolphs and renger @xcite . some theoretical treatments ( see for example ref .
@xcite ) express this model hamiltonian in terms of operators @xmath12 and @xmath13 that create and annihilate an exciton on chromophore @xmath5 . by applying the following substitutions @xmath14 we can write the electronic hamiltonian in eqs .
( [ eq : h0 ] ) and ( [ eq : u ] ) as an excitonic hamiltonian @xmath15 because these mappings are exact , these two hamiltonians are _
equivalent_. although the site energies @xmath10 and coupling energies @xmath11 will account for electron correlation in an average fashion if they are determined from quantum calculations and/or experimental data , these hamiltonians do not explicitly correlate the electrons within the chromophores .
the purpose of the present paper is to design an extension of the one - electron ( exciton ) model that treats the electron correlation within the chromophore explicitly both in the stationary - state solutions of the schrdinger equation and in the time - dependent ( dynamic ) solutions of the quantum liouville equation . here
we generalize this one - electron ( exciton ) representation by using the @xmath1-electron lmg model @xcite .
each chromophore is modeled as @xmath1 electrons in two energy levels that are each @xmath1-fold degenerate ; a pairwise interaction that excites two electrons from the lower level to the upper level or de - excites two electrons from the upper level to the lower level @xcite .
the hamiltonian of the 7 interacting lmg chromophores can be expressed as follows : @xmath16 where @xmath17 where the dimensionless parameter @xmath18 controls the strength of the electron interactions _ within _ each chromophore and
@xmath19 $ ] is a dimensionless screening parameter to be defined below .
the product @xmath20 is the interaction energy for the @xmath21 chromophore .
this definition ensures that for each chromophore the ratio of the interaction energy to the site energy is the same constant @xmath22 .
the new quantum number @xmath23 ( or @xmath24 ) denotes the @xmath1 degenerate states within each energy level , which are necessary to accommodate the @xmath1 electrons .
when the interaction strength @xmath22 equals zero , the @xmath1 electrons on each site are non - interacting , and with @xmath25 the model behaves the same as the one - electron - per - site models ; when @xmath22 is unequal to zero , we have a generalized model for light harvesting with a tunable electron correlation on the chromophores .
the @xmath1-electron lmg model has @xmath26 distinct `` molecular orbital '' configurations corresponding to the excitation of zero though @xmath1 electrons .
if @xmath1 is even , there are @xmath27 non - degenerate even excited - state configurations ( with an even number of excitations ) and @xmath27 non - degenerate odd excited - state configuration ( with an odd number of excitations ) .
the selection of @xmath1 in the lmg model , therefore , depends more on the number of configurations to be mixed rather than the number of electrons within the chromophore .
choosing @xmath28 is unsuitable because the single excitations can not mix with another odd configuration . in section [ sec : res ]
we choose @xmath29 to correlate the single excitations within the chromophore with the triplet excitations within the chromophore . while not shown , similar results are obtained with @xmath30 as well as with @xmath31 and @xmath32 .
[ fig : pops ] environmental effects of dephasing and dissipation as well as the transfer of energy to the reaction center ( sink ) can be incorporated by introducing a lindblad operator @xmath33 into the _ quantum liouville equation _ @xmath34 + { \hat l}(d)\ ] ] where @xmath35 is the many - electron density matrix .
the lindblad operator can be divided into three operators that account for dephasing , dissipation , and the sink @xmath36 where @xmath37 the state @xmath38 denotes the ground eigenstate of the hamiltonian in eq .
( [ eq : h2 ] ) , the states @xmath39 represent the excited eigenstates computed from this hamiltonian where the interaction @xmath2 between chromophores is set to zero , the state @xmath40 denotes the reaction center ( sink ) , and @xmath41 indicates the first excited state of the third chromophore multiplied by the ground states of the other chromophores .
non - markovian effects from the environment can also be added to the quantum liouville equation ( for example , refer to ref .
@xcite ) ; however , they will not qualitatively change the effect of strong electron correlation within chromophores on energy - transfer efficiency .
.the parameters @xmath42 , given below , are chosen to ensure that the excitations of the correlated lmg models agree with the excitation energies of the seven chromophores from ref .
@xcite . [ cols="^,^,^,^,^,^,^,^ " , ] the values for the site excitation energies are defined by the 7x7 hamiltonian of ref .
@xcite , which is derived from the data in ref .
@xcite ; when @xmath43 , the parameters @xmath44 are adjusted , as shown in table [ t : eps ] , to ensure that the excitations of the correlated lmg models agree with the excitation energies of the seven chromophores . based on acene - chain @xcite and metalloporphyrin - ring data @xcite , which imply that the population of the lowest unoccupied orbital ( unoccupied in a mean - field treatment ) is at least 20% , we estimate the interaction strength @xmath22 with @xmath29 to be 0.8 .
the estimate of @xmath22 is made by finding the value of @xmath22 that gives an 80% probability of finding an electron in the highest occupied orbital ( occupied in a mean - field treatment ) and a 20% probability of finding an electron in the lowest unoccupied orbital .
lower values of @xmath22 give less correlation and hence , less depletion of the highest mean - field occupied orbital . this estimate is conservative because : ( i ) the acene chains of a similar length typically reveal a nearly biradical filling ( @xmath45 50% in the highest occupied orbital ) , and ( ii ) the presence of the mg ion with its @xmath46 orbitals is expected to enhance the degree of correlation . because the coupling energies @xmath11 given in ref .
@xcite are `` dressed '' dipole - dipole interactions that account implicitly for both the electron correlation within the chromophores and the protein environment surrounding the chromophores , they require adjustment for the lmg chromophore model that contains an explicit electron - electron interaction . to prevent overcounting of the electron correction s effect on the coupling , we screen the coupling energies @xmath11 of ref . @xcite by selecting @xmath47 in eq .
( [ eq : h2 ] ) to be less than unity .
specifically , we set @xmath48 in _ all _ of the calculations reported in section [ sec : res ] to match experimental transfer efficiency to the sink with the lmg model when @xmath29 and @xmath49 . _ without screening _ the transfer efficiency at @xmath49 ( and @xmath50 ) is observed to be significantly greater than its experimental value because the electron correlation is counted _
twice _ : ( i ) implicitly in the coupling energies @xmath11 and ( ii ) explicitly in the lmg model of the chromophores .
the rate parameters @xmath51 , @xmath52 , and @xmath53 in the lindblad operators in eq .
( [ eq : lb ] ) are chosen in atomic units ( and wavenumbers ) to be @xmath54 a.u .
( 33.4 @xmath55 ) , @xmath56 a.u .
( 15.9 @xmath55 ) , and @xmath57 a.u .
( 0.00266 @xmath55 ) , respectively .
these definitions are similar to those employed in previous work with the one - electron / exciton model @xcite .
the exciton populations in chromophores 1 , 2 , and 3 as well as the sink population are shown as functions of time in femtoseconds ( fs ) in figs .
2a and 2b for @xmath29 and @xmath48 with @xmath58 and @xmath49 , respectively .
population dynamics of the excitation are generated by evolving the liouville equation in eq .
( [ eq : l ] ) from an initial density matrix with chromophore 1 in its first excited state and the other chromophores in their ground states . correlating the 4 electrons on each chromophore significantly accelerates the increase in the sink population with time . by 1
ps the sink population for @xmath58 is 0.114 while the population for @xmath49 is 0.287 . correlating the excitons on each chromophore also has the effect of shortening the periods of oscillation in chromophores 1 and 2 and accelerating the population decay in these chromophores , which is consistent with the change in the sink population .
the population dynamics with @xmath29 , @xmath48 , and @xmath49 in fig .
2b can also be compared with the dynamics of the one - electron ( exciton ) model hamiltonian ( equivalent to @xmath29 , @xmath59 , and @xmath25 ) , shown in fig .
1b of ref .
the two figures are nearly identical including the efficiency with which energy is transferred to the sink .
hence , the lmg model of the chromophores matches the experimental results but with an explicitly correlated treatment of the electrons on the chromophores .
the sink population as a function of time ( fs ) is shown in figs . 3 for a range of @xmath22 with @xmath29 and @xmath48 .
importantly , as @xmath22 increases , we observe a dramatic acceleration of the increase of the sink population . for @xmath22
increasing by the sequence 0.0 , 0.4 , 0.8 , and 1.2 , the sink population at 2 ps increases by the sequence 0.221 , 0.367 , 0.498 , and 0.547 . for @xmath29 correlating the electrons within the chromophores
significantly increases the efficiency of energy transfer to the reaction center ( sink ) by as much as 148% .
while we choose @xmath29 , the number @xmath1 of electrons per chromophore can model electron correlation for any @xmath60 .
the precise value of @xmath60 is unimportant because the effect of changing @xmath1 can be related to a rescaling of the interaction @xmath22 .
[ fig : sinks ] figure 4 examines the entanglement of excitons between the lmg - model chromophores for @xmath29 and @xmath48 with @xmath58 and @xmath49 .
we employ a measure of global entanglement in which the squared euclidean distance between the quantum density matrix and its nearest classical density matrix is computed @xcite : @xmath61 in some cases like the entanglement of the chromophores , the squared euclidean distance can be viewed as the sum of the squares of the _ concurrences _
@xcite , a measure of local entanglement . within the mathematical framework of bergmann distances
, the squared euclidean distance can also be related to _ quantum relative entropy _
@xcite , which is often applied as a global entanglement measure .
the squared euclidean distance @xmath62 is nonzero if and only if the excitons on different chromophores are entangled .
the correlation of electrons increases the degree of the entanglement between chromophores at early times and the frequency of its oscillation .
the greater entanglement at early times reflects the opening of additional channels between chromophores for quantum energy transfer .
[ fig : ent ] and @xmath48 with @xmath58 and @xmath49 .
the correlation of the excitons increases the degree of the entanglement between chromophores at early times and the frequency of its oscillation . ]
the chromophores interact through intermolecular forces , both dipole - dipole and london dispersion forces .
the correlated - model results presented here are consistent with the notion that nature enhances these intermolecular forces through strong electron correlation in the @xmath0-bonded networks of the chromophores to achieve the observed energy - transfer efficiency .
the two parameters of the lmg chromophore model provide the simplest approach to studying the effect of strong electron correlation @xmath22 on the effective coupling between chromophores . the one - electron or dipole models with their coupling energies @xmath11 can mimic the efficiency from such correlation within chromophores through an empirical inflation of the one - electron @xmath63 coupling ,
but they do not provide a mechanism for either isolating or estimating the magnitude of the enhanced coupling due to strong electron correlation .
orbital occupations from recent 2-rdm calculations of correlation in polyaromatic hydrocarbons @xcite suggest @xmath64 to be a conservative estimate of the correlation within the lmg models of the 7 chromophores . using this estimate with coupling energies screened to match experimental and computational data , we project a greater than 100% enhancement from the strong electron correlation relative to the theoretical limit in which electron correlation within the chromophore is absent . correlation - enhanced energy transfer can be compared with noise - assisted transfer .
theoretical models @xcite have shown that noise from the environment ( dephasing ) can assist energy transfer in the fmo complex by interfering with the coherence ( resonance ) between chromophores with similar energies , which facilitates the downhill flow of energy to the lowest - energy , third chromophore , connected to the reaction center . electron correlation on each chromophore , we have shown , enhances transfer by opening additional coherent channels between chromophores , which also accelerates energy transfer to the third chromophore .
photosynthesis can draw from both of these sources , strong electron correlation within chromophores and environmental noise , to increase the rates of energy transfer . some of the experimental energy - transfer efficiency attributed to noise in one - electron chromophores models may in fact be due to strong electron correlation .
briggs and eisfeld @xcite recently examined whether the energy - transfer efficiency from quantum entanglement might be matched by a purely classical process .
they conclude that if chromophores are approximated as dipoles , then quantum and classical treatments can achieve similar efficiency .
while their result might also be extendable to other dipole or one - electron approximations of the chromophores , many - electron models of the chromophores can not be represented within classical physics . neither the electron correlation , present in the lmg model of the chromophores when @xmath65 , nor the associated enhancement of energy - transfer efficiency can be mapped onto an analogous classical process .
many theoretical models @xcite have been designed to explore the energy transfer in recent light - harvesting experiments , but most of them treat each chromophore by a single electron with two possible energy states . in reality , however , the chromophores are assembled from chlorophyll molecules that contain an extensive network of conjugated carbon - carbon bonds surrounding magnesium ions , from which significant strong electron correlation , including polyradical character , has been shown to emerge @xcite .
recent experimental efforts , not yet published , provide further insights into the model developed here .
light - harvesting devices with artificial acene - like chromophores are being synthesized and shown to exhibit long - lived coherence .
once the details of these experiments are released , the present multi - electron model and future extensions will be useful in evaluating the role of strong electron correlation in these devices .
furthermore , the present model predicts that the coupling between artificial chromophores , and hence their energy - transfer efficiency , can be enhanced by changing their electronic structure to increase their electron correlation .
experimentally , this prediction can be tested through functional group substitutions or more fundamental changes in the synthetic chromophores that systematically modulate the degree of electron correlation .
based on the present model , we predict that significant changes in coupling and efficiency , statistically related to the amount of electron correlation , will be observed .
in this paper we have examined the effect of strong electron correlation and entanglement _ within _ chromophores through an extension of single - electron / exciton models of the chromophores to @xmath1-electron models , based on the lmg model @xcite .
we find that increasing the degree of electron correlation of each lmg - model chromophore significantly enhances the efficiency with which energy is transferred to the reaction center ( sink ) .
this model - based result is consistent with the notion that nature likely uses strong electron correlation to achieve its energy - transfer efficiency . by mixing and yet controlling the interplay of electron correlation within and between subunits ,
the model also has implications for the design of more energy- and information - efficient materials .
40ifxundefined [ 1 ] ifx#1 ifnum [ 1 ]
# 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | research into the efficiency of photosynthetic light harvesting has focused on two factors : ( 1 ) entanglement of chromophores , and ( 2 ) environmental noise .
while chromophores are conjugated @xmath0-bonding molecules with strongly correlated electrons , previous models have treated this correlation implicitly without a mathematical variable to gauge correlation - enhanced efficiency . here
we generalize the single - electron / exciton models to a multi - electron / exciton model that explicitly shows the effects of enhanced electron correlation within chromophores on the efficiency of energy transfer .
the model provides more detailed insight into the interplay of electron correlation within chromophores and electron entanglement between chromophores .
exploiting this interplay is assisting in the design of new energy - efficient materials , which are just beginning to emerge . |
recently , new methods were developed @xcite to construct symplectic 4-mani - folds with small topology and exotic smooth structures . moreover ,
the method proposed by j. park @xcite was later refined @xcite to produce interesting examples of minimal complex surfaces of general type . in this paper
we show how we can use these constructions in regard to the existence or non - existence of einstein metrics .
lebrun and catanese @xcite showed that there exist homeomorphic , but non - diffeo - morphic manifolds such that one of them admits an einstein metric of positive sign while the other admits an einstein metric of negative sign .
the manifold which has positive scalar curvature einstein metric is @xmath5 while the second manifold is a deformation of the barlow surface .
the key ingredient in their proof was to show that the barlow surface , a simply connected minimal surface of general type with @xmath6 can be deformed to a surface with ample canonical bundle .
in particular , they showed that there exists a simply connected complex surface of general type with @xmath7 and having ample canonical bundle .
the ampleness of the canonical bundle ensures the existence of a khler - einstein metric of negative scalar curvature .
based on ideas from @xcite , lee and park @xcite and more recently park , park , and shin @xcite constructed new examples of simply connected , minimal surfaces of general type with @xmath8 or @xmath9 we show that their examples satisfy the ampleness condition : [ amplu ] there exist simply connected complex surfaces of general type , with @xmath10 or @xmath11 and ample canonical bundle .
in his paper , park @xcite , used his examples and lebrun s obstruction theorem to exhibit a manifold homeomorphic , non - diffeomorphic to @xmath12 which does not admit an einstein metric .
we extend his results to @xmath13 and we also exhibit an infinite family of homeomorphic manifolds for which no einstein metric exists .
[ einstein ] each of the topological 4-manifolds @xmath0 for @xmath14 admits a smooth structure which has an einstein metric of scalar curvature @xmath2 , a smooth structure which has an einstein metric with @xmath3 and infinitely many non - diffeomorphic smooth structures which do not admit einstein metrics .
we can also prove non - existence results for an even smaller topological manifold : [ 5-blow - ups ] the canonical smooth structure on @xmath15 admits an einstein metric of positive scalar curvature , but @xmath16 also supports infinitely many exotic smooth structures which do nt have an einstein metric .
it is a well known fact that in dimensions @xmath17 and @xmath18 the sign of the einstein metric is a topological invariant .
this led besse @xcite to consider the conjecture that no smooth compact @xmath19manifold can admit einstein metrics with different scalar curvature signs .
as we see in the previous theorems , in dimension @xmath20 there are homeomorphic , non - diffeomorphic @xmath21manifolds which admit einstein metrics of opposite sign . for higher dimensions , as far as the authors know ,
lebrun and catanese constructed the only known counterexamples to besse s conjecture .
these examples are in dimensions @xmath22 where @xmath23 in these dimensions , we are able to provide many new examples with the same property : [ high ] let @xmath24 then the manifold @xmath25 obtained by taking the @xmath26fold products , @xmath27 of arbitrary @xmath28 or @xmath29 , admits two einstein metrics @xmath30 such that the signs of the scalar curvature are @xmath31 moreover , these metrics are khler - einstein with respect to two distinct complex structues @xmath32 as a consequence of the khlerian property of the metrics and of our method of construction , we have the following in peculiar fact : their corresponding volumes are equal @xmath33 [ rmk ] soon after this paper was finished , jongil park and his collaborators found @xcite an analogous example of a complex surface of general type homeomorphic to @xmath34 using this construction the results of theorems [ amplu ] , [ einstein ] and proposition [ high ] can be extended for @xmath34 the paper is organized as follows : in section [ ampl ] we discuss the ampleness of the canonical line bundles of some interesting examples of complex surfaces and the relation to the existence of einstein metrics , while in section [ exot ] we introduce the seiberg - witten invariants and use them to get results on non - existence of einstein metrics . in section
[ high ] we treat the higher dimensional case .
in general the existence or non - existence of einstein metrics on a given manifold is hard to prove . in the case @xmath35 in the unobstructed situations , i.e. for certain complex surfaces whose underlying differential structure is @xmath36 the existence of khler - einstein metrics was proved by siu @xcite , or tian and yau @xcite . a complete solution to the existence
was given by tian @xcite : [ tian ] a compact complex surface @xmath37 admits a compatible khler - einstein metric with @xmath38 if and only if its anti - canonical line bundle @xmath39 is ample and its lie algebra of holomorphic vector fields is reductive . in the case when @xmath40 a criterion for the existence of a khler - einstein metric was independently found by aubin @xcite and yau @xcite : [ ay ]
a compact complex manifold @xmath37 admits a compatible khler - einstein metric with @xmath41 if and only if its canonical line bundle @xmath42 is ample .
when such a metric exists , it is unique , up to an overall multiplicative constant . in this section , we discuss some examples of smooth structures on @xmath43 for each such @xmath44 the natural smooth structure as the complex projective plane blown - up at @xmath45 generic points , can be endowed @xcite by a khler - einstein metric of positive scalar curvature . on the other hand , lee and park @xcite and
more recently park , park and shin @xcite constructed new exotic smooth structures on @xmath46 for any @xmath47 moreover , these admit complex structures yielding interesting examples of minimal surfaces of general type .
we are going to prove that the canonical bundle of these surfaces is ample , and so they admit khler - einstein metrics of negative scalar curvature .
the examples constructed in these papers are very similar . in this article
we are going to treat just the more complicated example @xcite of a simply connected minimal surface of general type with @xmath48 and @xmath49 the other examples follow similarly .
we begin by a brief description of the example of a minimal surface of general type with @xmath48 and @xmath50 referring frequently to @xcite .
we start with an appropriate pencil of plane elliptic curves , and we resolve the base locus to obtain a rational elliptic surface @xmath51 with the following set of singular fibers : one @xmath52fiber , one reducible @xmath53fiber and two nodal fibers . moreover , this rational elliptic surface admits 4 sections .
now , @xmath51 is blown - up 21 times at well chosen points to obtain a new rational surface @xmath54 these points can be chosen in such a way @xcite that @xmath55 contains four particular disjoint chains of rational curves consisting of proper transforms of the singular fibers , proper transform of some sections and some of the exceptional divisors introduced in the blowing - up process . according to @xcite ,
these chains are @xmath56 and a chain of length one denoted by @xmath57 these chains are represented by continuous lines in figure [ figure ] .
, title="fig:",width=377 ] + they all satisfy artin s contractibility criterion @xcite . by contracting them ,
we obtain a projective surface @xmath58 with 4 singularities , each admitting 1-parameter @xmath59gorenstein smoothings .
let @xmath60 be the contraction map .
now , it is proved @xcite that the projective surface @xmath58 admits a @xmath59gorenstein smoothing , which is a smooth , simply - connected , minimal surface of general type with @xmath61 we denote by @xmath62 the seven smooth rational curves of self - intersection @xmath63 which are marked by discontinuous lines in figure [ figure ] and by @xmath64 the remaining exceptional divisor from the last singular fiber , drawn as a dotted line .
it is easy to see that the poincar duals of the curves @xmath65 together with all of the irreducible components of our 4 chains @xmath66 and @xmath67 form a basis of @xmath68 which is torsion - free and 31-dimensional .
first , we prove the ampleness of the canonical divisor of the singular surface @xmath69 note that since @xmath58 has only 1-parameter @xmath59gorenstein smoothing singularities , its canonical divisor is in fact @xmath59cartier .
a common feature of all of the examples appearing in @xcite is that the pullback of the canonical divisor of the singular variety @xmath58 to its minimal resolution is effective .
for instance , in the example described above , we can write : @xmath70 in @xcite , the coefficients are explicitely given , but for our purpose it is important to know only that @xmath71 and @xmath72 are strictly positive rational numbers .
for convenience , let @xmath73 denote the exceptional divisor of @xmath74 by a direct computation , in @xcite it is proved that @xmath75 and the nefness of @xmath76 suppose now that @xmath77 fails to be ample . since @xmath77 is already nef , according to the nakai - moishezon criterion @xcite
, there exists an irreducible curve @xmath78 such that @xmath79 the total transform of @xmath80 in @xmath55 is @xmath81 where @xmath82 is the strict transform and @xmath83 are non - negative rational numbers .
we should note here that @xmath82 is not numerically equivalent to @xmath84 to give an immediate proof for this assertion , note that @xmath85 for any ample line bundle @xmath86 on the projective surface @xmath69 but , if @xmath87 we would have @xmath88 and this is a contradiction .
now , a straightforward computation gives : @xmath89 now , obviously the intersection number of @xmath82 with any component of @xmath90 is greater or equal to @xmath91 with equality if and only if @xmath82 does not intersect any of the irreducible components of @xmath92 i.e. @xmath80 does not pass through the singular points of @xmath69 hence , it follows that @xmath93 thus @xmath94 in this case , either there is an @xmath95 such that @xmath96 or @xmath97 for all @xmath98 and @xmath82 does not meet any component of @xmath99 in the first case @xmath82 must coincide with @xmath100 but the computations in @xcite show that @xmath101 for all @xmath98 which is impossible by our assumption
. therefore @xmath82 must have vanishing intersection number with all of the @xmath102 s and also with the all of the components of @xmath99 but , as the poincar duals of the irreducible components of @xmath90 and of the @xmath103s generate @xmath104 must be numerically trivial on @xmath105 and this is a contradiction .
hence , @xmath58 has ample canonical bundle . as a consequence of theorem [ ay ] we get :
[ ample ] there exist smooth complex structures on @xmath0 for @xmath14 which have ample canonical line bundles and hence admitting khler - einstein metrics of negative scalar curvature .
as we have already mentioned earlier , we only discuss the case of @xmath106 the other cases follow similarly .
we know that a @xmath59gorenstein smoothing of @xmath58 exists @xcite and gives a smooth exotic structure on @xmath106 as we showed , the singular projective surface @xmath58 has ample @xmath59 cartier canonical divisor .
but , ampleness is an open property @xcite , so the canonical bundle of these smoothings is automatically ample .
this will allow us now to apply aubin - yau s theorem [ ay ] to prove the existence of khler - einstein metrics of negative scalar curvature on the minimal surfaces of general type described above .
[ barlow ] the examples of complex surfaces with ample canonical bundle we found are new to our knowledge , when @xmath107 or @xmath9 when @xmath108 lebrun and catanese @xcite were able to find a complex surface with ample canonical bundle homeomorphic to @xmath109 as a deformation of the barlow surface . as was pointed out by @xcite , it is not known whether the surfaces we found in corollary [ ample ] are deformation equivalent to the barlow surface , or even diffeomorphic to it .
the classical obstruction to the existence of an einstein metric on an oriented , compact , smooth , riemannian , four - manifold @xmath110 is the hitchin - thorpe inequality : @xmath111 with equality when the einstein manifold @xmath110 is covered by the hyperkhler @xmath112 surface or by the flat four - torus , or by the above with the reversed orientations . here
@xmath113 denote the euler characteristic and the signature of the manifold @xmath114 respectively . using the seiberg - witten theory ,
lebrun @xcite was able to find new obstruction to the existence of einstein metrics .
the novelty of his results is that the existence of einstein metrics depends on the differential structure considered .
the techniques developed by lebrun rely on the existence of a non - trivial solution of the seiberg - witten equations .
we briefly introduce the needed notions of the seiberg - witten theory and state the main obstruction theorem . for simplicity , we consider only the case when @xmath115 has no @xmath116torsion .
then there is a one - to - one correspondence between the set of @xmath117 structures and the set @xmath118 of characteristic classes in @xmath119 the seiberg - witten invariant is defined @xcite as an integer valued function @xmath120 we call @xmath121 a basic class if @xmath122 the seiberg - witten invariant is a diffeomorphism invariant in the case @xmath123 in the case of @xmath124 the invariant depends on the choice of an orientation of @xmath125 and @xmath126 but , as we are interested in manifolds for which the existence of an einstein metric is topologically unobstructed , the inequality ( [ ht ] ) must be strictly satisfied .
hence , in the case when the @xmath127 we have @xmath128 but for any arbitrary metric @xmath129 where we denote by @xmath130 the ( anti)-self - dual part of the harmonic representative 2-form .
thus @xmath131 and in this special situation ( see ( * ? ? ?
* thm 6.9.2 ) ) the seiberg - witten invariant does not depend on the choice of metric .
we are now ready to state lebrun s obstruction , see ( * ? ? ?
3.3 ) : [ obstr ] let @xmath58 be a compact oriented @xmath21manifold with a non - trivial seiberg - witten invariant and with @xmath132 then @xmath133 does not admit an einstein metric if @xmath134 in the last years there was a lot of work done on constructing exotic structures with small topology . in this paper
we will use a construction due to akhmedov , baykur and park @xcite .
similar results could also be found in papers of above authors or of baldridge and kirk .
here we are interested in some exotic structures on @xmath135 in ( * ? ? ?
* section 3.4 ) , the authors construct a minimal symplectic manifold @xmath58 homeomorphic to @xmath114 but not diffeomorphic . to give the reader a better understanding of this new manifold , we sketch its construction .
we start with two product manifolds @xmath136 and @xmath137 , where @xmath138 is a torus of dimension @xmath139 and @xmath140 is a riemann surface of genus @xmath141 if we consider the diagonal class @xmath142 + [ point\times t^2]\in h_2(t^4,\zz),$ ] then a smooth representative in this class is a riemann surface of genus @xmath17 with self - intersection @xmath141 blowing up two points on the smooth representative gives @xmath143 a smooth riemann surface of genus @xmath17 and @xmath144self - intersection .
if we endow the manifolds @xmath51 and @xmath145 with the canonical symplectic structures , we can choose the representative @xmath146 to be a symplectic submanifold .
we can now take the symplectic sum @xmath147 along the two genus @xmath17 symplectic submanifolds @xmath148 and @xmath149 @xmath150 is a symplectic manifold with topological invariants : euler characteristic @xmath151 and signature @xmath152 as @xmath150 is not simply connected we need to make six luttinger surgeries along well chosen tori ( see @xcite ) , which yield a new simply connected , minimal , symplectic manifold @xmath153 with the same topological invariants as @xmath154 notice that the definition of the seiberg - witten invariant in @xcite is for the poincar dual of the characteristic class .
the results are nevertheless the same .
the manifold @xmath58 has a unique basic class @xmath155 such that its seiberg - witten invariant @xmath156 moreover , @xmath58 contains a null - homologous torus @xmath157 with a preferred simple loop @xmath158 in section 4 , @xcite , it is shown that @xmath144surgery on @xmath159 with respect to @xmath160 yields a symplectic manifold @xmath161 which also has an unique basic class @xmath162 then @xmath163surgery@xmath164 on @xmath159 with respect to @xmath160 generates a family of manifolds @xmath165 which are homeomorphic to @xmath166 corresponding to the basic class @xmath167 of @xmath58 there is a unique basic class @xmath168 of @xmath165 for which the seiberg - witten invariant is non - zero , and it can be computed as follows : @xmath169 hence , the family @xmath170 consists of homeomorphic , pairwise non - diffeomorphic manifolds .
as @xmath171 is the unique basic class on @xmath172 and the seiberg - witten invariant of @xmath173 is not @xmath174 the manifold @xmath165 does not admit a symplectic structure .
of course , its basic class satisfies @xmath175 let @xmath176 then , the conditions of theorem [ obstr ] are satisfied and we have the following : [ non - einstein ] the manifolds @xmath177 for @xmath178 support infinitely many non - diffeomorphic exotic smooth structures none of which admits an einstein metric .
the only thing that needs to be argued is that the manifolds @xmath179 are homeomorphic but pairwise not diffeomorphic . for this
we use again the seiberg - witten invariant and the commutativity of our construction .
first we blow - up @xmath180 points on @xmath181 and denote by @xmath182 the exceptional divisors .
this gives a new symplectic manifold , which has basic classes @xmath183 @xmath144surgery on @xmath159 with respect to @xmath160 gives the manifold @xmath184 with corresponding basic classes @xmath185 we have a similar relation for the seiberg - witten invariant of @xmath186 : @xmath187 hence , for any fixed @xmath188 the manifolds @xmath186 are homeomorphic to @xmath189 but pairwise not diffeomorphic .
moreover , if @xmath190 they do nt support a symplectic structure .
we would like to note that since for all of the above smooth structures the seiberg - witten invariant is non - trivial the manifolds do nt admit riemannian metrics of positive constant scalar curvature . moreover , as the @xmath191 there is a bound @xcite on the scalar curvature : @xmath192 where @xmath193 is the volume form with respect to @xmath194 hence the manifolds @xmath186 do nt admit non - negative constant scalar curvature metrics , and their yamabe invariant is negative .
the results of theorem [ einstein ] are immediate by putting together the results from [ ay ] , [ ample ] , [ non - einstein ] .
let @xmath195 be a complex surface with ample canonical bundle and @xmath196 as an example of such surface , we can either take the one provided by corollary [ ample ] , or the example @xcite constructed by lebrun and catanese .
let @xmath197 also be complex surfaces of general type , with ample canonical line bundle and @xmath198 respectively , as in corollary [ ampl ] .
we are now ready to give the proof of proposition [ high ] .
let @xmath199 and let @xmath200 where @xmath201 is either @xmath202 or @xmath9 the manifolds @xmath203 and @xmath204 are homeomorphic , hence by a theorem of wall @xcite , they are h - cobordant .
these h - cobordisms induce an h - cobordism between @xmath25 and @xmath205 but as @xmath206 are simply connected manifolds of dimension greater than @xmath207 , smale s h - cobordism theorem tells us that they are diffeomorphic . we know that @xmath204 and @xmath203 admit khler - einstein metrics of positive , negative scalar curvature , respectively .
we can rescale these metrics such that the scalar curvatures are @xmath208 on @xmath206 we consider the product metrics associated to the corresponding @xmath209 as we take products of khler - einstein with the same scalar curvature , the new metrics are going to be khler - einstein .
the product complex structures on @xmath210 are of kodaira dimension @xmath211 and @xmath22 respectively .
as the dimension considered increases , the number of manifolds constructed in proposition [ high ] is @xmath212
the authors would like to thank claude lebrun for suggesting this problem and for his remarks and suggetions on an early version of this paper .
this paper was written while the first author was a cnrs post - doc at irma , strasbourg , and the second was visiting ihes as a ipde fellow .
we would like to thank these institutions for their hospitality . | we show that each of the topological 4-manifolds @xmath0 for @xmath1 admits a smooth structure which has an einstein metric of scalar curvature @xmath2 , a smooth structure which has an einstein metric with @xmath3 and infinitely many non - diffeomorphic smooth structures which do not admit einstein metrics .
we show that there are infinitely many manifolds homeomorphic non - diffeomorphic to @xmath4 which do nt admit an einstein metric .
we also exhibit new examples of manifolds carrying einstein metrics of both positive and negative scalar curvature .
the main ingredients are recent constructions of exotic symplectic or complex manifolds with small topological numbers . |
the shapley supercluster ( ssc ) is one of the largest known structures in the local universe @xcite .
recent measurements based on the tully - fisher relation give @xmath5 for the hubble flow recession velocity of a3558 , one of the central clusters in the ssc @xcite .
this corresponds to a redshift of @xmath6 , a luminosity distance of @xmath7 , and an angular diameter distance of @xmath8 . despite the great distance of the ssc , it has recently been argued that the ssc contributes significantly to the peculiar velocity of the local group ( lg ) based on the enhanced abundance of x - ray clusters in it @xcite .
the contribution of the ssc to the lg peculiar velocity depends critically on the mean matter density in the supercluster .
it is a non - trivial matter to relate the enhanced abundance of massive x - ray clusters in the region to its overdensity , since clusters are strongly biased with respect to the underlying distribution of matter @xcite .
because x - ray clusters correspond to rare density peaks , even mild enhancements in the mean matter density can result in the formation of many more clusters within the supercluster .
since this bias depends on the cluster mass , the relationship between a sample containing objects in a range of masses and the underlying matter can not be described accurately by a constant bias factor . in this paper
we study the dependence of the number of x - ray clusters in the ssc on the underlying matter overdensity in this region .
we use the hybrid model proposed by @xcite combining the extended press - schechter ( eps ) formalism @xcite with the structure formation prescription of @xcite ( st ) .
the model describes well published results from numerical simulations ( see discussion around fig .
4 of @xcite ) . using the mass - dependent , nonlinear bias for x - ray clusters ,
we calculate the matter overdensity in the ssc region from the number of observed x - ray clusters there . in [ method ] , we describe how the @xcite hybrid model can be used to calculate the mass in a large region enclosing collapsed objects .
we then specifically consider the ssc in [ clusters ] and describe the sample of x - ray clusters used in our calculation . in [ ssc ] we apply our formalism and calculate the matter overdensity , cluster overdensity , cluster bias , and total mass in the ssc region .
given these results , we then use the spherical collapse model in [ dynamics ] to consider the dynamics of the region . in [ pecvel ]
, we estimate the ssc contribution to the peculiar velocity of the lg . in
[ radial ] , we explore the radial dependence of our results in the ssc and , finally , investigate the robustness of our results in [ robust ] . we assume a flat , @xmath9cdm cosmology with the standard set of cosmological parameters @xmath10 @xcite .
according to the eps prescription @xcite , if the linear density fluctuations in the universe are extrapolated to their values today and smoothed on a comoving radius @xmath11 , a point whose overdensity exceeds a critical value of @xmath12 , belongs to a collapsed object with a mass @xmath13 if @xmath11 is the largest scale for which the criterion is met . here
, @xmath14 is the linear growth factor at redshift @xmath15 , @xmath16 is the closure density of the universe , and @xmath17 is the matter density parameter today .
the critical value of the overdensity is also known as the _ barrier_. for a gaussian random field of initial density perturbations , as indicated by measurements of cosmic microwave background ( cmb ) @xcite , the probability distribution of the extrapolated and smoothed over - density , @xmath18 , is also a gaussian : @xmath19 with zero mean and a variance given by : @xmath20 where @xmath21 is the linear power - spectrum of density fluctuations today as a function of comoving wave - number @xmath22 , and @xmath23 . since equation ( [ var ] )
is a monotonically decreasing function of @xmath11 ( or @xmath24 ) , the smoothing scale can be uniquely specified by the variance of the over - density field smoothed on that scale .
if we define @xmath25 to be the fraction of mass contained in halos in the mass range @xmath26 $ ] corresponding to @xmath27 $ ] at redshift @xmath15 , then the comoving number density of collapsed objects in that mass range is @xmath28 the eps prescription , which uses a constant barrier independent of mass with a value , @xmath29 , derived from the spherical collapse model , gives @xmath30 where @xmath31 is the number of standard deviations a density fluctuation on a scale @xmath32 must be above the mean to have crossed the critical density threshold . incorporating a moving ( i.e. scale dependent ) barrier generated from elliptical collapse with two free parameters , the st prescription gives @xcite @xmath33\,\exp\left(-\frac{a\,\nu^2}{2}\right)\,ds\ ] ] and results in a mass function that better matches simulations .
the best - fit values of the parameters are @xmath34 and @xmath35 , while the normalization factor is @xmath36 @xcite .
the unconditional mass function generated with @xmath37 represents an average over all regions of space ( or equivalently , over all realizations of density fourier mode amplitudes ) .
it assumes no prior knowledge of the overdensity on a given scale .
if we fix the average linear overdensity to be @xmath38 in a region smoothed on a particular scale , @xmath39 , we can generate a conditional mass function with a conditional form of @xmath40 . in eps ,
this can be done by substituting @xmath41 and @xmath42 and gives results that agree well with simulations .
the same substitution with @xmath43 , however , does not agree as well .
this is because the new barrier height is scale dependent @xcite .
however , @xcite suggested using a hybrid model in which the conditional mass function is generated from contributions by both @xmath44 and @xmath43 in regimes where they each fit best with simulations .
the resulting collapsed mass fraction per variance interval is @xmath45 this model describes well the numerical results from cosmological simulations @xcite . because @xmath46 in any region , the average number of collapsed objects , @xmath47 , with mass corresponding to @xmath32 in a region corresponding to @xmath39 is generated by equation ( [ mfdef ] ) with @xmath48 .
we use this prescription to map values of the overdensity in a region to the average number of collapsed objects contained in it .
variations between this conditional average number of objects , @xmath49 and the actually number , @xmath50 , of those counted in the region result only from poisson fluctuations . the cosmic variance has been taken out by stipulating @xmath51 in @xmath52 .
given an integer number , @xmath50 , of observed objects residing in a region @xmath39 , the differential probability distribution of values of @xmath49 that have resulted in @xmath50 is , @xmath53 where @xmath54 is the poisson distribution , @xmath55 is the unconditional distribution of overdensities in the region @xmath39 given by equation ( [ pdel ] ) .
the jacobian @xmath56 is derived from equations ( [ mfdef ] ) and ( [ fhybrid ] ) , and the coefficient @xmath57 is set so as to normalize the integral of equation ( [ pn ] ) over @xmath49 to unity . [ cols="<,^,^,^,^,^,^,^,^,^ " , ] col .
( 1 ) : source name . col .
( 2 ) : ra and dec ( j2000 ) .
( 3 ) : redshift .
( 4 ) : distance from a3558 in @xmath58 .
( 5 ) : @xmath59 in the @xmath60 energy band in units of @xmath61 .
( 6 ) : cluster average temperature in @xmath62 . col .
( 7 ) : @xmath63 calculated from equation ( [ m - t ] ) in units of @xmath64 .
( 8) : @xmath65 in units of @xmath64 .
we conservatively estimate errors in @xmath66 and @xmath63 to be about @xmath67 .
( 9 ) : literature reference for the cluster temperature : ( a ) @xcite ; ( b ) @xcite ; and ( c ) @xcite .
[ sample ] when applying this formalism to observations , there is an additional complication in that overdensities and the sizes of regions are observed in eulerian coordinates which evolve as the region breaks away from the hubble flow , while the eps and st prescriptions rely on initial values in lagrangian coordinates .
lagrangian sizes in the comoving frame do not change over time . as in @xcite , we use the spherical evolution model to calculate the lagrangian size , @xmath68 , corresponding to the observed eulerian size , @xmath69 , of a region containing the linear overdensity , @xmath38 in a @xmath70 universe .
the extent of the collapse depends on the magnitude of the overdensity .
the more overdense the region is , the larger it would have to be initially in order for it to collapse to the same value of @xmath69 .
similarly , a lower value of the overdensity would mean that the material inside @xmath69 came from a relatively smaller lagrangian size . integrating the equations of motion results in a mapping between the comoving eulerian size of a viewed region and the comoving lagrangian size of the region from where the same material originated in the early universe .
for a fixed value of @xmath69 , there is a one - to - one relationship between @xmath68 and the value of @xmath38 that collapses @xmath68 to @xmath69 .
@xmath68 is then associated with @xmath39 . since matter shells in the spherical collapse model do not cross until collapse , the amount of matter inside @xmath68 is the same as that inside @xmath69 .
the mass contained within the observed size @xmath69 is @xmath71 where the initial value of the overdensity is @xmath72 , @xmath73 is the initial redshift before the region begins to evolve nonlinearly , and @xmath74 is the observed redshift of the region .
the nonlinear matter overdensity is then @xmath75 , and the nonlinear bias is @xmath76 , where @xmath77 and @xmath78 .
@xmath79 is the unconditional , st mass function .
to calculate the mass and overdensity described in the previous section for the ssc , we need to know the number of the collapsed objects in the ssc , their minimum mass , and the size of the region in which they reside .
since the mass function given by eps and similar prescriptions ignores substructure , we must be careful to consider only the largest structures .
fortunately , the largest collapsed halos are also the most sensitive tracers of the mass function , being on its exponential tail .
we consider the most massive , x - ray luminous clusters in the supercluster as tracers of the most massive halos and make the assumption that each halo hosts one such cluster .
this is a good approximation given how well the average st halo mass function in the universe matches the cluster mass function @xcite .
our tendency to focus on only the rarest objects is moderated by the need to reduce poisson fluctuations on our sample . to balance these requirements
, we construct a sample of clusters in the ssc whose host halos have masses above @xmath80 .
the sample consists primarily of clusters studied by @xcite , but contains an additional cluster from the _ clusters in the zone of avoidance _ ( ciza )
sample @xcite .
we have approximated the distance between each of the clusters in our sample and the nominal center of the ssc , which we have taken to be a3558 . to estimate these distances , we assume that each cluster has the same peculiar velocity as a3558 , measured by @xcite .
this is nearly equivalent to assuming no relative peculiar velocity among the clusters . as we will see in [ dynamics ] this approximation is adequate for our purposes .
our entire sample extends to a radius of roughly @xmath81 around a3558 . at a distance of @xmath82 , @xmath1 perpendicular to
the line - of - sight corresponds to @xmath83 .
we assume that the clusters newly detected by @xcite have @xmath84 .
we make the same assumption about @xmath85 .
we exclude the object denoted as b8 by @xcite because it is not a confirmed cluster " .
table [ sample ] gives our resulting sample of @xmath86 clusters .
it includes a list of the clusters , positions , redshifts , estimated distances to a3558 , measurements of the x - ray flux and temperature ( @xmath87 and @xmath88 , respectively ) , and the resulting estimates of the host halo mass , @xmath63 and @xmath66 , which are defined below .
our sample is similar to the @xmath89 clusters studied by @xcite in the ssc in number and extent . here
, @xmath63 is the mass of the halo out to the radius within which the matter density is @xmath90 times the mean matter density used sometimes in the literature to refer to the halo mass out to the radius containing an average matter density equal to @xmath91 times the critical density ( eg .
@xcite ) . for our cosmology and at the redshift of the ssc , this notation corresponds to @xmath63 used here and elsewhere in the literature ( eg .
@xcite ) where the average matter density in the sphere is @xmath90 times the mean matter density of the universe . ] . to calculate @xmath63 , we make use of its relationship to x - ray temperature @xmath92 @xcite : @xmath93 for ciza j1410.4 - 4246
, we calculate @xmath63 from the x - ray flux via the relation @xcite : @xmath94 where the x - ray luminosity , @xmath95 , is in units of @xmath96 , @xmath97 is the luminosity distance , @xmath63 is in units of @xmath64 , and @xmath98 .
@xmath66 , on the other hand , is the mass for which eps - like prescriptions generate the mass function and corresponds roughly to the mass out to the radius within which the matter density is @xmath99 times the mean density .
the relationship between various definitions of mass can easily be calculated @xcite .
values of @xmath63 underestimate those of @xmath66 by @xmath100 ( i.e. @xmath65 ) .
we estimate errors in these masses to be about @xmath67 , resulting primarily from error in the temperature measurement and intrinsic scatter in the @xmath101 and @xmath102 relations from equations ( [ m - t ] ) and ( [ m - l ] ) . because the clusters under consideration are very massive and luminous
, we ignore the effect of the cluster selection function .
the flux of our dimmest cluster is bright enough to be seen even in regions where rosat sensitivity is slightly lower ( i.e. in the @xmath103 patch centered at ( ra : dec)@xmath104(199.295:-34.393 ) ) .
however , absorption by the galaxy could hide a few additional clusters . on the other hand , because the core of the ssc is clearly more overdense than the region out to @xmath105 , our assumption that each cluster represents one halo may not hold there .
multiple clusters in the core may lie within a single halo ( or multiple halos in the process of merging ) .
thus , the number of halos may be somewhat different than the @xmath86 we assume .
our analysis , however , is independent of the individual masses of these halos and depends only on the minimum halo mass of our sample and the total number of halos .
because the error in our mass measurements is only @xmath106 , errors in the minimum mass can be described by additional error in the number of clusters being above a fixed threshold .
in addition , we expect the variation in @xmath50 due to errors in @xmath66 to be small .
if it turns out that all of our mass estimates are systematically @xmath106 too high , then only two clusters below the limit would have made it into our sample undeserved . in [ robust ] ,
we show the dependence of our results on the exact value of @xmath50 .
in figure [ dist ] , we plot the distributions of several quantities calculated through the prescription outlined in [ method ] or derived from the resulting quantities .
panels ( a)-(i ) show results for the expected number of collapsed objects above @xmath107 , the linear matter overdensity , @xmath38 , the total mass of the ssc , @xmath108 , the nonlinear matter overdensity , @xmath109 , the cluster overdensity , @xmath110 , the cluster bias , @xmath111 , the lagrangian radius of the ssc , @xmath68 , the contribution of the ssc to the peculiar velocity of the lg , @xmath112 , and the peculiar radial velocity of the outer edge of the ssc , @xmath113 . the distribution of @xmath49 in panel ( a ) shows a mean clearly less than our assumed value of @xmath114 .
this is due to the @xmath115 factor in equation ( [ pn ] ) . if , in the absence of information about @xmath50 , each value of @xmath49 were equally likely , we would have @xmath116 . however , including the true prior probability distribution means that lower values of @xmath49 are more likely .
the linear overdensity can be interpreted given the value of @xmath117 on the scale of the supercluster . for @xmath118 ( @xmath119 ) , @xmath120 ( @xmath121 ) .
the mean linear overdensity of @xmath122 in panel ( b ) represents a @xmath123 ( @xmath124 ) fluctuation in the overdensity in the region . as a sanity check
, we note that the probability of a linear density as high or higher , assuming initial gaussian fluctuations , is approximately @xmath125 ( @xmath126 ) , and there are about @xmath127 ^ 3\sim60 $ ] ( @xmath128 ^ 3\sim40 $ ] ) such regions out to the distance of the ssc .
so it is not unlikely that we do observe such a region within @xmath129 , nor is it so unlikely that the supercluster remains one of the most overdense regions out to that distance .
our analysis gives @xmath0 for the mass of the ssc .
this is similar to the mass in the region estimated by @xcite , and only a bit lower than estimates by @xcite and @xcite .
panels ( d)-(f ) show the probability distributions of the matter overdensity , @xmath109 , the cluster overdensity , @xmath110 , and the cluster bias , @xmath111 .
the nonlinear overdensity distribution shows @xmath130 and @xmath131 .
larger values of @xmath109 correspond to lower values of @xmath110 and @xmath111 since the number of clusters is fixed and so more matter in the region gives closer agreement between the matter and cluster densities .
the bias is @xmath132 , however , the distribution is not gaussian and is skewed toward values higher than its peak at @xmath133 .
this peak value is just over @xmath134 higher than the maximum of the range of @xmath135@xmath136 that @xcite estimated ( for @xmath137 ) by assuming that their calculation of the peculiar velocity of the lg induced by their sample is equal to the true value .
we calculate only a @xmath138 chance that the bias falls into the @xmath135@xmath136 range . for reference
, we find that @xmath139 corresponds to @xmath140 .
the discrepancy in the bias may result from differences between the clusters associated with the scc and the entire @xcite sample .
the entire sample includes objects that are less intrinsically luminous ( i.e. less massive ) than those in the ssc .
while we used the spherical collapse model in previous sections to calculate @xmath68 corresponding to @xmath69 for each value of @xmath38 , we now look at the results from the model in detail about the dynamics of the ssc region .
we describe our analysis in [ collapse ] and possible tests for our results in [ testing ] .
the evolution of the physical size @xmath141 of a region containing mass @xmath108 is described by : @xmath142 @xmath143 where @xmath108 is the mass in the region given by equation ( [ mass ] ) , @xmath144 , @xmath141 is the evolving eulerian size of the region , and where the evolution of the hubble parameter is given by the friedmann equation , @xmath145 the evolution of the matter and dark energy density parameters ( @xmath146 and @xmath147 , respectively ) are given by @xmath148 subscripts 0 " refer to values today , and subscripts
i " refer to initial values at @xmath149 .
we take @xmath150 , set @xmath151 ( i.e. @xmath152 ) and @xmath153 as our initial conditions , and numerically integrate equation ( [ sphcol ] ) up to @xmath154 , corresponding to the hubble flow redshift of a3558 . the evolution of spheres around the center of the ssc as a function of the scale factor , @xmath155 , calculated from the spherical collapse model .
results are shown for radii of @xmath156 ( solid line ) , @xmath157 ( dot - dashed line ) , @xmath158 ( long - dashed line ) , and @xmath159 ( short - dashed line ) . since @xmath160 today ,
values of @xmath161 refer to time in the future .
the vertical line denotes the redshift of the ssc .
the mean values of the overdensity within the radii shown have been adopted in plotting the trajectories . ] in setting @xmath153 , as appropriate for a growing - mode perturbation , we have assumed that the region is not initially expanding with the hubble flow as in @xcite .
the @xmath162 term accounts for the peculiar velocity of the region due to its overdensity .
since we expect the @xmath163 to be small , however , the peculiar velocity term will also be small .
the probability distribution for the lagrangian radius , @xmath164 , is given in panel ( g ) of figure [ dist ] .
the distribution gives @xmath165 , indicating that the region of the ssc deviates from the hubble flow and collapses by only @xmath166 .
the more interesting result from applying the spherical collapse model to the ssc is that the region is still expanding at a radius of @xmath1 . while not expanding as fast as the hubble flow
, the ssc has not reached turn - around . moreover
, we find that , due to the repulsive effect of the cosmological constant , the region will never reach turn - around despite its overdensity but will continue to expand forever . this can be seen in the trajectory of the @xmath156 region plotted in figure [ collpro ] .
trajectories have also been plotted for the interior regions of the ssc ( see [ radial ] ) .
we arrive at the peculiar radial velocity of the region by subtracting the hubble velocity from the total expansion rate .
the distribution in the peculiar radial velocity of the @xmath105 shell , @xmath113 , is given in panel ( h ) of figure [ dist ] and gives @xmath167 . the minus sign indicates that the region is expanding somewhat slower than the @xmath168 hubble velocity at that radius . our prediction for the outward velocity of the @xmath105 shell could potentially be tested with the tully - fisher relation or by type ia supernovae @xcite . by determining the distance to an object of a known redshift , one may infer its peculiar velocity by subtracting off the hubble velocity at its inferred distance .
the quantity @xmath113 is the radial peculiar velocity with respect to the center of the ssc .
thus , the peculiar velocity of the ssc itself must be accounted for when using the tully - fisher relation to measure @xmath113 .
the latest results from @xcite provide a peculiar velocity for a3558 , the central cluster of the ssc , of @xmath169 .
however , the error on the peculiar velocity of a3558 is too large to constrain @xmath113 reliably .
below we explore alternative observational methods for future measurements of the predicted peculiar velocities .
the peculiar velocity of an x - ray cluster can be directly inferred from its contribution to the cmb through the kinetic sunyaev - zeldovich ( ksz ) effect in which cmb photons scatter off free elections moving along with the bulk velocity of the cluster .
the temperature fluctuation induced in this way by a cluster is given by : @xmath170 where @xmath171 is the optical depth for thomson scattering of the cmb by free electrons in the cluster , @xmath172 is the thomson cross - section , @xmath173 is the number density of electrons in the cluster , @xmath174 is its core radius , and @xmath175 is its peculiar velocity .
the minus sign indicates that velocities away from us result in temperature decrements . because a1644 is on the edge of the ssc ( albeit the edge on the sky
as opposed to along the line of sight ) , we use it as a test case to estimate @xmath176 due to @xmath113 of a hypothetical cluster on the @xmath1 shell and positioned along the line of sight to a3558 .
the radius of a1644 is @xmath177 , while its x - ray luminosity is @xmath178 and its temperature is about @xmath179k @xcite .
accounting for the x - ray luminosity through thermal bremsstrahlung we get @xmath180 .
if the hypothetical cluster under consideration ( whose properties are the same as a1644 ) is @xmath181 in front of a3558 directly along the line of sight so that @xmath113 is also along the line of sight , the cmb temperature fluctuation due to @xmath113 would be @xmath182 with @xmath183 . because the peculiar velocity is radially inward toward the center of the ssc , if the cluster is between us and a3558 , then this velocity is away from us and results in a temperature decrement . the ssc as a whole also induces cmb signals through the ksz effect from to its own peculiar velocity and through the rees - sciama effect @xcite . for an overdensity of @xmath184 , the optical depth to electron scattering through the ssc
is @xmath185 .
the resulting ksz signal due to the peculiar velocity of the structure is @xmath186 . in the rees - sciama effect ,
the evolution of the gravitational potential , @xmath187 , during the photon crossing time of the system introduces additional temperature fluctuations of order @xmath188 .
if the potential becomes deeper ( i.e. more negative ) as a photon passes through the region , the redshift experienced by the photon on its way out of the potential well is greater than the blueshift it receives on its way in and the cmb in that direction shows a temperature decrement . however , between @xmath189 and @xmath190 ( which spans the light crossing time for the ssc of @xmath191 $ ] ) , the ssc grows from @xmath192 to @xmath193 . for @xmath194 , the change in the potential results in @xmath195 .
the ksz and rees - sciama effects due to the ssc as a whole act against each other .
the predicted signal can probed by a number of upcoming cmb observatories such as _ planck _ ( http://planck.esa.int ) , the _
arcminute cosmology bolometer array reciever _ ( http://cosmology.berkeley.edu/group/swlh/acbar ) ,
the _ atacama pathfinder experiment _ ( http://www.mpifr-bonn.mpg.de/div/mm/apex ) , the _ south pole telescope _ ( http://spt.uchicago.edu ) , and the _ atacama cosmology telescope _ ( act , http://www.physics.princeton.edu/act ) . with a temperature noise of @xmath196 per pixel ,
wmap is not sufficiently sensitive to detect this effect . on the other hand
, act has a temperature noise of @xmath197 per pixel and a pixel size of @xmath198 . at a distance of @xmath199 ,
an a1644-like cluster would take up @xmath200 pixels , lowering the detectable level to @xmath201 .
therefore , the ksz signal due to this cluster is , in principle , detectable with act at the @xmath202 level .
though additional sources of error may make this detection more challenging @xcite , the major obstacle to measuring this signal is disentangling it from the primary anisotropy that is at least an order - of - magnitude larger and has an identical spectrum .
it is also important to note , that some type of distance measure from something like the tully - fisher relation or from type ia supernovae is still necessary to determine which clusters are on the @xmath105 shell .
the radial profile of the ssc . the number of observed clusters , @xmath50 , the matter overdensity , @xmath109 , and the amount of excess mass , @xmath203 within @xmath69 as functions of @xmath69 are plotted in the top three panels .
the bottom two panels show the peculiar velocity of the local group due to the overdensity contained within @xmath69 and the radial peculiar velocity of the shell with radius @xmath69 centered on a3558 .
solid lines indicate mean values , while the dashed curves denote @xmath134 bounds due to poisson fluctuations . ]
next we calaculate the contribution of the shapley suplercluster to the peculiar velocity of the lg , @xmath204 , given our result for the ssc mass .
the induced velocity in the linear regime ) is only valid in the linear regime .
however , since the ssc overdensity is modest , the formula gives a good approximation to the true value .
] is given by @xcite @xmath205 where @xmath206 and @xmath207 and @xmath208 point from the center of the ssc to each ssc mass element and to the lg , respectively . because we do not explore the mass distribution within the ssc , we assume a constant value of the overdensity out to @xmath105 . the distribution of values for @xmath112 , reflecting the distribution in the ssc overdensity , is plotted in panel ( i ) of figure [ dist ] .
we find @xmath209 , only @xmath210 of the @xmath211 peculiar velocity of the lg with respect to the cmb @xcite .
this is much less than the value estimated by @xcite of @xmath212 .
even @xmath139 and the corresponding value of @xmath213 ( see [ ssc ] ) produce only @xmath214 , but this is almost @xmath215 from our mean value .
thus , we agree qualitatively with @xcite , @xcite , and @xcite that the ssc does not contribute very significantly to the peculiar velocity of the lg .
the number of clusters in our sample decreases sharply with radius from the center of the ssc , and the clusters are not numerous enough to adequately sample the entire volume of the ssc .
nonetheless , we can obtain some interesting estimates about the radial distribution of the quantities we investigate by considering subsamples of clusters from table [ sample ] with distances from the center of the ssc within varying values of @xmath69 .
figure [ rprofile ] shows the radial profiles we calculate for the number of observed clusters , the matter overdensity and the amount of excess mass within @xmath69 as functions of @xmath69 .
also shown are the peculiar velocity of the lg due to the overdensity contained within @xmath69 and the radial peculiar velocity of the shell with radius @xmath69 centered on a3558 .
the plot of @xmath50 vs. @xmath69 depicts a step - like function where the value changes in discrete steps as observed clusters from the table [ sample ] become included in the subsample within @xmath69 .
the discreteness of the limited sample is manifested in the sharp transitions exhibited in the radial profiles of the other quantities in figure [ rprofile ] .
the true profiles are smooth curves that roughly follow those calculated and are bounded by the @xmath216 error brackets approximately @xmath217 of the time despite the sharp transitions . the uncertainty in the values of @xmath141 , the distance between each cluster and a3558 , shown in table [ sample ] , is not included in the errors plotted in the figure . since @xmath218 , the plots of @xmath109 and @xmath203 vs. @xmath69 are analogous , however , they illustrate two different points .
the radial profile of @xmath109 shows a roughly power - law rise in the overdensity as we move toward the center of the ssc .
the supercluster becomes quite nonlinear in the region interior to @xmath219 .
the radial profile of @xmath203 , on the other hand , demonstrates that the mass only rises significantly through sharp jumps that correspond to the adding of additional clusters to the subsample .
this can also be seen in the plot of @xmath112 vs. @xmath69 where the profile between jumps is extremely flat . moreover , while the contribution to the peculiar velocity of the lg rises most rapidly between @xmath220 and @xmath221 , @xmath112 seems to have leveled off by @xmath222 .
the evolution of @xmath113 with @xmath69 is very mild with its mean varying by only a factor of @xmath223 over the full range of @xmath69 shown .
in addition , the @xmath134 errors on @xmath113 are large , comparable to the range over which @xmath113 varies . however , it is clear that the general trend is toward increasingly negative values of @xmath113 as we move toward the center of the ssc .
this is what is expected from the spherical collapse model .
in addition , by tracing the future of regions at different radii with the spherical collapse model ( see [ collapse ] ) , we determine that the ssc is only a closed system at @xmath224 given the mean of the density probability distribution and at @xmath225 if @xmath134 fluctuations are considered .
the mass enclosed within this radius is comparable to that of a rich x - ray cluster of @xmath226 .
the trajectories of spheres with @xmath227 , @xmath158 , @xmath157 , and @xmath156 are plotted in figure [ collpro ] .
the mean values of the overdensity for each of these radii have been adopted in plotting the trajectories .
in [ method ] , we showed how the matter overdensity in a region can be determined by fitting the mass function in an overdense region to the cluster mass function in the region .
however , because of our limited sample , we only fit the normalization of the mass function and relied on the @xcite hybrid model to correctly fit the slope . in this section ,
we investigate the robustness of our results and their dependence on our sample . in [ nrobust ] , we consider the choice of @xmath50 obtained from the sample , while in [ hmrobust ] , we turn our attention to the slope of the mass function and ask whether different mass bins within the sample give consistent results . the dependence of the distributions for the nonlinear overdensity , @xmath109 , the total mass of the ssc , @xmath24 , the cluster bias , @xmath111 , the contribution of the ssc to the peculiar velocity of the lg , @xmath112 , and the radial peculiar velocity of the outer edge of the ssc , @xmath113 , on the value of @xmath50 .
the square points represent the mean values of the probability distributions , while the error bars show the @xmath216 spread .
the vertical line denotes @xmath114 .
@xmath228 and @xmath229 are held fixed for all values of @xmath50 . ]
the probability distributions of the matter overdensity , @xmath109 , ( upper panel ) and the cluster bias , @xmath111 , ( lower panel ) obtained with four different cluster subsamples from table [ sample ] .
the solid curve shows the results from the entire sample of @xmath86 clusters with masses above @xmath230 shown in figure [ pdel ] .
the dotted curve shows the results obtained using the @xmath89 clusters with masses above @xmath231 , while the @xmath232 clusters with masses above @xmath233 produced the results denoted by the dashed curve .
finally , the results denoted by the long - dashed curve were derived with only the @xmath234 clusters with host halo masses above @xmath235 .
the size of the region was fixed at @xmath229 for each distribution . ]
the number of collapsed halos , their minimum mass , and their extent are the three key parameters that we extract from our cluster sample for use in our structure formation analysis .
though it is probably correct to identify the number of x - ray luminous clusters with the number of collapsed halos , the two may not be exactly equivalent .
in addition , there may be undiscovered objects obscured by galactic absorption . as argued in [ clusters ] ,
the error in the mass estimates of each halo from the cluster x - ray properties can be expressed as an additional but small error in the number of halos above a fixed threshold .
the same can be said for the extent of the ssc , denoted @xmath69 .
small errors in the physical position of a cluster with respect to the center of the ssc can be represented as clusters being falsely included or excluded in our fixed selection of @xmath69 .
we thus explore the dependence of our results on the assumed number of collapsed halos .
figure [ neffect ] shows the mean values of @xmath109 , @xmath111 , @xmath24 , @xmath113 , and @xmath112 from figure [ dist ] , as well as the @xmath134 errors for different choices of @xmath50 . as @xmath50 changes , @xmath107 and @xmath69 are held fixed at the values assumed throughout this paper .
the results clearly do not change much as @xmath50 is varied over a relatively wide range of values . checking
the fit of the mass function explicitly is equivalent to asking whether samples with higher mass thresholds give consistent results for the overdensity in the ssc .
since there are many more lower mass objects than higher mass ones , significant deviations in the number of higher mass objects may not affect the results obtained using all of the clusters .
however , consistency between multiple mass bins shows that the results do not depend on the particular choices we made for our sample .
the probability distributions for the matter overdensity and the cluster bias are shown in figure [ hmbins ] calculated using each of four different mass thresholds .
these distributions are not completely independent because the lower mass thresholds include all of the objects at higher mass . however , since there is a steep drop - off in the number of objects with mass , the distributions are dominated by objects near the mass threshold .
all four mass thresholds give consistent results for the matter overdensity in the ssc , with the distributions clustered around @xmath236 , slightly higher than the mean of the @xmath228 distribution .
it is important to note that the mass threshold does not correlate with mean overdensity .
this is just what we would expect if there is no bias in the number of objects with mass since , according to the model , a single value of the average linear matter overdensity should describe the entire mass function . on the other hand
, we do expect a correlation between the bias and the mass of the objects in the sample .
the figure clearly show this correlation as the mean of each bias distribution increases with the mass threshold .
this is a manifestation of the fact that more massive clusters are more strongly clustered than less massive ones .
we have used the enhanced abundance of x - ray clusters to calculate the mass in the ssc , based on the eps - st model @xcite .
we constructed a sample of @xmath86 x - ray luminous clusters within a radius of @xmath237 in halos with masses above @xmath228 .
we then calculated probability distributions for the matter density contrast of the ssc region , the cluster bias for our sample , and the mass of the ssc .
we demonstrated that even mild values of the overdensity in a region can result in a significant over - abundance of massive clusters .
we found a mass of @xmath238 for the ssc .
our results are in good agreement with previous results @xcite , though we found that the cluster bias is probably higher than that estimated by @xcite .
we then used the spherical collapse model to investigate the dynamics of the ssc .
we found that the comoving size of the region has only collapsed by about @xmath239 from its initial value .
moreover , we concluded that the ssc is not bound at a radius of @xmath237 , despite the significant mass in the region .
the repulsive effect of the cosmological constant provides the extra push against gravity that will keep the region from ever collapsing .
the outer shell is moving radially away from the center with a velocity only @xmath240 slower than the @xmath241 hubble velocity at that radius .
this prediction could be tested with better tully - fisher data from ssc galaxies , or future cmb observations .
we also calculated the contribution of the ssc to the peculiar velocity of the lg , @xmath209 .
this value amounts to only @xmath210 of the lg peculiar motion , much smaller than a recent estimate by @xcite and more in agreement with @xcite , @xcite , and @xcite . despite the uncertainty in investigating the interior of the ssc using smaller subsamples of cluster , exploring the radial profiles of @xmath109 , @xmath112 , and @xmath113 indicated convergence , i.e. that we have included all of the mass in the region in excess of the universal average that contributes to the peculiar velocity of the lg .
moreover , while the @xmath1 region as a whole is only mildly nonlinear , the interior of the ssc becomes highly nonlinear .
the excess mass becomes enough to bind the sphere with radius only @xmath242 .
finally , we showed that our results are robust to errors in our input parameters and the mass threshold of our cluster sample .
subsamples of clusters with different mass thresholds give consistent results for the overdensity in the region and demonstrate an expected increase in bias with mass .
we would like to thank c. jones , k. masters , r. narayan , and especially a. vikhlinin for useful discussions .
thanks also to d. kocevski for his list of x - ray clusters in the ssc .
jm acknowledges support from a national science foundation graduate research fellowship . this research was supported in part by harvard university funds .
e. , dunkley j. , nolta m. r. , bennett c. l. , gold b. , hinshaw g. , jarosik n. , larson d. , limon m. , page l. , spergel d. n. , halpern m. , hill r. s. , kogut a. , meyer s. s. , tucker g. s. , weiland j. l. , wollack e. , wright e. l. , 2008 , arxiv e - prints , 803 | we calculate the density contrast of the shapley supercluster ( ssc ) based on the enhanced abundance of x - ray clusters in it using the extended press - schechter formalism .
we derive a total ssc mass of @xmath0 within a sphere of @xmath1 centered at a distance of about @xmath2 .
the nonlinear fractional density contrast of the sphere is @xmath3 relative to the mean matter density in the universe , but the contrast increases in the interior of the ssc . including the cosmological constant ,
the ssc region is found to be gravitationally unbound .
the ssc contributes only a minor portion ( @xmath4 ) of the peculiar velocity of the local group
. large scale structure of the universe local group |
the final state of a gravitational collapse is expected to be described by a black hole and not by a naked singularity . moreover , at late times , the system should settle down to a stationary regime and since the kerr black hole is expected to be the only stationary black hole in vacuum , the final state of all possible gravitational collapses should approach to a kerr black hole . for simplicity , in this discussion
we are not considering electromagnetic fields and also we are assuming that at some finite time all the matter fields have fallen into the black hole .
the above considerations roughly encompass what is known as the standard picture of the gravitational collapse which , in particular , includes the weak cosmic censorship conjecture . to prove that this heuristic picture is in fact
a consequence of the einstein field equations is one of the most relevant open problems in classical general relativity .
one fruitful strategy to study some aspects of this problem is the following . from the heuristic picture presented above it is possible to deduce some geometric inequalities on the initial conditions for gravitational collapse .
hence , initial conditions that violate these inequalities would automatically provide counter examples for the validity of the standard picture of the gravitational collapse .
in fact , the original intention of this strategy , proposed first by penrose @xcite , was to construct such counter examples .
however it was not possible to find them .
it was then natural trying to prove these inequalities .
such proofs provide an indirect but highly non trivial evidence that the heuristic picture of the gravitational collapse is correct ( see the discussion in @xcite ) .
this kind of inequalities are also interesting by themselves because they provide unexpected mathematical connections between geometric quantities . a prominent example of this idea is the penrose inequality which relates the mass with the area of the black hole horizon on the initial conditions .
an important special case of this inequality has been proved in @xcite @xcite ( see also the review article @xcite ) .
another example of this kind of inequalities is the inequality between mass and angular momentum .
this inequality , which constitute the main subject of the present article , arises as follows .
consider an axially symmetric gravitational collapse .
an important feature of axial symmetry is that axially symmetric waves can not carry angular momentum . in other words : in vacuum , angular momentum is a conserved quantity in axial symmetry .
let us assume that the heuristic picture presented above is correct .
denote by @xmath0 and @xmath1 the mass and angular momentum of the final kerr black hole .
the kerr black hole satisfies the inequality @xmath2 the kerr solution is well defined for any choice of the parameters @xmath0 and @xmath1 , it defines however a black hole only if inequality is satisfied .
let @xmath3 and @xmath4 be the total mass and total angular momentum of the initial conditions .
since gravitational waves carry positive mass we have @xmath5 ( this inequality is of course also valid without the assumption of axial symmetry ) . and because angular momentum is conserved in axial symmetry we have @xmath6 .
hence , in order to reach the inequality at a late time , every initial condition for axially symmetric collapse must satisfy @xmath7 see @xcite for a more detailed physical discussion .
this inequality involves only quantities defined on the initial conditions .
it is expected to hold for every axially symmetric vacuum ( not necessarily stationary ) black hole .
the inequality was studied in a series of articles @xcite @xcite @xcite and finally proved for the case of one black hole in @xcite and @xcite .
there exists however no proof for the case of multiple axially symmetric black holes .
this problem appears to be deeper ( and considerable more difficult ) than the single black hole case . in particular ,
it is related , as we discuss below , with the still open problem of the uniqueness of the kerr black hole among stationary black holes with disconnected horizons .
the main purpose of this article is to provide numerical evidence for the validity of for multiple black holes .
a naive method to test is to take some configuration of axially symmetric black holes and compute numerically the mass and the angular momentum of it . for a given configuration
the relevant parameters are the separation distance between the black holes and the individual angular momentum of them .
but , of course , these parameters ( or any other finite set of parameters ) do not characterize uniquely the initial conditions .
there exists infinitely many configurations with the same parameters , this essentially corresponds to the freedom of including gravitational waves surrounding the black holes .
then , either we find a counter example or this naive method will give a very poor evidence in favor of .
just some isolated points in the space of all possible initial conditions .
fortunately a different approach is possible .
it is based on the variational principle for the inequality presented in @xcite .
this variational principle states that the minimum of the mass of a given configuration with fixed angular momentum is achieved by the associated ( i.e. with the same parameters ) stationary and axially symmetric solution of einstein equations .
hence , in order to prove the inequality for a given configuration it is enough to compute the mass of the corresponding stationary and axially symmetric solution of einstein equations , which is characterized by the separation distance and individual angular momentum of the black holes .
the stationary and axially symmetric einstein equations are non linear elliptic equations . in this article
, we use a heat flow to numerically solve them .
this parabolic flow has two important properties , first for arbitrary data it converges ( as time goes to infinity ) to a stationary and axially symmetric solutions of einstein equations .
second , the mass is monotonically decreasing along the evolution and the angular momentum is conserved ( under appropriate boundary conditions ) .
hence , the flow provides an accurate procedure for computing the minimum of the mass of each possible configuration .
this method is interesting by itself as a method for solving numerically the stationary axially symmetric einstein equations with prescribed boundary conditions , which , up to the best of our knowledge , have not been used so far . for simplicity we will restrict ourselves to configurations with only two black holes , although our method applies for any number of black holes .
for this configuration , the most favorable case to violate the inequality is when the black holes have the same angular momentum pointing in the same direction .
this corresponds to a repulsive spin - spin force between them .
this is also the most favorable case for reaching a stationary solution representing two black holes at equilibrium , because it is in principle conceivable that the repulsive spin force balance the gravitational attraction .
this configuration has only two parameters , the separation distance and the angular momentum .
however , as we will see in the next section , due to the scale invariance of the equations we have only one non trivial parameter , which we chose to be the separation distance .
we can compute the mass for every choice of the separation distance and plot a curve . from the shape of the curve it is clear that , although we can compute only a finite range , the inequality will be satisfied for every separation distance .
for other configurations we proceed in similar way .
then , we obtain fairly strong numerical evidences that the inequality is satisfied for two black holes with any separation distances and any angular momentum .
as we mention above , the heat flow relaxes to a solution of the stationary and axially symmetric einstein vacuum equations .
an important open problem in general relativity is whether the kerr black hole is unique among stationary black holes ( see the recent article @xcite and reference therein ) .
this is essentially the same problem as whether is possible to achieve an equilibrium configuration of multiple black holes in general relativity .
this problem have been studied in @xcite@xcite @xcite@xcite@xcite @xcite . for some limit cases and also for cases with reflection symmetry it has been proved that equilibrium is not possible .
also , from a different perspective , the problem has been studied using exact solutions in @xcite .
again , the conclusion was that equilibrium is not possible for this class of solutions . using the heat flow , in this article we also provide numerical evidences that there is no regular equilibrium solution for two extreme black holes .
this case has not been analyzed previously in the literature .
the plan of the article is the following . in section [ sec : vari - probl - parab ]
we introduce the heat flow and analyze its main properties .
we also discuss the precise form of the conjecture regarding inequality and its relation with the black hole equilibrium problem mentioned above . in section [ sec : techniques ] we discuss the numerical techniques used to solve the parabolic heat flow equations . in section [ sec : results ]
we present our results and in section [ sec : conclusion ] we give some further perspective on the open problems .
finally , for the sake of completeness , we include an appendix [ sec : extr - kerr - solut ] with the explicit form of the extreme kerr solution used in our computations .
consider a vacuum , axially symmetric spacetime .
the axial killing vector defines two geometrical scalars , the square of its norm @xmath8 and the twist potential @xmath9 .
these scalars characterize the spacetime in the following sense .
take a foliation of cauchy surfaces on the spacetime with the corresponding time function .
an initial data set for the spacetime is determined by the value of the functions @xmath10 and the time derivatives @xmath11 on a cauchy surface . the einstein evolution equations essentially reduce to a non - linear system of wave equations for @xmath10 . in appropriate coordinates
, the total mass @xmath3 of the spacetime can be written as a positive integral on a cauchy surface in terms of @xmath10 and @xmath11 .
this integral is the non linear and conserved energy of the system of waves equations ( see @xcite for details ) .
an initial data set is called `` momentary stationary '' if @xmath11 vanished .
stationary data are a particular class of momentary stationary data for which the scalars @xmath12 satisfy a set of elliptic equations ( see equations ( [ eq : haf1])([eq : haf2 ] ) below ) . an important feature of the mass integral is that , for arbitrary data , the associated momentary stationary data has less or equal mass .
that is , there exists a lower bound for the mass that can be written in terms only on @xmath10 and no time derivatives @xmath11 are involved .
this lower bound for the mass plays a key role in order to reduce the proof of the inequality ( [ eq:14a ] ) to a pure variational problem .
it can be written as an integral in @xmath13 as follows ( for details see @xcite , @xcite , @xcite , @xcite ) .
let @xmath14 be cartesian coordinates in @xmath13 ( denoted also by @xmath15 , @xmath16 , @xmath17 ) and let @xmath18 be the associated cylindrical coordinates defined by @xmath19 , @xmath20 .
the positions of the black holes will be prescribed by a finite collection of points @xmath21 located at the @xmath22 axis .
more precisely , the points @xmath21 will represent extra asymptotic ends on the spacetime and they can be associated with the location of the black holes . for a given set of @xmath23 points
@xmath21 we define the separation intervals @xmath24 , @xmath25 , to be the open sets in the axis between @xmath21 and @xmath26 , and we define @xmath27 and @xmath28 as @xmath29 and @xmath30 respectively .
let @xmath31 be the length of @xmath24 for @xmath25 .
see figure [ fig:1 ] .
the length @xmath31 ( which are measured with respect to the cartesian coordinates introduced above ) will be associated with the separation distance between the black holes ( see the discussion in @xcite ) . from the square norm @xmath8 of the killing vector
we define the following function @xmath32 @xmath33 the lower bound of the mass is given by the following functional @xmath34 where @xmath35 is the volume element of @xmath13 , @xmath36 denotes partial derivatives with respect to cartesian coordinates @xmath14 and @xmath37 . as we mentioned above , for
an arbitrary axially symmetric initial data @xmath38 with mass @xmath3 we have that ( see @xcite ) @xmath39 the angular momentum @xmath40 of the end @xmath21 is given by @xmath41 the total angular momentum is defined by @xmath42 note the value of the function @xmath9 at the axis prescribe the angular momentum of the configuration .
the euler - lagrange equations of the functional @xmath43 are given by @xmath44 in these equations @xmath45 denotes the flat laplacian in @xmath13 .
an important property of the functional is that equations ( [ eq : el1])([eq : el2 ] ) correspond to the stationary axially symmetric einstein equations .
we are now in position to formulate the variational approach of the inequality .
the conjecture is the following : [ c:1 ] for arbitrary functions @xmath46 we have @xmath47 where @xmath4 is given by ( [ eq:54 ] ) .
moreover , the equality in implies that the functions @xmath48 correspond to the extreme kerr solution .
that is , for fixed total angular momentum @xmath4 , the extreme kerr solution is the unique absolute minimum of @xmath43 .
the inequality is a direct consequence of this conjecture and .
it is important to emphasize that the number of end points @xmath21 and their corresponding angular momentum @xmath40 are not fixed .
that is , the conjecture states that for fixed @xmath4 , extreme kerr is the unique absolute minimum among all possible functions @xmath46 and among all possible configurations of ends @xmath21 with individual angular momentum @xmath40 .
note that in order to have a non zero @xmath4 we need at least one end point .
this is a singular variational problem since a non zero @xmath4 implies ( by equation ) that at least one @xmath40 is non zero , then equation implies that @xmath9 is discontinuous at @xmath21 and hence has infinity gradient at this point . in order to make the second term in the integral finite
the function @xmath32 should diverge at @xmath21 to compensate the divergence of the gradient of @xmath9 .
also , the singularity of @xmath32 at @xmath21 can not be too severe because the first term in the integral should remain bounded .
in the formulation of the conjecture we did not specify the functional space of admissible functions @xmath46 for the variational problem . as we mentioned above
, the functions are typically singular at @xmath21 , and hence the prescription of the appropriate functional space can be quite subtle .
we will no discuss this issue here since it is beyond the scope of this article . for our present purpose ,
it is enough to assume some space of admissible functions which is regular enough in order that the integral is well defined but it is also compatible with the singular boundary conditions ( [ eq:54 ] ) ( for a discussion regarding this point see @xcite ) conjecture [ c:1 ] was proved for the case @xmath49 in @xcite and @xcite .
the case @xmath50 is open .
remarkably , for general @xmath51 in @xcite it has been proved that if the ends @xmath21 and the individual angular momentum @xmath40 are fixed then there exists a unique minimum of the functional @xmath43 .
this minimum satisfies the euler - lagrange equations ( [ eq : el1])([eq : el2 ] ) .
that is , for fixed @xmath21 and @xmath40 , there exists functions @xmath52 , solutions of ( [ eq : el1])([eq : el2 ] ) , where @xmath53 satisfies ( [ eq:27 ] ) , such that @xmath54 for all admissible functions @xmath46 where @xmath9 satisfies the boundary condition and we have defined @xmath55 what is not known is the value of @xmath56 . in particular , it is not known if this minimum satisfies the inequality ( [ eq : inm ] ) for @xmath50 .
a natural strategy to prove the conjecture is to prove that for arbitrary @xmath21 and @xmath40 the minimum @xmath56 satisfies ( [ eq : inm ] ) .
the main goal of this article is to compute numerically this value for different configurations , showing that it satisfies the inequality in all considered cases .
asymptotic ends[fig:1],width=151 ] in order to compute @xmath56 we need to calculate the solution @xmath57 of the euler - lagrange equations ( [ eq : el1])([eq : el2 ] ) with boundary conditions . as an efficient method for computing numerically both the solution and the value of the energy @xmath43
we propose a heat flow defined as follows .
we consider functions @xmath48 which depend on the coordinates @xmath14 and an extra parameter @xmath58 .
then , we define the following flow @xmath59 where a dot denotes partial derivative with respect to @xmath58 .
that is , we have added time derivatives to the right hand side of equations ( [ eq : el1])([eq : el2 ] ) . equations ( [ eq : haf1])([eq : haf2 ] ) represent the gradient flow of the energy ( [ eq:4 ] ) .
as @xmath60 we expect that the solution of the flow will reach a stationary regime ( i.e. @xmath61 and hence it will provide a solution of equations ( [ eq : el1])([eq : el2 ] ) .
the important property of the flow is that the energy @xmath43 is monotonic under appropriate boundary conditions .
this can be seen as follows .
consider the functional ( [ eq:4 ] ) defined on a bounded domain @xmath62 ( denoted in the following by @xmath63 ) for functions that are solutions of ( [ eq : haf1])([eq : haf2 ] ) and take a time derivative of @xmath63 .
integrating by parts and using the evolution equations ( [ eq : haf1])([eq : haf2 ] ) we obtain @xmath64 where @xmath65 is the area element of the boundary @xmath66 and @xmath67 denotes exterior normal derivative with respect to @xmath66 . by combining of homogeneous neumann boundary conditions @xmath68 or dirichlet boundary conditions @xmath69 for arbitrary time independent functions @xmath70 ( since in this case we get @xmath71 on the boundary ) will make the boundary term in ( [ eq:10 ] ) vanish .
and hence we get that @xmath72 when the domain @xmath62 is @xmath13 we need to prescribe appropriate fall off conditions in order to cancel the boundary term in ( [ eq:10 ] )
. however , as we will discuss in the next section , in the numerical calculations the domain @xmath62 is always bounded and hence the boundary conditions ( [ eq:42 ] ) and ( [ eq:11 ] ) will be used .
the procedure to compute the value of @xmath56 will be the following .
we begin with some arbitrary initial data @xmath73 at @xmath74 that satisfies the boundary condition for some fixed configuration of @xmath21 and @xmath40 .
then we solve numerically the flow equations ( [ eq : haf1])([eq : haf2 ] ) .
the mass @xmath75 will decrease with time , and it will reach the minimum value as @xmath60 . this minimum will be of course independent of the initial data .
that is , we expect the following behavior of the solution of the flow equations @xmath76 and @xmath77 note that ( [ eq:43 ] ) implies @xmath78 the monotonicity property ( [ eq:43 ] ) , together with the upper bound ( [ eq:15 ] ) , make the flow equations ideally suited for a numerical study of the inequality ( [ eq : inm ] ) . equations ( [ eq : el1])([eq : el2 ] ) are essentially harmonic maps with singular boundary conditions .
the first existence result for harmonic maps ( in compact manifolds without boundaries ) used a heat flow @xcite . in that reference the behavior ( [ eq:13 ] ) was in fact proved .
there exists also extensions of this result to include regular boundary conditions @xcite .
these works were the motivation for the flow equations ( [ eq : haf1])([eq : haf2 ] ) .
we emphasize however , that the existence results presented in @xcite @xcite for equations ( [ eq : el1])([eq : haf2 ] ) with the singular boundary conditions ( [ eq:27 ] ) ( which are based on @xcite ) do not use a heat flow , they use a direct variational method .
the numerical calculations presented in this article confirm ( [ eq:13 ] ) and hence suggest that a similar existence result can be proved using the present heat flow .
there is some freedom to construct a heat flow out of equations ( [ eq : el1])([eq : el2 ] ) in such a way that @xmath43 is monotonic under the evolution .
namely we can multiply by arbitrary positive functions the left hand side of ( [ eq : haf1])([eq : haf2 ] ) and we still have that @xmath79 is negative .
the choice made in ( [ eq : haf1])([eq : haf2 ] ) appears to be the simplest one because the principal part of the equations are given by heat equations . in effect
, we can write equations ( [ eq : haf1])([eq : haf2 ] ) as follows @xmath80 we also note that in equation ( [ eq : haf1 ] ) we can apply the maximum principle for parabolic equations ( see , for example , @xcite ) to conclude that @xmath81 will be positive for all @xmath58 if the initial data and boundary conditions are positive . in this article
, the flow ( [ eq : haf1])([eq : haf2b ] ) is used as an auxiliary method for computing a solution of the einstein stationary equations .
it is however interesting to point out the relation of this flow with einstein evolution equations . as we mention above , in axially symmetry einstein equations reduce , in an appropriate gauge , to a system of wave equations for @xmath82 .
more precisely , these equations have the structure of `` waves maps '' ( see , for example , @xcite for the definition of wave maps ) .
the initial conditions for these equations are essentially the value of @xmath82 and the value of the time derivative @xmath83 on a cauchy surface . for a typical collapse
initial data , the system will radiate gravitational waves and reach a final stationary black hole of mass @xmath0 . the initial energy of the system is given by the total mass @xmath3 and it is conserved along the evolution ( see @xcite for a discussion on this issue ) . the total angular momentum @xmath4 is also conserved along the evolution .
we always have @xmath5 .
these data can be also evolved using the heat flow . in this case
the data are only the value of @xmath82 at some time .
the total energy of the system if given by @xmath43 , and we have seen that @xmath84 with equality for momentary stationary data .
the system will dissipate energy and reach a final stationary regime with final energy @xmath56 .
we have that @xmath85 .
for the two cases , the system will reach a solution of the einstein stationary equations at late time .
these solutions are different , and there is a priori no obvious relation between them . in particular , there is no obvious relation between @xmath0 and @xmath56 .
the analogy presented above corresponds essentially to the relation between wave maps , heat flows and harmonic maps which represents a geometric generalization of the relation between wave equation , heat equation and laplace equation . for the case without symmetries
it is not possible to reduce einstein equation to a wave map but the analogy can still be made if we use the ricci flow instead of the heat flow .
note however , that in our case the parabolic equations , although non - linear , are much simpler than the ricci flow equations . for a further discussion about this analogy see @xcite .
the flow equations will provide a numeric solution @xmath73 of equations ( [ eq : el1])([eq : el2 ] ) .
as we will see below , the functions @xmath82 determine the complete metric of an stationary axially symmetric spacetime .
however , although the solution @xmath82 is always regular outside the ends @xmath21 , it turns out that the other components of the metric are , in general , not regular at the axis .
that is , not all solutions @xmath82 will produce a regular spacetime metric . in particular , it is expected that a solution @xmath82 that correspond to many black holes ( i.e. @xmath50 in our setting ) do not lead to a regular metric . as we mentioned in the introduction , this is a relevant point in the black hole uniqueness theorem .
this is precisely what we observe in the numerical computations presented in section [ sec : results ] .
we emphasize however that in order to test the inequality ( [ eq:14a ] ) we only need to compute the energy @xmath43 which depends only on @xmath82 and not on the other components of the spacetime metric .
in particular , the energy @xmath43 is not affected by the possible singular behavior at the axis of the other components of the metric . to reconstruct the spacetime metric from @xmath82 we follow @xcite .
assume that @xmath82 are solutions of equations ( [ eq : el1])([eq : el2 ] ) .
then , we can define , up to constants , the following functions @xmath62 and @xmath86 by @xmath87 and @xmath88 the spacetime metric , in coordinates @xmath89 , is given by @xmath90 where @xmath8 is given in terms of @xmath32 by and the functions @xmath91 , @xmath92 and @xmath93 are defined by @xmath94 all functions depend only on @xmath95 .
the two killing vectors of the metric are given @xmath96 and we have @xmath97 where @xmath98 are spacetime indexes .
we also have that @xmath99 is the twist potential of @xmath100 ( see @xcite ) .
the metric will be regular at the axis if the following condition is satisfied @xmath101 for arbitrary solutions @xmath46 this condition will not be satisfied and hence the metric will not define a regular solutions of einstein equations .
the singularities at the axis of these kind of metrics are interpreted as the forces needed to balance the gravitational attraction and keep the bodies in equilibrium ( see @xcite for details ) .
the regularity condition can be conveniently written in term of a function @xmath102 defined by , we have the relation @xmath103 @xmath104 this function satisfies the following equations @xmath105 condition implies that @xmath106 if the regularity condition fails , we can calculate the value of @xmath102 at each component of the axis @xmath107 these values are calculated integrating the gradients with an appropriate path , see figure [ fig:2 ] .
the force between the black holes is given by @xmath108 and @xmath109 .
the dashed line indicates a typical path for the integration of the function @xmath102.[fig:2],width=151 ] finally , we discuss an important property of the stationary equations ( [ eq : el1])([eq : el2 ] ) , namely their scale invariance ( see @xcite ) .
let @xmath110 be a real number .
given functions @xmath81 and @xmath99 we define the rescaled functions @xmath111 and @xmath112 by @xmath113 the functions @xmath114 define solutions of equations ( [ eq : el1])([eq : el2 ] ) with respect to the rescaled coordinates @xmath115 . under this scaling ,
the physical parameters rescale as @xmath116 and @xmath117 note that the quotient @xmath118 is scale invariant . in particular , for the case of two black holes , with parameters @xmath119 , @xmath1 and @xmath120 ,
the scale invariance of the solution implies that only two parameters are non - trivial .
we analyze in this section how to solve numerically the flow equations . although the problem has axial symmetry , equations are written in @xmath13 ( the laplace operator corresponds to the flat laplacian in @xmath13 ) .
we can solve these equations for arbitrary data , with or without axial symmetry
. the minimum will be axially symmetric for any choice of initial data .
this is of course possible because the boundary conditions ( [ eq:27 ] ) are axially symmetric .
the above considerations suggest that we can solve numerically the flow equations in @xmath13 .
this approach has the advantage that no extra boundary conditions on the axis are needed and also that the equations look more regular in these coordinates .
however , from the numerical point of view , this method has two major disadvantages .
the first one is that there is a significant loss of resolution because a 3-dimensional grid is used instead of 2-dimensional one .
second , the functions are singular at the end points @xmath21 and those points are inside the domain .
that is , there are grid points at both sides of a singularity and this is very problematic for the finite difference method .
we found that it is much more convenient to work in a 2-dimensional grid using cylindrical coordinates and imposing appropriate boundary conditions on the axis , as we describe in the following .
the only disadvantage of this approach is that we need to handle terms which are formally singular at the axis .
a typical example is the term @xmath121 which appears in the cylindrical form of the laplacian in @xmath13 , namely @xmath122 however , following @xcite @xcite , this kind of terms can be handle numerically in a very satisfactory manner as we will describe below .
consider @xmath123 with coordinates @xmath95 .
the domain of interest for our problem is the half plane @xmath124 .
the axis @xmath125 is a boundary of the domain . to simplify the notation and the discussion we will focus on the two black hole problem ( i.e. we will have only two end points @xmath126 and @xmath109 separated by a distance @xmath127 ) .
we emphasize however that the following discussion trivially extends to the general case . in order to handle the singular behavior of the functions at the points @xmath21 located on the axis
, we decompose the solution as follows .
let @xmath128 be the extreme kerr solution ( see appendix [ sec : extr - kerr - solut ] ) centered at the end @xmath126 with angular momentum @xmath1 . and
let @xmath129 be the extreme kerr solution centered at @xmath109 with angular momentum @xmath119 . instead of working with @xmath82 , which are singular at @xmath21
, we will work with @xmath130 defined by @xmath131
the idea is that all the singular behavior of the functions are contained in @xmath132 and @xmath133 .
we expect the functions @xmath134 to be regular during the evolution .
if we insert the ansatz into the flow equations and use the fact that each pair @xmath132 and @xmath133 are solutions of the stationary equations ( [ eq : el1])([eq : el2 ] ) , we obtain the following equations for @xmath134 @xmath135 and @xmath136 these are the equations that we actually solve .
let us analyze the boundary conditions for equations ( [ eq : haf1d])([eq : haf2d ] ) .
we begin with the axis .
the boundary conditions for the function @xmath137 at the axis are given by the regularity conditions .
that is , @xmath137 should be a regular function in @xmath13 and hence it should depend smoothly on @xmath138 ( see , for example , @xcite @xcite for a discussion on regularity conditions at the axis for axially symmetric problems ) and then at the axis it must satisfy @xmath139 we use equation as neumann boundary conditions at the axis . for the function @xmath140
the boundary conditions should be such that they do not change the angular momentum during the evolution . since the angular momentum is prescribed by the value of @xmath9 at the axis ,
the natural choice is that the angular momentum is fixed by the values of @xmath141 and @xmath142 at the axis .
hence the appropriate boundary condition for @xmath140 at the axis is the homogeneous dirichlet one @xmath143 if we consider the whole half plane @xmath124 as domain , then we need to prescribe fall off conditions for @xmath144 and @xmath145 at infinity compatible with the asymptotic flatness of the solutions ( see @xcite ) .
in particular , the solutions and its first derivative should go to zero at infinity . in our case
, since the grid is always finite , we need to consider a bounded domain
. the domain will be the rectangle @xmath146 and @xmath147 , where @xmath148 and @xmath149 are arbitrary positive constants ( see figure [ fig:2 ] ) .
let us denote by @xmath150 the part of the boundary that does not contain the axis @xmath125 . we need to prescribe boundary conditions on @xmath150 .
these boundary conditions should have two important properties .
first , they should imply that the energy on the domain is monotonic under the evolution .
second , in the limit @xmath151 they should be compatible with asymptotic flatness . that it , in this limit we want to recover the complete solution on the half plane
. for @xmath137 we can in principle chose between ( [ eq:42 ] ) or ( [ eq:11 ] ) .
note however that an homogeneous neumann boundary condition for @xmath32 translate into an inhomogeneous neumann boundary condition for @xmath137 ( since they are related by equations ( [ eq:23 ] ) ) .
hence , the simpler choice is the homogeneous dirichlet condition @xmath152 with this choice is also simpler to extend the function to the whole half plane as we will see below . for @xmath140 we can not prescribe neumann boundary condition on @xmath150 since
if we do so we can not control the value of @xmath140 at the points @xmath153 where @xmath150 touch the axis @xmath125 .
in particular , this will be incompatible with ( [ eq:3b ] ) .
hence , the only possibility is to prescribe an homogeneous dirichlet condition @xmath154 that is , our set of boundary condition for the numerical evolution is given by ( [ eq:3 ] ) , ( [ eq:3b ] ) , ( [ eq:7 ] ) and ( [ eq:7b ] ) . the variational problem formulated in section [ sec : vari - probl - parab ]
uses @xmath13 as domain , which is equivalent , by the axial symmetry , to the the half plane @xmath155 .
the fact that in every numerical computation only a finite domain can be used will of course introduce an error . in general it is not easy to measure this error . however , in our case , the variational characterization of @xmath56 implies that an upper bound of this quantity is always obtained even using a finite grid .
this can be seen as follows .
consider the functions @xmath134 obtained in the numerical evolution of the flow equations ( [ eq : haf1d])([eq : haf2d ] ) in the bounded domain @xmath62 .
these functions are , in principle , only defined in @xmath62 .
however , we can extend them to @xmath13 imposing that they vanish outside @xmath62 . and hence
, by ( [ eq:23 ] ) , we get functions @xmath73 defined in @xmath13 .
since @xmath156 vanish on @xmath66 ( by the boundary conditions ( [ eq:7])([eq:7b ] ) ) this extension will be continuous but of course , in general , it will not be differentiable at the boundary @xmath66 .
however the extension is weakly differentiable .
moreover the weak derivative is square integrable ( see , for example , @xcite for the definition of weak derivative and also for the proof of this fact ) .
hence , for the extended functions @xmath46 the integral is well defined in @xmath13 and they satisfy the boundary conditions ( [ eq:27 ] ) .
that is , they represent admissible test functions for the variational problem . since @xmath157 is a minimum we have @xmath158 where we emphasize that in this equation @xmath46 are the extended functions .
note that in @xmath159 we have @xmath160 and hence we can decompose the integral @xmath161 as follows @xmath162 the first integral will be the result of the numerical computations using the heat flow .
the second integral depend only on the explicit functions @xmath163 .
we computed this integral using maple 9.5 .
we now describe the way we carry out the numerical computations .
we use a finite difference scheme to solve the initial - boundary - value problem ( ibvp ) given by the equations ( [ eq : haf1d])([eq : haf2d ] ) and boundary conditions ( [ eq:3])([eq:7b ] ) .
we also perform numerical integrations on the computed solutions to evaluate both the mass @xmath164 and the function @xmath165 ( see eq .
( [ eq:31 ] ) ) used to evaluate the force ( [ eq:30 ] ) .
the ibvp is written in cylindrical coordinates @xmath95 on the domain @xmath166 and @xmath167 given two integers @xmath168 and @xmath169 we define the step - size in the @xmath170 and @xmath22 direction respectively as @xmath171 and @xmath172 our equations have singular coefficients at the points @xmath126 and @xmath109 on the @xmath125 axis .
these point will be placed , in all our runs , at positions @xmath173 @xmath174 the computational grid is defined so that the gridpoint at the site @xmath175 has coordinates @xmath176 in this way the uniform grid is half a step size displaced with respect to the coordinate axes and singular points .
one can think that the domain is broken into @xmath177 cells being each grid point with @xmath178 and @xmath179 placed at the center of a cell .
the gridpoints with @xmath180 and @xmath181 are gridpoints at the center of `` ghost cells '' used to impose boundary conditions at the @xmath182 parts of the boundary .
analogously , the gridpoints with @xmath183 and @xmath184 are gridpoints at the center of `` ghost cells '' used to impose boundary conditions at the @xmath185 parts of the boundary .
the four gridpoints at each corner of the grid are not used .
many of the problems we actually compute are symmetric or antisymmetric with respect to the @xmath186 plane . in these cases the symmetry is used explicitly to reduce the grid to half - size and so the computer time needed .
the grid covers a domain with @xmath187 ( see figure [ fig:4 ] ) . in our numerical scheme , the four partial derivatives @xmath188
@xmath189 @xmath190 and @xmath191 of @xmath137 and @xmath192 are approximated by the standard fourth - order accurate symmetric difference operators @xcite @xmath193 here @xmath194 and @xmath195 denote , as usual , the forward and backward difference operators , i.e. , if @xmath196 is a grid function on a 1-dimensional grid with step size @xmath197 , we have @xmath198 and @xmath199 to be more explicit we show the approximations to the derivatives with respect to @xmath200 if @xmath201 i.e. , @xmath202 denotes the grid function associated to the smooth function @xmath203 , then @xmath204 @xmath205 to carry out the time evolution we use the du fort - frankel method .
this method is known to be a good choice for solving parabolic problems because it is explicit and nevertheless unconditionally stable @xcite at least when applied for solving an initial value problem . in the notation of @xcite ( or (
7.3 ) ) we set @xmath206 the @xmath86 parameter in this case has to be chosen bigger than @xmath207 for the method to be stable .
the time step can not be chosen big though , and the reason is twofold .
first a big time step gives rise to an increasing parasitic solution @xcite and more important , the boundary conditions also impose stability restrictions . in the way we treat the boundary conditions ( explained below ) we have a scheme that is stable as can be seen explicitly in our runs , but this scheme is probably not unconditionally stable .
experimentally we did some runs with a big time step and could see how the solution diverges in few time steps starting at the boundaries ( around the singular points @xmath126 and @xmath109 ) . in
most of our computations we use @xmath208 and a time - step @xmath209 i.e. the square of the space step - size which is the normal ratio in explicit schemes for parabolic problems .
this time step is however , as the equations have singular coefficients , around ten times bigger than the time step we could use with other explicit schemes like 3rd order runge - kutta .
the du fort - frankel scheme is only second order accurate but this posses no inconvenience since we are looking for the stationary solution of the parabolic problem . in this case , the truncation error due to the time discretization vanishes when the solution approaches the time independent state .
all the boundary conditions we use are either homogeneous dirichlet or homogeneous neumann boundary conditions .
the boundary conditions are imposed to a grid function via the points at the ghost cells ( see for example @xcite ) .
we show , as example , how this is done for boundary conditions ( [ eq:3 ] ) and ( [ eq:3b ] ) . given the values of @xmath210 and @xmath211 in the interior of the domain , i.e. , for @xmath212 and @xmath181 , the values at the ghost cells with @xmath213 are defined as @xmath214 in this way the boundary conditions ( [ eq:3 ] ) and ( [ eq:3b ] ) are satisfied exactly to the accuracy order of our computations and the same difference operators can be used at all gridpoints inside the domain . as we are using the fourth order accurate operators defined in ( [ eq : op1 ] ) , ( [ eq : op2 ] ) , which have a span of @xmath215 gridpoints , we need two lines of ghost cells outside the domain for each part of the boundary .
we start the time evolution with with initial data that satisfies the right boundary conditions .
now , given the solution at time @xmath58 satisfying the right boundary conditions the right hand side of the equations can be computed in the interior of the domain and the time evolution algorithm computes the values of @xmath144 and @xmath216 in the interior of the domain , at the next time @xmath217 then the solution at this time is extended to the ghost cells so that it obeys the right boundary conditions and the process is iterated .
different criteria can be used to stop the time evolution when one is looking for the stationary state . as the main quantity we want to compute in each run is the mass @xmath164 of the final stationary solution
, we stop the run when the derivative of @xmath164 with respect to time becomes , in absolute value , smaller than a given small value .
to compute the mass @xmath164 and the value of @xmath165 we need to approximate two - dimensional and one - dimensional integrals . as the gridpoints are placed at the center of cells that cover the domain of our ibvp , the simplest appropriate rule to approximate these integrals is the midpoint rule .
the integrand in ( [ eq:4 ] ) , when written in cylindrical coordinates , have singular points at @xmath126 and @xmath218 however the midpoint rule provides good enough results .
for example , as in our runs we used vanishing initial data , i.e. , @xmath219 the integral in ( [ eq:4 ] ) becomes an integral of known , given functions .
thus , we could compare the value obtained with our code to the value obtained with a very precise integration rule implemented in maple ; in the worst case the relative difference between these values was smaller than @xmath220 all the solutions of the ibvp we computed can be divided into three groups .
the first group consists of symmetric configurations in which @xmath221 is placed at @xmath222 and @xmath223 placed at @xmath224 within this group we carried out runs for different values of @xmath225 and for different domain sizes .
it is clear in this case the solutions @xmath32 and @xmath9 of ( 13)(14 ) are , respectively , symmetric and antisymmetric as functions of @xmath226 moreover @xmath144 and @xmath192 satisfy the same symmetry during the whole time evolution of our ibvp even on a finite domain provided the domain itself and the initial data are symmetric .
the obvious initial data satisfying all boundary conditions and symmetry is @xmath227 this is what we used in all our runs .
the second group of solutions we computed correspond to antisymmetric configurations in which we placed @xmath221 at @xmath222 and @xmath228 at @xmath224 we carried out runs for different values of @xmath229 the solutions in this group also have a clear symmetry . in this case both @xmath144 and @xmath192 are symmetric as functions of @xmath226 the third group of solutions we computed correspond to asymmetric configurations in which we placed @xmath230 at @xmath231 and @xmath232 at @xmath233 within this group we carried out runs for various values of @xmath234 when computing solutions in the symmetric or antisymmetric configurations we need to compute the solution in half the domain only , @xmath235 $ ] and @xmath236.$ ] @xmath186 becomes a boundary and all we need is to use extra boundary conditions at @xmath186 that obey the symmetry of the solutions .
this boundary conditions are homogeneous neumann for @xmath144 and homogeneous dirichlet for @xmath192 in the symmetric case , and homogeneous neumann for both functions in the antisymmetric case . by using the symmetry of the solution we reduce to one half the computer time needed .
a main issue , from the point of view of the numerical calculations , is to determine the size of the domain where to compute @xmath237 at the same time we need to estimate error we commit in the determination of @xmath238 we attack these questions mainly by studying the symmetric case .
the time evolution of @xmath239 , for different values of @xmath127 , can be seen in figure [ fig : evolution ] .
the initial data in all the runs was set to zero .
the smaller the value of @xmath127 is the bigger the initial @xmath164 is , and also the stronger the equations dissipate so that the code runs for a longer time and the final `` stationary '' @xmath164 turns out to be smaller . for different values of @xmath127 ( number close to each curve ) . in this plot ,
all the runs were stopped when @xmath240 .
the detail shows the evolution for a short while @xmath241.$],title="fig:",width=302 ] with the purpose of evaluating the precision of the values of mass obtained and of determining a convenient domain size to carry out our computations we performed runs with the same physical parameters but on different domains ( and corresponding grids ) .
the results are shown in table [ table : domains ] for the two smallest values of @xmath127 . [ cols="^,^,^,^,^,^,^",options="header " , ] @xmath242 is located at @xmath231 and @xmath119 is located at @xmath233 the semi - domain is @xmath243\times[0,20].$ ] the continuous line is the lower bound according to the conjecture [ c:1],title="fig:",width=302 ]
the main result of this article is given in tables [ table : symmetric ] and [ table : asymmetric ] .
in all cases , we have verified the inequality ( [ eq : inm ] ) .
that is , we have provided strong numerical evidence that this inequality is true for two axially symmetric black holes .
moreover , we have computed a non zero force in the symmetric case ( table [ table : force ] ) and hence we have also provided numerical evidences that the equilibrium is not possible for two extreme black holes .
the monotonic dependence of the total energy @xmath43 in terms of the separation distance @xmath127 plotted in figure [ fig : main_plot ] suggests a possible strategy to prove analytically the inequality ( [ eq : inm ] ) .
namely , to study the neighborhood of @xmath244 of the energy .
in particular , the first step is to prove that @xmath245 at @xmath244 .
since the value of @xmath43 at @xmath244 is known , this will prove the inequality near @xmath244 .
the second step ( probably much more difficult ) will be to prove that @xmath245 for any @xmath127 .
finally , we have also shown that the heat flow equations ( [ eq : haf1])([eq : haf2 ] ) constitute an efficient and simple numerical method to construct solutions of the stationary and axially symmetric einstein equations .
we expect that this method can be used also with other kind of boundary conditions .
s. d. thanks piotr chruciel for useful discussions .
these discussions took place at the institut mittag - leffler , during the program `` geometry , analysis , and general relativity '' , 2008 fall .
s. d. thanks the organizers of this program for the invitation and the hospitality and support of the institut mittag - leffler . the authors want to thank luis lehner and steve liebling for useful discussions that took place at famaf during their visits in 2008 .
s. d. is supported by conicet ( argentina ) .
this work was supported in part by grant pip 6354/05 of conicet ( argentina ) , grant 05/b415 secyt - unc ( argentina ) and the partner group grant of the max planck institute for gravitational physics , albert - einstein - institute ( germany ) .
the extreme kerr black hole corresponds to the limit @xmath246 of the kerr metric , where @xmath3 is the total mass and @xmath4 is the angular momentum of the spacetime .
usually , instead of @xmath4 in the literature the parameter @xmath247 is used , the extreme limit correspond to @xmath248 . using the notation of section [ sec : vari - probl - parab ] , for the extreme kerr black hole we have only one end @xmath126 located at the origin .
the explicit form of the functions @xmath249 are given by @xmath250 where @xmath251 and @xmath252 in these equations , @xmath253 are spherical coordinates in @xmath13 related with the cylindrical coordinates @xmath254 by the standard formulas @xmath255 and @xmath256 . in this equations , @xmath4 is an arbitrary constant .
it gives the angular momentum and it is the only free parameter in this solution . in agreement with equation ( [ eq:27 ] ) , we have @xmath257 note that the angular momentum is given by @xmath258 , the other part of @xmath141 vanishes at the axis .
the important property of the functions @xmath249 is that they are solutions of equations ( [ eq : el1])([eq : el2 ] ) . in the above equations the end point @xmath126 is chosen to be at the origin of the coordinate system .
we have the obvious freedom to translate this point to an arbitrary location .
in particular , the extreme kerr solution centered at the point @xmath109 used in section [ sec : techniques ] is given by @xmath266 | we present numerical evidences for the validity of the inequality between the total mass and the total angular momentum for multiple axially symmetric ( non - stationary ) black holes .
we use a parabolic heat flow to solve numerically the stationary axially symmetric einstein equations . as a by product our method
, we also give numerical evidences that there are no regular solutions of einstein equations that describe two extreme , axially symmetric black holes in equilibrium . |
the process when two monomers separated by a large distance along the polymer chain come close enough to start interacting with each other is called looping .
the loop formation in macromolecules plays an important role in a number of biochemical processes , such as stabilization of globular proteins @xcite , transcriptional regularization of genes @xcite , dna compactification in the nucleus @xcite as well as in number of other processes that involve both synthetic and biological polymers .
statistics of long flexible polymer chains is known to be governed by a set of universal properties , independent on any details of microscopic chemical structure of macromolecules @xcite . in particular
, the averaged end - to - end distance @xmath3 of linear polymer chain obeys the scaling law : @xmath4 here @xmath5 means averaging over an ensemble of possible conformations of macromolecule , @xmath6 is a number of monomers and @xmath7 is the universal critical exponent . in the simplest case of idealized gaussian chain without any interactions between monomers
@xmath8 , whereas in presence of excluded volume interaction this exponent is depending on the space dimension @xmath0 only : @xmath9 @xcite , @xmath10 @xcite , @xmath11 .
moreover , a distance between any two monomers @xmath12 and @xmath13 along a polymer chain ( @xmath14 ) , which are not connected by a chemical bond , scales according to : @xmath15 with the same exponent @xmath7 .
let us consider the distribution function @xmath16 of the distance @xmath17 between any pair of monomers @xmath12 and @xmath13 along a polymer chain . in a simplified case of gaussian polymer
, @xmath16 adopts a gaussian form : @xmath18 with @xmath19 the loop in polymer chain corresponds to `` contact '' between two monomers @xmath12 and @xmath13 , such that @xmath20 . for the probability to find a loop of size @xmath21 ( cyclization probability ) in a gaussian chain
we immediately restore the result of jacobson and stockmayer @xcite : @xmath22 with @xmath23 .
the situation changes drastically when one considers a chain with excluded volume interactions .
the form of @xmath16 in this case is much more complicated and depends , in particular , on positions of monomers @xmath12 and @xmath13 along the chain @xcite .
the probability to find a loop of size @xmath21 in a polymer chain with excluded volume interactions scales as @xcite : @xmath24 with @xmath25 . here
, @xmath26 refers to the case when both @xmath12 and @xmath13 are end monomers , @xmath27 when @xmath12 is an end monomer and @xmath13 is inner one ( or vice versa ) , @xmath28 when both @xmath12 and @xmath13 are inner monomers ( see fig .
1 ) ; and @xmath29 are directly connected with the spectrum of vertex exponents @xmath30 @xcite : @xmath31 in was found , that end monomers of a chain have a higher probability of contact then inner ones @xcite . in particular , in @xmath32 the results of numerical simulations @xcite give : @xmath33 , @xmath34 , @xmath35 and thus : @xmath36 the refined analytical studies based on renormalization group approach give up to the first order of @xmath37 expansion @xcite : @xmath38 note , that in the case of idealized gaussian polymer one restores the result of jacobson and stockmayer ( [ expgaus ] ) : @xmath39 . the question of great importance in polymer physics is how the conformational properties of macromolecules are modified in presence of structural obstacles ( impurities ) .
one encounters such situations when considering polymers in gels or colloidal solutions @xcite or in the crowded environment of biological cells @xcite .
it was shown analytically @xcite and confirmed in numerical simulations @xcite , that structural disorder in the form of randomly distributed point - like defects does not alter the universality class of polymer macromolecules , unless the the concentration of impurities reaches the percolation threshold @xcite .
the density fluctuations of defects may lead to creation of complex fractal structures @xcite .
these peculiarities are captured within the model of so - called long - range correlated disorder . here , the defects are assumed to be correlated on large distances @xmath1 according to a power law with a pair correlation function @xcite : @xmath40 for @xmath41 , such a correlation function describes complex ( fractal ) defects extended in space .
the studies @xcite quantitatively reveal an extent of the effective size and anisotropy of both linear and closed ring macromolecules in presence of such a type of disorder .
schematic presentation of loops with different positions along the polymer chain . , width=302 ] in this concern , it is worthwhile to study a probability of loop formation in the environment with long range correlated disorder , which has not been considered so far . in the present paper ,
we analyze this problem analytically within the frames of continuous chain model , applying the direct polymer renormalization scheme .
the layout of the paper is as follows . in the next section ,
we introduce the model .
the direct polymer renormalization scheme is shortly described in section iii .
we present the results obtained for critical exponents in section iv and end up with giving conclusions in section v.
we start with considering a flexible polymer chain in solution in presence of long - range correlated disorder . within the edwards continuous chain model @xcite , the chain is presented as a continuous path of length @xmath42 , parametrized by @xmath43 , where @xmath44 changes from @xmath45 to @xmath42 . to describe the presence of loop with certain position along the chain ( see fig .
[ fig:37 ] ) , we consider the system of ring polymer connected with @xmath46 linear chains ( with @xmath47 ) , so that @xmath48 corresponds to a loop formed by two end monomers of the chains , @xmath49 a loop formed by inner monomer and end one , @xmath50 an inner loop . thus , our problem can be considered as a particular case of so - called rosette polymer structures , studied recently in ref .
@xcite .
the partition function of the system can thus be presented as : @xmath51 here , @xmath52 denotes functional path integrations over @xmath53 trajectories , the @xmath54-functions describe the fact that one trajectory is closed and that the starting point of all @xmath53 trajectories is fixed , and @xmath55 is a hamiltonian of the system : @xmath56 where the first term describes connectivity of trajectories , the second term describes short - range repulsion between monomers due to excluded volume effect governed by coupling constant @xmath57 and the last one arises due to the presence of disorder in the system and contains a random potential @xmath58 .
let us denote by @xmath59 the average over different realizations of disorder and assume @xcite : @xmath60 where @xmath61 is a corresponding coupling constant . performing the averaging of the partition function ( [ model - con ] ) over different realizations of disorder , taking into account up to the second moment of cumulant expansion and recalling ( [ avv0 ] ) we obtain @xmath62 with an effective hamiltonian : @xmath63 the probabilities of loop formations ( [ probloop ] )
can be thus calculated as : @xmath64 where @xmath65 is the total partition function of linear chain of length @xmath42 given by : @xmath66 .
\label{chain } \end{aligned}\ ] ]
we apply the direct renormalization method developed by des cloizeaux @xcite . the main idea of this technique is to eliminate various divergences , observed in the asymptotic limit of an infinitely long polymer chain ( corresponding to an infinite number of configurations ) by introducing corresponding renormalization factors directly associated with physical quantities .
in particular , the renormalization factor @xmath67 ( with @xmath68 being the set of bare coupling constants ) is introduced via : @xmath69 here , @xmath70 is the partition function of single polymer chain and @xmath71 is the partition function of an idealized gaussian model .
it is established that the number of all possible conformations of polymer chain scales with the weight of a macromolecule parametrised by @xmath42 as : @xmath72 where @xmath73 is the universal critical exponent depending only on the space dimension @xmath0 and @xmath74 is a non - universal fugacity . from the scaling assumption ( [ gamma1 ] )
one finds an estimate for an effective critical exponent @xmath75 governing the scaling behavior of the number of possible configurations as : @xmath76 in a similar way , in our problem we may introduce the factors @xmath77 via : @xmath78 with @xmath25 . recalling the scaling of loop probabilities ( [ probdef ] ) and remembering the fact that in gaussian approximation @xmath79 , one finds an estimate for the effective critical exponents @xmath80 : @xmath81 the critical exponents presented in the form of series expansions in the coupling constants are divergent in the asymptotic limit of large @xmath42 . to eliminate these divergences ,
the renormalization of the coupling constants is performed .
the critical exponents attain finite values when evaluated at a stable fixed point of the renormalization group transformation .
the flows of the renormalized coupling constants are governed by functions @xmath82 : @xmath83 reexpressing @xmath84 in terms of renormalized couplings @xmath85 , the fixed points of renormalization group transformations are given by common zeros of the @xmath86-functions .
stable fixed points govern the asymptotical scaling properties of macromolecules and gives reliable asymptotical values of the critical exponents ( [ nuexp ] ) .
we start with evaluating the partition function @xmath87 in the simplified gauusian case , which is then given by : @xmath88 where @xmath89 . using the fourier - transform of the @xmath54-functions
@xmath90 we receive @xmath91 in gaussian approximation the value of partition function does not depend on @xmath46 , and thus the looping probability is independent on the loop position along the chain according to ref .
@xcite .
diagrammatic presentation of contributions into the partition function @xmath92 up to the first order in the coupling constants.,width=226 ] taking into account the excluded volume effect and presence of disorder , we present the partition function as perturbation theory series in dimensionless coupling constants @xmath93 and @xmath94 : @xmath95 it is convenient to present the contributions into @xmath96 and @xmath97 using the diagrammatic technique ( fig .
[ fig:38 ] ) . here ,
dotted lines denote possible interactions between points @xmath98 and @xmath99 governed by couplings @xmath100 and @xmath101 , and integrations are to be performed over all positions of the segment end points . note that contributions into @xmath102 are given by only the first diagram and correspond to the partition function of the ring polymer @xcite . on the other hand ,
contributions into partition function @xmath103 are given by all diagrams except @xmath104 .
the analytical expressions corresponding to contributions of diagrams @xmath105-@xmath106 into @xmath96 read : @xmath107 performing the integrations , we get : @xmath108,\nonumber\\ & & { f_3}^u_f = z_u \left(2 \pi \frac{s}{1+f}\right)^{-d/2 } ( 2^{2-d/2}-2)\times \nonumber\\ & & \times\left(1-\frac{d}{2}\right)^{-1}\left(2-\frac{d}{2}\right)^{-1},\nonumber\\ & & { f_4}^u_f = z_u \left(2 \pi \frac{s}{1+f}\right)^{-d/2 } \left(1-\frac{d}{2}\right)^{-1}\left(2-\frac{d}{2}\right)^{-1},\nonumber\end{aligned}\ ] ] here , @xmath109 is euler beta - function and @xmath55 is hypergeometric function . as a result , we obtain contributions into the partition function in the form : @xmath110,\nonumber\\ & & z^u_{1;2}=z_u \left(2 \pi \frac{s}{3}\right)^{-d/2}\left [ b\left(1-\frac{d}{2},2-\frac{d}{2}\right)+\right.\nonumber\\ & & \left.+(2^{2-d/2})\left(1-\frac{d}{2}\right)^{-1}\left(2-\frac{d}{2}\right)^{-1}-\right.\nonumber\\ & & \left.-2\left(1-\frac{d}{2}\right)^{-1}b\left(2-\frac{d}{2},2-\frac{d}{2}\right)+\right.\nonumber\\ & & \left .
+ 2\left(1-\frac{d}{2}\right)^{-1 } \left(\frac{5}{4}\right)^{1-d/2}h(1/2;-1+d/2,3/2;1/5)\right].\nonumber\end{aligned}\ ] ] note , that corresponding contributions into @xmath97 are obtained simply by substituting @xmath100 by @xmath101 and @xmath0 by @xmath111 in expressions in square brackets in above relations . performing a double expansion over parameters @xmath112 , @xmath113 we finally receive the expressions for partition functions for the chains with a loop of certain type : @xmath114 in addition , the total partition function of linear chain of length @xmath42 is given by : @xmath115 probabilities of loops formations
are received using the definition ( [ probdef ] ) and read : @xmath116 recalling that ( [ nuexp ] ) can be presented in the form @xmath117 we find for the critical exponents governing the scaling of looping probabilities : @xmath118 taking into account that @xmath100 and @xmath101 in the first order of perturbation theory are proportional to @xmath119 , @xmath54 , and keeping only contributions up to @xmath119 and @xmath54 , the above expressions can be presented as : @xmath120 scaling exponents ( [ purell ] ) and ( [ lrll ] ) as function of @xmath113 in @xmath121.,width=340 ] we make use of results for fixed point values found previously for the linear polymer chains in long - range correlated disorder @xcite .
there are three distinct fixed points governing the properties of macromolecule in various regions of parameters @xmath0 and @xmath111:@xmath122 evaluating eqs .
( [ ll1 ] ) - ( [ ll ] ) at gaussian fixed point , we restore the result of jacobson and stockmayer @xcite : @xmath123 . for polymers with
excluded volume effect in pure solutions we restore results of duplantier @xcite : @xmath124 finally , for the considered case of a polymer chain in disordered environment we received a brand new results : @xmath125 to find the quantitative estimate for the exponents ( [ purell ] ) and ( [ lrll ] ) at @xmath126 , we evaluate the expressions at @xmath127 and various fixed values of @xmath54 ( see fig . [ fig:39 ] ) .
we find , that presence of long - range correlated disorder with any @xmath128 leads to an increase of exponents as compared with corresponding pure values , and thus the probabilities of loop formation decreases in presence of crowded environment . from physical point of view , we can interpret this as follows .
the presence of complex ( fractal ) obstacles in the system forces the macromolecule to avoid these extended regions of space , which results in effective elongation of polymer chain and makes the contact of two monomers along the chain less probable .
in present work we analyzed a probability of loop formation in flexible polymer chains in good solutions , which is known to be governed by scaling laws ( [ probloop ] ) with scaling exponents @xmath129 dependent on the position of loop along the macromolecule .
we considered the special case , when structural obstacles are present in the environment , which are assumed to be correlated on large distances @xmath1 according to a power law with a pair correlation function @xmath130 with @xmath41 @xcite .
the previous studies @xcite reveal the non - trivial impact of such a type of disorder on the universal conformational properties of both linear and closed ring macromolecules .
working within the frames of continuous chain model and applying the direct polymer renormalization scheme , we obtain the values of critical exponents @xmath129 up to the first order of perturbation theory in parameters @xmath112 , @xmath113 .
our results quantitatively reveal that presence of long - range correlated disorder with any @xmath131 leads to an increase of exponents as compared with corresponding pure values , and thus the probabilities of loop formation decreases in presence of crowded environment . | we analyze the probability of a single loop formation in a long flexible polymer chain in disordered environment in @xmath0 dimensions .
the structural defects are considered to be correlated on large distances @xmath1 according to a power law @xmath2 .
working within the frames of continuous chain model and applying the direct polymer renormalization scheme , we obtain the values of critical exponents governing the scaling of probabilities of loop formation with various positions along the chain as function of loops length .
our results quantitatively reveal that the presence of structural defects in environment decreases the probability of loop formation in polymer macromolecules . |
the crime and violence street gangs introduce into neighborhoods is a growing epidemic in cities around the world . today , over 1.23 million people in the united states are members of a _ street gang _ @xcite , which is a coalition of peers , united by mutual interests , with identifiable leadership and internal organization , who act collectively to conduct illegal activity and to control a territory , facility , or enterprise @xcite .
they promote criminal activities such as drug trafficking , assault , robbery , and threatening or intimidating a neighborhood @xcite .
moreover , data from the centers for disease control in the united states suggests that the victims of at least 1.3% of all gang - related homicides are merely innocent bystanders who live in gang occupied neighborhoods @xcite .
street gang members have established online presences coinciding with their physical occupation of neighborhoods .
the national gang threat assessment report confirms that at least tens of thousands of gang members are using social networking websites such as twitter and video sharing websites such as youtube in their daily life @xcite .
they are very active online ; the 2007 national assessment center s survey of gang members found that 25% of individuals in gangs use the internet for at least 4 hours a week @xcite .
gang members typically use social networking sites and social media to develop online respect for their street gang @xcite and to post intimidating , threatening images or videos @xcite .
this `` cyber- '' or `` internet banging '' @xcite behavior is precipitated by the fact that an increasing number of young members of the society are joining gangs @xcite , and these young members have become enamored with technology and with the notion of sharing information quickly and publicly through social media .
stronger police surveillance in the physical spaces where gangs congregate further encourages gang members to seek out virtual spaces such as social media to express their affiliation , to sell drugs , and to celebrate their illegal activities @xcite .
gang members are able to post publicly on twitter without fear of consequences because there are few tools law enforcement can use to surveil this medium @xcite .
police departments across the united states instead rely on manual processes to search social media for gang member profiles and to study their posts .
for example , the new york city police department employs over 300 detectives to combat teen violence triggered by insults , dares , and threats exchanged on social media , and the toronto police department teaches officers about the use of social media in investigations @xcite .
officer training is broadly limited to understanding policies on using twitter in investigations and best practices for data storage @xcite .
the safety and security of city neighborhoods can thus be improved if law enforcement were equipped with intelligent tools to study social media for gang activity .
the need for better tools for law enforcement can not be underscored enough .
recent news reports have shown that many incidents involving gangs start on twitter , escalate over time , and lead to an offline event that could have been prevented by an early warning .
for example , the media reported on a possible connection between the death of a teenage rapper from illinois and the final set of tweets he posted .
one of his last tweets linked to a video of him shouting vulgar words at a rival gang member who , in return , replied _
`` i m a kill you '' _ on social media . in a following tweet ,
the teenage rapper posted _
`` i m on 069 '' _ , revealing his location , and was shot dead soon after that post .
subsequent investigation revealed that the rivalry leading to his death began and was carried out entirely on social media .
other reporting has revealed how innocent bystanders have also become targets in online fights , leaving everyone in a neighborhood at risk .
this paper investigates whether gang member profiles can be identified automatically on twitter , which can enable better surveillance of gang members on social media . classifying
twitter profiles into particular types of users has been done in other contexts @xcite , but gang member profiles pose unique challenges .
for example , many twitter profile classifiers search for contextual clues in tweets and profile descriptions @xcite , but gang member profiles use a rapidly changing lexicon of keywords and phrases that often have only a local , geographic context .
this is illustrated in figure [ fig : twitterprofiles ] , which shows the twitter profile descriptions of two verified deceased gang members .
the profile of @osoarrogantjojo provides evidence that he belongs to a rival gang of the black disciples by ` # bdk ` , a hashtag that is only known to those involved with gang culture in chicago .
s profile mentions ` # pbg ` and our investigations revealed that this hashtag is newly founded and stands for the pooh bear gang , a gang that was formerly known as the insane cutthroat gangsters . given the very local , rapidly changing lexicon of gang members on social media , building a database of keywords , phrases , and other identifiers to find gang members nationally is not feasible .
instead , this study proposes heterogeneous sets of features derived not only from profile and tweet text but also from the emoji usage , profile images , and links to youtube videos reflecting their music culture .
a large set of gang member profiles , obtained through a careful data collection process , is compared against non - gang member profiles to find contrasting features .
experimental results show that using these sets of features , we can build a classifier that has a low false positive rate and a promising @xmath1-score of 0.7755 .
this paper is organized as follows .
section [ sec : rr ] discusses the related literature and positions how this work differs from other related works .
section [ sec : dc ] discusses the data collection , manual feature selection and our approach to identify gang member profiles .
section [ sec : eval ] gives a detailed explanation for evaluation of the proposed method and the results in detail .
section [ sec : con ] concludes the work reported while discussing the future work planned .
gang violence is a well studied social science topic dating back to 1927 @xcite .
however , the notions of `` cyber- '' or `` internet banging '' , which is defined as _
`` the phenomenon of gang affiliates using social media sites to trade insults or make violent threats that lead to homicide or victimization '' _
@xcite , was only recently introduced @xcite .
_ introduced the concept of `` internet banging '' and studied how social media is now being used as a tool for gang self - promotion and as a way for gang members to gain and maintain street credibility @xcite .
they also discussed the relationship between gang - related crime and hip - hop culture , giving examples on how hip - hop music shared on social media websites targeted at harassing rival gang members often ended up in real - world collisions among those gangs .
et al . _ and patton _ et al .
_ have also reported that street gangs perform internet banging with social media posts of videos depicting their illegal behaviors , threats to rival gangs , and firearms @xcite . the ability to take action on these discoveries
is limited by the tools available to discover gang members on social media and to analyze the content they post @xcite .
recent attempts to improve our abilities include a proposed architecture for a surveillance system that can learn the structure , function , and operation of gangs through what they post on social media @xcite .
however , the architecture requires a set of gang member profiles for input , thus assuming that they have already been discovered .
et al . _
@xcite devised a method to automatically collect tweets from a group of gang members operating in detroit , mi . however , their approach required the profile names of the gang members to be known beforehand , and data collection was localized to a single city in the country .
this work builds upon existing methods to automatically discover gang member profiles on twitter .
this type of user profile classification problem has been explored in a diverse set of applications such as political affiliation @xcite , ethnicity @xcite , gender @xcite , predicting brand loyalty @xcite , and user occupations @xcite .
however , these approaches may utilize an abundance of positive examples in their training data , and only rely on a single feature type ( typically , tweet text ) .
whereas most profile classifiers focus on a single _ type _ of feature ( e.g. profile text ) , we consider the use of a variety of feature types , including emoji , youtube links , and photo features .
this section discusses the methodology we followed to study and classify the twitter profiles of gang members automatically .
it includes a semi - automatic data collection process to discover a large set of verifiable gang member profiles , an evaluation of the tweets of gang and non - gang member posts to identify promising features , and the deployment of multiple supervised learning algorithms to perform the classification .
.number of gang member profiles captured . [ cols="<,^",options="header " , ] for each 10-fold cross validation experiment , we report three evaluation metrics for the ` gang ' and ` non - gang ' classes , namely , the precision = @xmath2 , recall = @xmath3 , and @xmath1-score = @xmath4 , where @xmath5 is the number of true positives , @xmath6 is the number of false positives , @xmath7 is the number of true negatives , and @xmath8 is the number of false negatives .
we report these metrics for the positive ` gang ' and negative ` non - gang ' classes separately because of class imbalance in our dataset .
table [ results : usertweetsuni ] presents the average precision , recall , and @xmath1-score over the 10 folds for the single - feature and combined feature classifiers .
the table includes , in braces ( ` \ { } ' ) , the number of gang and non - gang profiles that contain a particular feature type , and hence the number of profiles used for the 10-fold cross validation .
it is reasonable to expect that _ any _ twitter profile is not that of a gang member , predicting a twitter user as a non - gang member is much easier than predicting a twitter user as a gang member .
moreover false positive classifications of the ` gang ' class may be detrimental to law enforcement investigations , which may go awry as they surveil an innocent person based on the classifier s suggestion .
we thus believe that a small false positive rate of the ` gang ' class to be an especially important evaluation metric .
we say that a classifier is ` ideal ' if it demonstrates high precision , recall , and @xmath1-score for the ` gang ' class while performing well on the ` non - gang ' class as well .
the best performing classifier that considers single features is a random forest model over tweet features ( t ) , with a reasonable @xmath1-score of 0.7229 for the ` gang ' class
. it also features the highest @xmath1-score for the ` non - gang ' class ( 0.9671 ) .
its strong performance is intuitive given the striking differences in language as shown in figure [ fig : image2 ] and discussed in section [ sec : data_analysis_tweet_text ] .
we also noted that music features offer promising results , with an @xmath1-score of 0.6505 with a naive bayes classifier , as well as emoji features with an @xmath1-score of 0.6067 also achieved by a naive bayes classifier .
however , the use of profile data and image tags by themselves yield relatively poor @xmath1-scores no matter which classifier considered
. there may be two reasons for this despite the differences we observed in section [ sec : data_analysis ] .
first , these two feature types did not generate a large number of specific features for learning .
for example , descriptions are limited to just 160 characters per profile , leading to a limited number of unigrams ( in our dataset , 10 on average ) that can be used to train the classifiers .
second , the profile images were tagged by a third party web service which is not specifically designed to identify gang hand signs , drugs and guns , which are often shared by gang members . this led to a small set of image tags in their profiles that were fairly generic , i.e.
, the image tags in figure [ fig : imagetags ] such as ` people ' , ` man ' , and ` adult ' . combining these diverse sets of features into a single classifier yields even better results . our results for _ model(1 ) _ show that the random forest achieves the highest @xmath1-scores for both ` gang ' ( 0.7364 ) and ` non - gang ' ( 0.9690 ) classes _ and _ yields the best precision of 0.8792 , which corresponds to a low false positive rate when labeling a profile as a gang member . despite the fact that it has lower positive recall compared to the second best performing classifier ( a random forest trained over only tweet text features ( t ) ) , for this problem setting , we should be willing to increase the chance that a gang member will go unclassified if it means reducing the chance of applying a ` gang ' label to a non - gang member . when we tested _
model(2 ) _ , a random forrest classifier achieved an @xmath1-score of 0.7755 ( improvement of 7.28% with respect to the best performing single feature type classifier ( t ) ) for ` gang ' class with a precision of 0.8961 ( improvement of 6.26% with respect to ( t ) ) and a recall of 0.6994 ( improvement of 9.26% with respect to ( t ) ) . _
model(2 ) _ thus outperforms _
model(1 ) _ , and we expect its performance to improve with the availability of more training data with all feature types .
we also tested the trained classifiers using a set of twitter profiles from a separate data collection process that may emulate the classifier s operation in a real - time setting . for this experiment
, we captured real - time tweets from los angeles , ca and from ten south side , chicago neighborhoods that are known for gang - related activities @xcite using the twitter streaming api .
we consider these areas with known gang presence on social media to ensure that some positive profiles would appear in our test set .
we ultimately collected 24,162 twitter profiles : 15,662 from los angeles , and 8,500 from chicago .
we populated data for each profile by using the 3,200 most recent tweets ( the maximum that can be collected from twitter s api ) for each profile . since the 24,162 profiles are far too many to label manually , we qualitatively study those profiles the classifier placed into the ` gang ' class .
we used the training dataset to train our best performing random forest classifier ( which use all feature types ) and tested it on the test dataset .
we then analyzed the twitter profiles that our classifier labeled as belonging to the ` gang ' class .
each of those profiles had several features which overlap with gang members such as displaying hand signs and weapons in their profile images or in videos posted by them , gang names or gang - related hashtags in their profile descriptions , frequent use of curse words , and the use of terms such as _ my homie " _ to refer to self - identified gang members .
representative tweets extracted from those profiles are depicted in figure [ testset_examples ] .
the most frequent words found in tweets from those profiles were _ shit , nigga , got , bitch , go , fuck etc . _ and their user profiles had terms such as _ free , artist , shit , fuck , freedagang , and ripthefallen_.
they had frequently used emojis such as _ face with tears of joy , hundred points symbol , fire , skull , money bag , and pistol_. for some profiles , it was less obvious that the classifier correctly identified a gang member .
such profiles used the same emojis and curse words commonly found in gang members profiles , but their profile picture and tweet content was not indicative of a gang affiliation . in conclusion , we find that in a real - time - like setting , the classifier to be able to extract profiles with features that strongly suggest gang affiliation .
of course , these profiles demand further investigation and extensive evidence from other sources in order to draw a concrete conclusion , especially in the context of a law enforcement investigation .
we refrain from reporting any profile names or specific details about the profiles labeled as a ` gang ' member to comply with the applicable irb governing this human subject research .
this paper presented an approach to address the problem of automatically identifying gang member profiles on twitter . despite the challenges in developing such automated systems , mainly due to difficulties in finding online gang member profiles for developing training datasets , we proposed an approach that uses features extracted from textual descriptions , emojis , images and videos shared on twitter ( textual features extracted from images , and videos ) . exploratory analysis of these types of features revealed interesting , and sometimes striking differences in the ways gang and non - gang members use twitter .
classifiers trained over features that highlight these differences , were evaluated under 10-fold cross validation .
our best classifier achieved a promising @xmath1-score of 0.7755 over the ` gang ' profiles when all types of features were considered .
future work will strengthen our training dataset by including more gang member twitter profiles by searching for more location - independent keywords .
we also plan to develop our own image classification system specifically designed to classify images found on gang member profiles .
we would also like to experiment with building dictionaries that contain gang names to understand whether _
`` having a gang name in the profile description '' _ as a feature can improve our results .
finally , we would also like to study how can we further improve our classifier models using word embeddings @xcite and social networks of known gang members .
we are thankful to uday kiran yeda for helping us with data collection .
we acknowledge partial support from the national science foundation ( nsf ) award : cns-1513721 : `` context - aware harassment detection on social media '' , national institutes of health ( nih ) award : mh105384 - 01a1 : `` modeling social behavior for healthcare utilization in depression '' and grant no . 2014-ps - psn-00006 awarded by the bureau of justice assistance .
the bureau of justice assistance is a component of the u.s .
department of justice s office of justice programs , which also includes the bureau of justice statistics , the national institute of justice , the office of juvenile justice and delinquency prevention , the office for victims of crime , and the smart office .
points of view or opinions in this document are those of the authors and do not necessarily represent the official position or policies of the u.s .
department of justice , nsf or nih . | most street gang members use twitter to intimidate others , to present outrageous images and statements to the world , and to share recent illegal activities .
their tweets may thus be useful to law enforcement agencies to discover clues about recent crimes or to anticipate ones that may occur .
finding these posts , however , requires a method to discover gang member twitter profiles .
this is a challenging task since gang members represent a very small population of the 320 million twitter users
. this paper studies the problem of automatically finding gang members on twitter .
it outlines a process to curate one of the largest sets of verifiable gang member profiles that have ever been studied .
a review of these profiles establishes differences in the language , images , youtube links , and emojis gang members use compared to the rest of the twitter population .
features from this review are used to train a series of supervised classifiers .
our classifier achieves a promising @xmath0 score with a low false positive rate .
street gangs , twitter profile identification , gang activity understanding , social media analysis |
the concept of spin noise was first put forward by bloch @xcite as the incomplete stochastic cancellation of up- and down - spins in the case of a nuclear spin system .
this nuclear spin noise was later experimentally verified by sleator _
analogously to nuclear spins , the total spin of a classical gas of unpolarized atoms also fluctuates around a zero mean polarization .
this atomic spin noise can be mapped onto the intensity of a probe laser due to dichroic bleaching of the relevant optical transitions @xcite .
ultimatively , aleksandrov and zapasskii demonstrated that the spin noise of an alkali gas can also be recorded in a perturbation - free manner by off - resonant faraday rotation @xcite .
for sufficient detuning from the probed optical resonance , the off - resonant probing @xcite even establishes a quantum non - demolition measurement of the atomic spin @xcite and represents the building block of several light - matter and matter - matter entanglement experiments @xcite .
the relative noise signal is large in small spin ensembles @xcite which makes spin noise spectroscopy especially interesting for studying spin dynamics in semiconductor nanostructures @xcite . in semiconductors ,
optical excitations can fundamentally change the observed electron spin dynamics such that only spin noise spectroscopy employing below band gap light enables access to the intrinsic spin dynamics in these systems @xcite .
correspondingly , experimental as well as theoretical studies of spin noise spectroscopy in the case of atom as well as semiconductor physics usually focus onto the off - resonant probing regime and the weak scattering limit @xcite . however , in the case of a strong inhomogeneously broadened optical transition , as present in an ensemble of self - ensembled semiconductor quantum dots , donor bound carriers , or a hot gas of atoms , a significant fraction of the contributing spins is inevitably resonantly probed by spin noise spectroscopy @xcite . in the following , we present an experimental and theoretical assessment of spin noise under resonant as well as off - resonant probing conditions to reveal effects that accompany spin noise spectroscopy under resonant excitation .
we chose a classical rubidium gas as an archetypal sample system and the transmitted laser light acts simultaneously as both , probing and pumping beam , respectively .
the experiments reveal a pronounced non - linear behavior at the optical resonance @xcite .
the observations are well reproduced by extended bloch equations where the coherent coupling of the involved atomic states is crucial to explain the experimental results .
experiment and calculations reveal that the observed spin noise under resonant probing differs qualitatively depending on whether the probed optical transition is dominantly inhomogeneously or homogeneously broadened . in the case of an inhomogeneously broadening ,
we furthermore encounter significant spin noise contributions from excited states .
the complex experimental results can only partly be reproduced by a single - particle bloch analysis which indicates the presence of a many - body effect such as nonlinear self - rotation and polarization selective absorption of the probe light . ) for the case of @xmath0rb .
the bold arrow indicates the transition that corresponds to zero detuning in the following .
neighboring ground and excited states are coherently coupled .
( b ) experimental spin noise set - up .
a liquid crystal retarder ( lcr ) switches the polarization analysis between sensitivity on linear or circular birefringence , respectively ( details are found in the text ) . ]
an external magnetic field applied in transverse , i.e. , voigt , geometry leads to a splitting of the linear polarized @xmath1 and @xmath2 atomic transitions ( _ at _ ) in longitudinal excitation direction .
the classical , dichroic voigt effect could be observed if both transitions are addressed by linear polarized light , oriented @xmath3 with respect to the magnetic field axis . on the other hand ,
a pure faraday rotation @xmath4 of linear polarized light occurs even at zero magnetic field if the refractive indices @xmath5 for the @xmath6 and @xmath7 polarization components differ due to dispersive effects @xcite .
spin noise spectroscopy measures the stochastic spin polarization of the investigated system projected along the direction of light propagation via the faraday effect .
the spin noise arises due to the finite spin dephasing time @xmath8 and stochastic reorientation of the ensemble spin . a magnetic field transverse to the direction of light propagation
modulates the spin noise with the larmor frequency @xmath9 ( @xmath10 : magnetic field , @xmath11 : land factor depending on the total angular momentum quantum number @xmath12 , @xmath13 : bohr magneton ) and , correspondingly , facilitates spin noise measurements in the inevitable presence of white photon shot noise and common @xmath14 noise .
the power spectrum of the measured faraday rotation noise is in the case of mono - exponential spin dephasing and a single , purely homogeneously broadened optical transition described by a lorentz function centered at the corresponding larmor frequency and a full width at half maximum ( fwhm ) given by @xmath15 @xcite .
the detected faraday rotation angle is proportional to ( _ i _ ) the degree of spin polarization and ( _ ii _ ) the circular birefringence given by the real part of the refractive index of the specimen @xcite .
( _ i _ ) the time averaged spin polarization vanishes at thermal equilibrium but has a finite mean deviation .
the mean deviation of the stochastic spin polarization follows a gaussian probability distribution .
( _ ii _ ) the investigated optical transition in this work is the rubidium d2 @xmath16 transition which is exemplarily depicted in fig .
[ aufbau ] ( a ) for @xmath0rb including the hyperfine splitting . in off - resonant spin noise experiments
, it usually suffices to model the probed optical transition by a single lorentz oscillator .
however , the complex details of the optical transition are needed to understand spin noise spectroscopy under resonant or nearly resonant probing conditions .
the faraday rotation of the linear light polarization results from circular birefringence , i.e. , different refractive indices for the two circular polarization components which compile to linear polarized light . the corresponding dipole selection rules for circular polarized light demand @xmath17 and @xmath18 .
figure [ aufbau ] ( a ) illustrates that resonant probing of the d2 transition evokes coherences of the @xmath19 ground state as well as the excited @xmath20 states .
we show in the following that the detailed structure of the d2 transition becomes evident as pronounced non - linearities in the spin noise signal under resonant excitation and a non - symmetric response for positive and negative detuning , respectively .
we first present the experimental results ( sec .
[ experiment ] ) and subsequently reproduce the spin noise spectra by a theoretical model that is based on the calculation of the refractive index of the atomic system by extended bloch equations ( sec . [ theory ] ) .
[ [ experimental - technique . ] ] experimental technique .
+ + + + + + + + + + + + + + + + + + + + + + + and @xmath21 ( terrestrial magnetic field ) , respectively . ]
figure [ aufbau ] ( b ) depicts the experimental setup .
linear polarized laser light is transmitted through a rubidium vapor cell and acquires a stochastic faraday rotation .
the time - dependent faraday angle is measured by a polarization bridge consisting of a polarizing beam splitter ( pbs ) and a balanced photoreceiver .
the time - domain faraday rotation data is analyzed via fast fourier transformation ( fft ) spectral analysis .
the light source is an external cavity diode laser ( ecdl ) with a grating for wavelength tuning in littrow configuration .
the laser emission is linear polarized and has a spectral line width of less than 20 mhz .
an anamorphotic prism pair yields a round beam shape and an optical faraday isolator prevents unwanted back - reflections into the laser .
the laser beam is focused by a 500 mm focal length lens to a beam diameter of 240 @xmath22 at the center of the probed rubidium vapor cell .
the rayleigh range thereby corresponds to the length of the cell of @xmath23 .
the laser is swept with a tuning range of -7 to + 13 ghz relative to the d2 @xmath24 transition of @xmath0rb at around 780 nm while the absorption of the atomic system is measured simultaneously .
the spectral position of the laser is monitored with a fabry - prot interferometer .
a small magnetic field is applied transverse to the direction of light propagation by means of a helmholtz coil with the light polarization exactly perpendicular to the magnetic field axis . in this work ,
two different rubidium vapor cells are investigated : _ cell a _ with enhanced @xmath0rb concentration and _ cell b _ with natural abundance .
cell a also contains helium buffer gas at a pressure of 1 mbar .
natural rubidium comprises the different isotopes @xmath25rb ( 72.15% ) and @xmath0rb ( 27.85% ) .
the cells are heated to around 350 k. the measured quantity in noise spectroscopy is the spectral noise power density , i.e. , the amplitude variance per frequency .
subtraction of two acquired noise power spectra from each other remove the spectral noise background which itself results from strong white photon shot noise and residual electrical noise .
figure [ example ] depicts an typical spin noise spectrum that is acquired with cell a for the case of strong optical detuning .
the spin noise peak at 12.8 mhz corresponds to the larmor frequency for an applied magnetic field of 1.8 mt ( @xmath26 ) .
the negative peak close to zero frequency results from the subtraction of a noise power spectrum with no magnetic field applied , i.e. , a spectrum where the spin noise signal is centered around zero frequency .
the spin noise signal centered at zero frequency shows twice the amplitude compared to the modulated spin noise signal since the spin noise power which would theoretically appear in the second quadrant is folded into the detection window , i.e. , for both cases the integrated spin noise power is the same . in the case of
strongly broadened spin noise spectra ( cell b ) , the available magnetic field range is not affected by overlapping spectra . instead , reference spectra are acquired by switching the polarization state of the probe light via a liquid crystal retarder ( lcr ) .
this method is described in detail in ref .
the extracted fwhm of the spin noise peak in fig .
[ example ] amounts to @xmath27 khz , the corresponding spin dephasing time of @xmath28 ns results from the finite interaction time between the rubidium spins and the probe laser due to the spatial motion of the atoms while the intrinsic spin dephasing time of rubidium is of the order of 100 ms @xcite .
please note that the helium buffer gas of 1 mbar ( cell a ) reduces the efficiency of this time - of - flight broadening only slightly and in cell b a corresponding fwhm of around @xmath29 khz is measured .
nevertheless , saturation broadening in cell a is very efficient due to the helium buffer gas @xcite and the dominating mechanism for broadening of the d2 transitions changes from inhomogeneous doppler broadening at low probe laser intensities to the homogeneous saturation broadening at high probe laser intensities .
[ [ spin - noise - in - cell - a . ] ] spin noise in cell a. + + + + + + + + + + + + + + + + + + + + + next , we study the spin noise in cell a in dependence on the laser detuning and the probe laser intensity
. figures [ fig : rb87 ] ( a ) and ( b ) depict the detected spin noise power density in color coding as a function of the laser detuning ( abscissa ) and the noise frequency ( ordinate ) at probe laser intensities of @xmath30 and @xmath31 , respectively . the black line depicts the corresponding absorption spectra .
the intensity plots show the difference noise spectra between applied external magnetic fields of @xmath32 and @xmath33 , respectively , i.e. , different larmor frequencies centered at @xmath32 and @xmath34 .
the overall absorption spectrum at low probe intensities of the @xmath35 transitions show a fwhm on the order of 2 ghz where the doppler broadening contributes by an inhomogeneous line width of 560 mhz ( calculated value ) . the spin noise power density resulting from the @xmath35 transitions around 0 ghz detuning shows a pronounced asymmetric dependency for positive and negative detuning
. the asymmetry of the spin noise power spectrum decreases with increasing laser intensity [ compare fig . [
fig : rb87 ] ( a ) and ( b ) ] since the @xmath35 transitions are simultaneously excited at high intensities due to the stronger homogeneous saturation broadening .
we will show in section [ theory ] that n - type coherences are responsible for the asymmetric behavior and the spectral shift between absorption maximum and spin noise minimum .
[ [ spin - noise - in - cell - b . ] ] spin noise in cell b. + + + + + + + + + + + + + + + + + + + + + for different laser frequencies . at low detuning ,
spin noise of the excited @xmath36 states ( @xmath37 ) becomes significant .
( c ) spin noise spectra calculated by the extended bloch model corresponding to the experimental data shown in ( b ) . ] figure [ fig : rbnat ] ( a ) and ( b ) depict absorption and spin noise spectra for cell b acquired with a probe laser intensity of @xmath38 .
the absorption spectra are dominantly broadened by saturation broadening .
the spin noise spectra exhibit an additional noise signal due to the different land factor of the ground state of @xmath25rb ( @xmath39 ) compared to @xmath0rb .
most strikingly , the excited @xmath36-@xmath0rb states also contribute to the observed spin noise at high probe intensities and low detuning from the optical resonances .
this is indicated at the corresponding land - g - factor of @xmath37 in fig .
[ fig : rbnat ] ( b ) .
the excited state @xmath11-factors of @xmath25rb are @xmath40 , @xmath41 , @xmath42 and @xmath43 for @xmath44 and hence the excited state spin noise of @xmath25rb is spread over a huge frequency interval and thus does not interfere with the measured excited state noise of @xmath0rb .
[ [ bloch - equations . ] ] bloch equations .
+ + + + + + + + + + + + + + + + the detected spin noise results from a circular birefringence due to the fluctuating occupation of the degenerate @xmath45 spin states .
the equilibrium occupation probability of each @xmath19- and @xmath20-type hyperfine level @xmath12 is needed to determine the fluctuation strength and the transition probability of each transition involved and henceforth to calculate the circular birefringence .
all necessary quantities are obtained by solving the optical bloch equations .
the optical bloch equations represent the equation of motion for the density matrix @xmath46 of the investigated atomic system in the presence of a light field @xcite . in the following ,
the index @xmath47 ( @xmath48 ) denotes the ground ( excited ) states .
the ground states for the case of @xmath25rb ( @xmath0rb ) are @xmath49 @xmath50 ( @xmath51 ) ; the excited states are @xmath36 @xmath44 ( @xmath52 ) . in the rotating wave approximation ,
the bloch equations for the ground state occupancies @xmath53 read : @xmath54 here , @xmath55 is the spontaneous transition rate where @xmath56 denotes the fine structure oscillator strength , the term @xmath57 yields the branching ratio of the hyperfine transition with @xmath58 and @xmath59 is the resonance frequency of the corresponding optical transition .
the resonant rabi frequency for linear polarized light with intensity @xmath60 and light frequency @xmath61 reads @xmath62 here , @xmath63 is the electric field amplitude of the probe laser light and @xmath64 the transition dipole moment .
relaxation towards the thermal equilibrium by diffusion of atoms in and out of the excitation area is accounted by the diffusion rate @xmath65 , which is extracted from the off - resonant noise spectra and amounts to @xmath66 for cell a and @xmath67 for cell b. the occupation probability of the valence electron in the corresponding ground state scales with the degeneracy @xmath68 and @xmath69 is the two dimensional permutation tensor .
the excited state occupancies are given by @xmath70 .\ ] ] the coherences between ground and excited states are given by the off - diagonal elements of the density matrix and have to be included to correctly model the experimental findings , such as the asymmetric shape of the spin noise spectrum versus detuning : @xmath71 here the additional relaxation rate @xmath72 describes the quenching of the excited states after collisions with buffer gas atoms .
for 1 mbar helium gas pressure ( cell a ) , @xmath73 is extrapolated to @xmath74mhz from the data presented in ref @xcite .
the detuning is given by @xmath75 with @xmath76 for atoms with a velocity @xmath77 , the laser frequency @xmath78 , and resonance frequencies @xmath79 @xcite .
note that in eqs .
( [ eq : g - e - coh ] ) the ground - ground coherences ( @xmath80 ) and the excited state coherences ( @xmath81 ) are included , and hence the dynamics of these coherences must be carefully considered : @xmath82 \cdot \rho_{ii ' } \nonumber \\ & & + i \sum_{j } [ \omega_{ij } \cdot \rho_{i'j } - \omega_{i'j } \cdot \rho_{ji } ] , \end{aligned}\ ] ] @xmath83 \right ) -\gamma _ { b } + i\sum_{i } [ \delta_{j , i\pm { \rm 1,0 } } \cdot \delta_{i\pm { \rm 1,0\ ; } j } -\delta_{j',i\pm { \rm 1,0 } } \cdot \delta_{i\pm { \rm 1,0\ ; } j ' } ] \right\ } \cdot \rho_{jj ' } + i\sum _ { i}[\omega_{ij } \cdot \rho _ { ij ' } -\omega_{ij ' } \cdot \rho _ { ji } ] , \ ] ] where @xmath84 ensures the @xmath85 selection rules .
the steady state of this system of differential equations is solved numerically in terms of @xmath86 . the doppler broadening is included by convolution of the respective quantity of interest with the velocity distribution @xmath87 of atoms with the velocity @xmath77 and an average velocity @xmath88 for a given temperature @xmath89 .
note that every density matrix element @xmath90 is a function of @xmath77 .
[ [ absorption - spectrum . ] ] absorption spectrum .
+ + + + + + + + + + + + + + + + + + + + the absorption spectrum is calculated from the system s susceptibility @xmath91 including all @xmath1- and @xmath2-type transitions : @xmath92 with @xmath93 being the rubidium atom density . the doppler broadening is included by convolution of @xmath94 with the velocity distribution @xmath95 .
the absorption coefficient @xmath96 is calculated by @xmath97 .
[ [ spin - noise - spectrum . ] ] spin noise spectrum .
+ + + + + + + + + + + + + + + + + + + + calculating the spin noise spectrum is not as straight forward as the absorption spectrum since each hyperfine split optical transition @xmath98 needs to be weighted with the mean square spin imbalance @xmath99 , i.e. , the actual spin noise , and the spin noise of the distinct hyperfine levels @xmath12 appear at the corresponding larmor frequency @xmath100 in the noise spectrum . correspondingly , we model the complete spin noise spectrum as a sum of lorentz functions for all states .
the ground state spin relaxation rates increase at zero detuning due to optical pumping by : @xmath101 \right).\ ] ] the noise amplitude is given by : @xmath102 here @xmath103 is the amplification factor of the balanced receiver , @xmath104 the laser power , and @xmath105 the wave vector .
the faraday noise angle is calculated by : @xmath106 the sum of lorentz functions is finally convoluted with the velocity distribution to include doppler broadening and is valid for uncorrelated atoms . in eq .
@xmath107 is the dipole matrix element for each @xmath108 transition which is calculated separately for every hyperfine transition .
the corresponding expressions for the excited states are obtained by interchanging @xmath47 and @xmath48 in eqs .
( [ eq : noisewidth],[eq : noiseamplitude],[eq : noiseangle ] ) .
the excited state spin relaxation rates are calculated by @xmath109 .
further a factor of two is included in the calculation of the noise power of the excited states to account for stimulated emission .
[ [ results - and - comparison - with - experimental - data .
] ] results and comparison with experimental data .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + calculated absorption and spin noise spectra are depicted in figs .
[ fig : rb87 ] ( c , d ) and [ fig : rbnat ] ( a , c ) .
the calculated absorption spectra reproduce qualitatively , and to a large extent quantitatively , all different broadening mechanisms .
the saturation broadened absorption spectra show a non - negligible deviation in the width of the corresponding transitions between theory and experiment .
this discrepancy results from the probe laser beam waist that varies in the experiment over the cell length with a factor of @xmath110 .
however , this inaccuracy in the calculations is negligible since the spin noise at the focus point is weighted stronger than the spin noise at the cell endings .
indeed , the calculated and measured spin noise spectra for cell a ( pure @xmath0rb , fig .
[ fig : rb87 ] ) are in excellent agreement .
please note that there is no free parameter in the calculations to adjust the absolute noise power values .
especially , the asymmetric behavior around the @xmath35 transition is remarkably well reproduced for the two different noise spectra .
this asymmetry decreases in the calculations as well as in the experiment for higher probe light intensities due to additional homogeneous broadening .
the ratio between laser intensity and spin noise signal stays constant due to the larger homogeneous broadening and increasing spin relaxation rates for stronger laser light fields . in the case of cell b ( rb with natural abundance ) ,
the spin noise arising from the excited @xmath111-@xmath0rb states is qualitatively also well reproduced [ fig .
[ fig : rbnat ] ( b , c ) ] .
the excited state noise corresponds to noise frequencies around @xmath112 mhz denoted by @xmath37 .
we note , however , one significant discrepancy between the experimental and theoretical results in what concerns the strength of the excited state spin noise . the enhanced excited state spin noise measured at detunings close to the optical resonance and high probe laser intensities
is not recovered by the single - particle bloch analysis .
this discrepancy indicates a signature of a collective effect such as nonlinear self - rotation @xcite and polarization selective absorption of the probe light while traversing the rubidium gas cell .
these many - body effects are beyond the scope of the single particle model discussed in this paper and only appear at significant occupation of the excited state .
we have performed spin noise spectroscopy on a classical rubidium gas under resonant and non - resonant probing conditions .
the experimental results are modeled by extended bloch equations , which inherently include the populations and coherences between all participating states .
our careful analysis shows that the coherences between ground and excited states , and also ground - ground and excited - excited coherences , play a crucial role in spin noise spectroscopy at resonant and quasi - resonant probing conditions .
this is contrary to the case of off - resonant probing where coherences play no significant role .
we identify a characteristic asymmetric spin noise signal with respect to detuning which arises due the coherent coupling of the atomic levels under nearly resonant probing conditions .
the theoretical results are in very good qualitative , and to a large extent quantitative , agreement with our experimental findings .
however , a significant deviation is observed for the specific case of excited state spin noise for nearly resonant probing conditions .
interestingly , this discrepancy may indicate the appearance of collective effects , which are not covered by single - particle extended bloch equations .
we expect that these and similar effects also plays an important role in spin noise experiments with self - assembled semiconductor quantum dots or donor atoms .
we gratefully acknowledge financial support by the bmbf _ quahl - rep _ , the deutsche forschungsgemeinschaft in the framework of the priority program spp1285 _ semiconductor spintronics _ , and by the german excellence initiative via quest - _
center for quantum engineering and space - time research_. 37ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrev.70.460 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.55.1742 [ * * , ( ) ] @noop * * , ( ) http://ol.osa.org/abstract.cfm?uri=ol-18-21-1816 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.84.5292 [ * * , ( ) ] link:\doibase 10.1103/physreva.74.043819 [ * * , ( ) ] http://link.aip.org/link/?spj/54/64/1 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.18.577 [ * * , ( ) ] link:\doibase 10.1103/physreva.60.2346 [ * * , ( ) ] http://dx.doi.org/10.1038/35096524 [ * * , ( ) ] http://dx.doi.org/10.1038/nature03064 [ * * , ( ) ] http://dx.doi.org/10.1038/nature05136 [ * * , ( ) ] \doibase doi:10.1038/nature02804 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.216603 [ * * , ( ) ] \doibase doi : 10.1016/j.physe.2010.08.010 [ * * , ( ) ] http://link.aps.org/abstract/prl/v101/e206601 [ * * , ( ) ] link:\doibase 10.1103/physreva.75.042502 [ * * , ( ) ] link:\doibase 10.1103/physreva.75.063827 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.036601 [ * * , ( ) ] link:\doibase 10.1103/physreva.83.032512 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.74.1153 [ * * , ( ) ] link:\doibase 10.1364/josab.22.000007 [ * * , ( ) ] link:\doibase 10.1103/physreva.69.043809 [ * * , ( ) ] link:\doibase 10.1364/josa.69.000781 [ * * , ( ) ] link:\doibase 10.1364/josab.11.000409 [ * * , ( ) ] link:\doibase 10.1063/1.2794059 [ * * , ( ) ] http://link.aip.org/link/?apl/94/112105/1 [ * * , ( ) ] link:\doibase 10.1103/physreva.14.1711 [ * * , ( ) ] @noop _ _ , ed .
( , ) @noop * * , ( ) link:\doibase 10.1103/physreva.55.3710 [ * * , ( ) ] link:\doibase 10.1103/physreva.56.4569 [ * * , ( ) ] http://steck.us/alkalidata[__ ] ( , ) http://steck.us/alkalidata[__ ] ( , ) | high sensitivity faraday rotation spectroscopy is used to measure the fluctuating magnetization noise of non - interacting rubidium atoms under resonant and non - resonant optical probing conditions .
the spin noise frequency spectra in dependence on the probe light detuning with respect to the d2-transition reveals clear signatures of a coherent coupling of the participating electronic levels .
the results are explained by extended bloch equations including homogeneous and inhomogeneous broadening mechanisms .
our measurements further indicate that spin noise originating from excited states are governed at high intensities by collective effects . |
x - ray emission associated with radio jets in extragalactic sources now comes in a large variety of diverse characteristics .
when this sort of emission was first isolated ( e.g. the radio galaxy m87 with the einstein observatory , schreier , gorenstein , & feigelson 1982 ) and there were only a few examples , it was tempting to work on the assumption that the emission process was definable and would apply to all examples .
this notion persisted into the rosat era until a convincing case was made that the terminal hotspots of cygnus a represented synchrotron self - compton ( ssc ) emission ( harris , carilli , & perley , 1994 ) . with the advent of the chandra observatory ,
the number of sources has almost tripled ( from 7 to at least 19 ) yet there remain substantial problems in determining the emission process responsible for the x - rays in some sources .
although broken power laws connecting the radio , optical , and x - ray data are still viable spectral models for a few sources , a number of other sources are observed to have such a low flux density in the optical that the indicated cutoff in the spectrum would preclude a simple connection to the x - ray data . the introduction of the beaming model ( tavecchio et al .
2000 , celotti et al .
2001 ) in the particular case of pks0637 was presented as an escape from this dilemma . in this model , the enhancement of the x - ray emission relative to the synchrotron ( radio / optical )
is explained by hypothesizing a bulk velocity of the jet fluid which is relativistic even on large ( kpc ) scales .
if this were the case , in the frame of the jet the effective photon energy density of the cosmic microwave background ( cmb ) would be augmented by the square of the jet s lorentz factor , @xmath0 , and it was demonstrated that quite reasonable combinations of @xmath0 and the beaming factor , @xmath1 could be invoked to explain the observed intensities . in this paper
we review the common emission processes ( section [ sec : emission ] ) ; present a current list of jet sources and suggest a classification scheme ( section [ sec : classify ] ) ; describe a formulation of the beaming model which relies on an evaluation of the magnetic field strength from the equipartition field and includes the anisotropic nature of the ic scattering ( section [ sec : beam ] ) ; and examine the conflicting evidence and ramifications for the synchrotron and inverse - compton ( ic ) models ( section [ sec : conflict ] ) .
the appendix contains the details of our beaming formulation . for numerical results we use cgs units
unless stated otherwise and assume h@xmath2=50 km s@xmath3 mpc@xmath3 ; and q@xmath2=0 .
we follow the convention that the spectral index of a power law is defined by flux density , s@xmath4 .
for many sources which display convincing evidence that the x - ray and radio emission originate in the same volume , ( e.g. hotspot b of 3c390.3 , harris leighly & leahy 1998 ) it has been argued that thermal bremsstrahlung does not provide a satisfactory model for the x - ray emission because the required amount of hot gas is large ( @xmath5 ) , over - pressured , far from the parent galaxy , and the predicted faraday rotation and depolarization are not observed .
recent chandra results ( e.g. 3c 273 , marshall et al .
2001a ) confirm that the x - ray emission from jet features has no line emission and is best characterized by a power law .
synchrotron emission has been considered the ` process of choice ' for x - rays from knots in radio jets mainly because the optical polarization observed in sources such as m87 is convincing evidence that the optical emission as well as the radio emission comes from the synchrotron process .
demonstrations that the x - ray intensity was consistent either with a single power law extrapolation from radio and optical bands ( e.g. hotspot b of 3c 390.3 , harris , leighly , & leahy 1998 ) or with a broken power law ( e.g. knot a in m87 , biretta , stern , & harris 1991 ) were taken as circumstantial evidence that the x - rays were also generated by synchrotron emission . required for this model is the presence of electrons with lorentz factor @xmath6 ( cf .
values of 10@xmath7 for optical emission ) . in the typical equipartition fields of b @xmath8 g
, the radiation half lives @xmath9 of the x - ray emitting electrons would be of order 10 years ( however , see aharonian ( 2001 ) for problems associated with very fast cooling times ) .
there are , however , a number of sources for which the optical flux densities or limits preclude a simple construction of a broken power law . in general , assuming the usual shock acceleration processes and dominance of losses to the relativistic electrons that go with the energy squared , we expect that both the electron distribution and the resulting synchrotron emission spectrum will be concave downward when displayed on the usual log s@xmath10 vs. log @xmath11 plot . for this reason , x - ray intensities that lie well above the extrapolation of the radio / optical synchrotron spectrum
are taken to be strong evidence against the simple synchrotron model .
the emission of most of these sources however can still be explained with inhomogeneous synchrotron models .
spatially separated emission components which can not be resolved with the current x - ray detectors would be the consequence of diffusive shock acceleration with a time dependent high energy cut - off of accelerated particles .
while the low energy radio and optical emission is dominated by radiatively aged particle populations further downstream , the x - ray emission is only emitted by recently accelerated particles . in this scenario
the projected extension of the x - ray emission region is either given by the projected size of the accelerating shock , or by the distance the downstream plasma travels before the highest energy electrons cool away . for 3c 120
such an interpretation encounters the difficulty that the optical to x - ray energy spectrum is as hard as @xmath120.35 which indicates a spectrum of accelerated particles with a spectral index , @xmath13 ( even in the optimistic case that the x - ray electrons did not cool substantially ) ; harder than the canonical values of 2 expected for acceleration at strong non - relativistic shocks .
another class of models involves high energy protons .
these models and their difficulties are described in mannheim , krulls , & biermann ( 1991 ) and aharonian ( 2001 ) .
inverse compton emission has a distinct advantage over x - ray synchrotron emission in that extremely high energies for the emitting electrons are not required .
what is required are both enough electrons and enough energy density in photons of the proper energy to produce the desired scattered photons : @xmath14 . unlike synchrotron emission
for which the radio data often provide no clear indication about the higher energy parts of the electron spectrum , accurate predictions of source intensity can be calculated for ssc emission . given a reasonable estimate for the emitting volume , it is possible to calculate the photon energy density from the synchrotron spectrum and
then determine what b field is required to have the right number of electrons to give the observed ( radio ) synchrotron and the observed ( x - ray ) ic emissions .
although ssc emission is mandatory for all synchrotron sources , it is usually the case that the photon energy density ( @xmath15 ) is significantly smaller ( i.e. by factors of 100 or more ) than the energy density of the magnetic field ( @xmath16 ) and thus the major energy loss for all electrons is via synchrotron emission ( and the ssc component is too weak to observe or distinguish from other emissions ) .
there are however , four frii radio galaxies with convincing ssc x - ray emission from their ( terminal ) hotspots : cyg a ( harris et al . 1994 ; wilson , , young , and shopbell 2001a ) ; 3c 295 ( harris et al .
2000 ) ; 3c 123 ( hardcastle , birkinshaw , & worrall 2001a ) , and 3c 263 ( hardcastle et al . in preparation ) . for all of these sources , predicted ssc emission was calculated from the radio data , proposals were written , and the hotspots were detected as predicted . in all cases , the average magnetic field strengths derived from the ssc equations are consistent with equipartition fields for the case of little or no contribution to the particle energy density from relativistic protons .
the hotspots of these sources are , so far , the only resolved structures for which convincing ssc models have been published . except for high redshift sources , ic emission from the relativistic electrons scattering off
the cmb is incapable of producing the observed intensity of x - rays unless the beaming model is invoked with relativistic values of the bulk jet velocity ( see section [ sec : beam ] ) .
the beaming model is attractive because most or all of the x - ray emitting jets are one sided and the model permits the relative increase of the effective photon energy density to that in the magnetic field , thereby boosting the emission in the ic channel relative to that in the synchrotron channel .
external photons have been employed successfully in explaining x - ray emission from some radio lobes ( brunetti et al .
. however , it is unlikely to apply to knots far removed from the hypothesized highly luminous quasar - like core and brunetti s model provides ic enhancement on the receding side of the source , not the approaching side as observed in our sample .
celotti et al .
( 2001 ) discuss the possibility that emission from a high velocity ` spine ' of a jet may serve as seed photons for ic emission from a lower velocity sheath .
generally , a wide latitude of inventiveness is allowed for ic emission since the dominant contribution to the photon energy density may be anisotropic and hence not directly observable .
the contents of this section are ` subject to change ' since much of our information is incomplete .
the reason for including this source list is to provide an overview .
some aspects of the classification are subjective and we anticipate changes as new data become available .
we maintain current information at http://hea - www.harvard.edu / xjet/. table [ tab : sources ] contains a list of bona fide x - ray jet sources known to us as of 2001 march . in the last column
is the classification which is described here .
this category is perhaps more secure than the others because the data confirm the predictions rather well .
all 4 ssc sources ( sec . [ sec : ssc ] ) are frii radio galaxies and their terminal hotspots rather than brightness enhancements in the jets are the ssc emitters . for all of these ,
the magnetic field strengths in the hotspots are in the range 100 to 400 @xmath17 g , values consistent with the equipartition field for the case of little or no contribution to the particle energy density from protons .
synchrotron luminosities and lifetimes are unaffected by the ic losses and the electrons responsible for the observed x - rays are those emitting synchrotron emission in the normal radio band .
alternative models for these hotspots are the generation of high energy electrons via the pic ( proton induced cascade , mannheim et al .
1991 ) or proton synchrotron ( aharonian 2001 ) processes , both of which would require much stronger magnetic fields ( @xmath18 g ) .
mindful of ongoing debate about the evidence for spectral cutoffs from optical data ( sec .
[ sec : sync ] ) , we classify m87 knots d , a , and b as synchrotron emitters ( biretta , stern , and harris , 1991 ; marshall et al . ,
additional sources in this category are 3c 390.3 hotspot b ( harris et al .
1998 ) , knots a1 and b1 in the 3c 273 jet ( marshall et al .
2001a ) , and 3c 66b ( hardcastle , birkinshaw , & worrall , 2001b ) although for some of these , there is the problem that a cooling break in the emission spectrum has not been isolated .
the extension of the synchrotron spectrum from the optical to the x - ray has only a small effect on the total energy and on the calculation of the equipartition magnetic field .
however , this model requires the extension of the electron spectrum from @xmath19 to @xmath20 .
the primary observable consequence is that the electrons responsible for the x - rays have lifetimes of order 10 to 100 years . if a somewhat ad hoc additional spectral component is allowed ( sec .
[ sec : distinct ] ; i.e. a high energy component of the electron energy distribution with a flatter spectrum than that observed in the radio / optical domain ) , then synchrotron emission models can be devised for sources such as pictor a ( wilson , young , & shopbell 2001b , section 4.2.2.2 ) , 3c 273 knot d , pks 0637 , and 3c 120 ( harris et al . 1999 ; where a very hard spectral index of accelerated particles of @xmath21=1.7 would be required ) .
the ssc sources ( sec . [
sec : ssc ] ) could also be accomodated by synchrotron models if additional spectral components are allowed . in sections [ sec : beam ] and
[ sec : conflict ] we discuss many of the details and ramifications of this model .
while there is little debate concerning the utility of the beaming model in its ability to explain a larger ratio of x - ray to synchrotron emission than other models , there is no completely convincing evidence that kpc - scale jets actually have bulk velocities close to c. beaming models have been presented for pks0637 ( celotti et al , 2001 ; tavecchio et al . 2000 ) and for 3c 273 ( sambruna et al .
for the knots in these jets , viewing angles are generally required to be less than 10@xmath22 , @xmath0 values range from 5 to 20 , and equipartition fields are significantly smaller than for unbeamed synchrotron emission .
a very significant difference between beaming and synchrotron models is that the beaming model posits that the electrons responsible for the observed x - rays are at the very bottom of the electron distribution with @xmath23 in the range 20 - 150 .
this implies that no variability is expected in the x - ray intensity .
because there is a significant difference between the electron energies producing the x - rays and those that produce radio synchrotron emission observable from the earth with sufficient resolution to measure accurate spectral indices , there is a large uncertainty in the expected spectral shape of the electron distribution at low energies , and consequently , in the predicted spectral shape and the intensity of the x - ray emission .
although most of the sources in table [ tab : sources ] display spatial coincidence between the radio , x - ray , and optical morphologies , there are a few sources where this is not the case .
cen a , 3c 66b , and pks 1127 are the three examples of this behavior in our collection .
we do not believe that this effect results from limited angular resolution since it occurs in both the nearest ( cen a ) and furthest ( pks1127 ) source . in the case of cen
a , some offsets between the peak brightness of radio and x - ray are of order 2@xmath24 ( 34pc ) while other features are coincident . for pks1127 ,
an offset of 0.8@xmath24 observed for knot b corresponds to 9 kpc . for these jets , it is difficult to measure radio and x - ray intensities for well defined volumes .
it is also the case that the pks1127 jet is very weak in the radio ( the source is a vla calibrator ) and the published radio data for cen a do not have the combination of high resolution and s / n which are necessary for reliable inter - band comparisons . for 3c 66b and pks1127 ,
the peak x - ray brightness occurs upstream of the peak radio brightness . to obtain significant x - ray emission with little or no radio emission would require extreme beaming parameters or a distinct low energy electron population ( sec . [
sec : distinct ] ) for the beaming model or a flat high energy component of the electron spectrum for synchrotron models .
since data have not yet been published , we have not classified 3c 15 , 3c 31 , pks0521 , and ngc4261 in table [ tab : sources ] .
the beaming model introduced by tavecchio et al . ( 2000 ) and celotti et al .
( 2001 ) is appealing because it offers a method of increasing the ic emission relative to the synchrotron emission and because it was already clear from einstein and rosat results on m87 , 3c 120 , and 3c 390.3 that all known x - ray emitting jet features were one sided .
most of the new examples from chandra detections maintain the one sided nature of x - ray emission , a natural consequence of the beaming model .
the following formulation is based on the assumption of equipartition in the synchrotron source in order to evaluate the magnetic field strength and accounts for the anisotropic nature of the ic emission in the jet frame .
the details are given in the appendix .
we define the constraints on the bulk relativistic flow velocity of the jet fluid , @xmath0 ; the viewing angle of the line of sight of the observer with respect to the jet velocity vector , @xmath25 ; and the relativistic doppler factor , @xmath1 .
once these constraints have been evaluated , we examine a number of source parameters : * the segments of the electron energy spectrum responsible for both the synchrotron and ic emissions to make sure we end up with a self - consistent source model . * the halflives of the various energy electrons . * the basic parameters of the synchrotron source in the jet frame . over the years since its introduction ( burbidge , 1956 ) ,
the concept of minimum energy ( or equipartition ) has generated considerable debate and some confusion .
our view is that although it was originally borne of twin desperations ( the staggering amount of non - thermal energy required for large synchrotron radio sources and the realization that it was almost impossible to disentangle the two primary parameters of synchrotron emission ) , a reasonable case can be advanced that most extragalactic synchrotron sources are not more than a factor of a few from having their average magnetic field energy density equal to the average energy density in relativistic particles : u(p ) @xmath26 u(b ) .
obviously we are unconcerned with the condition u(p )
@xmath27 u(b ) ; it is the converse that has generated some interest as a means of increasing the relative emission in the ic channel compared to that in the synchrotron channel . for radio structures
far from their parent galaxy it seems that the only mechanism which serves to maintain the integrity of a radio lobe is the magnetic field , acting as a boundary between the non - thermal plasma and the external thermal plasma .
direct evidence that some synchrotron emitting plasmas maintain their identity is provided both by the detection of cavities in cluster gas ( carilli , perley , & harris 1994 ) and by the observation that the b field direction is normally tangent to the edges of radio lobes ( and not tangled as expected if turbulent mixing were occurring ) .
if the particle energy density were much greater than u(b ) it is unlikely that the field would confine the particles and radio sources would be transient phenomena without well defined boundaries .
evidence for equipartition has come from the analyses of x - ray emission from terminal hotspots . in the study of cygnus a hotspots , harris et al .
( 1994 ) argued that if the observed x - ray emission were to be ssc emission , then the average magnetic field strengths ( b@xmath28 ) would be very close to the equipartition fields estimated from the conditions : filling factor = 1 and no contribution to the particle energy density from relativistic protons .
since we have a fairly accurate estimate of the synchrotron photon energy density , we know that if there were more electrons than those corresponding to b@xmath29 , we would over - produce the x - ray emission .
thus the derived value of b@xmath30 is a firm lower limit and the only method of driving the hotspots out of equipartition with u(p ) @xmath31 u(b ) would be to add non - radiating particles .
the three recent additions to the collection of resolved ssc emitters strengthen this evidence .
a number of additional arguments have been presented in favor of synchrotron sources being close to equipartition .
readhead ( 1994 ) used the observed distribution of brightness temperatures in support of equipartition but did not deal with the physical mechanism responsible .
bodo , ghisellini & trussoni ( 1992 ) refined earlier work by syrovatskii ( 1969 ) and singal ( 1986 ) on the effects of spiraling electrons and their associated diamagnetic moments and stated `` the counter field produced by a cloud of relativistic electrons lead to an equilibrium between the magnetic field in the cloud and the energy density of radiating electrons . ''
although it is possible to imagine synchrotron sources shortly after the introduction of a large amount of energy in particles , it is beyond the scope of this paper to consider the time scale required to return to equipartition .
since we are dealing with large ( @xmath32kpc ) scale structures , we believe that most of the emitting regions considered in this paper will be close to equipartition .
however , questions are often raised such as `` what is the extent of departure from equipartition for some given condition ? ''
we note from equation [ eq : resulta ] that a given change in @xmath0 ( which involves the same change in @xmath1 ) will require a change in b(1 ) by this factor squared , and since u(b ) goes as b@xmath33 , the change to u(b ) goes as the fourth power . as an example , consider the beaming required for 3c120 : @xmath0=39 ( table [ tab : results ] ) .
how far from equipartition would we have to go to get @xmath0=10 ?
the answer is of order 230 times less energy in the magnetic field than in particles .
although this calculation is approximate , the formulation presented in the appendices relies on the equipartition assumption only to evaluate the average field strength , and if some other estimate is available , it can be used in place of b@xmath34 in eq .
[ eq : result ] .
we first take the synchrotron spectrum from the observer s frame back to the jet frame .
since the beaming factor is unknown , this process is done ( in principle ) for trial values of @xmath1 . in the jet frame
, we apply the usual synchrotron formulae ( e.g. pacholczyk , 1970 ) to solve for the equipartition field strength ( eq .
[ eq : beq ] ) . actually , the value of the average magnetic field for equipartition is inversely proportional to @xmath1 , so in practice , we need only calculate @xmath35 for @xmath1 = 1 .
we then derive the expression for the ic emission ( eq . [ eq : power ] ) using the effective energy density of the cmb as seen by the jet fluid ( eq . [ eq : unu ] ) .
we evaluate the quantity r ( the ratio of ic to synchrotron luminosities ) in two ways .
first , we compute r from the integrated ic energy flux to the integrated synchrotron energy flux ( eq . [ eq : robs ] ) .
this is done by converting observed synchrotron frequencies to electron energies ( which involves the magnetic field strength ) and integrating the ic emission from the same segment of the electron spectrum , taking into account the anisotropic radiation pattern of the ic component .
second , we compute the expected value of r from the ratio of the energy densities of the cmb to that in the magnetic field ( both in the jet frame ) ( eq .
[ eq : rexp ] ) . equating these two expressions , we are able to solve for the beaming parameters which satisfy the equality ( eq .
[ eq : result ] ) . since the jet parameters enter the r equations in rather complex ways because of the anisotropic nature of the ic emission , a numerical method is used .
this is demonstrated in fig .
[ fig : method ] and we obtain a result consistent with that of celotti et al .
( 2001 ) for pks0637 .
we use the following notation : we prime all quantities in the jet frame and , in cases where ambiguities could arise , we characterize jet parameters by the subscript `` j '' and electron parameters by the subscript `` e '' .
* we assume that all the energy density of the cmb occurs at @xmath36 , the peak of the black body distribution .
* we assume that in the jet frame equipartition holds between the energy densities of the relativistic particles and the magnetic field .
this allows us to estimate the average field strength in the source .
moreover , we take the volume filling factor to be one and the contribution to the particle energy density by protons to be nil .
if relativistic protons contribute significantly to the particle energy density , b@xmath29 increases and beaming parameters become more extreme ( larger values of @xmath0 and smaller angles ) .
* the spectral index measured for the synchrotron spectrum at the lowest available frequencies continues unchanged to frequencies much lower that those accessible to earth based telescopes .
the gross attributes of this extrapolation must be consistent with radio emission arising from @xmath37 and x - ray emission ( 0.2 - 10 kev ) arising from @xmath38 .
this extrapolation of the electron spectrum provides a large uncertainty in the beaming model and has been tacitly ignored by most previous formulations . at our current level of understanding , we do not have any convincing evidence on the amplitude or the power law index of the electron distribution at low energies . in this section ,
we provide some of the key synchrotron parameters and beaming descriptors for a few knots in radio jets . for the synchrotron parameters , we use the standard expressions ( e.g. pacholczyk , 1970 ) with observables
transformed back to the jet frame .
for the beaming parameters , our solution to eq .
[ eq : result ] requires only 4 numbers : * b(1 ) , the equipartition field from the synchrotron model when @xmath1=1 ( no beaming ) * r(1 ) , the ratio of amplitudes of the power law spectra : @xmath39 where both amplitudes must refer to the same spectral index * z , the redshift * @xmath40 , the spectral index table [ tab : results ] gives the results and it can be seen that the beaming parameters range from quite modest values ( e.g. pks1127 ) to rather unbelievable extremes ( 3c 120 ) .
we have plotted the key results in fig .
[ fig : results ] which is a representation of the beaming parameters as a function of the observables ( eq . [ eq : resulta ] ) .
in this section , we deal with the conflicting evidence for general beaming models and for synchrotron models . on the one hand
, some sort of beaming appears to be required by the observation that all of the known jet sources ( excluding of course the ssc terminal hotspots ) produce x - ray emission on only one side , and that is the side which has the only or dominant radio jet and for which relativistic effects have been demonstrated ( usually on vlbi scales ) . on the other hand , for knots such as a1/3c273 or b/3c390.3 ,
the observed x - ray intensity is accommodated by an extension of the power law connecting the radio and optical ( i.e. the synchrotron spectrum ) .
if the x - ray emission from these knots were to be ic / cmb emission enhanced by beaming , this situation would be only a coincidence .
in general , we expect synchrotron spectra to be concave downward when plotted on the usual log s vs. log @xmath11 plane . for high energy electrons ,
this is caused by the e@xmath33 losses .
although we do not have reliable observational evidence on the shape or amplitude of the electron spectrum for @xmath41 , our expectation is that the spectrum observed in the radio regime might flatten at lower frequencies , but it is unlikely to steepen .
this view is dictated by the requirement to avoid divergences in the total number of electrons and integrated energy in the electron spectrum .
reliable x - ray spectra are now available for only a few jets , but so far , the evidence is that the spectra are either the same as observed in the radio / optical ( e.g. 3c 273/a1 , marshall et al .
2001a ) , or steeper ( @xmath42(m87)[email protected] , bhringer et al .
2001 and @xmath42(3c66b)@xmath44 , hardcastle et al .
this supports the synchrotron model and is not in accord with our expectation for beaming from low energy electrons .
however , a relatively hard spectrum is found for pks1127 ( @[email protected] ; siemiginowska et al .
2001 ) , a jet that requires very modest beaming parameters .
similarly , chartas et al .
( 2001 ) find @[email protected] for the outer jet in pks0637 , another likely beamed ic source .
the x - ray spectral index may prove to be a key parameter in differentiating synchrotron from beaming sources .
@xmath45 may be indicative of beaming whereas @xmath46 is the condition we expect from homogeneous synchrotron models . by this
we mean a population of relativistic electrons which is distinct from that we measure in the radio / optical regime . the emitting volume could be quasi coincident with that of the known synchrotron component , or it could be embedded within the known volume . for synchrotron models , a small , flat spectrum component is required . in the case of the 25 knot in 3c 120
, the current optical upper limits demand that @xmath47 .
a small region within the radio knot would be an efficient method to produce the observed x - rays in the sense that the total energy of such an entity would be very modest ( harris et al . 1998 ) .
as noted in sec [ sec : xspec ] , the spectral data so far available do not support this sort of scenario . for the ic / cmb model without beaming , what is required is an excess population of electrons with @xmath23 in the range 500 - 2000 . for bulk
relativistic velocities , this energy range shifts to lower values since the peak of the cmb appears blue shifted to the jet .
the beaming model has untenable consequences for straight jets . in general ,
the angle , @xmath48 , between the jet axis and line of sight to the observer derived for jets such as 3c 273 are not very different from those found at vlbi scales , and this jet appears straight on the kpc scale .
if the jet fluid moves in a straight line along the axis of the jet , then the physical length of the jet is the projected length divided by sin @xmath48 . in table
[ tab : results ] we have listed these values which range up to a mpc for 3c 273 .
very few radio sources are directly observed ( i.e. in the plane of the sky ) to have such large physical sizes . these excessive lengths can be avoided in at least two ways .
we know that some jets bend ( e.g. 3c 120 ) .
if the apparently straight jets ( e.g. 3c 273 ) were to bend significantly in a plane normal to the plane of the sky , they would appear straight but the beaming required for a particular feature would not apply to the whole jet .
the other possibility is that the jet fluid responsible for the observed radiation would not be constrained to move along the jet axis , but , for example , might follow a helical path ( see e.g. meier , koida , and uchida , 2001 ) around the axis as suggested by the hst images of 3c 273 , or the emission regions might represent all or a portion of the flow being deflected toward the observer .
whichever of these possibilities might be operative , it should work in conjunction with the requirement that the jet seen is the one approaching the observer . are there reasonable expectations for kpc scale relativistic velocities aside from their utility in explaining x - ray emission ?
bridle ( 1996 ) marshals considerable evidence for bulk relativistic velocities at least for the inner section of fri jets , and all the way out to the terminal hotspots for frii radio galaxies and quasars .
although representative values for such velocities are not presented , we infer @xmath0 values of at least a few .
wardle & aaron ( 1997 ) , while agreeing that one - sidedness evidence supports relativistic velocities in jets , argue that the observed dispersion in the ratio of the flux density of the inner straight part of the jet ( the feature expected to be most enhanced by beaming ) to the flux density of the lobe ( no beaming ) is less than the expected dispersion for @xmath49 . thus they conclude that the characteristic jet speed is @xmath50 0.85 ( @xmath51 ) and that all jets decelerate between pc and kpc scales .
the consideration of large @xmath0 values for kpc scale jets reminds us of the problem of energy transport .
if electrons were to be a significant component of the jet , they can not escape the ic losses which increase as @xmath52 .
since these losses go as @xmath53 , we may expect that only low energy electrons survive for long distances .
thus the energy must be transported by poynting flux , protons , or cold electrons and positrons .
if the latter , it might be possible to explain the relatively smooth x - ray emission such as that underlying the 3c 273 knots or the pictor a inner jet , as ic beamed emission from essentially cold electrons which are characterized by @xmath54 .
the brighter knots would then rely on the conventional shock acceleration and/or pair production to generate ( in situ ) the usual power law distribution of relativistic electrons with @xmath23 values of normal synchrotron sources .
the synchrotron expectation is that lower frequency emission will persist further downstream since the higher energy electrons have shorter lifetimes .
however , at the site of the shock , there should be strong emission at all frequencies ( e.g. m87 , knot a ) .
in the beaming model , one way to accommodate a leading x - ray feature is to argue for the existence of a precursor shock which is incapable of producing electrons with @xmath55 .
in the optical , the 3c 273 jet has a series of well defined knots with comparable brightnesses and fluxes . however , at radio wavelengths , the knot intensities increase going out the jet and the situation is reversed for x - rays .
this x - ray and radio behavior is mimicked in pks1127 . as a natural consequence , the parameters for the beaming model become less restrictive ( smaller @xmath0 , larger @xmath48 ) as the distance from the core increases .
this sort of behavior is consistent with synchrotron models for which we might well expect that the properties of shocks which permit them to produce copious supplies of the highest energy electrons will diminish as we move out to larger distances .
the beaming model means that the underlying synchrotron components are significantly less energetic than for the stationery synchrotron model .
typical luminosities drop from 10@xmath56 erg s@xmath3 to values like 10@xmath57 erg s@xmath3 .
such entities must be very common in jets , so might we expect to be overwhelmed with x - ray emission from excessive numbers of weak synchrotron emitters which happen to be moving towards us ?
there is little doubt in our minds that thermal bremsstrahlung is not a major contributor to the x - ray emission from most of the jet features discussed in this paper .
we find the arguments in favor of ssc emission from terminal hotspots of a few frii radio galaxies to be convincing .
the failure of simple synchrotron models to fit the spectra of some features , together with the quasi - universality of one - sidedness almost demands the occurrence of relativistic bulk velocities . on the other hand , extreme values for the beaming parameters together with the excellent agreement of the x - ray spectra with the extrapolation of the radio - optical spectra for many features lend strong support to the synchrotron models .
thus the leading contender for most of the x - ray jets appears to be a single or multiple spectral component synchrotron model augmented by a modest beaming with @xmath58 to account for the one - sidedness of most of the recorded x - ray features .
the ic / beaming hypothesis can consistently explain the emission from several jets , and the low required beaming factors for some sources as e.g. pks 1127 make it probable that this emission component contributes substantially to the observed x - ray fluxes of at least some of the observed x - ray features . higher redshifts and steeper
radio spectra appear to favor beaming , while local sources with flatter radio spectra are probably dominated by synchrotron emission .
deh thanks w. tucker for extensive discussions during the early phases of this investigation , t. kneiske for discussions on comparison of beaming formulations , and a. bridle for discussing the radio aspects of jets .
we also thank m. hardcastle , m. birkinshaw , and r. sambruna for communicating results prior to publication and a. siemiginowska and h. marshall for sharing their proprietary data .
m. birkinshaw , d. worrall , w. forman , f. aharonian , and a. marscher provided helpful suggestions for improving the manuscript and the anonymous referee supplied useful criticisms .
hk thanks p. coppi for discussions on agn jets , and acknowledges support by nasa ( nas8 - 39073 and go 0 - 1169x ) .
the work at sao was partially supported by nasa contract nas8 - 39073 .
we first find the synchrotron parameters in the jet frame and then derive expressions for the ic emission .
then we obtain two expressions for the ratio of ic to synchrotron losses : one based on the observed data and one based on the energy densities of phtons and magnetic fields .
equating these two expressions provides the final beaming equation .
we use cgs units throughout except where stated otherwise .
the basic relationships for the parameters which describe the bulk motion of the jet fluid are the jet velocity , @xmath59 ; the lorentz factor , @xmath0 ; the angle of the fluid s velocity with respect to the line of sight of the observer , @xmath25 ; and the dopler beaming factor , @xmath1 .
@xmath60 @xmath61 the required observables are : * the size of the emitting volume * @xmath40 , the low frequency spectral index of the synchrotron spectrum . * @xmath62 and @xmath63 , the lower and upper limits of the synchrotron spectrum in the observer frame .
nb : it may be necessary to decrease @xmath62 to @xmath26 1 mhz so as to include low @xmath23 electrons required for the beaming model .
* @xmath64 at some @xmath65 , with @xmath64 the flux density observed at the earth , and @xmath65 the frequency at the receiver corresponding to @xmath64 .
this provides @xmath66 , the amplitude of the observed synchrotron spectrum : @xmath67 . * the luminosity distance and redshift .
we then move to the jet frame .
frequencies convert as : @xmath68 assuming a knotty jet and neglecting complications arising from different pattern and jet plasma velocities , we find from the lorentz invariant , @xmath69 ( see also eqs .
c7 and c11 of begelman , blandford , & rees 1984 ) , the monochromatic luminosity per solid angle in the jet frame : @xmath70 here , @xmath71 is the luminosity distance and @xmath72 is the @xmath73-correction .
this equation allows evaluation of @xmath74 at any convenient frequency in the jet frame .
integrating over frequency and solid angle , the total luminosity becomes : @xmath75 \left[\frac{k_s\,(\nu_2^{1-\alpha } - \nu_1^{1-\alpha } ) } { ( 1-\alpha)}\right]\ , = \,\frac{l_s}{\delta^4}~{\rm~erg~s^{-1 } } \label{eq : lum}\ ] ] the right term is the energy flux at the earth , @xmath76 computes the luminosity at the source , and @xmath77 transforms into the jet frame . to compute the equipartition magnetic field in the jet frame ( e.g. eq . 7.12 of pacholczyk , 1970 ) :
@xmath78^{2/7}~~{\rm~gauss } \label{eq : beq}\ ] ] v is the emitting volume ; @xmath79 is the volume filling factor ; and @xmath80 is a pacholczyk parameter which is a slowly varying function of @xmath81 , and @xmath82 ( e.g. @xmath80=5.7@xmath83 for @xmath84 ) . when the correct expression for @xmath80 is used in eq .
[ eq : beq ] , it introduces an extra factor of @xmath85 to the numerator within square brackets , so that together with the explicit @xmath77 from eq .
[ eq : lum ] which appears in the denominator , the final dependency within the square brackets goes as @xmath86 and hence : @xmath87 where b(1 ) is the equipartition field calculated for no beaming ( @xmath1=1 ) . from manipulation of the basic equations ( [ eq : gamma1 ] ,
[ eq : delta1 ] ) : @xmath88 since in the jet frame most of the cmb photons will come from the direction in which the jet plasma is moving and the scattered ic emission from a particular electron will be strongly beamed into the direction of the instantaneous velocity vector of the electron , the ic emission will not be isotropic in the jet frame .
take the angle between the l.o.s . to the observer and the jet axis as @xmath89 and @xmath90 in the observer and jet frames , respectively .
likewise , @xmath91 and @xmath92 are the corresponding cosines .
then : @xmath93 where @xmath94 is the velocity of the jet ( see e.g. pacholczyk 1970 , eq . 5.25 ) .
the inverse is also useful : @xmath95 to calculate the power emitted by an electron traveling at an angle @xmath96 to the seed photon direction , we use @xmath97 @xmath90 and follow rybicki & lightman ( 1979 , p. 199
, see also dermer et al
. 1992 ) to get : @xmath98 @xmath99 where @xmath100 is the thomson cross section ; @xmath101 is the energy of the electron ; @xmath102 is the doppler boosted energy density of the cmb ; and @xmath103 , @xmath104 are the lorentz factor and velocity of the electron in the jet frame
. the second term in eq.[eq : kappa ] will be negligible except when @xmath92 approaches minus one , which is the case for electrons moving directly away from the jet s velocity vector .
the energy density of the cmb ( as seen by the jet moving with @xmath0 ) : @xmath105 the peak frequency of the cmb ( again , as seen by the jet ) @xmath106 in the jet frame , the mean energy of photons scattered in the direction @xmath90 will be : @xmath107}~~{\rm~erg } \label{eq : icpower1}\ ] ] @xmath108 @xmath109~~{\rm~erg } \label{eq : icpower3}\ ] ] where @xmath110 is the mean energy of the photons before scattering : @xmath111 , and @xmath112 is the seed photon density in the jet frame , so that the denominator in eq .
( [ eq : icpower1 ] ) is the ( expected ) number of scatterings per unit time for one electron . we define r as the ratio of ic and synchrotron luminosities .
since both of these losses goes as the electron energy squared , all electrons will experience the same value of r. for @xmath40 = 1 , there is equal energy per decade and thus the integration band is not important so that the ratio of luminosities is equal to the ratio of amplitudes of the x - ray ( ic ) and radio ( sync ) spectra .
@xmath113 for @xmath114 , to obtain @xmath115 we change integration parameters from @xmath116 to @xmath117 .
these conversions are found by using eq .
( [ eq : icpower3 ] ) with @xmath104=1 and dropping the -1 : @xmath118~~{\rm~hz } \label{eq : fic}\ ] ] ( the last term is the peak frequency of the cmb in the jet frame ) @xmath119 then most of the factors of eq .
[ eq : lum ] including the integration itself , cancel and the remaining parameters are those resulting from the different conversions of @xmath11 to @xmath23 . after these manipulations ,
we find : @xmath120^{1-\alpha } \label{eq : robs}\ ] ] the ratio of the ic and synchrotron luminosities can be computed from the ratio of energy densities in photons and magnetic field .
this provides an expression for the expected value of r. in using eq.([eq : kappa ] ) , we have dropped the second term since it will normally be much smaller than the first term .
@xmath121}{b_{\rm eq}'^2/8\pi } \label{eq : rexp}\ ] ] where @xmath122 is the cmb energy density in the jet frame , modified by the deviation from the beamed emission power from that computed for an isotropic seed photon field . equating the two expressions for r ( eqs .
[ eq : robs ] and [ eq : rexp ] ) , @xmath123^{\alpha + 1 } \label{eq : result}\ ] ] thus with 4 parameters which depend on the observable data ( @xmath40 , z , b(1 ) , and r(1 ) ) we can solve eq .
[ eq : result ] numerically by finding the range of @xmath124 and @xmath0 pairs which satisfy the equation .
if we ignore the second @xmath0 term and replace b@xmath125 with eq .
[ eq : b1 ] , we obtain an approximate expression for the beaming parameters in terms of the observables ( see fig .
[ fig : results ] ) : @xmath126^{\frac{1}{1+\alpha } } ( 3.808~\times~10 ^ 4)^{\frac{1-\alpha}{1+\alpha}}}{(1+z)^{\frac{\alpha+3}{\alpha+1 } } } \label{eq : resulta}\ ] ]
once we have the allowed values of @xmath127 and @xmath0 , we need to determine where the emission bands occur and which segments of the electron energy spectrum are responsible for the observed radiation . to determine the ic emission frequency of electrons responsible for a particular synchrotron frequency : @xmath128 and the reverse : @xmath129 to find the electron energy responsible for a particular synchrotron frequency , we use eqs [ eq : shiftf ] , [ eq : b1 ] , and [ eq : fsy ] : @xmath130 for the electron energy responsible for an ic frequency we use eqs [ eq : shiftf ] and [ eq : fic ] : @xmath131 for the halflives , we use the normal expression : @xmath132 .
this results in : @xmath133 \right\ } } \label{eq : tauj}\ ] ] where @xmath134 is in @xmath17 g and @xmath135 is in years . making the usual approximations for quantities close to one , and since time intervals in the jet frame are observed at the earth as : @xmath136 an approximate expression for the observed halflife is : @xmath137
aharonian , f.a . 2001 , mnras ( in press ) astro - ph/0106037
the gamma range for the radio ( synchrotron spectrum ) corresponds to the frequency range of 1 mhz to the value of @xmath138(max ) given in a later column .
the synchrotron spectrum is assumed to extend down to 1 mhz in order to include the low energy electrons required for the beaming model . | the emission processes responsible for the observed x - rays from radio jets are commonly believed to be non - thermal , but in any particular case , it is unclear if synchrotron emission or one or more varieties of inverse compton emission predominates .
we present a formulation of inverse compton emission from a relativistically moving jet ( `` ic / beaming '' ) which relies on radio emitting synchrotron sources for which the energy densities in particles and fields are comparable .
we include the non - isotropic nature of inverse compton scattering of the relativistic electrons on photons of the cosmic microwave background ( cmb ) and provide beaming parameters for a number of jets .
a list of x - ray emitting jets is given and the jets are classified on the basis of their morphology and spectral energy distribution to determine their likely emission process .
we conclude that these jets have significant bulk relativistic velocities on kpc scales ; that higher redshift sources require less beaming because the energy density of the cmb is significantly greater than locally ; and that for some nearby sources , synchrotron x - ray emission predominates because the jet makes a large angle to the line of sight . |
xmmu j185330.7 - 012815 is an x - ray object that has its emission nature not yet completely confirmed .
it has been detected in the rosat all - sky survey ( rass ) . based on its extent inferred from the rass data ( i.e. 11 arcmin@xmath17 arcmin ) , schaudel ( 2003 )
has suggested that xmmu j185330.7 - 012815 , which was designated as g31.9 - 1.1 in their work , to be a candidate of supernova remnant ( snr ) in our galaxy .
schaudel ( 2003 ) have further analysed the x - ray properties of xmmu j185330.7 - 012815 by using an archival asca observation in which xmmu j185330.7 - 012815 located at an off - axis angle of @xmath2 arcmin .
although they found that xmmu j185330.7 - 012815 appears to be elongated in an asca gis image , the large off - axis angle precludes any constraining spatial analysis . examining the spectral data collected by asca , schaudel ( 2003 ) found that the x - ray spectrum of xmmu j185330.7 - 012815 is featureless and
can be fitted with an absorbed power - law model with a photon index of @xmath3 ( cf .
table 5.2 in schaudel 2003 ) .
together with the apparently centrally - brightened x - ray morphology , the author claimed that the power - law spectral fit strongly supports the interpretation of a center - filled snr or a crab - like snr .
however , with only @xmath4 source counts from the asca data , one can not unambiguously distinguish between the power - law model and a single - temperature thermal plasma model with @xmath5 kev ( schaudel 2003 ) .
also , the non - detection of any radio emission from the position of xmmu j185330.7 - 012815 makes the snr interpretation questionable . in an archival search for the galactic magnetars , muno et al .
( 2008 ) have made use of 506 archival chandra data and 441 archival xmm - newton data .
this search has included a dedicated xmm - newton observation of xmmu j185330.7 - 012815 with an off - axis angle of only @xmath6 arcmin .
interestingly , the authors have detected a signal with a period of @xmath0 s from this observation . with this discovery , instead of being a magnetar candidate , muno et al .
( 2008 ) have suggested this source is probably an accreting white dwarf .
the cataclysmic variable nature of this source is further supported by the optical spectroscopy performed by j. halpern & e. gotthelf ( private communication reported in muno et al .
2008 ) .
although muno et al .
( 2008 ) have identified xmmu j185330.7 - 012815 as a promising candidate of cataclysmic variable , they have not further analysed and discussed the nature of this object as this is out of the scope of their work . to confirm the x - ray emission properties of xmmu j185330.7 - 012815 , a detailed spectro - imaging and temporal analysis of the aforementioned xmm - newton observation is required and this provides the motivation of this investigation . in 2 , we are going to describe the details of this xmm - newton observation of xmmu j185330.7 - 012815 as well as the procedure of the data reduction .
the method and the results of the data analysis are presented in 3 .
finally , we will discuss the implication and the possible nature of xmmu j185330.7 - 012815 as an accreting white dwarf .
xmmu j185330.7 - 012815 was observed by xmm - newton on 25 - 26 october 2004 ( observation i d : 0201500301 ) .
the x - ray data used in this investigation were obtained with the * e*uropean * p*hoton * i*maging * c*amera ( epic ) on board xmm - newton ( jansen et al .
epic consists of two metal oxide semiconductor ( mos1/2 ) ccd detectors ( turner et al .
2001 ) of which half of the beam from two of the three x - ray telescopes is reflected to .
the other two halves of the incoming photon beams are reflected to a grating spectrometer ( rgs ) ( den herder et al .
the third of the three x - ray telescopes is dedicated to expose the epic - pn ccd detector solely ( strder et al .
the epic - pn ccd was operated in small - window mode with a medium filter to block optical stray light .
this data provide imaging spectral and temporal information .
all recorded events are time - tagged with a temporal resolution of 5.7 ms .
the mos1/2 ccds were setup to operate in full - window mode with a medium filter in each camera .
the mos1/2 cameras provide imaging , spectral and timing information , though the later with a temporal resolution of 2.6 s only .
the aimpoint of the satellite in this observation is ra=@xmath7 and dec=@xmath8 ( j2000 ) . with the most updated instrumental calibration
, we generate the event lists from the raw data obtained from all epic instruments with the tasks _ emproc _ and _ epproc _ of the xmm science analysis software ( xmmsas version 9.0.0 ) . examining the raw data from the epic - pn ccd
, we did not find any timing anomaly which was observed in many of the xmm - newton data sets ( cf . hui & becker 2006 and references therein ) .
this provides us with opportunities for an accurate timing analysis .
we then created filtered event files for the energy range 0.2 kev to 12 kev for all epic instruments and selected only those events for which the pattern was between @xmath9 for mos cameras and @xmath10 for the epic - pn camera .
we further cleaned the data by accepting only the good times when sky background was low and removed all events potentially contaminated by bad pixels . after the filtering
, the effective exposures are found to be 19.5 ks and 13.5 ks for mos1/2 and epic - pn respectively .
apart from the x - ray data , we have also made use of the optical monitor ( om , mason et al . 2001 ) data obtained in standard imaging mode with two filters uvw1 ( an effective wavelength of 2910 ) and uvm2 ( 2310 ) .
the om data are reduced by using the standard xmmsas _
omichain _ task . for imaging and source detection ,
the track history was created , bad pixels were removed and the resulting image was subsequently used for source detection . since individual photons were centroid to one eighth of detector pixel by onboard electronics which produces a noise pattern known as _
modulus 8 _ spatial fixed pattern noise and hence this noise was overcome using _ ommodmap _ task .
the source detection was performed by using the task _ omdetect _ and counts were converted into magnitudes for the corresponding filters with the aid of the task _ ommag_. the uv bandpass counts were converted into fluxes using the recipe provided by alice breeveld .
the composite mos1/2 image of a @xmath11 field around xmmu j185330.7 - 012815 is shown in figure [ mos_img ] .
xmmu j185330.7 - 012815 is observed as the brightest object in this field .
we determined its position and significance by means of a wavelet detection algorithm .
the x - ray position is found to be ra=@xmath12 , dec=@xmath13 [ j2000 ] , where the numbers in the parentheses indicate the @xmath14 statistical uncertainties of the last digit .
the signal - to - noise ratio of xmmu j185330.7 - 012815 is found to be @xmath15 . where @xmath16 is the background counts . ]
this mos1/2 image clearly rules out the claim of extended source as suggested by the rass data ( schaudel 2003 ) .
there are two serendipitous x - ray sources found in the vicinity of xmmu j185330.7 - 012815 , where are labeled as sources a and b in figure [ mos_img ] .
the wavelet detection reports the locations and the significances of these sources to be ( ra=@xmath17 , dec=@xmath18 [ j2000 ] ; @xmath19 ) and ( ra=@xmath20 , dec=@xmath21 [ j2000 ] ; @xmath22 ) for sources a and b respectively . since our focus is on characterising the emission nature of xmmu j185330.7 - 012815 , we will not discuss the properties of these two sources further in this paper .
we estimated the effects of pileup in all epic data by using the xmmsas task _ epatplot_. our results showed that all the epic data were not affected by ccd pileup . in order to maximize the signal - to - noise ratio for xmmu j185330.7 - 012815 , we extracted its source spectrum from circles with a radius of 50 and 30 in the mos1/2 and epic - pn cameras respectively .
this choice of extraction regions corresponds to the encircled energy fraction of @xmath23 in all cameras and at the same time it minimizes the contamination from the nearby x - ray sources .
the background spectra were sampled from the nearby regions with circles of a radius of 60 and 40 in mos1/2 and epic - pn respectively .
response files were computed for all datasets by using the xmmsas tasks _ rmfgen _ and _ arfgen_. after the background subtraction , we have @xmath24 cts , @xmath25 cts , and @xmath26 cts collected for the spectral analysis from mos1 , mos2 and epic - pn cameras respectively . in order to constrain the spectral parameters tightly , we fitted the data obtained from three cameras simultaneously . for the spectrum extracted from each camera , we grouped the data so as to have at least 50 counts per bin .
we found there are fluctuations in the spectral data below 0.3 kev that can be ascribed to the undesirable spectral response .
therefore , we limited all the spectral analysis in the energy range @xmath27 kev .
all the spectral fits are performed with xspec 12.5.1 .
the parameters of the best - fit models are summarized in table 1 .
all quoted errors are @xmath14 for 2 parameters of interest .
the x - ray spectrum of xmmu j185330.7 - 012815 as observed by xmm - newton is displayed in figure [ spectrum ] which has shown the line features suggesting k - lines of heavy elements , in particular the features at @xmath28 kev and @xmath29 kev .
this has led us to examine the spectrum with emission model of hot plasma that includes line emissions from different elements .
first , we attempted to fit the spectrum with mekal which is a code that models the plasma in collisional ionization equilibrium ( mewe et al .
1985 ) . with the metal abundances fixed at the solar values , a single temperature mekal model results in a hydrogen equivalent column density of @xmath30 @xmath31 and a plasma temperature of @xmath32 kev with a goodness - of - fit of @xmath33 ( 583 d.o.f .
the large @xmath34 indicates that this model is unlikely to be the adequate description of the data . in examining the fitting residuals ,
we have identified the systematic deviations at energies larger than @xmath35 kev and smaller than @xmath36 kev .
this prompts us to examine the spectrum with multi - temperature plasma model . with a two - temperature mekal model , the best - fit yields a similar absorption of @xmath37 @xmath31 and the plasma temperatures of @xmath38 kev and @xmath39 kev . in comparison with the single - temperature model , the goodness - of - fit , @xmath40 ( 581 d.o.f . ) ,
is found to be improved significantly .
statistically , the additional component is required at a confidence level @xmath41 .
although the systematic residuals at energies @xmath42 kev and @xmath43 kev are not observed in this two - temperature model , we notice that this model appears to overpredict the emission at the energies around @xmath36 kev and @xmath28 kev .
this leads us to speculate that either the abundances of the metals that give rise to the line emissions at these energies are different from their solar values or the data require an additional component for modeling .
for testing the first hypothesis , we further analysed the x - ray spectrum of xmmu j185330.7 - 012815 by taking metal abundance as free parameter in the fitting .
as the residuals at these energies are likely to be due to fe k@xmath44 and l lines , we take the abundance of iron as a fitting parameter . by entangling the free parameters of fe abundances in both components , we found the residuals at @xmath36 kev and @xmath45 kev can be minimized .
the goodness - of - fit is improved ( @xmath46 , 580 d.o.f . ) .
the model yields the column density of @xmath47 @xmath31 , the plasma temperatures of @xmath48 kev and @xmath49 kev , as well as an iron abundance of fe=@xmath50 .
although the residuals of the lines can be reduced with this model , with a more careful examination , we found that there are still discrepancies between the observed data and this model at energies larger than @xmath51 kev .
this leads us to consider the other possibility , namely whether there is any additional spectral component is required to model this data .
comparison between the best - fit three - temperature plasma model and the observed data is shown in figure [ spectrum ] .
there is no systematic deviation between the data and this model within the entire adopted energy range .
the goodness - of - fit is further improved ( @xmath52 , 579 d.o.f . ) .
different from the two - temperature model , we do not find any significant deviation of the fe abundance from the solar value in this adopted model .
therefore , we fixed all the metal abundances at the solar values in three - temperature plasma model .
it results in the column density of @xmath53 @xmath31 , the plasma temperatures of @xmath54 kev , @xmath55 kev and @xmath56 kev . the unabsorbed flux inferred by this model
is @xmath57 ergs @xmath31 s@xmath58 in @xmath27 kev .
we notice that there is a class of cataclysmic variable which contains a soft blackbody component ( cf .
evans & hellier 2007 ) .
statistically the soft blackbody component is only required at a confidence level of @xmath59 in this observed spectrum .
therefore , we do not consider this spectral contribution in this work .
the asca and xmm - newton spectra of xmmu j185330.7 - 012815 appear to be qualitatively different .
asca observations show no line emission features and the spectrum with a single power - law model and obtained a reasonable goodness - of - fit with the asca data ( @xmath60 , 40 d.o.f . )
( schaudel 2003 ) .
however , with the photon statistic improved by a factor of @xmath61 , the spectral data obtained by xmm - newton clearly show the presence of the emission line features . fitting a power - law to the epic spectrum
does not result in an acceptable goodness - of - fit ( @xmath62 , 583 d.o.f . ) .
similar to the single - temperature plasma model , it can not describe the data below @xmath63 kev and above @xmath35 kev . such discrepancy between the inferences drawn from these two observations
can be ascribed to the relatively poor quality of asca data . to confirm the x - ray period by muno et al .
( 2008 ) , we did a periodicity search by using an epoch folding method after converting the x - ray arrival times to the barycentric times of the solar system .
the best period , which was determined by fitting a gaussian function to the centroid of @xmath64-peak , is @xmath65 s for both pn and mos data under the assumption that the pulses are coherent ( where the errors in @xmath64 are included and the mos1 and mos2 data are combined ) .
this period is consistent with the result reported by muno et al .
if we take the possible irregularity of the pulses into account , then the error should increase to @xmath66 s in which the error is the standard deviation @xmath67 of the gaussian function fitted to the full @xmath68-profile .
muno et al .
( 2008 ) have also pointed out that the phases vary by @xmath69 cycles with a timescale of @xmath70 s which suggest the incoherent nature of the source .
we have also carried out the phase analysis by dividing the pn light curve into 4 segments with a timescale of @xmath70 s and folded each segment at the period of 238.1 s over the epoch mjd 53303.98674 ( figure [ shift_check ] ) . by fitting a sinusoidal model to the data ( see below ) , we obtained the following peak positions in time order : @xmath71 , @xmath72 , @xmath73 , and @xmath74 in phases ( where the errors are at the 90% confidence level ) . taking the uncertainties into account
, we are not able to unambiguously state whether there is a clear , systematic phase shift . furthermore , by running a @xmath68-test among each pair of the time - sliced light curves in figure [ shift_check ] , their distributions are found to be consistent with each other at a confidence level @xmath41 . therefore , we conclude that no convincing evidence for the incoherence can be found in our independent analysis . )
were folded at the period 238.1 s from the epoch mjd 53303.98674 .
the pulse phase has been repeated over two cycles .
a sine curve plus a constant ( @xmath75 ) was fitted to the profiles . ] a pulse or modulation profile folded at the period of 238.1 s shows a single broad peak over one cycle of the data .
it can be approximated by a sinusoidal model plus a constant unpulsed level : @xmath76 ) + c.\ ] ] to see how the profiles vary with energy , we investigate the energy - resolved profiles obtained in three different energy bands , [email protected] kev , [email protected] kev , and 3.0@xmath7710 kev .
for this analysis , we focus on the pn dataset as it provides the superior photon statistic in each of the considered energy bands .
figure [ pulse ] shows the profiles ( plus signs ) , where the backgrounds are not subtracted ( the extracted backgrounds from a photon deficit region in the same ccd chip are very small and they are at most @xmath78% level compared to their source count rates ) .
we fit the sinusoidal model to the profiles with allowing all the parameters to be free and obtain the following results from the best - fit parameters .
goodness of the model - fit is given in each panel of the figure [ pulse ] .
these profiles show no significant phase shift in the chosen energy bands . the pulse amplitude @xmath79 ( cf .
equation 1 ) is found to increase with increasing energy : @xmath80 ( 0.2 - 1.2 kev ) , @xmath81 ( 1.2 - 3.0 kev ) , and @xmath82 ( 3.0 - 10 kev ) . in the other words , a pulsed ( or modulation ) fraction ( i.e. @xmath83 ) increases from @xmath84 to @xmath85 with increasing energy .
all the quoted errors for the temporal results are at the 90% confidence level . ) was fitted to the profiles ( @xmath86 ) . ] within 10 from the x - ray position of xmmu j185330.7 - 012815 , we have detected two uv sources in the om data .
one uv source ( hereafter u1 ) has found to have the magnitudes of m@[email protected] and m@[email protected] with the corresponding significances of 68.31@xmath67 and 53.46@xmath67 respectively . on the other hand ,
the other source ( hereafter u2 ) has the magnitudes of m@[email protected] and m@[email protected] with the corresponding significances of 43.81@xmath67 and 6.70@xmath67 respectively .
the corrected count rate of u1 and u2 for both filters are c@xmath90=3.06@xmath88 0.06 cts / s ( uvw1 ) , c@[email protected] cts / s ( uvm2 ) and c@xmath91=1.92@xmath88 0.05 cts / s ( uvw1 ) , c@[email protected] cts / s ( uvm2 ) respectively .
we subsequently use these count rates to calculate the uv fluxes . for the conversion factors in the uvw1 and uvm2
bandpass of 4.76@xmath110@xmath92 erg @xmath31cts@xmath58@xmath58 and 2.20@xmath110@xmath93 erg @xmath31cts@xmath58@xmath58 , we obtain the fluxes of ( [email protected])@xmath1 10@xmath94 erg @xmath31 s@xmath58 and ( 4.16@xmath88 0.03)@xmath110@xmath94 erg @xmath31 s@xmath58 for u1 in the corresponding filters . for u2 ,
the fluxes are found to be ( [email protected])@xmath1 10@xmath94 erg @xmath31 s@xmath58 and ( [email protected])@xmath1 10@xmath94 erg @xmath31 s@xmath58 in uvw1 and uvm2 respectively . of these two
detected uv sources , u1 is found to have its position coincided with the nominal x - ray position of xmmu j185330.7 - 012815 .
furthermore , while there is no optical counterpart can be found for u2 , we have identified an optical source in @xmath95band at the position of u1 ( see 4 ) .
the x - ray - to - uv flux ratio is found to have the same order - of - magnitude as the x - ray - to - optical flux ratio . in view of these properties
, we suggest u1 is more likely to be the counterpart of xmmu j185330.7 - 012815 .
in this paper , we present a detailed spectro - imaging and temporal analysis of an archival xmm - newton observation of xmmu j185330.7 - 012815 , which has its emission nature not yet been identified unambiguously .
we found that the energy spectrum of xmmu j185330.7 - 012815 obtained by the epic can be well - described by an absorbed multi - temperature plasma model .
the inferred column density is @xmath96 @xmath31 which is far lower than the total galactic neutral hydrogen absorption , @xmath97 @xmath31 , in the direction towards xmmu j185330.7 - 012815 ( kalberla et al . 2005 ; dickey & lockman 1990 ) .
this leads us to rule out the extragalactic origin of xmmu j185330.7 - 012815 . the unabsorbed x - ray flux inferred from the best - fit model is found to be @xmath98 ergs @xmath31 s@xmath58 . together with the uv counterpart identified in om , the inferred x - ray - to - uv flux ratio is @xmath99 . to further constrain the source nature ,
we also search for any optical identification of xmmu j185330.7 - 012815 in the united states naval observatory ( usno)-b1.0 catalogue ( monet et al .
2003 ) . within the @xmath100 x - ray positional uncertainty of xmmu j185330.7 - 012815 ( see 3.1 ) , we have identified one optical counterpart with @xmath101 .
the position of this source is also found to be consistent with that of u1 . by using the @xmath102 inferred from the x - ray analysis to estimate the foreground extinction ,
we calculate the extinction - corrected optical flux in @xmath95band as @xmath103 ergs @xmath31 s@xmath58 which implies @xmath104 .
such low value of x - ray - to - optical flux ratio can safely rule out the possibility as an isolated neutron star which typically has @xmath105 ( cf .
haberl 2007 ) . through our independent search for the periodic x - ray signal from xmmu j185330.7 - 012815 ,
we have identified a peak at @xmath106 s in the power density spectrum .
this has confirmed the periodicity firstly reported by muno et al .
( 2008 ) . nevertheless , different from the result reported by muno et al .
( 2008 ) , we do not find any convincing phase shift in our phase analysis . in any case
, it is unlikely that this signal with such long period comes from rotation - powered pulsars ( cf .
manchester et al .
2005 ) . instead
, we speculate this periodic signal is possibly from a white dwarf . by comparing the uncovered period with the typical values of accreting white dwarf binaries , we further suggest that it is likely to be the spin period of the white dwarf and does not correspond to an orbital period .
this interpretation is further supported by the spectral properties .
the x - ray spectrum of xmmu j185330.7 - 012815 can be described by a three - temperature plasma model and clearly shows the presence of iron lines , which have often been seen in one type of accreting white dwarf binaries , namely the intermediate polars ( ips ) ( e.g. patterson 1994 ; cropper et al .
the additional frequencies in the x - ray power spectra which corresponds to orbital or the beat period is absent in xmmu j185330.7 - 012815 . in the context of an ip interpretation ,
this suggests that material accretes onto the pole via a disk .
although the spin period and the spectral properties strongly favor the ip interpretation , the observed property of monotonically increasing modulation with increasing energy in the light curves requires a further discussion .
in contrast to xmmu j185330.7 - 012815 , many ips show an opposite trend of decreasing modulation with energy ( e.g. norton & watson 1989 ) which is generally explained by the variation of photoelectric absorption in the accretion curtain across the observer s line of sight ( rosen , mason , & crdova 1988 ) .
the accretion rate is expected to have the maximum from the sector of the disc that is closest to the magnetic pole and fall off with deviation from this sector .
this gives rise to a continuous variation of column density along the cross - section of the curtain and hence results in the modulation by photoelectric absorption .
the minimum of the light curve is thus at the phase when the accreting pole , where the absorption is the largest , is pointing toward the observer .
as the effect of photoelectric absorption is more prominent in the soft band than in the hard band , a trend of decreasing pulsed fraction with increasing photon energy is not unexpected in this scenario ( cf .
figure 9 in rosen et al .
1988 ) .
although the behaviour of decreasing amplitude with increasing energy has been observed in many ips , a number of them do deviate from this general trend .
for example , a clear increase in amplitude modulation with increasing energy is also found in v2306 cygni ( wga j1958.2 + 3232 ) by the asca observation ( norton et al .
norton & mukai ( 2007 ) found that an ip candidate xy ari also shows an increasing modulation with increasing energy in the xmm - newton data , whereas it shows a decreasing behaviour with increasing energy in the rxte data ( see figure 1 and figure 5 in norton & mukai 2007 ) .
another ip that shows a modulation different from the general trend is igr j00234 + 6141 in the rxte observations ( see table 2 in butters et al .
however , the xmm - newton / epic - pn observation of this object shows no modulation above 2 kev ( anzolin et al .
there is another ip , pq gem , which shows an increasing modulation in sub - divided low energy bands and a decreasing modulation in high energy bands of asca and rxte observations ( james et al .
on the other hand , a few other ips show a constant modulation in rxte observations ( e.g. butters et al .
2007 , 2008 ) and some show no amplitude modulation in the epic - pn energy band ( e.g. de martino et al . 2005 ) .
since a diversity of temporal behaviour has been observed , energy dependency of amplitude modulation is a weaker criterion to constrain the property of an ip . to further probe the temporal behaviour of xmmu j185330.7 - 012815
, we have also analysed the archival rxte observations for this source ( obsids : 90070 - 04 - 01 - 00 , 90070 - 04 - 01 - 01 , 90070 - 04 - 01 - 02 , 90070 - 04 - 02 - 00 , 90070 - 04 - 02 - 01 and 90070 - 04 - 02 - 02 ) .
however , no pulsed signal from xmmu j185330.7 - 012815 was detected from all these datasets .
on the other hand , we found that the emission line features at @xmath45 kev in the rxte pca spectrum .
hence , we speculate that the non - detection of pulsation in the rxte data can be ascribed to the lack of imaging capability in resolving the source in a crowded region . the upcoming missions with focusing optics for hard x - rays , such as nustar ( harrison et al .
2010 ) , will certainly provide the appropriate instruments for further investigation of this interesting ip candidate .
the x - ray spectra of many ips are described by a complex absorption model as expected in the context of accretion curtain ( e.g. ramsay et al .
however , such additional absorption component is not required in modeling the spectrum of xmmu j185330.7 - 012815 .
the lack of complex and high absorption in xmmu j185330.7 - 012815 might indicate a low inclination angle such that the observed x - rays are not passing through the accretion curtains . on the other hand ,
the observed hydrogen column density ( i.e. @xmath107 @xmath31 ) is much lower than typically observed in other ips ( cf . ezuka & ishida 1999 ) .
the absence of high values of absorption can possibly due to the low accretion rate .
such low absorption have also been seen in several cases , such as ht cam ( de martino et al .
2005 ) , ex hya ( mukai et al . 2003 ) and v1025 cen ( hellier et al . 1998 ) along with a few other ip sources in which a strong soft component was always prominent ( evan & hellier 2007 ) .
nevertheless , such soft component is not observed in the case of xmmu j185330.7 - 012815 . in the scenario of a low inclination angle ,
the missing soft component may simply due to the geometrical effect ( evan & hellier 2007 ) .
alternatively , this component might be much softer than the spectral coverage of xmm - newton epic so that it is not revealed in this observation .
the spectral results of xmmu j185330.7 - 012815 suggest a relatively low shock temperature of @xmath108 kev ( cf .
table 1 ) .
very recently , yuasa et al .
( 2010 ) have investigated the shock temperatures of 15 ips and found two sources also with low temperatures , fo aqr ( @xmath109 14 kev ) and ex hya ( @xmath109 12 kev ) . in another systematic study of 23 ips ,
the lowest temperature is found to be @xmath109 12 kev for the source do dra ( brunschweiger et al . 2009 ) .
furthermore , we notice that the temperature of a new ip rx j0704 has exhibited a change from a high value of @xmath110 kev to @xmath10911 kev in eight months ( anzolin et al .
there is an ip , ae aqr , which shows exceptionally low temperatures plasma of 0.1 - 7 kev ( e.g. choi et al .
1999 ; choi & dotani 2006 ; mauche 2009 ) .
however , for ae aqr , it is widely believed that most of the transferred material from the companion does not reach the magnetic poles due to a magnetic propeller effect ( e.g. , wynn et al . 1997 ; choi & dotani 2006 ) . based on aforementioned temperature distribution in ips
, it is clear that the observed temperature of xmmu j185330.7 - 012815 is not significantly different from a few other ips and also the observed low plasma temperature may not be persistent .
a frequent monitoring of this source is encouraged . in summary , based on our temporal and spectral analysis
, we suggest that xmmu j185330.7 - 012815 belongs to the ip sub - classification of cataclysmic variables .
we would like to stress that , including xy ari , there are only 8 ips / ip candidates which show spin period less than 240 s. owing to this small population , the exact selection criteria to classify a source as an ip is not well - defined for this short - period sub - class , which so far can only be constrained on the basis of the six observational criteria given by patterson ( 1994 ) .
it is not certain if all six of these characteristics should be exhibited by an ip . for a further investigation of this new ip
, a multi - wavelength observation campaign is required to unveil the physical and geometrical configuration of this system .
in particular , the determination of the orbital period of this source through a dedicated optical observation will definitely play a key role in determining the system parameters such as orbital inclination .
furthermore , a dedicated series of x - ray observations is desirable to search for the possible eclipses from xmmu j185330.7 - 012815 . through the timing of the eclipse ingress and egress ,
one is able to determine the size of the x - ray emitting region and hence put an additional observational constraint on the emission nature of this system .
the authors would like to thank the anonymous referee for the useful comments .
cyh is supported by the research fund of chungnam national university in 2011 .
anzolin g. , de martino d. , bonnet - bidaud j. m. , mouchet m. , gansicke b. t. , matt g. , mukai k. , 2008 , a&a 489 , 1243 anzolin g. , de martino d. , falanga m. , mukai k. , bonnet - bidaud j .- m . , mouchet m. , terada y. , ishida m. , 2009 , a&a , 501 , 1047 brunschweiger j. , greiner j. , ajello m. , osborne m. , 2009 , a&a 496 , 121 butters o. w. , barlow e. j. , norton a. j. , mukai k. , 2007 , a&a , 475 , l29 butters o. w. , norton a. j. , hakala p. , mukai k. , barlow e. j. , 2008 , a&a , 487 , 271 butters o. w. , norton a. j. , mukai k. , tomsick j. a. , 2011 , a&a 526 77 choi c. s. , dotani t. , agrawal p. c. , 1999 , apj , 525 , 399 choi c. s. , dotani
t. , 2006 , apj , 646 , 1149 cropper m. , ramsay g. , hellier c. , mukai k. , mauche c. , pandel d. , 2002 , proc . r. soc .
london a , 360 , 1951 den herder et al .
, 2001 , a&a , 365 , 7 de martino d. , matt g. , mukai k. , bonnet - bidaud j .- m .
, gansicke b. t. , gonzalez perez j. m. , haberl f. , mouchet m. , solheim j .- e . , 2005 , a&a , 437 , 935 dickey j. m. , lockman f. j. , 1990 , araa , 28 , 215 evans p. a. , hellier c. , 2007 , apj , 663 , 1277 ezuka h. , ishida m. , 1999 , apjs , 120 , 277 haberl f. , 2007 , ap&ss , 308 , 181 harrison f. a. , et al . , 2010 ,
spie , 7732 , 27 hellier c. , beardmore a. p. , buckley , d. a. h. , 1998 , mnras , 29 , 851 hui c. y. , becker w. , 2006 , a&a , 454 , 543 james cynthia h. , ramsay g. , cropper m. , branduardi - raymont g. , 2002 , mnras , 336 , 550 jansen f. , et al .
, 2001 , a&a , 365 , l1 kalberla p. m. w. et al . , 2005 , a&a , 440 , 775 manchester r. n. , hobbs g. b. , teoh a. , hobbs m. , 2005 , aj , 129 , 1993 mason k. o. , et al . , 2001 , a&a , 365 , l36 mauche c. w. , 2009 , apj , 706 , 130 mewe r. , gronenschild e. h. b. m. , van den oord g. h. j. , 1985 , a&as , 62 , 197 monet d. , et al . , 2003 ,
aj , 125 , 984 mukai k. , kinkhabwala a. , peterson j. r. , kahn s. m. , paerels f. , 2003 , apj , 586 , l77 muno m. p. , gaensler b. m. , nechita a. , miller j. m. , slane , p. o. , 2008 , apj , 680 , 639 norton a. j. , watson m. g. , 1989 , mnras , 237 , 853 norton a. j. , quaintrell h. , katajainen s. , lehto h. j. , mukai k. , negueruela i. , 2002 , a&a , 384 , 195 norton a. j. , mukai k. , 2007 , a&a , 472 , 225 patterson j. , 1994 , pasp , 106 , 209 ramsay g. , wheatley p. j. , norton a. j. , hakala p. , baskill d. , 2008 , mnras , 387 , 1157 rosen s. r. , mason k. o. , crdova f. a. , 1988 , mnras , 231 , 549 schaudel d. , 2003 , phd thesis , ludwig - maximilians universitt mnchen strder l. , et al . , 2001 , a&a , 365 , l18 turner m. j. l. , et al . , 2001 , a&a , 365 , l27 wynn g. a. , king a. r. , horne , k. , 1997 , mnras , 286 , 436 yuasa t. , nakazawa k. , makishima k. , saitou k. , ishida m. , ebisawa k. , mori h. , yamada s. , 2010 , a&a , 520 , 25 + + @xmath111the abundance of iron relative to the solar photospheric values .
+ @xmath112the normalization of mekal model is expressed as @xmath113 where @xmath114 is the source distance in cm and @xmath115 and @xmath116 are the electron and hydrogen densities in @xmath117
. + @xmath118the normalization of the power - law model ( pl ) at 1 kev in units of photons kev@xmath58 @xmath31 s@xmath58 . | we report the results from a detailed spectro - imaging and temporal analysis of an archival xmm - newton observation of a new intermediate polar xmmu j185330.7 - 012815
. its x - ray spectrum can be well - described by a multi - temperature thermal plasma model with the k - lines of heavy elements clearly detected .
possible counterparts of xmmu j185330.7 - 012815 have been identified in optical and uv bands .
the low value of the inferred x - ray - to - uv and x - ray - to - optical flux ratios help to safely rule out the possibility as an isolated neutron star .
we confirm the x - ray periodicity of @xmath0 s but , different from the previous preliminary result , we do not find any convincing evidence of phase - shift in this observation .
we further investigate its properties through an energy - resolved temporal analysis and find the pulsed fraction monotonically increases with energy .
[ firstpage ] binaries : close cataclysmic variables stars : individual ( xmmu j185330.7 - 012815 , g31.9 - 1.1 ) x - rays : stars |
single - molecule magnets ( smms ) @xcite have been proposed as candidates for applications in molecular spintronics .
@xcite especially enticing is the prospect of using an individual smm as a base component of a spintronic circuit which would be capable of storing @xcite or processing @xcite classical and quantum information . in general
, the essential prerequisite for this is a magnetic bistability which in smms stems from a large molecular spin and a strong _ easy - axis _ magnetic anisotropy , given by a parameter @xmath1 .
this tends to fix the spin along an axis determined by the molecular structure , without favoring any specific direction along this axis . in consequence ,
an energy barrier @xmath2 protects the spin of the molecule against reversal between the two opposite , energetically degenerate orientations . from this point of view
, detection of the additional _ transverse _ magnetic anisotropy , characterized by the parameter @xmath3 in the hamiltonian @xmath4 , is crucially important .
such transverse anisotropy can impair the bistability by opening under - barrier quantum tunneling channels for spin reversal .
@xcite these quantum tunneling processes are also of fundamental interest since the spin - dynamics displays hitherto , most techniques aiming to extract the transverse anisotropy parameter @xmath5 are based on the detection of the tunnel splittings it induces , which display a characteristic magnetic field dependence .
@xcite the major challenge for all such approaches is that these splittings are complicated functions of @xmath5 , and even more , the splitting for high - spin states and low magnetic fields are smaller than the parameter @xmath5 itself by several orders of magnitude . using landau - zener spectroscopy
the tunnel splittings have been accurately determined in bulk fe@xmath6 by measuring their pronounced berry - phase oscillations .
@xcite also in bulk crystals and solutions of smms the parameter @xmath5 has been established by different methods , such as high - frequency electron paramagnetic resonance @xcite and inelastic neutron scattering .
@xcite these methods , however , probe large assemblies of molecules , and thus are not designed for investigating the magnetic properties of an individual smm . as a result , little is known about the transverse anisotropy of individual smms in spintronic devices .
in this paper we propose an approach for extracting the parameter @xmath5 of a single molecule by employing electronic transport measurements .
we study a fe@xmath0 smm captured in a gateable junction ( for details see app .
[ sec : experiment ] ) a geometry close to envisaged device structures which is a unique tool for addressing the spin in different redox states of a molecule .
@xcite we show that , as a consequence of the mixing of the spin eigenstates of the smm , the transverse anisotropy significantly manifests itself in transport .
in particular , we predict and experimentally observe characteristic variations of the coulomb peak _ amplitude _ with the magnetic field from which the parameter @xmath5 can be estimated .
importantly , the method proposed here works at temperatures and electron tunnel broadenings @xmath7 exceeding @xmath5 by many orders of magnitude , while @xmath5 in its turn much exceeds the tunnel splittings .
a scheme of a three - terminal smm junction is shown in fig . [ fig_1](a ) .
an smm bridges the source and drain gold electrodes .
an underlying aluminium electrode separated by a few nanometers of aluminium oxide allows for electrical gating of the molecule and , thus , accessing different redox states , see also app .
[ sec : three - terminal_junctions ] .
the chip containing the junctions is mounted on a piezo - driven rotator that enables to change _ in situ _ the orientation between the external magnetic field @xmath8 and the magnetic anisotropy axes of the molecule , which is characterized by angles @xmath9 and @xmath10 as illustrated in fig .
[ fig_1](b ) .
all the measurements are performed at @xmath11 k. smm bridging the junction .
( b ) spatial orientation of an external magnetic field with respect to the principal axes set by the magnetic anisotropy of an smm .
( c ) differential - conductance map , @xmath12 , measured as a function of gate @xmath13 and bias @xmath14 voltages showing two charge states @xmath15 ( neutral ) and @xmath16 ( charged ) for sample a. ( d ) representative coulomb peaks [ corresponding to linear conductance @xmath17 e.g. , marked by dashed line in ( c ) ] measured at different values of the external magnetic field @xmath18 . the bold arrowed lines and color dots serve as a guide for eyes to indicate the non - monotonic change in the coulomb peak height . ] the differential conductance plotted in fig .
[ fig_1](c ) shows the standard signatures of sequential electron tunneling ( set ) through a molecule with two competing charge states tuned by a gate voltage .
@xcite strong high - conductance resonance lines separate adjacent charge - stable coulomb blockade regions , labeled @xmath15 and @xmath16 , from the set regions where transport is possible .
importantly , several fingerprint features of the stable fe@xmath0 smm can be identified : ( i ) high charging energies expected for an individual molecule ; ( ii ) a strong set excitation at approximately 4.8 mev , @xcite specific to fe@xmath0 as it corresponds to the predicted transition energy between the ground ( @xmath19 ) and the first - excited ( @xmath20 ) spin multiplets for the neutral molecule ; @xcite ( iii ) a non - linear shift of the degeneracy peak in the presence of magnetic field as described by gate - voltage spectroscopy ( for details see ref . [ ] and
[ sec : gate - voltage_position ] ) .
moreover , depending on the strength of tunnel coupling @xmath7 , split kondo zero - bias anomalies in coulomb blockade regimes of _ subsequent _ charge states can be observed , which show the zero - field splitting ( zfs ) at the values expected for the fe@xmath0 smm .
@xcite these features also indicate that the molecule is in an intermediate coupling regime with the electrodes , with its upper - limit estimated to be @xmath7=1.6 mev
obtained from the full width at half maximum of the crossing ( degeneracy ) point of the coulomb edges at zero bias , the coulomb peak , for further discussion see app .
[ sec : gate - voltage_position ] .
in a magnetic field the position of the coulomb peak ( cp ) depends both on the _ magnitude _ and the _ orientation _ of an external magnetic field @xmath8 .
@xcite in short , the cp marks the transition between the ground states of two spin multiplets , with spin values @xmath21 and @xmath22 , for the two neighboring charge states .
the energy difference between these states is then a function of @xmath8 , and in particular , it translates into a shift of the linear response degeneracy point in @xmath13 , as shown in fig .
[ fig_1](d ) .
from such a shift one can infer that the ground spin - multiplets of the two charge states evolve differently in the applied field ; therefore , the shift provides information about the magnetic properties of the system .
for example , in simple quantum dots the shift corresponds just to the _ linear _ zeeman effect which is _
isotropic_. @xcite on the other hand , for magnetically anisotropic molecules , like the smms discussed here , not only does the cp shift depend on the relative sample - field orientation , allowing us to extract the value of the angle @xmath9 , but it also provides information about the uniaxial magnetic anisotropy ( @xmath1 ) .
@xcite however , the gate - voltage position of the peak , determined by the low - energy spectrum , is insensitive to the small tunnel splitting corrections induced by the transverse magnetic anisotropy .
below we show that information about the transverse magnetic anisotropy ( @xmath5 ) can instead be inferred from a nonmonotonic dependence of the peak _ amplitude _ @xmath23 , such as in fig .
[ fig_1](d ) , which relies on transition probabilities between different spin states .
we have measured around 200 junctions of which 17 showed clear molecular signatures . from those , 9 samples displayed a clear coulomb peak suitable to perform gate spectroscopy and a magnetic field modulation of @xmath23 .
further discussion of statistics together with differential - conductance maps for several devices are presented in app .
[ sec : statistics ] .
k : ( a ) dependence of the coulomb peak ( cp ) height @xmath23 [ i.e. , the maximal value of @xmath24 , cf .
[ fig_1](d ) ] on magnetic field @xmath18 shown for two different samples where the orientation of the magnetic field lies in the hard plane ( @xmath25 ) .
( b ) analogous to ( a ) for a single sample , except that now @xmath9 is varied and @xmath10 is unknown . note that the evolution of the cp position in magnetic field , and not @xmath23 , was previously analyzed in ref .
[ ] for samples a and c. _ bottom panels : _ theoretical predictions for evolution of the cp height with magnetic field @xmath18 kept in the hard plane : ( c ) for indicated values of @xmath26 and @xmath27 , whereas in ( d ) for several angles @xmath10 and the fixed value of @xmath26 estimated from ( a ) .
bold dashed lines represent the case of @xmath28 for @xmath27 ( c ) and @xmath29 ( d ) .
notice that the shape of @xmath23 for @xmath28 is independent of @xmath10 due to the rotational symmetry around the molecule s easy axis . ] in fig .
[ fig_2](a ) the amplitude @xmath23 of the coulomb peak , normalized to its value at @xmath30 , is plotted as function of @xmath18 for two different samples . for both samples , the gate - voltage analysis of the peak position allows us to conclude that the magnetic field lies in the hard plane ( @xmath31 ) , see app . [ sec : gate - voltage_position ] .
interestingly , @xmath32 for the two samples exhibits a significantly different behavior .
if only uniaxial magnetic anisotropy was present ( @xmath33 ) , the transport properties of the molecule would be left unaffected upon rotation of the field in the hard plane . on the contrary , for @xmath34 this rotational symmetry
is broken .
the dissimilar behavior of the amplitude @xmath23 as observed in fig .
[ fig_2](a ) is therefore attributed to different values of the angle @xmath10 in the presence of a non - zero @xmath5 .
similar curve shapes have been observed in additional samples , as shown in fig .
[ fig_9 ] .
although the values of @xmath5 for bulk samples / monolayers of smms are typically small ( for fe@xmath0 @xmath35 ) @xcite , the linear conductance through a molecule appears to be measurably influenced by it .
a similar change in the field - evolution of @xmath23 is also observed in a single sample c , shown in fig .
[ fig_2](b ) , by rotating the sample holder relative to the magnetic field .
in order to understand how the transverse magnetic anisotropy could qualitatively affect the linear conductance through an smm ( i.e. , the cp amplitude ) , while hardly influencing its gate - voltage position , we use a minimal molecular quantum - dot model based on two giant - spin hamiltonians , @xcite @xmath36 , \ ] ] one for each charge state . here , @xmath37 accounts for the magnetic anisotropy of the smm in the @xmath38th charge state , @xmath39 , \ ] ] with the first / second term representing the uniaxial / transverse magnetic anisotropy , and @xmath40 is the zeeman term ( @xmath41 ) . we combine this with a master equation description of the set transport to nonmagnetic electrodes with tunnel coupling @xmath7 . @xcite the essential steps of this approach are provided in app .
[ sec : transport ] .
the appearance of a clear cp in the experiment restricts @xmath42 ( otherwise spin - blockade would be seen ) @xcite . for the fe@xmath0 smm we can estimate @xmath19 and @xmath43 @xmath44ev for the neutral state , whereas from the cp position dependence we obtain @xmath45 , and fix @xmath46 @xmath44ev with approximately collinear easy axes for both charge states , all in agreement with previous measurements , @xcite see also app . [ sec : gate - voltage_position ] .
we assume that upon charging only the overall energy scale of the magnetic anisotropy changes , i.e. , @xmath47 , leaving just a single parameter @xmath48 for the transverse anisotropy . in fig .
[ fig_2](c ) we plot the calculated cp amplitude @xmath23 for @xmath25 and @xmath27 as a function of the applied field @xmath18 .
surprisingly , the calculations reveal that a non - zero value of @xmath5 significantly influences the current through the molecule . by adjusting the parameter @xmath26 , qualitative agreement with the measured amplitude variation
is obtained for sample a when @xmath49 .
the dissimilar behavior of @xmath23 between samples a and b is then qualitatively reproduced when assuming strongly differing values of the angle @xmath10 as shown in fig .
[ fig_2](d ) . from the shape of the curves we estimate the value of @xmath10 to be @xmath50 for sample
a and @xmath51 for sample b. note that the minimum of @xmath23 for @xmath29 appears in fig .
[ fig_2](d ) at a somewhat larger @xmath18 field value than for sample b which signifies larger @xmath26 , cf .
[ fig_13]-[fig_15 ] . therefore , combining the information from figs .
[ fig_2](c)-(d ) , the cp amplitude could be used to estimate the values of @xmath5 and @xmath10 .
the obtained rough estimate @xmath52 is larger than the values reported for bulk samples , @xcite as also suggested by xmcd experiments on fe@xmath0 monolayers deposited on gold .
@xcite @xmath44ev and @xmath53 and @xmath8 along the hard axis ( @xmath25 and @xmath27 ) : ( a ) conductance @xmath32 traces for various temperatures over the range @xmath54 k at intervals of 0.2 k. ( b ) occupation probabilities for several lowest - energy states in the spin multiplets for @xmath15 and @xmath16 at @xmath11 k. here , @xmath55 ( @xmath56 ) labels the states in order of increasing energy for @xmath15 ( @xmath16 ) , with @xmath57 ( @xmath58 ) denoting the ground state .
( c ) relevant transition energies @xmath59 for @xmath60 determining the set processes at the coulomb resonance ( note that @xmath61 is restored for each @xmath18 by tuning @xmath13 )
. different colors of lines are used to distinguish groups of transitions with respect to possible combinations of indices @xmath56 and @xmath55 ( see the main text ) . for the association of these lines with specific transitions as well as the energies of individual levels
see fig .
[ fig_4 ] .
( d ) evolution of the current vs. magnetic field at @xmath11 k calculated by including a restricted number of states per spin - multiplet up to @xmath62 , where @xmath63 , showing that for small @xmath62 significant deviations are found compared to the calculation involving all the states ( dashed line ) , used in all other plots .
for a precise definition of the current @xmath64 see app .
[ sec : transport ] . ] to gain deeper insight into the mechanism leading to a modulation of @xmath23 we analyze in fig .
[ fig_3](a ) how the calculated @xmath18-traces of the cp amplitude evolve with temperature .
the appearance of a maximum at around @xmath65 t ( marked by the vertical dashed line ) and its enhancement with increasing temperature suggests that this feature is build up from contributions of many excited states of the smm .
this is indeed confirmed by inspection of the evolution of the occupation probabilities shown in fig .
[ fig_3](b ) for the experimental temperature @xmath11 k. to obtain this figure we first find the eigenstates of @xmath37 , given by eq .
( [ eq : h_n ] ) .
for @xmath66 we obtain two sets of eigenspectra , @xmath67 and @xmath68 . here ,
@xmath55 and @xmath56 label the states in order of increasing energy , starting from @xmath57 ( @xmath58 ) for the neutral ( charged ) _ ground _ state . using these energies and states ,
we calculate the probabilities from the master equation .
one should note that the energies ( not shown ) and occupation probabilities of corresponding states ( @xmath69 ) for different charge are very similar . from fig .
[ fig_3](b ) , however , it is not clear which of the maxima of the probabilities is responsible for the maximum of the @xmath32 curves , indicated by the vertical dashed line .
( c ) , but now for each transition - energy line we specify the initial and final states , with respective energies @xmath70 and @xmath71 , between which the transition occurs . recall that @xmath56 is an index which numbers states in a given spin multiplet with respect to energy , with @xmath58 denoting the ground state . moreover , by labelling the lines with @xmath72 we mean that @xmath56 refers to the final state of a charged smm ( @xmath16 ) whereas @xmath55 represents the initial state of a neutral smm ( @xmath15 ) .
we note that information shown in ( a ) can not be readily seen from energies @xmath73 ( @xmath66 ) of the individual levels , which for the completeness of the present discussion are plotted in ( b ) .
observe that since energies in ( b ) are calculated at the coulomb resonance , the curves for @xmath58 overlap . ] instead , to understand the @xmath32 dependence in fig .
[ fig_3](a ) one has to consider the transition energies @xmath59 between levels of different charge states .
this is demonstrated in fig .
[ fig_3](c ) , where the horizontal dashed lines represent the available thermal energy .
the transition energies fall into three generic groups : ( i ) low - energy transitions ( @xmath69 green lines ) ; ( ii ) transitions of low energy for small @xmath18 but high energy for large @xmath18 ( @xmath74 or @xmath75 orange lines ) ; ( iii ) high - energy transitions ( remaining @xmath56 and @xmath55 pairs blue lines ) .
importantly , the temperatures used in fig .
[ fig_3](a ) lie just below the group of transition - energy curves exhibiting a minimum at finite magnetic fields roughly between 2 - 4 t ( blue curves in fig .
[ fig_3](c ) ) . as the magnetic field
is augmented from zero , these curves thus initially approach the thermal energy ( horizontal dashed lines ) before moving away at higher fields towards their high - field asymptotes .
this leads to an enhancement of @xmath23 for @xmath76 t , followed by a steady decrease , i.e. , the characteristic non - monotonic behavior experimentally observed in fig .
[ fig_2](a ) .
we emphasize that the above mechanism does not constitute a purely spectroscopic method : the current and probabilities depend on both the _ energies _ and _ quantum states _ , which determine the tunnel rates .
the importance of including many excited states in the calculation is quantified in fig .
[ fig_3](d ) , where we show how the non - monotonic behavior can be strongly overestimated when including too few excited states , see also figs .
[ fig_13]-[fig_16 ] .
we note that some additional remarks regarding signatures of the transverse anisotropy parameter @xmath5 in the peak amplitude of @xmath23 are discussed in app .
[ sec : signatures_e ] .
finally , worth of note is the larger - than - predicted modulation of the cp amplitude observed in the experiments .
we briefly comment on the verifications to rule out some other contributions that could lead to such an amplification .
first , the master equation analysis was constrained to a weak tunnel - coupling @xmath7 as compared to temperature .
we verified that higher - order tunnel processes that lead to broadening and inelastic tunneling do not increase the scale of the modulation of the cp height .
for this we employed a perturbative approach including next - to - leading tunneling processes @xcite and non - perturbative numerical renormalization group ( nrg ) method .
@xcite second , we assumed symmetric tunnel - coupling of the smm to both electrodes with the same energy @xmath7 .
one can show that a junction asymmetry gives rise to an overall constant factor suppressing the conductance @xmath23 .
thus , this can not change its field dependence .
third , the addition of higher - order magnetic anisotropy terms to the smm model , eq .
( [ eq : h_n ] ) , is also not likely to affect the magnitude of the modulation .
we checked , for instance , the effect of the 4th order transverse anisotropy of the form @xmath77 $ ] , for a range of values of the parameter @xmath78 for which this term competes with the 2nd order transverse term . we thus conclude that the intensity of the modulation may rely on some intrinsic amplification mechanism not captured by our model , i.e. , going beyond the giant - spin model , @xcite when considering a single electron interacting with the molecule .
we summarize here in a few steps how to determine magnetic anisotropy of an individual smm ( see eqs .
( [ eq : h_smm])-([eq : h_n ] ) and app .
[ sec : model ] ) by exploiting the information contained both in the coulomb peak _ position _ as well as in the magnetic field evolution of its _ amplitude_. in particular , the method under discussion allows for finding both the magnetic anisotropy constants @xmath79 and @xmath80 in two charge states ( i.e. for @xmath66 ) of an smm , and the orientation of an external magnetic field relative to the molecule s principle axes , given by the angles @xmath9 and @xmath10 .
a. let us first consider only the coulomb peak _ position _ , shown in the left panel of fig .
[ fig_5 ] . as explained in ref .
[ ] , by analyzing the position of the coulomb peak one can immediately conclude whether a molecule captured in the junction exhibits magnetic anisotropy at all . if the molecule is _ spin - isotropic _ , one observes a linear dependence on the magnetic field [ see dashed line in figs .
[ fig_5](a)-(b ) ] that reflects the linear zeeman effect . on the other hand , if the molecule is _ spin - anisotropic _ , this dependence becomes nonlinear , and the uniaxial magnetic anisotropy parameter @xmath79 together with the angle @xmath9 can be estimated from it .
this , in turn , permits for systematic adjustment of the magnetic field s orientation so that the field is kept perpendicular to the molecule s easy axis @xmath81 , which corresponds to @xmath25 . b. the transverse magnetic anisotropy breaks the molecule s rotational symmetry around the easy axis @xmath81 ( see also app .
[ sec : magnetic_anisotropy ] ) . in consequence ,
one expects that such a symmetry breaking should manifest itself in different transport characteristics of the system occurring for various orientations of the magnetic field in the hard plane ( i.e. , the plane perpendicular to the easy axis ) . from figs .
[ fig_5](a)-(b ) it is clear that the sole position dependence in practice does not allow one to derive reliably either the transverse magnetic anisotropy constant @xmath80 or the angle @xmath10 . for this purpose ,
also the _ amplitude _ of the coulomb peak has to be taken into consideration . c. the presence of transverse magnetic anisotropy
can be confirmed by observation of how the field dependence of the coulomb peak amplitude changes when rotating the field orderly in the hard plane , or in other words by varying the angle @xmath10 .
specifically , one should notice then _ two _ significantly different shapes of the amplitude showing up every 90@xmath82 , cf .
red lines with others in the right panel of fig .
[ fig_5 ] .
these two limiting cases represent the situation when the magnetic field lies either along the molecule s hard axis @xmath83 ( @xmath27 or @xmath84 ) , fig .
[ fig_5](c , e ) , or along the molecule s intermediate axis @xmath85 ( @xmath29 or @xmath86 ) , fig .
[ fig_5](d , f ) .
consequently , this enables one to determine the approximate value of the angle @xmath10 .
d. the effect of transverse magnetic anisotropy on the coulomb peak amplitude should be most pronounced for the magnetic field aligned along the molecule s hard axis @xmath83 , see app .
[ sec : magnetic_anisotropy ] and figs .
[ fig_2](c)-(d ) of the main article . for a sufficiently high temperature @xmath87
[ see fig .
[ fig_5](f)-(g ) and fig .
[ fig_16 ] ] one observes then formation of local maxima , marked by red arrows in figs .
[ fig_5](c , e ) , from whose position the value of the transverse magnetic anisotropy constant @xmath80 can be numerically estimated .
importantly , the method under discussion relies on a simultaneous fitting of _ position _ ( sensitive to @xmath79 ) and the _ amplitude _ ( sensitive both to @xmath79 and @xmath80 ) of the coulomb peak .
this strictly limits the freedom of the parameters choice , basically leaving @xmath80 to be determined from the field value at which the maximum amplitude is acquired . for instance , making the parameters @xmath79 smaller by 25% than the one used above ( given the fixed experimental temperature @xmath11 k ) , while assuming @xmath88 , may also produce a maximum , see green lines in figs .
[ fig_5](c)-(f ) .
however , not only does it result in peak positions at completely wrong magnetic fields [ cf .
position of green and red arrows in fig .
[ fig_5](e ) ] , but also the amplitude shape remain unaltered upon changing the orientation of the field in the hard plane [ cf .
red and green lines between figs .
[ fig_5](e ) and ( f ) ] .
this restriction , combined with the sensitivity of the qualitative curve shape of the conductance to the parameters is advantageous for extracting the anisotropy parameters of smms _ in situ_.
in conclusion , we have proposed a new method of probing the transverse magnetic anisotropy of an individual smm embedded in a three - terminal device .
it exploits the information contained in the spin states of the molecule through the analysis of the magnetic field evolution of the linear conductance amplitude @xmath23 .
we found that the evolution of @xmath23 in a magnetic field could only be reproduced when including a sufficient number of excited states .
estimates for the transverse anisotropy of the fe@xmath0 smm yield @xmath89 @xmath44ev , a value of @xmath5 significantly larger than the observed bulk / monolayer values .
this is expected for a molecule captured in the low symmetry environment of a transport junction .
importantly , the technique does not rely on the small induced tunneling effects and hence works well at temperatures by far exceeding the tunnel splittings and even @xmath5 itself .
our measurements find larger modulation of @xmath23 than calculated and the origin of this enhancement requires further study .
this method may facilitate the detection of _ in - situ _ mechanical tuning @xcite or excitation @xcite of magnetic anisotropy of a single molecule .
this work was supported by nwo ( veni ) and ocw , and by the eu fp7 project 618082 acmol and advanced erc grant ( mols@mols ) .
m.m . acknowledges the financial support from the alexander von humboldt foundation .
k.p . was supported by u.s .
national science foundations dmr-1206354 .
single - molecule magnet : ( a ) sketch of the magnetic core of the fe@xmath90 smm .
( b ) ground - state spin multiplet ( @xmath19 ) of the fe@xmath90 smm in a neutral charge state @xmath15 for further explanation see app .
[ sec : magnetic_anisotropy ] .
( c ) depiction of the fe@xmath90 smm illustrating the orientation of the phenyl rings [ omitted in ( a ) ] that terminate the molecule .
note that both in ( a ) and ( c ) hydrogen atoms are disregarded for clarity . ]
we used an fe@xmath90 smm with formula [ fe@xmath90(l)@xmath91(dpm)@xmath92@xmath93et@xmath91o where hdpm is 2,2,6,6-tetramethyl - heptan-3,5-dione and h@xmath94l is the tripodal ligand 2-hydroxymethyl-2-phenylpropane-1,3-diol , which carries a phenyl substituent .
@xcite in the bulk phase , the crystallographic symmetry is @xmath95 .
@xcite the magnetic core of the fe@xmath90 smm is made of 4 fe@xmath96 ions ( each with spin @xmath97 ) as illustrated in fig .
[ fig_6](a ) .
the antiferromagnetic exchange interaction between the central and peripheral ions yields a large molecular spin @xmath98 in the ground state .
magnetic anisotropy due to the interaction with the crystal field lifts the degeneracy of the spin multiplet into five doublets and one singlet that are distributed over an energy barrier as shown in fig .
[ fig_6](b ) for further discussion see app .
[ sec : magnetic_anisotropy ] .
the height of the barrier , which hinders the spin reversal , is given by @xmath99 , where @xmath1 is the uniaxial magnetic anisotropy parameter .
in the case of bulk fe@xmath90 the height is @xmath100=1.4 mev .
@xcite the ` _ _ zero - field splitting _ _ ' ( zfs ) , defined as the energy difference between the two lowest - lying doublets ( @xmath101 and @xmath102 ) is 0.5 mev .
the low symmetry of the molecule induces a transverse magnetic anisotropy @xmath5 that , in bulk , is @xmath103 @xmath44ev from epr measurements .
@xcite finally , we note that the molecule contains two axial tripodal ligands @xmath104 which hold the core together and six peripheral dpm@xmath105 ligands that create an hydrophobic envelope , see fig .
[ fig_6](c ) .
the three - terminal junctions are fabricated on a silicon substrate covered by 280 nm of sio@xmath106 .
the schematics of the fabrication process is described in fig .
[ fig_7](a ) .
the gate electrode is fabricated by e - beam lithography and subsequent e - beam deposition of al . in the next step ,
the oxidation of the gate in a controlled oxygen atmosphere produces a dielectric coating layer of 2 - 3 nm of al@xmath106o@xmath107 .
the source and drain electrodes are fabricated by self - breaking , controlled electromigration of a au nanobridge deposited by e - beam lithography on top of the oxidized gate .
the self - breaking technique prevents the formation of gold nano - grains in the junction that could mimic the behavior of a molecule .
figure [ fig_7](b ) shows a scanning electron microscope image of a device before electromigration .
the molecules are deposited onto the chip by drop casting a @xmath108 m solution in toluene into a liquid cell containing the chip with the junctions .
the electromigration of the bridge and subsequent self - breaking are carried out in solution to maximize the yield of junctions with a molecule .
the molecule - electrode coupling @xmath7 is estimated from the broadening of the coulomb edge at low bias .
in particular , the full - width at half - maximum ( fwhm ) of the coulomb peak is used for this purpose .
we find 1.6 mev , 2.0 mev and 1.4 mev for samples a , b and c respectively .
note , however , that these values are an upper limit for @xmath7 since we can not resolve the presence of additional components for the broadening such as thermal energy or the contribution of other molecular levels very close in energy . , calculated from the giant - spin hamiltonian , eqs .
( [ eq : h_smm])-([eq : h_n ] ) . from the fit
we get the following values for sample a in ( a ) : @xmath109 @xmath44ev , @xmath110 and @xmath111 ; for sample b in ( b ) : @xmath112 @xmath44ev , @xmath113 and @xmath114 ; for sample c : in ( c ) @xmath115 and @xmath116 , whereas in ( d ) @xmath117 and @xmath118 , with @xmath119ev in both cases .
we note that the evolution of the coulomb peak position in magnetic field , and not @xmath23 , for samples a and c was previously analyzed in ref .
[ ] . also note that in the fitting for sample a we included @xmath120 and @xmath121 obtained in fig .
[ fig_2 ] . ]
figure [ fig_8 ] shows the coulomb peak ( cp ) position in gate voltage @xmath13 as a function of the magnetic field for the samples a , b and c described in the main text .
the values of @xmath13 are multiplied by the gate coupling @xmath122 to obtain energy units ( @xmath123 ) and subsequently re - scaled to make @xmath124 for @xmath30 .
the non - linearity of the field dependence is a clear signature of the magnetic anisotropy as described in the text ( see also ref . [ ] ) .
moreover , the low - field ` flatness ' of @xmath123 observed in figs .
[ fig_8](a)-(c ) is indicative of a high value of @xmath9 in contrast with fig .
[ fig_8](d ) .
the solid lines in fig .
[ fig_8 ] are a fit of the data to @xmath125 as defined by the giant - spin hamiltonian , eqs .
( [ eq : h_smm])-([eq : h_n ] ) , and also discussed in detail in app .
[ sec : model ] .
the cp position is mainly insensitive to @xmath5 ( see also supporting information of ref . [ ] ) , and therefore we can independently extract the parameters @xmath1 and @xmath9 related to the uniaxial anisotropy . note that we fix the value of @xmath126 ( neutral state ) to the bulk value @xmath127 @xmath44ev and thus the free parameters are @xmath128 , @xmath129 and @xmath130 .
see the caption of fig .
[ fig_8 ] for the fitting values of these parameters .
we measured around 200 electromigrated junctions from which 17 showed molecular signatures .
a total of 9 molecular junctions displayed a clear coulomb peak suitable for further analysis by means of the gate - voltage spectroscopy method , from which the junctions were proven to exhibit magnetic anisotropy .
importantly , all these junctions displayed a modulation of the peak amplitude @xmath131 as a function of the magnetic field .
a total of 6 of these samples could be rotated or were close to @xmath132 . from those , one sample was close to @xmath121 ( hard axis ) , and it is referred to as sample a. figure [ fig_9 ] shows the differential - conductance maps , @xmath12 , and corresponding magnetic field evolutions of @xmath23 for different fe@xmath0 molecular junctions , that is other than samples a , b and c discussed in the main text .
the top panel [ ( a)-(c ) ] of fig . [ fig_9 ] presents samples for which the gate spectroscopy yields low values of @xmath9 .
worthy of note is that for @xmath133 t a decrease of @xmath23 is observed with increasing @xmath134 . on the other hand , the bottom panel [ ( d)-(f ) ] of fig .
[ fig_9 ] shows examples where @xmath135 ( i.e. , close the the hard plane ) . the shape of @xmath23 for ( d ) and ( e ) indicates that the magnetic field is close to the intermediate axis ( @xmath136 ) , which follows from the analysis carried out in the main text . for the last sample ,
[ fig_9](f ) , the field is most likely at an intermediate angle @xmath10 in the hard plane . in order to discard the influence of universal conductance fluctuations induced by the magnetic field in the measurements , in fig .
[ fig_10 ] we plot @xmath131 as a function of @xmath18 for the samples shown in figs .
[ fig_2](a)-(b ) for both positive and negative polarities of magnetic field .
we note that the main features , like the minima or maxima around 4 t , are reproducible under inversion of the field polarity .
universal conductance fluctuations are not expected to be symmetric by changing the @xmath18 polarity .
some additional symmetric structure appears also in the measurements .
the analysis of this smaller contribution is interesting but beyond the scope of this work . :
dependence of the scaled coulomb peak height @xmath137 on magnetic field @xmath18 for the samples discussed in the main text , cf .
[ fig_2](a)-(b ) , showing that the curves are symmetric upon reversal of the field polarity . ] , measured as a function of magnetic field @xmath18 at two different points : ( a ) @xmath138 v and @xmath139 mv and ( b ) @xmath138 v and @xmath140 mv , which correspond to the co - tunneling background in the left - hand charge state of sample a , cf . fig .
[ fig_1](c ) . ]
if present , conductance fluctuations would equally appear in the zero - bias and the higher bias conductance .
therefore , in order to rule out their presence , we have analyzed the magneto - resistance at higher biases and different gate voltages .
figure [ fig_11 ] shows differential conductance , @xmath12 , as a function of @xmath18 measured at two different and fixed gate @xmath13 and bias @xmath14 voltages in the coulomb blockade in sample a. we observe an almost flat response of @xmath12 with peak - to - peak variation of the order of 0.1 ns .
this magnitude is not comparable to the modulations we attribute to the presence of the transverse anisotropy .
moreover , note that these two spectra are not symmetric by reversing the magnetic field polarity .
thus , we conclude that the universal conductance fluctuations are not significant in our measurements .
the central element of the theoretical description of the gate - spectroscopy technique is a proper choice of the model capturing essential features of an smm .
as introduced in the main article , the molecule is represented by a model based on two giant - spin hamiltonians .
@xcite this allows us to take into account the fact that by tuning a gate voltage @xmath13 the molecule can be switched between two different charge states , @xcite referred to as a _ neutral _ ( @xmath15 ) and _ charged _ ( @xmath16 ) one . in general , each of this states can be characterized not only by different values of molecular ground - state spin ( @xmath21 and @xmath22 ) , but also uniaxial ( @xmath126 and @xmath128 ) and transverse ( @xmath141 and @xmath142 ) magnetic anisotropy constants . using the spin raising
/ lowering operators @xmath143 , the hamiltonian of an smm in the charge state @xmath38 and subject to an arbitrarily oriented external magnetic field @xmath8 takes the form given by eqs .
( [ eq : h_smm])-([eq : h_n ] ) and the zeeman term explicitly given by @xmath144 , \end{gathered}\ ] ] with the angles @xmath9 and @xmath10 defined as illustrated in figure 1b .
noteworthily , by keeping the same value of @xmath9 and @xmath10 for both charge states , we implicitly assume that the orientation of the molecule s principle axes set by magnetic anisotropy is not affected by charging .
this assumption not necessarily holds for real systems as shown in refs .
[ ] and [ ] .
however , since the tilting , if observed , usually does not exceed few degrees , we do not include such an effect into the present considerations .
[ [ secmagnetic_anisotropyhow - does - magnetic - anisotropy - affect - the - energy - spectrum - of - a - large - spin ] ] [ sec : magnetic_anisotropy]how does magnetic anisotropy affect the energy spectrum of a large spin ? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ before we analyze how electronic transport probes the transverse magnetic anisotropy of a molecule , it may be instructive first to discuss the consequences of the transverse magnetic anisotropy and external magnetic field for the smm s energy spectrum . to begin with , as long as the transverse magnetic anisotropy is vanishingly small the system can be described simply by the first term of the hamiltonian ( [ eq : h_n ] ) . as a result ,
the eigenvalues @xmath145 of the spin operator @xmath146 become good quantum numbers for labelling the eigenstates of @xmath147 , that is @xmath148 .
for @xmath149 the energy spectrum of an smm takes the form of an inverted parabola with an energy barrier of height @xmath150 for spin reversal , which basically corresponds to the indirect transition between the ground states @xmath151 and @xmath152 by climbing the barrier _ via _ the intermediate states @xmath153 ( for @xmath154 ) , see figs .
[ fig_12](a)-(b ) .
importantly , the excitation energy between the ground state @xmath155 and the first excited state @xmath156 , the so - called _ ` zero - field splitting ' _ @xmath157 , sets the threshold energy scale for the reversal process to take place .
note that transition energies between neighboring excited states @xmath153 and @xmath158 with @xmath159 are characterized by energies @xmath160 ( for @xmath161 ) that are smaller than the zfs , and these states remain generally unpopulated until the ground - to - first excited state transition becomes energetically permitted .
this bottleneck behavior manifest then in electronic transport through an smm , where it can be observed as a step - like feature in the conductance only at bias voltages @xmath162 .
@xcite the relatively simple picture presented above is not valid , however , if the transverse magnetic anisotropy ( or an external magnetic field perpendicular to the molecule s easy axis ) is significant .
when @xmath34 , the second term of the hamiltonian ( [ eq : h_n ] ) breaks the system s rotational symmetry around the easy axis @xmath81 , so that @xmath145 is no longer a good quantum number .
in fact , each of the @xmath163 eigenstates of @xmath164 $ ] is now a linear combination of the eigenstates @xmath153 , which , in turn , underlies the origin of the quantum tunneling of magnetization .
@xcite in particular , each of these eigenstates is formed from states @xmath153 belonging to one of two uncoupled , time - reversed sets , as shown in figs .
[ fig_12](c)-(d ) . for an _ integer _
spin @xmath165 , the transverse magnetic anisotropy leads to splitting of energy levels , usually referred to as _ ` tunnel - splittings ' _ , fig .
[ fig_12](c ) , whereas for a _ half - integer _
spin @xmath165 ( in the absence of magnetic field ) according to the kramers theorem each energy level is doubly degenerate , fig .
[ fig_12](d ) .
interestingly , if one applies an external magnetic field in the direction perpendicular to the system s easy axis @xmath81 , periodic changes of these tunnel - splittings can be observed if the field is oriented along or close the hard axis @xmath83 , figs . [
fig_12](e)-(f ) , and they disappear as the field gets rotated towards the direction of the medium axis @xmath85.@xcite for a weak tunnel - coupling between an smm and electrodes , transport in the single electron tunneling ( set ) regime can be considered in the leading - order perturbative approach ( fermi golden rule combined with a master equation).@xcite we describe metallic , _
nonmagnetic _ electrodes [ @xmath166 as reservoirs of noninteracting electrons , whose tunneling processes to / from a molecule are modelled by the following hamiltonian @xmath167 with @xmath168 where @xmath169 is the tunneling matrix element , @xmath170 represents creation of an electron with spin @xmath171 in the molecular orbital @xmath172 , and @xmath173 denotes the annihilation operator for the @xmath174th electrode with @xmath56 standing for an orbital quantum number . note that the molecular state has been expanded in the basis of eigenvectors @xmath175 and @xmath176 of @xmath177
next , we express the molecular eigenstates @xmath178 and @xmath179 with respect to the basis of angular momentum ( spin ) eigenstates . in principle , an arbitrary molecular state can be decomposed as @xmath180 . as a result , one obtains @xmath181 the key problem one encounters when analyzing the above equation is that the operator @xmath170 involves two degrees of freedom , namely , the orbital one ( @xmath172 ) and the spin one ( @xmath171 ) .
consequently , it may seem that in the next step we need to calculate @xmath182 explicitly .
this complication , however , can be avoided by making use of the the wigner - eckart theorem , @xcite which basically allows for finding matrix elements of an operator with respect to angular momentum eigenstates , @xmath183 the first factor of the rhs is a clebsch - gordan coefficient for adding spins @xmath21 and @xmath184 to get @xmath22 .
this depends only on how the system is oriented with respect ot the @xmath81 axis . on the other hand , the second factor ,
the so - called _ reduced matrix element _ , remains independent of the spatial orientation , as it does not contain the magnetic quantum numbers @xmath185 , @xmath186 or @xmath171 .
thus , we get @xmath187 with @xmath188 and the term @xmath189 regarded in calculations as a _
free parameter to be adjusted for each electrode . specifically , assuming a symmetric coupling between the molecule and two identical electrodes ( @xmath190 ) , the tunnel coupling takes the from @xmath191 , where @xmath192 and @xmath193 denotes the constant , spin - independent density of states in electrodes .
the stationary current @xmath194 flowing through a molecule is calculated as @xmath195 , where @xmath196 ( for @xmath197 ) stands for the current flowing from the @xmath174th electrode to the molecule , @xmath198 where @xmath199 , and @xmath200)^{-1}$ ] is the fermi - dirac function of the @xmath174th electrode , with @xmath87 and @xmath201 standing for temperature and the relevant electrochemical potential , respectively .
the probabilities @xmath202 of finding an smm in a specific state @xmath203 are then derived from a stationary master equation .
@xcite finally , since smms are typically characterized by long spin coherence and spin relaxation times as a result of a weak spin - orbit and hyperfine coupling to the environment,@xcite we neglect relaxation of the spin states due to processes other than due to the electron tunneling . in fig .
[ fig_3](d ) , and also in figs .
[ fig_13][fig_16 ] , we present the current @xmath204 which includes first @xmath62 lowest - in - energy states in the spin multiplet of each charge state .
we use this to show that many excited states in both charge state have to be taken into account in order to describe current correctly .
we define @xmath205 in the following way @xmath206 with denoting summation over states @xmath203 in the charge state @xmath38 that is limited only to first @xmath62 states of lowest energy .
[ [ secsignatures_esignatures - of - the - transverse - anisotropy - parameter - e - without - the - berry - phase - oscillations ] ] [ sec : signatures_e]signatures of the transverse anisotropy parameter @xmath5 without the berry phase oscillations ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in figs .
[ fig_2](a ) and [ fig_3 ] we discuss the initial increase of the current with magnetic field followed by a decrease . the key insight of our calculations using the method described in the previous section ( app .
[ sec : transport ] ) is that the mechanism for this effect is significantly enhanced and modified for @xmath34 giving rise to the characteristic @xmath23 curves shown in fig .
[ fig_2 ] .
since this is at the basis of our scheme of detection , it deserves a further comment . in particular , the relation to the berry phase oscillations which underlay most of the previously used techniques for determining the parameter @xmath5 .
a. upon increase of @xmath5 the minima of the transition - energy curves are shifted to higher field values and the value achieved at the minimum is lowered , cf .
[ fig_3](c ) with fig .
[ fig_13](e)-(h ) . for a fixed temperature
, this leads to a more pronounced maximum conductance attained at a higher field value .
b. generally , the transition energies in fig .
[ fig_3](c ) show sharp features ( i.e. , oscillations below @xmath207 t ) due to berry - phase interference on which several techniques for extracting @xmath5 rely by analyzing the field dependence of the tunnel splitting between two selected states .
@xcite however , the detection of such behavior in the conductance requires very specific low temperature conditions .
this is in contrast to the present experimental conditions where these berry - phase features are averaged out when taking into account multiple accessible states .
this leaves only the large scale , collective variations of the transition energy spectrum caused by @xmath5 which as we have shown suffice for estimation of @xmath5 . in fig .
[ fig_3](d ) we illustrate the importance of taking into account many excited states for both charge states to describe current correctly .
c. finally , fig .
[ fig_2](c ) shows the relative cp amplitude for increasing @xmath26 .
a qualitative distinction from the @xmath208 limit is the appearance of an additional shoulder close to @xmath209 t. it is tempting to see such a shoulder in the sample a curve of fig .
[ fig_2](a ) , although the sample b curve exhibits features of similar size where it should theoretically be smooth . in summary ,
the calculations certainly show that a sizeable @xmath5 term leads to fingerprints in the linear conductance as clear as those for the @xmath1 term , even for relatively high temperatures . and @xmath27 ) :
analogous to figs .
[ fig_3](b)-(d ) with each column corresponding now to a different value of @xmath26 : ( a)-(d ) occupation probabilities for several lowest - in - energy states in the spin multiplets for @xmath15 and @xmath16 at @xmath11 k ; ( e)-(h ) transition energies @xmath59 relevant for the set processes at the coulomb resonance ( i.e. , @xmath61 is restored for each @xmath18 by tuning @xmath13 ) for @xmath60 .
different colors of lines are used to distinguish groups of transitions with respect to possible combinations of indices @xmath56 and @xmath55 ( see the discussion regarding figs . [
fig_3](c ) and [ fig_4 ] ) ; ( i)-(l ) energies @xmath73 for @xmath66 at the coulomb resonance
observe that the curves for @xmath58 overlap ; ( m)-(p ) dependence of the current on the number of spin - multiplet states @xmath62 included from each charge state .
the left / right most column represents the case of absent / significant transverse magnetic anisotropy .
importantly , each column shows a detailed analysis of selected conductance curves from fig .
[ fig_2](c ) .
we note that transition - energy lines in ( e)-(h ) can be easily identified with the use of fig .
[ fig_4](a ) .
it can be seen that increasing @xmath26 results in shifting the minima of the transition - energy curves in ( e)-(h ) towards higher values of the field .
such a behavior , in turn , affects the occupation probabilities ( a)-(d ) , so that the probability of finding an smm either in the ground ( @xmath58 ) or first excited ( @xmath210 ) state for both charge states @xmath15 and @xmath16 remain equal for a larger magnetic - field range ( compare the outermost columns ) . recall that the position of the coulomb peak is fixed mostly by @xmath1 , see fig .
[ fig_8 ] .
, scaledwidth=85.0% ] and @xmath29 ) : generally , this figure is analogous to fig .
[ fig_13 ] except that now the external magnetic field is rotated to align with the molecule s intermediate ( @xmath85 ) axis . to begin with ,
we note that the results shown in the leftmost column ( i.e. for @xmath28 ) are identical to those in the leftmost column of fig .
[ fig_13 ] , which is the manifestation of the molecule s rotational symmetry about the easy ( @xmath81 ) axis in the absence of transverse component of magnetic anisotropy . unlike for the case of @xmath27 ,
the consequence of the increase of @xmath26 is the displacement of the transition - energy curves minima ( e)-(h ) towards smaller values of the field .
interestingly enough , in the situation under discussion one thus observes a more abrupt decrease of the current [ see dashed lines in ( m)-(p ) ] for larger @xmath26 occurring at smaller values of @xmath18 .
, scaledwidth=85.0% ] ) for @xmath53 : analogous to figs .
3(b)-(e ) of the main article with each column corresponding now to a different value of @xmath10 .
note that the case of @xmath27 is presented in figs .
3(b)-(e ) of the main article .
furthermore , here each column shows a detailed analysis of selected conductance curves from fig .
2(d ) of the main article .
, scaledwidth=85.0% ] ): this figure serves to illustrate the fact that even if the transverse magnetic anisotropy is absent , by making the uniaxial magnetic anisotropy parameter @xmath1 smaller ( keeping a fixed temperature ) one can eventually also produce a maximum as for @xmath34 .
however , this maximum occurs at a completely different ( smaller ) value of magnetic field .
moreover , the shape of @xmath32 remains invariant under rotation of the field in the hard plane , this is when the angle @xmath10 is varied .
none of these are the case in the experiment under discussion .
( a)-(b ) dependence of @xmath32 on the value of the uniaxial magnetic anisotropy parameter @xmath211 ( and @xmath212 ) for an external magnetic field applied along the molecule s hard axis ( @xmath25 and @xmath27 ) .
a detailed analysis of selected curves from ( a)-(b ) is carried out in ( c)-(s ) , with each column corresponding to the indicated value of @xmath1 . , scaledwidth=85.0% ] 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 _ _ , , vol .
( , , ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) | by means of electronic transport , we study the transverse magnetic anisotropy of an individual fe@xmath0 single - molecule magnet ( smm ) embedded in a three - terminal junction .
in particular , we determine in situ the transverse anisotropy of the molecule from the pronounced intensity modulations of the linear conductance , which are observed as a function of applied magnetic field .
the proposed technique works at temperatures exceeding the energy scale of the tunnel splittings of the smm .
we deduce that the transverse anisotropy for a single fe@xmath0 molecule captured in a junction is substantially larger than the bulk value . |
it is known that all massive galaxies have a central supermassive black hole ( smbh ) ( e.g. , * ? ? ?
yet it is unknown why only a small fraction of these smbhs are actively accreting as active galactic nuclei ( agn ) .
several theories attempt to explain the fueling mechanism of agn : major mergers ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , minor mergers and gravitational instabilities within the host galaxy ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , nuclear starbursts ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , supernova explosions within nuclear starbursts ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and collisions of warm halo clouds with the nuclear region @xcite .
given the large range of observed agn properties , it is likely that different fueling mechanisms come into play for different agn populations .
current observations suggest that powerful quasars are triggered by major mergers but moderate luminosity agn are more likely to be triggered by stochastic fueling incidents @xcite .
however , it is not yet clear which fueling mechanisms are dominant in which portions of the agn population . the unified model of agn @xcite explains the different observed levels of agn obscuration as a simple geometric effect .
this model assumes that the central engines of all agn are identical and the level of obscuration is dependent on the line of sight between the observer and the central engine .
thus type 2 agn are obscured by column densities @xmath4 @xmath5 10@xmath6 @xmath7 , because the observer is looking through the dusty torus , in contrast to type 1 agn which are unobscured ( @xmath4 @xmath2 10@xmath6 @xmath7 ) because the observer is looking down the throat of the dusty torus . since the central engines are identical , it is expected that , on average , the host galaxies of various spectral types of agn will be similar . however , different black hole fueling mechanisms are likely to lead to different relationships between the various spectral types of agn . in the major merger agn fueling paradigm , the agn is triggered while deeply embedded in gas and dust and thus when the agn first turns on , it is highly obscured .
eventually the radiation pressure pushes away the remaining gas and dust revealing an unobscured quasar ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in the major merger paradigm , type 1 and type 2 agn have host galaxies which are in different evolutionary stages .
conversely , in the nuclear starburst fueling paradigm , type 1 and type 2 agn hosts can be similar since on average the unified model can hold @xcite .
this shows that tests of the unified model can be used to explore which processes are viable options for agn fueling and in which sections of the agn population different mechanisms are relevant . classifying the host galaxies of compton thick ( ct ) agn , agn with @xmath4 @xmath5 10@xmath8 @xmath7 , is of special interest .
it is expected that a large fraction of agn with ct levels of obscuration are in a young evolutionary stage characterized by rapid black hole and host bulge growth @xcite .
thus , understanding the host properties of ct agn may offer special insight into the agn triggering process and the formation of massive galaxies .
large samples of ct agn are difficult to identify because , by definition , ct agn suffer from extreme levels of obscuration ( e.g. , * ? ? ?
. therefore , previous studies of agn host galaxies have not been able to study the stellar populations of ct agn .
galaxy optical colors are often described in terms of the red sequence and the blue cloud , where objects located on the red sequence are characterized by massive older stellar populations and objects on the blue cloud are less massive and have young stellar populations and current star formation . between the red sequence and the blue cloud
is a lightly populated region referred to as the green valley .
some studies find that agn feedback is connected to the shut down of host star formation and thus certain types of agn hosts preferentially reside in the green valley ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus agn activity is potentially a stage in galaxy evolution where feedback from the agn could shut down star formation , causing the host galaxy to age from the blue cloud , across the green valley , and onto the red sequence @xcite .
however , other studies find that the colors of an agn and its host are more closely related to the amount of available obscuring material rather than the evolutionary stage of the host stellar population @xcite . indeed ,
@xcite show that agn are intrinsically bimodal in color and that many agn which appear to be on the red sequence are actually located in the blue cloud once dust extinction is taken into account .
thus further investigation is necessary to establish the nature of agn host galaxy stellar populations and the role of agn feedback in regulating star formation .
this point is especially salient as several possible agn fueling mechanisms include either causal or concurrent star formation .
some studies even show that agn obscuration may in part be due to nuclear starburst disks ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in order to understand how host galaxy processes affect the central smbh and the role of the central smbh in host galaxy evolution , the nature of stellar populations of agn hosts must be well understood .
as agn hosts exhibit size - able object - to - object variability , it is difficult to elucidate trends in the agn population by fitting individual objects .
thus , in order to study larger trends in agn
host galaxy interactions , the average stellar properties of a large ensemble of agn hosts is investigated here . to this end
the stellar synthesis models of @xcite are used to explore the stellar populations of agn hosts . as in previous studies @xcite ,
the agn seds are calculated using the photoionization code cloudy @xcite .
the cosmic x - ray background ( cxrb ) synthesis modeling framework is used to characterize the agn population .
the model seds are then used to move the cxrb framework into other wavelength regions .
this allows the space density and evolution of agn host galaxies to be determined by the most comprehensive census of agn activity . by comparing these models against various observations in the optical through mid - infrared ( mid - ir ) spectral ranges ,
constraints are placed on the average stellar populations of type 1 , compton thin type 2 ( here referred to simply as type 2 ) , and ct agn at various redshifts . in section [ sect : model ] the agn spectral model and stellar population model are described . section [ sect : results ] presents the results of the model . in sections
[ sect : disc ] and [ sect : sum ] the results are discussed and summarized .
ab magnitudes are used , unless otherwise stated , and @xmath9 = 0.7 , @xmath10 = @xmath11 = 0.7 is assumed as necessary .
in order to study the average properties of agn host galaxy stellar populations , a model is created which combines knowledge of agn space density and evolution from deep x - ray surveys and cxrb population synthesis models with stellar population studies .
a model sed is computed which includes the emission from the agn , the host galaxy stellar population , and ongoing dust enshrouded star formation .
this agn and host sed covers the wide spectral range of hard x - ray to mid - ir .
the calculation of the agn model sed is discussed in section [ sub : agn ] and the details of the host stellar population model sed calculation are discussed in section [ sub : bc03 ] . to account for dust enshrouded star formation ,
the templates of @xcite are used , which extend from 5 @xmath0 m to 30 cm , and are determined by averaging empirical seds of local purely star forming galaxies . while dust enshrouded star formation may contribute to the galaxy sed at wavelengths shorter than 5 @xmath0 m , generally the near ir agn and host seds are dominated by either emission from the agn or the @xcite portion of the host sed .
examples of the rest frame agn and host seds are shown in figure [ fig : sed ] , where the dot - dashed lines show agn seds , the dashed lines show stellar population seds , and the dotted lines show the @xcite templates . a gaussian stellar mass ( @xmath12 ) distribution was used with @xmath13 = 10@xmath14 m@xmath15 and @xmath16 = 10@xmath17 m@xmath15 .
the average @xmath12 was set at 10@xmath18 m@xmath15 with a standard deviation of 0.4 dex , in agreement with the sample of x - ray selected , @xmath1 @xmath19 1 agn hosts presented by @xcite .
it is assumed that the mass - to - light ratio of the host galaxy is constant within each waveband and thus the host galaxy flux scales linearly with the host @xmath20 .
therefore , the stellar population sed is calculated for @xmath12 = 1 m@xmath15 and then re - normalized to @xmath12 = 10@xmath14 10@xmath17 m@xmath15 before the stellar population and agn seds are combined . creating host seds at several masses in the agn host mass spectrum
allows the consideration of agn and host colors for hosts of various masses . when considering space densities , however , it is important to consider not only the range of agn host masses but also the probability that an agn host be a certain @xmath12 .
the space density calculations presented are therefore weighted sums across the mass spectrum of agn hosts where the weight is determined by the gaussian @xmath12 distribution described above . specifically , when calculating luminosity functions and number counts the weighted sum over the gaussian @xmath12 distribution is used ; however , when considering flux - flux plots , only the range of @xmath20 is used . in agreement with current observational results
, the @xmath12 distribution used does not evolve with redshift @xcite .
a variety of broad band data from the optical through mid - ir is used to constrain the host models .
these model constraints are described in section [ sub : constrain ] , and include mid - ir number counts , the j band agn and host space density as a function of redshift and the absolute j band magnitude , @xmath21 , and the type 1 agn and host b band luminosity function .
we begin by assuming that the unified model ( e.g. , * ? ? ?
* ) holds and therefore , since all agn are essentially the same system viewed along different lines of sight , there is only one class of agn host galaxies .
it is known that the cxrb encodes the accretion history of smbhs and as such is a powerful tool for scientific inquiry into accretion processes @xcite .
the cxrb synthesis model framework can be used in wavelength regions outside of the x - ray spectral region by modeling the seds of agn and their host galaxies .
thus the advances in cxrb synthesis modeling ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) can be used to further the understanding of agn at all wavelengths . particularly , cxrb synthesis models provide constraints on the distribution of @xmath4 , the fraction of type 2 agn , @xmath22 , and the fraction of ct agn , @xmath23 . here a simple @xmath4 distribution is assumed where type 1 agn are evenly distributed among @xmath24 = 20.0 , 20.5 , 21.0 , 21.5 , type 2 agn are evenly distributed among @xmath24 = 22.0 , 22.5 , 23.0 , 23.5 , and ct agn are evenly distributed among @xmath24 = 24.0 , 24.5 , 25.0 .
@xcite showed that this @xmath4 distribution is consistent with cxrb observations and provides results which are negligibly different from the observed @xmath4 distribution of @xcite @xcite .
the shape of the cxrb combined with deep x - ray surveys has also been used to constrain @xmath22 ( e.g. , * ? ? ?
at a given @xmath25 and redshift , @xmath22 @xmath26 @xmath27 where @xmath28 = 0.4 @xcite and @xmath29 = 4.7
. the normalization of @xmath22 is set such that at @xmath1 = 0 and @xmath30 = 41.5 , the type 2 to type 1 ratio is 4:1 ( see section 2.2 of * ? ? ?
both the original and the composite model of ct agn evolution of @xcite are considered .
the original , non - evolving model assumes that ct agn are a simple extension of the compton thin type 2 agn population and that , when assuming the @xcite hard x - ray luminosity function , @xmath1944@xmath31 of all obscured agn are ct . in this model
there is a population of ct agn which are nearly as common as compton thin type 2 agn and which are in all ways similar to compton thin type 2 agn except for the presence of more obscuration along our line of sight .
contrastingly , the composite model assumes that ct agn are a population of agn distinct from the compton thin type 2 agn population .
simulations show that gas rich galaxy mergers will cause dust and gas to be funneled into the nuclear region of the galaxy , triggering star formation and accretion onto the central smbh ( e.g. , * ? ? ?
due to the large reservoir of material , the smbh will accrete very rapidly and be very highly obscured @xcite .
thus it is expected that ct agn are high eddington ratio sources .
@xcite found that if all ct agn are rapidly accreting sources , than the local space density of ct agn is under - predicted , and thus a population of low eddington ratio ct agn are necessary to explain the observed space density of ct agn .
these low accretion rate ct agn are likely obscured by molecular clouds within the host bulge .
observational evidence of this low luminosity ct agn population has been found in the local universe @xcite .
furthermore , @xcite showed that if ct agn have moderate eddington ratios , the space density of ct agn with @xmath25 @xmath3 10@xmath32 erg s@xmath33 at @xmath1 @xmath5 0 is greatly over - predicted .
therefore , the evolving model of ct agn has an eddington ratio dependent @xmath23 with @xmath1986@xmath31 of all agn which are accreting at @xmath390@xmath31 of their eddington rate being ct and @xmath1960@xmath31 of all agn which are accreting at @xmath21@xmath31 of their eddington rate being ct . in this model
ct agn are a distinct population of agn associated with specific stages of agn evolution and @xmath23 evolves with both @xmath25 , the agn 210 kev luminosity , and redshift .
both the evolving model and the non - evolving model are consistent with the cxrb and the local space density of ct agn with @xmath25 @xmath3 10@xmath32 erg s@xmath33 @xcite . in order to compute the model agn seds , the photoionization code cloudy version c08.00 @xcite
is used , following the same procedure as in section 2 of @xcite .
these seds cover the wavelength range of very hard x - ray through far - ir and include the transmitted agn emission , the diffuse emission emitted along the line of sight by the obscuring material around the agn , and the reflected emission off the inner face of the obscuring cloud .
the inner radius of the obscuring material is assumed to be @xmath1910 pc .
as in @xcite , the obscuring clouds are assigned neutral hydrogen densities , @xmath34 , in agreement with typical molecular clouds ; the compton thin clouds have @xmath34 = 10@xmath35 @xmath36 and the compton thick clouds have @xmath34 = 10@xmath37 @xmath36 . the host galaxy stellar populations
are modeled using galaxev @xcite .
this model calculates stellar population seds over the wavelength range of 91 160 @xmath0 m . using an isochrone synthesis technique , galaxev can evolve stellar populations with three stellar evolution library options .
we assume the stellar evolution models of the padova 1994 library and the stellar spectra of the stelib / basel 3.1 semi - empirical library .
galxev models can be computed for six different metallicities in the range @xmath38 = 0.005z@xmath15 - 2.5z@xmath39 .
both the @xcite and @xcite initial mass functions ( imfs ) are available in the galaxev code . here
we assume solar metallicity @xcite and the @xcite imf @xcite . the sensitivity of the results to these assumptions is assessed in section [ sect :
results ] .
galaxev also allows for a variety of star formation histories , including constant star formation rate , instantaneous bursts , and exponentially declining star formation rate .
an exponentially declining star formation history @xcite with an @xmath40-folding time of 0.5 gyr is used here .
this star formation history allows for a simple parametrization of the average age of the stellar population while not requiring the entire stellar population to be a single age .
galaxev does not include re - radiation of energy absorbed by dust ; however , the majority of this re - emission by cool dust occurs at wavelengths longer than those considered here .
the stellar population seds are computed in four different redshift bins : @xmath1 @xmath2 1 , 1 @xmath2 @xmath1 @xmath2 2 , 2 @xmath2 @xmath1 @xmath2 3 , and 3 @xmath2 @xmath1 @xmath2 5 , where @xmath1 is the redshift .
we assume that the host galaxy has a dominate stellar population ( dsp ) which is slightly younger than the average age of the universe in each redshift bin , and @xmath12 is given by the gaussian distribution described in section [ sub : method ] . a younger stellar population ( ysp ) which accounts for @xmath19 0.1@xmath12 @xcite is also included .
the age of the ysp is selected such that @xmath41(4000 ) @xmath42 1.5 @xcite , using @xmath41(4000 ) = @xmath43 , where @xmath44 is the flux in the 40004100 wavelength range and @xmath45 is the flux in the 38503950 wavelength range @xcite .
for the stellar population models used here , @xmath41(4000 ) @xmath42 1.5 corresponds to a stellar population age of @xmath192 gyr .
the age of the ysp and dsp in each redshift bin is summarized in table [ ages ] . at @xmath1 @xmath3 2 ,
the older and younger stellar populations are the same age and thus at 2 @xmath2 @xmath1 @xmath2 3 the entire stellar population is 2 gyr old and at 3 @xmath2 @xmath1 @xmath2 5 the entire stellar population is 1 gyr old .
extinction due to dust , as described by @xmath46 , is allowed to vary between redshift bins , but not within redshift bins .
while in some objects the agn emission is likely to be extincted by extended dust structures within the host galaxy ( e.g. , * ? ? ?
* ) , this is not true for all objects .
as this study focuses on modeling the average properties of agn and their hosts , a detailed modeling of the complex geometry of obscuring material in individual sources is beyond the scope of this work .
thus , it is assumed here that all of the extinction suffered by the agn emission occurs within a few tens of parsecs of the central engine . the dust which extincts the stellar population
is not allowed to further extinct the agn emission .
a variety of data in a broad range of spectral regions was used to constrain the stellar population models .
the number counts are computed at 3.6 , 5.7 , 8.0 and 24 @xmath0 m and compared to the agn observed by _
spitzer space observatory _ in the goods fields with @xmath47 @xmath5 1 @xmath48 10@xmath49 erg s@xmath33 @xmath7 , where @xmath47 is the 28 kev flux @xcite . this x - ray flux limit is taken into consideration in the number counts calculated here , thus the number counts presented are computed for the same population probed by the observational data points .
these number counts are dominated by type 2 agn and their hosts and therefore are primarily useful to constrain the type 2 agn host population . the number of sources at wavelength @xmath50 per square degree with flux greater than @xmath51 , @xmath52(@xmath53 ) , is found by @xmath54^{1/2 } } d\log l_x dz , \label{eq : counts}\ ] ] where @xmath55 = 3.05 @xmath48 10@xmath56 deg@xmath57 sr@xmath33 , @xmath58 is the hard x - ray luminosity function of @xcite or , when referring to the evolving model , is the evolving eddington ratio space density calculated by @xcite , in units of mpc@xmath59 , @xmath60 is the luminosity distance , and @xmath61 is the 210 kev rest - frame luminosity which corresponds to the observed - frame flux @xmath51 at redshift z. as the number counts are dominated by type 2 agn and their hosts , a separate data set must be used to constrain the type 1 agn host stellar populations .
the most obvious choice is the optical luminosity function for type 1 seyferts and qsos and their hosts . using the x - ray luminosity function , the type 1 agn and host b band luminosity function
, @xmath62 can be calculated using the following equation : @xmath63 this luminosity function is considered in three redshift bins @xmath1 @xmath2 0.4 , 1.0 @xmath2 @xmath1 @xmath2 1.55 , and 1.55@xmath2 @xmath1 @xmath2 2.1 .
further constraints were considered in order to rule out possible degeneracies between stellar population age and extinction due to absorption by dust .
the first of these additional constraints was the agn and host j band space density as a function of redshift in several @xmath21 bins .
first , the agn and host j band luminosity function , @xmath64 , must be calculated .
this is done in the same manner as @xmath62 above . therefore , @xmath65 the normalization constant @xmath66 depends on the agn spectral type being considered . for type 1 agn , @xmath66 = @xmath67 , while for type 2 agn , @xmath66 = @xmath22 , and for ct agn , @xmath66 = @xmath23 . for comparison with the j band agn and host space density measured by @xcite , @xmath64
is then binned into six @xmath21 bins .
additionally , the optical colors were considered in the form of the u - b versus m@xmath68 color - magnitude diagram ( cmd ) .
also , it is well documented that agn and their hosts tend to have x - ray to optical flux ratios of 0@xmath691 , with ct agn hosts generally falling below this ratio . here
the ratio between the soft x - ray flux , @xmath70 , and the r band flux , @xmath71 , @xmath72 = 0@xmath691 is used to ensure that the absolute optical fluxes are in agreement with observations . with all the necessary ingredients in place , we begin by assuming that the unified model holds , and therefore , on average , type 1 , type 2 , and ct agn hosts are identical .
the agn space density , @xmath4 distribution , type 1/type 2 agn ratio , and @xmath23 are set by the cxrb model used .
the age of the stellar populations is assigned such that the dominant stellar population age is slightly less than the mean hubble time in each redshift bin and a younger stellar population is assigned the age corresponding to the average observed @xmath41(4000 ) for agn host galaxies .
the stellar population ages are summarized in table [ ages ] .
the host @xmath12 distribution is gaussian within the observed agn host @xmath12 range .
the only free parameters are the host galaxy dust enshrouded star formation rate and the extinction due to dust .
these two free parameters are used to fit the near and mid - ir agn and host number counts .
the dust enshrouded star formation rate is determined by cycling through the @xcite templates in order of increasing infrared luminosity in steps corresponding to a change in sfr of @xmath191 m@xmath15 yr@xmath33 , until the 24 @xmath0 m number counts are over - estimated at all flux levels .
the template of the highest luminosity to not over - estimate the mid - ir number counts is selected as the best fit template .
the dust extinction in the @xcite models is set using the total effective v band optical depth obscuring young stars , @xmath73 .
seds are computed with @xmath74 = 0.0 , 1.0 , 2.0 , 5.0 , 7.5 , 10 , 15 , 20 , and 25 .
this corresponds to @xmath46 = 0.0 - @xmath191.0 , with the exact values of @xmath46 depending on the age of the stellar population .
the @xmath46 with a reduced @xmath75 , @xmath76 , closest to 1.0 , with respect to the 35 number counts data points at 3.6 @xmath0 m , 5.7 @xmath0 m , 8.0 @xmath0 m , and 24 @xmath0 m , is selected as the best fit model . the resulting model is then compared to the suite of optical and near ir observations described in section [ sub : constrain ] in order to evaluate the appropriateness of the model fit to observations . in this manner the average agn host galaxy properties
are elucidated without the complications which arise from object - to - object variability .
next , the assumption of the unified model is tested by considering the specific constraints for different spectral types .
multiple star formation scenarios and the evolution of @xmath22 are also considered .
in this section , it is assumed that the unified model holds and therefore type 1 , type 2 , and ct agn have , on average , identical host galaxies .
first , the maximum dust enshrouded star formation rate ( sfr ) allowed by the near and mid - ir number counts for x - ray selected agn observed by _
spitzer _ in the goods fields , as reported by @xcite , is determined .
figure [ fig : sed ] shows that the mid - ir emission is due primarily to the agn and dust enshrouded star formation ; therefore , fitting the 24 @xmath0 m number counts can be used to set an upper limit on the average agn host dust enshrouded sfr . in order to not over - estimate the agn and host mid - ir number counts
, the average agn host must have a sfr @xmath77 2 m@xmath15 yr@xmath33 .
this sfr is consistent with the findings of @xcite , who found that an agn host sfr @xmath42 1 m@xmath15 yr@xmath33 provides a good fit to _ spitzer _
s measurement of the agn contribution to mid - ir portion of the cosmic infrared background , and is similar to the sfr of local normal spiral galaxies ( e.g. , * ? ? ?
according to the @xcite relation , this sfr corresponds to an infrared star formation luminosity @xmath78 @xmath42 10@xmath79 l@xmath15 .
next , the dust content of the average agn host galaxy is fixed by fitting the near ir agn and host number counts .
it is found that if the average agn host has @xmath46 @xmath42 0.5 , the near ir number counts are under - predicted . however , if the average agn host has @xmath46 @xmath42 0.4 , the near and mid - ir number counts are over - predicted . as it is known that higher redshift galaxies tend to be dustier than local galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , it is assumed that @xmath1 @xmath2 1 agn hosts contain less dust than @xmath1 @xmath3 1 agn hosts .
if the average agn host at @xmath1 @xmath2 1 has @xmath46 @xmath42 0.25 and the average @xmath1 @xmath3 1 agn host has @xmath46 @xmath42 0.5 , the predicted number counts are in good agreement with the near and mid - ir number counts observations , with @xmath76 = 1.2 . if the average @xmath1 @xmath2 1 agn host has @xmath46 @xmath2 0.25 , the mid - ir number counts are over - predicted .
in contrast , the model constraints are not very sensitive to the age of the older , dominant stellar population , which can be changed by @xmath19 1 gyr with only minor effects .
the age of the younger stellar mass population is constrained to within @xmath190.5 gyr .
this unified model of agn hosts is also in good agreement with the other model constraints .
the j band agn and host space density is in decent agreement with observations , as shown in figure [ fig : jun ] .
the data points in figure [ fig : jun ] show the results of @xcite . at @xmath1 @xmath2 0.75 the ir and x - ray selected agn sample of @xcite
was chosen using a specific selection criterion for optically extended sources . as the models used here do not account for the spatial extent of the agn host galaxies ,
this exact selection criterion can not be replicated .
this gives rise to an obvious discrepancy between the model predictions presented here and the observed j band space density reported by @xcite .
thus at @xmath1 @xmath2 0.75 we over - predict the agn and host space density in magnitude bins of medium brightness and under - predict the fainter magnitude bins .
type 2 agn hosts dominate the j band space density at @xmath21 @xmath3 -26 for @xmath1 @xmath5 0.75 , with type 1 agn hosts dominating the brightest magnitude bins . according to figure 6 of @xcite , type 1 and type 2 agn
tend to be found in the 0.55 @xmath2 u - b @xmath2 1.4 and -23.5 @xmath2 @xmath80 @xmath2 -19 region of the cmd .
this region of the cmd is well populated by the model adopted here as many combinations of agn @xmath25 and host galaxy @xmath12 fit this criteria .
furthermore , the agn and hosts fit the expected @xmath81 ratio , as shown by the green triangles in figure [ fig : fxtororig ] .
thus the unified model of agn hosts explains the average trends of agn and host observations in the optical through mid - ir spectral regions . in order to understand
how well constrained these findings are , the same procedure is followed using a supersolar metallicity , @xmath38 = 2.5@xmath82 , a subsolar metallicity , @xmath38 = 0.2@xmath82 , and the @xcite imf . in all of these scenarios
, it is found that the stellar ages described in section [ sub : bc03 ] are able to fit the model constraints well ; however different @xmath46 values are necessary .
the columns labeled `` unified / type 2 agn host '' in table [ models ] summarize the @xmath46 values for the different models considered .
as the @xcite templates are independent of the galaxev stellar population models , the average agn host dust enshrouded sfr is not affected by changing the metallicity or imf of the host galaxy stellar population .
overall , the near and mid - ir agn and host number counts , as well as the j band space density , optical colors , and x - ray to optical ratio of agn and their hosts are in good agreement with observations if type 1 , type 2 , and ct agn have , on average , similar host galaxies . also , the near and mid - ir number counts require agn hosts at @xmath1 @xmath3 1 be dustier than agn hosts at @xmath1 @xmath2 1 .
we now investigate if this finding holds when spectral type specific model constraints are considered . in order to study the host galaxies of type 1 agn , we begin by calculating the type 1 agn contribution to the type 1 agn and host b band luminosity function .
the predicted agn only b band luminosity function is shown as the dot - dashed lines in figure [ fig : bband ] , where black lines and data refer to @xmath1 @xmath2 0.4 , blue lines and data refer to 1.0 @xmath2 @xmath1 @xmath2 1.55 , and red lines and data refer to 1.55 @xmath2 @xmath1 @xmath2 2.1 .
the @xmath76 takes into account the 50 b band type 1 agn and host b band luminosity function data points shown in figure [ fig : bband ] .
when only the agn contribution to the b band luminosity function is considered , @xmath76 = 14 .
it is clear that except for at the very brightest magnitudes , a contribution from the host galaxy is necessary in order to fit the observed type 1 agn and host b band luminosity function . if the same stellar population described in section [ sub : unified ] is used to calculate the type 1 agn and host b band luminosity function , the predicted @xmath1 @xmath3 1 luminosity functions are not in agreement with observations .
the unified model gives a @xmath76 = 13 fit to the b band luminosity function .
this is shown by the dashed lines in figure [ fig : bband ] . for the @xmath1 @xmath2 0.4 type 1 agn and hosts
, the lower limit @xmath46 of the unified agn host model b band luminosity function ( solid black line in figure [ fig : bband ] ) is in decent agreement with observations . as the type 2 agn hosts are constrained by the ir agn and host number counts as described in section [ sub : unified ] , this suggests that at @xmath1 @xmath2 1 , agn hosts can be described by the unified agn host model , but at @xmath1 @xmath3 1 the unified agn host model can not properly predict the type 1 agn and host b band luminosity function .
thus at @xmath1 @xmath3 1 , type 1 and type 2 agn must be hosted by two distinct galaxy populations .
the lowest @xmath76 with respect to the b band luminosity function is found when type 1 agn hosts have @xmath46 @xmath42 0.05 at @xmath1 @xmath3 1 . for this model @xmath76 = 4.3 , which is not formally a good fit to the data points , but is dominated by the over - prediction of the two faintest @xcite data points at both 1.0 @xmath2 @xmath1 @xmath2 1.55 and 1.55 @xmath2 @xmath1 @xmath2 2.1 . if these points are not included @xmath76 = 2.5 . as these data points refer to observations of faint agn at high redshift ,
it is likely that the observational sample is not complete at these fluxes .
thus , at @xmath1 @xmath2 1 type 1 and type 2 agn hosts are similarly dusty . however , at @xmath1 @xmath3 1 , type 1 agn hosts are less dusty than type 2 agn hosts . also , on average , it appears that type 1 and type 2 agn hosts have similarly aged stellar populations , at least at @xmath1 @xmath2 2 , where the model can be well constrained .
the type 1 agn host stellar population described above is in good agreement with the other model constraints considered .
figure [ fig : ctsorig ] shows the number counts for type 1 agn and their hosts as dotted lines .
this model has a @xmath76 = 1.3 with respect to the observed number counts .
for the unified model , a @xmath76 = 1.2 was found with respect to the ir number counts ; thus reducing the dust extinction in type 1 agn hosts at @xmath1 @xmath3 1 has little effect on the total ir number counts .
type 1 agn and hosts make a considerable contribution to the shorter wavelength number counts , but at mid - ir wavelengths they make a minimal contribution .
again shown as dotted lines , figure [ fig : jorig ] demonstrates that type 1 agn and hosts dominate the j band space density for @xmath21 @xmath2 -26 and make a significant contribution in the -26 @xmath2 @xmath21 @xmath2 -24 magnitude bin .
however , at fainter absolute magnitudes , the type 1 agn hosts make only a nominal contribution to the j band agn and host space density .
the region of the cmd where type 1 agn and their hosts tend to be found , as shown by the upside down triangles in figure 6a of @xcite , is -23.5 @xmath2 @xmath80 @xmath2 -19 and 0.55 @xmath2 u - b @xmath2 1.4 .
this region of the cmd is well populated by a variety of agn @xmath25 and host @xmath20 model combinations .
the majority of type 1 agn and host model seds used here are in agreement with the @xmath81 ratio , as exhibited by the blue circles in figure [ fig : fxtororig ] . at fainter soft x - ray fluxes
it appears that the type 1 agn and hosts fall below the expected ratio .
these sources are located at @xmath1 @xmath3 2.0 where the observed r band flux measures rest frame uv emission . due to the observed frame r band filter redshifting out of optical wavelengths , it is unclear whether the @xmath81 = 0 @xmath69 1 relation should hold at high redshift .
these high redshift type 1 agn and hosts can be brought into agreement with @xmath81 = 0 @xmath69 1 by increasing the extinction due to dust in the host stellar population or by increasing the age of the host stellar population .
thus if the locally observed @xmath81 ratio holds at high redshift , this suggests that at the peak of quasar activity , type 1 agn hosts either had older stellar populations than type 2 agn hosts at the same redshift or were dustier than 1.0 @xmath2 @xmath1 @xmath2 2.0 type 1 agn hosts .
if the stellar populations of type 1 agn hosts at @xmath1 @xmath3 2 are older than the stellar populations of type 2 agn hosts at the same epoch , this implies an evolutionary scenario where recently triggered agn are obscured and then blow out the obscuring gas and dust in order to reveal an unobscured agn and an older host .
it is also possible that , since the fraction of gas rich galaxies at @xmath1 @xmath3 2 is larger than at 1 @xmath2 @xmath1 @xmath2 2 @xcite , type 1 agn hosts located at @xmath1 @xmath3 2 contain more dust than type 1 agn hosts located at 1 @xmath2 @xmath1 @xmath2 2 .
both of these possibilities are consistent with quasars being triggered by major mergers , which is expected to be the dominate quasar fueling mechanism at 2 @xmath2 @xmath1 @xmath2 3 .
when considering the subsolar metallicity , supersolar metallicity and @xcite imf models , the values for @xmath46 which supply the best fit for the type 1 agn and host b band luminosity function change , but the general trend that type 1 and type 2 agn host stellar populations are similar at @xmath1 @xmath2 1 and different at @xmath1 @xmath3 1 remains .
the @xmath46 values for type 1 agn hosts in the different models are shown in the columns labeled `` type 1 agn host '' in table [ models ] . in all models
the type 1 agn hosts are similar in dust content to the type 2 agn at @xmath1 @xmath2 1 and the type 1 agn hosts are less dusty than the type 2 agn hosts at @xmath1 @xmath3 1 .
thus , regardless of the metallicity or imf , type 1 and type 2 agn hosts are similar at @xmath1 @xmath2 1 and intrinsically different at @xmath1 @xmath3 1 .
the emission due to obscured star formation is taken into account using the @xcite star formation templates .
the mid - ir agn and host number counts suggest that the average agn host has a dust obscured sfr @xmath42 2 m@xmath15 yr@xmath33 , the same sfr as normal local spiral galaxies ( e.g. , * ? ? ?
however , recent studies suggest that a larger , but still modest , agn host sfr is expected ( e.g. , * ? ? ?
* ; * ? ? ?
furthermore , studies also find that the agn host sfr tends to increase with redshift @xcite , agn luminosity @xcite , or both @xcite .
as several agn fueling mechanisms , such as the starburst disk model of @xcite , require processes related to star formation , it is important to consider if it is possible for agn hosts to have average sfr @xmath5 2 m@xmath15 yr@xmath33 .
thus several sfr evolutions are considered here .
the @xcite sfr evolution depends on both agn @xmath25 and redshift , finding @xmath83 when the @xcite sfr evolution is used , the average @xmath1 @xmath2 1 sfr must still be @xmath77 2 m@xmath15 yr@xmath33 in order to not over - predict the mid - ir number counts .
however samples of agn hosts with average sfrs an order of magnitude higher than this ( @xmath19 18 - 41 m@xmath15 yr@xmath33 ) have been observed @xcite . in order to explain these observations ,
a population of enhanced star formation sources are considered . the normalization factor for the proportionality in equation [ eq : sfr ]
is set such that the average sfr for agn hosts at @xmath1 @xmath2 1 is @xmath1920 m@xmath15 yr@xmath33 . in order to not over - estimate the faint end of the 24 @xmath0 m number counts , enhanced star formation sources can account for at most @xmath1915@xmath31 of the agn population .
similar fractions of enhanced star formation sources are found using other sfr evolutions .
@xcite used _ herschel space observatory _ observations of type 1 sdss selected quasars to study the evolution of sfr in quasar hosts , finding that agn host sfr displays strong luminosity dependent evolution with redshift .
this leads to a sfr evolution of the form @xmath84 where @xmath85 @xmath42 -1.9@xmath86 - 42 , where @xmath86 is the absolute i band magnitude of the quasar . if the average @xmath1 @xmath2 1 sfr is set at @xmath1920 m@xmath15 yr@xmath33 , enhanced star formation sources can at most be @xmath195@xmath31 of the agn population .
if agn host sfr does not evolve with luminosity , and instead only evolves with redshift , then the average agn host sfr can be considerably higher .
if the average agn host sfr redshift evolution found by @xcite for moderate luminosity agn , sfr @xmath26 @xmath87 , is applied to all agn hosts , an average @xmath1 @xmath2 1 sfr @xmath42 16 m@xmath15 yr@xmath33 for both type 1 and type 2 agn is in good agreement with the mid - ir agn and host number counts , with @xmath76 = 1.3 .
this average sfr corresponds to sfr(@xmath1=0.0 ) = 0.5 m@xmath15 yr@xmath33 , which is in excellent agreement with @xcite who , using a sample of local type 1 sdss quasars , found that the average local agn host sfr @xmath42 0.5 m@xmath15 yr@xmath33 .
furthermore , in the redshift only sfr evolution , for an average @xmath1 @xmath2 1 sfr @xmath42 20 m@xmath15 yr@xmath33 , @xmath1980@xmath31 of agn hosts can be enhanced star formation sources .
thus , if the agn host sfr evolution is dependent on luminosity , either through luminosity evolution or luminosity dependent redshift evolution , enhanced star formation sources are @xmath195 - 15@xmath31 of the agn population , but if agn host sfr only evolves with redshift , the average agn host sfr can be up to @xmath1916 m@xmath15 yr@xmath33 .
this finding is fully consistent with the 15 @xmath0 m agn luminosity function at @xmath1 @xmath19 0.7 observed by @xcite .
furthermore , @xcite find that for moderate luminosity agn , the sfr does not evolve with @xmath25 and that only at the highest quasar luminosities does the agn host sfr seem to depend on the agn luminosity .
therefore , the fraction of enhanced star formation sources found here is a lower limit to the true fraction of enhanced star formation sources .
observational and theoretical evidence suggests that ct levels of obscuration of an agn may be due to an evolutionary stage where the smbh and host bulge are both in a phase of rapid growth ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
according to galaxy merger simulations , gas rich mergers will ignite a burst of star formation and rapid black hole growth ( e.g. * ? ? ? * ) . in this scenario
, it is expected that ct agn hosts would be a subset of the ultraluminous infrared galaxy ( ulirg ) population , characterized by @xmath78 @xmath3 10@xmath17 l@xmath15 , where @xmath78 is the 81000 @xmath0 m luminosity . in order to test this scenario , we attempt to model the ct agn hosts with stellar populations similar to those found in ulirgs hosting agns .
@xcite find that ulirgs hosting agns have an average stellar population age of @xmath190.3 gyr which dominates the stellar mass of the galaxy with an average @xmath20 @xmath42 10@xmath88 m@xmath15 and standard deviation of @xmath190.35 dex and an average @xmath46 @xmath42 0.6 .
this scenario is tested using both the non - evolving and evolving models described in section [ sub : xrb ] . due to the observational challenges of identifying ct agn , there is no data set specific to the optical or near infrared properties of ct agn hosts .
thus the near and mid - ir agn number counts measured from an x - ray flux limited sample of agn , with @xmath47 @xmath5 1 @xmath48 10@xmath49 erg s@xmath33 @xmath7 @xcite , are used to place limits on the average ct agn host galaxy .
the x - ray flux limit of the _ spitzer _ goods agn sample is taken into account in the number counts calculation .
thus , despite the fact that a large fraction of ct agn are missed by deep x - ray surveys @xcite , the dot - dashed lines in figures [ fig : ctsorig ] and [ fig : ctscomp ] show the ct agn which would have been selected by the @xcite selection criteria . the j band space density , x - ray to optical flux ratio , and optical colors are also used to ensure the model ct agn hosts are consistent with observations of the agn population as a whole .
the non - evolving model of @xcite assumes that ct agn evolve like less obscured type 2 agn . in order to match the peak of the cxrb at @xmath1930 kev
, it is required that @xmath1944@xmath31 of obscured agn are ct .
when the ct agn host stellar populations are modeled in agreement with the findings of @xcite , the ir number counts are greatly over - predicted .
this suggests that on average , ct agn in the non - evolving model can not be hosted by ulirgs .
in fact , when using the non - evolving model , the average ct agn host can not even be a luminous infrared galaxy ( lirg ) , which is characterized by @xmath78 @xmath3 10@xmath89 l@xmath15 . in order for the near and mid - ir number counts
to not be over - predicted at the faint end , the average non - evolving model ct agn host must have @xmath78 @xmath77 10@xmath90 l@xmath15 , corresponding to a star formation rate of @xmath77 10 m@xmath15 yr@xmath33 , according to the @xcite relation .
the youngest the non - evolving model ct agn host average stellar population can be , and not over - predict the mid - ir number counts , is @xmath19 1 gyr old which requires @xmath46 @xmath42 1.0 .
the contribution of the non - evolving model ct agn and their hosts to the near and mid - ir number counts is shown as the dot - dashed line in figure [ fig : ctsorig ] . in the non - evolving model ,
ct agn dominate the low redshift , faint @xmath91 region of the j band space density .
figure [ fig : jorig ] shows the ct agn and host contribution to the j band space density as the dot - dashed lines . as expected by the higher dust content , ct agn hosts occupy a region of the cmd which is on average slightly redder and fainter than the type 1 and type 2 agn hosts .
the ct agn tend to lie below the @xmath81 ratio , as the majority of the soft x - ray flux of ct sources is absorbed by the ct obscuring material .
figure [ fig : fxtororig ] shows the @xmath81 ratio of ct agn and their hosts as red squares .
changing the metallicity or imf used for the host galaxy stellar population does not change these results .
when the enhanced star formation sources are considered , the ct agn hosts must be similar to the type 2 agn hosts , regardless of host galaxy stellar population metallicity and imf .
similarly , when the redshift evolution of agn host sfr is included , the ct agn hosts must be similar to the type 2 agn hosts .
thus , if @xmath23 does not evolve with agn eddington ratio , then ct agn and their hosts are expected to be a simple extension of the less obscured type 2 agn population .
this would require that agn triggered by mergers be a small minority of the quasar population .
the evolving model allows ct agn to evolve independently of the less obscured type 2 agn .
instead , @xmath23 is assumed to be eddington ratio dependent . in order to fit the peak of the cxrb at @xmath1930 kev , the local ct agn space density , and the @xmath1 @xmath3 1 ir ct agn space density
, it is found that @xmath1986@xmath31 of agn with eddington ratios @xmath30.9 are ct , @xmath1960@xmath31 of agn with eddington ratios @xmath2 0.01 are ct , and @xmath190@xmath31 of agn with intermediate eddington ratios are ct . as agn with weaker accretion rates are observed to have older stellar populations @xcite , the low eddington ratio agn are assumed to have the same mass distribution as the type 1 and type 2 agn but an older stellar population and very little star formation . the age of the low eddington ratio ct agn host stellar populations are summarized in table [ ages ] .
if the low eddington ratio ct agn hosts have the same @xmath46 as the type 1 and type 2 agn hosts , the near and mid - ir number counts are over - predicted . increasing the dust in the low eddington ratio ct agn hosts to @xmath46 @xmath42 1.0 brings the model predictions into agreement with the faint end of the ir number counts , as shown by the blue dot - dashed lines in figure [ fig : ctscomp ] .
the larger amount of dust in the ct agn hosts as compared to the type 2 agn hosts makes sense as only galaxies containing a large amount of dust will be able to host an agn with ct levels of obscuration . in the evolving model , the high eddington ratio ct agn hosts should be galaxies in a phase of rapid star formation and black hole growth .
therefore , the high eddington ratio ct agn hosts should be lirgs or ulirgs .
indeed we find that if the stellar populations of the high eddington ratio ct agn hosts have @xmath78 = 10@xmath17 l@xmath15 from star formation , @xmath46 @xmath42 0.6 , with a stellar population of age @xmath190.3 gyr which dominates @xmath20 , in agreement with the average stellar population of ulirgs which host an agn @xcite , the near and mid - ir number counts predictions are in agreement with deep observations .
the high eddington ratio ct agn and host number counts are shown as the red dot - dashed lines in figure [ fig : ctscomp ] .
the over - prediction at the bright end of the number counts is due to the incompleteness of the survey at bright fluxes .
as the goods fields cover only 0.1 deg@xmath92 @xcite , goods misses bright , rare objects . if the stellar population is assumed to be younger than @xmath190.3 gyr ,
the number counts are over - predicted .
similarly if the @xmath46 @xmath2 0.6 or if the average @xmath20 @xmath3 10@xmath89 m@xmath15 , the number counts are over - predicted .
thus , the stellar population described here is the upper limit for how bright the high eddington ratio ct agn hosts can be without over - predicting the ir number counts .
the ct agn host galaxies are also in good agreement with the other model constraints considered here .
the ct agn host contribution to the j band space density is shown as the dot - dashed lines in figure [ fig : jcomp ] , where the red lines refer to the high eddington ratio ct agn and their hosts and the blue lines refer to the low eddington ratio ct agn and their hosts .
as expected the high eddington ratio sources dominate the ct agn and host contribution in the brighter magnitude bins while the low eddington ratio sources dominate at the fainter magnitude bins .
the optical colors are also in agreement with observations .
the low eddington ratio ct agn hosts are on average redder and fainter than the type 1 and type 2 agn hosts , while the high eddington ratio ct agn hosts are on average as bright or slightly brighter than the type 1 and type 2 agn hosts but on average a little redder than the type 1 and type 2 agn hosts . as with the non - evolving model ,
most of the ct agn hosts lie below the empirical average @xmath81 ratio , as expected by the high levels of soft x - ray absorption fundamental to ct agn . in figure
[ fig : fxtorcomp ] the @xmath93 versus @xmath70 for ct agn and their hosts are shown as the red squares with the low eddington ratio objects shown as filled red squares and the high eddington objects shown as open red squares .
the same result is found when the host galaxy imf or metallicity is varied . as the high eddington ratio ct agn are a small fraction of the overall agn population , the evolving model ct agn hosts are not affected by the consideration of the enhanced star formation sources .
thus the ct agn hosts of the evolving model are in agreement with the agn evolution scenario where major mergers trigger nuclear starbursts and highly obscured agn activity .
the evolution of @xmath22 , the type 2 agn fraction , is important for understanding the agn life cycle and how agn and their host galaxies interact ( e.g. , * ? ? ?
* ) . in the unified model
, @xmath22 is the covering factor of the dusty torus , and thus the evolution of @xmath22 shows a fundamental evolution of the torus parameters @xcite .
several studies suggest that @xmath22 evolves with @xmath25 and possibly also with redshift ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . however , other studies find that it is not necessary for @xmath22 to evolve with @xmath25 ( e.g. , * ? ? ? * ) nor with redshift ( e.g. , * ? ? ?
* ; * ? ? ?
* ) in order to explain observations .
indeed , we find that the number counts and j band space density can be fit with a constant @xmath22 . for @xmath1 @xmath2 0.4 , the type 1 agn and host b band luminosity function
can also be satisfactorily fit with a constant @xmath22 .
however , fitting the type 1 agn and host b band luminosity function at @xmath1 @xmath5 1 requires that @xmath22 is not constant .
if @xmath22 is constant , the best fit @xmath76 @xmath3 7.0 and the type 1 agn and host b band luminosity function predicts a considerably larger population of b band magnitude , @xmath95 @xmath19 23 , 1 @xmath2 @xmath1 @xmath2 2 type 1 agn than observed .
thus , it is found that in order to fit the type 1 agn and host b band luminosity function at @xmath1 @xmath3 1 , @xmath22 must evolve with @xmath25 .
the argument for evolution of @xmath22 with redshift is less conclusive .
figure [ fig : bf2lx ] shows the type 1 agn and host b band luminosity function assuming that @xmath22 does not evolve with redshift , which gives @xmath76 = 6.5 . when @xmath22 does evolve with redshift , the best fit @xmath76 = 4.3 . while the @xmath76 for both the evolving and the non - evolving @xmath22 models do not represent formal good fits to the observed data points , the model in which @xmath22 does evolve with redshift provides a better fit to the data . in the scenario where @xmath22 does not evolve with redshift ,
it is necessary that at higher redshift the type 1 agn host galaxies be dustier than locally .
this is in contrast to the scenario where @xmath22 does evolve with redshift and type 1 agn hosts are less dusty at higher redshift compared to the local population . in order to not over - predict the ir number counts
, the type 2 agn hosts must also be dustier at higher redshift than locally .
assuming that @xmath22 does not evolve with redshift has minimal affect on the j band space density , optical colors , and @xmath81 ratio .
a comparison of figure [ fig : bband ] and figure [ fig : bf2lx ] , suggests that @xmath22 not only evolves with @xmath25 , but also with redshift .
also , the model where @xmath22 evolves with both @xmath25 and redshift provides a better fit to the type 1 agn and host b band luminosity function has measured by @xmath76 .
as neither the evolving nor non - evolving @xmath22 models provide a formally good fit to the observed luminosity function , the possibility that @xmath22 does not evolve with redshift can not be conclusively ruled out ; however , the evolution of the b band type 1 agn and host luminosity function suggests that @xmath22 does evolve with redshift . in order to fit the type 1 agn and host b
band luminosity function @xmath22 must evolve with @xmath25 . using a variety of optical and near and mid - ir data ,
we have constrained the stellar populations of agn host galaxies .
table [ work ] summarizes the average agn host galaxy for different agn spectral types .
it is found that at @xmath1 @xmath2 1 type 1 and type 2 agn hosts are similar , but at @xmath1 @xmath3 1 type 1 agn hosts are less dusty than type 2 agn hosts .
the majority of agn have an average sfr @xmath77 2 m@xmath15 yr@xmath33 , however there is evidence of a population of enhanced star formation sources which account for @xmath5 515@xmath31 of the agn population and has sfr @xmath42 20 m@xmath15 yr@xmath33 .
also , it is found that if ct agn evolve like type 2 agn , then ct agn hosts are similar to type 2 agn hosts and if ct levels of obscuration are indicative of specific evolutionary stages in the agn life cycle , then ct agn hosts are also in specific evolutionary stages .
furthermore , it is shown that @xmath22 evolves with @xmath25 and @xmath22 is likely to evolve with redshift .
for both the evolving and non - evolving models , the ct agn host stellar populations suffer from at least as much dust extinction as the type 2 agn host stellar populations .
observations suggest that at least some of the observed extinction of heavily obscured agn may be due to extended dust structures or molecular clouds within the host galaxy @xcite , so it is expected that the stellar populations of ct agn will be enshrouded in dust .
it is within the limits imposed by the model constraints for the high eddington ratio ct agn hosts to have an average @xmath46 which is in agreement with ulirgs hosting agn @xcite .
the agn evolution scheme , in which mergers trigger large nuclear starbursts and agn activity , claims that the agn activity will initially be very highly obscured while the black hole grows very rapidly .
as the black hole grows , the radiation pressure on surrounding dusty gas will increase until the agn feedback blows out the obscuring material and halts the star formation in the host nuclear region @xcite . in this scheme
ct agn should have young stellar populations and possibly high levels of on going star formation .
the non - evolving model places a lower limit on the stellar population age of 1 gyr .
however , when the enhanced star formation sources are included , the non - evolving model ct agn hosts have similar stellar ages to the type 2 agn hosts , in contrast to the expectations of the agn evolution scheme . for the high eddington ratio ct agn in the evolving model ,
the lower limit on the stellar population age is 0.3 gyr , the average stellar age of ulirgs hosting an agn @xcite .
the low eddington ratio ct agn hosts of the evolving model can be of similar age as the type 1 and type 2 agn , but the near and mid - ir predicted number counts are in better agreement with the observations if these agn hosts have slightly older stellar populations than the average type 1 and type 2 agn hosts .
it has been demonstrated that galaxies hosting agn with lower [ oiii ] luminosities have larger values for @xmath41(4000 ) compared to galaxies hosting agn with high [ oiii ] luminosities @xcite .
thus it is expected that agn with lower eddington ratios are in hosts with older stellar populations , in agreement with the findings of this study .
the stellar populations of ct agn hosts for the evolving model and the non - evolving model are in agreement with the expectations from the agn evolution scheme .
the non - evolving model finds that the average ct agn host has recently ( @xmath51 gyr ago ) undergone a large burst of star formation , but that current star formation rates are more modest .
the evolving model finds that high eddington ratio ct agn hosts have recently ( @xmath50.3 gyr ago ) undergone a large burst of star formation and that current star formation rates may also be elevated .
if the enhanced star formation sources are included , the non - evolving model agn hosts have stellar populations of similar age as the type 2 agn , and thus are in better agreement with the orientation based unified model than with the agn evolution scheme . the evolving model ct agn hosts are consistent with the agn evolution scheme regardless of the inclusion of the enhanced star formation sources . in summary ,
the non - evolving model ct agn hosts are a simple extension of the type 2 agn host population while the evolving model ct agn are consistent with the paradigm where major mergers cause both intense starbursts and agn activity . in the evolving model
, low eddington ratio ct agn hosts will appear as galaxies with old and dusty stellar populations while the high eddington ratio ct agn hosts will appear as ir bright starburst galaxies .
it is found that @xmath195 - 15@xmath31 of agn hosts can be enhanced star formation sources with an average sfr @xmath42 20 m@xmath15 yr@xmath33 , a factor of 10 higher than the majority of agn hosts .
this is expected to be a lower limit of the fraction of enhanced star formation sources as some sources may have such highly embedded star formation that the majority of the reprocessed emission due to star formation is at wavelengths longer than 24 @xmath0 m .
observations at longer wavelengths , such as in the far - ir with the _ herschel _ or the millimeter / sub - mm regime with the atacama large millimeter / submillimeter array ( alma ) , are necessary for uncovering the evolution of star formation rates in agn hosts @xcite . even deep radio observations
are a useful tool in determining the highly embedded star formation rates of agn hosts @xcite .
indeed , by stacking sub - mm observations of x - ray selected agn , @xcite found that the average agn host sfr @xmath42 30 m@xmath15 yr@xmath33 .
furthermore , @xcite found that the agn host sfr evolves strongly with redshift but with evidence of luminosity dependent evolution only for the highest luminosity agn . in the scenario where agn host sfrs are not luminosity dependent ,
it is found that more than half of agn hosts can be enhanced star formation sources .
the fact that the mid - ir agn and host number counts are over - predicted by an average sfr @xmath3 2 m@xmath15 yr@xmath33 , despite observational evidence that there is a population of agn with an average sfr an order of magnitude higher than this upper limit , suggests that there are two populations of agn .
the majority of agn hosts , at least at @xmath1 @xmath2 1 , have sfrs similar to local spiral galaxies and @xmath55 - 15@xmath31 of agn hosts have markedly higher sfrs .
the existence of these two populations of agn hosts can be interpreted in two complimentary ways .
the first interpretation is that the agn hosted by galaxies with enhanced sfrs are being fueled by different mechanisms than the lower star formation objects . at @xmath1 @xmath2 1 , the lower star formation objects are likely dying quasars which were triggered by major mergers while the enhanced star formation objects are likely seyferts which are both fueled and obscured by circumnuclear starburst disks , as investigated by @xcite .
the other interpretation is that all agn are fueled by processes related to nuclear starbursts , whether those starbursts are triggered by mergers or through secular processes , and on average , the nuclear starburst phase overlaps with the active agn phase for @xmath55 - 15@xmath31 of the agn lifetime .
further exploration of these two populations of agn hosts is necessary to determine the processes which trigger and fuel agn activity .
it is well documented that the integrated emission of agn observed in deep x - ray surveys is insufficient to account for the intensity of the cxrb at @xmath1930 kev and that the shape of the cxrb necessitates that the missing population of agn be highly obscured ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
given the uncertainties of the normalization of the cxrb and the agn hard x - ray luminosity functions , different models predict vastly different numbers of missing ct agn ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as these highly obscured agn are missed in deep x - ray surveys @xcite , it is important to consider other methods to identify the elusive ct agn population .
one possibility is that the majority of x - ray bright optically inactive galaxies ( xbongs ) host ct agn @xcite .
xbongs are x - ray sources found in deep surveys which have no optical counterparts with @xmath96 @xmath77 25.5 and make up a substantial portion of deep survey x - ray sources @xcite .
it is thought that these sources are either heavily obscured agn and/or high redshift quasars @xcite .
considering the sources above the dotted horizontal line in figures [ fig : fxtororig ] and [ fig : fxtorcomp ] suggests that a small fraction of optically faint x - ray sources may be ct agn , but the majority of this population is @xmath1 @xmath5 1 type 2 agn .
thus it is unlikely that xbongs host the majority of the missing ct agn population .
another method used to identify ct agn candidates is to search for infrared bright sources which have an infrared excess , usually defined by @xmath97 @xmath5 1000 , where @xmath98 is the 24 @xmath0 m flux ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
some concern has been raised that this method will also pick out lower redshift type 2 agn masquerading as high redshift ct agn @xcite .
we therefore investigate the population of agn selected by the infrared excess criteria .
@xcite suggest that using the criteria @xmath97 @xmath5 1000 and @xmath96 - @xmath99 @xmath3 4.5 selects a distinct class of sources , the majority of which are ct agn . however , according to both the evolving and non - evolving model ct agn and host seds developed here , the @xcite criteria selects low to moderately x - ray bright agn at moderate to high redshift .
the vast majority of the selected agn are obscured , but a significant fraction are still compton thin .
similarly , @xcite find that the @xcite criteria is likely to select dusty star forming templates and low x - ray flux agn .
however , @xcite find no evidence that the selected agn are ct .
a more effective method of identifying ct agn candidates is based on the selection criteria of @xcite : @xmath97 @xmath5 1000 and @xmath98 @xmath3 1.0 mjy .
furthermore , it appears that samples with a lower @xmath98 limit , such as @xmath98 @xmath3 700 @xmath0jy or even @xmath98 @xmath3 550 @xmath0jy , also contain a large fraction of highly obscured agn ( * ? ? ?
* and references therein ) .
fainter samples of infrared bright galaxies are found to be predominately powered by star formation rather than by agn ( e.g. , * ? ? ?
figure [ fig : f24tor]a shows that representatives of all spectral types of agn can be found with @xmath97 @xmath5 1000 and @xmath98 @xmath3 550 @xmath0jy .
these sources will be located at all redshifts . however , the majority of type 1 and type 2 agn are bright in the soft x - ray . for both the evolving and non - evolving models , the vast majority of infrared excess agn with @xmath98 @xmath3 550 @xmath0jy and f@xmath100 @xmath2 10@xmath101 erg @xmath7 s@xmath33 are ct agn , as shown in figure [ fig : f24tor]b .
figure [ fig : f24torbyz]b , which shows the redshift distribution of the ct agn with f@xmath100 @xmath2 10@xmath101 erg @xmath7 s@xmath33 , illustrates that these x - ray faint ct agn are located at all redshifts .
for the evolving model the vast majority of these ct agn are high eddington ratio sources .
this criteria will select a small number of type 2 agn with enhanced star formation , but the majority of sources selected in this manner are indeed ct . if an infrared excess agn has @xmath70
@xmath77 10@xmath49 erg @xmath7 s@xmath33 and @xmath98 @xmath3 550 @xmath0jy , than the agn is ct and located at @xmath1 @xmath77 2 .
this is true for both the evolving and non - evolving models .
an interesting consequence of this is that a size - able population of @xmath1 @xmath3 2 ct agn should have 10@xmath101 @xmath3 @xmath70 @xmath3 10@xmath49 erg s@xmath33 @xmath7 .
according to the evolving model , nearly all of the @xmath70 @xmath77 10@xmath49 erg @xmath7 s@xmath33 sources are high eddington ratio sources and according to the non - evolving model these low x - ray flux sources are nearly all enhanced star formation sources .
it is therefore expected that x - ray stacking of ir bright sources with @xmath97 @xmath5 1000 and low soft x - ray flux will yield a large fraction of the @xmath1 @xmath77 2 high luminosity ct agn population .
this shows that combining observations in multiple spectral regions , such as mid - ir , optical , and x - ray observations , is the most efficient way of identifying ct agn candidates .
identifying and characterizing the elusive ct agn population is a necessary part of understanding the history of accretion and galaxy evolution .
however , due to the observational challenges of studying highly obscured sources , multi - wavelength investigations are necessary to identify and understand the nature of ct agn .
several recent studies suggest that there are two distinct processes which lead to agn activity , secular evolution and merger events ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the latter paradigm ,
quasar activity is activated by galaxy mergers which cause gas and dust to be funneled into the nuclear region ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , whereas moderate luminosity agn are fueled by gravitational instabilities internal to the host galaxy or through minor interactions ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is probable that if different forms of agn activity are caused by contrasting fueling mechanisms , the relationship between various agn spectral types may be different for the high and moderate luminosity populations .
this study finds that the average type 1 and type 2 agn hosts have similar stellar populations and similar levels of dust attenuation at @xmath1 @xmath2 1 . at @xmath1 @xmath3 1 type 1 and type 2 agn hosts
appear to be fundamentally different .
even though the high redshift type 1 and type 2 hosts have similar stellar populations , the type 1 agn and host b band luminosity function requires that type 1 agn hosts have significantly lower levels of dust extinction than what is required for type 2 agn hosts to be in agreement with the observed near and mid - ir number counts . at @xmath1 @xmath3 1 , type 1 agn hosts are intrinsically less dusty than type 2 agn hosts .
this suggests that the orientation based unified model works well for describing the local seyfert population , but may not be appropriate for the high redshift , high luminosity quasar population .
this conclusion is consistent with the findings of several other recent studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . for galaxies which are evolving secularly
, the unified model appears to be an apt description .
however , the violent growth experienced by black holes and their host bulges during major merger events does not appear to fit into the orientation based unified scheme .
it is likely that different agn fueling mechanisms will result in different relationships between agn spectral types . in the merger scenario ,
young agn are highly obscured and old quasars are unobscured ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
different spectral types of agn fueled by secular processes related to nuclear starbursts are expected to be in agreement with the orientation based unified model due to the disky nature of the nuclear starburst which is likely to both fuel and obscure the agn ( e.g. , * ? ? ?
in the evolving model , it is found that mergers can play a strong role in fueling high @xmath25 , @xmath1 @xmath3 1 quasars .
however , in the non - evolving model , when the enhanced star formation sources are considered , it appears that the dominant fueling mechanism is not mergers .
instead , the difference between spectral types may only be that some galaxies have less dust and gas than others at @xmath1 @xmath3 1 . in the non - evolving @xmath23 model
, it appears that secular processes are the dominate agn fueling mechanism at all redshift . for both the evolving and non - evolving model , at @xmath1 @xmath2 1 the processes which lead to agn activity are most likely secular
however , the findings of this study suggest that the dominant agn fueling process changes at @xmath1 @xmath19 1 , since at @xmath1 @xmath2 1 the average type 1 and type 2 agn hosts are more similar than at higher redshift . in order to understand the mechanisms which trigger and fuel agn , it is important for future studies to pay careful attention to which fueling mechanisms are dominant in different subsets of the agn population .
thus , it has been shown that the relationship between different agn spectral types is a helpful tool for understanding the dominant processes responsible for triggering and fueling agn activity .
it is found that at @xmath1 @xmath2 1 the orientation based unified model holds , suggesting that the dominant agn fueling mechanisms at @xmath1 @xmath2 1 are secular processes . at @xmath1 @xmath3 1 ,
the orientation based unified model does not seem to hold .
the evolving ct agn model suggests that at @xmath1 @xmath3 1 mergers are an important agn fueling mechanism ; however , the non - evolving ct agn model suggests that even at @xmath1 @xmath3 1 the dominant agn fueling mechanisms are secular processes and that mergers play only a minor role in fueling agn at all redshifts .
understanding the evolution of @xmath22 is an important step in understanding the nature of agn obscuration and the interplay between an agn and it s host galaxy .
if @xmath22 evolves with agn @xmath25 , this suggests that the obscuring material is close enough to the central engine that dust sublimation @xcite and/or radiation pressure fed winds @xcite affect portions of the obscuring material . if @xmath22 evolves with redshift this would indicate that the evolution of the obscuring material is somehow connected to the evolution of the host galaxy @xcite .
while most model constraints used in this study are not very sensitive to the evolution of @xmath22 , the type 1 agn and host b band luminosity function does prove to be a valuable test of the evolution of @xmath22 .
this study finds that @xmath22 must evolve with agn @xmath25 and that @xmath22 probably evolves with redshift .
if @xmath22 does not evolve with @xmath25 , the predicted 1 @xmath2 @xmath1 @xmath2 2 type 1 agn and host b band luminosity functions are not in agreement with observations . using the host galaxy stellar population age and @xmath46 as free parameters
does not allow for an appropriate fit to the data .
the over - prediction of @xmath95 @xmath19 23 sources at 1 @xmath2 @xmath1 @xmath2 2 is not due to observational bias as @xmath95 @xmath19 23 is significantly brighter than the depth of b band coverage accessible to surveys ( e.g. , b band 5@xmath102 level for cosmos is @xmath95 = 26.7 ; * ? ? ?
@xcite argue that @xmath22 does not evolve with @xmath25 and that apparent evolution of @xmath22 with @xmath25 is due to x - ray observational biases . however , the @xmath1 @xmath3 1 type 1 agn and host b band luminosity function data used here is based on optical selection criteria @xcite .
the @xcite type 1 agn and host b band luminosity function does use a sample defined through x - ray selection , and the @xmath1 @xmath2 0.4 predicted type 1 agn and host b band luminosity function is in reasonable agreement with observation even if @xmath22 does not evolve with @xmath25 .
this suggests that x - ray observational biases do not create an artificial evolution of @xmath22 with @xmath25 .
therefore , at least for @xmath1 @xmath3 1 , @xmath22 must evolve with @xmath25 . while this study can not rule out the possibility that @xmath22 is constant with redshift
, our findings suggest that @xmath22 does evolve mildly with redshift .
while some studies suggest that @xmath22 must evolve with redshift , the evidence is still tentative .
if agn are fueled by different mechanisms , one might assume that the type 1/type 2 ratio will be different for the different mechanisms .
therefore , if different mechanisms dominate during different epochs , the type 1/type 2 ratio should also include some redshift evolution .
it is possible that the slower secular evolution processes of seyfert galaxies do not require evolution of @xmath22 with redshift or that the evolution is very mild , since this form of galaxy evolution is governed by stochastic processes .
however , in the quasar regime , where agn fueling is initiated by major mergers , a stronger evolution of @xmath22 with redshift may be necessary .
indeed , the type 1 agn and host b band luminosity function which is least well fit by a non - evolving @xmath22 is the highest redshift bin considered , which is the redshift bin closest to the peak of quasar activity .
while this study does not conclusively show that @xmath22 evolves with redshift , the findings presented here suggest that @xmath22 does exhibit some evolution with redshift .
it is found that @xmath22 evolves with @xmath25 and @xmath22 is likely to evolve with redshift .
this suggests that the obscuring medium is close enough to the central engine to be affected by dust sublimation @xcite and/or radiation pressure fueled winds @xcite .
also , the type 1 agn and host b band luminosity function is an excellent x - ray independent tool to test the fraction of type 1 agn predicted by x - ray observations .
by applying observational constraints at optical through mid - ir wavelengths to agn host stellar population models , the average stellar properties of agn hosts have been constrained and hard x - ray through mid - ir seds have been developed which include emission from the agn and the host galaxy stellar population and dust enshrouded star formation .
the findings of this study are summarized as follows .
the orientation based unified model provides a good description of the @xmath1 @xmath2 1 seyfert population .
the unified model does not seem to hold at @xmath1 @xmath3 1 , where quasar activity triggered by major mergers becomes more prevalent .
multi - wavelength data is consistent with the paradigm in which ( high eddington ratio ) ct agn have recently undergone intense star formation . in the non - evolving model
the agn activity does not start until the starburst is @xmath51 gyr old and the stellar population is highly obscured by dust . in the evolving model , it is possible that the average high eddington ratio ct agn host galaxy is a lirg or ulirg , with the agn activity starting one average @xmath50.3 gyr after the starburst .
if the enhanced star formation sources are included , the evolving model ct agn hosts are unaffected , but the non - evolving model ct agn hosts are not consistent with the agn evolution scheme and instead are consistent with the orientation based unified scheme .
thus , if ct agn are similar to type 2 agn , then ct agn host galaxies are similar to type 2 agn host galaxies ; however , if ct agn are a distinct population of agn in a special evolutionary stage , then the host galaxies of ct agn are also in a special evolutionary stage . in order to fit the type 1 agn and host b band luminosity function
it is necessary for @xmath22 to evolve with @xmath25 .
however , future work studying the dependence of @xmath22 on redshift is necessary to elucidate the connection between agn and their hosts . understanding the redshift evolution of @xmath22 will offer insight into how agn are fueled and the nature of agn obscuration . while the actual values of @xmath46 depend on the metallicity and imf used in the @xcite host galaxy population synthesis model , the qualitative findings of this study are independent of , and robust against the uncertainties in , the host galaxy metallicity and imf .
aird , j. , et al . , 2010 , , 401 , 2531 alexander , d.m . ,
brandt , w. n. , hornschemeier , a. e. , garmire , g. p. , schneider , d. p. , bauer , f. e. , & griffiths , r. e. , 2001 , , 122 , 2156 alexander , d.m .
, et al . , 2008 , , 687 , 835 antonucci , r. , 1993 , ara&a , 31 , 473 assef , r.j . , et al . , 2011 , , 728 , 56 ballantyne , d.r . ,
everett , j.e . , & murray , n. , 2006a , , 639 , 740 ballantyne , d.r . , shi , y. , rieke , g.h .
, donley , j.l , papovich , c. , & rigby , j.r . , 2006b , , 653 , 1070 ballantyne , d.r . , 2008 , , 685 , 787 ballantyne , d.r . , 2009 , , 698 , 1033 balogh , m.l . ,
morris , s.l . ,
yee , h.k.c . ,
carlberg , r.g .
, & ellingson , e. , 1999 , , 527 , 54 bluck , a.f.l .
, conselice , c.j . ,
almaini , o. , laird , e.s . ,
nandra , k. , & grtzbauch , r. , 2011 , , 410 , 1174 brand , k. , et al . , 2007 , , 663 , 204 brandt , w.n .
, et al . , 2001 , , 122 , 1 brusa , m. , et al . , 2009 , , 507 , 1277 bruzual , g. & charlot , s. , 2003 , , 344 , 1000 cardamone , c.n . ,
urry , c.m . ,
schawinski , k. , treister , e. , brammer , g. , & gawiser , e. , 2010 , , 721 , l38 chabrier , g. , 2003 , , 115 , 763 chen , y .- m . , wang , j .-
m . , yan , c .- s . ,
hu , c. , & zhang , s. , 2009 , , 695 , l130 crenshaw , d.m . ,
kraemer , s.b .
, & gabel , j.r .
, 2003 , , 126 , 1690 croom , s. m. , smith , r. j. , boyle , b. j. , shanks , t. , miller , l. , outram , p. j. , & loaring , n. s. , 2004 , , 349 , 1397 dahlen , t. , mobasher , b. , dickinson , m. , ferguson , h. c. , giavalisco , m. , kretchmer , c. , & ravindranath , s. , 2007 , , 654 , 172 davies , r.i .
, mller snchez , f. , genzel , r. , tacconi , l.j . , hicks , e.k.s . ,
friedrich , s. , & sternberg , a. , 2007 , , 671 , 1388 della ceca , r. , zamorani , g. , maccacaro , t. , setti , g. , & wolter , a. , 1996 , , 465 , 650 donley , j.l . ,
rieke , g.h . ,
alexander , d.m . ,
egami , e. , & prez - gonzlez , p.g . , 2010 , , 719 , 1393 donley , j.l . ,
rieke , g.h .
, prez - gonzlez , p.g . , & barro , g. , 2008 , , 687 , 111 draper , a.r . &
ballantyne d.r . , 2011 , , 729 , 109 draper , a.r .
& ballantyne d.r .
, 2010 , , 715 , l99 draper , a.r . & ballantyne d.r . , 2009 , , 707 , 778 dunne , l. , eales , s.a . , & edmunds , m.g . , 2003 , , 341 , 589 faber , s.m .
, et al . , 2007 , , 665 , 265 fabian , a.c . , 1999 , , 308 , l39 fabian , a.c .
& barcons , x. , 1992 , ara&a , 30 , 429 fabian , a.c . , vasudevan , r.v .
, gandhi , p. , 2008
, , 385 , l43 fabian , a.c . , vasudevan , r.v . , mushotzky , r.f . , winter , l.m . , & reynolds , c.s . , 2009 , , 394 , l89 ferland , g.j . , korista , k.t . ,
verner , d.a . ,
ferguson , j.w .
, kingdon , j.b . , & verner , e.m . , 1998 , pasp , 110 , 761 fiore , f. , et al . , 2008 , , 672 , 94 fiore , f. , et al . , 2009 , , 693 , 447 fu , h. , et al . , 2010 , , 722 , 653 gandhi , p. & fabian , a.c . , 2003 , , 339 , 1095 georgakakis , a. , clements , d.l . ,
bendo , g. , rowan - robinson , m. , nandra , k. , & brotherton , m.s .
, 2009 , , 394 , 533 georgakakis , a. , rowan - robinson , m. , nandra , k. , digby - north , j. , prez - gonzlez , p. g. , & barro , g. , 2010 , , 406 , 420 ghisellini , g. , haardt , r. , & matt , g. , 1994 , , 267 , 743 gilli , r. , comastri , a. , hasinger , g. , 2007 , , 463 , 79 gonzlez - martn , o. , masegosa , j. , mrquez , i. , & guainazzi , m. , 2009 , , 704 , 1570 griffiths , r.l . & stern , d. , 2010 , , 140 , 533 hartwick , f.d.a & schade , d. , 1990 , ara&a , 28 , 437 hasinger , g. , 2008 , , 490 , 905 hickox , r.c . , et al . , 2009 , , 696 , 891 hopkins , p.f .
& hernquist , l. , 2009 , , 694 , 599 hopkins , p.f . ,
hernquist , l. , cox , t.j . , dimatteo , t. , roberson , b. , & springel , v. , 2006 , , 163 , 1 kauffmann , g. , et al .
, 2007 , , 173 , 357 kauffmann , g. , white , s.d.m . , heckman , t.m . , mnard , b. , brinchmann , j. , charlot , s. , tremonti , c. , & brinkmann , j. , 2004 , , 353 , 713 kauffmann , g. , et al .
, 2003 , , 346 , 1055 kennicutt , r.c . , 1998 , ara&a , 36 , 189 kim , m. , ho , l.c .
, i m , m. , 2006 , , 642 , 702 knigl , a. & kartje , j.f . , 1994 , , 434 , 446 kormendy , j. & kennicutt , r.c . , 2004 , ara&a , 42 , 603 kormendy , j. & richstone , d. , 1995 , , 33 , 581 kumar , p. & johnson , j.l . , 2010 , , 404 , 2170 la franca , f. , et al .
, 2005 , , 635 , 864 lawrence , a. , 1991 , , 252 , 586 lawrence , a. & elvis , m. , 2010 , , 714 , 561 lee , j.c . , et al . , 2009 , , 706 , 599 lutz , d. , et al . , 2010 , , 712 , 1287 mainieri , v. , et al . , 2005 , , 437 , 805 martnez - sansigre , a. , et al . , 2010 , , 706 , 184 mckernan , b. , ford , k.e.s .
, & reynolds , c.s . , 2010a , , 407 , 2399 mckernan , b. , maller , a. , & ford , k.e.s . , 2010b , , 718 , l83 melbourne , j. , et al . , 2011 , , in press nardini , e. & risaliti , g. , 2011 , , in press page , m.j . , stevens , j.a . , ivison , r.j . , & carrera , f.j . , 2004 , , 611 , l85 pierce , c.m . , et al . , 2010 , , 408 , 139 pierce , c.m . , et al . , 2007 , , 660 , l19 polletta , m. , weedman , d. , hnig , s. , lonsdale , c.j . , smith , h.e . , & houck , j. , 2008 , , 675 , 960 polletta , m.d.c . , et al . , 2006 , , 642 , 673 pope , a. , et al . , 2008 , , 689 , 127 rieke , g.h . ,
alonso - herrero , a. , weiner , b.j .
, prez - gonzlez , p.g . , blaylock , m. , donley , j.l .
, & marcillac , d. , 2009 , , 692 , 556 rigopoulou , d. , et al . , 2009 , , 400 , 1199 risaliti , g. , maiolino , r. , & salvati , m. , 1999 , , 522 , 157 rodrguez zaurn , j. , tadhunter , c.n . , & gonzlez delgado , r.m . , 2010 , , 403 , 1317 rovilos , e. , fotopoulou , s. , salvato , m. , burwitz , v. , egami , e. , hasinger , g. , & szokoly , g. , 2011 , , in press rovilos , e. , georgantopoulos , i. , akylas , a. , & fotopoulou , s. , 2010 , , 522 , 11 salpeter , e.e . , 1955 , , 121 , 161 sanders , d.b . , soifer , b.t . , elias , j.h . ,
madore , b.f . ,
matthews , k. , neugebauer , g. , & scoville , n.z .
, 1988 , , 325 , 74 santini , p. , et al .
, 2010 , , 518 , l154 schawinski , k. , thomas , d. , sarzi , m. , maraston , c. , kaviraj , s. , joo , s .-
yi , s.k . , & silk , j. , 2007 , , 382 , 1415 scoville , n. , et al . , 2007 , , 172 , 38 serjeant , s. , et al . , 2010 , , 518 , l7 serjeant , s. & hatziminaoglou , e. , 2009 , , 397 , 265 seymour , n. , et al . , 2011 , , 413 , 1777 shapiro , k.l . , et al . , 2008 , , 682 , 231 shi , y. , rieke , g.h . , ogle , p. , jiang , l. , & diamond - stanic , a.m. , 2009 , , 703 , 1107 silverman , j.d . , et al . , 2009 , , 695 , 171 simpson , c. , 2005 , , 360 , 565 smoli , v. & riechers , d.a . , 2011 , , in press terashima , y. , & wilson , a.s . , 2003 , , 583 , 145 thompson , g.d . , levenson , n.a . , uddin , s.a . , sirocky , m.m .
, 2009 , , 697 , 182 thompson , t.a . , quataert , e. , & murray , n. , 2005 , , 630 , 167 treister , e. , et al . , 2006 , , 640 , 603 treister , e. , et al . , 2009 , , 706 , 535 treister , e. , urry , c.m . , &
virani , s. , 2009 , , 696 , 110 treister , e. , natarajan , p. , sanders , d.b . ,
urry , c.m . ,
schawinski , k. , & kartaltepe , j. , 2010 , science , 328 , 600 treister , e. & urry , c.m . , 2006 , , 652 , l79 treister , e. , & urry , c.m . , 2005 , , 630 , 115 ueda , y. , akiyama , m. , ohta , k. , & miyaji , t. , 2003 , , 598 , 886 wilman , r.j .
, jarvis , m.j . ,
mauch , t. , rawlings , s. , & hickey , s. , 2010 , , 405 , 447 winter , l.m . , mushotzky , r.f . , reynolds , c.s . , &
tueller , j. , 2009 , , 690 , 1322 wold , i. , sheinis , a.i . ,
wolf , m.j . , & hooper , e.j . , 2010 , , 408 , 713 ccccc @xmath1 @xmath2 0.5 & 2.0 & 7.0 & 4.5 & 10 + 0.5 @xmath2 @xmath1 @xmath2 1 & 2.0 & 7.0 & 4.5 & 7.0 + 1 @xmath2 @xmath1 @xmath2 2 & 2.0 & 4.5 & 4.5 & 4.5 + 2 @xmath2 @xmath1 @xmath2 3 & 2.0 & 2.0 & 2.0 & 2.0 + 3 @xmath2 @xmath1 & 1.0 & 1.0 & 1.0 & 1.0 + lcccc type 1 agn & col 2 & 3 table [ ages ] & 0.25 & 0.05 & 2.0 ( 20 ) + unified / type 2 agn & col 2 & 3 table [ ages ] & 0.25 & 0.5 & 2.0 ( 20 ) + + * non - evolving model ct agn * & & & & + w / o enhanced star formation sources & 1.0 & 1.0 & 1.0 & 10 + w/ enhanced star formation sources & col 2 & 3 table [ ages ] & 0.25 & 0.5 & 2.0 ( 20 ) + + * evolving model ct agn * & & & & + low eddington ratio ct agn & col 4 & 5 table [ ages ] & 1.0 & 1.0 & 1.0 + high eddington ratio ct agn & 0.3 & 0.6 & 0.6 & 175 + | studying the average properties of active galactic nuclei ( agn ) host stellar populations is an important step in understanding the role of agn in galaxy evolution and the processes which trigger and fuel agn activity .
here we calculate model spectral energy distributions ( seds ) that include emission from the agn , the host galaxy stellar population , and dust enshrouded star formation . using the framework of cosmic x - ray background population synthesis modeling , the model agn hosts are constrained using optical ( b band ) and near infrared ( j band , 3.6 @xmath0 m , 5.7 @xmath0 m , 8.0 @xmath0 m , and 24 @xmath0 m ) luminosity functions and number counts .
it is found that at @xmath1 @xmath2 1 , type 1 and type 2 agn hosts have similar stellar populations , in agreement with the orientation based unified model and indicative of secular evolution . at @xmath1 @xmath3 1 , type 2 agn hosts are intrinsically different from type 1 agn hosts , suggesting that the simple orientation based unified model does not hold at @xmath1 @xmath3 1 .
also , it is found that if compton thick ( ct ) agn evolve like less obscured type 2 agn , then , on average , ct agn hosts are similar to type 2 agn hosts ; however , if ct obscuration is connected to an evolutionary stage of black hole growth , then ct agn hosts will also be in specific evolutionary stages .
multi - wavelength selection criteria of ct agn are discussed . |
cogent ( coherent germanium neutrino technology ) is a program aiming to exploit the characteristics of p - type point - contact germanium detectors in areas as diverse as the search for low - mass dark matter candidates , coherent neutrino - nucleus elastic scattering , and @xmath2ge double - beta decay @xcite .
data collected from a first cogent detector at a shallow underground location demonstrated sensitivity to low - mass ( @xmath3 gev / c@xmath1 ) dark matter particles @xcite .
in particular , it appeared cogent was particularly well suited to address the dama / libra @xcite modulation result . following the identification of several sources of internal background in this prototype ,
a second cogent detector was installed in the soudan underground laboratory ( sul ) during 2009 with the goal of improving upon the dark matter sensitivity reach of the 2008 result @xcite .
the first 56-days of operation of the cogent detector at sul showed an unexpected excess of events @xcite above the anticipated backgrounds for ionization energies below 2 kev .
further data collection from this detector continued until an interruption imposed by a fire in the access shaft to the laboratory halted the initial run in march of 2011 .
analysis of the accumulated data set @xcite , spanning 442 live days over the period 4 december 2009 to 6 march 2011 , showed a @xmath4 significance modulation of the monthly event rate in the low - energy region that is compatible with the dark matter signature described in @xcite .
the fitting procedure generating this low - significance modulation result used unconstrained phase , period , and amplitude variables .
time - stamped data have been made publicly available , allowing for a number of independent analyses and interpretations . in this paper
we provide a more in - depth description of the apparatus and data analysis , concentrating on aspects of instrument stability , data cuts , uncertainties , and background estimation .
the data set employed for this discussion is the same as in @xcite , and all energies are in kevee ( kev electron equivalent , i.e. , ionization energy ) , unless otherwise stated . following the three - month outage resulting from the soudan fire , this detector has taken data continuously , starting 7 june 2011 .
an additional body of data is to be released in the near future . the design and expectations for cogent-4 ( c-4 ) , a planned expansion aiming at an increase in active mass by a factor of ten , featuring four large ppc detectors with a reduced energy threshold and lower background , are discussed in a separate publication @xcite .
the present cogent detector is located at the soudan underground laboratory ( soudan , minnesota , usa ) at a vertical depth of 2341 feet ( 689 feet below sea level ) , providing 2090 meters of water equivalent ( m.w.e . ) overburden as shielding against cosmic rays and associated backgrounds .
the detector shield is placed on a floor built on top of base i - beams that once supported the soudan-2 proton decay experiment @xcite .
the detector element is a single modified bege germanium diode .
bege ( broad energy germanium ) is the commercial denomination used by the manufacturer ( canberra industries ) for their line of ppc detectors .
the technical characteristics of this ppc are shown in table [ tab : begecharacteristics ] .
the detector is contained within an ofhc copper end cap cryostat , and mounted in an ofhc copper inner can connected to an ofhc copper cold finger .
internal detector parts were custom manufactured in either ofhc copper or ptfe .
all internal parts were etched to remove surface contaminations using ultra - pure acids in class 100 clean room conditions , following procedures similar to those described in @xcite .
a commercial stainless steel horizontal cryostat encloses the rear of the assembly , providing electrical feed - through to a side - mounted canberra dprp pulse - reset preamplifier typically used in high - resolution x - ray detectors ( figure [ fig : begeinnershield ] ) .
.[tab : begecharacteristics ] characteristics of the cogent high purity ppc germanium detector at sul . [ cols="<,>",options="header " , ] similar to figure [ fig : muoninduced ] , with expanded ranges : energy spectrum of the simulated @xmath5u , @xmath6th , and @xmath7k resistor background ( dotted line ) compared to cogent data ( solid ) . in the energy range
displayed the estimated resistor backgrounds are by far dominant .
the resistor background spectrum is for metal film resistors , the same used in the cogent front end .
also shown are other background contributions and their sum .
contributions from @xmath8pb bremsstrahlung and radioactivity in ptfe and ofhc cryostat parts are comparatively negligible . ] as a further consistency check we examined the existing cogent data out to an energy of 300 kevee .
the statistics in this range are limited ( 5 days of dedicated exposure , see sec.iii-a ) .
figure [ fig : data300gammas ] shows possible 238 kev @xmath9pb ( @xmath6th chain ) and 295 kev ( @xmath5u chain ) gamma lines . due to their relatively - low energy ,
their source would be near the crystal , within the inner lead cavity .
if they are considered as a measure of the @xmath5u and @xmath6th chain contamination in front - end resistors , a 14@xmath107 bq / kg for @xmath5u contamination and [email protected] bq / kg for @xmath6th contamination is obtained for the resistors .
this activity would provide @xmath11937 events in the 0.5 - 3.0 kevee region , in good agreement with the measured flat component of the spectrum .
the statistical evidence for these lines is however slim , and their presence is seen to be mutually exclusive when examining the uncertainties associated to the energy scale extrapolation used for this short run . existing cogent data in the range up to 300 kev , with possible weak @xmath9pb ( 238 kev ) and @xmath12pb ( 295 kev ) gamma lines indicated by arrows .
the extrapolated energy scale can only be considered approximate .
the energy binning corresponds to the approximate fwhm resolution for these two lines .
see text for a discussion on a possible origin for these putative lines in the front - end resistors .
notoriously absent are a @xmath8pb peak at 46.5 kev and excess lead x - rays , a result of the radiopurity of the inner lead layers in the shield ( sec.ii-a ) and detector surfaces ( sec.v - c ) . ]
while the smallness of neutrino cross - sections indicate that their contribution to the cogent spectrum should be negligible , the signal from coherent neutrino - nucleus scattering @xcite from several sources ( e.g. solar , atmospheric , diffuse supernova , and geo - neutrinos ) would be highly concentrated at low energies .
we engage here in the exercise of providing a few estimated upper limits for these contributions .
inferring from a recent analysis on solar and atmospheric neutrinos @xcite , a germanium detector with 0.33 kg active mass and a @xmath112 kev nuclear recoil threshold ( as in the present cogent detector ) would observe a rate of just @xmath110.012 counts / year from coherent neutrino - nucleus scattering from @xmath13b and @xmath14he - proton fusion ( hep ) solar neutrinos , the only solar sources able to produce a signal above threshold .
diffuse supernova background neutrinos and atmospheric neutrinos might also contribute , however their rate is reduced by factors of @xmath15 10@xmath16 @xcite and @xmath15 10@xmath17 @xcite , respectively .
geoneutrinos , having energies less than 4.5 mev @xcite , can not produce nuclear recoil energies above the cogent threshold . each of these sources may also induce direct electron scattering . however , the neutrino - electron scattering rate is suppressed by @xmath11 10@xmath17 relative to the neutrino - nucleus coherent scattering rate @xcite .
therefore this other channel can not significantly contribute even taking into account the factor of 32 increase in scattering targets , the absence of a quenching factor , and the higher electron recoil energies .
we notice however that interaction rates large enough to be of interest can be generated by solar neutrinos with enhanced baryonic currents @xcite .
additional mechanisms @xcite are able to generate a phenomenology involving diurnal and yearly modulations in rates .
cogent is the first detector technology specifically designed to look for wimp candidates in the low mass range around 10 gev / c@xmath1 , an area of particular interest in view of existing anomalies in other dark matter experiments , recent phenomenological work in particle physics , and possible signals using indirect detection methods @xcite .
however , investigation of the largely unexplored @xmath11few kev recoil energy range brings along new challenges in the understanding of low - energy backgrounds .
the experience accumulated during the ongoing cogent data - taking at sul demonstrates that ppc detectors have excellent properties of long - term stability , simplicity of design , and ease of operation .
this makes them highly suitable in searches for the annual modulation signature expected from dark matter particles forming a galactic halo .
besides their excellent energy resolution , low energy threshold and ability to reject surface backgrounds , ppcs compare well to other solid - state detectors under several criteria : a ) the relative simplicity of cogent s data analysis results in comparable irreducible spectra regardless of analysis pipeline , b ) the response to nuclear recoils is satisfactorily understood , resulting in a reliable nuclear recoil energy scale , c ) uninterrupted stable operation of ppc detectors can be expected over very long ( several year ) timescales .
we plan to continue improving this technology and our understanding of low - energy backgrounds within the framework of a cogent expansion , the c-4 experiment @xcite .
we are indebted to jeffrey de jong and alec habig ( minos collaboration ) for sharing with us information on radon and muon rates at sul , and to all sul personnel for their constant support in operating the cogent detector .
work sponsored by nsf grants phy-0653605 and phy-1003940 , the kavli foundation , and the pnnl ultra - sensitive nuclear measurement initiative ldrd program ( information release number pnnl - sa-90298 ) .
and t.w.h . are supported by the doe / nnsa stewardship science graduate fellowship program ( grant number de - fc52 - 08na28752 ) and the intelligence community ( ic ) postdoctoral research fellowship program , respectively . as a consistency test ,
the same corrected spectrum at the bottom panel of figure [ fig : steps ] can be recovered by ignoring the 90% c.l .
fast pulser cut , and applying in lieu of it the overall fraction of fast events derived from fits like those in figure [ fig : surface_fits ] . following a sufficiently long exposure ,
the distribution in rise time of surface and bulk events becomes self - evident , making possible to calculate bulk sa and surface br for any choice of rise time cuts .
this supersedes the electronic pulser method , which is limited by only providing a bulk sa , but no surface br information .
e. fernandez - martinez and r. mahbubani , jcap * 1207 * ( 2012 ) 029 ; r. bernabei _ et al .
j. * c72 * ( 2012 ) 2064 ; s. chang , j. pradler and i. yavin , phys . rev .
* d85 * ( 2012 ) 063505 ; j. pradler , arxiv:1205.3675 . | cogent employs p - type point - contact ( ppc ) germanium detectors to search for weakly interacting massive particles ( wimps ) . by virtue of its low energy threshold and ability to reject surface backgrounds ,
this type of device allows an emphasis on low - mass dark matter candidates ( @xmath0 gev / c@xmath1 ) .
we report on the characteristics of the ppc detector presently taking data at the soudan underground laboratory , elaborating on aspects of shielding , data acquisition , instrumental stability , data analysis , and background estimation .
a detailed background model is used to investigate the low energy excess of events previously reported , and to assess the possibility of temporal modulations in the low - energy event rate .
extensive simulations of all presently known backgrounds do not provide a viable background explanation for the excess of low - energy events in the cogent data , or the previously observed temporal variation in the event rate
. also reported on for the first time is a determination of the surface ( slow pulse rise time ) event contamination in the data as a function of energy .
we conclude that the cogent detector technology is well suited to search for the annual modulation signature expected from dark matter particle interactions in the region of wimp mass and coupling favored by the dama / libra results . |
the transport mechanism and in turn the cause of resistance is of fundamental importance . various model , especially for thin films ,
exist to understand the contribution from different scattering mechanisms .
the film resistance , however , may be due to a combination of three mechanisms , namely ( i ) due to scattering from phonons , impurities and point defects etc .
, ( ii ) from film surface and ( iii ) due to grain boundaries which would be predominant in polycrystalline films@xcite .
different models exist to explain the dependence of resistance on film thickness . in the case of scattering from the film surface
the variation from film resistivity with film thickness was given by the fuchs- sondheimer ( f - s ) relationship@xcite @xmath1 \nonumber\end{aligned}\ ] ] where @xmath2 is the mean free path , d is the film thickness and @xmath3 the resistivity of the bulk sample .
the constant p indicates the fraction of electrons being reflected from the surface .
the value indicates the scattering mechanism , for example p=1 indicates specular reflection .
a similar relationship was established by mayadas and shatzkes ( m - s ) @xcite to explain the scattering from grain boundaries , with a very similar functional dependence with film thickness .
however , the model is limited to very thin films with an added restriction that the grain size are of the order of film thickness .
the grain boundary is defined as region between two grains where crystal orientation changes .
the transport properties of @xmath0 films like resistivity , hall coefficient , mobility and seeback coefficient have been extensively studied @xcite , and reports in literature indicate films to be p - type with narrow band gap .
however , all these reports are silent on the mechanism of scattering and in turn the source of resistivity . only damodara das et al @xcite
have reported resistivity as a function of thickness @xmath4 .
the article states average grain size to be of the order of film thickness and indicates the scattering mechanism to be that of grain boundary scattering in accordance to the m - s model .
however , no report exists on the variation of resistance or resistivity of @xmath0 films with grain size in thicker films .
this article investigates variation of resistance in films whose thickness is enough to assume that the variation in resistivity is independent of defects and specular scattering .
films of @xmath0 were grown on glass substrates kept at room temperature , using thermal evaporation method .
@xmath0 ingot of high purity ( 99.99% ) supplied by aldrich ( usa ) were used as the starting material .
the crushed ingot were evaporated from molybdenum boat at a vacuum better than @xmath5torr .
the film thickness was measured using dektek iia surface profiler .
the movement of the mechanical stylus across the edge of the film determines the step height or the films thickness .
indium contacts were grown on the glass substrates before they were placed in the chamber , such that a strip of @xmath0 film of dimensions @xmath6 could be fabricated on these contacts using a mask .
the i - v characteristics of the films were measured by four probe method .
it was found to be linear between 25mv-24v , showing the ohmic nature of indium contacts as well as the polycrystalline film for applied field .
the films resistance were measured by an high input impedance digital multimeter . the structural and compositional analysis of these films
were done using phillips pw1840 x - ray diffractometer and shimadzu esca750 ( electron spectroscopy for chemical analysis ) .
the films were found to be stoichiometrically uniform over the area 5 cm x 5 cm as determined by esca carried out in various regions of the film .
the morphological analysis was done with joel 840 scanning electron microscope ( sem ) .
the as grown films showed tendency of ageing @xcite , where the resistance of the film varied with time and saturated to a constant value in couple of weeks .
the results presented in this article are of films which had achieved such saturation .
the average grain size was determined from both sem micrographs and x - ray diffractograms .
the grain size was calculated using the full width at half maxima ( fwhm ) of x - ray peaks @xcite .
the results of grain size found by both methods were in agreement .
a plot between the film thickness and grain size shows no trend ( fig 1 ) .
this variation in grain size with film thickness may be a result of not having perfectly identical conditions during film evaporation .
it also represents the randomness of the growth process .
this shows that the average grain size is not proportional to the film thickness and resistance or resistivity will have to be studied both as a function of thickness and grain size to resolve the main contributor in scattering mechanism .
the f - s theory shows that the contributions from the surface leads to an inverse proportionality with thickness ( equation 1 ) , where the model is restricted to cases where the charge carriers mean free path is of the order of the film thickness @xmath7 .
however , since the samples in our study have thickness between 130 - 500 nm , the film thickness is far greater then the mean free path . beyond this limit
one can assume the films resistivity to be same as that of the bulk , showing no further change with increasing film thickness .
thus , film s resistance in this limit should only fall inversely with thickness .
figure 2 shows the resistance of aged @xmath0 films falling linearly with increasing thickness .
it can be understood trivially , that for the resistance of the film to vary linearly with thickness as shown in figure 2 , the resistivity would have to show a parabolic relationship with film thickness .
another important contributor to resistance is the grain boundary .
however , that too requires an inverse proportionality with thickness .
this lack of trend may be due to the assumption in m - s theory that the grain size is proportional to the film thickness , which is not the case here .
it is clear that in the present study the surface scattering and grain boundary scattering do not contribute to the film resistance .
hence , to investigate the influence of the grain size on transport properties , variation of resistance with grain size was studied .
figure 3 shows the variation of film resistance with grain size .
as stated earlier the average grain size was determined from both sem micrographs and x - ray diffractograms .
the grain boundary is defined as region between two grains where crystal orientation changes .
the representative micrographs of @xmath0 in figure 4 however , show large distances between two grains .
the grains tend to have the resistivity of the bulk , however , even if there is an inter - connectivity between two neighbouring grains this region will have high resistivity by purely geometry of narrowing @xcite . these voids , hence would definitely contribute differently from the defined grain boundary in m - s theory .
volger s model @xcite assumes the film to be made up of cubical grains of edge size a arranged in an ordered manner , as shown in fig .
5a , with equal spacing between the neighbouring grains .
the inter - grain distances are different along x , y and z directions and are same along any one direction .
consider the film has q number of grains arranged regularly at equal inter- grain spacing @xmath8 along the length l and r and p grains arranged along the width and thickness of the film . also , the resistance is measured along the length of the film by taking the contacts across the cross - section in the yz plane , then the points a - b , c - d etc . shown in fig .
5a are at equal potential
. the equivalent dc circuit of this arrangement of measurement would be as shown in figure 5b , where @xmath9 represents the high resistance of the inter - grain voids @xcite . as can be seen in figure 5b ,
the whole film can be considered to be a parallel combination of pr resistive elements , where resistance of each element is given by @xcite @xmath10 thus , the net resistance along the length of the film between the two contacts would be given as @xmath11 seto @xcite made a similar simplification step by assuming the problem to be that of one dimension .
@xmath9 , the high resistance of the inter - grain voids , is a function of @xmath8 which in turn would depend on the mechanism by which charge carriers would cross the inter - grain boundary .
many suggestions have been made for explaining the cross over , such as ohmic conduction , tunnelling or thermionic emission @xcite .
it may also be a combination of these , depending on the actual inter - grain distances .
the resistance of such a film , assuming ohmic conduction in between grains is given as @xcite @xmath12 where @xmath13 is a proportionality constant , given as @xmath14 in true sense @xmath13 is not a constant since @xmath15 will depend on the grain size , as also film dimensions , including it s thickness .
however , @xmath15 is assumed to be a slow varying function of film thickness , or constant .
the constant k represents a ratio of the inter - grain region s resistivity to the grain s resistivity .
since the void resistivity is large , k obviously is a very large entity .
the variable is a ratio of the inter - grain length and the grain edge or @xmath16 where a is the grain size , assuming as in volger s model , the grains to be cubic in nature .
the inter - grain distance , @xmath8 is extent of void in x direction ( along length of the film strip ) , since the resistance is measured along the length of the film .
the inter - grain distance varies as a function of the grain size depending the mechanism of grain growth . since , the @xmath0 films aged to a hexagonal crystal state , with @xmath17 , it should show easier grain growth along the length and width as compared with that along restrictive film thickness .
the films hence aged with the c - axis aligned normal to the substrate plane @xcite . as per volger s model
q , the grain number along the length , would be decreasing more rapidly than p , that along the film s thickness , leading to a general trend of decrease in resistance .
thus , the variation of inter - grain distance with grain size for the films in consideration would be given as @xcite @xmath18 where @xmath19 .
thus , it can be seen that the variable , is a function of the grain size , a. the increase in void size with increasing grain size can be appreciated from the representative sem micrographs .
equation 2 , is physically valid for positive values , which requires @xmath20 considering a extreme case of @xmath21 along with the stated assumption that @xmath22 is constant , then equation 2 maybe written as @xmath23 the variable x required for equation ( 1 ) can then be expressed as @xmath24 this increase in inter - grain distance with growing grain size was discussed in our earlier work @xcite .
hence , using equation ( 3 ) the films resistance given by equation ( 1 ) , can be expressed as @xmath25 equation ( 4 ) fits quite well to the experimental observations as shown by the solid line in figure 4 .
the values of the constants evaluated by fitting are @xmath13 = 2308 @xmath26 , @xmath27 and @xmath28 . as stated earlier , the constant k is a ratio of the high resistivity of the inter - grain region as to the low resistivity of the grains .
the numerical value shows the resistivity of the inter - grain region will be nearly @xmath29 times that of the low resistivity region , which is consistent with the with the assumption that inter - grain region can be assumed to be a path of high resistance .
the voids between neighbouring grains would present itself as a barrier which the charge carriers would have to transverse to establish current flow .
the magnitude of the barrier height can be computed from the slope of the plot between @xmath30 and temperature inverse ( 1/t in kelvin ) .
the barrier height was calculated using this method for various film thickness .
the variation is shown in figure 6 .
the variation in barrier height with film thickness maybe due to one or a cumulative effect of the following ( i ) variation in the grain size of the polycrystalline film , ( ii ) a large density of dislocations , ( iii ) quantum size effects and ( iv ) change in film stoichiometry . since the film thickness of this study is large the quantum size effect
is immediately ruled out .
careful growth technique followed by ageing would minimise the contribution due to dislocation and off stoichiometric compositions , however , can not be completely ruled out .
the major contribution hence would be due to the size of the grains .
slater @xcite estimated the barrier height as a function of grain size by modelling grain boundary as a pn type of structure . the variation is given as @xmath31 where @xmath32 is the doping concentration , n the carrier concentration , k the dielectric constant of the material .
the barrier height increases with grain size for @xmath33 , which would be the case in pure samples @xmath34 .
a fit for the experimental data using equation 3 , which states that the barrier width in proportional to the square of the grain size , along with the estimated proportionality constant ( b ) and equation 5 is shown by the continuous line in figure 6 .
the fit shows reasonable agreement , however it does indicate possible contributions from dislocations etc .
the values of the coefficients of equation 5 are @xmath35=5.66mev , @xmath36 = 90 and @xmath37 .
the ratio of @xmath36 may appear to be very small , however , it should be noted that the curve fitting was done using the earlier estimated proportionality constant ( @xmath38 ) .
thus , the influence of the inter - grain voids and grain size on the magnitude of barrier height and various characterising parameters of the film is evident .
this though expected is different from an earlier study by rajagopalan et al @xcite on films of @xmath0 with thickness between 160 nm and 800 nm which reported that the barrier height was independent of the film thickness . in the next section we investigate the role of the inter - grain voids on another parameter used to characterise the transport properties of a material .
the resistance of the film is a function of temperature .
the variation of resistance with temperature in general is expressed as @xmath39 thus the temperature coefficient of resistance or tcr @xcite is given as @xmath40 while tcr is positive for metals , it is negative in case of semiconductors . the f - s model for very thin films states
that the variation of tcr with film thickness follows an identical form as expressed by equation ( 1 )
. however , there seems to be no model in the literature to explain the variation of tcr with either film thickness or grain size for films with thickness greater then the mean free path of their charge carriers .
a plot between tcr and grain size seems scattered ( not shown ) .
however , the plot between tcr and barrier height is a straight line , figure 7 , with slope @xmath41 and intercept @xmath42 .
it immediately follows from the linearity between tcr and @xmath43 along with equation ( 5 ) that the tcr would be a polynomial function of the grain size .
this explains the seemingly scattered data points of tcr with grain size as discussed .
it also explains the lack of any model or theory on the variation of tcr with grain size .
the relationship shows that with increasing barrier height , the rate of change of resistance with temperature ( dr / dt or tcr , equation 6 ) becomes increasingly smaller .
the negative temperature coefficient of resistance in semiconductors is a result of increasing charge carriers due to breaking covalent bonds .
an increased barrier height implies its more difficult for the charge carriers to escape into the voids from the grains , confining the increased number of carriers inside the grain itself .
this rapidly brings down the resistance of the grain contributing to a negative tcr proportional to the barrier height .
the dc transport properties of @xmath0 films with thickness between 130 - 500 nm have been discussed .
the properties showed no size effects as was expected , since the film thickness was far greater then the charge carriers mean free path .
the films resistance , the barrier height and temperature coefficient of resistance ( tcr ) showed a strong dependence on both the grain size and inter- grain void .
the inter - grain void was approximated to vary with increasing grain size , enabling to study the above properties of the films as a function of grain size . | resistance of vacuum deposited @xmath0 films of thickness between 100 - 500 nm has been measured in vacuum .
it is found that the resistance of the polycrystalline films strongly depends on the grain size and inter - granular voids .
the charge carrier are shown to cross this high resistivity inter- granular void by ohmic conduction .
the barrier height as well as temperature coefficient of resistance are also shown to depend on the grain size and inter- grain voids . |
in a previous paper @xcite we introduced the so - called simplicial lusternik - schnirelmann category of a simplicial complex .
this invariant , denoted by @xmath0 , is defined directly from the combinatorial structure of the complex @xmath1 , instead of considering the topological ls - category of the geometric realization @xmath2 , by replacing the notion of homotopy by that of contiguity . with this approach it turned out that the simplicial ls - category is an invariant of the strong homotopy type , as defined by barmak and minian in @xcite .
also , simplicial ls - category is closely related to the ls - category of the finite @xmath3-space represented by the hasse diagram of @xmath1 @xcite .
it is worth noting that although the idea of contiguous simplicial maps is a classic one , the corresponding theory of lusternik - schnirelmann category had not been developed until now .
then , many natural questions remained unsolved in this new context .
the aim of this paper is to answer some of them : for instance , the relationship between @xmath0 and @xmath4 , the whitehead formulation of simplicial category , or an analogue of varadarajan s formula for fibrations , among others .
other results include a direct proof of the formula @xmath5 , for the barycentric subdivision , or a counterexample of the analogue of the homotopy extension property for subcomplexes
. a final important result will be that simplicial category generalizes to arbitrary simplicial complexes the well known notion of arboricity of a graph .
the contents of the paper are as follows .
we start ( section [ 2 ] ) by introducing the definition of simplicial lusternik - schnirelmann category @xmath0 of the simplicial complex @xmath1 .
the idea is to define an analogue of the classical ls - category of a topological space , by replacing the notion of homotopic maps by that of contiguous simplicial maps .
we then prove some basic properties of this new notion of category , and we recall our previous result @xcite that @xmath6 is an invariant of the strong homotopy type . in particular , the category of the simplicial complex @xmath1 equals that of its core @xmath7 ( the minimal subcomplex of @xmath1 obtained by strong collapses ) . as a corollary
we prove that @xmath0 is bounded above by the number of vertices and the number of maximal simplices of @xmath7 . in section
[ 3 ] we study the effect of barycentric subdivision on simplicial category and we give a direct proof of @xmath8 . for
that we need to prove that if @xmath9 , @xmath10 are contiguous maps , then the induced maps @xmath11 , @xmath12 , are in the same contiguity class .
the only proof that we found in the literature is as a consequence of the results given in @xcite via finite posets . in section
[ 4 ] we compare the simplicial category of the complex @xmath1 with the classical ls - category of the geometric realization @xmath2 , proving that @xmath13 .
this implies that @xmath14 , for any iterated barycentric subdivision @xmath15 of @xmath1 , although there are examples where this inequality is strict ( we show one of these kind of examples ) .
section [ 5 ] deals with the study of simplicial products .
being more precise , we consider the categorical product , denoted by @xmath16 .
then , we prove that @xmath17 .
we are now ( section [ 6 ] ) in a position to discuss the simplicial analogue of the whitehead formulation of the ls - category .
it is known that for topological spaces @xmath18 with `` good '' properties the ls - category can be computed as the least integer @xmath19 such that the diagonal map @xmath20 factors ( up to homotopy ) through the so - called fat wedge @xmath21 .
we try to adapt this result to abstract simplicial complexes .
first , we define the @xmath22-th fat wedge @xmath23 for any pointed complex @xmath1 and we study its behaviour under contiguity .
we then define a simplicial whitehead category @xmath24 , which is an invariant of the pointed strong homotopy type and we are able to prove that @xmath25 .
however , unlike the continuous case , the other inequality is not true , as we show by an example . in section [ 7 ]
we show how the partial failure of the whitehead formulation is related to the contiguity extension property .
we discuss a new notion of cofibration in the simplicial setting .
it is well known that if @xmath26 is a subcomplex of a cw - complex @xmath18 then the pair @xmath27 has the homotopy extension property .
therefore , if @xmath28 is a subcomplex of a simplicial complex @xmath1 , the pair @xmath29 has the ( topological ) homotopy extension property .
we define a purely combinatorial analogue for a simplicial pair @xmath30 , but we show counterexamples where the contiguity extension property fails to be true . finally , section [ 8 ] is focused on the study of the simplicial ls - category in the one - dimensional case , that is , on graphs . the well known graph - theoretical notion of _ arboricity _ will play a central role in this study .
the arboricity @xmath31 of the graph @xmath32 is the cardinality of a minimal decomposition of @xmath32 into disjoint spanning forests , i.e. , acyclic subgraphs which are non - necessarily connected and cover all the vertices .
the aim of this section is to prove that arboricity coincides with simplicial category , being more precise , we prove that @xmath33 .
the two main ideas are that categorical subcomplexes are precisely forests and that the notion of contiguity is a very rigid one in the setting of graphs , allowing only a limited number of moves .
as a final comment , we emphasize that simplicial ls - category is a new strong homotopy invariant , defined in purely combinatorial terms , that generalizes to arbitrary simplicial complexes the well known notion of arboricity of a graph , and that allows to develop all the machinery of algebraic topology which is costumary in the classical theory of lusternik - schnirelmann category .
[ [ section ] ] we recall the notion of simplicial ls - category , introduced by the auhors in @xcite
. let @xmath1 and @xmath28 be abstract simplicial complexes .
remember ( see @xcite ) that two simplicial maps @xmath34 are _ contiguous _ , denoted by @xmath35 , if for any simplex @xmath36 of @xmath1 , the set of vertices @xmath37 is a simplex of @xmath28 .
more generally , two simplicial maps @xmath38 are in the same _ contiguity class _ , denoted by @xmath39 , if there is a sequence of simplicial maps @xmath40 , @xmath41 , such that @xmath42 [ full ] let @xmath38 be two simplicial maps . let @xmath43 be a full subcomplexes such that @xmath44 .
if @xmath9 and @xmath10 are contiguous then their restrictions @xmath45 are contiguous too .
the concept of contiguity class provides a simplicial analogue of homotopy classes of continuous maps between topological spaces . by means of the notion of contiguity we define the key concept of categorical subcomplex .
let @xmath1 be an abstract simplicial complex .
we say that the subcomplex @xmath46 is _ categorical _ ( in @xmath1 ) if there exists a vertex @xmath47 such that the inclusion @xmath48 and the constant map @xmath49 are in the same contiguity class ( denoted by @xmath50 ) .
in other words , the inclusion of @xmath51 factors through the vertex @xmath47 up to `` simplicial homotopy '' ( meaning contiguity class ) . notice that a categorical subcomplex may not be connected .
[ relative]let @xmath1 be an abstract simplicial complex .
the _ simplicial ls - category _ of @xmath1 , denoted by @xmath0 , is the least integer @xmath52 such that there exist @xmath53 categorical subcomplexes @xmath54 of @xmath1 which cover @xmath1 , that is , such that @xmath55 .
for instance , as we shall explain in subsection [ properties ] , @xmath56 if and only if @xmath1 is strongly collapsible to a point , in the sense given by barmak and minian @xcite . [ [ section-1 ] ] we now show that the simplicial ls - category can be computed by taking into account only maximal simplices .
the proof is inspired on an idea from j. strom for finite topological spaces .
[ subcat]let @xmath57 be subcomplexes of @xmath1 .
if @xmath51 is categorical ( in @xmath1 ) then @xmath58 is categorical ( in @xmath1 ) .
if the inclusion @xmath59 verifies @xmath50 for some constant map @xmath60 , and @xmath61 is the inclusion , then @xmath62 is the inclusion , @xmath63 is a constant map , and @xmath64 .
[ max ] in definition [ relative ] one may assume that : 1 .
each categorical subcomplex in the covering @xmath54 is a union of maximal simplices of @xmath1 ; 2 .
each maximal simplex of @xmath1 is contained in only one of the elements of the covering .
let @xmath65 be a categorical covering .
we shall replace each subcomplex @xmath66 by another subcomplex @xmath67 ( may be empty ) defined as the union of the simplices @xmath68 which are maximal in @xmath1 .
first , @xmath69 is a covering of @xmath1 , because if @xmath70 is a simplex of @xmath1 , it must be contained in some maximal simplex @xmath71 , which in turns is contained in some @xmath66 .
then @xmath72 , so @xmath73 . moreover , @xmath67 is categorical ( by lemma [ subcat ] ) , because @xmath74 .
the second part follows from the fact that if we suppress each maximal simplex from all excepting one of the @xmath67 , then , the resulting subcomplexes are still categorical , by lemma [ subcat ] , and they cover @xmath1 . [
[ properties ] ] we state several general properties of simplicial ls - category . in @xcite , see also @xcite , barmak and minian introduced the notion of _ strong collapse _ , a particular type of simple collapse which is specially adapted to the simplicial structure .
actually , it can be modelled as a simplicial map , in contrast with the standard concept of collapse .
a vertex @xmath75 of a simplicial complex @xmath1 is _ dominated _ by another vertex @xmath76 if every maximal simplex that contains @xmath75 also contains @xmath76 .
equivalently , the link of @xmath75 is a simplicial cone with vertex @xmath76 . an _ elementary strong collapse _ consists of removing the open star of a dominated vertex @xmath75 from a simplicial complex @xmath1 . the inverse of a strong collapse is called a strong expansion . then , two simplicial complexes @xmath77 have the same _ strong homotopy type _ , denoted by @xmath78 , if they are related by a sequence of strong collapses and expansions .
surprisingly , this turns out to be intimately related to the classical notion of contiguity .
more precisely , having the same strong homotopy type is equivalent to the existence of simplicial maps @xmath79 and @xmath80 such that @xmath81 and @xmath82 ( * ? ? ?
* corollary 2.12 ) .
the strong homotopy type gives a simplicial analogue to the homotopy type of topological spaces .
notice that @xmath56 if and only if @xmath1 is _ strongly collapsible _ , that is , there is a finite sequence of elementary strong collapses reducing it to a vertex .
the simplicial complex @xmath83 in figure [ examplecone ] is not strong collapsible , in fact @xmath84 . with @xmath84,height=94 ]
let @xmath1 be a finite simplicial complex .
the cone over @xmath1 , @xmath85 is strong collapsible .
that is , @xmath86 .
this result can be proven just by taking into account that @xmath87 @xcite .
let @xmath1 be a finite simplicial complex such that @xmath88 , and @xmath89 the complex given by only two @xmath90- simplices and no other simplices .
we define the suspension of @xmath1 as @xmath91
. then @xmath92 .
next theorem was proved in ( * ? ? ?
* theorem 3.4 ) .
the simplicial ls - category is an invariant of the strong homotopy type , that is , @xmath78 implies @xmath93 .
[ score ] therefore a simplicial complex @xmath1 and its _ core _ @xmath7 have the same simplicial ls - category .
remember that the core of a complex is the minimal subcomplex obtained by eliminating dominated vertices ( see for instance barmak s book @xcite ) .
more precisely , if a vertex @xmath75 is dominated by another vertex @xmath76 , then the collapse @xmath94 is a strong equivalence . under a finite number of steps one attains a complex @xmath7 which is minimal ,
that is , it does not have dominated vertices .
this minimal complex is unique up to simplicial isomorphism .
next result establishes two combinatorial upper bounds for the simplicial category which do not exist for the usual ls - category .
[ bounds ] @xmath0 is strictly bounded from above by the number of vertices and the number of maximal simplices of its core @xmath7 .
the star of a vertex @xmath95 , @xmath96 , is a strong collapsible subcomplex of @xmath7 because all vertices in @xmath95 are dominated by @xmath75 .
therefore , the family @xmath97 provides a cover of @xmath7 by categorical subcomplexes , and so we have that @xmath98 , where @xmath99 is the number of vertices of @xmath7 . since the value of @xmath0 is a strong homotopy invariant , @xmath100 . on the other hand ,
let @xmath101 be the number of maximal simplices of @xmath7 .
then , from proposition [ max ] it follows that @xmath102 because any maximal simplex is a strong collapsible subcomplex of @xmath7 .
[ [ geomcat ] ] in @xcite , the authors also introduced the notion of _ geometric _ simplicial category @xmath103 of the complex @xmath1 , which is the analogue of the geometric ls - category of a topological space introduced by fox in @xcite , see also @xcite .
the difference with @xmath0 is that each subcomplex in a categorical covering is required to be strongly collapsible in itself , rather than in the ambient complex ; in other words , the identity , rather than the inclusion , is in the contiguity class of a constant map .
clearly , @xmath104 .
geometric simplicial category has a very different behaviour from that of the ordinary category .
in particular it is neither hereditary nor homotopically invariant .
however , the bounds of corollary [ bounds ] are still true for @xmath103 because @xmath105 , as proven in @xcite .
let @xmath1 and @xmath28 be two finite simplicial complexes .
we define the join @xmath106 as the simplicial complex with set of vertices @xmath107 and with simplices the simplices of @xmath1 , the simplices of @xmath28 and the simplices given by @xmath108 , @xmath71 simplex in k and @xmath109 simplex in l. let @xmath1 and @xmath28 be two finite simplicial complexes , then @xmath110 .
suppose that @xmath111 , then there are @xmath53 strong collapsible subcomplexes @xmath112 .
consider the subcomplexes @xmath113 covering @xmath1 .
they are strong collapsible and they provide a cover of @xmath106 .
[ [ section-2 ] ] we now study the behaviour of @xmath6 under barycentric subdivisions . our main result states that @xmath6 is decreasing under such kind of subdivisions .
[ subdiv]let @xmath114 be the first barycentric subdivision of @xmath1
. then @xmath115 .
this theorem was proved in ( * ? ? ?
6.7 ) using results of barmak and minian ( precisely , ( * ? ? ?
* prop . 4.11 and prop . 4.12 ) ) about finite spaces .
we shall reformulate them in order to give a direct proof .
if @xmath1 is an abstract simplicial complex , its first barycentric subdivision can be defined formally as the complex @xmath114 whose vertices @xmath116 are identified to the simplices @xmath36 of @xmath1 , while the simplices of @xmath114 are the sequences @xmath117 of simplices of @xmath1 such that @xmath118 ( see @xcite ) .
let @xmath119 be a simplicial map .
the induced map @xmath120 is defined as @xmath121 clearly @xmath11 is a simplicial map , @xmath122 and @xmath123 .
[ sdcont]if the simplicial maps @xmath38 are in the same contiguity class , @xmath124 , then @xmath125 are in the same contiguity class , @xmath126 . without loss of generality
, we may assume that the maps @xmath9 and @xmath10 are contiguous , @xmath127 , which means that @xmath128 is a simplex , for any simplex @xmath129 .
let @xmath130 be the map given by @xmath131 we shall prove that @xmath132 by increasing step by step the size of the set @xmath133 note that @xmath134 , while @xmath135 .
if @xmath136 there is nothing to prove .
otherwise , there exists @xmath137 such that @xmath138 , or equivalently , @xmath139 is strictly contained in @xmath140 .
let us take @xmath70 to be of maximal dimension with this property ; in this way @xmath141 when @xmath70 is a proper face of @xmath71 .
now , we can define a new map @xmath142 as @xmath143 it follows : \(1 ) the map @xmath144 is simplicial .
in fact , if @xmath117 is a simplex of @xmath114 , then @xmath145 if @xmath146 while @xmath147 if @xmath148 in both cases the image is a simplex of @xmath149 . note that in the second case , by the maximality of @xmath70 cited above , it follows that @xmath150 if @xmath151 .
\(2 ) we have @xmath152 , because if @xmath153 then @xmath154 which equals the simplex @xmath155 where @xmath156 is the lowest index such that @xmath157 , if such a @xmath156 exists .
\(3 ) finally , @xmath158 , by the definition of @xmath144 .
by repeating this construction we shall obtain a sequence of contiguous maps @xmath159 which shows that @xmath132 . using the same argument for @xmath10
, we can prove that @xmath160 .
thus @xmath161 , as claimed .
however , subdivision does not preserve the strong homotopy type , as shown by the following example taken from ( * ? ? ?
* example 5.1.13 ) . consider @xmath1 to be the boundary of a @xmath162-simplex and @xmath114 its barycentric subdivision as in figure [ erica3 ] .
they are both minimal complexes because they have no dominated vertices , but they are not isomorphic , therefore they do not have the same strong homotopy type ( see lemma [ minid ] ) .
that has not the same strong homotopy type of its subdivision @xmath114.,height=94 ] let @xmath163 , and take a categorical covering @xmath54 of @xmath1 .
consider the subcomplexes @xmath164 , which cover @xmath114 .
since each inclusion @xmath165 is in the contiguity class of some constant map @xmath166 , denoted by @xmath167 , it follows from proposition [ sdcont ] that @xmath168 .
but it is clear that @xmath169 is the inclusion @xmath170 , while @xmath171 is the constant map @xmath172 .
then each @xmath173 is a categorical subcomplex and @xmath174 .
[ sd1 ] let @xmath1 be a finite simplicial complex and let @xmath114 be the first barycentric subdivision .
then , @xmath84 implies that @xmath175 . in ( * ? ? ?
* theorem 4.15 ) it is proved that a complex @xmath1 is strongly collapsible if and only if @xmath114 is strongly collapsible . in other words , @xmath56 if and only if @xmath176 . jointly with our theorem [ subdiv ] this ends the proof .
[ subdivgraph ] the following example , suggested to the third author by j. barmak , shows a complex where the inequality of proposition [ subdiv ] is strict .
let @xmath1 be the complete graph @xmath177 ( see figure [ k5 ] ) considered as a @xmath178-dimensional simplicial complex .
let us consider the following categorical cover of @xmath1 : @xmath179 verifies @xmath180.,height=188 ] therefore @xmath181 .
moreover , there is no cover of two categorical subcomplexes .
in fact , if it happens , then one of the two subcomplexes has to contain at least @xmath182 edges .
moreover , since any tree contains one more vertex than the number of edges , a forest contains more vertices than edges .
hence , any forest with @xmath182 edges should have at least @xmath183 vertices , but it is impossible in our complex .
we conclude that there is not any categorical subcomplex with at least five vertices , therefore there is no covering of @xmath177 given by two categorical subcomplexes .
hence we have @xmath184 . on the other hand , the first barycentric subdivision @xmath114 has a covering with two categorical subcomplexes given , for example , by the two subcomplexes that are showed in figure [ sdk5 ] : @xmath185 , drawn with continuous edges , and @xmath186 , drawn with dashed edges .
since @xmath187 is not strongly collapsible we can conclude that @xmath188 . with two categorical subcomplexes.,height=188 ]
note that @xmath189 equals the _ arboricity _ of @xmath190 minus one ; later we shall prove that this is a general result for any graph ( theorem [ arbcat ] ) .
also , we shall show later that for graphs the simplicial category equals the topological category of the geometric realization ( corollary [ aborcat ] ) .
it is interesting to point out that theorem [ subdiv ] can be extended to the geometrical simplicial category , that is , @xmath191 with an analogous argument .
in addition , notice that the complex @xmath192 ( figure [ sdk5 ] ) can be covered with two trees and thus , @xmath193 .
[ [ section-3 ] ] a natural question is to compare the simplicial category of the complex @xmath1 with the ls - category of @xmath194 , the so - called geometric realization of @xmath1 @xcite .
we recall the classical definition of lusternik - schnirelmann category of a topological space . for general properties of this invariant
we refer to @xcite .
[ catop ] an open subset @xmath51 of the topological space @xmath18 is called _ categorical _ if @xmath51 can be contracted to a point inside the ambient space @xmath18 .
the _ ls - category _ of @xmath18 , denoted by @xmath195 , is the least integer @xmath52 such that there is a covering of @xmath18 by @xmath53 categorical open subsets .
it is known ( see proposition 1.10 of @xcite ) that , when the space @xmath18 is a normal anr , the categorical sets in definition [ catop ] can be taken to be _ closed _ instead of open . in particular , this is the case for the geometric realization @xmath194 of a finite abstract simplicial complex ( see ( * ? ? ?
* ii.4 ) , also @xcite and the references and comments in @xcite ) . on the other hand ,
o. randal - wallis pointed out to the second author the following example , showing that categorical sets can be rather pathological .
let @xmath196 $ ] be the unit interval and let @xmath197 be the cantor set .
then @xmath197 is contractible in @xmath18 and has not the homotopy type of a finite cw - complex .
this is because @xmath197 is totally disconnected but non - discrete ( see @xcite ) .
anyway , the following theorem shows that , when @xmath194 is the geometric realization of an abstract simplicial complex @xmath1 , the ls category of @xmath18 can be computed by means of a closed categorical covering whose sets are subcomplexes of @xmath1 in a certain subdivision .
this result is essentially stated , but without proof , in fox s paper @xcite .
[ realization]if @xmath194 then @xmath198 if and only if there exist subcomplexes @xmath199 of some subdivision @xmath200 of @xmath1 , such that each @xmath201 is contractible in @xmath18 and @xmath202 .
the proof of the above theorem [ realization ] can be sketched as follows : if we take @xmath203 , the lebesgue number of the considered categorical covering and we consider a generalized subdivision of @xmath1 , @xmath114 , such that @xmath204 .
then , by considering the subcomplexes obtained with simplices contained on each open element of this covering , a categorical covering by subcomplexes is obtained .
[ [ section-4 ] ] we now state the precise relation between the simplicial category of a complex and the topological ls - category of its geometric realization .
let @xmath1 be a finite simplicial complex
. then @xmath13 .
let @xmath163 and let @xmath205 be a categorical simplicial covering of @xmath1 , that is , each inclusion @xmath206 is in the same contiguity class that a constant map @xmath207 .
therefore , the induced maps @xmath208 between the geometric realizations are homotopic , @xmath209 @xcite .
it is clear that @xmath210 is the inclusion @xmath211 , and @xmath212 is a constant map .
thefore the subspaces @xmath213 form a categorical closed cover of @xmath2 . as we commented before , since @xmath2 is a normal anr we can consider closed covers instead of open covers .
then @xmath214 .
[ sd ] @xmath215 , for any iterated barycentric subdivision @xmath216 of @xmath1 . the geometric realizations @xmath217 and @xmath2 are homeomorphic ( see for instance ( * ? ? ?
* prop . 2.33 ) ) .
the inequality in corollary [ sd ] may be strict .
for instance , the complex @xmath1 in figure [ mother ] is topologically contractible ( it is collapsible ) , that is @xmath218 , but all the barycentric subdivisions have simplicial category one .
this happens because , as proven by barmak and minian in ( * ? ? ?
* theorem 4.15 ) , a complex is strongly collapsible if and only if its barycentric subdivision is strongly collapsible . in other words , @xmath219 if and only if @xmath220 . but
the complex @xmath1 is not strongly collapsible ( it has no dominated vertices ) and can be covered by two categorical subcomplexes , so @xmath221 . by applying theorem [ subdiv ]
we have @xmath222 for all @xmath223 .
[ [ prelimprod ] ] we recall the definition of categorical product ( see remark [ catvscart ] below for an explanation of our notations ) . [ prod ] let @xmath77 be two abstract simplicial complexes . the _ categorical product _
@xmath16 is defined as follows .
the vertices of @xmath16 are the pairs @xmath224 of vertices with @xmath47 and @xmath225 .
the simplices of @xmath16 are the sets of vertices @xmath226 such that @xmath227 is a simplex of @xmath1 and @xmath228 is a simplex of @xmath28 .
we fix some notation . for @xmath229
, we shall denote @xmath230 . by definition ,
the projections @xmath231 , @xmath232 , onto each factor are simplicial maps .
the diagonal map @xmath233 defined by @xmath234 is simplicial too .
a map @xmath235 is simplicial if and only if the compositions with all the projections are simplicial maps .
if @xmath236 are simplicial maps , the map @xmath237 defined by @xmath238 is simplicial .
if @xmath79 is a simplicial map then we denote by @xmath239 the map @xmath240 which is simplicial .
[ catvscart ] our definition [ prod ] is like definition 4.25 in @xcite but we changed the notation @xmath241 to @xmath16 for the sake of simplicity .
as an example , @xmath242 .
note that @xmath243 is not homeomorphic to @xmath244 .
hence the categorical ( or simplicial ) product should not be confused with the so - called cartesian product ( * ? ? ? *
ii.8 ) , which depends on some ordering of the vertices and does not verify the universal property of a product .
however , as kozlov proves in ( * ? ? ?
15.23 ) , the geometric realizations of both products have the same homotopy type , then @xmath245 as topological spaces ( see also ( * ? ? ?
* theorem 10.21 ) ) . [ stuff ] 1 .
let @xmath38 and @xmath45 be simplicial maps such that @xmath127 and @xmath246 . then @xmath247 2 .
let @xmath248 and @xmath249 , then @xmath250 [ contprod ] let @xmath78 be two complexes with the same strong homotopy type .
then @xmath251 .
the following theorem establishes the simplicial analogue of a well known result on the ls - category of a product of topological spaces .
let @xmath1 and @xmath28 be finite simplicial complexes
. then @xmath252 suppose that @xmath163 and @xmath253 , therefore there exists a categorical covering @xmath254 of @xmath1 and a categorical covering @xmath255 of @xmath28 .
consider the subcomplexes @xmath256 , for @xmath257 and @xmath258 .
we want to show that @xmath259 form a categorical covering of @xmath16 .
each inclusion map @xmath260 is in the same contiguity class of a constant map @xmath261 where @xmath262 is a vertex of @xmath1 ; analogously , each inclusion @xmath263 is in the same contiguity class of a constant map @xmath264 where @xmath207 is a vertex in @xmath28 . by proposition
[ stuff ] the map @xmath265 is in the same contiguity class of the map @xmath266 . clearly , @xmath267 and @xmath268 , where @xmath269 is a vertex of @xmath270
. therefore the subcomplexes @xmath259 are categorical .
now , we shall prove that @xmath271 is a covering of @xmath16 .
if the simplex @xmath272 is in @xmath16 then @xmath273 is contained in a subcomplex @xmath274 of @xmath1 and @xmath275 is contained in a subcomplex @xmath276 of @xmath28 .
then @xmath272 is contained in @xmath259 .
thus , we conclude that @xmath277 .
[ [ section-5 ] ] it is well known that for topological spaces @xmath18 with `` good properties '' there is the following so - called whitehead characterization of the topological ls - category ( see ( * ? ? ? * theorem 1.55 ) ) : @xmath198 if and only if the diagonal map @xmath20 factors ( up to homotopy ) through the so - called _ fat wedge _ @xmath21 .
this result is a very useful tool for computing ls - category . in this section
we shall try to adapt it to abstract simplicial complexes .
first we define a simplicial version of the topological fat wedge @xmath278 of a topological space @xmath18 @xcite .
also , we shall briefly develop the notion of _ pointed _ contiguity class .
let @xmath1 be an abstract simplicial complex and fix some vertex @xmath279 of @xmath1 as a base point .
for each @xmath232 let @xmath280 be the subcomplex of @xmath281 spanned by the vertices whose @xmath156-th coordinate is the base point @xmath279 . for @xmath229
the _ @xmath22-th fat wedge _
@xmath23 is the subcomplex @xmath282 .
for instance , @xmath283 is a point and @xmath284 is the wedge @xmath285 .
note that @xmath23 is not a full subcomplex of @xmath281 . a pointed map ( that is , a simplicial map preserving the base points ) @xmath286 induces a simplicial map @xmath287 , which is the restriction of @xmath288 .
let @xmath38 be two contiguous simplicial maps preserving the base points .
then the induced maps @xmath289 are contiguous .
the maps @xmath290 are contiguous by proposition [ stuff ] .
moreover they send each subcomplex @xmath291 into itself .
[ invpoint]let @xmath292 be two pointed simplicial complexes with the same pointed strong homotopy type .
that is , we assume that the homotopy equivalences @xmath293 between @xmath1 and @xmath28 , as well as the sequences of contiguous maps defining the relations @xmath81 and @xmath82 , preserve the base points . then @xmath294 . [
[ section-6 ] ] we are now in position to discuss the whitehead formulation of the simplicial ls - category . in order to be systematic we follow the approach of @xcite , by defining a so - called _ simplicial whitehead category _ @xmath24 and trying to compare it to the simplicial ls - category @xmath0 .
[ white]we say that @xmath295 if the diagonal map @xmath296 factors through the fat wedge @xmath297 up to contiguity class . that is , there exists some simplicial map @xmath298 such that @xmath299 , where we denote by @xmath300 the inclusion ( see figure [ whitehead ] ) . @xmath301^{\ i\ } & \
k^{n+1}\\ k\ar@{-->}[u]^{\delta}\ar[ru]_{\delta}&}$ ] [ ineq ] @xmath25 .
assume that @xmath302 , and let @xmath298 be as in definition [ white ] .
let @xmath303 be the projections onto each factor .
let @xmath304 , @xmath305 , be the subcomplexes defined in ( [ subcom ] ) . since @xmath299
, there exists a sequence @xmath306 of maps such that @xmath307 and @xmath308 are contiguous .
call @xmath309 the preimages of the @xmath310 defined above .
clearly , @xmath311 .
it only remains to show that each subcomplex @xmath312 is categorical , that is , each inclusion map @xmath313 is in the same contiguity class that a constant map .
let us prove it : since @xmath314 it follows that @xmath315 now , @xmath316 , which is the constant map @xmath317 because @xmath318 . on the other hand , @xmath319 is the inclusion @xmath320
. then @xmath321 .
we have found @xmath53 categorical subcomplexes covering @xmath1 , therefore @xmath322 .
[ [ section-7 ] ] in this section we shall prove that the converse inequality of theorem [ ineq ] is not true , by exhibiting an example of a complex @xmath1 such that @xmath84 while @xmath323 .
the proof of the next propositions is inspired by a result about finite co - h - spaces proved in @xcite .
we thank j. oprea for pointing out this reference to us and r.d .
helmstutler for some explanations about his paper .
first we prove that the simplicial whitehead category is an invariant of the pointed strong homotopy type . if @xmath324 is a pointed strong equivalence as in corollary [ invpoint ] , then @xmath325 .
let @xmath326 .
consider the diagram in figure [ whiteheadkl ] and the strong equivalences @xmath327 ( corollary [ contprod ] ) and @xmath328 ( corollary [ invpoint ] ) .
@xmath329^{\varphi^n}&\ l^n\ar@<+1ex>[l]^{\psi^n}&&\\ & & & \\ { \ t^{n+1}k\ } \ar@<+1ex>[ruu]^{\ i_k}\ar@<+1ex>[r]^{t^{n+1}\varphi}&{\ t^{n+1}l\ } \ar@<+1ex>[l]^{t^{n+1}\psi } \ar[ruu]^{\ i_l}&&\\ & & & \\
k\ar@{-->}[uu]^{\delta_k}\ar[ruuuu]_(0.3){\ \delta_k}|{{\begin{matrix}&&\cr&&\cr & & \cr&&\cr \end{matrix}}}\ar@<+1ex>[r]^{\varphi}&l\ar@<+1ex>[l]^\psi\ar[uu]^{\delta_l}\ar@<-1ex>[ruuuu]_(0.3){\delta_l\ } & & } $ ] with the obvious notations , from the relations @xmath330 it follows that @xmath331 as well as @xmath332 moreover , from the definitions , we have @xmath333 as well as @xmath334 define @xmath335 since , from hypothesis , @xmath336 , it follows @xmath337 then @xmath338 .
the other inequality is proved in the same way .
clearly , we may assume that the base point @xmath279 is in the core @xmath7 , hence eliminating dominated vertices is a pointed equivalence @xmath339 .
[ core]the simplicial whitehead category of a complex equals that of its core , @xmath340 .
we are now in a position to prove the main results of this section .
2.7)[minid ] let @xmath7 be a minimal complex and let @xmath341 be a simplicial map which lies in the same contiguity class as the identity .
then @xmath9 is the identity .
[ onezero]let @xmath1 be a simplicial complex such that @xmath342 .
then @xmath1 is strongly collapsible , which is equivalent to @xmath56 .
if @xmath343 then the result follows from theorem [ ineq ] . if @xmath344 we have , from corollary [ core ] , that @xmath345
this means that there exists a simplicial map @xmath346 such that @xmath347 , where @xmath348 is the diagonal map . let @xmath349 be the projections of the categorical product .
then @xmath350 because @xmath7 is a minimal complex ( lemma [ minid ] ) .
remember that in the proof of theorem [ ineq ] we denoted by @xmath351 ( respectively , @xmath352 ) the subcomplex @xmath353 ( resp .
@xmath354).then , for @xmath355 we have @xmath356 because @xmath357 , which shows that @xmath358 .
analogously @xmath359 proves that @xmath360 .
then the core @xmath361 is a point , that is , the complex @xmath1 has the strong homotopy type of a point .
the complex @xmath1 in figure [ mother ] has @xmath84 so it is not strongly equivalent to a point . from theorem [ onezero ] its whitehead simplicial
category is at least @xmath162
. then @xmath362 .
[ [ section-8 ] ] we now briefly discuss the notion of cofibration in the simplicial setting .
the `` homotopy extension property '' is a very important notion in topology .
a cofibration is a map @xmath363 which satisfies the homotopy extension property with respect to all spaces .
it is well known ( see @xcite or @xcite ) that if @xmath26 is a subcomplex of a cw - complex @xmath18 then the pair @xmath27 has the homotopy extension property .
therefore , if @xmath364 is a subcomplex of a simplicial complex , the pair @xmath29 of the geometric realizations has the ( topological ) homotopy extension property .
we want to define a purely combinatorial analogue for a simplicial pair @xmath30 .
a simplicial map @xmath365 has the _ contiguity extension property _ if given two simplicial maps @xmath366 which lie in the same contiguity class , @xmath124 , and given an extension of @xmath9 ( that is , a simplicial map @xmath367 such that @xmath368 ) , there exists an extension @xmath369 of @xmath10 with @xmath370 ( cf .
figure [ fig ] ) .
@xmath371_i\ar@<+1ex>[rr]^{\ \varphi\ } \ar@<-1ex>[rr]_{\ \psi\ } & & \ m\\ & & \\
k\ar@<-1ex>[rruu]^{\ \tilde\varphi}\ar@{-->}@<-3ex>[rruu]_{\tilde\psi\ } & & } $ ] let @xmath7 be the core of the complex @xmath1 .
then the inclusion @xmath372 is a simplicial cofibration .
in fact , as explained before , there is a simplicial retraction @xmath373 such that @xmath374 and @xmath375
. then one can take @xmath376 because @xmath377 . as an application of theorem [ onezero ]
, we shall show that there are simplicial pairs @xmath30 that do not have the extension property .
let @xmath1 be a connected simplicial complex with @xmath163 .
assume that there is a categorical covering @xmath378 of @xmath1 such that all the pairs @xmath379 have the contiguity extension property
. then @xmath302 .
due to theorem [ ineq ] we only have to prove that @xmath380 . by hypothesis ,
each inclusion @xmath381 is in the same contiguity class that some constant map , whose image is a vertex @xmath207 .
since two such constant maps are contiguous if and only if the vertices @xmath382 lie on the same simplex ( lemma [ twopoints ] ) , and @xmath1 is connected , we can suppose that all the vertices are equal , say to some base point @xmath279 .
taking @xmath383 , @xmath384 , @xmath385 and @xmath386 , the simplicial extension property gives maps @xmath387 such that @xmath388 .
define the map @xmath389 it is well defined , because each vertex @xmath75 is contained in some @xmath390 , hence @xmath391 , meaning that the @xmath156-th coordinate of @xmath392 is the base point .
moreover @xmath203 is a simplicial map .
in fact , for any simplex @xmath129 we have that @xmath393 is a simplex in @xmath281 ( proposition [ prelimprod ] ) .
moreover , @xmath71 must be contained in some @xmath390 , and @xmath394 .
then @xmath395 .
now , let @xmath396 be the sequence of contiguous maps connecting @xmath397 and @xmath398 ( clearly we may assume that the length @xmath99 does not depend on @xmath156 ) .
then the maps @xmath399 define a sequence of contiguous maps between @xmath400 and @xmath401 .
hence @xmath402 , which means that @xmath295 . as a corollary ,
some categorical covering of the complex @xmath1 in figure [ mother ] fails to verify the hypothesis of the preceding theorem , because @xmath84 but @xmath323 .
the subcomplex @xmath403 , of the complex @xmath1 in figure [ mother ] , where @xmath404 , has not the contiguity extension property .
let @xmath75 be the upper vertex of @xmath1 ( see figure [ mother ] ) . since @xmath28 is strongly collapsible to @xmath75 , the inclusion @xmath405 and the constant map @xmath406 lie in the same contiguity class .
fix @xmath407 to be the identity .
if there exists @xmath408 such that @xmath409 there must be some sequence @xmath410 of contiguous maps .
however , since @xmath411 , the simplicial map @xmath412 must be constant , and we obtain that @xmath398 is contiguous to a constant map , which is a contradiction because @xmath413 . .
the subcomplex @xmath414 has not the contiguity extension property .
, height=151 ] a simpler example was communicated to us by n. scoville : let @xmath1 be the @xmath178-dimensional complex in figure [ triang ] , with vertices @xmath415 and let @xmath71 be the @xmath178-simplex joining the vertices @xmath416 and @xmath417 .
then the subcomplex @xmath418 has not the contiguity extension property .
has not the contiguity extension property .
, height=94 ]
[ [ section-9 ] ] this section is focused on the study of the simplicial ls - category in the one - dimensional case , that is , on graphs . the well known graph - theoretical notion of _ arboricity
_ will play a central role in this study .
basically , arboricity is based on the cardinality of minimal decompositions of a graph into disjoint spanning forests , i.e. , acyclic subgraphs , which are non - necessarily connected and cover all the vertices .
the aim of this section is to prove that arboricity coincides ( up to one ) with both simplicial and geometric simplicial categories . in @xcite ,
aaronson and scoville introduced a so - called discrete ls - category in the simplicial seting .
they proved that , for the @xmath178-dimensional case , it is equivalent to arboricity .
let us start with some basic notions on graph theory .
a general reference is harary s book @xcite .
let @xmath32 be a graph .
a cycle in @xmath32 is an alternating sequence of distinct vertices and edges , @xmath419 , where the incident vertices of each edge @xmath420 are @xmath421 and @xmath422 respectively , and such that @xmath423 . under a topological point of view , cycles are triangulations of the circumference @xmath424 . a _ forest _ is a graph without cycles , alternatively it can be called acyclic graph .
a _ tree _ is a connected forest .
the _ arboricity _ of a graph @xmath32 , denoted by @xmath31 , is the minimum number of edge - disjoint spanning forests into which @xmath32 can be decomposed .
nash - williams @xcite determined the arboricity of a general graph : ( * ? ? ?
9.10 ) let @xmath32 be a nontrivial graph and let @xmath425 be the maximum number of edges in any subgraph of @xmath32 with @xmath22 vertices
. then @xmath426 for the particular case of complete graphs ( see figure [ k5 ] ) it follows the following formula : @xmath427 [ [ section-10 ] ] now we prove some results which will be used later in this section .
every standard elementary collapse in a graph is the deletion of a so - called _ leaf vertex _ @xmath75 and the unique edge @xmath428 containing @xmath75 .
thus @xmath75 is dominated by @xmath429 and hence , in graphs , every elementary collapse is an elementary strong collapse .
[ cycle ] let @xmath32 be a connected graph and let @xmath430 be a subgraph containing at least one cycle @xmath431 . if @xmath432 is a simplicial map contiguous to the inclusion @xmath433 , then @xmath434 contains the cycle @xmath431 .
moreover @xmath435 and @xmath436 is the inclusion @xmath437 .
every edge @xmath438 in @xmath28 satisfies that @xmath439 is a simplex in @xmath32 ( for graphs , `` simplex '' means a vertex or an edge ) , so , equivalently @xmath440 ; therefore , either @xmath441 or @xmath442 is one of the extreme vertices of @xmath438 .
this implies that @xmath443 .
let @xmath431 be a cycle contained in @xmath28 and let us consider the restrictions to @xmath431 of @xmath9 and the inclusion @xmath444 , denoted by @xmath445 and @xmath446 respectively , which are also contiguous as consequence of the fact that the composition of contiguous maps is contiguous ( see @xcite ) .
so , if @xmath9 maps every edge of the cycle @xmath431 onto itself then @xmath447 and hence @xmath434 contains the cycle @xmath431 .
otherwise , there exists an edge @xmath448 in @xmath431 such that @xmath449 ; we can suppose without loss of generality that @xmath450 .
now let us consider the edge @xmath451 which is adjacent to @xmath448 in @xmath452 . since the map @xmath9 is simplicial , we have @xmath453 , which is a contradiction with @xmath454 . finally , we conclude that all the edges in @xmath431 remain fixed by @xmath9 so @xmath434 contains at least one cycle .
let @xmath455 be a a path in a graph , such that there is a simplicial map @xmath456 contiguous to the inclusion @xmath457 .
by the same argument of the proof above , any edge @xmath438 contained in @xmath458 satisfies that either @xmath441 or @xmath459 , where @xmath460 is one of the extreme vertices of @xmath438 with degree @xmath178 , i.e. @xmath460 is a so called leaf vertex of @xmath458 .
thus , we conclude that the only possible reductions induced by a simplicial map contiguous to the inclusion are given by standard collapses .
the next result establishes the equivalence for graphs between the notions of categorical subcomplex and acyclic subgraph .
[ catfor ] let @xmath32 be a connected graph and let @xmath461 be a subgraph .
then @xmath28 is categorical in @xmath32 if and only if @xmath28 is a forest .
let us suppose that there exists a categorical but non - acyclic subgraph @xmath430 . by definition , there exists a vertex @xmath462 such that the inclusion @xmath463 and the constant map @xmath464 are in the same contiguity class , which gives a sequence @xmath465 of directly contiguous maps @xmath466 .
now , since @xmath467 and @xmath28 contains at least one cycle @xmath431 , by taking into account the previous lemma we conclude that @xmath468 contains at least the cycle @xmath431 .
moreover , @xmath469 is the inclusion @xmath470 .
now , @xmath471 implies , by composing with the inclusion @xmath472 , that @xmath473 .
that means , by applying the lemma again , that @xmath474 is the cycle @xmath431 , and by repeating the argument we shall arrive to a cycle that can be deformed into a point , which is impossible .
hence we conclude that @xmath28 is an acyclic subgraph of @xmath32 , that is , @xmath28 is a forest .
conversely , let us assume that @xmath28 is a forest , so @xmath28 is a disjoint union of trees @xmath475 with @xmath476 .
it is clear that each inclusion map @xmath477 is in the same contiguity class as the constant map sending the tree @xmath475 onto one of its vertices @xmath422 .
since @xmath32 is connected , there is at least one path in @xmath32 joining every vertex @xmath422 with a given vertex @xmath279 .
then the inclusion of @xmath28 is in the same contiguity class as the constant map @xmath75 .
hence @xmath28 is a categorical subcomplex in @xmath32 .
[ arbol ] let @xmath32 be a connected graph .
for every covering of @xmath32 by forests there exists a covering of @xmath32 by trees with the same number of elements .
let @xmath478 be a covering of @xmath32 by forests .
since @xmath32 is connected , it follows that given two trees @xmath479 and @xmath480 in a forest @xmath481 and any vertices @xmath482 and @xmath483 , they can be linked by means of a path in @xmath32 . by adding such a path to @xmath481 , and removing , if necessary , the edges of the path which create cycles and are not contained in @xmath481 , we link the trees @xmath479 and @xmath480 of @xmath481 ( see figure [ bosque ] ) . repeating the argument for the remaining trees , we obtain one tree @xmath475 containing @xmath481 . finally , carrying out the same procedure on every forest of @xmath32
, we obtain a covering of @xmath32 by trees @xmath484 .
[ [ section-11 ] ] we now state the main results of this section . [ arbcat ] let @xmath32 be a connected graph
. then @xmath485 .
let us suppose that @xmath486 , so by definition there exists a covering of @xmath32 with @xmath487 edge - disjoint spanning forests . by means of lemma [ arbol ] we can construct a covering of @xmath32 with @xmath487 trees . since the trees are strongly collapsible we conclude that @xmath488 .
conversely , let us assume that @xmath489 .
it means that there is a covering of @xmath32 with @xmath487 strongly collapsible subsomplexes . by theorem [ catfor ] , these complexes are trees @xmath490 .
starting from a tree @xmath475 , we can obtain a spanning forest @xmath481 by adding all the isolated vertices which are not covered by @xmath475 .
next , if an edge is contained in several forests , in order to obtain a covering by edge - disjoint forests , we remove it from all these forests but one .
hence @xmath491 .
[ aborcat]for any graph @xmath32 , we have @xmath492 .
apply again lemma [ cycle ] and theorem [ catfor ] separately on each connected component of @xmath32 .
[ [ section-12 ] ] in section 2.4 , the behaviour of @xmath6 under barycentric subdivisions was studied .
actually , it was proved that @xmath115 . in the one - dimensional case
there are examples where this inequality is strict ( see example [ subdivgraph ] ) .
also there are examples of @xmath162-dimensional complexes where @xmath493 remains constant for all @xmath223 and is strictly greater than the topological ls - category of the geometric realization @xmath4 ( see example [ subdivgraph ] ) .
in contrast , in the one - dimensional case it holds the geometric category of the first barycentric subdivision always reaches the topological category of the geometric realization .
in fact , the following result states that this is possible for a certain kind of `` local '' barycentric subdivision . remember that any finite cw complex @xmath18 satifies @xmath494 , therefore @xmath495 for a non - contractible graph .
let @xmath32 be a connected graph . if @xmath32 is a tree then @xmath496 . otherwise , let @xmath497 be the subdivision obtained from @xmath32 by only bisecting those edges out of a spanning tree in @xmath32
. then @xmath498 .
let us consider a spanning tree @xmath479 in @xmath32 .
it is well known ( * ? ? ?
1a.2 ) that there is a bijection between the edges out of a spanning tree and the basic cycles generating the one - dimensional homology of a graph . now ,
after every edge out of @xmath479 is barycentrically subdivided , the subdivision @xmath497 of @xmath32 can be covered by two different ( non disjoint ) spanning trees @xmath499 and @xmath500 constructed as follows ( see figure [ sdk5 ] ) : @xmath499 is an expansion of @xmath479 which is obtained by adding one edge ( one of the subdivided ones ) on every leaf vertex of @xmath479 ; analogously , @xmath500 is obtained by adding the other edge ( not previously added to construct @xmath499 ) to all the leaf vertices of @xmath479 . by definition both @xmath499 and @xmath500 are spanning trees covering @xmath497 and hence @xmath498 .
notice that for certain graphs , as the complete ones @xmath501 , the simplicial ls - category equals the topological ls - category of the realization of the graph by bisecting a fewer number of edges than the stated one by the above proposition .
for example , for @xmath190 only three edges must be bisected in order to get @xmath502 ( see figure [ tresedges ] ) .
[ geomgraph ] let @xmath32 be a connected graph .
then @xmath503 . taking into account the above results ,
it is interesting to point out that , in the one - dimensional case , the difference @xmath504 can be arbitrarily large , as can be checked by considering the complete graph @xmath501 .
we thank j. barmak for his valuable suggestions and m.j .
pereira - sez for many useful discussions . | the simplicial ls - category of a finite abstract simplicial complex is a new invariant of the strong homotopy type , defined in purely combinatorial terms .
it generalizes to arbitrary simplicial complexes the well known notion of arboricity of a graph , and allows to develop all the machinery of algebraic topology which is costumary in the classical theory of lusternik - schnirelmann category . |
the studies of chaotical dynamics of closed isotropic cosmological model has a long story .
they were initiated by papers @xcite where the possibility to avoid a singularity at the contraction stage in such a model with a minimally coupled massive scalar field was discovered .
later it was found that this model allows the existence of periodical trajectories @xcite and aperiodical infinitely bouncing trajectories having a fractal nature @xcite .
this result was reproduced in other terms in our papers @xcite . in @xcite
the set of periodical trajectories was studied from the viewpoint of dynamical chaos theory .
it was proved that the dynamics of a closed universe with a massive scalar field is chaotic and an important invariant of the chaos , the topological entropy , was calculated . on the other hand , in our paper
@xcite we found that deformations of the scalar field potential may change the dynamical regime from chaotic to regular one .
this fact poses a question : is the chaotic behaviour closely connected with a concrete form of the potential ( @xmath0 ) used in the previous papers or is it a more general phenomenon ?
a numerical analysis shows that the situation for another natural scalar field potential @xmath1 is qualitatively the same .
but most of modern scenarios based on ideas of the string theory and compactification naturally lead to another forms of potential which are exponential or behave as exponential for large @xmath2 ( see , for example , ref .
this steepness of the potential apparently changes the possibilities of escaping the singularities and alters the structure of infinitely bouncing trajectories . under some conditions , which will be studied below in detail ,
the chaotic behaviour can completely disappear .
another interesting problem concerning asymptotic regime near the singularity for such potentials was recently studied in @xcite .
the structure of the paper is the following : in sec .
2 we recall the structure of chaos in the case of a massive scalar field with a special attention paid to the statements which are valid for an arbitrary scalar field potential . in sec .
3 the conditions for the chaotic behaviour in models with exponential potentials are investigated both analytically and numerically . in sec .
4 this analysis is extended to steeper potentials .
sec . 5 provides a brief summary of the result obtained .
we shall consider a cosmological model with an action @xmath3 for a closed friedmann model with the metric @xmath4 where @xmath5 is a cosmological radius , @xmath6 a lapse function and @xmath7 is the metric of a unit 3-sphere and with homogeneous scalar field @xmath2 the action ( 1 ) takes the form @xmath8 + \frac{\dot{\varphi}^{2}}{2 n^{2 } } - v(\varphi)\right\}.\ ] ] now choosing the gauge @xmath9 we can get the following equations of motion @xmath10 @xmath11 in addition , we can write down the first integral of motion of our system @xmath12 it is easy to see from eq . ( 6 ) that the points of maximal expansion and those of minimal contraction , i.e. the points , where @xmath13 can exist only in the region where @xmath14 sometimes , the region defined by inequalities ( 7 ) is called euclidean , and the opposite region is called lorentzian .
this definition is not good enough , because for this dynamical system there are no classically forbidden regions at all ( see @xcite for detail ) , but we shall use it for brevity .
now we would like to distinguish between the points of minimal contraction where @xmath15 @xmath16 and those of maximal expansion where @xmath15 @xmath17 .
assuming @xmath18 , one can express @xmath19 from eq .
( 6 ) as @xmath20 substituting ( 8) and @xmath18 into ( 4 ) we have @xmath21 from eq .
( 9 ) one can easily see that the possible points of maximal expansion are localized inside the region @xmath22 while the possible points of minimal contraction lie outside this region ( 10 ) being at the same time inside the euclidean region ( 7 ) .
it is easy to see that the value of @xmath23 on this separating curve is @xmath24 times smaller than the corresponding @xmath23 on the euclidean boundary for a given value of the scalar field independently on the concrete form of @xmath25 . in @xcite
it was shown that this ratio remains constant for some cases in a more general situation with a non - minimally coupled scalar field . here
we would like to describe briefly the approach presented in @xcite .
the main idea consists in the fact that in the closed isotropical model with a minimally coupled scalar field satisfying the energodominance condition all the trajectories have the point of maximal expansion .
then the trajectories can be classified according to localization of their points of maximal expansion .
the points of maximal expansion are all located inside the euclidean region .
a numerical investigation shows that this area has a quasi- periodical structure .
narrow zones starting from which the trajectory has the point of bounce are separated by wide zones containing the initial conditions of trajectories falling into a singularity .
each zone of bounces contains a periodical trajectory with the point of full stop ( @xmath26 @xmath27 ) on the euclidean boundary .
then studying the substructure of these zones from the point of view of possibility to have two bounces one can see that this substructure reproduce on the qualitative level the structure of the whole region of possible points of maximal expansion .
continuing this procedure _ ad infinitum _ yields the fractal set of infinitely bouncing trajectories .
this kind of fractal , nonattracting invariant set is typical for chaotic systems without dissipation ( see , for example , @xcite ) . in the theory of dynamical chaos it is called strange repellor .
numerical investigations show also that the structure of periodical trajectories for the system eqs ( 4)- ( 6 ) have two important properties : \1 .
all the simple periodical trajectories ( i.e. having only one bounce per period ) have a full stop point on the euclidean boundary .
this property allows us to use the points of the euclidean boundary and zero velocities as an initial conditions for searching a possible strange repellors .
trajectories , going from the boundary into the euclidean region has a point of maximal expansion almost immediately , and then go towards a singularity .
so , periodical trajectory approaches their bounce point on the boundary from the lorentzian side .
the only exception is a single peculiar periodical trajectory existing in the case of nonzero cosmological constant term ( see below ) .
this two properties were found in numerical analysis and the question whether or not they are satisfied for an arbitrary scalar field potential requires more investigation .
our researches show that it is true at least for potentials steep as power - law and steeper .
if the aforesaid is satisfied , some analytical approach is possible .
indeed , periodical trajectories with a full stop point on the euclidean boundary penetrate into the lorentzian region .
that is why all such full stop points lie to the left from a critical point introduced by page @xcite in the configuration space @xmath28 .
it is a point on the euclidean boundary separating trajectories going into lorentzian and euclidean regions . by definition ,
the page s point is the point where the direction of motion at the initial moment coincides with the direction of the tangent to the curve given by equality in ( 7 ) , i.e. the point where @xmath29 using eqs .
( 4 ) , ( 5),and ( 7 ) one can find for the case @xmath0 @xmath30 except for the trivial solution @xmath31 , @xmath32 .
trajectories starting from @xmath33 and going into the euclidean region reach the point of maximal expansion almost immediately after crossing the separating curve ( see @xcite where we studied such set of maximal expansion points ) . after that point the trajectory goes towards singularity .
so it looks like a small zigzag but not a `` true '' bounce . to define a really significant bounce ,
cornish and shellard in @xcite used a condition that in the point of bounce @xmath34 in addition to @xmath35 @xmath36 .
this is just the condition that bounces lie to the left from the page point in configuration space @xmath37 .
it is also possible to use the criterion that the trajectory must return to the lorentzian region after the bounce .
qualitatively the picture of regions containing bouncing trajectories under this criterion is the same as for the former one , with the only exception that the width of the regions is somewhat smaller .
this criterion can be treated as more direct one , but it requires numerical integration while the study of page s points may be done analytically . indeed , it is easy to obtain an analog to ( 12 ) for an arbitrary potential during the same procedures .
the result is the following equation for the @xmath2-coordinate : @xmath38 it can be easily derived from eq.(13 ) that a qualitative picture of massive scalar field case do not change for any pow - law potential with the even index : there is one non - trivial page s point with the full stop points of periodical trajectories lying to the left from it on the euclidean boundary . in the next section
we will see that the situation changes significantly for the exponentially steep potentials .
we finish this section by the descriptions of another important modification of the scalar field potential introducing a constant term ( so called cosmological constant ) .
a non - zero value of the cosmological constant @xmath39 makes possible an existence of trajectories without points of maximal expansion , i.e. the trajectories which begin and end in desitter regime @xmath40 or the trajectories which begin in the singularity and end in desitter expansion or vice versa . here , in eq .
( 14 ) @xmath41 denotes hubble constant @xmath42 the chaos in the model with a massive scalar field and a cosmological constant @xmath43 was studied in @xcite . before recalling some results of this paper ,
let us look on page s points for this potential .
the equation ( 13 ) has the form @xmath44 it is easy to see from ( 16 ) that for @xmath45 there are no such points : all trajectories from the euclidean boundary go into the lorentzian region ( and then fall into desitter regime ) . for @xmath46 there are two page s points .
the left point has the same properties as for the potential without @xmath39 , while properties of the right point are inverse : the trajectories with an initial @xmath47 go into the lorentzian region while trajectories with @xmath48 ( but , of course,@xmath49 ) to the euclidean one .
a numerical analysis confirms that all trajectories with @xmath47 fall to desitter regime and can not belong to the chaotic repellor .
so , the role of the 2-nd page s point appear to be in restricting of the chaotic area in the configuration space while the existence of chaos seems to be connected with the 1-st one .
an unstable periodical trajectory , found in @xcite via numerical studies that restricts the area of bounce intervals in @xmath23 from the large values of the scale factor lies to the left from the 2-nd page s point .
it crosses the @xmath2-axis at some point with a scale factor @xmath50 .
in addition to discussion in @xcite , we give here a simple analytical estimation of the location of this periodical trajectory in the configuration space .
the typical value of @xmath50 , corresponding to this trajectory can be easily estimated for large @xmath23 . in this case
we may assume the harmonic oscillations to be a solution for the field @xmath2 dynamics : @xmath51 for calculating @xmath52 consider a point where @xmath53 .
taking the derivative with respect to time and using the constraint equation ( 6 ) , we get @xmath54 substituting these two expressions into ( 4 ) and integrating over one period @xmath55 of the scalar field oscillation , we can find the velocity @xmath56 after one oscillation : @xmath57 it vanishes for @xmath58 trajectories with @xmath59 have increasing scale factors during the scalar field oscillations .
they inevitably reach the desitter regime and can not belong to the strange repellor . as we found in @xcite , if the cosmological constant is more than about one half of the mass square of the scalar field ( a more accurate value is @xmath60 ) the chaotic structure disappear and the dynamics becomes regular .
this critical ratio @xmath61 is about @xmath62 times lower than that corresponding to the disappearance of page s points ( @xmath63 ) .
surprisingly , the estimation ( 17 ) is correct even for @xmath39 near the critical value when corresponding @xmath50 can not be treated as large .
numerical values of @xmath50 differ from ( 17 ) by several percents at most .
the possible values of @xmath39 allowing the chaotic dynamics for potential ( 15 ) are also restricted from below for a negative cosmological constant .
the corresponding critical value is @xmath64 ( see @xcite for detail ) .
we start with one note about the pure exponential potential @xmath65 .
it is easy to see from eq.(13 ) that there are no page s point in this case .
the direction of a trajectory starting with zero initial conditions from the euclidean boundary is fully determined by the value @xmath66 .
if @xmath67 _ all the trajectories _ from the euclidean boundary go into the lorentzian region while if @xmath68 all of them go into the euclidean region .
now we consider the potential @xmath69 . in the limit
@xmath70 it looks like @xmath71 with @xmath72 , while for large @xmath2 it looks like the pure exponential one .
the equation for page s point is now @xmath73 we can see that page s point exists only if @xmath67 . in the opposite case , again all the trajectories from the euclidean boundary
go into the euclidean region .
it means that there are no periodical trajectories with the full stop points in this case . remembering the first property of our chaotic system listed in previous section
, we may expect that the strange repellor is absent at all when @xmath68 .
= results of the numerical integration confirms this analytical considerations . in fig.1
several typical trajectories are plotted .
although points with @xmath74 @xmath75 are possible , trajectories containing such points have a zigzag - like form and can not return to the lorentzian region .
periodical trajectories are absent and the dynamics , in contrast to previously considered cases , is regular : it is impossible to avoid a singularity even on zero - measure set of initial condition .
all the trajectories fall into singularity .
for @xmath67 the structure of trajectories is similar to the massive scalar field case .
again , if we consider points of maximal expansion , we can find regions corresponding to the bouncing trajectories . to distinguish between bounces and zigzags
, we use an additional condition that the value of @xmath2 at the point of bounce is greater than @xmath76 ( or we can use the condition that the trajectory after the bounce returns to the lorentzian region ) . then in limit @xmath77 the width of bouncing regions tends to zero . it should be noted that when @xmath66 crosses the critical value the entire structure of periodical trajectories disappears in a jump - like manner .
now we turn to a more general case @xmath78 .
it corresponds to an effective cosmological constant @xmath79 .
let us study the case of positive @xmath80 .
page s points exist if the transcendent equation @xmath81 have roots .
= if @xmath68 this equation has one root .
the page s point corresponding to this sole root looks like the 2-nd one for a massive scalar field with a cosmological constant : trajectories with @xmath82 tend to the lorentzian region while with @xmath83 to the euclidean one . and the dynamical picture obtained from the numerical integration have similar features : one unstable periodical orbit near @xmath84 and nothing else .
the discrepancy between the estimation ( 17 ) and the numerical solution grows with @xmath66 reaching it s critical value .
the case @xmath67 is fully similar to the massive one with a cosmological constant : there exist two or zero page s points depending on the values of @xmath66 and @xmath39 ( the long - dashed curve in fig.2 ) .
the chaos disappears for @xmath39 ( the lower solid curve in the same plot ) which are somewhat larger than needed for page s points to disappear . for completeness we also present the area of values of @xmath85 leading to the chaotic dynamics ( it is bounded by the upper solid curve in fig.2 ) .
it essentially depends on the trajectory behaviour for large values of the scale factor and hence can not be understood by studying page s points .
we see from this plot that the massive case relation @xmath86 is still valid with a good accuracy .
for potentials more steep than exponential we face a situation which differs from the massive scalar field case significantly .
it can be shown from studying the analog of ( 11 ) that trajectories from the euclidean boundary for large values of @xmath2 go into euclidean region .
so the area of possible periodical trajectories is limited from the side of large @xmath2 .
this leads to limiting the possible values of the scale factor on such trajectories . in fig.3
we schematically show the number of oscillations of the field @xmath2 for periodical trajectories depending on the location of their full stop points on the euclidean boundary .
in contrast to the previous cases , this number is restricted from above by some number @xmath6 .
the numerical results show that the number of bounce intervals is also finite and restricted by some number @xmath87 .
in addition , there exists a complicated system of rules determining the substructure of intervals because the number of subintervals is now depends on the ordinal number of the interval . in particular , for an interval with ordinal number @xmath88 ,
the number of subintervals is always less than @xmath6 .
it is interesting that this dynamical picture looks very similar to that described in @xcite for the system with a massive scalar field and a hydrodynamical matter .
= numerical investigations also show the transitions to regular dynamics which can be understood by studying the page points configuration .
let us illustrate this picture using some concrete one - parameter family of potentials @xmath89 depending on the parameter @xmath66 , two or zero nontrivial page , s points can appear .
the critical value of @xmath66 obtained numerically is @xmath90 . for @xmath2 larger than this value
we have two page s points .
the area of initial points for trajectories penetrating into the lorentzian region ( i.e. possible full stop points of periodical trajectories ) lies between them .
so the properties of the page s points are inverse with respect to those for the case of a massive scalar field with a cosmological constant . in reality
, the chaotic behaviour disappears for a somewhat larger value @xmath91 .
it is however remarkable that this quite simple analytical expressions ( with computer used only to solve corresponding transcendent equation ) gives us a rather good estimation of values of @xmath66 which allows a chaotic regime
. the concrete form of potential for large @xmath2 does not affect the properties of chaos and therefore the steepness of the potential can be arbitrary high .
the picture described above is valid even for cases with infinitely high potential walls .
we have investigated a simple case @xmath92 in this case the condition of the existence of page s points has the exact solution @xmath93 , while the numerical result for the existence of chaotic dynamics is @xmath94 .
we have studied the dynamics of closed friedmann - robertson - walker universes with a scalar field .
it was found that there exists a rather wide class of scalar field potentials , more steep than power - law , for which the dynamics of a scale factor is regular and there is no possibility to escape a singularity at a contraction stage . on the other side , the class of potentials which allow the chaos , associated with chaotic oscillations of the scale factor , is also sufficiently wide .
this chaos manifests itself in the presence of a strange repellor - a fractal set of unstable periodical orbits .
their existence is connected with a possibility for this system to have a bounce .
taking the point of maximal expansion as the initial one , the regions of the initial conditions in configuration space that lead to bouncing trajectories have a rather regular and obvious structure : the n - th region contains trajectories which have a bounce after n oscillations of field @xmath2 .
more investigations is still required to study the possibility of existence of more complicated chaos for some another class of scalar field potentials .
the formal definition of the bounce ( @xmath36 , @xmath95 ) is not good enough .
the reason is that trajectories like in fig.1 is very natural for our dynamical system .
they satisfy the formal bounce criterion , but it is intuitively clear that zigzags on trajectories shown in fig.1 can not change significantly the fate of the trajectories falling into singularity . on the other hand , the condition we used previously @xcite ( the value of scale factor at the next point of maximal expansion is greater than at previous one ) rejects trajectories with decreasing but lying inside the area of the strange repellor value of @xmath23 at the points of maximal expansion .
the condition that the trajectory after the bounce returns to the lorentzian region or that the scale factor @xmath23 at the point of bounce is less than the scale factor of the chaos generating page s point may be used as a more suitable additional criterion .
varying the potential , several types of transitions from chaotic to regular behaviour can be distinguished : ( 1 ) the number of bounce regions in the initial condition space can remains infinite , but the area of the strange repellor in the configuration space tends to zero ( for the potential like @xmath96 ; @xmath97 ) .
( 2 ) the number of the bounce regions becomes finite and tends to zero while the area of the strange repellor remains unbounded ( the same potential but for @xmath98 ) .
( 3 ) the number of the bounce regions diminishes and the strange repellor shrinks ( for very steep potentials or for a mass - like potential in a more general case of a scalar field with a hydrodynamical matter @xcite ) ( 4 ) the width of the bounce regions tends to zero while their number and the area of the strange repellor remain unchanged ( for @xmath99 ) .
the study of the transition from chaos to order for such dynamical systems may be interesting from the mathematical point of view .
using the fact that the region in the configuration space where bounces are possible represents a sufficiently narrow area near the euclidean boundary , we have proposed that studying trajectories having full stop points at this curve can provide us with an important information about the strange repellor as a whole .
this suggestion realizes at least for the type of chaos which satisfies two properties , established by numerical investigations for steep potentials and described in sec.2 .
their validity for less steep potentials will be the goal of our future work .
in addition , the transition from chaos to order appear to be closely connected with the change of structure of page s point configuration at the euclidean boundary at least for positive scalar field potentials .
this work was supported by russian foundation for basic research via grants no 96 - 02 - 16220 and no 96 - 02 - 17591 .
99 l. parker and s.a .
fulling , phys .
d * 7 * , 2357 ( 1973 ) ; a.a .
starobinsky , sov .
. lett . * 4 * , 82 ( 1978 ) .
s.w . hawking , in _ relativity , groups and topology ii _
, les houches , session xl 1993 eds . b.s .
dewitt and r. stora , ( north holland , amsterdam , 1984 ) .
d.n . page , class .
quantum grav . * 1 * , 417 ( 1984 ) .
kamenshchik , i.m .
khalatnikov and a.v .
toporensky , int . j. mod
d * 6 * , 673 ( 1997 ) .
kamenshchik , i.m .
khalatnikov and a.v .
toporensky , int . j. mod
d * 7 * , 129 ( 1998 ) .
cornish and e.p.s .
shellard , phys .
81 * , 3571 ( 1998 ) u.gunther and a.zhuk , phys.rev .
* d56 * , 6391 ( 1997 ) s.foster , gr - qc/9806113 p.gaspard and s.a.rice , j. chem . phys . * 90 * , 2225 , ( 1989 ) .
n.j.cornish and j.j.levin , phys .
d * 55 * , 7489 , ( 1997 ) .
kamenshchik , i.m .
khalatnikov and a.v .
toporensky , int . j. mod .
d * 6 * , 649 ( 1997 ) .
kamenshchik , i.m .
khalatnikov , s.v.savchenko and a.v .
toporensky , phys .
d * 59 * : 123516 , ( 1999 ) . | the dynamics of closed scalar field frw cosmological models is studied for several types of exponentially and more than exponentially steep potentials .
the parameters of scalar field potentials which allow a chaotic behaviour are found from numerical investigations .
it is argued that analytical studies of equation of motion at the euclidean boundary can provide an important information about the properties of chaotic dynamics .
several types of transition from chaotic to regular dynamics are described .
+ pacs : 05.45.+b 98.80.hw 98.80.cq + + |
the top quark was expected in the standard model ( sm ) of electroweak interactions as a partner of the b - quark in a su(2 ) doublet of the weak isospin , in the third family of quarks .
searching for the top quark was the primary physics objective in run - i .
the first published evidence appeared in a cdf @xcite paper in 1994 , and its observation ( discovery ) was reported by cdf @xcite and d0 @xcite in the same issue of prl in 1995 .
both experiments have finished their analyses of run - i data for some time now , and only a few new results on top quark are presented in this paper . a summary of the top quark s mass and the and single top production cross section measurements is presented . a closer look at the analysis techniques used and a perspective view on top quark physics after its first 7 years ( or so ) is the subject of this paper . in anticipation of run - ii , in which the number of reconstructed events is expected to be at least 20x larger than in run - i , the question of whether all available results are consistent with the simplest hypothesis , adopted by both cdf and d0 experiments in run - i , that data contains just the events and the standard model backgrounds is re - visited .
at fermilab tevatron energy , @xmath10 = 1.8 tev , the dominant production mechanism for top quarks is @xmath11 pair production by a quark - antiquark or gluon - gluon initial state via the strong interaction ; for top quark masses above m@xmath12 the @xmath13 fusion dominates . assuming the standard model decays , there are three classes of final states , all with two b - quarks jets : i ) _ di - leptons
, when both w decays are leptonic , with 2 jets and missing transverse energy ( ) , @xmath14 for @xmath15 final states ; ii ) _
lepton+jets _ , when one w decays leptonically and the other into quarks , with 4 jets and , @xmath16 for @xmath15 ; iii ) _ all - hadronic _
, when both w s decay into quarks , with 6 jets and no , @xmath17 .
the backgrounds to the possible @xmath18 signal coming from the @xmath19+jets process of qcd have been feared to be much larger than the top signal and must be addressed .
the qcd multijet production background is even more important for all - hadronic final states , in which both w bosons decay into quarks .
the two dominant electroweak processes leading to a single top quark production are : a ) s - channel w * production and its subsequent decay into t and b quark / antiquarks , leading to a final state with a w and two b - quark jets ; b ) t - channel w - gluon fusion process , leading to a final state with a w and two jets but only one of them being due to the b - quark .
the techniques used in cdf and d0 are variations of simple event counting .
both experiments follow identical steps : i ) identify events with the expected top signature ; ii ) calculate the expected sm backgrounds ; iii ) count events above the expected backgrounds ; iv ) apply corrections for the acceptance , reconstruction inefficiencies and other biases .
this paper reports on measurements of the pair - production cross section and the single top production cross section .
one should remember two facts : i ) it is _ assumed _ that the selected sample of events contains just the events and the sm background ; this is the simplest and the most natural hypothesis since the top quark is expected in the sm ; ii ) some of the acceptance corrections are strongly varying functions of the top quark mass , m@xmath20 , and , consequently , the value of the measured cross section depends on the value of m@xmath20 , which has to be determined independently .
all mass measurement techniques used by cdf and d0 assume that each event in the selected sample contains a pair of massive objects of the same mass ( quarks ) which subsequently decay as predicted in the sm .
information about the kinematics of the event is used in a variety of fitting techiniques . a one - to - one mapping between the observed leptons and jets and the fitted partons
is assumed .
again , there are two things to remember : i ) it is _ assumed _ that the selected sample of events contains just the events and the sm background ; ii ) the combinatorics of the jets - lepton(s ) combinations ( only one of many possible combinations is correct ) adds to the complexity of the problem .
precision measurements of various electro - weak parameters , whose values depend on m@xmath20 indirectly ( via radiative corrections ) , are compared with the corresponding values predicted by the theoretical calculations in the consistency checks of the sm .
data from lep - i , lep - ii , sld , cdf , d0 , @xmath21-scattering results and other precision experiments , including or excluding the direct measurements of the top quark mass , can be used to yield the most likely top quark mass , consistent with the predicted values of the measured electroweak observables .
results are model dependent , as one has to assume a particular theory ( e.g. sm or minimal supersymmetric standard model ) to make such comparisons possible .
an additional uncertainty comes from the unknown higgs boson mass , which also enters into calculations of the radiative corrections .
the results of such global fits are summarized in section 8 , where the question of overall consistency of all electroweak measurements is examined within the framework of the standard model .
all cdf and d0 searches impose stringent identification , selection and transverse energy ( ) cuts on leptons and jets to minimize the sm and misidentification backgrounds . except for di - lepton samples , in which backgrounds are expected to be small , various techniques of tagging b
- quarks are employed to reduce backgrounds .
soft - lepton " tagging is used by both cdf and d0 , and the secondary vertex tagging , which uses a silicon vertex detector ( svx ) , by cdf .
d0 , not equipped with a svx , makes much greater use of various kinematic variables to reduce backgrounds .
the largest sm background is the qcd w+jets production .
both cdf and d0 use vecbos @xcite calculations to estimate shapes of the background distributions due to this process . presently available samples of the top event candidates are small , and the top cross section and mass measurements are still dominated by the statistical errors .
this will no longer be true in run - ii .
.results of d0 @xcite and cdf @xcite direct top searches . [ cols="<,^,^,^,^",options="header " , ] perhaps even more importantly , the and single top events constitute background to any new physics . as a consequence of the large top mass ,
the event selection cuts in top analyses are virtually identical to those applied in many analyses looking for physics beyond the sm ( supersymmetry , technicolor , et cetera ... ) .
the measured cross section values depend on the top quark mass , whose value has been determined in cdf and d0 using various kinematical fitting techniques @xmath22 the assumption that events are just the events and the sm background .
if the sample is not not exclusively due to the events and the sm background , the mass measurements may be incorrect . if an additional process were present , the number of observed events would not agree then with the mc predictions obtained with the measured value of m@xmath20 .
it is thus imperative to compare various distributions of the reconstructed top quarks , and especially those of the -system , with the sm predictions .
discrepancies could indicate new physics .
both cdf and d0 made numerous comparisons .
no significant disagreements were found , as perhaps should be expected given the still limited statistics .
however , there exist a few hints that the simplest hypothesis ( that the top candidate events are just the events and sm background ) may not be entirely correct . with the luminosity of 2 fb@xmath0 per experiment they should be monitored carefully , as they may be offering us glimpses of new physics .
i. cdf cross section seems a little high compared to the theoretical predictions ; it would be more consistent with the lower value of m@xmath20 ; the indirect measurements of m@xmath20 , based on the global checks of the sm _ excluding _ the direct m@xmath20 measurements , alo prefer lower m@xmath20 ( @xmath23 157 - 169 gev / c@xmath24 ) ii .
there is an excess of w+2jet and w+3jet events ( 13 where 4.4@xmath25 are expected ) with double tagged jets ( tagged both with svx and slt ) in the tagged jet multiplicity distribution in the cdf .
in addition , the kinematical properties of those events do nt agree well with the sm predictions @xcite .
there is a hint of an increase of the reconstructed top quark mass with the number of jets in an event .
iv . two ( out of 9 ) cdf di - lepton events poorly fit the hypothesis and have unexpectedly large + @xmath26 .
one such event exists in the d0 sample . v. the distributions of the mass , in both cdf and d0 , seem to have a few more events than expected in the high mass region .
the transverse momentum distribution of the system for the sample of 32 cdf tagged lepton+jets events , seems a little harder than expected , based on the monte carlo calculations .
d0 data does not show any deviations from sm expectations .
the rapidity distribution ( figure 6 ) of the system for the sample of 32 cdf tagged lepton+jets events has a strikingly different shape than that based on mc simulations .
the rapidity variable probes directly the _ fitted _ longitudinal component of the neutrino momenta and , as such , is perhaps more sensitive than other variables to the correctness of the original hypothesis that the fitted events are the events .
however , the d0 pseudorapidity plot ( figure 7 ) is in good agreement with expectations @xcite .
i would like to thank the conference organizers for their very warm hospitality and for making the epiphany 2003 an exceptionally well organized conference .
it was an especially memorable experience for me ; i still vividly remember my times as a student , and later an assistant , at the institute of physics of jagiellonian university where the conference was held .
berends , h. kuijf , b. tausk , w.t .
giele , ( both cdf and d0 experiments used the vecbos w+3jets calculation , with an additional jet added in the fragmentation process simulated using herwig , rather than the more appropriate vecbos w+4jets calculation because of prohibitive amount of cpu required by the latter . ) top mass and cross section results from cdf and d0 at the fermilab tevatron " , k. sliwa , invited talk at 13th topical conference on hadron collider physics , tata institute of fundamental research , mumbai , india , january 14 - 20 , 1999 ; in proceedings , p 169 ; world scientific , 1999 . | measurements of the top quark mass and the and single top production cross sections , obtained by cdf and d0 collaboration at the tevatron , are presented .
the methodology of cdf and d0 top quark analyses and their underlying assumptions are summarized .
the cdf and d0 top mass averages , based on about 100 pb@xmath0 of data from collisions of @xmath1 at @xmath2 = 1.8 tev collected by each experiment in run - i , and obtained from measurements in several channels , are m@xmath3 and m@xmath4 , respectively . the combined tevatron top quark mass is m@xmath5 .
the cdf measurement of the cross section ( assuming m@xmath6 ) is @xmath7 pb , and the d0 value ( assuming m@xmath8 ) is @xmath9 pb . in anticipation of the increased amount of data in run - ii , prospects are presented .
the fact that top quark analyses are among the best windows to new physics beyond the standard model is emphasized .
pacs number:14.65 ha .
= 1 = by -1 |
the na49 collaboration has recently performed a series of measurements of pb - pb collisions at 20 , 30 , 40 , 80 and 158 @xmath1 gev beam energies @xcite . combining these results with measurements at lower beam energies from the ags @xcite
they reveal a sharp variation with beam energy in the @xmath9 , with @xmath10 , and in the k@xmath0 ratios . such a strong variation with energy does not occur in pp collisions and therefore indicates a major difference in heavy - ion interactions .
a strong variation with energy of the @xmath9 ratio has been predicted on the basis of arguments put forward in @xcite .
it has also been suggested recently in ref .
@xcite that this might be a signal of the critical point in the qcd phase diagram at high baryon density . in this paper
we explore another , less spectacular , possibility for the sharp maximum .
the arguments for this will be presented in two steps . in step ( i ) the statistical model of a hadron gas is studied . for this purpose ,
we investigate various quantities along the freeze - out curve @xcite as a function of @xmath4 .
it is shown that the entropy of the hadronic gas is dominated at the lower energies by baryons and at the higher ones by mesons .
the transition occurs just at the energy where the sharp maximum has been observed . in step ( ii )
we speculate whether at these incident energies the freeze - out might happen earlier .
the consequences are discussed and arguments based on hbt results are presented to support this idea .
the experimental data from heavy - ion collisions show that the k@xmath0 ratio rises from sis up to ags .
it is larger for ags than at the highest cern - sps energies @xcite and decreases even further at rhic @xcite .
this behavior is of particular interest as it could signal the appearance of new dynamics for strangeness production in high - energy collisions .
it was even conjectured that this property could indicate an energy threshold for quark - gluon - plasma formation in relativistic heavy - ion collisions @xcite . in the following
we analyze the energy dependence of strange to non - strange particle ratios in the framework of a hadronic statistical model . in the whole energy range ,
the hadronic yields observed in heavy - ion collisions resemble a system in chemical equilibrium characterized by two parameters , the temperature @xmath2 and the baryon chemical potential @xmath3 .
these two parameters define the `` freeze - out curve '' which seems determined by the condition of fixed average energy / particle @xmath11 1 gev @xcite . as the beam energy increases in the energy domain of sis and ags , @xmath2 rises and
@xmath3 decreases slightly .
above ags energies , @xmath2 exhibits only a moderate change and converges to its maximal value in the range of 160 to 180 mev , while @xmath3 is strongly decreasing . rather than studying the k@xmath0 ratio
, we use the ratios of strange to non - strange particle multiplicities ( wroblewski factor ) @xcite defined as @xmath12 where the quantities in angular brackets refer to the number of newly formed quark - antiquark pairs , i.e. it excludes all quarks that were present in the target and the projectile nuclei . applying the statistical model to particle production in heavy - ion collisions
calls for the use of the canonical ensemble to treat the number of strange particles , particularly for data in the energy range from sis up to ags @xcite .
the calculations for au - au and pb - pb collisions are performed using a canonical correlation volume @xcite .
the quark content used in the wroblewski factor is determined at the moment of _ chemical freeze - out _ ,
i.e. from the hadrons and , especially , hadronic resonances before they decay .
this ratio is thus not an easily measurable observable unless one can reconstruct all resonances from the final - state particles .
the results are shown in fig . 1 left as a function of @xmath4 .
( for definition see text ) from strange baryons , strange mesons , and mesons with hidden strangeness .
the sum of all contributions is given by the full black line .
right : lines of constant wroblewski factor @xmath13 in the @xmath14 plane ( solid lines ) together with the freeze - out curve ( dashed line ) @xcite.,title="fig:",width=283 ] ( for definition see text ) from strange baryons , strange mesons , and mesons with hidden strangeness .
the sum of all contributions is given by the full black line .
right : lines of constant wroblewski factor @xmath13 in the @xmath14 plane ( solid lines ) together with the freeze - out curve ( dashed line ) @xcite.,title="fig:",width=283 ] [ wrob ] the solid line ( marked `` sum '' ) in fig .
1 describes the statistical - model calculations in complete equilibrium along the unified freeze - out curve @xcite with the energy - dependent parameters @xmath2 and @xmath3 . from fig .
1 we conclude that around @xmath4 = 8.2 gev corresponding to an incident energy of 30 @xmath1 gev , the relative strangeness content in heavy - ion collisions reaches a clear and well pronounced maximum .
the wroblewski factor decreases towards higher energies and reaches a limiting value of about 0.43 .
for details see ref .
@xcite .
the appearance of the maximum can be traced to the specific dependence of @xmath3 and @xmath2 on the beam energy as also pointed out in ref .
figure 1 ( right ) shows lines of constant @xmath13 in the @xmath14 plane .
as expected , @xmath13 rises with increasing @xmath2 for fixed @xmath3 . following the chemical freeze - out curve , shown as a dashed line in fig . 1
, one can see that @xmath13 rises quickly from sis to ags energies , then reaches a maximum at @xmath15 mev and @xmath16 mev .
these freeze - out parameters correspond to 30 @xmath1 gev laboratory energy . at higher incident energies
the increase in @xmath2 becomes negligible but @xmath3 keeps on decreasing and as a consequence @xmath13 also decreases .
the importance of finite baryon density on the behavior of @xmath13 is demonstrated in fig .
1 left , showing separately the contributions to @xmath17 coming from strange baryons , from strange mesons and from hidden strangeness , i.e. from hadrons like @xmath18 and @xmath19 . as can be seen in fig . 1
, the origin of the maximum in the wroblewski ratio can be traced to the contribution of strange baryons .
the production of strange baryons dominates at low @xmath4 and loses importance at high incident energies when the yield of strange mesons increases . however
, strange mesons also exhibit a broad maximum .
this is due to associated production of e.g. kaons together with hyperons at the lower incident energies .
+ .,title="fig:",width=377 ] + .,title="fig:",width=491 ] figure [ lambda_pi ] shows the comparison of the statistical model @xcite and @xmath20 data ( except for rhic data , above @xmath4 = 130 gev ) .
as can be understood from the arguments above , the ratio @xmath21 exhibits the most pronounced maximum , k@xmath0 a weaker one and k@xmath5 has no maximum at all .
it is worth noting that the maxima in the ratios for multi - strange baryons occur at ever higher beam energies .
this can be seen clearly in fig .
[ lambda_pi ] for the @xmath7 ratio which peaks at a higher value of the beam energy .
the ratio @xmath8 also shows a ( very weak ) maximum , as can be seen in fig .
[ lambda_pi ] . the higher the strangeness content of the baryon , the higher in energy is the maximum .
this behavior is due to a combination of the facts that the baryon chemical potential decreases rapidly with energy and the multi - strange baryons have successively higher thresholds . it is to be expected that if these maxima do not all occur at the same temperature , i.e. at the same beam energy , then the case for a phase transition is not very strong .
the observed behavior seems to be governed by properties of the hadron gas .
more detailed experimental studies of multi - strange hadrons will allow the verification or disproval of the trends shown in this paper .
it should be clear that the @xmath8 ratio is very broad and shallow and it will be difficult to find a maximum experimentally
. in general , the model gives a good description of the data .
it shows a broad maximum in the k@xmath0 ratio while the data exhibit a sharp peak .
the drop towards 158 @xmath1 gev is most pronounced when using 4@xmath22 yields as done in these figures .
this decrease is less pronounced for midrapidity values @xcite .
while the statistical model can not explain the sharpness of the peak in the k@xmath0 ratio , there are nevertheless several phenomena giving rise to the rapid change which warrant a closer look at the model . as a function of the beam energy as calculated in the statistical model using thermus @xcite.,width=396 ] [ entropy_s ] to get a better estimate of the thermal parameters in the transition region we show in fig .
[ entropy_s ] the entropy density as a function of beam energy following the freeze - out curve given in @xcite .
the separate contribution of mesons and of baryons to the total entropy is also shown in this figure .
there is a clear change of baryon to meson dominance around @xmath4 = 8.2 gev . above this value
the entropy is carried mainly by mesonic degrees of freedom .
it is remarkable that the entropy density divided by @xmath23 is constant for all incident energies except for the low - energy region corresponding to the sis energy region .
in this section , we explore the possibility that freeze - out might happen earlier in the transition region . for this interpretation
, we show in fig .
[ kp_pip_t_mub ] the calculated values of the k@xmath0 ratio for various combinations of @xmath2 and @xmath3 as contour lines with the corresponding values given in the figure .
the thick solid line reflects the locations of the freeze - out given by the condition of ref .
if freeze - out happens around an incident energies of 30 @xmath1 gev at higher @xmath2 , then the ratio k@xmath0 will be higher .
this ratio can never exceed a value of 0.25 in an equilibrium condition .
ratio for combinations of @xmath2 and @xmath3 are given by the contour lines and the corresponding values .
the thick line refers to the freeze - out curve @xcite.,width=351 ] [ kp_pip_t_mub ] it turns out that other particle ratios are less affected by a different freeze - out scenario , as their variation in the @xmath14 plane is very different @xcite .
[ ceres ] an early freeze - out is supported by results from hbt studies @xcite .
figure 5 shows the extracted volumina as a function of @xmath4 . between top ags and the lowest sps energies
a minimum can be seen . as the fireball is expanding ,
a smaller volume reflects an earlier time .
the authors of ref .
@xcite relate this minimum to a change in the interaction from @xmath24 to @xmath25 .
it has been shown that the statistical model yields a maximum in the relative strangeness content around 30 @xmath1 gev .
this is due to a saturation in the temperature @xmath2 while the chemical potential keeps decreasing with incident energy .
since the chemical potential plays a key role , it is clear that baryons are strongly affected .
indeed , all hyperon/@xmath22 ratios yield maxima . in contrast
, the k@xmath5 ratio shows a continuously rising curve as expected from the arguments above .
the k@xmath0 ratio , however , exhibits a maximum , as k@xmath26 mesons are sensitive to the baryo - chemical potential due to their associate production with hyperons .
the model predicts that for different hyperon/@xmath22 ratios the maxima occur at different energies .
if experiments prove this , the case for a phase transition is strongly weakened .
the energy regime around 30 @xmath1 gev seems to have specific properties .
it is shown that the entropy production occurs below this energy mainly via creation of baryons , while at the higher incident energies meson production dominates . in a third section , we speculated on the impact of a change in the freeze - out condition which might lead to an early freeze - out , thus deviating from the usual freeze - out condition .
such a scenario would increase the k@xmath0 while leaving other particle ratios essentially unchanged .
hbt studies show that around 30 @xmath1 gev a minimum in the extracted volumina occurs .
this could be interpreted as an earlier kinetic freeze - out and might indicate also another freeze - out for chemical decoupling . +
this work was supported by the german ministerium fr bildung und forschung ( bmbf ) and by the polish committee of scientific research .
10 m. gadzicki , ( na49 collaboration ) , j. phys .
g : nucl . part . phys . * 30 * ( 2004 ) s701 .
afanasiev et al . , ( na49 collaboration ) , phys . rev .
* c66 * ( 2002 ) 054902 .
t. anticic et al . , phys .
* 93 * ( 2004 ) 022302 ; c. blume et al . , j. phys .
g : nucl . part .
* 31 * ( 2005 ) s685 .
l. ahle et al . , ( e802 collaboration ) , phys . rev .
* c57 * ( 1998 ) 466 .
l. ahle et al . , ( e802 collaboration ) , phys . rev .
* c60 * ( 1999 ) 044904 and 064901 .
l. ahle et al . , ( e866/e917 collaboration ) , phys . lett .
* b490 * ( 2000 ) 53 .
s. albergo et al . , phys .
* 88 * ( 2002 ) 062301 .
s. ahmad et al . , phys . lett .
* b381 * ( 1996 ) 3 . j. klay et al .
, ( e895 collaboration ) , phys . rev .
* c68 * ( 2003 ) 054905 .
m. gadzicki and m.i .
gorenstein , acta phys .
polonica b * 30 * ( 1999 ) 2705 .
r. stock , j. phys .
g : nucl . part .
phys . * 30 * ( 2004 ) s633 .
j. cleymans and k. redlich , phys . rev .
* 81 * ( 1998 ) 5284 .
j. cleymans , h. oeschler , k. redlich and s. wheaton , contribution to these proceedings .
l. ahle et al . , ( e-802 collaboration ) , phys . rev .
* c58 * ( 1998 ) 3523 ; phys .
c60 ( 1999 ) 044904 ; l. ahle et al .
, e866/e917 collaboration , phys .
b476 ( 2000 ) 1 ; phys.lett .
b490 ( 2000 ) 53 . j.c .
dunlop and c.a .
ogilvie , phys .
c61 ( 2000 ) 031901 and references therein ; c. a. ogilvie .
talk presented at qm2001 , stony brook , january 2001 , nucl .
a * 698 * 3c ; j.c .
dunlop , ph.d.thesis , mit , 1999 .
i. bearden et al . , ( na44 collaboration ) , phys .
b471 ( 1999 ) 6 . j. harris , ( star collaboration ) , talk presented at qm2001 , stony brook , january 2001 , nucl
a*698 * 64c .
a. wroblewski , acta physica polonica b16 ( 1985 ) 379 .
j. cleymans , h. oeschler , k. redlich , phys . rev .
* c59 * ( 1999 ) 1663 .
p. braun - munzinger , j. cleymans , h. oeschler , k. redlich , nucl .
phys . a 697 ( 2002 ) 902 .
s. kabana , eur .
j. * c21 * ( 2001 ) 545 .
s. wheaton and j. cleymans , thermus - a thermal model package for root , hep - ph/0407174 .
s. wheaton , ph.d .
thesis , university of cape town , 2005 .
d. adamova et al . , ( ceres collaboration ) , phys .
* 90 * ( 2003 ) 022301 . | the recently discovered sharp peak in the excitation function of the k@xmath0 ratio around 30 @xmath1 gev in relativistic heavy - ion collisions is discussed in the framework of the statistical model . in this model ,
the freeze - out of an ideal hadron gas changes from a situation where baryons dominate to one with mainly mesons .
this transition occurs at a temperature @xmath2 = 140 mev and baryon chemical potential @xmath3 = 410 mev corresponding to an energy of @xmath4 = 8.2 gev .
the calculated maximum in the k@xmath0 ratio is , however , much less pronounced than the one observed by the na49 collaboration .
the smooth increase of the k@xmath5 ratio with incident energy and the shape of the excitation functions of the @xmath6 , @xmath7 and @xmath8 ratios all exhibiting maxima at different incident energies , is consistent with the presently available experimental data .
the measured k@xmath0 ratio exceeds the calculated one just at the incident energy when the freeze - out condition is changing .
we speculate that at this point freeze - out might occur in a modified way .
we discuss a scenario of an early freeze - out which indeed increases k@xmath0 ratio while most other particle ratios remain essentially unchanged .
such an early freeze - out is supported by results from hbt studies . |
biomembranes exert various types of repulsive and attractive forces on each other .
the van der waals forces are weakly attractive
these vary as @xmath3 for close separations and transition to @xmath4 at larger distances @xcite . here
@xmath0 is the mean distance between the interacting membranes .
the notable aspect of the attractive force is that it is long - ranged .
a somewhat ambiguous term _ hydration forces _
@xcite , is used to denote the repulsive force that mediates at very small inter - membrane distances .
the underlying mechanisms of hydration forces are still under active research @xcite it suffices to simply indicate here that they are quite short ranged and drop off exponentially with distance .
helfrich , in a pioneering work @xcite , showed that two fluctuating fluid membranes exert a repulsive force on each other .
biomembranes are generally quite flexible and a single membrane fluctuates both freely and appreciably at physiological temperatures . as two membranes approach each other , they _ hinder _ or diminish each others out - of - plane fluctuations .
this hindrance decreases the entropy and the ensuing overall increase of the free - energy of the membrane system , which depends on the inter - membrane distance , leads to a repulsive force that tends to push the membranes apart .
helfrich @xcite , using a variety of physical arguments and approximations , postulated that the entropic force varies as @xmath3 .
in contrast to the other known repulsive forces , this behavior is long - ranged and competes with the van der waals attraction at all distances @xcite . since helfrich s proposal @xcite
, biophysicists have used the existence of this repulsive force to explain and understand a variety of phenomena related to membrane interactions .
helfrich s work has been reexamined and extended in ref .
@xcite ( among others ) and most recently by freund @xcite .
see also @xcite for a an overview of freund s work .
freund clearly highlights some of assumptions made in helfrich s work and provides a fresh perspective on this problem @xcite .
freund controversially finds that within a range of @xmath0 values the force law between two fluctuating membranes is proportional to @xmath5 rather than the well - accepted result of helfrich : @xmath3 . to settle this issue ,
we have reexamined this problem both analytically and through recourse to carefully conducted monte carlo simulations .
as was initially pointed out by helfrich @xcite , due to reflective symmetry , the evaluation of the force between two membranes in a periodic stack may be replaced by a single membrane confined between two rigid walls ( figure 1 ) . throughout this article
, we will emphasize the differences between our work and those of ref .
@xcite .
a pair of fluctuating membrane may be replaced by a single membrane confined between two walls separated from each other by a distance @xmath6 . ]
consider a membrane as depicted in fig .
[ fig:1 ] . assume that the membrane occupies @xmath7 ^ 2 $ ] on the @xmath8-plane and the thermodynamic state of the membrane is described by @xmath9 , where @xmath10 is membrane mid - plane deviation along the @xmath11-axis . in the helfrich model , the hamiltonian is @xmath12 = \int_s \dd ^2\xv \frac{\kappa}{2}(\partial^2u)^2 .
\label{eqn:1}\ ] ] to address the thermal fluctuation we make the following assumptions : ( i ) the membrane consists of @xmath13 molecules located at @xmath14 , where @xmath15 is the molecule s size .
in other words , @xmath16 is the total degrees of freedom of the system , and ( ii ) microscopically the out - of - plane deviation @xmath17 of each molecules is quantized and can only take values from @xmath18 , where @xmath19 is a small spacing along the deviation direction .
then from the definition @xcite the partition function of the system can be written as a functional integral : @xmath20 where @xmath21 and @xmath22 is any differentiable function interpolating the discrete molecules deviation @xmath23 , @xmath24 .
the point - wise hindrance condition that @xmath25 is enforced throughout this work .
this constraint is in fact the key obstacle in the closed - form evaluation of the partition function .
freund @xcite modified the partition function by introducing adjustable integration limits , and then minimized the resultant free energy with respect to these limits to determine the change in free energy with respect to inter membrane separation ( and hence the entropic force law ) . in the conclusions section ,
we discuss freund s approach further . here
, we adopt the form of the partition function in eq.([eqn:2 ] ) which , notwithstanding the analytical intractability of its integration limits ( i.e. pointwise constraint on @xmath26 ) , is _ exact _ within the present formulation .
we remark here that helfrich @xcite avoids the point - wise @xmath27 condition , @xmath28 , and instead imposes a weaker constraint where the mean square membrane displacement is required to be bounded by @xmath29 , i.e. @xmath30 . to gain new insights into the partition function , two dimensionless quantities , @xmath31 and @xmath32 , are introduced as @xmath33 by changes of variables we may rewrite the partition function as ( @xmath34 ^ 2 $ ] , @xmath35 ) @xmath36 where @xmath37 .
let @xmath38 be a dimensionless variable . by a change of variable back and forth , @xmath39
, we obtain @xmath40 where @xmath41 as a consequence , the free energy density has the form @xmath42 it should be noted here that our free energy density has a subtle difference from the one in ref
the first term in the free energy does not depend on @xmath0 and vanishes upon differentiation with respect to @xmath0 .
it follows that the steric pressure is @xmath43 where @xmath44 .
janke and kleinert @xcite have performed monte carlo calculations and found that @xmath45 is a constant in the thermodynamic limit @xcite .
the value of this constant was also found by kleinert using a variational approach @xcite .
we remark that in the past works that have performed monte carlo calculations of this problem , the veracity of the helfrich s pressure law has been implicitly embraced and the focus has not been on examining the dependence of the entropic pressure on inter membrane distance but rather calculation of @xmath45 _ assuming _ that helfrich s inverse cube law is correct .
this is the reason , we believe , that freund s result and ( now ours ) has not been noted until now . as evident from the definition in eq.([eqn:8 ] ) , @xmath46 is a function of the separation between the rigid walls confining the membrane .
it is interesting to consider the energy variation , and consequently the pressure , at some small distances . as freund has shown , this limit is analytically tractable .
below we reproduce the result using a slightly different procedure .
consider the scaled partition function in eq.([eqn:6 ] ) , by fourier transformation we introduce @xmath47 where @xmath48 is the reciprocal lattice of @xmath49 . in matrix notations
the above equations can be rewritten as @xmath50 where @xmath51 ( resp .
@xmath52 ) denotes the column vector formed by @xmath53 , @xmath54 ( resp .
@xmath55 , @xmath56 ) , and @xmath57 is a unitary matrix satisfying @xmath58 . since @xmath59 , we may convert an integral over @xmath60 as a summation over @xmath61 : @xmath62 , and consequently @xmath63 . applying the parseval s theorem ,
we rewrite @xmath64 where @xmath65 is the diagonal matrix with entries @xmath66 , @xmath67 . defining a dimensionless variable @xmath68 ,
the scaled partition function is @xmath69 for @xmath70 i.e. @xmath71 , one has @xmath72 , \end{split } \label{eqn:14}\ ] ] where the identity @xmath73 was used . since @xmath74 , and the inner product of @xmath75 with the identity matrix yields its trace , the reduced partition function in the asymptotic limit is @xmath76 .
\end{split } \label{eqn:15}\ ] ] it is clear from the definition @xmath77 that the steric pressure has the leading order term of @xmath78 .
the next correction term can be obtained using the identity @xmath79 $ ] , where @xmath80 .
the steric pressure in the limit @xmath70 ( @xmath71 ) is thus @xmath81 .
\label{eqn:16}\ ] ] this pressure law has a very different @xmath82dependence than any known theories or simulations to date @xcite except of course the work by freund @xcite .
another interesting point obtained from eq.([eqn:16 ] ) is that the first term in the right - hand side has the form of pressure of the ideal gas .
physically , this implies that the membrane fluctuates similar to an ideal gas in the limit @xmath71 , with a correction term which is in the order of @xmath83 ( the second term in eq.([eqn:16 ] ) ) .
this ideal - gas contribution does not depend on the bending modulus @xmath84 , hence every type of membranes should exhibit the @xmath85 pressure law at the limit @xmath71 . in order to elucidate the full @xmath82dependence including the limit @xmath71 , a natural step is to take recourse in numerical monte carlo simulations which are discussed in the next section .
in the monte carlo simulations , the spatial coordinates are replaced by a square grid @xmath86 with lattice constant @xmath87 .
the membrane displacement along the @xmath88direction is specified by @xmath89 , where it is also discretized to a grid of space @xmath19 .
this scheme has been shown to be sufficient in prior monte carlo calculations @xcite .
the hamiltonian over these lattice points is @xmath90 where the discretized laplacian @xmath91 , \label{eqn:18}\ ] ] and @xmath92 denotes the displacement vectors to the four nearest neighbors to the site @xmath93 .
@xmath94{fig2a.eps } \\ \includegraphics[scale=0.45 , clip]{fig2b.eps}\\ \end{array } $ ] our simulation code was fully parallelized using spatial decompositions .
the lattice was divided into strips so that each strip can be updated independently using the usual metropolis algorithm @xcite .
since each strip is updated in parallel , some care has been taken in order to account for the lattice sites at the boundary .
figure [ fig:2](a ) illustrates a typical geometry in the simulations .
updating the central point requires the knowledge of the values at all points inside the dashed square .
it is forbidden to update any other site in this neighborhood until the update of the central site has been completed .
this can be accomplished by an appropriate choice of strip lengths along the @xmath95axis and performing a row - by - row update . in our simulations ,
the minimal lattice length along the @xmath95axis is 5 , so that there is no overlap between dashed square in figure [ fig:2](a ) .
a message passing routine was employed in updating the top and bottom two rows of the strips , while along the @xmath96axis a simple periodic boundary condition was used . in this way , our choice of lattice sizes are 50 by 50 , 100 by 100 , 600 by 600 , and 1000 by 1000 . an mc realization is shown in figure [ fig:2](b ) , where the parameters used for its generation are listed in the caption .
in addition to the ensemble average of the energy @xmath97 , the second physical quantity needed for this problem is the pressure .
the pressure can be derived by differentiating the change in free energy with respect to the inter membrane separation .
alternatively and more conveniently , the pressure can be related to the ensemble average of the hamiltonian density @xmath98 . rewriting eq.([eqn:4 ] ) as @xmath99 where @xmath100 is a scaled hamiltonian .
the membrane configuration @xmath101 is parameterized by the set of variable @xmath102 .
differentiation of the free energy can be carried out straightforwardly by using these rescaled parameters @xmath103 ; the concept which also appears in the derivation for the hellmann - feynman forces in quantum mechanics @xcite .
it follows that @xmath104 e^{-\frac{1}{\tau } \tilde{h}(\{v_{\yv}\})}\right.\\ & \left .
+ \frac{n}{d } ( cd)^n \int_{-1}^{1 } \prod_{\yv\in\tilde{{{\cal l } } } } \dd v_{\yv } e^{-\frac{1}{\tau } \tilde{h}(\{v_\yv\ } ) } \right ] \end{split } \label{eqn:20}\ ] ] using the concept of ensemble average and eq.([eqn:19 ] ) , the derivative of the free energy is @xmath105 in term of the energy density @xmath106 and @xmath107 for a continuos membrane , or @xmath108 and @xmath109 for a discretized membrane , where @xmath13 is the number of molecules and @xmath110 is a square encompassing each of them @xmath111 consequently for a membrane situated between two rigid plates of separation @xmath6 , the steric pressure @xmath112 where @xmath113 .
the above pressure is exact , and makes the computation of its value straightforward .
furthermore , the errors on the pressure @xmath114 stems from only the errors on @xmath115 , which for sufficiently long monte carlo time steps become small compared to the value of @xmath115 .
the validity of eq.([eqn:23 ] ) for both the asymptotic and general form of the pressure can be easily checked . in the asymptotic limit @xmath116 or @xmath71 , eq.([eqn:15 ] ) gives the free energy density @xmath117 of the form @xmath118 .
\label{eqn:24}\ ] ] it then follows that @xmath119 and @xmath120 in agreement with eq.([eqn:16 ] ) . in the general case
, it can be deduced from eq.([eqn:7 ] ) that @xmath121 as a result , the pressure obtained from eq.([eqn:23 ] ) is @xmath122 which is again in agreement with eq.([eqn:8 ] ) . a typical heating of a confined membrane .
circles represent mc results , while the dashed line corresponds to the heating of one - dimensional ideal gas .
the rigid wall separation is @xmath123 , while the membrane discretized spacing is @xmath124 .
the bending modulus is @xmath125 .
the membrane starts to experience the presence of the walls at about @xmath126 .
the errors of the energy is about @xmath127 .
for example , the energy on the far right point is @xmath128 ] @xmath129{fig4a.eps}\\ \includegraphics[scale=0.5 , clip]{fig4b.eps}\\ \includegraphics[scale=0.5 , clip]{fig4c.eps}\\ \end{array } $ ] @xmath130{fig5.eps } \end{array } $ ] in most simulation runs , at least @xmath131 mc time steps were performed .
a monte carlo time step is defined as the number of monte carlo updates divided by the number of grid points of the membrane .
the first 3000 time steps were omitted for thermalization .
the calculated statistical errors of a few percent of the energy estimators are obtained . a typical simulation result for heating of a membrane
is shown in figure [ fig:3 ] .
the size of the membrane in this figure is 120 by 120 , where @xmath124 , @xmath132 , and @xmath133 . to give a sense about the errors on the average energy , the measured internal energy density of figure [ fig:3 ] at @xmath134 is @xmath135 .
the pressure - distance relation for the separation ranging from @xmath136 is shown in figure [ fig:4 ] .
the simulation parameters are @xmath137 , @xmath138 , and @xmath139 , while the bending moduli are @xmath140 , @xmath141 , and @xmath142 , respectively .
the size of the membrane shown are @xmath143 , @xmath144 , and @xmath145 . from these log - log plots ,
it is clear that the relation @xmath1 is applicable within a small range @xmath0 , while there is a transition to @xmath2 as the separation distance increases .
the transition from @xmath3 to @xmath5-dependence can be estimated from eq.([eqn:13 ] ) as @xmath146 which for the choice of parameters @xmath147 , @xmath125 , @xmath148 the corresponding transition length is @xmath149 or about five times the spacing between molecule ( @xmath150 ) . for more realistic values of @xmath151 and @xmath152 @xmath153 @xcite ,
the transition length is @xmath154 . is the transition from @xmath5 to @xmath3-pressure law a result of finite size effects ?
our mc results in figure [ fig:4 ] for the membrane as large as @xmath155 indicate otherwise . as evident , the smaller membrane of size
@xmath143 exhibits the same transition length @xmath156 to those of much larger ones . since in real biological membranes
the number of possible excitation modes are perhaps much larger than in our model , the transition from the ideal gas @xmath5 to @xmath3 pressure law could possibly represent some new physics
the exploration of this is deferred to future work .
our results are easily interpreted within the context of the free energy of a tightly confined membrane . at small @xmath157
, the bending energy contribution to the free energy is negligible compared to the entropic contribution .
the suppression of the elastic effects hence leads to the limiting pressure law of the ideal gas .
nevertheless , it should be noted that the transition length @xmath156 depends on the bending modulus , temperature , and intermolecular spacing as indicated in eq.([eqn:29 ] ) .
the pressure law transition from @xmath5 to @xmath3 dependence may have some important implication in the interactions among living cells , potentially opening a new avenue in reevaluating the conventional understandings about how biological cells mechanically interact .
a larger range of the @xmath158 dependence is shown in figure [ fig:5 ] .
the sizes of the membrane , which are @xmath159 by @xmath159 , are shown in the subfigures . at large @xmath0
the pressure follows an exponential decay relation of type @xmath160 resulting in the log - log plots of the form @xmath161 where @xmath162 , @xmath163 , and @xmath164 .
this exponential decaying regime has not yet been confirmed by any simulations or theory thus far , despite a speculation in ref .
. an interesting conclusion of this result is that the helfrich entropic force is not _ really _ as long - ranged as previously believed .
in summary , we conclude that freund s @xcite conclusions are correct for short inter - membrane distances and in that regime , his result is a major modification of the well - accepted entropic force law due to helfrich @xcite .
however , helfrich is correct for intermediate distances and finally , for large separations , the entropic force decays exponentially . at the time of writing this manuscript
, we became aware of a pre - print by t. auth and g. gompper , who ( at least for short and intermediate membrane separations ) have reached similar conclusions as us .
p. sharma gratefully acknowledge helpful discussions with professor ben freund and his encouragement to pursue this work .
y. hanlumyuang thanks dr .
xu liu and professor aiichiro nakano for answering several questions regarding parallel computations , and dr .
dengke chen for countless insightful discussions . | the complex interplay between the various attractive and repulsive forces that mediate between biological membranes governs an astounding array of biological functions : cell adhesion , membrane fusion , self - assembly , binding - unbinding transition among others . in this work , the entropic repulsive force between membranes which originates due to thermally excited fluctuations
is critically reexamined both analytically and through systematic monte carlo simulations .
a recent work by freund @xcite has questioned the validity of a well - accepted result derived by helfrich @xcite .
we find that , in agreement with freund , for small inter - membrane separations ( @xmath0 ) , the entropic pressure scales as @xmath1 , in contrast to helfrich s result : @xmath2 . for intermediate separations ,
our calculations agree with that of helfrich and finally , for large inter - membrane separations , we observe an exponentially decaying behavior . |
we consider a line lightning protection device ( llpd ) consisting of a series of arc gaps between a power line and the ground . during normal grid operation ,
the device works as an insulator since the voltage over the device is insufficient to cause dielectric breakdown and arcing . in the case of a lightning strike ,
on the other hand , the large overvoltage will ignite a series of electric arcs in the device .
this allows the ligthning current to be redirected to the ground , thereby protecting the insulation of the power line .
a typical llpd is shown in fig .
[ fig : streamer_device ] . an important characteristic of the llpd is its ability to quench the current following a lightning strike .
experiments performed on real devices have consistently demonstrated two modes of current quenching@xcite : in the case of _ current zero quenching _ ( zq ) , current will flow through the device until it passes through zero due to the vanishing grid voltage .
this results in arcing times of the order to 10 ms . in the case of _
impulse quenching _ ( iq ) , the current is suppressed within less than 0.1 ms after the lightning strike , leading to significantly less erosion of the arcing chambers and the electrodes .
the two modes of quenching are similar to the behavior of circuit breakers@xcite .
the arc voltage in high - voltage breakers is much smaller than the grid voltage and the current will continue to flow unimpeded through the arc until the next current zero , corresponding to zq . in the case of low voltage breakers ,
typically equipped with splitter plates , the total arc voltage is higher than the grid voltage and the arc current is limited , corresponding to iq .
there is one very important difference between lightning protection devices and circuit breakers , however . in the case of a circuit breaker
, the grid current has to be interrupted .
the llpd , on the other hand , only has to transmit the current from the lightning pulse .
any follow current due to the grid voltage is detrimental to the device and should be quenchend as quickly as possible .
the purpose of this paper is to provide a physical explanation of the two modes of current quenching and to develop a simulation model suitable for virtual product development .
we begin by looking at a simple circuit with the arc gap represented by the cassie - mayr arc model .
a dimensional analysis of this model leads to a simple criterion for iq in terms of the grid voltage , the maximum follow current , and the power loss of the arc . in a second step ,
we develop a complete 3d arc simulation based on the assumption of a thermal plasma and use it to simulate both zq and iq .
this model allows us to examine the shape of the arc in the real arcing chamber geometry and to study the interaction of the arc with the circuit .
the difference between impulse quenching and current zero quenching can be illustrated easily in the laboratory using the setup depicted in fig .
[ fig : circuit - arc ] . the left part of the circuit represents the oscillating grid voltage and the right part represents the arcing device @xmath3 in series with the footing resistance @xmath4 of the tower and the ground .
( 0,0 ) to[capacitor , l^=@xmath5 ( 0,4 ) ( 3,4 ) to[inductor , l^=@xmath6 ( 3,0 ) ( 0,0 ) ; ( 3,4 ) to [ resistor , l^=@xmath4 ] ( 7,4 ) to [ resistor , l^=@xmath3 ] ( 7,0 )
( 3,0 ) ; the differential equations describing this system are @xmath9 and @xmath10 where @xmath11 represents the current through the coil and @xmath12 the current through the two resistances . the resistance due to arcing in the llpd is modelled by a simple form of the cassie - mayr equation@xcite , @xmath13 this model has the advantage that it describes the arc behavior using two physically meaningful parameters : @xmath14 is the intrinsic time scale of the arc , determining how fast it responds to a change in the current , and @xmath15 is the power loss of the arc due to convection and radiation .
an often used approximation for the loss - function is to assume that the current density of the arc is constant , i.e. the radius is proportional to @xmath16 , and that the arc only loses energy from its surface@xcite .
this leads to @xmath17 with @xmath18 .
another way of understanding the importance of the exponent @xmath19 is to consider the stationary arc voltage from ( [ cmeq1 ] ) .
for @xmath20 we obtain @xmath21 the important point is that that the arc shows _ negative differential resistance _ , meaning that the arc voltage decreases with increasing current for any @xmath22 . for @xmath23 ,
the arc voltage is approximately independent of the current .
the derivation below is valid for any @xmath22 and we shall use @xmath24 for illustration . before solving the circuit equations , it makes sense to write them in dimensionless form using the circuit parameters .
the charging voltage of the capacitor @xmath1 determines the voltage level and the time scale @xmath25 is defined through @xmath26 the corresponding scale for the current follows from @xmath27 the unit for resistance will then be @xmath28 and the unit for power @xmath29 we write all physical quantities as a product of the physical scale and a dimensionless quantity according to @xmath30 with these definitions , the equations for the circuit are given by @xmath31 @xmath32 and @xmath33 [ eqdl3 ] where @xmath34 is a dimensionless constant determining power loss of the arc , i.e. , the loss - function is written as @xmath35 this new set of equations ( [ eqdl1])-([eqdl4 ] ) only contains dimensionless constants @xmath36 , @xmath34 and @xmath37 .
typical examples of zero quenching ( zq ) and impulse quenching ( iq ) are illustrated in fig .
[ fig : iq - zq ] .
the arc is ignited by switching the value of @xmath38 from a very large constant value ( in our case @xmath39 ) to @xmath40 at time @xmath41 , where the capacitor is fully charged . in the case of zq
, the arc will continue to burn until the grid voltage across the arc gap vanishes and the arc is extinguished . in the case of iq ,
the grid voltage is insufficient to maintain the arc and it vanishes at a time scale of the order to @xmath14 . [ cols="^,^ " , ]
we have analyzed the two modes of current quenching appearing in llpds and provided a theoretical explanation for the phenomena .
a dimensional analysis shows that if the cooling power of the arc is proportinal to the product of the grid voltage an the maximum follow current ( @xmath0 ) , we can guarantee impulse quenching with a very short follow current , leading to considerably less erosion of the the device .
this has important consequences for applications .
furthermore , we have demonstrated that quenching power of the arc can be simulated very well using fully resolved 3d simulations based on the magnetohydrodynamic equations for a thermal plasma .
as expected , radiation is the main mechanism by which the arc loses energy .
this means that the current - voltage characteristics of the arc are sensitive to the absorption spectrum of the plasma .
more effort will have to go into the definition and computation of effective absorption coefficients for the plasma and this will be subject of an upcoming publication .
in addition , more detailed modeling will also have to include the effects of contact erosion and ablation of wall material , changing both the pressure in the arcing chamber and the composition of the plasma .
the 3d simulations of the arcing phenomenon will make it possible to design future llpds on the computer .
10 g. v. podporkin , e.yu .
enkin , e. s. kalakutsky , v. e. pilshikov , and a. d. sivaev .
lightning protection of overhead lines rated at 335 kv and above with the help of multi - chamber arresters and insulator - arresters . in _ electromagnetic compatibility ( apemc ) , 2010 asia - pacific symposium on _ , pages 12471250 , 2010 .
g. v. podporkin , e.yu .
enkin , e. s. kalakutsky , v. e. pilshikov , and a. d. sivaev .
development of multi - chamber insulator - arresters for lightning protection of 220 kv overhead transmission lines . in _ lightning protection
( xi sipda ) , 2011 international symposium on _ , pages 160165 , 2011 . yongjoong lee , henrik nordborg , yongsug suh , and peter steimer .
arc stability criteria in ac arc furnace and optimal converter topologies . in _
pec 07 - twenty - second annual ieee applied power electronics conference and exposition _ ,
pages 12801286 , 2007 . | we develop a consistent model for a line lightning protection device and demonstrate that this model can explain the two modes of current quenching impulse quenching and current zero quenching observed in such devices .
a dimensional analysis shows that impulse quenching can always be obtained if the power loss from the electric arcs is large enough as compared to @xmath0 , where @xmath1 is the grid voltage and @xmath2 is the maximum follow current after a lightning strike .
we further show that the two modes of quenching can be reproduced in a full 3d arc simulations coupled to the appropriate circuit model .
this means the arc simulations can be used for optimization and development of future llpds .
june 2016 _ keywords _ : arc simulation , lightning protection , current quenching , radiation |
since 2001 when superconductivity in mgb@xmath0 was first discovered @xcite , many calculations of its bulk and surface electronic structure have been performed . in general , these calculations showed a good agreement with arpes experiments , suggesting that mgb@xmath0 is not a strongly correlated system , and that superconductivity is probably a consequence of some strong electron - phonon coupling mechanism .
in other words , mgb@xmath0 appears to be an extreme case of an eliashberg superconductor .
indeed , many studies show strong coupling of @xmath6 electrons and @xmath9 phonons in the b layer , and a wide superconducting gap in the @xmath6 band has been measured @xcite .
also , since the recent fabrication of pristine ultra - thin mgb@xmath0 films @xcite , surface enhanced superconductivity has become an increasingly active area of research @xcite .
these theoretical investigations greatly contributed to the understanding of the peculiar bulk / surface electronic structure of the seemingly simple mgb@xmath0 crystal .
bulk electronic structure calculations @xcite show that mgb@xmath0 has five main bands , namely completely filled @xmath1 b , partially filled @xmath10 , @xmath4 and @xmath11 b , along with completely unoccupied @xmath7 mg bands .
bonding within the b layers is mostly covalent ( @xmath12 hybridization ) and ionic / metallic between b
b layers .
due to the mg between the b planes , the nature of in - plane covalent bonding is not typical , as in graphite for example .
specifically , the presence of mg enhances the inter - layer overlap and induces strong hybridization between dangling b @xmath11 and mg @xmath7 orbitals ( @xmath2 bond ) .
this causes an upward shift of the mg @xmath7 and downward shift of the b @xmath11 band . in other words
, this induces an emptying of the mg @xmath7 and filling of the b @xmath11 bands . in this way
mg atoms donate electrons interstitially ( between b mg layers ) and become positively ionized .
this causes an additional downward shift of the b @xmath13 band relative to the b @xmath6 bands and charge transfer from the @xmath6 to the @xmath13 band .
in this way mgb@xmath0 has a small hole doping of the higher energy @xmath10 and @xmath14 bands ( at the @xmath15 point ) and bonds within the b layers become mixed covalent / metallic .
such hole doping in the b layers enhances the superconductivity .
this is similar to the case of intercalated graphites @xcite , where doping ( inter - plane intercalation ) by alkali metals causes the graphite to become superconducting .
the above description is also consistent with charge density distributions calculated in ref . , where it was seen that mg atoms are strongly ionized , but electrons are donated interstitially , rather than directly to the b layers , and inter - layer bonding is much more metallic than ionic .
also , it has been shown that less charge is participating in the in - plane b
b bonding in mgb@xmath0 compared to the c c bonding in primitive graphite .
this suggests that mg induces a covalent bonding in the b layer which is not completely saturated and is weakened , i.e. slightly metallic . on the other hand ,
the electronic structure of the mgb@xmath0 surface is much more delicate .
due to the weak dispersion of bulk b @xmath6 bands in the ( 0001 ) direction , there exist two wide gaps in the projected bulk band structure of mgb@xmath0 at the @xmath16 point .
this allows for the formation of many different types of surface states .
indeed , several studies of the mgb@xmath0(0001 ) surface @xcite showed the existence of various surface and subsurface states , whose positions and existence depend on crystal termination and coverage @xcite . for the case of a b terminated surface , every bulk band has its own surface state band which follows the upper edge of the corresponding projected bulk band .
all these states are mostly localized in the first few b layers .
for example , the surface state band for b terminated mgb@xmath0 reported in ref . has @xmath11 symmetry and is located on the upper edge of the projected @xmath11 bulk band .
it s energy at the @xmath16 point in the surface brillouin zone ( spz ) is about @xmath17 ev and its charge density is mostly distributed throughout the first three b layers . in the case of the mg terminated surface ,
there is one surface state whose energy at the @xmath16 point is about @xmath18 ev .
this surface state has @xmath2 symmetry and is mostly localized above the surface plane in the vacuum region ( supra - surface ) .
however , arpes measurements which followed @xcite showed a good agreement with bulk electronic structure calculations @xcite , although the peak at @xmath19 ev in the normal photo - emission does not agree with any of the above mentioned surface states .
the calculations of servedio _ et al . _
@xcite showed good agreement but are not particularly relevant , as the surface potential barrier was there modeled by a step potential , so that the binding energy of the surface state was strongly dependent on the position of the potential step . on the other hand , calculations by profeta _
@xcite showed that the @xmath2 surface state at the @xmath16 point strongly depends on the termination and coverage of the mgb@xmath0 crystal , indicating that in the experiment @xcite the surface termination may have been mixed or had significant surface contamination .
@xcite paid much more attention to surface preparation , and in their measurement the normal photo - emission peak appeared at @xmath20 ev . in refs . an @xmath21 monolayer ( ml ) thick ( @xmath21 ml b and @xmath21 ml mg layers ) mgb@xmath0 film was deposited an a mg(0001 ) substrate using the molecular beam epitaxy ( mbe ) co - deposition technique . following this it was determined ( using several techniques including xps , leed+xrd , x ray absorption ) that indeed , there was no surface contamination , and arpes measurements were then performed .
the results showed an excellent agreement of the projected bulk band structure with bulk calculations @xcite .
agreement with surface band structure calculations @xcite was also quite good .
for example , peaks at about @xmath22 ev and @xmath23 ev are measured which correspond to @xmath2 surface states in mg and b terminated surfaces respectively .
measurements in refs . also suggested that the most stable surface termination is mg , as was anticipated in refs . .
recent experimental realizations of the mgb@xmath0 surface have been brought to an increasingly high level of precision @xcite .
these experiments were aimed at investigating the fine superconductivity properties of mgb@xmath0 , e.g. , multiple superconductivity gaps @xcite , and surface enhanced superconductivity @xcite .
however , such clean mgb@xmath0 surfaces , with a very well defined electronic structure , can also be an excellent starting point for the investigation of both optical and dielectric mgb@xmath0 surface properties .
it also should be noted that in ref .
ultrathin mgb@xmath0 films down to @xmath24 nm were fabricated , which is quite close to the thicknesses used in the calculations shown herein .
interplay between 2d and 3d electronic structures give rise to the unique dielectric @xcite and extraordinary optical properties @xcite of the mgb@xmath0 crystal . on the other hand , because of the plethora of surface , localized and subsurface states , the mgb@xmath0 surface dielectric and optical properties may be even more interesting . before providing such theoretical calculations
we require an accurate description of the surface electronic structure .
for this reason , we will focus here on the investigation of three ( b , mg and li ) terminated mgb@xmath0 surface electronic structures by comparing the results of three different ( two plane wave and real space ) _ ab initio _ density functional theory ( dft ) methods . in so doing
, we may demonstrate how by changing the surface termination , charging , and doping the mgb@xmath0(0001 ) we may tune both the localization and energy range of the surface and subsurface states , with potential applications in the areas of plasmonics and superconductivity .
the paper is organized as follows . in sec .
[ compmeth ] we describe the _ ab initio _ methods used for the surface electronic structure calculations , the description of surface formation energies , and develop criteria to distinguish between surface , subsurface , vacuum , and bulk states .
the numerical results are presented in sec .
[ numres ] for b , mg , and li terminated surfaces , with results for each termination in separate subsections . these results are then discussed in more detail in sec .
[ conclus ] , where we show how surface termination and electro - chemical doping may be used to tune mgb@xmath0(0001 ) surface states .
this is followed by a concluding section .
structural optimization and electronic structure calculations for each mgb@xmath0(0001 ) surface have been performed by combining three different _ ab initio _ codes . specifically , we compare our own pseudopotential based plane wave dft code ( sppw ) and the plane wave self - consistent field dft code ( pwscf ) belonging to the quantum espresso ( qe ) package @xcite , with a real - space projector augmented wave function method dft code ( gpaw ) @xcite . in each case
we have employed the perdew - zunger local density approximation ( lda ) for the exchange correlation ( xc)-potential @xcite .
an electronic temperature of @xmath25 ev was used to converge the kohn - sham wavefunctions , with all energies extrapolated to 0 k. the electronic density was calculated using a @xmath26 monkhorst - pack special @xmath27-point mesh , i.e. by using @xmath28 special points in the irreducible brillouin zone . in both sppw and pwscf , lda based pseudopotentials for li , mg and b@xcite were used , and the energy spectrum was found to be converged with a @xmath29 ry plane wave cutoff . for gpaw we employed a grid spacing @xmath30 in the mgb@xmath0(0001 ) surface plane , and @xmath31 normal to the surface , which yielded converged results .
( 0001 ) surface unit cells with b , mg , and li termination , repeated twice in each direction . ]
as initial atomic coordinates we consider the system to be a hexagonal lattice with lattice parameters @xmath32 and @xmath33 taken from the bulk experimental values , @xmath34 a.u .
@xmath35 and @xmath36 a.u .
@xmath37 from ref . , which were previously shown to change by less than 0.6% upon structural optimization of the mgb@xmath0 bulk @xcite .
structural optimization is performed within qe @xcite , and the system is structurally optimized until a maximum force below @xmath38 ry / a.u .
@xmath39 ev / was obtained .
schematics of the relaxed mgb@xmath0(0001 ) surface supercell models with b , mg , and li termination are shown in fig .
[ mgb2_surfaces ] .
the b terminated surface was modeled using a supercell consisting of @xmath40 mg and @xmath41 b alternating layers , while the mg terminated surface was obtained by adding mg layers on top of each surface , with the li terminated surface formed by replacing the two mg surface layers with two li layers . for all terminations four unit cells of vacuum , i.e. @xmath42 a.u .
@xmath43 , were used to separate the surfaces .
this was found to be sufficient to ensure a good description of both surface and bulk states .
.gpaw calculated formation energies @xmath44 for b , mg , and li terminated mgb@xmath0(0001 ) surface in ev / @xmath45 and j / m@xmath45 relative to bulk mgb@xmath0 and the chemical potential of the terminating metal in the bulk for species @xmath46 , @xmath47 , taken from ref . . [ cols="<,^,^,^",options="header " , ] to understand how surface charging , or electro - chemical doping might affect mgb@xmath0(0001 ) surface states , we have directly replaced mg by li in the first atomic layer ( mg@xmath48li ) . in effect , this approximates the removal of one electron from each of the mg terminated surfaces . in other words , replacing mg by li resembles a positively charged mg terminated ( mg@xmath8 ) surface .
this allows us to directly probe the effect of surface charging of the mg terminated surface without the need for compensating background charges in the calculation .
we have then performed structural optimization of the li terminated mgb@xmath0(0001 ) surface to model how li termination changes the surface states . from table
[ table3 ] we see that the smaller atomic radius of li induces a substantial relaxation , with the relaxed li atomic layer moved down by 52.4 pm towards the b atomic layer . both the surface charging and the crystal structure modification strongly change the electronic structure of the surface states .
figure [ fig5 ] shows that the lowest b @xmath1 surface resonance band for the structurally unoptimized surface ( mg@xmath48li@xmath49mg@xmath8 ) , is slightly separated from the upper edge of the bulk b @xmath1 band .
the density distributions of such states are shown in fig .
[ fig6](a , b , c , f ) , where we see that b @xmath1 surface resonances are mainly localized in the four topmost b layers .
structural optimization causes a slight downward shift of the surface resonance b @xmath1 bands and induces a strong interaction / hybridization with the bulk b @xmath1 band , so that they decay into the bulk states .
i.e. , it increases the amplitude of the propagation of the resonance state into the bulk , as can be seen at the @xmath16 point in fig .
[ fig6](a ) . for @xmath50 and @xmath51 points
, the b @xmath1 surface resonances can not even be distinguished from the normal bulk states and are not shown .
this is seen clearly in fig .
[ fig_gpaw_li ] , with the relaxed li terminated b @xmath1 surface states becoming increasingly bulk - like between @xmath16 and k , until @xmath52 . relaxation also causes a @xmath53 ev reduction of the binding energy of the @xmath2 supra - surface state band .
this is expected because relaxation causes contraction of the topmost li - b interstitial region ( _ cf .
_ table [ table3 ] ) where the @xmath2 surface state is mostly located ( _ cf . _ fig .
[ fig6 ] ) .
the binding energy of the @xmath2 surface state at @xmath16 point is strongly reduced , from about @xmath54 ev to @xmath55 ev , comparing with the b terminated surface .
this suggests that the binding energy of the @xmath2 surface state at the @xmath16 point may be strongly influenced by surface contamination or doping of the b terminated surface , as we will discuss in sec .
[ discuss ] .
we also find the more localized b @xmath10 and @xmath14 bands are strongly modified by li termination . as can be seen in fig .
[ fig5 ] for mg@xmath48li@xmath49mg@xmath8 , the b @xmath3 and @xmath14 surface bands are always slightly separated from the upper edge of the bulk b @xmath3 and @xmath14 bands , with their density highly localized in the first one or two b layers , as seen in fig .
[ fig6](g , e , h ) .
these states resemble the highly localized b @xmath3 and @xmath14 surface states in the b terminated surface . however , surface relaxation shifts the localized b @xmath10 and @xmath14 bands down in energy .
this causes a strong interaction between localized and bulk @xmath6 states , with the surface state decaying into the bulk states .
this is clearly illustrated by the densities of these states for the relaxed li terminated structure shown in fig .
[ fig6](e , h ) . here
we see that the @xmath6 localized state now penetrates substantially into the crystal . by plotting the same distributions for an extended crystal structure ( _ cf .
_ insets of fig .
[ fig6](e , h ) ) we see that these states are actually localized state resonances or even pure bulk states at the @xmath50 point . from fig .
[ fig_gpaw_li ] we see that the b @xmath3 and @xmath14 surface states strongly hybridize with the bulk @xmath6 states , with @xmath52 over much of the sbz . from this
we make two observations .
first , removing charge from the mg terminated surface would strongly affect the surface and subsurface states .
second , for the relaxed li terminated surface all subsurface states hybridize into localized state resonances or bulk states , while the @xmath2 surface state survives . we will discuss how these affects may be understood and controlled , with the goal of tuning the surface states of mgb@xmath0 , in the following section .
point , @xmath56 versus surface termination and doping of the mgb@xmath0(0001 ) surface by b , mg@xmath48li@xmath49mg@xmath8 , li , and mg , in ev relative to the fermi level @xmath57 , for ( a ) b @xmath3 and @xmath14 , ( b ) @xmath2 , and ( c ) b @xmath1 states .
results from plane wave pwscf calculations before ( ) and after ( ) structural optimization are compared with real - space gpaw calculations for the relaxed structures ( @xmath58 ) .
the b @xmath59 and b @xmath1 bulk band regions are shaded gray
. fits to the surface state positions ( * * ) are provided as guides to the eye . ] comparing the relative positions of the b @xmath6 and @xmath2 surface states for b , mg , and li terminations , shown in figs .
[ fig_gpaw_b ] , [ fig_gpaw_mg ] , and [ fig_gpaw_li ] , respectively , clear trends emerge .
these suggest that changing surface termination or doping electro - chemically can tune both the energy and localization of the mgb@xmath0(0001 ) surface states . to demonstrate this point , we replot in fig . [ trends ] the energies of the surface and subsurface states at the @xmath16 point as a function of surface termination .
as we move from left to right in fig .
[ trends ] , charge is transferred to the b atomic layer , as we move from neutral ( b ) , to charging of less than @xmath60 ( mg@xmath48li@xmath49mg@xmath8 ) , of about @xmath60 ( li ) , and finally , of about @xmath61 ( mg ) . as charge
is donated to the b atomic layer , both the degenerate b @xmath3 and @xmath14 levels ( shown in fig . [ trends](a ) ) and b @xmath1 levels ( shown in fig . [ trends](c ) ) are filled and move down in energy relative to the bulk mgb@xmath0 fermi level , in a quasi - linear fashion .
further , when surface states cross the bulk b @xmath6 bands , they hybridize , and lose much of their weight on the surface .
this is clearly seen in fig .
[ fig_gpaw_li ] , where for the b @xmath3 and @xmath14 states @xmath62 , and the b @xmath1 state shown in fig .
[ fig6](a ) , also has more weight in the bulk .
this strongly suggests that by electro - chemically doping the surface , we may tune both the energy and localization of the b @xmath6 surface states . for the higher energy b @xmath3 and @xmath14 surface states
this is particularly interesting , as they resemble quantum well ( qw ) states .
this is because the occupying electrons effectively feel the potential of a qw centered around the b layer .
moreover , electrons in such states are trapped in the topmost b layer but are `` free '' to move within the band . in this way , the b @xmath6 band may be considered a 2d electron gas decoupled from the surrounding 3d electron gas of the bulk mgb@xmath0 .
such a two - component electron plasma allows different types of plasma dispersions relations .
for example , a noble acoustic plasmon may be present , whose dispersion resembles that of the acoustic plasmons .
such a plasmon may affect ( screen ) the electron - phonon coupling . on the other hand , for the @xmath2 state shown in fig .
[ trends](b ) , the energy of the surface state does not correlate with charging of the b atomic layer . instead , to understand the behavior of the @xmath2 state
, we must first consider its projected density @xmath63 .
for each termination we find that the @xmath2 state is an ionic / metallic bonding state between the mg / li and b atomic layers .
this is clear from figs .
[ fig2](d ) , [ fig4](a ) , and [ fig6](d ) , where we see that the @xmath2 state has mostly @xmath11 character localized in the inter - layer region .
specifically , for b termination the @xmath2 has significant weight in the bulk , while for mg and li terminations the @xmath2 state has more @xmath7 character , with significant weight in the vacuum region .
this is also clear from the band structures shown in figs .
[ fig_gpaw_b ] , [ fig_gpaw_mg ] , and [ fig_gpaw_li ] .
specifically , for b termination ( _ cf .
_ fig .
[ fig_gpaw_b ] ) the @xmath2 state is at the top of the broad @xmath13 band , having mostly @xmath11 character , and is filled throughout most of the sbz . for the mg terminated surface ( _ cf . _
[ fig_gpaw_mg ] ) the @xmath2 state has more @xmath7 character , is separated from the bulk @xmath13 bands although still following the @xmath13 bands through the sbz , is shifted up in energy , and partially emptied . when we directly replace mg by li to model a mg@xmath8 surface , we find that the @xmath2 surface state has mostly mg / li @xmath7 character ( _ cf .
_ fig .
[ fig6](d ) ) , is shifted up in energy between the @xmath13 and mg @xmath7 bulk states , and is almost completely empty . for the relaxed li
terminated surface ( _ cf . _ fig .
[ fig_gpaw_li ] ) the @xmath2 state is nearly emptied , and has significant @xmath7 character , as shown in fig . [ fig6](d ) . together
, this suggests that by altering the termination of the mgb@xmath0(0001 ) surface , the @xmath2 surface state is changed from a bulk - like state with @xmath11 character for b termination , to an increasingly localized and emptied surface state with more metallic @xmath7 character from mg termination , to mg@xmath48li ( or mg@xmath8 ) termination , and finally the structurally optimized li termination .
this is consistent with previous calculations for the na terminated mgb@xmath0(0001 ) surface @xcite , which found the @xmath2 surface state between that calculated here for mg and mg@xmath48li terminations . in effect , by varying the surface termination or electro - chemically doping the surface , one can tune both the localization and energy range of mgb@xmath0(0001 ) @xmath2 surface states . in this way
, one could then modify the density or fermi velocity of the charge carriers on the surface and consequently surface plasmons dispersions .
this means that surface plasmon characteristics may be tuned rather easily by surface charging or doping .
in this paper we have investigated mgb@xmath0(0001 ) surface and subsurface states , and the influence of surface termination ( by b , mg , and li ) and surface charging ( replacing mg by li to model mg@xmath8 ) using both plane wave and real - space _ ab initio _ methods .
generally we find for the mgb@xmath0(0001 ) surface there exist three types of surface states which are often well separated from bulk bands , namely , the b @xmath1 surface band , the @xmath2 supra - surface state band , and the b @xmath3 and @xmath64 qw bands .
if charge is added to the b layer by either changing the surface termination or surface charging , the b @xmath6 surface bands shift shift down in energy .
as these bands cross the bulk b @xmath6 bands , they lose their surface state character due to hybridization . in effect , by adjusting the charging of the b subsurface atomic layer , we may tune both the localization and energy range of the b @xmath6 surface states . on the other hand , the addition of a mg , mg@xmath8 , or li layer above the b surface layer gives the @xmath2 surface state band an increasingly metallic @xmath7 character as it is emptied .
we find these results are quite robust , and independent of the _ ab initio _ methodology employed ( plane wave or real - space ) and relaxation of the surface .
overall , these results suggest that by changing surface coverage and charging , one may tune both the density and fermi velocity of the charge carriers at the surface .
this would make possible the controlled tuning of the surface plasmon dispersions , e.g. the slope of the acoustic surface plasmon , which may then influence ( modify or enhance ) `` surface '' superconductivity .
this has profound implications for both the superconducting behaviour and surface or acoustic surface plasmon dispersion of mgb@xmath0 .
therefore , because of the diversity of mgb@xmath0 s surface electronic structure , with combinations of various types of quasi-2d and 3d plasmas , it seems that the mgb@xmath0 surface should have very unusual dielectric and optical properties . to clarify these unique properties
, our next step will be the investigation of the mgb@xmath0 surface dielectric response .
we acknowledge funding through the spanish `` juan de la cierva '' program ( jci-2010 - 08156 ) , spanish ministerio de ciencia e innovacon ( fis2010 - 21282-c02 - 01 , fis2010 - 19609-c02 - 01 ) , spanish `` grupos consolidados upv / ehu del gobierno vasco '' ( it-319 - 07 , it-366 - 07 ) , and aci - promociona ( aci2009 - 1036 ) .
the european theoretical spectroscopy facility is funded through etsf - i3 ( contract number 211956 ) .
10 , n. nakagawa , t. muranaka , y. zenitani , and j. akimitsu , , , 63 ( 2001 ) , doi : http://dx.doi.org/10.1038/35065039[10.1038/35065039 ] . ,
y. machida , t. sato , t. takahashi , h. matsui , s .- c .
wang , h. ding , a. kaminski , j. c. campuzano , s. sasaki , and k. kadowaki , , , 65 ( 2003 ) , doi : http://dx.doi.org/10.1038/nature01619[10.1038/nature01619 ] .
, z. lin , q. dai , d. li , y. wang , y. zhang , y. wang , and q. feng , , , 015013 ( 2011 ) , doi : http://dx.doi.org/0.1088/0953 - 2048/24/1/015013[0.1088/0953 - 2048/24/1/015013 ] . , m. v. schilfgaarde , and v. p. antropov , , , 092503 ( 2001 ) , doi : http://dx.doi.org/10.1103/physrevb.64.092503[10.1103/physrevb.64.092503 ] . , i. i. mazin , k. d. belashchenko , v. p. antropov , and l. l. boyer , , , 4656 ( 2001 ) , doi : http://dx.doi.org/10.1103/physrevlett.86.4656[10.1103/physrevlett.86.4656 ] . and
v. p. antropov , , , 49 ( 2003 ) , doi : http://dx.doi.org/10.1016/s0921 - 4534(02)02299 - 2[10.1016/s0921 - 4534(02)02299 - 2 ] . and
w. e. pickett , , , 4366 ( 2001 ) , doi : http://dx.doi.org/10.1103/physrevlett.86.4366[10.1103/physrevlett.86.4366 ] .
, a. d. bronnikov , o. v. zharikov , and a. v. palnichenko , , , 1445 ( 1987 ) , doi : http://dx.doi.org/10.1016/0038 - 1098(87)90355 - 3[10.1016/0038 - 1098(87)90355 - 3 ] .
, j. i. lee , b. i. min , and a. j. freeman , , , 020508 ( 2001 ) , doi : http://dx.doi.org/10.1103/physrevb.64.020508[10.1103/physrevb.64.020508 ] . ,
j. yang , j. g. hou , and q. zhu , , , 100507 ( 2002 ) , doi : http://dx.doi.org/10.1103/physrevb.65.100507[10.1103/physrevb.65.100507 ] .
, e. v. chulkov , and p. m. echenique , , , 172512 ( 2001 ) , doi : http://dx.doi.org/10.1103/physrevb.64.172512[10.1103/physrevb.64.172512 ] .
drechsler , and t. mishonov , , , 140502 ( 2002 ) , doi : http://dx.doi.org/10.1103/physrevb.66.140502[10.1103/physrevb.66.140502 ] .
, a. continenza , f. bernardini , and s. massidda , , , 184517 ( 2002 ) , doi : http://dx.doi.org/10.1103/physrevb.66.184517[10.1103/physrevb.66.184517 ] .
, k. m. shen , s. lee , a. damascelli , d. h. lu , d. l. feng , z .- x .
shen , and s. tajima , , , 157002 ( 2002 ) , doi : http://dx.doi.org/10.1103/physrevlett.88.157002[10.1103/physrevlett.88.157002 ] .
, r. macovez , m. sancrotti , l. petaccia , r. larciprete , s. lizzit , and a. goldoni , , , 976 ( 2004 ) , doi : http://dx.doi.org/10.1063/1.1777413[10.1063/1.1777413 ] . , c. cepek , m. sancrotti , a. goldoni , l. petaccia , r. larciprete , and s. lizzit , , , s3451 ( 2004 ) , doi : http://dx.doi.org/10.1088/0953 - 8984/16/33/005[10.1088/0953 - 8984/16/33/005 ] . , c. cepek , s. lizzit , r. larciprete , r. macovez , m. sancrotti , and a. goldoni , , , 12 ( 2006 ) , doi : http://dx.doi.org/10.1088/1367 - 2630/8/1/012[10.1088/1367 - 2630/8/1/012 ] .
, e. v. chulkov , p. m. echenique , and v. m. silkin , , , 224502 ( 2008 ) , doi : http://dx.doi.org/10.1103/physrevb.78.224502[10.1103/physrevb.78.224502 ] .
, a. balassis , p. m. echenique , and e. v. chulkov , , , 054521 ( 2009 ) , doi : http://dx.doi.org/10.1103/physrevb.80.054521[10.1103/physrevb.80.054521 ] . ,
w. e. pickett , r. t. scalettar , and a. g. eguiluz , , , 057001 ( 2002 ) , doi : http://dx.doi.org/10.1103/physrevlett.88.057001[10.1103/physrevlett.88.057001 ] .
, p. c. chow , o. d. restrepo , y. takano , k. togano , h. kito , h. ishii , c. c. chen , k. s. liang , c. t. chen , s. tsuda , s. shin , c. c. kao , w. ku , and a. g. eguiluz , , , 176402 ( 2006 ) , doi : http://dx.doi.org/10.1103/physrevlett.97.176402[10.1103/physrevlett.97.176402 ] .
, a. b. kuzmenko , d. van der marel , s. m. kazakov , n. d. zhigadlo , and j. karpinski , , , 104509 ( 2006 ) , doi : http://dx.doi.org/10.1103/physrevb.73.104509[10.1103/physrevb.73.104509 ] .
, a. akyurtlu , a. semichaevsky , and n. angkawisittpan , , , 195123 ( 2007 ) , doi : http://dx.doi.org/10.1103/physrevb.76.195123[10.1103/physrevb.76.195123 ] . ,
s. baroni , n. bonini , m. calandra , r. car , c. cavazzoni , d. ceresoli , g. l. chiarotti , m. cococcioni , i. dabo , a. d. corso , s. de gironcoli , s. fabris , g. fratesi , r. gebauer , u. gerstmann , c. gougoussis , a. kokalj , m. lazzeri , l. martin - samos , n. marzari , f. mauri , r. mazzarello , s. paolini , a. pasquarello , l. paulatto , c. sbraccia , s. scandolo , g. sclauzero , a. p. seitsonen , a. smogunov , p. umari , and r. m. wentzcovitch , , , 395502 ( 2009 ) , doi : http://dx.doi.org/10.1088/0953 - 8984/21/39/395502[10.1088/0953 - 8984/21/39/395502 ] .
, l. b. hansen , and k. w. jacobsen , , , 035109 ( 2005 ) , doi : http://dx.doi.org/10.1103/physrevb.71.035109[10.1103/physrevb.71.035109 ] .
, c. rostgaard , j. j. mortensen , j. chen , m. duak , l. ferrighi , j. gavnholt , c. glinsvad , v. haikola , h. a. hansen , h. h. kristoffersen , m. kuisma , a. h. larsen , l. lehtovaara , m. ljungberg , o. lopez - acevedo , p. g. moses , j. ojanen , t. olsen , v. petzold , n. a. romero , j. stausholm - mller , m. strange , g. a. tritsaris , m. vanin , m. walter , b. hammer , h. hkkinen , g. k. h. madsen , r. m. nieminen , j. k. nrskov , m. puska , t. t. rantala , j. schitz , k. s. thygesen , and k. w. jacobsen , , , 253202 ( 2010 ) , doi : http://dx.doi.org/10.1088/0953 - 8984/22/25/253202[10.1088/0953 - 8984/22/25/253202 ] . and a. zunger , , , 5048 ( 1981 ) , doi : http://dx.doi.org/10.1103/physrevb.23.5048[10.1103/physrevb.23.5048 ] . and j. l. martins , , , 1993 ( 1991 ) , doi : http://dx.doi.org/10.1103/physrevb.43.1993[10.1103/physrevb.43.1993 ] .
, , 7th ed .
( john wiley & sons , inc . , new york , n.y . , 1996 ) .
, j. i. martnez , f. calle - vallejo , j. rossmeisl , k. s. thygesen , k. w. jacobsen , and j. k. nrskov , , , 2244 ( 2011 ) , doi : http://dx.doi.org/10.1021/jp110489u[10.1021/jp110489u ] . , k. burke , and m. ernzerhof , , , 3865 ( 1996 ) , doi : http://dx.doi.org/10.1103/physrevlett.77.3865[10.1103/physrevlett.77.3865 ] . | surface state localization and hybridization on b , mg , and li terminated mgb@xmath0(0001 ) surfaces are studied within density functional theory ( dft ) . for the b
terminated surface we find a low energy b @xmath1 surface state , a @xmath2 surface state , and b @xmath3 and @xmath4 quantum well states , which are @xmath5 localized in the topmost b layer .
our results demonstrate that by charging the b atomic layer , either by changing the surface termination or through electro - chemical doping , the b @xmath6 surface states are shifted down in energy , filled , and delocalize through hybridization as they cross the bulk mgb@xmath0 bulk bands . on the other hand
, the @xmath2 surface state should be shifted up in energy , emptied , and gain an increasingly metallic @xmath7 character by adding a mg , mg@xmath8 , or li terminating atomic layer .
these results clearly show both the robust nature of mgb@xmath0(0001 ) surface states , and how their localization and energy range may be tuned by surface termination and charging , with implications for the superconducting and plasmonic behaviour of mgb@xmath0 . |
let @xmath0 be a ( strongly continuous ) unitary representation of the ( possibly infinite dimensional ) lie group @xmath1 ( with an exponential function ) .
let @xmath10 be its subspace of smooth vectors . on this space
we obtain by @xmath14 the _ derived representation _ of @xmath15 which we extend naturally to a representation of the enveloping algebra @xmath16 , also denoted @xmath17 .
we call an operator @xmath18 _ smoothing _ if @xmath19 ( @xcite ) .
a closely related concept is that of a _ schwartz operator _ , which means that , for all @xmath20 ( the enveloping algebra of the lie algebra @xmath15 of @xmath1 ) , the sesquilinear form @xmath21 on @xmath10 extends continuously to @xmath22 ( ( * ? ? ?
3.4 , p. 349 ) , @xcite ) .
this note grew out of the question to understand the relation between smoothing and schwartz operators .
this is completely answered by theorem [ thm:2.4 ] which asserts , for any smooth representation of a frchet lie group @xmath1 and @xmath23 , the following are equivalent : * @xmath24 is schwartz .
* @xmath24 and @xmath25 are smoothing . *
the map @xmath26 is smooth .
smoothing operators are of particular importance for unitary representations of finite dimensional lie groups which are _ trace class _ in the sense that , for each @xmath27 , the operator @xmath28 is trace class .
actually we show in proposition [ prop:1.6 ] that every smoothing operator is trace class if and only if @xmath7 is trace class .
this connection was our motivation to compile various characterizations of trace class representations scattered in the literature , mostly without proofs ( @xcite ) .
surprisingly , this also led us to some new insights , such as the fact that , if @xmath7 is trace class , then there exists an @xmath4 such that all operators @xmath3 , @xmath5 , are trace class . as a consequence , the corresponding distribution character @xmath6 is of finite order .
this is contained in theorem [ thm:1.3 ] which collects various characterizations of trace class representations .
one of them is that , for every basis @xmath29 of @xmath15 and @xmath30 , the positive selfadjoint operator @xmath31 has some negative power which is trace class .
this is analogous to the nelson stinespring characterization of ccr representations ( all operators @xmath3 , @xmath32 , are compact ) by the compactness of the inverse of @xmath31 .
locally compact groups for which all irreducible unitary representations are trace class have recently been studied in @xcite , and for a characterization of groups for which all irreducible unitary representations are ccr , we refer to ( * ? ? ?
* thm . 2 ) .
in the measure theoretic approach to second quantization , the fock space of a real hilbert space is realized as the @xmath33-space for the gaussian measure @xmath34 on a suitable enlargement of @xmath11 .
combining our characterization of trace class representations with results in @xcite , we see that the trace class condition is equivalent to @xmath10 being nuclear , which in turn is equivalent to the realizability of the gaussian measure on the dual space @xmath12 of distribution vectors .
* notation : * throughout this article , @xmath35 . for a unitary representation @xmath36 of @xmath1 ,
let @xmath37 for @xmath38 denote the infinitesimal generator of the one - parameter group @xmath39 by stone s theorem .
set @xmath40 and @xmath41 .
in this section @xmath1 will be a finite dimensional lie group and @xmath15 will be the lie algebra of @xmath1 .
we fix a basis @xmath29 of @xmath15 and consider the corresponding nelson
laplacian @xmath42 , considered as an element of the enveloping algebra @xmath16 .
we write @xmath43 for the @xmath44th schatten ideal in the algebra @xmath45 of bounded operators on a hilbert space @xmath11 and @xmath46 for the ideal of compact operators .
recall that a unitary representation @xmath36 is called trace class if @xmath47 for every @xmath48 . for every unitary representation @xmath0 , the subspace @xmath10 of smooth vectors can naturally be endowed with a frchet space structure obtained from the embedding @xmath49 , where @xmath50 .
its range is the closed subspace of smooth equivariant maps in the frchet space @xmath51 .
this frchet topology on @xmath10 is identical to the topology obtained by the family of seminorms @xmath52 , where @xmath53 for @xmath54 .
lem:1.1 if @xmath0 is a unitary representation of the lie group @xmath1 , then @xmath55 for every @xmath56 . in view of (
4.4 ) , the representation @xmath57 of @xmath1 on the frchet space @xmath10 is smooth .
hence , for every @xmath58 and @xmath56 , the continuous compactly supported map @xmath59 has a weak integral @xmath60 .
then , for every @xmath61 , @xmath62 and therefore @xmath63
. lem:1 let @xmath64 be a frchet space , @xmath65 be a metrizable vector space and @xmath66 be a sequence of continuous linear maps @xmath67 for which @xmath68 exists for every @xmath69 .
then @xmath70 is continuous .
since @xmath64 is a baire space and @xmath65 is metrizable , it follows from ( * ? ? ?
ix , 5 , ex .
22(a ) ) that the set of discontinuity points of @xmath70 is of the first category , hence its complement is non - empty .
this implies that @xmath70 is continuous .
the following theorem generalizes ( * ? ? ?
2.6 ) in a bourbaki expos of p. cartier which states the equivalence of ( iii ) and ( v ) , but unfortunately without giving a proof or a reference to one .
thm:1.3 for a unitary representation @xmath36 of @xmath1 , the following are equivalent : * there exists an @xmath4 such that @xmath71 and the corresponding map @xmath72 is continuous .
* @xmath73 and the map @xmath74 is continuous . *
@xmath7 is a trace class representation , i.e. , @xmath73 .
* @xmath75 .
* there exists a @xmath76 such that @xmath77 is trace class .
let @xmath78 , where @xmath79 denotes the derived representation , extended to the enveloping algebra .
recall that @xmath80 is a non - positive selfadjoint operator on @xmath11 ( @xcite ) .
\(i ) @xmath81 ( ii ) @xmath81 ( iii ) @xmath81 ( iv ) are trivial implications .
\(iv ) @xmath81 ( iii ) : according to the dixmier
malliavin theorem ( ( * ? ? ?
* thm . 3.1 ) ) , we can write every @xmath27 as a finite sum of products @xmath82 with @xmath83 . hence the assertion follows from @xmath84 .
\(iii ) @xmath81 ( ii ) : the assertion in ( * ? ? ?
1.4 ) comes close to this statement but does not assert the continuity of the @xmath85-valued map .
let @xmath86 be a @xmath87-sequence in @xmath88 , i.e. , @xmath89 and @xmath90 converges to @xmath91 in the sense that , for every @xmath92-neighborhood @xmath93 in @xmath1 , we eventually have @xmath94 .
then @xmath95 for every @xmath27 holds in @xmath96 ( and even in @xmath88 ) . for every @xmath97 ,
the linear map @xmath98 is continuous because the linear maps @xmath99 are continuous .
here we use that @xmath100 . in view of lemma [ lem:1 ] , it suffices to show that , for every @xmath27 , we have @xmath101 holds in @xmath85 . using the dixmier
malliavin theorem ( ( * ? ? ?
* thm . 3.1 ) ) , we write @xmath102 with @xmath103
. then @xmath104 since the right multiplication maps @xmath105 are continuous and @xmath106 in @xmath45 , it follows that @xmath107 for every @xmath27 .
now the assertion follows from lemma [ lem:1 ] .
\(ii ) @xmath81 ( v ) : let @xmath108 be a compact @xmath92-neighborhood in @xmath1 and @xmath109 then @xmath110 is a banach space for each @xmath111 , and the frchet space @xmath112 is the projective limit of the banach spaces @xmath110
. therefore the continuity of the seminorm @xmath113 on @xmath112 implies the existence of some @xmath4 such that the map @xmath114 extends continuously to @xmath110 .
this implies that @xmath115 .
next we observe that by an argument similar to the proof of a lemma by m. duflo ( ( * ? ? ?
* lemma 3.2.3 , p. 250 ) ) , there exists for every @xmath116 a positive integer @xmath117 , an open @xmath92-neighborhood @xmath118 in @xmath1 , and functions @xmath119 such that @xmath120 where @xmath121 is the dirac distribution in @xmath92
. then @xmath122 holds as an identity of linear operators on @xmath10 ( lemma [ lem:1.1 ] ) , and since both sides are bounded on @xmath11 , we obtain @xmath123 by the preceding argument , both summands on the right are trace class , so that @xmath124 is trace class as well .
\(v ) @xmath81 ( i ) : for @xmath27 , we have @xmath125 since the first factor on the right is trace class and @xmath126 , it follows that @xmath73 .
moreover , the continuity of the linear operator @xmath127 and the density of @xmath88 in @xmath128 imply that the identity holds for all @xmath129 .
we conclude that @xmath130 , and continuity of the integrated representation @xmath131 implies that the corresponding map @xmath132 is continuous . along the same lines one obtains the following characterization of completely continuous representations ( ccr ) from (
( nelson stinespring ) for a unitary representation @xmath36 of @xmath1 , the following are equivalent : * @xmath133 .
* @xmath134 .
* @xmath135 is a compact operator .
the equivalence of ( i ) and ( ii ) follows from the density of @xmath88 in @xmath96 .
we now use the same notation as in the preceding proof .
\(i ) @xmath81 ( iii ) : from the relation @xmath136 we derive the existence of some @xmath76 for which @xmath137 is compact , but this implies that @xmath138 is compact as well .
\(iii ) @xmath81 ( ii ) : for @xmath27 , we have @xmath139 therefore the compactness of @xmath138 implies ( ii ) . for a unitary representation @xmath0 of a lie group @xmath1 , an operator @xmath18 is called _ smoothing _ if @xmath140 .
we write @xmath141 for the subspace of smoothing operators in @xmath45 .
it is shown in ( * ? ? ?
* thm . 2.11 ) that for the class of frchet lie groups , which contains in particular all finite dimensional ones , an operator @xmath8 is smoothing if and only if it is a smooth vector for the representation @xmath142 of @xmath1 on @xmath45 . if @xmath7 is not norm continuous , then this representation is not continuous because the orbit map of the identity operator is not continuous , but it defines a continuous representation by isometries on the norm - closed subspace @xmath143 by grding s theorem , @xmath144 for every @xmath145 . applying the dixmier
malliavin theorem ( * ? ? ? * thm .
3.3 ) to the continuous representation @xmath146 , we see that @xmath147 it follows in particular that all smoothing operators are trace class if @xmath7 is a trace class representation .
alternatively one can use the factorization @xmath148 for every smoothing operator @xmath8 to see that @xmath8 is trace class because @xmath137 is trace class for some @xmath117 . from grding s
theorem we obtain another characterization of trace class representations : prop:1.6 a unitary representation @xmath0 of @xmath1 is trace class if and only if @xmath149 , i.e. , all smoothing operators are trace class .
prop:1.7 if @xmath0 is a trace class representation of @xmath1 , then the space of smoothing operators coincides with the subspace of smooth vectors of the unitary representation @xmath150 defined by @xmath142 .
since the inclusion @xmath151 is smooth , every @xmath152 has a smooth orbit map @xmath153 , hence is smoothing .
if , conversely , @xmath8 is smoothing , then shows that @xmath8 is a finite sum of operators of the form @xmath154 , @xmath27 , @xmath155 .
since @xmath156 is a continuous linear map by theorem [ thm:1.3 ] , the right multiplication map @xmath157 is continuous , and the map @xmath158 is smooth , the relation @xmath159 implies that @xmath154 has a smooth orbit map in @xmath160 .
we conclude that the same holds for every smoothing operator .
the equivalence of the statements in the first two parts of the following corollary can also be derived from the vastly more general theorem [ thm:2.4 ] , but it may be instructive to see the direct argument for trace class representations as well .
[ cor:1.8 ] for a trace class representation of @xmath1 and @xmath18 , the following are equivalent : * @xmath8 is a schwartz operator .
* @xmath8 and @xmath13 are smoothing . *
@xmath161 and the map @xmath162 is smooth .
\(i ) @xmath81
( ii ) : if @xmath8 is schwartz , then in particular the operators @xmath163 , @xmath164 , are bounded on @xmath10 , and thus from ( * ? ? ?
* thm 2.11 ) it follows that @xmath13 is smoothing .
furthermore , boundedness of @xmath165 for every @xmath164 entails in particular that @xmath166 , so that by ( * ? ? ?
* thm 2.11 ) we obtain that @xmath8 is also smoothing .
\(ii ) @xmath81 ( iii ) : next assume that @xmath8 and @xmath13 are smoothing .
then proposition [ prop:1.6 ] implies that @xmath167 , and proposition [ prop:1.7 ] implies that the maps @xmath168 are smooth . for the unitary representation of @xmath169 on @xmath160 defined by @xmath170
this implies that the matrix coefficient @xmath171 is smooth , so that @xmath8 is a smooth vector for @xmath172 by ( * ? ? ?
\(iii ) @xmath81 ( i ) : finally , assume that @xmath161 and the map @xmath173 is smooth . since the linear embedding @xmath174 is continuous , the orbit map @xmath175 is also smooth . from (
* lem 2.9 ) and ( * ? ? ? * lem 2.10 ) , and by considering suitable partial derivatives at @xmath176 of the map , we obtain boundedness of the operators @xmath177 where @xmath178 . for the last result of this section we need the following lemma , which appears in ( * ? ? ?
* thm 1.3(b ) ) without proof .
[ lem : cartier1.3 ] let @xmath36 be a unitary representation of @xmath1 and let @xmath12 denote the space of distribution vectors , i.e. , the anti - dual of @xmath10 .
then every @xmath179 is a sum of finitely many anti - linear functionals @xmath180 of the form @xmath181 , where @xmath182 and @xmath164 .
continuity of @xmath183 is straightforward .
next fix @xmath184 .
the map @xmath185 is a topological embedding , where @xmath186 is equipped with the product topology .
thus by the hahn banach theorem , we can extend @xmath70 to a continuous anti - linear functional on @xmath186 . since the continuous anti - dual of @xmath11 is identical to the continuous dual of the complex conjugate hilbert space @xmath187 , and the continuous dual of a direct product is isomorphic to the direct sum of the continuous duals
, we obtain that @xmath188 for some @xmath189 and @xmath190 .
let @xmath191 denote the space of schwartz operators of a unitary representation @xmath36 .
if @xmath36 is trace class , then from corollary [ cor:1.8 ] it follows that @xmath192 is the space of smooth vectors of the unitary representation of @xmath193 on the hilbert space @xmath160 , defined by @xmath194 . in this case
we equip @xmath192 with the usual frchet topology of the space of smooth vectors .
the next proposition characterizes the topological dual of @xmath192 .
[ prp : linfunc ] let @xmath36 be a trace class representation of @xmath1 .
every continuous linear functional on the frchet space @xmath192 of schwartz operators can be written as a sum of finitely many linear functionals @xmath195 from corollary [ cor:1.8 ] we know that the space @xmath196 of schwartz operators coincides with the space of smooth vectors of the unitary representation @xmath197 given by @xmath198 .
for @xmath199 and every smooth vector @xmath200 for @xmath172 , we have @xmath201 by lemma [ lem : cartier1.3 ] if now follows that , for every @xmath202 , we can write @xmath70 as @xmath203 where @xmath204 and @xmath205 for @xmath206 . combining theorem [ thm:1.3 ] with ( * ? ? ?
* cor . 4.18 )
we obtain : prop:3.18 for a unitary representation @xmath0 , the following are equivalent : * @xmath7 is trace class . *
the frchet space @xmath10 is nuclear .
* there exists a measure @xmath34 on the real dual space @xmath12 of @xmath10 , endowed with the @xmath207-algebra generated by the evaluations in smooth vectors , whose fourier transform is @xmath208 for @xmath58 . the main idea in the proof of (
* cor . 4.18 )
is that @xmath10 coincides with the space of smooth vectors of the selfadjoint operator @xmath209 and that properties ( b ) and ( c ) can now be investigated in terms of the spectral resolution of this operator .
the equivalence of ( a ) and ( b ) is also stated in ( * ? ? ?
2.6 ) without proof .
in this section we prove a characterization of schwartz operators in terms of smoothing operators , namely that @xmath24 is schwartz if and only if @xmath24 and @xmath25 are smoothing for any smooth unitary representation of a frchet lie group .
we shall need the following result from interpolation theory ( ( * ? ? ?
9 , p. 44 ) : prop : inter1 let @xmath11 be a hilbert space and @xmath210 be positive selfadjoint operators on @xmath11 with possibly unbounded inverses .
suppose that the bounded operator @xmath211 satisfies @xmath212 then @xmath213 with @xmath214 we consider a smooth unitary representation @xmath0 of the ( locally convex ) lie group @xmath1 and we assume that @xmath1 has a smooth exponential function .
the next lemma provides an equivalent definition of schwartz operators . for @xmath215 and @xmath97 ,
we consider the selfadjoint operator @xmath216 note that ( * ? ? ?
* lemma 4.1(b ) ) implies that @xmath217 coincides with the closure of the operator @xmath218 on @xmath10 .
prop:1 if @xmath23 is a smoothing operator whose adjoint @xmath25 is smoothing as well , then @xmath24 is a schwartz operator , i.e. , for @xmath20 , the operators @xmath219 defined on @xmath10 are bounded , i.e. , extend to bounded operators on @xmath11 .
since @xmath16 is spanned by the elements of the form @xmath220 , @xmath215 , we have to show that , for @xmath221 and @xmath222 , the operator @xmath223 is bounded . from ( * ? ? ?
* thm . 2.11 ) we know that @xmath224 is bounded , and from ( * ? ? ?
2.8(a ) ) it follows that @xmath225 is bounded on @xmath226 .
next we observe that the operators @xmath227 are all bounded .
writing @xmath228 as @xmath229 , we obtain the estimate @xmath230 therefore proposition [ prop : inter1 ] implies that for @xmath231 we have @xmath232 for @xmath233 , this leads to @xmath234 so that @xmath235 is bounded on @xmath236 .
as @xmath237 is bounded , it follows that the following operator is bounded : @xmath238 and this implies the boundedness of @xmath223 , more precisely @xmath239 we now consider the representation of @xmath169 on @xmath45 by @xmath240 rem:1 ( a ) suppose that @xmath8 is a continuous vector for the left multiplication representation @xmath70 and also for the right multiplication action @xmath241 .
then @xmath242 implies that @xmath8 is a continuous vector for @xmath172 .
we write @xmath243 for the closed subspace of @xmath172-continuous vectors in @xmath45 and note that since @xmath172 acts by isometries , it defines a continuous action of @xmath1 on the banach space @xmath243 .
\(b ) suppose that @xmath8 is a @xmath244-vector for @xmath70 and @xmath241 and @xmath221 .
since all operators @xmath245 are contained in @xmath243 , the closedness of @xmath243 implies that @xmath246 and @xmath247 are also @xmath172-continuous .
we claim that @xmath8 is a @xmath244-vector for @xmath172 .
in fact , the map @xmath248 is partially @xmath244 , so that its differential @xmath249 exists .
this map is given by @xmath250 since @xmath172 defines a continuous action on @xmath243 , the continuity of @xmath251 follows from the continuity of the corresponding linear map @xmath252 which follows from the assumption that @xmath8 is a @xmath244-vector for @xmath70 and @xmath241 .
this shows that @xmath251 is continuous and hence that @xmath8 is a @xmath244-vector for @xmath172 .
thm:2.4 for a smooth unitary representation of a frchet lie group @xmath1 and @xmath23 , the following are equivalent : * @xmath24 and @xmath25 are smoothing . * @xmath24 is schwartz . *
@xmath24 is a smooth vector for @xmath172 , i.e. , the map @xmath253 , @xmath254 is smooth .
that ( i ) implies ( ii ) is proposition [ prop:1 ] .
\(ii ) @xmath81 ( iii ) : for @xmath255 , the operators @xmath256 and @xmath257 on @xmath10 are bounded , so that ( * ? ? ?
* thm . 2.11 ) implies that @xmath24 and @xmath25 are smoothing , hence in particular @xmath244-vectors for @xmath172 by remark [ rem:1 ] and @xmath258 it follows in particular that @xmath259 is schwartz as well ( because @xmath24 is schwartz if and only if @xmath260 is bounded on @xmath10 for every @xmath261 ) .
thus we obtain inductively that @xmath262 for every @xmath97 . since @xmath1 is frchet , ( * ? ? ?
1.6(ii ) , cor .
1.7 ) now imply that @xmath24 is a smooth vector for @xmath172 .
\(iii ) @xmath81 ( i ) follows from the characterization of smoothing operators ( ( * ? ? ?
* thm . 2.11 ) ) . if the lie group @xmath1 is only assumed to be metrizable , the additional quantitative information from proposition [ prop:1 ] can still be used to obtain the equivalence of ( i ) and ( iii ) in the preceding theorem .
this is done in theorem [ thm:2.5 ] below .
first we need a lemma .
rmk - hadi - xy let @xmath36 be a smooth unitary representation of a ( locally convex ) lie group @xmath1 with a smooth exponential map .
let @xmath263 be a schwartz operator , and set @xmath264 with domain @xmath10 , where @xmath265
. then @xmath8 is bounded , and therefore @xmath266 .
we now show that @xmath267 . indeed for @xmath182 ,
if @xmath268 is a sequence such that @xmath269 in @xmath11 , then @xmath270 and from boundedness of @xmath271 with domain @xmath10 ( recall that @xmath24 is schwartz ) it follows that the sequence @xmath272 is convergent . but @xmath273 is closed , hence @xmath274 .
[ thm:2.5 ] let @xmath36 be a smooth unitary representation of the lie group @xmath1 and assume that @xmath15 is metrizable . for @xmath23 ,
the following are equivalent : * @xmath24 and @xmath25 are smoothing .
* @xmath24 is a smooth vector for @xmath172 , i.e. , the map @xmath253 , @xmath254 is smooth .
\(ii ) @xmath81 ( i ) follows from ( * ? ? ?
* thm . 2.11 ) .
now assume that @xmath24 and @xmath25 are smoothing .
then @xmath24 is a schwartz operator by proposition [ prop:1 ] . according to we
have @xmath275 by ( * ? ? ?
* thm . 2.11 )
the map @xmath276 is smooth with @xmath277 in particular , @xmath278 is @xmath117-linear and continuous .
similarly the smoothness of @xmath279 shows that @xmath280 is continuous in @xmath281 .
therefore entails that @xmath282 is bounded for @xmath283 in a neighborhood of @xmath284 . since @xmath16 is spanned by the elements of the form @xmath285 , @xmath215 , @xmath286 , and @xmath287
is bilinear , polarization implies that the @xmath288-linear map @xmath289 @xmath290 is bounded near @xmath291 and therefore continuous for every @xmath292 . next we show that ( the unique extension to @xmath11 of ) @xmath260 lies in @xmath243 for every @xmath265 .
the proof is inductive , namely , we assume that @xmath293 where @xmath264 , and we show that for all @xmath294 , the unique extensions to @xmath11 of @xmath295 and @xmath296 are in @xmath243 .
remark [ rmk - hadi - xy ] and ( * ? ? ?
2.9 ) imply that @xmath297 for any @xmath298 , and @xmath299 . now
an argument similar to remark [ rem:1](b ) yields @xmath300 .
furthermore , @xmath301 and it is straightforward to verify that @xmath302 where @xmath303 is the principal anti - involution of @xmath16 defined by @xmath304 for @xmath298 .
since obviously @xmath25 is schwartz , the operator @xmath13 is the unique extension to @xmath11 of the bounded operator @xmath305 , hence for any @xmath306 we have by remark [ rmk - hadi - xy ] that @xmath307 . now
* lem 2.8(a ) ) yields boundedness of @xmath308 , and ( * ? ? ? *
lem 2.10 ) implies that @xmath309 .
next we observe that for @xmath310 , the partial derivatives of @xmath311 exist and are continuous ( see ( * ? ? ?
* lemmas 2.9/10 ) and remark [ rmk - hadi - xy ] , and recall from above that for @xmath312 , the operator @xmath308 is bounded ) .
this yields @xmath313 and @xmath314 hence we can prove @xmath315 by induction .
the continuity of the maps @xmath316 and ( * ? ? ? * cor .
1.7(ii ) ) now implies that @xmath24 is a smooth vector for @xmath172 .
recall that @xmath192 denotes the space of schwartz operators of a unitary representation @xmath36 .
the next proposition is an application of theorem [ thm:2.4 ] .
[ prp : funccalc ] let @xmath36 be a smooth unitary representation of a frchet lie group @xmath1 .
let @xmath202 .
assume that @xmath200 is a non - negative self - adjoint operator .
then @xmath317 .
since @xmath318 is self - adjoint , by theorem [ thm:2.4 ] it is enough to show that it is smoothing .
next choose @xmath54 such that @xmath319 .
let @xmath303 denote the principal anti - involution of @xmath16 , defined by @xmath304 for @xmath298 .
then @xmath320 thus the operator @xmath321 is bounded on @xmath10 . from (
* thm 2.11 ) it follows that @xmath318 is smoothing .
schwartz operators have also been studied in @xcite for nilpotent lie groups , and more generally in @xcite . note that from ( * ? ? ?
2.11 ) it follows that there is redundancy in the definitions given in ( * ? ? ?
1.2 ) and ( * ? ? ?
3.1 ) . from (
3.1 ) it follows that smooth vectors of the @xmath193-action on @xmath160 are schwartz operators .
this is weaker than theorem [ thm:2.4 ] above .
furthermore , ( * ? ? ?
3.1 ) gets close to proposition [ prop:1.6 ] and corollary [ cor:1.8](iii ) , but in @xcite it is not proved that being trace class is equivalent to nuclearity of the space of smooth vectors ( see proposition [ prop:3.18 ] ) .
finally , proposition [ prop:1.6 ] implies that what is proved in ( * ? ? ?
3.1 ) for irreducible unitary representations of nilpotent lie groups indeed holds for all trace class representations of general finite dimensional lie groups . in this section
we investigate the connection between our results and those of @xcite more closely .
in particular , we will show that several of the results of @xcite are special cases of the results of our paper , when applied to the schrdinger representation .
let @xmath322 be a @xmath323-dimensional real symplectic space and let @xmath324 denote the heisenberg group associated to @xmath322 , that is , @xmath325 with the multiplication @xmath326 let @xmath327 denote the lie algebra of @xmath324 , and let @xmath328 denote the universal enveloping algebra of @xmath327 . by the stone
von neumann theorem , to every nontrivial unitary character @xmath329 we can associate a unique irreducible unitary representation @xmath330 of @xmath324 for which the center acts by @xmath331 . in the schrdinger realization , @xmath330 acts on the hilbert space @xmath332 , where @xmath333 is a polarization of @xmath64 , and @xmath334 is the lebesgue measure on @xmath335 .
the action of @xmath330 is given by @xmath336 and @xmath337 , where @xmath338 , @xmath339 , @xmath340 , and @xmath341 .
the following result is a special case of the general theory of unitary representations of nilpotent lie groups ( e.g. , see @xcite ) .
prop : pichi the representation @xmath330 is trace class , the space of smooth vectors of @xmath330 is the schwartz space @xmath342 , and @xmath343 is equal to the algebra of polynomial coefficient differential operators on @xmath344 . from proposition [ prop : pichi ]
it follows that the operators defined in ( * ? ? ?
3.1 ) are the schwartz operators for @xmath330 in the sense of our paper . from corollary [ cor:1.8 ]
it follows that @xmath345 is the space of smooth vectors for the action of @xmath346 on @xmath160 , and therefore it can be equipped with a canonical frchet topology .
it is straightforward to verify that this frchet topology is identical to the one described in ( * ? ? ?
next we show that if @xmath347 , then @xmath348 for every @xmath349 .
this is proved in ( * ? ? ?
* lemma 3.5(b ) ) , but the argument that will be given below applies to any trace class representation . from proposition [ prop:1.7 ]
it follows that the map @xmath350 given by @xmath351 is smooth .
since the bilinear map @xmath352 is continuous , the map @xmath353 is also smooth .
thus by proposition [ prop:1.7 ] , the operator @xmath354 is smoothing .
a similar argument shows that @xmath355 is also smoothing , and corollary [ cor:1.8 ] implies that @xmath354 is schwartz .
proposition [ prop:1.6 ] implies that every schwartz operator for @xmath330 is trace class .
this is also obtained in ( * ? ? ?
* lemma 3.6 ) .
theorem [ thm:2.4 ] applied to @xmath330 gives ( * ? ? ?
* thm . 3.12 ) .
proposition [ prp : funccalc ] implies ( * ? ? ?
3.15 ) , and proposition [ prp : linfunc ] implies ( * ? ? ?
5.12 ) .
it is possible that the relation between the weyl transform and schwartz operators that is investigated in ( * ? ? ?
3.6 ) is a special case of more general results in the spirit of our paper , at least for nilpotent lie groups .
finally , it is worth mentioning that the paper @xcite studies ( among several other things ) the class of representations of infinite dimensional lie groups with the property that their space of smooth vectors is nuclear . by proposition [ prop:3.18 ] , when @xmath1 is finite dimensional this condition is equivalent to the representation being trace class . in the infinite dimensional case
this is an interesting class of representations which deserves further investigation .
we hope to come back to these problems in the near future .
cartier , p. , _ vecteurs diffrentiables dans les representations unitaires des groupes de lie _ , sem .
bourbaki , vol .
1974/75 , exp .
454 , lect .
notes math . * 514 * , springer , berlin heidelberg new york , 1976 | in this note we collect several characterizations of unitary representations @xmath0 of a finite dimensional lie group @xmath1 which are trace class , i.e. , for each compactly supported smooth function @xmath2 on @xmath1 , the operator @xmath3 is trace class . in particular we derive the new result that , for some @xmath4 , all operators @xmath3 , @xmath5 , are trace class . as a consequence the corresponding distribution character @xmath6 is of finite order .
we further show @xmath7 is trace class if and only if every operator @xmath8 , which is smoothing in the sense that @xmath9 , is trace class and that this in turn is equivalent to the frchet space @xmath10 being nuclear , which in turn is equivalent to the realizability of the gaussian measure of @xmath11 on the space @xmath12 of distribution vectors .
finally we show that , even for infinite dimensional frchet lie groups , @xmath8 and @xmath13 are smoothing if and only if @xmath8 is a schwartz operator , i.e. , all products of @xmath8 with operators from the derived representation are bounded .
+ msc2010 : 22e45 , 22e66 |
magnetic flux tubes existing in the photosphere and lower chromosphere of the sun are considered to be narrow bundles of strong magnetic field lines that rapidly expand with height in the solar atmosphere ( e.g. , solanki 1993 ) .
the fundamental modes of linear oscillations of these flux tubes are typically identified with longitudinal , transverse and torsional tube waves ( e.g. , defouw 1976 ; roberts & webb 1978 ; roberts 1979 , 1981 ; spruit 1981 , 1982 ; priest 1982 ; hollweg 1985 ; roberts 1991 ; roberts & ulmschneider 1997 ) , with the two latter waves being alfvn - like waves .
observational evidence for the existence of alfvn - like waves in different regions of the solar atmosphere was given by high - resolution observations performed by the solar optical telescope ( sot ) and the x - ray telescope ( xrt ) onboard the hinode solar observatory . according to de pontieu et al .
( 2007 ) and cirtain et al .
( 2007 ) , signatures of alfvn waves were observed by the sot and xrt instruments , respectively . moreover , alfvn waves were also reported by tomczyk et al .
( 2007 ) , who used the coronal multi - channel polarimeter of the national solar observatory .
interpretations of these observations were given by van doorsselaere et al .
( 2008 ) and antolin et al .
( 2009 ) , who concluded that the reported observational results describe kink waves . as of today
, the most convincing observational evidence for the existence of alfvn waves in the quiescent solar atmosphere was given by mcintosh et al .
( 2011 ) , who reported indirect evidence for such waves found in observations by solar dynamic observatory ( sdo ) . moreover , fujimura & tsuneta ( 2009 ) used sot observations to study fluctuations in pores and intergranular magnetic structures , and concluded that such fluctuations could be explained by the existence of longitudinal ( sausage ) and transverse ( kink ) waves propagating along magnetic flux tubes embedded in the solar photosphere .
they found oscillation periods of 3 - 6 min for the pores and 4 - 9 min for the intergranular magnetic elements .
more recently , okamoto and de pontieu ( 2011 ) used data from sot and presented observational evidence for the existence of high - frequency transverse waves propagating along spiclues in the solar atmosphere .
similar high - frequency tranverse waves were reported by yurchyshyn et al .
( 2012 ) , who observed type ii spicules using the new solar telescope .
there is also observational evidence for the existence of torsional alfvn waves in the solar atmosphere as reported by jess et al .
( 2009 ) , who interpreted data obtained with high spatial resolution by the swedish solar telescope ( sst ) .
alfvn - like motions were also observed by bonet et al .
( 2008 ) , who found vortex motions of g band bright points around downflow zones in the solar photosphere , and by wedemeyer - bhm & rouppe van der voort ( 2009 ) , who used data from the sst to demonstrate that more disorganized relative motions of photospheric bright points can also induce swirl - like motions in the solar chromosphere .
more recently , observational evidence for alfvn - like waves was found by bonet et al .
( 2010 ) , who used the ballon - borne sunrise telescope , and by de pontieu et al .
( 2012 ) , who used the sst to find torsional motions in spicules .
moreover , magnetic swirls in the solar atmosphere were reported by wedemeyer - bhm et al .
( 2012 ) . among the above described observational results , the observations perfomed by fujimura & tsuneta ( 2009 ) , okamoto and de pontieu ( 2011 ) , and yurchyshyn et al .
( 2012 ) are the most relevant because the authors directly refer to transverse tube waves , which are the main topic of this paper .
our goal here is to determine the propagation conditions for these waves by calculating the so - called cutoff frequencies . in general
, there are global cutoff frequencies , which are the same along the entire length of the tube , and local cutoff frequencies that are height dependent .
if , under certain simplifying approximations , the cutoff frequency is constant along the entire length of the tube , it is called a global cutoff .
otherwise , if it is height dependent , it is a local cutoff . in section 6 , we compare these cutoff frequencies to the observational results .
the global cutoff frequency represents the natural frequency of linear oscillations of the magnetic flux tubes , and its value restricts the wave propagation to only those frequencies that are higher than this cutoff . for isothermal and thin flux tubes , the global cutoff frequencies for longitudinal and transverse tube waves were first determined by defouw ( 1976 ) and spruit ( 1981 ) , respectively .
these global cutoff frequencies are ratios of the characteristic wave speeds to the pressure ( density ) scale height . since the scale height and the wave speeds are constant along isothermal and exponentially expanding magnetic flux tubes , the resulting cutoffs are the same along the entire length of the tubes . the fact that there is no global cutoff frequency for torsional waves propagating along isothermal and thin magnetic flux tubes
was demonstrated by musielak , routh & hammer ( 2007 ) . a method to derive the global cutoff frequency for longitudinal tube waves
was introduced by rae & roberts ( 1982 ) and musielak et al .
( 1987 , 1995 ) , who demonstrated that the wave equation for these waves can be transformed into its standard form ( also referred to as the klein - gordon equation ) , which directly displays the global cutoff frequency .
a similar method was used by musielak & ulmschneider ( 2001 , 2003 ) to determine the global cutoff frequency for transverse tube waves .
these global cutoff frequencies are important in studies of atmospheric oscillations ( e.g. , roberts 1991 ; hasan & kalkofen 1999 ; hasan 2003 ; musielak & ulmschneider 2003 ; hasan 2008 ) and chromospheric and coronal heating ( e.g. , narain & ulmschneider 1996 ; ulmschneider & musielak 2003 ) .
the existence of global cutoff frequencies is restricted to thin and isothermal magnetic flux tubes .
however , in more general cases when the flux tubes are either wide and isothermal , or thin and non - isothermal , or wide and non - isothermal , the resulting cutoff frequencies depend on atmospheric height , which means that they are local quantities ( e.g. , routh , musielak & hammer 2007 ) .
a method to determine such local cutoff frequencies was developed by musielak , musielak & mobashi ( 2006 ) .
this method requires that wave equations are cast in their standard forms ( i.e.,without terms with the first - order derivatives ) and that the oscillation theorem ( e.g. , kahn 1990 ) is used to obtain the cutoff frequencies .
the form of the oscillation theorem used in this paper is that given by kahn ( 1990 ) .
different forms of this theorem were considered by mathematicians depending on specific differential equations and the applied boundary conditions .
the basic idea was originally introduced by sturm ( 1836 ) , who considered a second - order ordinary differential equation written in its standard form and established a comparison criterion , which allowed him to determine whether solutions to the equation were oscillatory or not , without formally solving the equation .
another comparison criterion was developed by kneser ( 1893 ) and later generalized by fite ( 1918 ) , hille ( 1948 ) , wintner ( 1949 , 1957 ) , and many others .
the most important generalization of kneser s criterion was done by leighton ( 1950 , 1962 ) . in more recent work ,
some oscillation theorems were established for linear ( e.g. , li & yeh 1995 , 1996 ) and nonlinear ( e.g. , li 1998 ; lee et al .
2005 ; tyagi 2009 ) second - order differential equations .
the oscillation theorem presented by kahn ( 1990 ) , which is used in this paper , is equivalent to kneser s criterion , which is also known as kneser s theorem ( e.g. , bohner & nal 2005 )
. first applications of this theorem to solar physics problems were done by musielak & moore ( 1995 ) , who considered the propagation of linear alfvn waves in an isothermal solar atmosphere .
then , schmitz & fleck ( 1998 ) used the theorem to establish criteria for acoustic wave propagation in the solar atmosphere .
more recently the theorem was used by schmitz & fleck ( 2003 ) , musielak et al .
( 2006 , 2007 ) , routh et al .
( 2007 , 2010 ) , and hammer et al .
( 2010 ) . in this paper
, we use this method to derive the cutoff frequency for transverse waves propagating along a thin and non - isothermal magnetic flux tube embedded in the solar atmosphere .
the effects of temperature gradients on the cutoff frequency are studied for several power - law temperature models as well as for the reference mean solar atmosphere model c given by vernazza , avrett & loeser ( 1981 ) .
the height dependence of the cutoff frequency in these models is calculated , and it is shown that the value of this cutoff at a given atmospheric height determines the frequency that transverse tube waves must exceed in order to be propagating at this height .
the results are compared to those previously obtained for a thin and isothermal magnetic flux tube .
we also briefly discuss implications of our results for the energy and momentum balance of the solar atmosphere .
our paper is organized as follows : the governing equations and derivation of our wave equations are given in section 2 ; the global cutoff frequency for linear transverse waves propagating along a thin and isothermal magnetic flux tube is obtained in section 3 ; the local cutoff frequency for the wave propagation along a thin and non - isothermal magnetic flux tube is derived in section 4 ; the local cutoff frequencies for different power - law temperature distributions and for the val c solar atmosphere model are presented and discussed in sections 5 and 6 , respectively ; and conclusions are given in section 7 .
we consider a thin and non - isothermal magnetic flux tube that is embedded in the solar atmosphere . the tube axis is assumed to have a circular cross - section and an unperturbed tube axis oriented vertically along the @xmath0-axis , so that gravity @xmath1 .
the tube density , pressure , temperature , and magnetic field are respectively given by @xmath2 , @xmath3 , @xmath4 , and @xmath5 .
moreover , the density , pressure and temperature of the external non - magnetic ( @xmath6 ) atmosphere are represented by @xmath7 , @xmath8 and @xmath9 , respectively .
we assume that the tube is in thermal equilibrium with its surroundings , which means that at each atmospheric height @xmath10 . in the thin flux tube approximation ( e.g. , ferriz - mas et al .
1989 ) , the horizontal pressure balance for the tube becomes @xmath11 the equation for transverse linear oscillations of the tube was originally derived by spruit ( 1981 ) , who obtained @xmath12 where @xmath13 is the horizontal displacement of the tube , and @xmath14 is the characteristic wave speed given by @xmath15 the relationship between the displacement @xmath13 and the corresponding magnetic field perturbation @xmath16 was obtained by stix ( 1991 ) and bogdan et al .
( 1996 ) , and it can be written as @xmath17 defining the velocity perturbation @xmath18 and using @xmath19 where @xmath20 is the pressure scale height , with @xmath21 being the ratio of specific heats and @xmath22 being the speed of sound given by @xmath23 , we write equation ( [ s2eq2 ] ) in the following form @xmath24 then we combine equations ( [ s2eq2 ] ) and ( [ s2eq4 ] ) , use @xmath25 , and obtain @xmath26 { { \partial b_x } \over { \partial z } } = 0\ . \label{s2eq8}\ ] ] the derived wave equations describe the propagation of linear transverse waves along a thin and non - isothermal magnetic flux tube .
since the wave equations for @xmath27 and @xmath16 are different , there is a phase difference between the wave variables .
physical consequences of the existence of this phase shift will be explored in the next sections .
we now consider the special case of a thin and isothermal magnetic flux tube by taking @xmath28 = const , @xmath29 = const and @xmath30 , which gives @xmath14 = const and @xmath31 = const .
this allows us to write the wave equations ( see equations ( [ s2eq7 ] ) and ( [ s2eq8 ] ) ) as @xmath32 and @xmath33 comparison of the above wave equations shows that they are different , which means that the behavior of the wave variables @xmath27 and @xmath16 is different even for a thin and isothermal flux tube . the wave equation for @xmath27 was originally obtained by spruit ( 1981 , 1982 ) and he showed that the propagation of transverse waves along a thin and isothermal magnetic flux tube is affected by a cutoff frequency . a new result presented in this paper
is that we also derive the wave equation for @xmath16 , show that it is different than the one obtained for @xmath27 , and demonstrate how to obtain the cutoff frequency for both wave variables . to obtain the cutoff frequency ,
we follow musielak & ulmschneider ( 2001 ) and cast equations ( [ s3eq1 ] ) and ( [ s3eq2 ] ) into their standard ( or klein - gordon ) forms ( e.g. , roberts 1981 ; rae & roberts 1982 ) . the required transformations are @xmath34 and @xmath35 , and the resulting wave equations can be written as @xmath36\ [ v_1 ( z , t ) , b_1 ( z , t)]\ = 0\ , \label{s3eq3}\ ] ] where @xmath37 is called here spruit s cutoff frequency . since @xmath14 = const and @xmath31 = const in the isothermal case , @xmath38 is also constant , which means that it is a global quantity . with the coefficients of equation ( [ s3eq3 ] ) being constant
, we can make fourier transforms in time and space , and derive the global dispersion relation : @xmath39 , where @xmath40 is the wave frequency and @xmath41 is the wave vector along the tube axis . based on this dispersion relation , @xmath38 is the global cutoff frequency for transverse tube waves . according to the dispersion relation , the waves are propagating when @xmath42 and @xmath43 is real , and non - propagating when either @xmath44 with @xmath45 or @xmath46 with @xmath43 being imaginary ; in the latter case , the waves are called evanescent waves . according to musielak et al .
( 2006 , 2007 ) and petukhov ( 2006 ) , gradients of restoring forces , which can introduce characteristic scale heights , are responsible for the origin of cutoff frequencies for different waves . using this argument ,
musielak et al . (
2006 ) argued about the origin of the acoustic cutoff frequency and musielak et al .
( 2007 ) showed that the propagation of torsional tube waves is cutoff - free .
moreover , petukhov ( 2006 ) demonstrated that horizontal , non - uniform but potential magnetic fields have no influence on a cutoff frequency for magneto - acoustic waves and , as a result , the cutoff is identical to lamb s cutoff frequency , @xmath47 , which was obtained by lamb ( 1908 , 1975 ) for acoustic waves propagating in a stratified and isothermal medium .
the results obtained here for transverse tube waves clearly show that the vertical and non - uniform ( exponentially diverging ) magnetic field of a thin and isothermal flux tube is responsible for the origin of spruit s cutoff frequency .
actually , the existence of this cutoff is caused by the horizontal pressure balance , which relates the tube magnetic pressure to the gas pressure . since the latter introduces the pressure scale height @xmath31 that also determines the rate with which the field diverges with the atmospheric height , the tube magnetic field and its characteristic scale height are the main physical reasons for the existence of spruit s cutoff frequency .
finally , we may relate the spruit cutoff @xmath38 to the lamb cutoff @xmath48 by writing @xmath49 which shows that the ratio of @xmath14 to @xmath22 determines how much @xmath38 differs from @xmath48 . for the special case @xmath50
, we have @xmath51 .
hence , @xmath38 can be much larger than @xmath48 when @xmath52 and much smaller when @xmath53 .
we now consider a thin but non - isothermal magnetic flux tube whose internal temperature distribution is @xmath54 , and assume that at each atmospheric height the tube is in thermal equilibrium with its surroundings , which means that @xmath55 .
the presence of temperature gradients makes both @xmath14 and @xmath31 to be functions of @xmath0 , and this leads to wave equations for @xmath27 and @xmath16 that have non - constant coefficients ( see equations ( [ s3eq1 ] ) and ( [ s3eq2 ] ) ) . since these wave equations can not be solved by using fourier transforms in space , a different method to obtain cutoff frequencies is required .
such a method was originally developed by musielak et al .
( 2006 ) for acoustic waves propagating in non - isothermal media .
routh et al . ( 2007 , 2010 ) and hammer et al . (
2010 ) used the method to determine cutoff frequencies for the propagation of torsional waves in thick ( isothermal ) and thin ( non - isothermal ) magnetic flux tubes . here
, this method will be used to derive local cutoff frequencies for transverse waves propagating along a thin flux tube with different temperature profiles .
the method is based on the oscillation theorem given by kahn ( 1990 ) , which follows the original work of sturm ( 1836 ) and kneser ( 1893 ) .
actually , there are numerous oscillation theorems developed by mathematicians for various differential equations and their boundary conditions ( e.g. , swanson 1968 ; teschl 2011 , and references therein ) .
however , most of these theorems can not be directly applied to euler s equation ( e.g. , wong 1996 ; aghajani & roomi 2012 ) , which is used in this paper .
fortunately , solutions to euler s equations are well - known ( e.g. , murphy 1960 ) , and they form the basis for our method used in this paper . the method begins with introducing a new variable @xmath56 the physical meaning of this variable becomes obvious after both sides of the above equation are integrated ( see below ) . then , @xmath57 is the actual wave travel time between a height at which a wave source is located and a given height @xmath0 along the axis of the magnetic flux tube .
we express equations ( [ s2eq7 ] ) and ( [ s2eq8 ] ) in terms of the variable @xmath58 and obtain the following transformed wave equations @xmath59 v_x ( \tau , t)\ = 0 \label{s5eq2}\ ] ] and @xmath60 b_{x } ( \tau , t)\ = 0\ , \label{s5eq3}\ ] ] where @xmath61 , @xmath62 and @xmath63 . to convert the transformed wave equations into their standard forms , we use @xmath64 \label{s5eq4}\ ] ] and @xmath65\ , \label{s5eq5}\ ] ] and obtain @xmath66 v ( \tau , t ) = 0 \label{s5eq6}\ ] ] and @xmath67 b ( \tau , t ) = 0\ , \label{s5eq7}\ ] ] where @xmath68 and @xmath69 with @xmath70 and @xmath71 .
the frequencies @xmath72 and @xmath73 are known as the critical frequencies ( musielak et al .
1992 , 2006 ) .
we make the fourier transform in time @xmath74 = [ \tilde v ( \tau ) , \tilde b ( \tau ) ] e^{- i \omega t}$ ] , where @xmath40 is the wave frequency , and write equations ( [ s5eq6 ] ) and ( [ s5eq7 ] ) as @xmath75 \tilde v ( \tau ) = 0 \label{s5eq10}\ ] ] and @xmath76 \tilde b ( \tau ) = 0\ .
\label{s5eq11}\ ] ] using the oscillation theorem given by kahn ( 1990 ) and comparing the above equations to euler s equation ( e.g. , murphy 1960 ) , we obtain the following conditions for the wave propagation @xmath77 and @xmath78 note that applications of the oscillation theorem to other wave propagation problems relevant to solar physics were previously considered by musielak & moore ( 1995 ) and schmitz & fleck ( 1998 ) ; see also routh et al . ( 2010 ) for more recent results .
the idea of using euler s equation to determine turning - point frequencies was first introduced by musielak & moore ( 1995 ) .
we follow musielak et al .
( 2006 ) and define the turning - point frequencies as @xmath79 and @xmath80 where @xmath58 is the actual wave propagation time ( see routh et al .
2010 ) and is given by @xmath81 with @xmath82 being an integration constant to be evaluated when flux tube models are specified ( see sections 6 and 7 ) .
it is important to point out that the turning - point frequencies were obtained by making the derived wave equations ( see equations ( [ s5eq10 ] ) and ( [ s5eq11 ] ) ) equivalent to euler s equation , so that the well - known solutions of the latter can be directly applied to these wave equations .
having obtained the turning - point frequencies ( see equations ( [ s5eq14 ] ) and ( [ s5eq15 ] ) ) , we now express them in terms of the @xmath0 variable , which requires using equation ( [ s5eq16 ] ) to convert @xmath58 to @xmath0 .
we begin with the critical frequencies @xmath83 and @xmath84 given by equations ( [ s5eq8 ] ) and ( [ s5eq9 ] ) . using the following expressions @xmath85 @xmath86 and @xmath87 we obtain @xmath88 - { 1 \over 2}c_\mathrm{k } \frac { d^{2 } c_\mathrm{k } } { d z^{2 } } + { 1 \over 4}\left(\frac{d c_\mathrm{k}}{dz}\right)^2\ , \label{s5eq20}\ ] ] and @xmath89 + \frac{c_\mathrm{k}}{2 h } \frac{dc_\mathrm{k } } { dz } \\
& + { 1 \over 2}c_\mathrm{k}\frac{d^{2 } c_\mathrm{k}}{d z^{2 } } + { 1 \over 4 } \left(\frac{d c_\mathrm{k}}{dz}\right)^2\ , \end{split}\ ] ] we now use equation ( [ s5eq16 ] ) to express the conditions for wave propagation and the turning - point frequencies as functions of @xmath0 , and obtain @xmath90 > { 1 \over 4 } \left [ \int^z { { d \tilde z } \over { c_\mathrm{k } ( \tilde z)}}\ + \tau_c \right ] ^{-2 } \label{s5eq22}\ ] ] and @xmath91 > { 1 \over 4 } \left [ \int^z { { d \tilde z } \over { c_\mathrm{k } ( \tilde z)}}\ + \tau_c \right ] ^{-2}\ . \label{s5eq23}\ ] ] the turning - point frequencies are @xmath92^{-2 } \label{s5eq24}\ ] ] and @xmath93^{-2}\ .
\label{s5eq25}\ ] ] we shall use the turning - point frequencies given above to determine the cutoff frequency for transverse tube waves . we follow musielak et al .
( 2006 ) and take the larger of the two turning - point frequencies to be the cutoff frequency @xmath94 , @xmath95\ .
\label{s5eq26}\ ] ] our selection of @xmath94 is physically justified by the fact that in order to have propagating transverse tube waves , the wave frequency @xmath40 must always be higher than any turning - point frequency . in other words
, the choice guarantees that propagating wave solutions are obtained for both wave variables , and that the cutoff frequency does separate the propagating and non - propagating wave solutions ( see also hammer et al .
hence , the condition for propagating waves is @xmath96 .
based on our definition of the turning - point frequencies , the condition for non - propagating ( evanescent ) waves is @xmath97 .
the above results show that the cutoff frequency can only be determined when a flux tube model is specified , so that the turning - point frequencies can be obtained ( see sections 5 and 6 ) .
in addition , one must keep in mind that the conditions given by equation ( [ s5eq26 ] ) must be checked at each height because in some regions along the tube @xmath98 could be larger than @xmath99 , and the opposite could be true in other regions ( see section 6 ) .
let us consider the following temperature distribution inside the tube @xmath100 where @xmath101 is the distance ratio , with @xmath102 being a fixed height in the model , @xmath103 is the temperature at @xmath102 , and @xmath104 can be any real number .
note that in all power - law models discussed below a wave source is assumed to be located at @xmath105 , which means that in all calculations @xmath106 .
in addition , for all models we take @xmath107 km , @xmath108 k ,
@xmath109 km / s , and for gravity we take its solar value .
the resulting temperature distributions for @xmath104 being a positive integer that ranges from @xmath110 to @xmath111 are presented in figure 1 . to describe the process of deriving a local cutoff frequency , we begin with the simplest case of @xmath112 , which corresponds to the temperature varying linearly with @xmath13 .
we calculate @xmath113 , @xmath114 , @xmath115 and @xmath14 as functions of @xmath13 , and use equation ( [ s5eq16 ] ) to obtain @xmath116 where @xmath117 is the value of @xmath14 at @xmath102 and @xmath82 is the integration constant . to determine this constant , we assume that @xmath118 , which gives @xmath119 .
we express the turning - point frequencies given by equations ( [ s5eq24 ] ) and ( [ s5eq25 ] ) in terms of @xmath13 .
this gives @xmath120 \xi^{-1}\ , \label{s6eq3}\ ] ] where @xmath121 , with @xmath122 being the scale height at @xmath102 , and @xmath123 and @xmath124 \xi^{-1}\ .
\label{s6eq5}\ ] ] the difference between the turning - point frequencies is @xmath125 which shows that @xmath126 is larger than @xmath127 . hence , @xmath98 becomes the local cutoff frequency , so we write @xmath128 and @xmath129^{1/2 } \xi^{-1/2}\
. \label{s6eq7}\ ] ] the local cutoff frequency @xmath94 is plotted as a function of @xmath13 in figure 2 .
it is seen that the cutoff frequency decreases with the atmospheric height in the model with @xmath112 . in this case , @xmath130 is given by @xmath131 where @xmath82 is the integration constant determined from the assumption that @xmath118 .
this gives @xmath132 .
the turning - point frequencies are @xmath133 \label{s6eq9}\ ] ] and @xmath134\ , \label{s6eq10}\ ] ] where @xmath135 here @xmath136 is always larger than @xmath137 .
thus , we choose @xmath98 as the local cutoff frequency and write @xmath138 , or @xmath139^{1/2}\ .
\label{s6eq12}\ ] ] this cutoff frequency is plotted as a function of @xmath13 in figure 2 , which shows that the cutoff remains almost constant in the model with @xmath140 . in this general case of @xmath141
, we obtain @xmath142 where the integration constant @xmath82 is evaluated by taking @xmath143 ; note that our choice of @xmath144 gives the same physical parameters at @xmath145 for all the power - law models . after evaluating @xmath82 ,
we calculate @xmath146 \ , \end{split}\ ] ] where @xmath147 and @xmath148 \ .
\end{split}\ ] ] we now calculate the difference between these turning - point frequencies and obtain @xmath149 since in this model @xmath150 , @xmath141 , and @xmath151 , @xmath152 is larger than @xmath153 .
this indicates that the local cutoff frequency is @xmath154 , or @xmath155^{1/2}\ .
\end{split}\ ] ] the cutoff frequency calculated for the power - law temperature models with @xmath156 , @xmath157 and @xmath111 is plotted versus the distance ratio in figure 2 .
it is seen that this cutoff frequency always increases with the atmospheric height in the models with @xmath141 and that its increase is much faster for higher values of @xmath104 .
the effects of different temperature gradients on the cutoff frequency for transverse tube waves are presented in figure 2 .
since the cutoff frequency is a local quantity , its value at a given atmospheric height determines the frequency that the waves must exceed in order to be propagating waves at this height .
our results demonstrate that the conditions for wave propagation strongly depend on the temperature profiles .
if the temperature increases linearly with height ( @xmath112 ) , the cutoff frequency starts with its maximum at @xmath145 and then decreases with height .
for the temperature model with @xmath140 , the cutoff frequency remains practically constant with height .
however , the cutoff frequency always increases with height in the temperature models with @xmath158 ; the higher the value of @xmath104 , the steeper an increase of the local cutoff frequency with height is observed .
the main purpose of using the power - law temperature models was to demonstrate the dependence of the local cutoff frequency on the increasing steepness of the temperature models .
obviously , the power - law models do not properly describe the temperature gradients in the solar atmosphere .
therefore , we now consider a more realistic model of the solar atmosphere .
we assume that a thin flux tube is embedded in the reference mean solar atmosphere model val - c ( vernazza et al . 1981 ) . in this model
, the height @xmath159 km corresponds to unity optical depth , where the temperature is @xmath160 k. beyond the temperature minimum @xmath161 k , which is located at @xmath162 km , the model extends through the chromosphere well into the transition region to the corona , up to a temperature of @xmath163 k. at the base of the transition region the val - c model exhibits a plateau with a temperature around @xmath164 k , caused by ly@xmath165 emission .
this plateau was later shown to disappear if the physics of the transition region is treated more realistically ( fontenla et al .
therefore we restrict ourselves to the height range from @xmath166 km ( optical depth unity ) to the just below the plateau ( @xmath167 km ) . to calculate the characteristic wave speed @xmath14 as a function of @xmath0 , we have to know the magnetic field @xmath168 , which is not given in the val - c model .
since the tube is assumed to be in in thermal equilibrium with its surroundings at each atmospheric height , we use the local value of the pressure scale height , which is the same outside and inside the tube , to evaluate @xmath169 ; the calculation begins with an assumed surface magnetic field @xmath170 g ( see equation ( [ s2eq1 ] ) ) . having obtained the distribution of the tube magnetic field with height , we then calculate the gas pressure and density inside the tube , and the characteristic wave speed @xmath171 ( see equation ( [ s2eq3 ] ) ) . in the lower panel of figure 3
we plot @xmath14 versus the atmospheric height in the model ; for comparison we also plot the sound speed @xmath22 .
the results show that @xmath14 is smaller than @xmath22 in almost the entire model , except in the upper chromosphere and lower transition region , where @xmath14 becomes comparable to , or even slightly larger than , @xmath22 . in order to calculate the wave propagation time @xmath58
, we use equation ( [ s5eq16 ] ) . for the integration
constant @xmath172 we choose a value of 100 s , thus assuming that the waves have traveled for 100 s before they reach the base @xmath102 of the model .
for typical wave speeds ( cf . lower panel of figure 3 ) this means that the waves are generated about two scale heights below @xmath102 . to determine the cutoff frequency @xmath173 according to section 4.5 , we calculate both turning - point frequencies @xmath174 and @xmath175 and select the larger one as @xmath176 ( see equations ( [ s5eq20 ] ) , ( [ s5eq21 ] ) , and ( [ s5eq24])([s5eq26 ] ) )
to perform these calculations , we must evaluate the first derivative of @xmath14 and @xmath31 , and the second derivative of @xmath14 .
these calculations have to be done numerically because the model consists of tabulated data .
the upper panel of figure 3 shows the resulting cutoff frequency @xmath176 as well as spruit s cutoff frequency @xmath177 ( cf .
equation ( [ s3eq4 ] ) ) , where the latter is again treated here as a local , height - dependent quantity .
a striking property of figure 3 is that the cutoff frequency for the nonisothermal case , @xmath176 , is wiggly .
this is mostly caused by the dependence of the turning - point frequencies on the second derivative of the tube speed @xmath14 ( cf .
equations ( [ s5eq20 ] ) ([s5eq21 ] ) ) . since the val - c model is specified by a number of tabulated distinct points , the curves in figure 3 must be plotted by finding interpolating or fitting functions through these points and determining the first derivative of the scale height , and the first and second derivatives of the tube speed , in order to calculate the terms in equations ( [ s5eq20])([s5eq21 ] ) .
such derivatives of interpolated curves tend to show a wiggly behavior .
as shown in the middle and lower panels of figure 3 , these wiggles are stronger near the temperature minimum and in particular at the foot of the transition region , where the gradients change rapidly ( see also fawzy & musielak 2012 ) .
one could avoid the wiggles by approximating the data by simple functions with known smooth derivatives .
there are also automatic algorithms that can smooth out such wiggles more or less completely ( e.g. , chartrand 2011 ) .
however , we show them here explicitly in order to illustrate the dependence of the cutoff frequency , via these derivatives , on the nature of the solar atmosphere , which in reality is less homogeneous than the val - c model and in addition temporally variable . therefore , we chose the well - proven algorithm by reinsch ( 1967 ) , which fits the data points by a natural cubic spline , thus giving consistent first and second derivatives .
a moderate amount of smoothing can be applied , so that the spline needs not go exactly through the points .
since the value of @xmath94 is different at each atmospheric height , the frequency @xmath40 of a transverse tube wave must be higher than the cutoff at a given height @xmath0 in order to be propagating at this height .
overall , the new cutoff frequency @xmath94 is seen to be slightly higher than @xmath38 throughout the chromosphere .
this is mostly caused by the fact that the chromospheric temperature increases with height , leading to a positive first derivative of the tube speed @xmath14 , which increases both turning - point frequencies ( cf .
equations ( [ s5eq20 ] ) and ( [ s5eq21 ] ) ) and thus also their maximum , the cutoff frequency @xmath94 ( equation ( [ s5eq26 ] ) ) .
another difference between the spruit cutoff and our new result lies in the behavior at the foot of the transition region . here
@xmath38 decreases with height ( which becomes more evident if one continues the calculation to higher levels of the val - c model ) , while @xmath94 increases rapidly as we approach the steep temperature rise in the transition region .
formally this means that a wave with arbitrary frequency tends to lie above @xmath38 at the transition region ( implying propagation in the case of an isothermal treatment of the cutoff ) , but below our value ( implying reflection off the temperature gradient in our nonisothermal treatment ) .
this may have implications for the contribution of such transverse waves to the heating and dynamics of the overlying corona .
however , we put not much weight on such differences in the high chromosphere and transition region , fow two reasons .
first , we have not provided a complete study of the reflection properties of the transition region , which would require a more complex analysis of the complete solar atmosphere . and
second , a basic approximation underlying all these theories breaks down much earlier , already in the mid chromosphere - namely , that the flux tubes can be approximated as thin .
the thin flux tube approximation , which is the basis for the results presented in this paper and for spruit s results as well , is no longer valid at those heights .
thus the results obtained for the upper atmosphere must be taken with caution .
we have extended the calculations up to the foot of the transition region in order to determine the approximate behavior of the cutoffs in these layers , even though this is beyond the formal limit of validity of our analytical results , because the presented results demonstrate that the temperature gradient in the upper solar chromosphere and in the solar transition region will have a major influence on the propagation of transverse tube waves .
the increase of temperature with height requires atmospheric heating , which is typically identified with acoustic - gravity and flux tube waves , including transverse tube waves , or with phenomena related to magnetic reconnection ( e.g. , priest 1982 ; narain & ulmschneider 1996 ; ulmschneider & musielak 2003 ) .
our results presented in figure 3 give constraints on the range of frequencies of transverse tube waves that are propagating in different parts of the solar atmosphere .
our results show that in the upper photosphere and lower chromosphere @xmath178 , or the cutoff period is 500 s , and that in the middle and upper chromosphere @xmath179 , which corresponds to the cutoff period of 350 s. as already mentioned in section 1 , fujimura & tsuneta ( 2009 ) observed transverse tube waves with periods of 240 - 540 s , or the corresponding frequency range @xmath180 , in intergranular magnetic elements .
comparison of these observational results to our theoretical cutoff frequencies shows that most of the observed waves are freely propagating in the solar photosphere and lower chromosphere because @xmath181 , however , in the middle and upper parts of the solar chromosphere the waves with frequencies @xmath182 are not propagating .
now , the high frequency waves with periods of 45 s , or @xmath183 , observed by okamoto & de pontieu ( 2011 ) along spicules do satisfy the condition @xmath184 and they are obviously freely propagating waves in the solar chromosphere .
similarly , the high frequency oscillations with periods ranging from 30 s to 180 s , which corresponds to the frequency range @xmath185 , detected by yurchyshyn et al .
( 2012 ) in type ii spicules also satisfy the condition @xmath184 , and therefore they are freely propagating waves in the solar chromosphere .
the above constraints can be used to assess the role of transverse tube waves in the heating and dynamics of the solar atmosphere . as discussed by hasan & kalkofen ( 1999 ) , musielak & ulmschneider ( 2003 ) , and hasan ( 2003 )
, transverse tube waves may be responsible for the excitation of solar atmospheric oscillations observed in magnetically active regions near sunspots .
the results obtained in this paper can be used to determine the natural frequency of the solar atmosphere inside thin and non - isothermal magnetic flux tubes and the effects of temperature gradients on solar atmospheric oscillations .
our method to determine the cutoff frequency for transverse waves propagating along a thin and non - isothermal flux tube requires that integral transformations are used to cast the wave equations for both wave variables in the standard forms .
then the conditions for the existence of propagating wave solutions are established by using the oscillation theorem .
the theorem is also used to obtain the turning - point frequency for each wave variable .
the larger among the two turning - point frequencies is selected as the cutoff frequency .
to study the effects of temperature gradients on the cutoff frequency , we used different power - law temperature distributions .
for the temperature that increases linearly with height , the cutoff frequency is large at the bottom of the model and then decreases with height .
for the temperature profile with the power of @xmath186 , the cutoff frequency remains practically constant with height .
however , for powers higher than @xmath186 , the cutoff frequency always increases with height ; and the higher the power , the steeper the increase of the cutoff frequency with height
. we also calculated the cutoff frequency @xmath94 as a function of height @xmath0 in the val - c model and compared it to spruit s global cutoff frequency @xmath187 that was treated as a height - dependent quantity .
the comparison shows that @xmath94 exceeds @xmath187 , and that the differences are especially prominent in the upper parts of the model , where however the thin flux tube approximation is not valid any longer .
the differences in the lower atmosphere , where the thin flux tube approximation is valid , may be important for the energy carried by transverse tube waves from the solar convection zone , where the waves are generated , to the overlying solar atmosphere , where the wave energy is deposited .
the cutoff frequency @xmath94 calculated for the val - c model gives constraints on the range of frequencies of transverse tube waves that are propagating in different parts of the solar atmosphere .
we compared @xmath94 to the range of frequencies determined from observations by fujimura & tsuneta ( 2009 ) , okamoto & de pontieu ( 2011 ) , and yurchyshyn et al .
based on this comparison , we established that most of the waves observed by fujimura & tsuneta ( 2009 ) in intergranular magnetic elements are freely propagating in the solar photosphere and lower chromosphere , however , in the middle and upper solar chromosphere the lower frequency part of these waves corresponds to non - propagating waves . on the other hand , all the waves detected by okamoto & de pontieu ( 2011 ) and yurchyshyn et al .
( 2012 ) in spicules are freely propagating waves in the solar chromosphere .
these constraints are important for understanding the role played by transverse tube waves in the heating of the solar atmosphere and in the acceleration of plasma , e.g. in spicules and the solar wind .
our results can be used to determine the natural frequency of the solar atmosphere inside thin and non - isothermal flux tubes and to study the effects of the temperature stratification on the excitation of atmospheric oscillations inside solar magnetic flux tubes .
qqq aghajani , a. , & roomi , v. 2012 , turk . j. math .
36 , 273 antolin p. , shibata k. , kudoh t. , shiota d. , & brooks d. 2009 , in asp conf .
415 , the second hinode science meeting : beyond discovery - toward understanding , ed .
b. lites , m. cheung , t. magara , j. mariska , & k. reeves ( san francisco : asp ) , 247 bogdan , t. j. , hindman , b. w. , cally , p. s. , & charbonneau , p. 1996 ,
apj , 465 , 406b bohner , m. , & nal , m. 2005 , j. phys .
a : math . gen .
, 38 , 6729 bonet , j. a. , marquez , i. , sanchez almeida , j. , cabello , i. , & domingo , v. 2008 , apj 687 , 431 bonet , j. a. , et al .
2010 , apj , 723 , l139 chartrand , r. 2011 , isrn appl math , vol .
2011 , 164564 cirtain , j. w. , et al .
2007 , science , 318 , 1580 defouw , r. j. 1976 , apj , 209 , 226 de pontieu , b. , et al .
2007 , science , 318 , 1574 de pontieu , b. , et al .
2012 , apj , 752 , l12 fawzy , d. & musielak , z. e. , 2012 , mnras , 412 , 159 ferriz - mas , a. , schssler , m. , & anton , v. 1989 , a&a , 210 , 425 fite , w. b. 1918 , trans .
soc . , 19 , 341 fontenla , j. m. , avrett , e. h. & loeser , r. , 1990 , apj , 355 , 700 fujimura , d. , & tsuneta , s. 2009 , apj , 702 , 1443 hammer , r. , musielak , z. e. , & routh , s. 2010 , an , 331 , 593 hasan , s. s. , 2003 , in lectures on solar physics , ed .
h. m. antia , a. bhatnagar , & p. ulmschneider ( lecture notes in physics , vol .
619 ; berlin : springer ) , 173 hasan , s. s. , 2008 , adspr , 42 , 86 hasan , s. s. , & kalkofen , w. 1999 , apj , 519 , 899 hille , e. 1948 , trans .
amer . math .
soc . , 64 , 234 hollweg j. v. 1985 , in : advances in space plasma physics , b. buti , ed . , world scientific publ . , singapore , p. 77
jess , d. b. , mathioudakis , m. , erdelyi , r. , crokett , p. j. , keenan , f. p. , & christian , d. j. 2009 , science 323 , 1582 kahn , p. b. 1990 , mathematical methods for scientists and engineers , john wiley , new york kneser , a. 1893 , math .
ann . , 42 , 409 lamb , h. 1908 , proc .
soc . 2 , 122 lamb , h. 1975 , hydrodynamics , cambridge uni .
press , cambridge lee , c .- f .
, yeh , c .- c . , & gau , c .- y .
2005 , czechoslovak math .
j. , 55 , 845 leighton , w. 1950 , duke math .
j. 17 , 57 leighton , w. 1962 , proc .
amer . math .
soc . 13 , 603 li , w. 1998 , j. math . anal .
217 , 1 li , h. j. , & yeh , c.c .
1995 , proc .
, 125a , 1193 li , h. j. , & yeh , c.c .
1996 , math .
nachr . , 182 , 295 mcintosh , s. w. , et al .
2011 , nature 475 , 477 murphy , g. m. 1960 , ordinary differential equations and their solutions , d. van nostrand company , new york musielak , z. e. , fontenla , j. m. , & moore , r. l. 1992 , phys .
fluids , 4 , 13 musielak , z. e. , & moore , r. l. 1995 , apj , 452 , 434 musielak , z. e. , musielak , d. e. , & mobashi , h. 2006 , phys . rev .
e , 73 , 036612 - 1 musielak , z. e. , rosner , r. , gail , h. p. , & ulmschneider , p. 1995 ,
apj , 448 , 865 musielak , z. e. , rosner , r. , & ulmschneider , p. 1987 , in lecture notes in physics 291 : cool stars , stellar systems and the sun , eds .
j. l. linsky & r. e. stencel , berlin : springer , 66 musielak , z. e. , routh , s. , & hammer , r. 2007 , apj , 659 , 650 musielak , z. e. , & ulmschneider , p. 2001
, a&a , 370 , 541 musielak , z. e. , & ulmschneider , p. 2003 ,
a&a , 406 , 725 narain , u. , ulmschneider , p. 1996
, space sci .
, 75 , 453 okamoto , t. j. , & de pontieu , b. , 2011 , apj , 736 , l24 petukhov , m. yu . 2006 , astron .
32 , 418 priest , e. r. 1982 , solar magnetohydrodynamics , dordrecht - reidel rae , i. c. , & roberts , b. 1982 , apj , 256 , 761 reinsch , c. r. 1967 , num .
, 10 , 177 roberts , b. 1979 , sol . phys . ,
61 , 23 roberts , b. 1981 , in physics of sunspots , l. e. cram & j. h. thomas , eds .
, sacramento peak , new mexico , 360 roberts , b. 1991 , geophys . astrophys .
fluid dynamics , 62 , 83 roberts , b. , & ulmschneider , p. 1997 , in : solar and heliospheric plasma physics , simett g. m. , alissandrakis c. e. , vlahos l. , eds . , springer - verlag , berlin , p. 75 roberts , b. , & webb , a.r .
1978 , sol .
56 , 5 routh , s. , musielak , z. e. , & hammer , r. 2007 , sol .
246 , 133 routh , s. , musielak , z. e. , & hammer , r. 2010 , apj , 709 , 1297 schmitz , f. , & fleck , b. 1998 , a&a 337 , 487 schmitz , f. , & fleck , b. 2003 , a&a 399 , 723 solanki , s. k. 1993 , space sci .
, 63 , 1 spruit , h. c. 1981 , a&a 98 , 155 spruit , h. c. 1982 , sol .
phys . 75 , 3 stix , m. 1991 , the sun - an introduction , 2nd ed .
( berlin : springer ) sturm , c. 1836 , j. math .
pures appl . , 1 , 106 swanson , c. a. 1968 , comparison and oscillation theory of linear differential equations , academic press , new york and london teschl , g. 2011 , ordinary differential equations and dynamical systems , ( providence , ri : ams ) tomczyk , s. , et al . 2007 , science , 317 , 1192 tyagi , j. 2009 , electronic j. diff .
eqs 19 , 1 ulmschneider , p. , & musielak , z.e .
2003 , in 21st nso / sp workshop on current theoretical models and future high resolution solar observations : preparation for atst , a.a .
pevtsov and h. uitenbrock ( eds ) , asp conf . ser .
, 286 , 363 van doorsselaere t. , nakariakov v. m. , verwichte e. , 2008 , apj , 676 , l73 vernazza , j. e. , avrett , e. h. , & loeser , r. , 1981 , apjs , 45 , 635 wedemeyer - bhm , s. , & rouppe van der voort , l. , 2009 , a&a , 507 , l9 wedemeyer - bhm , s. , et al .
2012 , nature 486 , 505 wintner , a. 1949 , quat .
, 7 , 115 wintner , a. 1957 , math .
scand . 5 , 255 wong , j. s. w. 1996 , meth .
analysis 3 , 476 yurchyshyn , v. , kilcik , a. , & abramenko , v. 2012 , arxiv:1207.6417v1 | it is a well - established result that the propagation of linear transverse waves along a thin but isothermal magnetic flux tube is affected by the existence of the global cutoff frequency , which separates the propagating and non - propagating waves . in this paper ,
the wave propagation along a thin and non - isothermal flux tube is considered and a local cutoff frequency is derived .
the effects of different temperature profiles on this local cutoff frequency are studied by considering different power - law temperature distributions as well as the semi - empirical val c model of the solar atmosphere .
the obtained results show that the conditions for wave propagation strongly depend on the temperature gradients
. moreover , the local cutoff frequency calculated for the val c model gives constraints on the range of wave frequencies that are propagating in different parts of the solar atmosphere .
these theoretically predicted constraints are compared to observational data and are used to discuss the role played by transverse tube waves in the atmospheric heating and dynamics , and in the excitation of solar atmospheric oscillations . |
in the absence of a solar antineutrino flux , the superkamiokande ( sk ) background is expected to be isotropic . in the presence of solar antineutrinos , positrons emitted by i@xmath4d contribute to the background , which , due to the angular dependence of i@xmath4d cross section , should have a non - zero angular slope proportional to the antineutrino flux .
a linear fit to the counting yield , @xmath12 ( in the angular region where events from the @xmath2 interactions can be neglected , see fig .
[ fig1 ] ) , provides the antineutrino flux @xmath13 through the relation : @xmath14 where @xmath15 is the number of free protons , t is the exposure time , @xmath16 is the ( assumed constant ) detection efficiency , @xmath17 the minimal detectable antineutrino energy and @xmath18 is the cross section averaged over the antineutrino spectrum for @xmath19 . in order to provide a quantitative illustration of the previous points we used data from the first 101.9 operational days of sk , as reported in fig.3 of @xcite , corresponding to @xmath20 .
the reported background does not show an angular dependence . according to equation ( 3 ) , to extract an upper limit on the solar antineutrino flux
, we must know the average cross section @xmath18 , which can be determined within two approches : @xmath21_assuming _ that the antineutrino spectrum has the same shape as that of @xmath22b solar neutrinos , one has @xmath23 .
this gives , as a final result , @xmath24 , to the 95% c.l .
this bound corresponds to a fraction x=3.5% of the solar neutrino flux ( in the energy range @xmath25 ) predicted by the ssm @xcite .
@xmath26 as @xmath27 is an increasing function of @xmath11 , one has @xmath28 .
this lower limit to @xmath18 gives a _ model independent _ bound @xmath29 to the 95% c.l .
we remark two points of the method just presented : * the sensitivity to antineutrinos increases as statistics accumulates . within three years of data taking
, the sensitivity to @xmath1 transition probability will reach the 1% level , thus allowing for a definite test of several theoretical models . *
the determination of the angular slope @xmath30 provides a mean for _ detecting _ antineutrinos from the sun ( and not only for deriving upper bounds ) .
a non vanishing slope for sk background would be , in fact , a clear signature of a solar antineutrino flux . .
if a type ii supernova explosion occurs near the center of our galaxy about 4000 @xmath0 events and 300 @xmath31 events will be produced in the superkamiokande detector @xcite .
the detection of supernova neutrinos and antineutrinos will provide important information on the supernova mechanism .
in particular we remark that : * from the strong correlation of the elastic scattering events one can reconstruct the supernova direction with an accuracy of about four degrees @xcite ; * from a fit to experimental data one can determine the number of neutrinos and antineutrinos events , thus providing important inputs to check supernova modelling .
the anisotropy of i@xmath4d is so small that it looks reasonable to neglect it when discussing the detector sensitivity @xcite . however , let us investigate how it can affect the data analysis . with this aim
we have simulated a supernova explosion in the galactic plane , assuming that it produces @xmath32 and @xmath33 events in the detector .
we took into account a @xmath34 angular resolution @xcite and repeated the simulation two - hundred times .
we analysed the data looking for the galactic longitude @xmath35 of the supernova and the ratio @xmath36 of reconstructed events , see fig.2 .
if the analysis is performed neglecting the anisotropy ( i.e. @xmath37 ) the supernova direction is correctly determined within the predicted accuracy , however the reconstructed @xmath38 is badly underestimated ( @xmath39 , the correct value being @xmath40 ) and the fit is poor ( @xmath41 on the average of the 200 simulations , using 20 angular bin ) .
we checked that when the anisotropy is included , both the supernova direction and the ratio @xmath38 are correctly reconstructed , and the quality of the fit improves significantly .
these features are clearly understood : * as the positron angular distribution is symmetrical around the supernova direction , it does not alter its reconstruction . *
the depleted forward positron production mimics a reduced number of @xmath2 events .
99 g. fiorentini , m. moretti , f.l .
villante , phys .
b , in press , 1997 y. totsuka , `` first result from super - kamiokande '' , presented at texas symposium ( 1996 ) .
j.n . bahcall and m.h .
pinsonneault , rev .
61 ( 1992 ) 885 . k. nakamura ,
superkamiokande , presented at 3rd nestor international workshop , pylos , greece , 1993 | the anisotropy of the positrons emitted in the reaction @xmath0 has to be taken into account for extracting an antineutrino signal in superkamiokande . for the sun
, this effect allows a sensitivity to @xmath1 transition probability at the 3% level already with the statistics collected in the first hundred days . for a supernova in the galaxy ,
the effect is crucial for extracting the correct ratio of @xmath2 to @xmath3 events .
as well known , the specific signature of antineutrinos in hydrogen containing materials is through the inverse beta decay ( i@xmath4d ) , @xmath0 , which produces _ almost isotropically _ distributed monoenergetic positrons ( @xmath5 ) . for energy above a few mev , the differential cross section is : @xmath6 where : @xmath7 @xmath8 ( @xmath9 ) is the vector ( axial ) coupling of the neutron and @xmath10 is the total cross section for antineutrino energy @xmath11 .
the goal of this paper is to assess the importance of the angular dependence of the cross section in two situations of practical interest : * the search for solar antineutrinos ; decay or magnetic moment .
_ see @xcite for details and references _ ] * the detection of a supernova in the galaxy . |
for a proper scheme @xmath1 over a perfect field , the picard scheme @xmath2 representing the functor @xmath3 exists , and its connected component @xmath4 is separated and of finite type ( * ? ? ?
* ii 15 ) . by chevalley
s structure theorem @xcite , the reduced connected component @xmath5 is an extension of an abelian variety @xmath6 by a linear algebraic group @xmath7 : @xmath8 the commutative , smooth affine group scheme @xmath7 is the direct product of a torus @xmath9 and a unipotent group @xmath10 .
the following theorem completely characterizes @xmath9 : [ torus ] if @xmath11 is proper over a perfect field , then the cocharactermodule @xmath12 of the maximal torus of @xmath2 is isomorphic to @xmath13 as a galois - module . to analyze the unipotent part
, we let @xmath14)_{[1]}$ ] be the typical part , i.e. the subgroup of elements @xmath15 of @xmath14)$ ] such that the map @xmath16\to x[t]$ ] , @xmath17 sends @xmath15 to @xmath18 .
[ npic ] let @xmath11 be proper over a perfect field .
then @xmath14)_{[1]}$ ] is isomorphic to the group of morphisms of schemes @xmath19 satisfying @xmath20 for every @xmath21 .
in particular , @xmath22)_{[1]}$ ] , and this is an equality in characteristic @xmath23 . to get another description of @xmath10 , we assume that @xmath11 is reduced ( the map on the picard scheme induced by the map @xmath24 is well understood by the work of oort @xcite ) .
the semi - normalization @xmath25 is the largest scheme between @xmath11 and its normalization which is strongly universally homeomorphic to @xmath11 in the sense that the map @xmath25 induces an isomorphism on all residue fields .
a theorem of traverso @xcite implies that @xmath14)_{[1]}$ ] , hence @xmath10 , vanishes if @xmath11 is reduced and seminormal .
we use this to show [ unipotent ] let @xmath11 be reduced and proper over a perfect field .
\a ) there is a short exact sequence @xmath26 and @xmath27 .
\b ) the group scheme @xmath10 represents the functor @xmath28 _ notation : _ for a field @xmath29 , we denote by @xmath30 its algebraic closure , and for a scheme @xmath11 over @xmath29 we let @xmath31 . unless specified otherwise , all extension and homomorphism groups are considered on the large etale site .
this paper was written while the author was visiting t.saito at the university of tokyo , whom we thank for his hospitality .
we are indebted to g.faltings for pointing out a mistake in a previous version , and the referee , whose comments helped to improve the exposition and to give more concise proofs .
[ oiu ] let @xmath32 and @xmath33 be abelian group schemes over a scheme @xmath34 , and assume that @xmath35 .
then every extension of @xmath33 by @xmath32 of sheaves of abelian groups on the big etale site of @xmath34 is representable , i.e. the group of extensions of @xmath33 by @xmath32 in the category of group schemes and in the category of sheaves agree .
_ proof . _ given an extension of sheaves @xmath36
, it suffices to show that @xmath37 is representable as a sheaf of sets . by hypothesis
, we can find a section @xmath38 with @xmath39 .
define a ( set - theoretic ) splitting @xmath40 to @xmath41 by sending a section @xmath42 of @xmath43 to @xmath44 .
but then @xmath45 is an isomorphism of sheaves .
@xmath46 + [ extvanish ] let @xmath29 be algebraically closed .
then @xmath47 vanishes for every smooth , connected , commutative affine group scheme @xmath32 over @xmath29 .
in particular , @xmath48 for every proper scheme @xmath1 . _ proof . _
the group scheme @xmath32 has a filtration consisting of schemes of the form @xmath49 and @xmath50 .
this is shown in sga3 xvii prop.4.1.1 and thm.7.2.1 for the flat topology , and follows with ( * ? ? ?
* thm.3.9 ) for the etale topology .
looking at long exact sequences , it suffices to show @xmath51 .
now @xmath52 , and @xmath53 since higher cohomology of a quasi - coherent sheaf on an affine scheme vanishes .
thus by the lemma it suffices to show that the corresponding extension groups in the category of group schemes vanish .
but there are no extensions of groups schemes between @xmath49 and @xmath49 , or @xmath49 and @xmath50 by sga 3 xvii thm . 5.1.1 . if @xmath11 is proper over @xmath29 , then by the stein factorization , we can write @xmath41 as the composition @xmath54 , where @xmath55 is the spectrum of an artin @xmath29-algebra , and @xmath56 . in particular , @xmath57 , and
we can assume that @xmath11 is finite of some degree @xmath58 over @xmath29 .
in this case , the weil - restriction @xmath59 is an open subscheme of @xmath60 by ( * ? ? ?
* prop.7.6/2 ) , hence smooth , connected and affine , and we can apply the above .
@xmath46 + _ remark . _ the hypothesis that @xmath29 is algebraically closed is necessary .
for example , if @xmath61 is the anisotropic form of @xmath49 , then @xmath62 , generated by the weil - restriction @xmath63 . _ proof .
_ ( theorem [ torus ] ) since the maps defined below are natural , we can assume that @xmath29 is algebraically closed .
we can also assume that @xmath11 is reduced , because @xmath64 , and because the map @xmath65 has unipotent kernel and cokernel ( * ? ? ?
* cor.page 9 ) .
it suffices to calculate @xmath66 , because there are no homomorphisms from @xmath49 to commutative group schemes other than tori @xcite . by yoneda
s lemma , the latter group is isomorphic to the group of homomorphisms of sheaves on the big ( etale ) site @xmath67 . from the duality of diagonal group schemes and
locally constructible sheaves , we get @xmath68 and @xmath69 , hence the spectral sequence ( * ? ? ?
* iii thm.1.22 ) @xmath70gives an isomorphism @xmath71 .
the leray spectral sequence @xmath72 gives an exact sequence @xmath73 by proposition [ extvanish ] , @xmath74 is injective .
if @xmath11 is normal , then @xmath75 by ( * ? ?
* thm.2.1 ) , hence @xmath76 is the zero - map
. we will show by induction on the dimension of @xmath11 that @xmath76 is the zero - map in general .
let @xmath77 be the normalization of @xmath11 , and consider the conductor square @xmath78 where @xmath79 is the closed subscheme ( of smaller dimension ) where @xmath80 is not an isomorphism .
it suffices to show that the lower horizontal map in the following commutative diagram is injective .
@xmath81 this will follow if @xmath82 , where @xmath83 is the cokernel of the injection @xmath84 .
but this cokernel is a quotient of a commutative , smooth , connected linear algebraic group , hence is a commutative , smooth , connected linear group itself , and we can apply proposition [ extvanish ] .
@xmath46 + _ remark .
_ the example in ( * ? ?
8.2 ) shows that the map @xmath85 is not an isomorphism for @xmath86 .
one can ask if it is an isomorphism if one replaces @xmath87 by the eh - cohomology group @xmath88 of @xcite . _ example . _ if @xmath11 is the node over an algebraically closed field , then @xmath89 , and @xmath90 . let @xmath11 be a node with non - rational tangent slopes at the singular point .
base changing to the algebraic closure , one sees that @xmath91 , with galois group acting as multiplication by @xmath92 , hence @xmath9 is an anisotropic torus . using the theorem
, we are able to recover the torsion of @xmath9 , @xmath6 and the diagonalizable part of @xmath93 in terms of etale cohomology : [ detail ] let @xmath11 be proper over a perfect field @xmath29 . then we have canonical isomorphisms @xmath94 _ proof . _
taking the colimit of the isomorphism @xmath95 of ( * ? ? ?
* prop.4.16 ) or @xcite , we obtain @xmath96 .
since there are no extensions of group schemes of @xmath97 by a smooth commutative group scheme @xmath37 , and since @xmath98 , we can apply lemma [ oiu ] to conclude that @xmath99 .
in particular , theorem [ torus ] implies that @xmath100 .
consider the commutative diagram : @xmath101 the middle column is the short exact coefficient sequence , and the left column and middle row are short exact by the above .
a diagram chase shows that @xmath80 is injective , and the right vertical map is an isomorphism .
the corollary follows because the group @xmath102 is divisible and the group @xmath103 is finite .
@xmath46 + the above result should be compared to ( * ? ? ?
* prop.6.2 ) , where we show that , for every proper scheme over an algebraically closed field , the higher chow group of zero - cycles @xmath104 is the pontrjagin dual of @xmath105 .
this implies a short exact sequence @xmath106 for @xmath107 the dual abelian variety of @xmath6 , and @xmath108 the character module of @xmath9 .
however , in this case the contribution from the torus and from the abelian variety are not compatible with the coefficient sequence @xmath109 as in corollary [ detail ] .
looking at tangent spaces , the previous corollary gives a dimension formula : let @xmath110 be a prime different from @xmath111
. then @xmath112
let @xmath113)\xrightarrow{0^ * } { \operatorname{pic}}(x)\big)$ ] . since @xmath114 induces @xmath115 on the typical part , @xmath14)_{[1]}$ ] is a subgroup of @xmath116 . in @xcite
, weibel shows that for every scheme there is a direct sum decomposition @xmath117)\cong{\operatorname{pic}}(x)\oplus n{\operatorname{pic}}(x)\oplus n{\operatorname{pic}}(x)\oplus h^1_{{\text{\rm et}}}(x,{{{\mathbb z}}}).\ ] ] _ proof . _ ( theorem [ npic ] ) .
we show first that @xmath118 .
since there are no non - trivial morphisms of schemes from @xmath119 to an abelian variety , a torus , an infinitesimal group , or a discrete group , we see that the kernel of @xmath120 agrees with the kernel of @xmath121 .
let @xmath1 and @xmath122 be the structure morphisms . then the leray spectral sequence gives a commutative diagram @xmath123 and it suffices to show that the outer vertical maps are isomorphisms .
let @xmath124 be the stein factorization of @xmath41 , such that @xmath125 and @xmath55 is the spectrum of an artinian @xmath29-algebra .
since @xmath126 is flat , @xmath127 , and @xmath128 is the stein factorization of @xmath129 .
we obtain @xmath130 and @xmath131 .
hence the terms on the left vanish because @xmath132 . to show that @xmath133 is an isomorphism ,
we can assume that @xmath55 is a local artinian @xmath29-algebra with ( perfect ) residue field @xmath134 . by ( * ? ? ?
* iii rem.3.11 ) we are reduced to showing that @xmath135 is an isomorphism , and this can be found in ( * ? ? ?
* iv ex.2.20 ) .
given an element @xmath15 of @xmath116 , the condition @xmath136)_{[1]}$ ] implies that the corresponding @xmath137 satisfies @xmath20 for all @xmath138 .
if @xmath29 has characteristic @xmath23 , then @xmath139 for some @xmath140 , and the map @xmath141 corresponds to a morphism of hopf algebras @xmath142\to k[t]$ ] .
if @xmath143 , then @xmath144 only if @xmath145 for all @xmath138 , hence @xmath146 . @xmath46 + _ proof .
_ this follows from @xmath151 , because any unipotent , connected , smooth affine group is an affine space as a scheme , hence admits a non - trivial morphism from @xmath152 which sends @xmath23 to @xmath23 if it is non - trivial .
@xmath46 + the kernel and cokernel of @xmath153 has been described in @xcite , hence we will from now assume that @xmath11 is reduced .
if @xmath154 is the semi - normalization of @xmath11 , then the map @xmath155 is an injection of sheaves on the same topological space . for @xmath154
reduced and semi - normal , @xmath156 by traverso s theorem @xcite together with ( * ? ? ?
hence the corollary implies that @xmath157 , and that @xmath158 ( for curves , this recovers ( * ? ? ?
* prop.9.2/10 ) . )
indeed , by corollary [ detail ] , the map @xmath159 induces an isomorphism on the torus and abelian variety part , because it induces an isomorphism on etale cohomology .
_ proof . _
a ) since any scheme @xmath166 over @xmath29 is flat , we have by flat base change @xmath167 , where @xmath168 is the projection . in particular , @xmath169 and it suffices to show that @xmath170 . since @xmath171 , we can assume that @xmath29 is algebraically closed and that @xmath11 is connected .
but @xmath25 is a universal homeomorphism , so that @xmath172 because @xmath11 and @xmath154 are reduced , proper , connected , and have a @xmath29-rational point .
\b ) we proceed by induction on the dimension of @xmath11 .
if @xmath173 , then @xmath165 for @xmath164 because @xmath174 is exact for finite maps , and to show that @xmath163 is a smooth , connected , unipotent group scheme , we can assume that @xmath175 is a finite field extension of @xmath29 , and that @xmath176 is a local artin algebra over @xmath55 with residue field @xmath55 .
since the weil - restriction is compatible with extensions , and since the weil - restriction of @xmath50 along a finite flat map of degree @xmath58 is @xmath177 , we can assume that @xmath178 and @xmath179 with @xmath180 a nilpotent ideal .
we can find a sequence of local artin algebras @xmath181 such that @xmath182 , @xmath183 and @xmath184 with @xmath185 . since
an extension of smooth , connected , unipotent commutative group schemes is of the same type , we can assume that @xmath179 with @xmath186 .
for a given scheme @xmath187 , @xmath188 hence the cokernel of @xmath189 is represented by the vector group @xmath180 .
if @xmath190 , we consider the conductor square . by hypothesis , @xmath191 is a bijection of schemes of lower dimension which induces isomorphism on residue fields , and @xmath79 is still reduced . from the injectivity of @xmath192 and the short exact sequence @xmath193
we conclude that the map @xmath194 is injective with cokernel equal to the cokernel of @xmath195 .
the proposition follows because @xmath196 .
@xmath46 + \b ) recall that @xmath168 , and consider the diagram @xmath198 since @xmath199 is an isomorphism as in proposition [ bef]a ) , the outer maps are isomorphisms , and it suffices to calculate @xmath200 .
let @xmath201 and @xmath202 , and consider the tautological map @xmath203 it suffices to show the following statements : \b ) let @xmath207 with @xmath208 . since @xmath209 is flat , we get an injection @xmath210 .
we claim that the inverse of @xmath209 in @xmath211 is the sheaf associated to the presheaf @xmath212 .
this can be checked on an affine covering , and then it is ( * ? ? ?
* lemma 2.2(2 ) ) .
\c ) let @xmath209 and @xmath213 be subsheaves of @xmath211 which are invertible in @xmath211 and isomorphic as abstract invertible sheaves . multiplying with the inverse of @xmath213 inside @xmath211 , it suffices to show that if @xmath209 is a subsheaf of @xmath211 , and @xmath214 an isomorphism , then @xmath215 .
but @xmath216 is a global unit of @xmath217 , and by proposition [ bef]a ) , @xmath218 .
hence @xmath219 .
@xmath46 + j.p.murre , _ on contravariant functors from the category of pre - schemes over a field into the category of abelian groups ( with an application to the picard functor ) _ , inst .
hautes etudes sci .
23 ( 1964 ) 543 . | we describe the maximal torus and maximal unipotent subgroup of the picard variety of a proper scheme over a perfect field .
@=11 @ i#1i#2i@checki .. i .. i & & @ e#1e#2e@ @ @@@ @h@xmath0 @>@ @@
ifnotempty#2@ne @ne>@ @@ne @ @@ |
the charmonium family is a great laboratory for precision tests of the quark model , because of their relative immunity from complications like relativistic effects and the large value of the strong coupling constant @xmath5 . in this talk ,
a brief review of recent experimental results on charmonium @xmath0 , @xmath1 and @xmath2 is presented .
although these states were predicted just after the discovery of @xmath6 , their properties were not very clear for a long period .
@xmath3 is the mostly recently discovered charmonium states ; recent studies uncovered its production and properties .
@xmath1 is the lowest - lying @xmath7-wave spin - singlet charmonium state and has been observed through various processes .
its `` inconsistent '' lineshapes in different production modes inspired a couple of precise measurements in the last few years .
the @xmath2 is the first radial excitation of the @xmath8 charmonium ground state .
after 30 years of searching , it was recently observed in charmonium transitions , having been observed in @xmath9 decays and @xmath10 fusion previously .
results in this talk come from besiii , cleo , babar and belle .
besiii and cleo study charmonium from @xmath4 decays and @xmath11 annihilation near @xmath12 threshold .
they provide very clean and simple environments .
babar and belle are @xmath13factories , which produce charmonium via @xmath10 fusion and @xmath9 decays .
the advantage of studying charmonium in a @xmath13factory is the relatively large statistics and reconstruction efficiency .
of the charmonium states below @xmath12 threshold , the @xmath14 is experimentally the least accessible . that is because it can not be produced directly in @xmath11 annihilation , or appear in the electric dipole transition process of a @xmath15 charmonium state .
statistics and photon detection also made it very challenging for early experiments to observe @xmath3 in charmonium transitions .
information about the spin - dependent interaction of heavy quarks can be obtained from precise measurement of the @xmath16 hyperfine mass splitting @xmath17 , where @xmath18 mev/@xmath19 @xcite is the spin - weighted centroid of the @xmath20 mass and @xmath21 is the mass of the singlet state @xmath3 .
a non - zero hyperfine splitting may give indication of non - vanishing spin - spin interactions in charmonium potential models @xcite .
the first evidence of the @xmath3 state was reported by the fermilab e760 experiment @xcite and was based on the process @xmath22 .
this result was subsequently excluded by the successor experiment e835 @xcite , which investigated the same reaction with a larger data sample .
e835 also studied the process @xmath23 , in this case finding an @xmath3 signal . soon after this the cleo collaboration observed the @xmath3 and measured its mass @xcite by studying the decay chain @xmath24 in @xmath11 collisions .
cleo subsequently presented evidence for @xmath3 decays to multi - pion final states @xcite .
since these data were collected in 2009 , besiii has put lots of effort into measuring the properties of @xmath3 . to study the decay @xmath25 , three methods have been used : * inclusive : in the inclusive mode , only the @xmath26 is detected and the @xmath3 are recognized as a peak in the @xmath26 recoil mass spectrum .
the @xmath26 momentum in @xmath27 is about 85 mev .
the inclusive yield is used to extract the absolute branching ratio of @xmath27 .
this mode has the largest background . *
e1-tagged : detecting the @xmath26 and the e1 transition @xmath28 from the @xmath29 ( 500mev ) .
the e1-tagged signal yield is proportional to the product branching ratio @xmath30 . combining with the inclusive measurement , e1-tagging also provides the absolute branching ratio of @xmath29 .
the background for this mode is smaller than that for the inclusive mode . *
exclusive : reconstructing @xmath26 , e1@xmath28 and all of the decay products of the @xmath1 in @xmath25 . here
all final - state particles are detected and energy - momentum conserving kinematic fits can be used to improve the resolution .
this method has small background and provides the best @xmath3 mass and width measurement .
the yield is proportional to @xmath31 , where the @xmath32 refer to specific final states in the @xmath1 decay .
the observation of @xmath3 from cleo used the e1-tagged and exclusive modes . using inclusive and e1-tagged modes ,
besiii first measured the absolute branching ratios @xmath33 and @xmath34 @xcite .
these results are consistent with theoretical expectations and make it possible to extract absolute @xmath3 cross sections / branching ratios @xcite@xmath35@xcite . in the same paper ,
besiii also determined the mass and width of @xmath3 to be @xmath36 mev and @xmath37 mev ( @xmath38 confidence level upper limit is 1.44mev ) , which agree with cleo s observation . with this mass value ,
the @xmath39wave hyperfine splitting is @xmath40 mev , consistent with zero .
[ hc - fitdata ] shows the @xmath3 signals and fits in the inclusive and e1-tagged @xmath26 recoil mass spectrum in @xmath4 decays @xcite .
measurement from besiii : ( a ) the @xmath26 recoil mass spectrum and fit for the @xmath41-tagged analysis of @xmath42 ; ( b ) the @xmath26 recoil mass spectrum and fit for the inclusive analysis of @xmath43 .
fits are shown as solid lines , background as dashed lines .
the insets show the background - subtracted spectra.,title="fig:",width=302 ] ( -40,180 ) * ( a ) * ( -40,90 ) * ( b)*(-230,70)(-140,-5)*@xmath44 recoil mass ( gev/@xmath19 ) * a detailed study of 16 exclusive channels is in progress at besiii @xcite . the aim is to obtain the most precise @xmath3 resonance parameters and study @xmath1 line - shape parameters in the @xmath41 transition @xmath29
. preliminary results of this study are @xmath45 and @xmath46 mev . the summed fit results are shown in fig . [
fig : fithcpi0tot ] .
cleo has confirmed besiii s inclusive measurement result @xcite . by rejecting very asymmetric
@xmath47 decays , cleo also observed @xmath3 signal in the inclusive @xmath26 spectrum in @xmath4 decays ( figure [ fig : fig4 ] ) .
@xmath48 is determined to be @xmath49 , which is consistent with the besiii measurement .
beyond @xmath27 , cleo has made an important discovery of @xmath3 production in @xmath50 at @xmath51mev using 586 pb@xmath52 of @xmath11 annihilation data .
10@xmath53 signal for @xmath3 was found in the decay @xmath54 12 decay modes .
this result demonstrats a new prolific source of @xmath3 and has inspired the belle collaboration to search for @xmath55 in @xmath11 annihilations at = 10.685 gev using the same technique .
cleo also finds evidence for @xmath56 at @xmath57mev at the @xmath58 level , and sees hints of a rise in the @xmath59 cross section at @xmath60mev @xcite .
the @xmath61 cross sections measured by cleo at different center - of - mass energies are summarized in fig .
[ fig : scan ] .
@xmath1 is the lowest - lying @xmath7-wave spin - singlet charmonium state .
although it has been known for about thirty years @xcite , its resonance parameters are still interesting . for a long period ,
the measurements of the @xmath1 width from @xmath13factories and from charmonium transitions were inconsistent . in pdg10
@xcite , confidence level of the global fit is only 0.0018 for the @xmath1 mass and only 0.0001 for the @xmath1 width .
these discrepancies can be attributed to poor statistics and inadequate consideration of interference between @xmath1 decays and non - resonant backgrounds .
the experimental confusion introduced difficulties in the determination of the charmonium @xmath62 mass hyperfine @xmath63 . with old @xmath1 parameters in pdg10 , @xmath62 hyperfine mass
splitting is @xmath64 mev , away from theoretical predictions @xcite .
recent studies by belle , babar , cleo , and besiii @xcite , with large data samples and careful consideration of interference , obtained similar @xmath1 width and mass results in two - photon - fusion production and @xmath4 decays . in 2009 , cleo observed a distortion in the @xmath1 line shape in @xmath65 .
cleo concluded that the distortion is caused by photon - energy dependence of the magnetic dipole transition rate ( hindered - m1 transition ) @xcite .
this observation inspired besiii s @xmath1 line shape study via @xmath65 with a 106m - event @xmath4 sample @xcite . in besiii
s @xmath1 analysis , @xmath1 is reconstructed with six decay modes : @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , and @xmath71
. a simultaneous fit to these channels is performed .
the @xmath1 breit - wigner is weighted by an @xmath72 factor to account for the energy dependence of the hindered-@xmath73 transition .
interference with background from non - resonant @xmath4 decays is also considered .
the new besiii mass and width values are @xmath74 and @xmath75 .
they agree well with results from @xmath13factories .
using only the new besiii @xmath1 mass value , the @xmath76 hyperfine mass splitting is @xmath77 mev , which agrees better with theory calculations .
figure [ fig : metac ] shows the data and fit for each channel in this analysis .
the result from besiii provides strong evidence that previous inconsistent @xmath1 parameters in radiative charmonium decays and two - photon collisions@xmath13meson decays are caused by the hindered-@xmath73 transition and non - resonant interence .
there are also new measurements of @xmath1 from @xmath13factories . with a data sample of 535 million @xmath78-meson pairs , belle measured the @xmath1 lineshape via @xmath79 @xcite .
compared to the two - photon process , the advantages of studying @xmath1 in @xmath9 decays are the relatively large reconstruction efficiency , small background , and the fixed quantum numbers of the initial state . in belle s analysis , a 2d - fit to the @xmath80 and @xmath81 distributions
is performed to obtain the interference contributions , where @xmath81 is defined in fig .
[ pic:7 ] . to reduce the uncertainty from the interference , p- and
d - waves are separated from the s - wave in the non - resonant background .
[ pic : fit1 ] shows projections of the fit .
@xmath1 parameters measured in this analysis are @xmath82 mev/@xmath19 and @xmath83 mev/@xmath19 , which are consistent with recent results from besiii @xcite and babar @xcite .
.,height=4 ] [ cols="^,^,^ " , ] and @xmath2 decays to the @xmath84 and the @xmath85 in two - photon interactions from babar : fit to ( a ) the @xmath84 and ( c ) the @xmath85 mass spectrum .
the solid curves represent the total fit functions and the dashed curves show the combinatorial background contributions .
the background - subtracted distributions are shown in ( b ) and ( d ) , where the solid curves indicate the signal components . , title="fig : " ] and @xmath2 decays to the @xmath84 and the @xmath85 in two - photon interactions from babar : fit to ( a ) the @xmath84 and ( c ) the @xmath85 mass spectrum .
the solid curves represent the total fit functions and the dashed curves show the combinatorial background contributions .
the background - subtracted distributions are shown in ( b ) and ( d ) , where the solid curves indicate the signal components .
, title="fig : " ] babar studied @xmath1 and @xmath2 in the two - photon processes @xmath86 and @xmath87 using a data sample of 519.2 fb@xmath52 near the @xmath88 ( @xmath89 ) resonances .
@xmath90 was found with a significance of @xmath91 @xcite .
babar also obtained the @xmath2 mass and width by performing to a fit to the @xmath92 spectrum ( fig .
[ fig : fit ] ) , and obtained the values @xmath93
mev/@xmath19 and @xmath94 mev/@xmath19 .
on the other hand , the search for @xmath2 through a radiative transition from the @xmath4 is very hard and has been stuck for many years .
the difficulty comes from the detection of the low - energy radiative photon in the @xmath95 .
the branching ratio and mechanism of this process has been predicted by many papers , but the absence of experimental result had made it impossible to discriminate among them @xcite@xmath35@xcite . for a long period , this transition has been looked for by crystal ball @xcite , bes @xcite , cleo @xcite and besiii @xcite , but none of them could provide a convincing observation .
most recently , besiii made the first observation of this process using the 106 m @xmath4 sample .
analyses of @xmath96 with @xmath97 and @xmath98 gave a significance greater than @xmath99 .
in addition to the excellent low energy photon detection , smart use of the kinematic fitting plays a key role in this observation .
[ pic_fit_etacp ] shows data and fits .
numerical results are @xmath100 mev/@xmath19 ) , @xmath101 mev and @xmath102 .
the branching ratio of @xmath95 is determined to be @xmath103 .
these results are consistent with results from @xmath13factories .
in summary , experimental studies of @xmath3 , @xmath1 and @xmath2 have made major progress and still face nontrivial challenges : * @xmath3 + the key branching ratios @xmath27 and @xmath104 have been nailed down , so the absolute @xmath3 cross sections / branching ratios are available .
a new prolific production mode of @xmath3 has been found : @xmath50 . because the @xmath105 is about @xmath106
, the remaining decays of @xmath3 should be large enough to be observed .
further measurement of these unclear decays will be helpful to understand the property of @xmath3 and the transition mechanism between @xmath3 and other charmonium . *
@xmath1 + the mass and width are more consistent in @xmath4 decays , @xmath9 decays and @xmath10 production than previously , and the charmonium @xmath62 mass hyperfine splitting from experiments agrees better with theory . the @xmath8 lineshape in @xmath3 is not as distorted as in charmonium/@xmath9 decays and @xmath10 fusion , because of the small non - resonant interfering background .
ultimately , with a large @xmath4 sample , this channel will be best suited to determine @xmath8 resonance parameters .
* @xmath2 + finally , @xmath2 was observed in charmonium transitions after thirty years of searching .
finding @xmath2 has proved hard enough , but understanding its decay properties and measuring its lineshape are proving more difficult . due to the lower production rate , @xmath2 is hard to reconstruct and is affected by interference more than @xmath1 .
further measurement of @xmath2 will require large statistics , a convincing theoretical model and a sophisticated fitting method .
j. l. rosner _
[ cleo collaboration ] , phys .
lett . * 95 * , 102003 ( 2005 ) [ hep - ex/0505073 ] .
s. dobbs _ et al . _
[ cleo collaboration ] , phys .
* 101 * , 182003 ( 2008 ) .
g. s. adams _ et al . _
[ cleo collaboration ] , phys . rev .
d * 80 * , 051106 ( 2009 ) .
m. ablikim _ et al . _
[ besiii collaboration ] , phys .
lett . * 104 * , 132002 ( 2010 ) .
s. l. olsen [ besiii collaboration ] , arxiv:1203.4297 [ nucl - ex ] .
kuang , s. f. tuan , and t. m. yan , phys . rev .
d * 37 * , 1210 ( 1988 ) .
p. ko , phys .
d * 52 * , 1710 ( 1995 ) .
kuang , phys .
d * 65 * , 094024 ( 2002 ) .
s. godfrey and j. rosner , phys .
d * 66 * , 014012 ( 2002 ) .
j. y. ge _ et al .
_ [ cleo collaboration ] , phys . rev .
d * 84 * , 032008 ( 2011 ) .
t. k. pedlar _ et al .
_ [ cleo collaboration ] , phys . rev . lett . *
107 * , 041803 ( 2011 ) .
t. himel , g. trilling , g. s. abrams , m. s. alam , c. a. blocker , a. blondel , a. boyarski and m. breidenbach _ et al .
_ , phys .
* 45 * , 1146 ( 1980 ) .
a. vinokurova _ et al .
_ [ belle collaboration ] , phys .
b * 706 * , 139 ( 2011 ) .
del amo sanchez _ et al .
_ [ babar collaboration ] , phys .
d * 84 * , 012004 ( 2011 ) .
r. e. mitchell _ et al .
_ [ cleo collaboration ] , phys . rev .
lett . * 102 * , 011801 ( 2009 ) [ erratum - ibid . *
106 * , 159903 ( 2011 ) ] .
m. ablikim _ et al . _
[ besiii collaboration ] , phys .
lett . * 108 * , 222002 ( 2012 ) . | recent experimental results on charmonium @xmath0 , @xmath1 and @xmath2 from belle , babar , cleo and besiii are reviewed .
@xmath3 production and properties , the @xmath1 lineshape and the observation of @xmath2 in @xmath4 decays are discussed . |
like the lhc , the rhic collider is a dual function machine providing collisions of protons as well as heavy ions . unlike the lhc
, the experiments at rhic have been designed completely with the goals of heavy ion physics in mind and then subsequently used also for measurements with ( polarized ) proton colliding beams . in many cases
the choice of parameters in detector design have had interesting consequences for the types of measurement that can be carried out with pp collisions but nowhere is this so clear as in the case of forward instrumentation .
the primary goals of forward instrumentation at rhic are * to measure absolute luminosity * to provide a truly minimum bias trigger * to characterize the initial conditions of individual events(reaction plane orientation , impact parameter ) outside the mid - rapidity region the interaction region geometry , depicted in fig . 1 , provides a location at z=@xmath0 m of maximum momentum dispersion where spectator fragments " from the colliding nuclei are resolved primarily according to their charge - to - mass ratio ( z / a ) . an intermediate impact parameter event is shown in fig . 2 .
rhic experiments were designed to characterize the reaction plane and impact parameter of individual events . the number of forward going spectator nucleons " leaving the collision is anti - correlated with the number of participants : @xmath1 in neutron rich ion species stored at rhic ( au ) and lhc ( pb ) most of the information is carried by neutrons which are detected in the rhic zero degree calorimeters ( zdc s ) @xcite that were installed in each experiment .
protons are also measured in the phenix experiment using a large area hadron calorimeter recycled from ags experiment e865 . the centrality measurement by spectators is partially diluted by a significant component of recombination / coalescence of the spectator nucleons ( ie @xmath2 d ) which are neither detected in the zdc or the fcal .
even for central collisions @xmath3 of the protons coalesce to higher mass fragments @xcite .
so in phenix the participant number is sampled by the multiplicity in the @xmath4 region and combined with zdc energy to form the centrality algorithm .
early studies of rhic collisions @xcite showed that spectator neutrons are detected in the zdc with @xmath5 in nuclear interactions at all centralities .
even in ultraperipheral collisions ( upc ) where @xmath6 , beam energy neutrons are evaporated ( and detected in the zdc s ) due to the intense electromagnetic interactions between high - z nuclei @xcite .
the zdc is therefore the ideal minimum biased event trigger for nuclear interaction physics . for upc , the zdc trigger ( tagged " upc events - see below ) is used to form rapidity gap " triggers and is the only practical way to access this physics . because the dominant upc process -mutual coulomb dissociation - is calculable with an accuracy of @xmath7 @xcite and cleanly measured @xcite , it is used as the basis for all absolute luminosity measurements in rhic .
the reaction plane orientation is best measured in the zdc using a position sensitive scintillator array at a depth of @xmath8 in the calorimeter because of the large directed flow , @xmath9 , among nucleons at this rapidity @xcite .
since the orientation is measured independently in both zdc arms(left and right ) the resolution can be directly derived and we find @xmath10
ultraperipheral collisions become an important aspect of heavy ion physics at the large c.m .
energies available at colliders . at the lhc , with pb+pb collisions at @xmath11 tev
, processes can be studied that are equivalent to 100 tev quasi - real photon collisions with a stationary nuclear target @xcite opening up a wide range of opportunities @xcite . the principal method for analyzing such collisions , the method of equivalent quanta , originated with fermi @xcite who showed that interactions of electrons and @xmath12particles with atoms can be reduced to a convolution of an equivalent photon spectra and the known @xmath13a interaction cross sections .
the method was extended to the extreme relativistic case by weizsacker and williams and has found many applications in high energy and nuclear physics . at ion colliders
the characteristic strength of interaction @xmath14 is large so that multiple photon exchanges lead to the phenomenon of tagging .
additional photon exchanges can cause photo - nuclear breakup of one nucleus which emits a soft neutron @xcite detected in the zdc .
originally the tagging requirement was thought of as a penalty factor required to access this physics with a clean experimental trigger .
however , as shown in ref .
@xcite the tagging fraction is correlated with impact parameter and strongly biases the tagged photon distribution toward higher equivalent energies .
we now calculate tagged and untagged photon spectra separately , using this nomenclature below . during the @xmath15 gev initial au+au run , star recorded data with a two - prong track requirement ( consistent with diffractive @xmath16 decays ) as well as a minibias zdc trigger @xcite .
the two data samples were examined for diffractive @xmath17 production as well as low mass @xmath18candidates .
the results are summarized in table . 1 , reproduced from ref .
@xcite where results are also compared with a calculation of the @xmath17 production cross sections for both the tagged and untagged cases @xcite .
the cross sections ( subscripts 0n and xn denote 0 and 1-or more neutron tags , respectively ) are in reasonable agreement with the calculation .
note that the experimental errors are smallest in the tagged sample .
.@xmath17 photoproduction results reproduced from ref .
@xcite . [ cols="<,^,^",options="header " , ] during the most recent ( 2004 ) au+au run , phenix implemented a diffractive trigger sensitive to high mass @xmath19 pairs using a zdc tag and vetoing on events with activity in the @xmath4 ( = bbc ) region .
the phenix experiment has excellent identification capabilities for high momentum electrons due to the highly segmented ( @xmath20 ) electromagnetic ( em ) calorimeter and ring imaging cerenkov counters .
the diffractive @xmath19 trigger also required at least one cluster of calorimeter energy with greater than 0.8 gev .
the full trigger is @xmath21 this loose rapidity gap trigger ( _ ie _ a leading neutron and no activity in at least one bbc arm ) is very effective in heavy ion collisions and had a rate of @xmath22 of the min bias rate .
a total of 8.5 m upc triggers were recorded .
electron candidates were reconstructed requiring consistent em calorimeter energy and magnetic spectrometer momentum as well as a cerenkov match . the resultant @xmath19 invariant mass spectrum @xcite
is shown in fig .
all candidates have a net @xmath23 of @xmath24 mev , consistent with diffractive photo - production and the expected au nuclear form factor .
this should be compared to the @xmath25 gev / c we find in non - diffractive @xmath26+x data .
it should be noted that this spectrum is highly preliminary since , at the time of this initial pass through the data , there has been very little prior experience in phenix reconstructing events with such low multiplicities ( _ ie _ where the interaction vertex is not well constrained ) .
our calculation of the expected j/@xmath27 and @xmath19 continuum yield is higher than observed by about a factor of 2 .
nevertheless , even at this very early stage in the analysis , it is clear that a diffractive signal of high mass pairs has been seen and the capability for small cross section diffractive physics has been demonstrated .
mass spectrum.,width=283 ]
immediately following the first high luminosity au+au run rhic operated as a d+au collider ( at @xmath30 gev ) .
the primary purpose of this run was to directly measure the contribution of initial state nuclear effects on inclusive particle spectra observed in d - au@xmath31+x for example . in order to measure inclusive particle cross sections , it is critical to calibrate the d - au luminosity . also , since particle spectra are compared for an equivalent number of binary collisions , derived from a glauber model calculation , it is useful to validate this calculation by confirming the corresponding minimum bias cross section . during the d+au run star and
phenix ran with different min.bias triggers : @xmath32 which , as pointed out in ref .
@xcite , have different acceptances for @xmath33 collisions with a rapidity gap and hence different cross sections .
in phenix we have measured directly the inelastic d+au cross section by normalizing to the theoretical calculation @xcite for deuteron photodissociation of @xmath34 b. phenix recorded 12 m events with three different triggers enabled : the 2 inelastic triggers of eqn . 4 in or " with a trigger requiring @xmath35 gev .
this last trigger is @xmath36 efficient for deuteron breakup events since , with a reasonable deuteron input wavefunction , all neutrons fall within the zdc aperture .
dissociation events have a particularly clean signature in the phenix forward detectors since , in the d beam direction , there is an easily resolved 100 gev energy peak simultaneously in both the zdc and the proton calorimeter .
also there is no activity in the central detector or the au beam zdc . ref .
@xcite predicts a factor of 0.83 suppression of the min.bias(phenix ) cross section relative to the naive glauber value of @xmath37 b. instead we measure , applying a small correction for bbc efficiency : @xmath38 which differs by @xmath39 st.dev . from ref .
the error on this preliminary result will probably reach @xmath40 after further study .
large @xmath42 inclusive cross sections are relevant for analysis of super - gzk cosmic ray candidates and also for lhc luminometry , where this process will be used for beam diagnostics . at rhic
it has a special role in polarimetry since there is a large ( @xmath43 ) left - right asymmetry of the neutron angular distribution relative to the up - down polarization of the incident proton .
also a large ( @xmath44 ) @xmath45+x signal is observed which is not reproduced by most event generators ( _ ie _ dpmjetiii ) .
this signal is the basis of all rhic @xmath46 luminosity monitoring .
further study of these forward physics processes is now under way .
the atlas experiment is now constructing zdc s to be installed in the tan absorber at @xmath47 m from the ip1 interaction region .
full geant simulation of this detector with @xmath11 tev pb+pb collisions has demonstrated a similar energy resolution to that observed at rhic
. the atlas zdc s will be the primary trigger for diffractive physics and will be combined with low-@xmath23 jet or secondary vertex triggers . a likely area of research is the study of photo - produced dijets , @xmath13jet pairs and heavy quarks ( @xmath48 ) and also the corresponding diffractive photo - production processes ( which have a rapidity gap in _ both _ beam directions ) .
we have carried out @xcite a calculation of rates for the above processes and the jet - jet case is presented in fig .
4 for a 1 month run at nominal lhc pb - pb luminosities .
even with a jet threshold of @xmath49 gev the total event rate exceeds @xmath50 .
we are now studying the atlas capability for accessing low @xmath23 jets in events with the very low underlying multiplicities expected for these diffractive processes .
it is likely that soft heavy quark production can be tagged through a secondary vertex at lower @xmath23 .
this would probe the lowest @xmath51 parton distributions(abscissa in fig .
4)- possibly down to @xmath52 .
we also indicate in fig .
4 the equivalent incident photon energy and the average rapidity(@xmath53 ) corresponding to each bin . for the most part , the produced jets are very central and it should be possible to require a large rapidity gap along the direction of the beam emitting the photon .
the atlas forward calorimeters are unique in providing sensitivity up to @xmath54 for @xmath55 gev and this makes it the most promising experiment to carry out rapidity gap physics with heavy ions .
i am grateful to the organizers of diffraction 2004 " for a very stimulating and pleasant meeting .
9 a. j. baltz , c. chasman and s. white , nucl .
instr . and meth .
a417 , 1 ( 1998 ) . c. adler et al .
, nucl . instr . and meth .
a470 , 488 ( 2001 ) .
m. chiu et al .
( 2002 ) 012302 , nucl - ex/0109018 .
h. appelshauser et al.[na49 collaboration ] , eur .
phys . j. a2 , 383 ( 1998 ) .
r. snellings et al .
lett . * 84 * 2803 ( 2000 ) .
s. white applications of the equivalent photon approximation to heavy ion collisions " in electromagnetic probes of fundamental physics " w. marciano and s. white , eds .
world scinetific ( 2003 ) .
g. bauer et al . hot topics in ultraperipheral heavy ion collisions " , ibid .
e. fermi ( translation by m. gallinaro and s. white ) on the theory of interactions between atoms and electrically charged particles " , ibid .
g. baur et al .
nucl.phys.a729(2003 ) 787 and nucl - th/0307031 .
c. adler et al . phys.rev.lett . 89 ( 2002 ) 272302 and nucl - ex/0206004 .
a. j. baltz , s. r. klein , and j. nystrand , phys .
lett . * 89 * 012301 ( 2002 ) . for details
see d. silvermyr in www.phenix.bnl.gov/@xmath56silvermy/dnp04.pdf .
s. klein and r. vogt , phys . rev .
c 68 ( 2003 ) 017902 . and nucl - ex/0303013 .
b. kopeliovich , phys .
c 68 ( 2003 ) 044906 and nucl - th/0306044 .
m. strikman , r. vogt and s. white manuscript in preparation . to be published in cern yellow report on ultraperipheral collision physics and r. vogt , hep - ph/0407298 . | the heavy ion physics approach to global event characterization has led us to instrument the forward region in the phenix experiment at rhic . in heavy ion collisions
this coverage yields a measurement of the spectator " energy and its distribution about the beam direction .
this energy flow is the basis of event - by - event determination of the centrality and reaction plane which are key to analyzing particle production in heavy ion collisions .
these same tools have also enabled a unique set of measurements on inelastic diffraction with proton , deuteron and gold ion beams in the phenix experiment .
we present first new results on this topic and discuss briefly the opportunity for diffractive physics with heavy ion beams at the lhc . |
we study the dynamics of an abstract dynamical system modeling a network composed with @xmath1 neurons , where @xmath2 is arbitrarily large , that are reciprocally coupled by inhibitory synapsis .
each neuron is modeled as a pacemaker of the type integrate and fire @xcite , including the case of relaxation oscillators .
the internal variable @xmath3 describing each neuron s potential , for @xmath4 evolves increasingly on time @xmath5 during the interspike intervals : @xmath6 has positive first derivative @xmath7 and negative second derivative @xmath8 being the solution of a deterministic autonomous differential equation of a wide general type .
when the potential @xmath9 reaches a given threshold value , the neuron @xmath10 produces an spike , its potential @xmath9 is reseted to zero . it
is suppose that all the neurons are inhibitory , there are no delays and the network is totally connected .
when the neuron @xmath10 spikes , not only its potential @xmath9 changes , being reseted to zero , but also , through the synaptical connections , an action potential makes the other @xmath11 neurons @xmath12 , suddenly change their respective potentials @xmath13 with a jump of amplitude @xmath14 .
the instants of spiking are defined by the evolution of the system itself , and not predetermined by regular intervals of observation of the system .
we analyze the state of the system immediately after each spike , in the sequence of instants of spiking of the network .
we prove that the state of the system after each spike is a function @xmath15 of the state after the prior spike .
this function @xmath15 is the so called poincar map .
the study of the dynamics by iteration of the poincar map is not an artificial discretization of the real time dynamics . on the contrary ,
the dynamics of @xmath15 and its properties ( for instance periodicity , chaotic attractors ) are equivalent to those of the system evolving in real time . in @xcite the technique of the first return poincar map to a section transversal to the flux
was first applied to study neuron networks , in that case , an homogeneous network of excitatory coupled pacemakers neurons . in @xcite the same technique is applied to networks of inhibitory cells , analyzing the real time dynamics via a discrete poincar map @xmath15 . this map is locally contractive and piecewise continuous in a compact set of @xmath16 .
we include the proof of these properties in the section 2 of this paper . in section [ seccionperiodico ] , we prove theorem [ teorema1 ] , which is the main abstract mathematical result : _ locally contractive piecewise continuous maps with the separation property generically have only periodic asymptotic behavior , with up to a finite number of limit cycles that are persistent under small perturbations of the map . _
generic systems have a topological meaning in this paper : they include an open and dense family of systems .
_ _ as a consequence we obtain the following applied result : _ generic neuron networks composed by @xmath1 inhibitory cells exhibit only periodic behavior with a finite number of limit cycles that are persistent under small perturbations of the set of parameter values . _ _ _ this is a result generalizing the conclusions obtained for two neurons networks in @xcite and @xcite .
on the other hand non generic dynamics are structurable unstable : they are destroyed if the system is perturbed , even if the perturbation is arbitrarily small .
we refer to those as bifurcating systems .
the results of this paper are proved in an abstract and theoretical context , using the classical qualitative mathematical tools of the topological dynamical systems theory .
the systems have discontinuities and evolve in finite but large dimension @xmath2 .
it is still mostly unknown , the dynamics of discontinuous and large dimensional systems .
that is why in sections 3 and 4 we include an abstract theory of topological dynamical systems with discontinuities .
the piecewise continuity and the local contractiveness help us to obtain the thesis of persistent periodicity in the lemma [ lema ] of this paper .
on the other hand , the separation property will play a fundamental role to obtain the thesis of density , and genericity of the periodic behavior in the theorem [ teorema1 ] . even being our mathematical analysis theoretically abstract , we observe that the conclusions about the dynamics of our model of neuron networks , fit with those obtained by experiments in computer simulations with mutually coupled identical neurons in networks of up to @xmath17 cells , as reported in the following papers : in @xcite it was observed the transition among different periodic activity , indicating that the simulation data perfectly fit to experimental and clinical observations . in the computer
simulated experiments the alterations of the discharge patterns when passing from one periodic cycle to another , arise from changes of the network parameters , changes in the connectivity between cells , and also of external modulation . in @xcite
the computer simulated experiment shows the dynamics of the network of a large number of coupled neurons .
it was observed to be significantly different from the original dynamics of the individual cells : the system can be driven through different synchronization states .
our thesis of theorem [ teorema1 ] is only applicable to deterministic systems .
nevertheless their conclusions also qualitatively fit with computer simulations of neural systems with randomness @xcite , @xcite , which also show the generation of detectable preferred firing sequences .
we include the detailed proof of the mathematical translation from a physical model of @xmath2 inhibitory pacemaker neurons network to the dynamics of iterations of a piecewise continuous contractive map @xmath18 , locally contractive and with the separation property , as first posed in @xcite .
the model is applicable for any finite number @xmath1 of neurons in the network .
the phase space of the system is the compact cube @xmath19^n \subset \mathbb{r}^n$ ] .
a point in the phase space is @xmath20 , describing the potential @xmath9 of each of the neurons @xmath21 .
we assume that the the phase space is normalized : the threshold level of each of the neurons potentials is 1 , so 1 is the maximum of @xmath9 .
also the minimum @xmath9 is normalized to @xmath22 , and that the reset value of @xmath23 , after a spike of the neuron @xmath10 , is @xmath24 [ definicionmodeloinicial ] _ * the physical model . *
the point @xmath25 in the phase space @xmath26 evolves on time @xmath5 , during the interspike intervals of time , according to an autonomous differential equation and changes without delay in a discontinuous fashion in the exact spiking instants , according to a reseting - synaptical rule .
the two regimes , during the interspike interval , and in the spiking instants respectively , are precisely defined according to the following assumptions : _ * [ definicionmodeloinicial].1 * * inter - spike regime assumptions .
* @xmath6 is the solution of a differential equation @xmath27 \mapsto \mathbb{r } , \ ; \ ; \gamma _ i \in { \cal c}^1 , \ ; \ ; \gamma_i ( v_i ) > 0 , \ ; \gamma _ i ' ( v_i ) < 0 \ ; \ ; \forall \ , v_i \in [ -1,1].\ ] ] where @xmath28 denotes the space of real functions in @xmath29 $ ] , continuous and derivable with continuous derivative in @xmath29 $ ] .
the assumption @xmath30 reflects that each neuron potential in the inter - spike interval is strictly increasing while it does not receive interactions from the other neurons of the network .
this comes from the hypothesis that each isolated neuron @xmath10 is of pacemaker type , i.e. from any initial state @xmath31 , the potential spontaneously reaches the threshold level @xmath32 for some time @xmath33 , if none inhibitory synapsis is received in the time interval @xmath34 $ ] .
the assumption @xmath35 , which we call the _ dissipative hypothesis _ reflects that the cynetic energy @xmath36 is decreasing on time while the potential freely evolves during the interspike intervals .
in fact : @xmath37 _ _ the most used example of this type of inter - spike evolution is the relaxation oscillator model of a pacemaker neuron , for which @xmath38 where @xmath39 are constants . for this type of cells the differential equation ( [ ecuaciondiferencial ] )
is linear , and its solution can be explicitly written : @xmath40 exp ( - \alpha _ i t).\ ] ] * [ definicionmodeloinicial].2 * * consequences of the inter - spike regime assumptions . *
we define the flux @xmath41 as the solution with initial state @xmath20 of the differential equations system given by ( [ ecuaciondiferencial ] ) . precisely : @xmath42 as @xmath43 we can apply the general theory of differential equations to deduce the following results , as a consequence of the assumptions in ( [ ecuaciondiferencial ] ) : * two different orbits by the flux do not intersect . *
if @xmath0 and @xmath44 are two @xmath45-dimensional topological and connected sub - manifolds of @xmath46 transversal to the vector field @xmath47 , then the flux transforms homeomorphically any set of initial states in @xmath0 onto its image set of final states in @xmath44 . * for each constant time @xmath5 it holds the louville formula : @xmath48 * [ definicionmodeloinicial].3 * * spiking instants computations .
* for each initial state @xmath49 the first spiking instant @xmath50 in the network is defined as the first positive time such that at least one of the neurons of the network reaches the threshold level 1 .
this means that @xmath51 @xmath52 is the set of neurons that reach the threshold level simultaneously at the instant @xmath50 .
it is standard to prove that for an open and dense set of initial states there is a single neuron @xmath10 reaching the threshold level first , i.e. @xmath53 . * [ definicionmodeloinicial].4 * * spiking - synaptical assumptions . * in the spiking instant @xmath54 the reseting and inhibitory synaptical interaction without delay produces an instantaneous discontinuity @xmath55 in the state of the system , according to the following formulae : * if @xmath56 then @xmath57 and : @xmath58 + where @xmath59 is constant , depending only on @xmath60 , and gives the instantaneous discontinuity jump in the potential of neuron @xmath12 produced through the inhibitory synaptical connection from neuron @xmath10 to neuron @xmath61 * if @xmath62 then @xmath63 is multiply defined , having @xmath64 possible vectorial values @xmath65 , where @xmath66 is defined according to formulae ( [ ecuacionsinapsis1 ] ) and ( [ ecuacionsinapsis2 ] ) .
we also assume that the network whose nodes are the cells and whose sides are the synaptical inhibitory interactions @xmath67 , is a complete bidirectionally connected graph .
precisely : @xmath68 * [ definicionmodeloinicial].5 * * relative large dissipative assumption .
* we assume the following relations between the functional parameters @xmath69 in the differential equations ( [ ecuaciondiferencial ] ) governing the dissipative interspike regime , and the real parameters @xmath70 in the formula ( [ ecuacionsinapsis2 ] ) governing the spiking - synaptical regime . @xmath71 } |\gamma'_i(v_i)|}{4 } \label{ecuacioneslargedissipativeness2 } \\
\frac{\max_{i \neq j}{h_{ij}}}{\min_{i \neq j}{h_{ij } } } -1 & < & \frac { \min _ { i } \min _ { v_i \in [ 1/4,3/4 ] } |\gamma'_i(v_i)|}{4 \ , \max _ { i } \ ; \gamma_i ( 3/4 ) } \label{ecuacioneslargedissipativeness3 } \end{aligned}\ ] ] condition ( [ ecuacioneslargedissipativeness1 ] ) assumes that the discontinuity synaptical jumps @xmath67 are not relatively as large as the widest range @xmath72 $ ] of the potential of the cells when they act as oscillators between the reset value 0 and the threshold level 1 , free of synpatical interactions .
the hypothesis ( [ ecuacioneslargedissipativeness2 ] ) and ( [ ecuacioneslargedissipativeness3 ] ) verify for instance for homogeneous networks in which all the functions @xmath73 and all the synaptic interactions @xmath67 are constant independent of the neurons @xmath74 .
but as they are open conditions , they also verify if the network is not homogeneous but the neurons and the synaptical jumps are not very different .
finally they also verify for networks that are very heterogeneous , but the dissipative parameter of the system @xmath75 } |\gamma'_i(v_i)|$ ] is large enough .
the assumptions above can be also possed for some number @xmath76 instead of @xmath77 in inequality ( [ ecuacioneslargedissipativeness1 ] ) , the number @xmath78 instead of @xmath79 in the values of @xmath9 , and @xmath80 instead of the denominator @xmath81 , of inequalities ( [ ecuacioneslargedissipativeness2 ] ) and ( [ ecuacioneslargedissipativeness3 ] ) . nevertheless , and without loss of generality , in the computations of this work we will take the assumptions above with @xmath82 to fix the numerical bounds .
[ comments ] * comments about the physical model . * the hypothesis ( [ ecuacioneslargedissipativeness1 ] ) , ( [ ecuacioneslargedissipativeness2 ] ) and ( [ ecuacioneslargedissipativeness3 ] ) will allow to prove the so called separation property in theorem [ teoremasparationproperty ] in this paper .
this property is essential to prove that the family of all the systems which exhibit a limit set formed only by a finite number of limit cycles is dense , which leads to the topological genericity of such systems .
we observe that the assumptions in ( [ ecuaciondiferencial ] ) , ( [ ecuacionsinapsis1 ] ) and ( [ ecuacionsinapsis2 ] ) are more general that what they a priori seem .
in fact , if instead of the variables @xmath9 which describe the electric potentials of each of the neurons , we used other equivalent variables , the vector field @xmath83 of the differential equation ( [ ecuaciondiferencial ] ) , and the synaptical vectorial interaction @xmath63 given by ( [ ecuacionsinapsis1 ] ) and ( [ ecuacionsinapsis2 ] ) , would have other coordinate expressions .
for instance , each isolated cell @xmath10 acts as an oscilator , whose potential @xmath9 varies in the interval @xmath72 $ ] .
we can diffeomorphically change the variable @xmath9 to a new one @xmath84 $ ] , called the _ phase _ of the oscilator , which by definition , evolves _ linearly _ with the time @xmath5 , during a time constant @xmath85 . in the new variables
the differential equation governing the phase state @xmath86 will be @xmath87 and the flux will be linear in @xmath26 . _ _ _ _ in @xcite it is developed the model in such phase variables @xmath86 for which the flux is linear , and it is defined the synaptical inhibitory interaction jumps @xmath88 in the phase state @xmath89 , when the phase @xmath86 reaches the threshold level 1 . to be equivalent to the constant jumps @xmath90 in the old variables @xmath13 , it is showed in @xcite that the interaction jumps @xmath88 in the new phase variables @xmath89 , must be functions @xmath91 , strictly increasing with @xmath92 and such that @xmath93 is also strictly increasing . in a widest model
the functions @xmath94 are continuous but not necessarily differentiable . in resume ,
up to a change of variables , the model assumed in this paper in hypothesis ( [ ecuaciondiferencial ] ) , ( [ ecuacionsinapsis1 ] ) and ( [ ecuacionsinapsis2 ] ) , includes for instance the model in @xcite in which the flux is linear during the interspike interval regime , and the synaptic jumps in the spiking instants adequately depend of the phase of the postsynaptic neuron .
[ definicionmodelomatematico ] * the mathematical model . * _ in this subsection we will define a poincar section @xmath95 of the dynamical system modeling physically the network of @xmath2 inhibitory neurons defined in [ definicionmodeloinicial ] .
we then shall define the first return poincar map @xmath96 .
we will prove that this map is piecewise continuous , locally contractive and has the separation property .
these properties justify the definition [ definicionmodeloabstracto ] , at the end of this section , in which we will model and analyze this kind of inhibitory neuron networks through the abstract mathematical discrete dynamical system defined by the iterates of its poincar map @xmath15 .
_ * [ definicionmodelomatematico].1 . * * the poincar section @xmath0 .
* let @xmath97^n$ ] be the compact @xmath45-dimensional set defined as follows : @xmath98 the topology in @xmath0 is defined in each @xmath99 as the induced by its inclusion in the @xmath45 dimensional subspace @xmath100 of @xmath46 .
each @xmath99 is transversal to the flux defined in [ definicionmodeloinicial].2 solution of the system of differential equations ( [ ecuaciondiferencial ] ) , because the vector field @xmath101 in the second term of this differential equations has all its components strictly positive .
after each spike , the state of the system is in @xmath0 , due to the reset rule in equality ( [ ecuacionsinapsis1 ] ) .
so the system returns infinitely many times to @xmath0 from any initial state @xmath49 . *
the geometric illustration of the three neurons system . * in figure [ fig1 ]
the cube @xmath26 is represented for @xmath102 dimensions .
the three coordinate axis @xmath103 are the three edges of the cube that are hidden in dotted lines , at the rear part of the cube , intersecting pairwise orthogonally in the unique hidden vertix @xmath104 of the cube .
the axis of @xmath105 goes from @xmath104 to the front , the axis of @xmath106 to upwards , and the axis of @xmath107 to the right .
the poincar section @xmath0 is formed by the three faces of the cube that correspond to at least one of the potentials @xmath103 equal to zero .
@xmath0 is the union of the three faces of the cube at the rear part , that would not be seen if the cube were not transparent , each one in a plane , orthogonal in the origin , to the respective coordinate axis .
the increasing orbits of the flux inside the cube , are in that figure , parallel lines orthogonal to the plane of the figure .
due to the perspective each of the orbits is seen as a black dot in the figure .
each black dot , for instance a " , represents a linear segment of an orbit .
the dot a " is the orbit of the flux @xmath108 from the initial state @xmath109 , in the left face of the cube ( that would not be seen if the cube were not transparent ) , to the front side of the cube , where the neuron @xmath110 reaches the threshold level 1 .
in that moment , the neuron @xmath111 , whose potential increased to reach one ( the state of the system is in the front vertical face of the cube ) , resets to zero , and the state of the system goes ( through a dotted horizontal line parallel to the axis @xmath112 ) , from the vertical front face of the cube ( where @xmath113 ) to the parallel vertical rear face in the poincar section @xmath0 , where @xmath114 .
( in the figure : from the point
a " at front , to the point b " at back ) . in that vertical rear face belonging to @xmath0 ,
the other two neurons @xmath115 , whose potentials @xmath116 were not reset , suffer a reduction of their potentials , of amplitudes @xmath117 and @xmath118 respectively , due to the inhibitory synaptic rule . that is why , the system does not stay in the point
b " of the figure , but jumps to c " , always in the rear face of the cube , corresponding to @xmath114 . at that instant , immediately after the first spike , from the point
c " in the backward rear face of the cube , the system starts to evolve again according to the differential equation , in the inter - spike regime , moving on an orbit inside the cube . in the figure this orbit corresponds to a segment orthogonal to the plane of the observer , collapsed in the black dot c " due to the perspective .
this new orbit arrives to the upper face of the cube , ( also in the black dot c " of the figure ) , meaning that neuron @xmath119 arrived to the threshold level one .
one could believe that figure [ fig1 ] is too particular , because the flux is linear inside the cube @xmath26 , with orbits that are parallel lines .
but , due to the tubular flux theorem , any flux tangent to a vector field @xmath120 such that @xmath121 , after an adequate differentiable ( generally non linear ) change of variables @xmath122 in the space , becomes a linear flux whose orbits are parallel lines , and project orthogonally to a certain plane .
one could transform all the dynamical system in these new coordinates @xmath123 , but in this case the synaptic jumps @xmath67 in the old variables @xmath13 would become synaptic jumps @xmath124 in the new variables @xmath125 .
maybe the matrix @xmath126 becomes dependent of the new variables @xmath127 , similarly to what was remarked in [ comments ] .
the tubular flux theorem is also valid in any dimension @xmath128 , and this geometric model has all the data that we will develop analytically in this paper . in particular , the tubular flux theorem and the linearization of the orbits , are used and analytically written in the proof of theorem [ teoremacontraccion ] .
the observer of the systems does not need to see " the orbits inside the 3-dimensional cube @xmath26 of the figure [ fig1 ] to study its dynamics .
he or she just see a point jumping inside the plane hexagon on which the cube @xmath26 projects orthogonally to the flux , and orthogonally to the plane of the draw , from the viewpoint of the observer .
there is a transformation @xmath15 from this hexagon to itself , giving the position of the point
c " from the initial state
a " , then the position of e " from c " , etc .
the dynamics by iterates of this transformation @xmath15 describes exactly the same dynamics of the system , which indeed evolves with continuous time @xmath5 , but is disguised as discrete .
it is not the system which was discrete nor the observer who made it discrete .
the observer positioned in the adequate viewpoint , without modifying the system , just to see it in its discrete disguise .
nevertheless there is a problem with this discrete system , if two or more neurons got the threshold level simultaneously .
for instance if the point a " would be in one of the three frontal sledges of the cube , that are sledges inside the hexagon , marked as not dotted lines in figure [ fig1 ] , there would be more than one possible consequent state @xmath129 , depending on which frontal face the system chooses to reset and apply the synaptic rule .
that is why the transformation @xmath15 has discontinuities , and is not a uniquely defined map in the frontal sledges , or lines of discontinuities .
all the arguments in this section could be obtained geometrically in the @xmath2-dimensional cube @xmath26 , just after its projection on an adequate poligon on an hyperplane , in which the state of the system evolves accordingly to a discrete transformation @xmath15 . *
[ definicionmodelomatematico].2 . * * the partition of @xmath0 in the continuity pieces @xmath130 . * recalling the definition of the spiking instant @xmath131 in equalities ( [ ecuaciontiempospike ] ) , and the definition of the set @xmath132 of all the neurons that reach the threshold level at time @xmath131 , in equality ( [ formulaconjuntoj ] ) , we define the following subset @xmath130 of the poincar section @xmath0 , for any @xmath4 : @xmath133 in other words , the set @xmath130 is formed by all the initial states @xmath25 in the poincar section @xmath0 such that the neuron @xmath10 reaches the threshold level before or at the same instant than all the other neurons of the network , from the initial state @xmath25 . from the implicit equation at right of formulae ( [ ecuaciontiempospike ] ) , we deduce that @xmath130 is compact , and that its interior @xmath134 is formed by all the initial states for which @xmath135 for all @xmath12
. then @xmath136 .
as the flux is strictly increasing inside @xmath26 , from any initial state @xmath137 there exists a finite time @xmath131 defined by equalities ( [ ecuaciontiempospike ] ) .
therefore @xmath138 for some not necessarily unique @xmath139 .
then @xmath140 and the family of subsets @xmath141 is a topological finite partition of @xmath0 ( i.e. it is a covering of @xmath0 with a finite number of compact sets whose interiors are pairwise disjoint . ) the compact sets @xmath130 are called _ continuity pieces . _
_ _ we define the _ separation line @xmath142 _ of the partition @xmath141 , or _ line of discontinuities__as the union of the topological frontiers @xmath143 of its subsets @xmath130 .
precisely : @xmath144 _ _ _ _ _ * [ definicionmodelomatematico].3 . * * the first return poincar map @xmath15 . * the first return map @xmath96 to the poincar section @xmath145 is the finite collection of maps @xmath146 defined as @xmath147 where @xmath148 is the solution flux defined in [ definicionmodeloinicial].2 of the system of differential equations ( [ ecuaciondiferencial ] ) , @xmath131 is the spiking instant defined by equalities ( [ ecuaciontiempospike ] ) and @xmath149 is the synaptic vectorial map defined in [ definicionmodeloinicial].4 .
* for simplicity we denote @xmath150 and , when it is previously clear that @xmath138 , we simply denote @xmath151 to refer to the uniquely well defined map @xmath152 . *
we observe that @xmath15 is uniquely defined in @xmath153 , and multi - defined in the separation line @xmath142 .
applying the formulae ( [ ecuaciontiempospike ] ) , ( [ ecuacionsinapsis1 ] ) and ( [ ecuacionsinapsis2 ] ) , we deduce : @xmath154 where by convenience we agree to define @xmath155 . *
[ definicionmodelomatematico].4 .
* * piecewise continuity of the poincar map @xmath15 .
* the formula ( [ formulareturnmapf ] ) implies that @xmath156 is continuous , and , as @xmath130 is compact , then @xmath157 is also compact .
the formula ( [ formulareturnmapf ] ) changes when one passes from @xmath130 to @xmath158 with @xmath159 , so @xmath15 is multidefined in the points of @xmath160 . besides @xmath15 may be discontinuous in @xmath161 because @xmath162 is not necessarily equal to @xmath163 .
we agree to define the image set @xmath164 of a point @xmath137 as @xmath165 .
the image set @xmath164 is a single point if @xmath166 because @xmath167 does not intersect @xmath168 for @xmath12 .
the image set @xmath169 of a set @xmath170 is by definition @xmath171 . *
[ definicionmodelomatematico].5 . * * the positive reduced poincar section @xmath172 .
* we will define a subset @xmath173 such that @xmath174 for all @xmath175 large enough .
our aim is to study the limit set of the orbits , therefore the last property allows us to restrict @xmath15 to @xmath172 .
@xmath176 where @xmath177 is the minimum of the absolute values of the synaptic interactions @xmath59 for @xmath178 , as assumed in [ definicionmodeloinicial].4 , equality ( [ formulaepsilo0grafocompleto ] ) .
also , by hypothesis ( [ ecuacioneslargedissipativeness1 ] ) we have @xmath179 * [ definicionmodelomatematico].6 .
* * properties of the positive reduced poincar section @xmath172 .
* the set @xmath172 is _ homeomorphic to a compact ball in @xmath180 , _
( property that the whole poincar section @xmath0 does not have ) .
in fact , @xmath172 is the union of @xmath2 compact squares @xmath181^{n-1 } \subset \mathbb{r}^{n-1}$ ] such that , for @xmath182 : @xmath183 is formed only by the @xmath184-dimensional lines @xmath185 in the frontiers of both squares . _ _ the dynamics properties of @xmath15 restricted to the positive poincar section @xmath172 , which precisely justify the restriction to @xmath172 , will be stated and proved in theorems [ teoremabmasinvariante ] , [ teoremafinjective ] , [ teoremasparationproperty ] and [ teoremacontraccion ] .
they justify the definition of the abstract mathematical model in [ definicionmodeloabstracto ] , whose dynamics in the future and attractors will be studied in the following sections . *
[ definicionmodelomatematico].7 . * * the positive continuity pieces @xmath186 of the poincar map .
* we define@xmath187 where @xmath172 is the positive reduced poincar section defined in [ definicionmodelomatematico].5 and @xmath130 are the continuity pieces of the poincar map @xmath15 , defined in [ definicionmodelomatematico].2 , equality ( [ ecuacionbsubi ] ) , and in [ definicionmodelomatematico].4 .
* remark : * it is rather technical to prove that _ @xmath186 is homeomorphic to a compact ball in @xmath188 . _
we sketch here a proof , leaving the technical details : @xmath189 is the pre - image in @xmath172 by the flux @xmath190 , of the @xmath45 square @xmath191 .
the flux is injective and continuous from @xmath172 onto its image @xmath192 , because it is transversal to @xmath44 and to @xmath172 and two different orbits of the flux do not intersect .
any two orbits that intersect @xmath172 in different points do intersect @xmath44 in different points .
continuous and injective maps @xmath148 from a ( homeomorphic ) ball @xmath172 in @xmath188 onto a set @xmath193 , are homeomorphisms , due to the theorem of the invariance of the domain .
therefore @xmath194 is homeomorphic to a @xmath45-dimensional compact ball , contained in @xmath44 . _
_ also @xmath195 is . to prove this last assertion , be aware that the intersection of two compact homeomorphic balls is not necessarily a single homeomorphic ball , but it holds in our model , because the frontiers of @xmath196 and @xmath197 have some symmetric properties due to the fact that the flux @xmath148 has @xmath2 components @xmath198 , each one depending only on the respective single variable @xmath199 .
then , @xmath200 is the homeomorphic image by @xmath201 of the compact ( homeomorphic ) ball @xmath195 . @xmath202 .
[ teoremabmasinvariante ] * the return map to the positive poincar section @xmath172 . *
the positive reduced poincar section @xmath173 defined in [ definicionmodelomatematico].5 , is forward invariant by the poincar map @xmath203 , and it is reached from any initial state in @xmath0 .
even more , @xmath204 and there exists @xmath205 such that @xmath206 _ ( recall that the constant @xmath207 defined in equality ( [ formulaepsilo0grafocompleto ] ) verifies the hypothesis ( [ epsilonceromenoruncuarto ] ) . ) _ _ _ to prove theorem [ teoremabmasinvariante ] we will use the following lemma : [ lematiempominimo ] there exists a constant positive minimum time @xmath208 @xmath209 such that , if @xmath137 verifies @xmath210 , then the interspike interval @xmath211 . _
proof : _ according to the formula ( [ ecuaciontiempospike ] ) : @xmath212 where @xmath213 is the solution of the implicit equation @xmath214 .
we integrate the differential equation ( [ ecuaciondiferencial ] ) with initial condition @xmath9 , and recall that @xmath215 , while the real solution @xmath216 is strictly increasing with @xmath217 ( for @xmath9 constant ) and it is the solution of an autonomous differential equation . using the hypothesis @xmath218 , and applying the inequality ( [ epsilonceromenoruncuarto ] ) ,
we obtain : @xmath219 @xmath220 @xmath221 where @xmath222 and @xmath223 , being @xmath224 the time that takes the flux @xmath225 to be equal to @xmath226 from the initial state @xmath227 .
recall that @xmath215 is strictly decreasing with @xmath9 : @xmath228 @xmath229 @xmath230 \leq \gamma_i(3/4 ) \ ; t_i(v_i)\ ] ] @xmath231 _ * proof of theorem [ teoremabmasinvariante ] : * _ it is enough to prove the following two assertions : * assertion [ teoremabmasinvariante].a : * @xmath232 ( even if @xmath233 ) , @xmath234 there exists a constant @xmath235 such that for all @xmath138 , if @xmath236 for some @xmath12 , then @xmath237 .
note that the assertion [ teoremabmasinvariante].b states its thesis in particular if @xmath233 , and also if @xmath238 and @xmath239 . recall that from the formulae ( [ formulareturnmapf ] ) of the poincar map @xmath15 : @xmath240 for all @xmath138 .
observe that , being @xmath241 for all @xmath137 , from the assertion [ teoremabmasinvariante].b we deduce that the first number @xmath205 of iterates of @xmath15 such that @xmath174 is at most equal to @xmath242 . to prove the assertion [ teoremabmasinvariante].a , apply the formulae ( [ formulareturnmapf ] ) of the return poincar map @xmath15 , and recall the assumptions ( [ formulaepsilo0grafocompleto ] ) , ( [ ecuacioneslargedissipativeness1 ] ) .
if @xmath138 then @xmath243 to prove the assertion [ teoremabmasinvariante].b , fix @xmath138 such that , for some @xmath12 @xmath244 use the formulae ( [ formulareturnmapf2 ] ) .
we assert that @xmath245 in fact , if it were greater or larger than @xmath77 , as @xmath246 due to hypothesis ( [ ecuacioneslargedissipativeness1 ] ) , the formulae ( [ formulareturnmapf2 ] ) would imply that @xmath247 contradicting our hypothesis ( [ equationefejotanopositivo ] ) .
due to the hypothesis of the differential equation ( [ ecuaciondiferencial ] ) , the function @xmath248 is strictly decreasing with @xmath13 , and the flux @xmath249 is strictly increasing with @xmath5 . use the integrate expression ( [ ecuaciondiferencialintegrada ] ) of the differential equation , to compute @xmath250 , the inequality ( [ ecuacionphimenoruncuarto ] ) and the lemma [ lematiempominimo ] , to deduce : @xmath251 @xmath252 recalling the integral equation ( [ ecuaciondiferencialintegrada ] ) and the formula ( [ formulareturnmapf2 ] ) of the return map @xmath15 , we deduce : @xmath253 @xmath254 to end the proof it is enough to show that @xmath235 .
recall from equality ( [ formulaepsilo0grafocompleto ] ) that @xmath255 @xmath256 @xmath257 where we have taken @xmath258 and @xmath64 such that @xmath259 .
@xmath260 @xmath261 applying the mean value theorem of the derivative of @xmath262 , which is negative due to the dissipation hypothesis of the differential equation in assumption ( [ ecuaciondiferencial ] ) , we obtain @xmath263 } \geq \frac{\min _ i \ , \min_{v_i \in [ 1/4 , 3/4 ] } |\gamma'_i(v_i)|}{2}\ ] ] the last inequalities and the assumptions ( [ ecuacioneslargedissipativeness2 ] ) and ( [ ecuacioneslargedissipativeness3 ] ) of relative large dissipative , imply : @xmath264 } |\gamma'_i(v_i)| } { \max_{k } \gamma_k ( 3/4 ) } \left ( \frac{1}{2 } - \frac{1}{4 } - \frac{1}{4 } \right ) = 0 \ ; \ ; \box\ ] ] _ * formula of the poincar map in @xmath172 . *
_ as a consequence of theorem [ teoremabmasinvariante ] , from now on we will restrict the poincar map @xmath15 to the positive section @xmath172 .
in fact , from the statements of theorem [ teoremabmasinvariante ] it is deduced that the forward dynamics and the limit set of the orbits to the future , of the restricted @xmath15 , will be the same as those of @xmath15 in the whole poincar section . due to theorem [ teoremabmasinvariante ] if @xmath238 then @xmath265 .
therefore @xmath266 . using the formula ( [ formulareturnmapf ] ) we can rewrite the expression of the poincar map , without the maximum : @xmath267 [ teoremafinjective ]
* local injectiveness of the poincar map . * the poincar map @xmath151 defined in formulae ( [ formulareturnmapf3 ] ) , restricted to each of its positive continuity pieces @xmath268 defined in _
[ definicionmodelomatematico].7 , _ is injective . _
_ _ proof : _ fix a continuity piece @xmath186 of @xmath15 in the positive poincar section @xmath172 .
the piece @xmath186 will remain fixed along this proof
. therefore we will denote @xmath15 instead of @xmath152 .
take @xmath269.such that @xmath270 .
we must prove that @xmath271 .
due to the formulas ( [ formulareturnmapf3 ] ) of the poincar map : @xmath272 and @xmath273 as @xmath67 is constant , we deduce that @xmath274 therefore the vectorial flux @xmath275^n $ ] defines an orbit from the initial @xmath238 that intersects the orbit from the initial state @xmath276 .
two different orbits of the flux do not intersect .
then , the two orbits are the same . if necessary changing the roles of @xmath25 and @xmath277 , we deduce that @xmath278 @xmath279 , so @xmath25 has at least one component @xmath280 and all of them not negative . but
@xmath281 is the strictly increasing in time solution of the differential equation @xmath282 , with @xmath283 for all @xmath284 .
we deduce that if @xmath285 , and if @xmath286 for @xmath287 , then @xmath288 , and therefore @xmath289 .
as we know that @xmath276 and @xmath290 , we conclude that @xmath291 , and then @xmath292 [ definicionseparationproperty ] * the separation property . * _ we say that @xmath15 verifies _ the separation property _ if @xmath293 where @xmath294 , are the continuity pieces of @xmath15 in the positive poincar section @xmath172 , as defined in [ definicionmodelomatematico].7 . ,
and @xmath152 is the continuous expression of @xmath295 according to the formulae ( [ formulareturnmapf3 ] ) . _ _ _ note that @xmath200 is compact for all @xmath10 , and @xmath15 is continuous in each @xmath200 .
therefore the image @xmath296 is a compact set .
then , the separation property implies that _ there exists a minimum positive distance @xmath297 between the images by @xmath15 of two different continuities pieces . _ _ _ [ teoremasparationproperty ] the poincar map @xmath15 verifies the separation property . _
proof : _ take @xmath200 and @xmath298 with @xmath178 .
the formulae ( [ formulareturnmapf3 ] ) of the poincar map @xmath299 and the theorem [ teoremabmasinvariante ] imply that @xmath300 @xmath301 then @xmath302 [ remarkgloballyinjectiveness ] * global injectiveness of the poincar map . * _ from theorems [ teoremafinjective ] and [ teoremasparationproperty ] it is deduced that the poincar map @xmath15 is globally injective in @xmath172 . in fact , if @xmath303 are in the same continuity piece @xmath186 , then @xmath304 because @xmath305 is injective . and
if @xmath303 respectively belong to two different continuity pieces @xmath186 and @xmath306 for @xmath178 , then @xmath307 because @xmath308 , due to the separation property .
we deduce that if @xmath303 then @xmath309 , where the image set @xmath164 of a point is defined in the remark [ definicionmodelomatematico].4 . _ [ teoremacontraccion ] * local contractiveness .
* the poincar map @xmath310 is uniformly contractive , but not infinitely contractive , in each of its continuity pieces @xmath200 . precisely , there exist two constant real numbers @xmath311 and a distance @xmath312 in the positive poincar section @xmath313 , such that , for all @xmath314 : @xmath315 where @xmath316 is the continuous restriction of @xmath15 to @xmath186 , according with formulae ( [ formulareturnmapf3 ] ) . * remark : * the distance @xmath312 of theorem [ teoremacontraccion ] induces the same topology in @xmath172 as homemorphic to a compact ball of @xmath317 .
in fact , along the proof of the theorem [ teoremacontraccion ] we will construct a linear projection @xmath318 and a diffeomorphism @xmath319 of @xmath320 class , such that : @xmath321 _ proof of the theorem _ [ teoremacontraccion ] : the continuity piece @xmath186is fixed .
for simplicity of the notation , along this proof we will use simply @xmath15 to denote @xmath152 .
the existence of the distance @xmath312 and the contraction rate @xmath322 is proved in the theorem 3 of @xcite . for a seek of completeness
we include here some pieces of the proof of @xcite , adding to them the existence of the bound contraction rate @xmath323 , @xmath324 .
due to the tubular flux theorem there exists a @xmath28 diffeomorphism which is a spatial change of variables @xmath325 from @xmath326 onto @xmath327 , such that @xmath328 and the solutions of the differential equation ( [ ecuaciondiferencial ] ) in @xmath26 verify @xmath329 in @xmath330 , where @xmath331 is a constant vector with positive components .
it verifies : @xmath332 define in @xmath333 the ortogonal projection @xmath334 onto the @xmath45-dimensional subspace @xmath335 the flux of the differential equation ( [ ecuaciondiferencial ] ) , after the change @xmath122 of variables in the space , is ortogonal to that subspace , and is transversal to @xmath336 ( recall that @xmath337 is the identity map ) .
consider any real function @xmath338 : @xmath339 it is left to prove that @xmath316 is contractive with this distance .
let us apply @xmath152 to @xmath25 and @xmath340 in @xmath341 .
we use the equalities ( [ formulareturnmapf3 ] ) .
we shall use the liouville derivation formula of the flux of the differential equation respect to its initial state : @xmath342 define : @xmath343 } \ ; \ ; { \gamma'_j ( v_j)}\ ; \ ; < 0 , \ ; \ ; \ ; \ ; \ ; \ ; -\alpha ^ * = \min_{j } \ , \min _ { v_j \in [ -1,1 ] } \ ; \ ; { \gamma'_j ( v_j)}\ ; \ ; < 0\ ] ] use the lemma [ lematiempominimo ] to bound uniformly above zero the inter - spike intervals @xmath131 : @xmath344 recall that @xmath131 is the solution of the @xmath320 implicit equation @xmath345 .
then @xmath346 is a continuous real function of @xmath238 , and @xmath172 is a compact set .
so , @xmath50 is also upper bounded by a constant : @xmath347 derive the formulae ( [ formulareturnmapf3 ] ) to obtain : @xmath348 _ { 1 \leq j \leq
n } & \nonumber \\ & & \nonumber \\ & { \partial f_j } /{\partial v_j } \ , = \left . ( { d \phi_j^{t}(v_j ) } /{d v_j } ) \right |_{t = \overline t ( v ) } = \exp \left ( \int _ 0 ^{\overline t(v ) } \gamma'_j ( \phi_j^s ( v_j ) ) \ , ds ) \right ) \in [ e^{-\alpha ^ * t^ * } , e^{- \alpha t } ] \nonumber & \\ & & \label{formulanombre}\\ & { \partial f_j } /{\partial v_i } \ , = \left .
( { d \phi_j^{t}(v_j ) } /{dt } ) \right |_{t = \overline t ( v ) } \cdot ( { d t_i ( v_i)}/{dv_i})= g(v ) \cdot \gamma_j(\phi_j^{\overline t ( v)}(v_j ) ) & \nonumber \end{aligned}\ ] ] where @xmath349 is the real function obtained deriving respect to @xmath9 the implicit equation given in ( [ ecuaciontiempospike ] ) : @xmath350 .
call @xmath351 to the @xmath352th .
vector of the canonic base in @xmath333 and join all the results above : @xmath353 we define the numbers @xmath354 and @xmath355 : @xmath356 and observe from the computations in ( [ formulanombre ] ) that : @xmath357 applying the definition of the differential distance @xmath312 in ( [ definiciondistanciacontractiva ] ) , and the equality ( [ ecuaciondiferencialderho ] ) , we obtain : @xmath358 integrating by formula ( [ definiciondistanciacontractiva ] ) we conclude : @xmath359 * local homeomorphic property of the poincar map . * each continuity piece @xmath152 of the poincar map in @xmath186 is an homeomorphism onto its image .
[ remarkhomeomorfismo ] _ it is an immediate consequence of theorem [ teoremacontraccion ] and the global injectiveness of @xmath15 .
even more , the continuous restriction @xmath360 , is lipschitz with constant @xmath361 and its inverse ( defined from @xmath362 ) is also lipschitz with constant @xmath363 .
then @xmath152 is an _ homeomorphism _ onto its image . @xmath364
_ _ _ we note that the same result can be obtained without using the theorem [ teoremacontraccion ] . due to theorem [ teoremafinjective ] , @xmath152 is injective , and due to the formulae ( [ formulareturnmapf3 ] ) , @xmath152 is continuous . due to the theorem of the invariance of the domain ,
any continuous and injective function from a ball in @xmath180 to @xmath180 is an homeomorphism onto its image . in the following corollary
we resume all the conclusions of this section : [ corolario ] if the network of @xmath2 inhibitory neurons verifies the assumptions of the physical model , _ evolving with real time @xmath5 in the phase space @xmath326 as stated in ( [ ecuaciondiferencial ] ) , ( [ ecuacionsinapsis1 ] ) , ( [ ecuacionsinapsis2 ] ) , ( [ formulaepsilo0grafocompleto ] ) , ( [ ecuacioneslargedissipativeness1 ] ) , ( [ ecuacioneslargedissipativeness2 ] ) and ( [ ecuacioneslargedissipativeness3 ] ) , _ then there exists a poincar section @xmath172 _ that is homeomorphic to a @xmath45-dimensional compact ball , _ and a return map @xmath365 , _ which has the following properties : _ _ _ _ _ * a ) * _ @xmath15 is piecewise continuous . _
precisely : there exist a finite partition @xmath366 of the poincar section @xmath172 , formed by compact sets @xmath186 homeormorphic to compact balls of @xmath188 , with pairwise disjoint interiors , and there exist @xmath2 continuous maps @xmath367 being @xmath368 for all @xmath238 . _
_ as a consequence @xmath15 is univoquely defined as @xmath152 in the interior of its continuity piece @xmath130 , and multidefined as @xmath369 in @xmath370 , if @xmath178 , @xmath371 .
* b ) * _ @xmath15 is locally uniformly contractive and not infinitely contractive , _
i.e. for some metric @xmath312 in @xmath172 the exist constants @xmath311 such that for all @xmath372 : @xmath373 . _
_ * c ) * _ @xmath15 has the separation property , _
i.e. @xmath308 if @xmath178 .
therefore , there exists @xmath374 . _ _
note that from b ) and c ) , it is deduced that @xmath15 is globally injective in @xmath172 , as proved in remark [ remarkgloballyinjectiveness ] . also from b )
it is deduced that @xmath375 is an homeomorphism onto its image .
* comments about the mathematical model . * due to corollary [ corolario ] , all the general results that we will prove for abstract piecewise continuous maps @xmath15 verifying a ) , b ) , c ) , are applicable to the networks of inhibitory neurons in the assumptions of the physical model stated in [ definicionmodeloinicial ]
. nevertheless the reciprocal of the corollary [ corolario ] does not hold . given a map @xmath15 verifying a ) ,
b ) , c ) there does not necessarily exist a network of inhibitory neurons in the hypothesis of the physical model stated in [ definicionmodeloinicial ] for which @xmath15 is its first return poincar map .
nevertheless we can wide our scenario of possible inhibitory neuronal networks models .
in fact , the properties a ) , b ) c ) are open ( in the uniform @xmath376 topology of the finite family of maps @xmath152 ) . thus they are not only verified by systems for which the differential equations ( [ ecuaciondiferencial ] ) are independent in the @xmath2 variables @xmath9 , but also if the system is of the form @xmath377 , where @xmath378 is a @xmath28 vector field , near enough the given @xmath379 , even if @xmath380 does not verify all the hypothesis stated in ( [ ecuaciondiferencial ] ) .
also the matrix @xmath381 of synaptic interactions in the network can be substituted for any matrix @xmath382 , not necessarily constant , but functions near the constant matrix @xmath381 and so , still verifying the assumptions ( [ formulaepsilo0grafocompleto ] ) , ( [ ecuacioneslargedissipativeness1 ] ) , ( [ ecuacioneslargedissipativeness2 ] ) , ( [ ecuacioneslargedissipativeness3 ] ) .
therefore , without changing the synaptical rules in equations ( [ ecuacionsinapsis1 ] ) and ( [ ecuacionsinapsis2 ] ) , but allowing the synaptic interactions slightly depend of the postsynaptic potentials , we will obtain a poincar map @xmath15 still verifying the thesis a ) , b ) , c ) of the corollary [ corolario ] .
besides , as observed in the subsection [ comments ] , the physical model includes looser hypothesis than those specified in [ definicionmodeloinicial ] , modulus any differentiable change of the variables of the system .
so , also in those models the properties a ) , b ) , c ) are verified by an open family of systems .
finally , the properties a ) b ) c ) of the corollary [ corolario ] are verified by many other models , in which the interspike regime is stated as a dynamical system depending continuously on time @xmath5 and on the initial state @xmath25 , but not necessarily as regular as to verify a differential equation .
the dynamics of the potential @xmath9 in the inter - spike interval may be given by a flux @xmath383 defined continuously in time @xmath5 , strictly increasing on @xmath5 , continuous but not necessarily differentiable respect to @xmath5 nor to the initial state .
but not all such general models are in the aim of this work .
they must be posed some hypothesis , to get the properties a ) b ) and c ) of the poincar section and its return map @xmath15 .
the arguments above justify to wide the abstract mathematical model of a network of @xmath2 inhibitory neurons , according to the following definition : [ definicionmodeloabstracto ] * the abstract mathematical model . * _ we say that a map @xmath365 , in a set @xmath172 homeomorphic to a compact ball of @xmath180 , _ models a generalized network of @xmath384 inhibitory neurons _ if it verifies the statements a ) , b ) , c ) of the corollary [ corolario ] . _ _ _
let @xmath385 be a compact set , homeomorphic to a compact ball of @xmath180 . in particular @xmath130 is connected .
[ particion ] _ a _ finite partition _ of @xmath0 is a finite collection @xmath386 of compact non empty sets @xmath130 of @xmath0 , homeomorphic to compact balls of @xmath188 , such that @xmath387 and @xmath388 , for @xmath389 . _ _ _ denote @xmath390 , and call @xmath142 the _ separation line , _ or _ line of discontinuities _ , although it is not a line in the usual sense , but the union of the topological frontiers of @xmath130 , each one homeomorphic to some @xmath184-dimensional manifold . _ _ _ _ _ [ piecewisecontinuous ] _ given a finite partition @xmath391 of @xmath0 , we call @xmath15 a _ piecewise continuous map on @xmath392 with the separation property _ if @xmath15 is a finite family @xmath393 of _ homeomorphisms _ @xmath394 , such that @xmath395 if @xmath178
. we note that @xmath15 is multi - defined in the separation line @xmath142 . _
_ _ _ each @xmath130 shall be called a _ continuity piece _ of @xmath15 . _ _ [ remarkfinversa ] _ a piecewise continuous map @xmath15 with the separation property is globally injective because it is an homeomorphism in each continuity piece and two different continuities pieces have disjoint images
. therefore @xmath396 exists , uniquely defined in each point of @xmath397 .
in fact : _ for any point @xmath398 , its backward first iterate is uniquely defined as @xmath399 , where @xmath10 is the unique index value such that @xmath400 .
nevertheless @xmath396 is not necessarily injective because @xmath15 is multidefined in @xmath401 .
@xmath396 is continuous in @xmath402 , because @xmath403 and @xmath152 is an homeomorphism due to the definition [ piecewisecontinuous ] . _
[ definicioncontraccionlocal ] _
we say that @xmath15 is uniformly _ locally contractive _ if there exists a constant @xmath404 , called an _ uniform contraction rate for @xmath15 , _ such that @xmath405 , for all @xmath406 and @xmath407 in the same @xmath130 , for all @xmath408 .
_ _ _ _ given a point @xmath409 , its image set is @xmath410 . if @xmath411 , its image set is @xmath412 .
we have that @xmath413 .
the second iterate of the point @xmath409 is the set @xmath414 .
analogously is defined the @xmath352th .
iterate as the set @xmath415 for any @xmath416 .
we convene to define @xmath417 and @xmath418 . _ [ definicionatomo ] for any natural number @xmath419 , we call _ atom of generation @xmath64 _ to @xmath420 where @xmath421 and @xmath422 is the subset of @xmath423 where the composed function above is defined . (
if @xmath422 were an empty set , then the atom is empty . )
abusing of the notation we write the atom as : @xmath424 _ _ _ we note that each atom of generation @xmath64 is a compact , not necessarily connected set , whose diameter is smaller than @xmath425 .
the set @xmath426 is a compact set , formed by the union of all the not empty atoms of generation @xmath64 .
there are at most @xmath427 and at least @xmath428 not empty atoms of generation @xmath64 , where @xmath428 is the number of continuity pieces of @xmath15 .
[ limitset ] _ _ given @xmath429 , a future orbit @xmath430 is a sequence of points @xmath431 , starting in @xmath432 , such that @xmath433 . due to the multi - definition of @xmath15 in the separation line @xmath142 , the points of @xmath142 and those that eventually fall in @xmath142 may have more than one future orbit .
a point @xmath407 is in the limit set @xmath434 of a future orbit of @xmath432 if there exists @xmath435 such that @xmath436 .
the limit set @xmath437 is the union of the limit sets of all its future orbits .
the _ limit set _
@xmath438 of the map @xmath15 , also denoted as @xmath439 , is the union of the limit sets of all the points @xmath409 .
_ _ _ _ _ [ remarkconjlimiteinvariante ] due to the compactness of the space @xmath0 the limit set @xmath434 of any future orbit , _ is not empty .
_ _ _ _ also , it is standard to prove that _
@xmath434 is compact _ ( because it is closed in the compact space @xmath0 )
. nevertheless @xmath437 may be not compact , if the point @xmath432 has infinitely many different future orbits . _ _
finally , we assert that _
@xmath434 is invariant : _
@xmath440 . _
_ proof : _ consider @xmath441 .
we have @xmath442 if @xmath443 .
@xmath444 is a continuous uniquely defined function in the compact set @xmath402 ( see remark [ remarkfinversa ] ) .
then @xmath445 , so @xmath446 proving that @xmath447 let us prove the converse inequality : @xmath448 .
@xmath449 is defined and continuous in each of its finite number of pieces @xmath130 that are compact and cover @xmath0 .
then there exists some @xmath450 and a subsequence ( that we still call @xmath451 ) , such that @xmath452 there exists @xmath453 such that @xmath454 .
in other words , @xmath455 .
this last assertion was proved for any @xmath441 .
therefore @xmath456 as wanted .
@xmath457 _
we say that a point @xmath406 is _ periodic of period @xmath175 _ if there exists a first natural number @xmath458 such that @xmath459 .
this is equivalent to @xmath406 be a periodic point in the usual sense , for the uniquely defined map @xmath396 , i.e. @xmath460 for some first natural number @xmath205 . _
_ _ we call the backward orbit of @xmath406 ( i.e. @xmath461 ) , a periodic orbit with period @xmath175
. we will prove in lemma [ lemaprevio ] that the limit set @xmath438 is contained in the compact , totally disconnected set @xmath462 .
it could be a cantor set .
but generically @xmath463 shall be the union of a finite number of periodic orbits , as we shall prove in theorem [ teorema1 ] .
_ [ finallyperiodic ] _
we say that @xmath15 is _ finally periodic _ with period @xmath175 if the limit set @xmath464 is the union of _ only _ a finite number of periodic orbits with minimum common multiple of their periods equal to @xmath175 .
in this case we call _ limit cycles _ to the periodic orbits of @xmath15 . _ _ _ _ _ _ we call basin of attraction of each limit cycle @xmath465 to the set of points @xmath409 whose limit set
@xmath466 is @xmath465 .
* topology in the space of piecewise continuous locally contractive maps in @xmath0 . * let @xmath467 and @xmath468 be finite partitions ( see definition [ particion ] ) of the compact region @xmath0 with the same number @xmath428 of pieces .
we define the distance between @xmath469 and @xmath470 as @xmath471 where @xmath472 denotes the hausdorff distance between the two sets @xmath44 and @xmath0 .
i.e. @xmath473 _ [ topologia ] let @xmath474 and @xmath475 be locally contractive piecewise continuous maps on @xmath392 and @xmath476 respectively .
given @xmath477 we say that @xmath478 is a _ @xmath479-perturbation of @xmath15 _ if @xmath480 where @xmath481 denotes _ the uniform contraction rate _ of @xmath15 in its continuity pieces , defined in [ definicioncontraccionlocal ] , and @xmath482 denotes the @xmath483 distance in the functional space of continuous functions defined in a _ compact _ set @xmath484 : @xmath485 _ _ _ _ _ _ _ [ definicionpersistencia ] _ we say that the limit cycles of a finally periodic map @xmath15 ( see definition [ finallyperiodic ] ) are _ persistent _ if : _ _ _ for all @xmath486 there exists @xmath487 such that all @xmath479-perturbations @xmath488 of @xmath15 are finally periodic with the same finite number of limit cycles ( periodic orbits ) than @xmath15 , and such that each limit cycle @xmath489 of @xmath488 has the same period and is @xmath490-near of some limit cycle @xmath491 of @xmath15 ( i.e. the hausdorff distance between @xmath492 and @xmath493 verifies @xmath494 ) . _ [ generico ] _
denote @xmath495 to the _ space of all the systems that are piecewise continuous with the separation property and locally contractive , _ according with the definitions [ piecewisecontinuous ] and [ definicioncontraccionlocal ] . _ _
we say that a property @xmath496 of the systems in @xmath495 ( for instance being finally periodic as will be shown in theorem [ teorema1 ] ) is _ ( topologically ) generic _ if it is verified , at least , by an _
open and dense _ subfamily of systems in the functional space @xmath495 , with the topology ( in @xmath495 ) defined in [ topologia ] .
_ _ _ _ precisely , being _
generic _ means : _ _
\1 ) the _ openness _ condition : for each piecewise continuous map @xmath15 that verifies the property @xmath496 there exist @xmath487 such that all @xmath479-perturbation of @xmath15 also verifies @xmath496 . _ _
\2 ) the _ denseness _ condition : for each piecewise continuous map @xmath15 that does not verify the property @xmath496 , given @xmath487 , arbitrarily small , there exist some @xmath479-perturbation @xmath488 of @xmath15 such that @xmath488 verifies the property @xmath496 . _
_ the openness condition implies that the property @xmath496 shall be robust under small perturbations of the system .
it is robust under small changes , not only of a finite number of real parameters , but also of the functional parameter that defines the model itself .
so the system should be structurally stable .
when this robustness holds , the property @xmath496 is still observed when the system , the model itself , does not stay exactly fixed , but is changed , even in some unknown fashion , remaining near the original one .
the density condition combined with the openness condition , means that the only behavior that have chance to be observed under not exact experiments are those that verify the property @xmath496 .
in fact , if the system did not exhibit the property @xmath496 , then some arbitrarily small change of it , would lead it to exhibit @xmath496 robustly . _
_ the denseness condition implies that if the property @xmath496 were generic , then the opposite property ( non-@xmath496 ) has null interior in the space of @xmath495 of systems , i.e. non-@xmath496 is not robust : some arbitrarily small change in the system will lead it to exhibit @xmath496 .
that is why we define the following : _ [ bifurcating ] if the property @xmath496 is generic , we say that any system that does not exhibit @xmath496 is _ bifurcating , _ and non-@xmath496 is a _ not persistent _ property .
[ teorema1 ] let @xmath15 be a locally contractive piecewise continuous map with the separation property . then generically @xmath15 is finally periodic with persistent limit cycles . _
_ to prove theorem [ teorema1 ] we shall use the following lemma : [ lema ] if there exists an integer @xmath419 such that the compact set @xmath497 does not intersect the separation line @xmath142 of the partition into the continuity pieces of @xmath15 , then @xmath15 is finally periodic and its limit cycles are persistent . _
proof : _ by hypothesis , @xmath498 , because @xmath484 and @xmath142 are disjoint compact sets . on the other hand @xmath499 , where @xmath500 denotes the family of all the atoms of generation @xmath64 . as the diameter of each of
the finite number of atoms of generation @xmath64 is smaller than @xmath501 , it converges to zero when @xmath502 .
thus , for all @xmath64 large enough , it is smaller than @xmath503 .
we assert that each atom @xmath44 of such generation @xmath64 , is contained in the interior of some continuity piece @xmath130 .
in fact , fix a point @xmath504 .
as the continuities pieces cover the space @xmath0 , there exists some ( a priori not necessarily unique ) index @xmath10 such that @xmath505 .
it is enough to prove that @xmath506 for all @xmath507 ( including @xmath406 itself ) .
we argue in the compact and connected metric space @xmath0 , using known properties of any general compact and connected metric space , for instance the triangular property , and also the property asserting that the distance of a point @xmath407 to a set , is the same that the distance of @xmath407 to the frontier of that set .
we denote @xmath508 to the complement of @xmath130 in @xmath0 , and in the topology relative to @xmath0 we denote : @xmath509 to the closure of @xmath508 , i.e the complement of @xmath134 , and @xmath143 to the frontier of @xmath130 in @xmath0 , @xmath510 : @xmath511 @xmath512 therefore @xmath513 proving the assertion . we deduce that given an atom @xmath514 , there exists and is unique a natural number @xmath515 such that @xmath516 . therefore @xmath169 is a single atom of generation @xmath517 . from the definition of atom in [ definicionatomo ] , we obtain that any atom of generation larger than @xmath64 is contained in an atom of generation @xmath64 . but
each atom of generation @xmath64 is in the interior of a piece of continuity of the partition @xmath518 .
we deduce that there exists a sequence of natural numbers @xmath519 , such that @xmath520 and the successive images of the atom @xmath44 of generation @xmath64 , are single atoms of generation @xmath521 .
therefore , the successive images of the atom @xmath44 , in the sequence ( [ equationimagenesdeatomo ] ) , are contained in a sequence of atoms : @xmath522 all of generation @xmath64 .
the same property holds for any of these atoms of generation @xmath64 , and each of them is contained in the interior of a continuity piece of @xmath15 , so @xmath15 is uniquely defined there and we have : @xmath523 the family of atoms of generation @xmath64 is finite , so we conclude that there exists two first natural numbers @xmath524 such that @xmath525 . note that , @xmath526 is uniquely defined as @xmath527 , because we are considering sets contained in the interior of the continuity pieces of @xmath15 . due to the uniform contractiveness of @xmath152 in each of its continuities piece , @xmath528 , is uniformly contractive .
the brower theorem of the fixed point states that in a complete metric space , any uniformly contractive map from a compact set to itself , has an unique fixed point , and all the orbits in the set converge to this fixed point in the future .
therefore , there exists in @xmath529 a periodic point by @xmath15 of period @xmath205 , and all the orbits with initial states in @xmath529 have the periodic orbit @xmath465 of @xmath175 , as their limit set . by construction @xmath529 was the image of @xmath44 by an iterate @xmath530 uniquely defined .
so we conclude that the limit set of all the points in the atom @xmath44 is @xmath465 .
the construction above can be done starting with any initial atom @xmath44 of generation @xmath64 . and
they are a finite family .
we conclude that there exists one and at most a finite number of periodic limit cycles , attracting all the orbits of @xmath531 .
the last assertion implies that the limit set of @xmath0 is formed by that finite family of periodic limit cycles .
finally it is left to prove that the limit cycles are persistent according to the definition [ definicionpersistencia ] .
the condition of the hypothesis of this lemma is open in the topology defined in [ topologia ] , because @xmath484 and @xmath142 are compact and at positive distance .
we assert that the itinerary of each of the atoms @xmath44 of generation @xmath64 , for @xmath64 fixed and large enough , remains unchanged when substituting @xmath15 by @xmath488 , being @xmath488 a @xmath479-perturbation of @xmath15 for @xmath487 small enough .
in fact @xmath44 and @xmath532 ( and also all the other atoms of generation @xmath64 , with @xmath64 fixed ) are contained in the images by @xmath533 or by @xmath534 respectively , of some of their one - to - one corresponding continuity pieces .
with @xmath64 fixed , if @xmath535 is sufficiently small , they remain at distance larger than the number @xmath536 from the separation lines of @xmath15 and of @xmath488 respectively , being @xmath537 .
thus , if the generation @xmath64 is chosen so the atoms have diameter smaller than @xmath538 , repeating the argument at the beginning of this proof we show that @xmath44 is in the interior of some continuity piece of @xmath15 , and @xmath532 is in the interior of the respective correspondent continuity piece of @xmath488 . on the other hand , the future iterates of any atom of generation @xmath64 by @xmath15 , and also by @xmath488 , are contained in the atoms of generation @xmath64
. therefore the images of an atom @xmath44 or @xmath532 of generation @xmath64 , by all the future iterates of @xmath15 or of @xmath488 respectively , are in the interior of their respective one - to - one correspondent continuity pieces .
then the itineraries are the same as we asserted . as a consequence , the indexes @xmath539 in the finite chain of atoms denoted in ( [ equationimagenesdeatomo ] ) and ( [ ecuacionchain2 ] ) ,
remain unchanged , and therefore we deduce the following statement : * a * : _ the number of periodic orbits in the atoms of generation @xmath64 , and their periods , remain unchanged , when substituting @xmath15 by any @xmath479-perturbation @xmath488 , if @xmath487 is sufficiently small .
_ _ _ it is standard to prove by induction on @xmath419 that for any @xmath479-perturbation @xmath488 of @xmath15 , such that @xmath540 , each atom @xmath532 of generation @xmath64 for @xmath488 , is at distance smaller than @xmath541 of the respective atom @xmath44 for @xmath15 with the same itinerary .
therefore we deduce the following statement : * b * : _ any periodic point found in an atom @xmath542 of generation @xmath64 for @xmath488 , is at distance smaller than @xmath543 than the respective periodic point found in the correspondent atom @xmath44 for @xmath15 with the same itinerary . _
_ the statements a and b imply that the limit cycles are persistent according to definition [ definicionpersistencia].@xmath544 _ in the proof of lemma [ lema ] we did not use the separation property @xmath545 . at the end of the proof of lemma [ lema ] we obtained that the piecewise continuous and locally contractive systems verifying the thesis of the lemma [ lema ]
, even if they do not have the separation property , contain an open family of systems in the topology defined in [ topologia ] .
then : _
_ in the space of all the piecewise continuous and locally contractive systems ( even if they do not have the separation property ) , those whose limit set is formed by a finite number of persistent limit cycles form an open family . _ _ _ nevertheless , to prove the genericity of the periodic persistent behavior , we need to prove that the family of periodic maps is dense in the space of systems . in the following proof , to obtain the density we shall restrict to the space of systems @xmath495 that verify the separation property . [ remarkperiodosgrandes ] _ from the proof of lemma [ lema ] , the first integer @xmath419 such that @xmath546 may be very large , and so the period @xmath175 may be very large . _ in fact , if the system has @xmath547 neurons , and if no neuron becomes dead , i.e. it does not eventually remain forever under the threshold level without giving spikes , then the periodic sequences @xmath548 , defined as the itinerary of the periodic limit cycles , have inside the period @xmath175 , at least once each of all the indexes @xmath549 . then @xmath550 . as we have shown in the proof of the lemma [ lematiempominimo ]
, there exists a minimum time @xmath551 between two consequent spikes .
suppose for instance that @xmath552 $ ] and @xmath547 .
the lasting time of the periodic sequence could be approximately @xmath553 = 10 ^ 9 [ s ] \geq 31 $ ] years .
so , if most of the neurons did not become dead , the observation of the theoretical periodic behavior of the inhibitory system in the future , could not be practical during a reasonable time of experimentation , and only the irregularities inside the period could be registered , showing the system as virtually chaotic . *
_ proof of theorem [ teorema1 ] . _ * due to lemma [ lema ] the existence of a finite number of limit cycles attracting all the orbits of the space is verified at least for those systems in the hypothesis of [ lema ] . at the end of the proof of lemma [ lema ] we showed that its hypothesis is an open condition . to prove its genericity it is enough to prove now that the hypothesis of lemma [ lema ] is also a dense condition in the space of piecewise continuous contractive maps with the separation property .
take @xmath15 being not finally periodic .
we shall prove that , for all @xmath487 there exists a @xmath554 perturbation @xmath488 of @xmath15 that verifies the hypothesis of lemma [ lema ] , and thus @xmath488 is finally periodic with persistent limit cycles .
let be given an arbitrarily small @xmath487 .
the contractive homeomorphisms @xmath152 of the finite family @xmath555 , with contraction rate @xmath556 , can be @xmath557 extended to @xmath558 , where @xmath559 is an homeomorphism onto its image defined in compact neighborhoods @xmath560 in @xmath0 ( i.e. @xmath561 where the closures and interiors of the sets are taken in the relative topology of @xmath562 ) , such that @xmath563 , and such that @xmath559 is still contractive , with a contraction rate @xmath564 such that @xmath565 .
the extended map @xmath566 , is now multidefined on @xmath567 .
the separation property is an open condition , thus the extension @xmath568 still verifies @xmath569 for all @xmath570 , if the neighborhoods @xmath571 and @xmath572 are chosen at a sufficiently small hausdorff distance from their respective pieces @xmath130 and @xmath168 , and @xmath535 is small enough . call @xmath573 to a positive real number smaller or equal than @xmath479 , and also smaller or equal than the distance from @xmath130 to the complement of @xmath571 , for all @xmath574 .
precisely @xmath575 consider the compact sets : @xmath576 @xmath577 define the _ extended atoms _ of generation @xmath578 for @xmath579 that form @xmath580 as @xmath581 , where @xmath582 is a word of length @xmath64 formed by symbols in @xmath583 . _ _ the diameter of each extended atom is smaller that @xmath584 . therefore ,
for sufficiently large @xmath578 all the extended atoms of generation @xmath64 that form @xmath580 have diameters smaller that @xmath585 .
we assert that the extended atoms of generation @xmath419 are pairwise disjoint : in fact , for two different @xmath570 the images are disjoint : @xmath586 .
so the atoms of generation 1 are pairwise disjoint .
two extended atoms of generation @xmath64 are @xmath581 and @xmath587 .
they can intersect if and only if @xmath588 because each @xmath559 is an homeomorphism onto its image .
so , they intersect if and only if they coincide . by construction , @xmath560 and @xmath563 .
therefore each of the atoms of generation @xmath64 for @xmath15 , is contained in the respective extended atom of generation @xmath64 for @xmath568 , that has the same finite word @xmath589 . if none of the extended atoms of generation @xmath64 intersects @xmath142 , then none of the atoms of generation @xmath64 for @xmath15 intersects @xmath142 , and the system verifies the hypothesis of lemma [ lema ] .
so , in this case , there is nothing to prove , because the given system @xmath15 verifies the thesis of the lemma [ lema ] and thus , it is finally periodic with persistent limit cycles . if some of the extended atoms of generation @xmath64 intersects @xmath142 , consider a new finite partition @xmath590 of @xmath0 such that the distance , defined in ( [ definicionhausdorfdistance ] ) , between @xmath470 and the given partition @xmath591 of @xmath15 , is smaller than @xmath235 : @xmath592 we shall besides choose the new partition @xmath470 such that the new separation line @xmath593 does not intersect the extended atoms of generation @xmath64 of @xmath580 .
this last condition is possible because the diameters of the generalized atoms are all smaller than @xmath594 , they are compact pairwise disjoint sets , and the distance between the two partitions @xmath469 and @xmath470 ( which is smaller than @xmath573 ) can be chosen larger than @xmath594 , defined in ( [ definicionhausdorfdistance ] ) as the maximum hausdorff distance between their respective pieces .
( we note that the old , and principally the new , separation lines @xmath595 and @xmath596 , are not necessarily @xmath320 nor even lipschitz manifolds in the space @xmath0 , and even if they are , they do not need to be @xmath320 or lipschitz near one from the other , to be near with the hausdorff distance ) .
the first condition @xmath597 , joined with the assumption @xmath598 , where @xmath130 is the @xmath599th piece of the partition @xmath469 , implies that the respective piece @xmath600 of the partition @xmath470 verifies @xmath601 .
therefore the extension @xmath602 in @xmath603 can be restricted to @xmath600 .
define @xmath604 where @xmath605 . by construction @xmath488 and @xmath15
coincide in @xmath606 , the distance between the respective partitions @xmath469 and @xmath470 is smaller than @xmath607 , and the difference of their respective contraction rates @xmath608 and @xmath322 is also smaller than @xmath479 .
so @xmath488 is a @xmath609-perturbation of the given @xmath15 , according to the definition [ topologia ] .
it is enough to prove now that @xmath488 is finally periodic with persistent limit cycles .
consider the limit set @xmath610 of @xmath488 as follows : @xmath611 as @xmath488 is a restriction of @xmath612 to the sets @xmath601 , we have that @xmath613 , and in particular for all @xmath419 the atoms of generation @xmath64 for @xmath488 , i.e. @xmath614 , are contained in the extended atoms of generation @xmath64 for @xmath568 . by construction the separation line @xmath615 among the continuity pieces @xmath600 of @xmath488 is disjoint with the extended atoms of generation @xmath64 of @xmath568 .
therefore , it is also disjoint with the atoms of generation @xmath64 of @xmath488 .
then @xmath616 and , applying lemma [ lema ] , @xmath488 is finally periodic with persistent limit cycles . @xmath364
it is possible ( but not immediate ) to construct , in a compact ball @xmath0 of any dimension @xmath617 , piecewise continuous systems , uniformly locally contractive and with the separation property , as defined in section [ definicionesabstractas ] , that do not verify the thesis of the theorem [ teorema1 ] , and thus their limit set is not composed only by periodic limit cycles .
suppose that the system had @xmath618 regularity , with @xmath619 or with @xmath620 , i.e. the continuity pieces @xmath130 and the separation lines that form @xmath142 , are bi - lipschitz homeomorphic ( or @xmath618 diffeomorphic respectively ) to @xmath45 or @xmath184 dimensional balls or manifolds , and the homeomorphisms @xmath152 in their continuity pieces @xmath621 , are bi - lipschitz ( or @xmath618-diffeomorphisms respectively ) . with this additional assumption of regularity of @xmath15 , it is an open question to construct examples that do not verify the thesis of theorem [ teorema1 ] .
in other words , assuming more regularity , it is unknown if the system has to exhibit a limit set always formed only by periodic orbits . on the other hand ,
it is also unknown if the existence of a finite number of limit cycles , as in theorem [ teorema1 ] , is generic for lipschitz or @xmath320 regular , locally contractive and piecewise continuous systems with the separation property .
the discontinuities of the poincar transformation @xmath15 , due to spike phenomena in the neural network , play an essential role to study these systems , although it is an obstruction to apply mostly previously known results of the theory of dynamics systems , which is mostly developed for continuous dynamics . due to the non - genericity of the bifurcating case , which is a consequence of theorem [ teorema1 ] , those dynamic
would never be seen in experiments : in fact , arbitrarily small perturbations in the parameters of the system will lead it to periodic or quasi- periodic dynamics .
these perturbations stabilize the system , to exhibit a limit set composed only by limit cycles .
also we showed that the inter - spike interval is bounded away from zero .
it means that , generically , when the system is periodic , in spite of having preferred periodic patrons of discharges , the neurons do not synchronize in phase .
on the other hand , if the number of neurons in the system is very large , the limit cycles of the network may have a very large period @xmath175 , much larger than the observation time , or even than the life time of the biological system .
therefore , in spite of being asymptotically periodic , these systems may never show its regularity .
these two facts : extremely large periods , and irregularity inside the period , allow us to assert that those persistent systems with very large period @xmath175 shall be in fact non - predictible for the experimenter , and will be perceived as virtually chaotic . | we prove that a topologically generic network ( an open and dense set of networks ) of three or more inhibitory neurons have periodic behavior with a finite number of limit cycles that persist under small perturbations of the structure of the network .
the network is modeled by the poincar transformation which is piecewise continuous and locally contractive on a compact region @xmath0 of a finite dimensional manifold , with the separation property : it transforms homeomorphically the different continuity pieces of @xmath0 into pairwise disjoint sets .
pacs 2008 codes : 87.19 lj , 87.19 ll , 87.19 lm msc 2000 92b20 , 34c25 , 37g15 , 24c28 keywords : neural network , piecewise continuous dynamics , limit cycles . |
the notion of a ` bare ' particle loses its significance once it is strongly coupled to a medium .
landau introduced the notion of a quasiparticle whose properties may be very different from those of a bare particle @xcite .
the most prominent example is an electron moving in a crystal : the electron displaces the nearby ions and carries this distortion with it .
the presence of the phonon cloud changes the mass and energy of the electron , that is dubbed as ` polaron ' @xcite .
more generally , a polaron arises whenever a quantum impurity is strongly coupled to an environment .
these quantum - mechanical quasiparticles play a key role in the low - energy behavior of a macroscopic quantum liquid . in recent years
, the field of ultracold atoms has provided an exciting framework for studying polaronic effects .
a key idea is that models designed for describing the rich and non - trivial structure of the solid state , can be emulated in a clean and controllable manner with ultracold atoms .
for example , so - called fermi polarons @xcite , spin - down impurities that are strongly coupled to a spin - up fermi sea , can be created in a degenerate two - component atomic fermi gas when going to the limit of strong spin - imbalance close to a feshbach resonance .
the impurity is coherently dressed with particle - hole excitations of the fermi sea .
the properties of the fermi polaron are important for the quantitative understanding of a strongly imbalanced fermi gas @xcite . in this paper , we focus on the ` attractive fermi polaron ' , with an attractive interaction between the impurity and the fermions of the bath .
a recent experiment using an ultracold gas of @xmath1li atoms in three dimensions revealed the existence of fermi polarons through a narrow quasiparticle peak in the impurities radio - frequency ( rf ) spectrum @xcite . at a critical interaction strength ,
the disappearance of this peak was interpreted as a transition from polaronic to molecular binding , when the impurity and an atom of the sea form a two - body bound state .
such a transition had theoretically been predicted in three dimensions ( 3d ) by prokofev and svistunov @xcite . to determine the transition point
, they developed a diagrammatic monte carlo technique ( diagmc ) capable of solving the fermi polaron model@xcite .
calculations of the ground - state energy showed that for a sufficiently strong attraction between the impurity atom and the atoms of the spin - up fermi sea , a molecular state becomes energetically favorable .
the crossing point was found at an interaction strength @xmath2 , with @xmath3 the fermi momentum of the spin - up sea and @xmath4 the @xmath5-wave scattering length . a variational treatment developed by chevy based on an expansion up to single particle - hole excitations on top of the unperturbed fermi sea turned out to be remarkably accurate @xcite .
a combination of chevy s ansatz with a variational wave function in the molecular limit @xcite also revealed the polaron - to - molecule transition , very close to the diagmc result . in the present work
, we study the quasiparticle properties of the fermi polaron problem in 3d with the diagmc technique @xcite .
this technique evaluates a series of feynman diagrams for the one - particle and two - particle proper self - energies .
a full description of the diagmc algorithm was presented in ref .
building on the work of ref .
@xcite we have implemented the diagmc algorithm independently .
we explore various diagmc schemes @xcite and series resummation methods to check the robustness of the results against the possible uncertainties of summing the series .
first , we confirm the transition point .
next , we calculate the quasiparticle residue which we compare to experimental data and variational results . the quasiparticle residue , or @xmath0-factor , gives the overlap of the non - interacting wave function and the fully interacting one , @xmath6 with @xmath7 the fully interacting ground state and @xmath8 a free spin - down atom carrying momentum @xmath9 in a non - interacting fermi sea @xmath10 of @xmath11 spin - up atoms .
the spin - up atoms are non - interacting since @xmath12-wave scattering is negligible .
the residue reflects the impurity s probability of free propagation .
the outline of the paper is as follows . in section [ sec1 ]
we introduce the model and the structure of the feynman diagrammatic expansion . in section [ sec2 ]
we discuss the results of the numerical calculations .
thereby , we investigate on how the results depend on the choices made with regard to the diagrammatic series , like the use of bare versus dressed propagators .
also the resummation of the diagrammatic series is discussed in depth .
the results for the quasiparticle properties , like the residue , are the subject of section [ sec3 ] .
we consider a dilute two - component gas of ultra - cold fermionic atoms interacting via the van der waals - potential .
the hamiltonian has a kinetic and interaction term @xmath13 the operators @xmath14 ( @xmath15 ) create ( annihilate ) fermions with momentum @xmath16 and spin @xmath17 .
the spin-@xmath17 fermions have mass @xmath18 and dispersion @xmath19 , and @xmath20 is the volume of the system . we take @xmath21 throughout the paper , and consider the mass - balanced case @xmath22 .
all the theoretical considerations are for zero temperature ( or @xmath23 with @xmath24 the fermi temperature ) .
the diluteness of the system ensures that the range @xmath25 of the potential is much smaller than the typical inter - particle distance @xmath26 , or @xmath27 , with @xmath3 the fermi momentum of the spin - up sea , and therefore the details of the interaction potential become irrelevant .
accordingly , without loss of generality , one can model the short - ranged interaction as a contact interaction , @xmath28 , in combination with the standard ultra - violet divergence regularization procedure described below .
the one - body and two - body propagators provide access to information about the ` polaron ' and ` molecule ' channel respectively .
the polaron and molecule are two distinct objects belonging to different charge sectors .
the one - body and two - body propagators are discussed in sections [ subseca ] and [ subsecc ] .
the adopted regularization procedure for the renormalized interaction is the subject of section [ subsecb ] .
the diagmc method is introduced in section [ subsecd ] .
the polaron quasiparticle properties can be extracted from the impurity s green s function defined as @xmath29 with @xmath30 the annihilation operator in the heisenberg picture , @xmath31 the propagator @xmath32 is written in the momentum imaginary - time representation , @xmath33 is a free parameter , @xmath34 is the number operator for spin-@xmath17 particles , and @xmath35 is the chemical potential of the spin - up sea .
the state @xmath36 consists of the spin - down vacuum and the non - interacting spin - up fermi sea . since we are dealing with an impurity spin - down atom ,
@xmath37 is only non - zero for times @xmath38 .
the ground - state energy and @xmath0-factor can be extracted from the green s function of eq .
( [ eq : gdown ] ) . inserting a complete set of eigenstates @xmath39 of the full hamiltonian ( [ eq : ham ] ) for one spin - down particle and @xmath11 spin - up particles into eq .
( [ eq : gdown ] ) yields for @xmath40 @xmath41 with @xmath42 the energy of the polaron , @xmath43 the energy eigenvalues of the hamiltonian ( [ eq : ham ] ) and @xmath44 the energy of the ideal spin - up fermi gas , with @xmath45 the fermi energy .
the difference between the polaronic and molecular state is embedded in the factors @xmath46 in eq .
( [ eq : gdownlehman ] ) . for situations where the polaron is a well - defined quasiparticle in the ground state @xmath47
, we have eq . ( [ eq : z ] ) for the @xmath0-factor and @xmath48 .
if , on the other hand , the ground state @xmath49 is a dressed molecule the overlap @xmath50 is zero @xcite .
this is clear from the expansion of the molecular state in the number of particle - hole excitations , @xmath51 the coefficients @xmath52 are variational parameters , and the primes indicate that the sums on @xmath16 , @xmath53 and @xmath54 are restricted to @xmath55 and @xmath56 . even if a molecule is formed in the ground state , the polaron can be a well - defined excited state ( in the sense of a narrow peak in the spectral function ) , and @xmath57 can be non - zero . for vanishing interactions @xmath58 the impurity green s function of eq .
( [ eq : gdown ] ) becomes @xmath59 the one - body propagator for the spin - up sea is defined as @xmath60 ~ | \psi^{n_{\uparrow}}_0 \rangle \ ; , \label{eq : gup}\ ] ] with @xmath61 the time - ordering operator . without interactions ,
one obtains the free propagator @xmath62 our goal is to calculate the @xmath37 of eq .
( [ eq : gdown ] ) to extract @xmath42 by means of eq .
( [ eq : gdownlehman ] ) .
this is achieved by summing all irreducible one - particle self - energy diagrams with the diagmc algorithm ( which works in momentum - imaginary - time representation ) .
the irreducible self - energy @xmath63 in imaginary - frequency representation is obtained after a numerical fourier transform , and inserted into dyson s equation to give @xmath37 , @xmath64^{-1}= \big [ g^{0}_{\downarrow } ( \mathbf{k } , \omega)\big]^{-1 } - \sigma(\mathbf{k } , \omega ) \ ; , \label{eq : dyson}\ ] ] with @xmath65 the imaginary frequency
. a graphical representation of the dyson equation is shown in the top panel of fig .
[ fig : dyson ] . as was shown in ref .
@xcite , the polaron energy @xmath42 and @xmath0-factor @xmath57 can be extracted directly from the self - energy @xmath66 , @xmath67 the effective mass @xmath68 of the polaron is evaluated with the estimator@xcite @xmath69 with @xmath70 \ ; , \end{aligned}\ ] ] which can conveniently be estimated by expanding @xmath71 in legendre polynomials .
one obtains @xmath72 and the integral can be evaluated during the mc simulation .
the upper limit of integration ( @xmath73 ) is optimized to minimize the statistical noise while avoiding a systematic error at too large @xmath73 .
we also used an alternative way by calculating the quasiparticle spectrum @xmath74 and fitting @xmath68 via @xmath75 .
we introduce the @xmath5-wave scattering length @xmath4 for collisions between spin - up and spin - down particles .
one of the advantages of working with feynman diagrams is that one can work directly in the zero - range limit @xmath76 ( or , equivalently , @xmath77 with @xmath78 an ultraviolet momentum cut - off ) while keeping @xmath79 constant .
thereby , the ultra - violet physics can be taken into account by means of a summation over all feynman ladder diagrams . in momentum - imaginary - frequency representation @xmath80 ,
one obtains for the partially dressed interaction @xmath81 with @xmath82 the two - particle self - energy consisting of one ` bare ' ladder @xmath83 where the momentum cutoff @xmath78 is required to keep the sum finite .
the bare coupling constant @xmath84 can be eliminated in favor of the physical scattering length @xmath4 by using standard scattering theory @xmath85 the @xmath86 from eq .
( [ eq : dysongamma0tilde ] ) can be expressed in terms of the @xmath5-wave scattering length @xmath4 , by taking the limit @xmath87 and @xmath88 with @xmath4 fixed .
in this zero - range limit , one gets @xmath89^{-1 } = [ \tilde{\gamma}^0(\mathbf{p } , \omega ) ] ^{-1 } - \bar{\pi}(\mathbf{p } , \omega ) \label{eq : gamma0tilde } \ ; , \end{aligned}\ ] ] with @xmath90 here , we have taken the thermodynamic limit ( @xmath91 and @xmath92 fixed ) .
the integral in eq .
( [ eq : pibar ] ) can be evaluated analytically , and the dressed interaction in vacuum is given by @xmath93^{-1 } = \frac{m}{4 \pi a } - \frac{m}{8\pi } \sqrt{p^2 - 4 m ( i \omega + \mu + \epsilon_f ) } \ ; , \ ] ] for @xmath94 or @xmath95 , and assuming the principal branch . for @xmath96 $ ]
, the fourier transform to imaginary time can be done analytically , producing @xmath97 with @xmath98 the complementary error function . as in ref .
@xcite , we use @xmath99 as a partially dressed interaction vertex in the diagrammatic series , instead of the bare interaction vertex @xmath100 .
this dressed vertex is calculated here in imaginary time representation by performing the fourier transform of eq .
( [ eq : gamma0tilde ] ) numerically . in a next step ,
the interaction vertex will be fully dressed by calculating the two - particle self - energy @xmath101 and plugging it into the bethe - salpeter equation , @xmath102^{-1 } = [ \gamma^0(\mathbf{p } , \omega ) ] ^{-1 } - \pi(\mathbf{p } , \omega ) \ ; .
\label{eq : bs}\ ] ] a graphical representation of this equation is shown in fig .
[ fig : dyson ] .
the self - energy @xmath101 contains all connected two - particle diagrams that are irreducible with respect to cutting a single @xmath103 propagator .
to avoid double counting , the diagrams for @xmath101 should not contain any ladders , since those have been summed in @xmath103 by means of the eq .
( [ eq : dysongamma0tilde ] ) .
this rule also holds when summing diagrams for the one - body self - energy @xmath104 , built from free propagators @xmath105 and @xmath103 . here
, we consider the pair annihilation operator , @xmath106 with @xmath107 the momentum representation of the wave function @xmath108 for relative motion of the two fermions of opposite spin .
the two - particle propagator is defined as @xmath109 where we included the fact that the impurity spin-@xmath110 atom propagates forward in time .
inserting the complete basis @xmath111 for ( @xmath112 ) spin - up particles and one spin - down particle , gives @xmath113 with @xmath114 the molecule energy and @xmath115 the molecule @xmath0-factor .
if the molecule is a well - defined quasiparticle in the ground state , we have @xmath116 and @xmath117 .
note that the value of @xmath115 depends on the wave function @xmath107 .
the functional from of this pair wave function depends on the nature of experiment used to probe the molecule . in practice
, it is easier to calculate the molecule energy from the fully dressed interaction @xmath118 ( see eq .
( [ eq : bs ] ) ) .
this function is closely related to the pair correlation function , namely , @xmath119 with @xmath120 the pair correlation function .
the field operators @xmath121 create a spin-@xmath17 fermion at position @xmath122 . in eq .
( [ eq : pcorr ] ) the pair of particles is created at the same position ( which corresponds to @xmath123 in eq .
( [ eq : pair ] ) ) .
the structure of the fully dressed interaction @xmath118 and the two - particle propagator @xmath124 now implies that both structures have the same poles ( see eq .
( [ eq : gammap ] ) ) .
therefore , the exponential tail of the function @xmath125 can conveniently be used for estimating the molecule energy , rather than the tail of @xmath126 .
this is equivalent with looking for this pole of the bethe - salpeter equation ( [ eq : bs ] ) .
the molecule s energy @xmath114 is given by the parameter @xmath33 that satisfies the equation @xmath127^{-1 } = \pi(\mathbf{p}=\mathbf{0 } , \omega=0 ) \ ; , \ ] ] where the left - hand - side is known analytically , and the right - hand - side is evaluated with the diagmc algorithm in imaginary - time domain .
diagmc evaluates the series of feynman diagrams for the self - energy in a stochastic way .
we deal with both the one - body and two - body self - energies . in a first step ,
the self - energy is built from the free propagators @xmath105 and the partially dressed interaction @xmath103 ( obtained through summation of @xmath128 ladders ; as discussed in section [ subsecb ] ) .
we will refer to this series as the ` bare series ' .
[ fig : series_bare ] shows the one - body and two - body self - energy diagrams up to order @xmath129 in the bare scheme .
the order of a diagram is @xmath130 when there are @xmath130 dressed interactions @xmath118
( i.e. , @xmath130 boxes ) present in the @xmath104-diagram , and @xmath131 boxes in the @xmath101-diagram .
note that the diagrams can not contain ladders since these have been taken into the vertex function @xmath103 . to illustrate the factorial growth with order , the number of one - body self - energy diagrams for given order @xmath132 is given in the second column of table [ tab : numbers ] for the bare series . in a second step , we will use dressed propagators or ` bold lines ' in the diagrams .
such dressed ( skeleton ) series are evaluated with the bold diagmc technique @xcite .
we consider the case with only dressed @xmath37 propagators while keeping @xmath103 of eqs .
( [ eq : gamma0tilde ] ) and ( [ eq : pibar ] ) as renormalized interaction , and the case whereby both the one - body propagators and interactions are dressed .
we will refer to these skeleton series as ` bold @xmath133 ' and ` bold @xmath133-@xmath118 ' , respectively . in the latter case ,
the bold diagmc algorithm is constructed as follows : given approximate one - body and two - body self - energies @xmath104 and @xmath101 , the dyson and bethe - salpeter equation are solved to deliver the one - body propagator @xmath37 and the dressed interaction @xmath118 ( see eqs .
( [ eq : dyson ] ) and ( [ eq : bs ] ) ) . in a next step ,
these are used to dress the series for @xmath104 and @xmath101 , which are evaluated stochastically with diagmc up to order @xmath134 .
this self - consistent cycle is repeated until convergence is reached .
[ fig : series_bold ] shows the skeleton ( bold @xmath133-@xmath118 ) series for the one- and two - body self - energies up to order 4 . evidently , when dressing the lines in the self - energies , one has to keep track of two - particle reducibility , and systematically avoid any double counting .
this typically means that at any order @xmath130 the numbers in the second column of table [ tab : numbers ] ( bare ) are an upper limit of the number of diagrams in the third and fourth column . at @xmath135 and @xmath136 , however , the number of diagrams increases due to the fact that ladders should be included again once @xmath37 is bold .
all the diagrams of table [ tab : numbers ] are summed explicitly during the ( bold ) diagmc simulation . .
factorial increase of the number of feynman diagrams . at fixed order @xmath130
, the number of one - body self - energy diagrams is given for different types of series : the bare series , the skeleton series with dressed @xmath37 ( bold @xmath133 ) , and the skeleton series with dressed @xmath37 and @xmath118 lines ( bold @xmath133-@xmath118 ) . [ cols="<,>,>,>",options="header " , ]
we have considered the fermi - polaron system in three dimensions , in which a single spin - down impurity is strongly coupled to a non - interacting fermi sea of spin - up particles .
although this system contains strongly interacting fermions , it can be solved with the diagrammatic monte - carlo method .
this method is based on the stochastic evaluation of a series of feynman diagrams . to extract ground - state properties
, one has to overcome a factorial complexity due to the increase of the number of diagrams .
nonetheless , extrapolation to infinite diagram order becomes possible when the diagrams cancel each other better than the factorial increase in number . at interaction strength @xmath137
, we find such perfect cancellation ( within our statistical errors ) .
when considering the series built on bare propagators on the bec side , however , oscillations with diagram order remain and prevent a controlled extrapolation to the infinite diagram order . we have followed two strategies around this problem : the first is to consider skeleton series ( built on dressed propagators ) , and the second the use of resummation techniques . though dressed series can be evaluated to higher orders
, we have found that in some cases dressing can destroy a favorable cancellation of diagrams .
for all interaction strengths we found that the ( skeleton ) series of the one - body and two - body self - energy is resummable by means of abelian resummation .
bare series , skeleton series and resummed series give robust answers in their respective region of applicability ( i.e. where the infinite diagram order extrapolation is controlled ) .
we have identified classes of dominant diagrams for the one - body and two - body self - energy in the crossover region of strong interaction .
the dominant diagrams turn out to be the leading processes of the strong - coupling limit : scattering between a dimer and a spin - up fermion , which is diagrammatically represented by the three - body t - matrix diagrams . including just these dominant diagrams
gives a quantitatively good correction to the lowest order result , even away from the strong - coupling limit .
we have shown that not only the polaron and molecule energies agree very well with a variational ansatz from weak to strong attraction , but also the polaron residue or @xmath0-factor .
though this agreement must be due to strong cancellation of diagrams , we only observed convergence for the bare series at @xmath137 . a full explanation for the success of the variational ansatz is still missing , and it is therefore unclear in which cases the ansatz is appropriate .
this work is supported by the fund for scientific research - flanders .
the authors would like to thank c. lobo , n. prokofev , b. svistunov , f. werner and m. zwierlein for the helpful discussions and suggestions .
we thank r. combescot , s. pilati , n. prokofev , m. punk , b. svistunov and m. zwierlein for sending us their data .
landau , phys .
z. sowjetunion * 3 * , 664 ( 1933 ) .
lifshitz and l.p .
pitaevskii , _ statistical mechanics _ , part 2 ( pergamon press , new york , 1980 ) .
prokofev and b.v .
svistunov , phys .
b * 77 * , 020408 ( 2008 ) .
a. schirotzek , c.h .
wu , a. sommer and m.w .
zwierlein , phys .
* 102 * , 230402 ( 2009 ) .
s. nascimb ` ene , n. navon , k.j .
jiang , l. tarruell , m. teichmann , j. mckeever , f. chevy and c. salomon , phys .
. lett . * 103 * , 170402 ( 2009 ) .
s. pilati and s. giorgini , phys .
* 100 * , 030401 ( 2008 ) . s. pilati , private communication . n.v . prokofev and b.v .
svistunov , phys .
b * 77 * , 125101 ( 2008 ) .
f. chevy , phys .
a * 74 * , 063628 ( 2006 ) . c. mora and f. chevy , phys .
a * 80 * , 033607 ( 2009 ) .
m. punk , p.t .
dumitrescu and w. zwerger , phys .
a * 80 * , 053605 ( 2009 ) .
r. combescot , s. giraud and x. leyronas , epl * 88 * , 60007 ( 2009 ) .
prokofev and b.v .
svistunov , phys .
* 99 * , 250201 ( 2007 ) .
hugenholtz , physica * 23 * , 533 ( 1957 ) .
r. combescot and s. giraud , phys .
* 101 * , 050404 ( 2008 ) .
s. giraud , ph.d .
thesis , universit paris vi , 2010 , http://tel.archives-ouvertes.fr/tel-00492339 x. leyronas and r. combescot phys .
lett . * 99 * , 170402 , ( 2007 ) .
skorniakov and k.a .
ter - martirosian , zh .
* 31 * , 775 ( 1956 ) [ sov .
jetp * 4 * , 648 ( 1957 ) ] .
hardy , _ divergent series _ ( oxford university press , new york , 1956 ) .
k. van houcke , f. werner , e. kozik , n. prokofev , b. svistunov , m.j.h .
ku , a. t. sommer , l.w .
cheuk , a. schirotzek and m.w .
zwierlein , nature phys .
* 8 * , 366 ( 2012 ) .
p. dennery and a. krzywicki , _
mathematics for physicists _ ( dover , new york , 1996 ) . y. ohashi and a. griffin , phys . rev .
a * 72 * , 013601 ( 2005 ) .
p. massignan , g.m .
bruun and h.t.c .
stoof , phys .
a * 77 * , 031601(r ) ( 2008 ) .
c. trefzger and y. castin , arxiv:1210.8179 . | we report on a study of a spin - down impurity strongly coupled to a spin - up fermi sea ( a so - called fermi polaron ) with the diagrammatic monte - carlo ( diagmc ) technique .
conditions of zero temperature and three dimensions are considered for an ultracold atomic gas with resonant interactions in the zero - range limit .
a feynman diagrammatic series is developed for the one - body and two - body propagators providing information about the polaron and molecule channel respectively .
the diagmc technique allows us to reach diagram orders that are high enough for extrapolation to infinite order .
the robustness of the extracted results is examined by checking various resummation techniques and by running the simulations with various choices for the propagators and vertex functions .
it turns out that dressing the lines in the diagrams as much as possible is not always the optimal choice .
we also identify classes of dominant diagrams for the one - body and two - body self - energy in the region of strong interaction .
these dominant diagrams turn out to be the leading processes of the strong - coupling limit .
the quasiparticle energies and @xmath0-factor are obtained as a function of the interaction strength .
we find that the diagmc results for the molecule and polaron properties are very similar to those obtained with a variational ansatz .
surprisingly , this variational ansatz gives very good predictions for the quasiparticle residue even when this residue is significantly smaller than one . |
several decades have passed since the kondo effect was discovered experimentally in the 1930s @xcite and was first explained theoretically in the 1960s @xcite
. however , the kondo problem and related subjects are still quite attractive in both theoretical and experimental studies @xcite .
quantum impurity systems show universal critical behavior at low temperature , which is characterized by the ratio of the impurity magnetic susceptibility @xmath3 to the impurity specific heat @xmath4 with temperature @xmath5 @xcite @xmath6 we call it the wilson ratio .
it was exactly shown by the bethe ansatz that it is given by 2 for the kondo model @xcite .
the effect of an impurity embedded in a one - dimensional interacting quantum system or in a tomonaga - luttinger liquid @xcite has been one of the most attractive topics during the 1990s and also in the last decade .
it has been investigated in a variety of systems by different methods such as renormalization group techniques @xcite , conformal field theories ( cft ) @xcite , numerical techniques with cft @xcite , and the bethe - ansatz method @xcite .
recently , it is studied by functional renormalization group @xcite and also experimentally in a quasi - one - dimensional conductor @xcite .
however , it is still rare that the finite - temperature thermodynamic behavior is explicitly and exactly shown by a theoretical method for a large but finite lattice system without making any approximation or assumption . in this paper
we show explicitly the finite - temperature behavior of an integrable model of the spin-1/2 xxz periodic chain with a spin-1/2 impurity .
we remark that the spin part of the kondo model is quite similar to the heisenberg spin chain , i.e. the xxx model .
it is thus interesting to study the integrable spin - chains with the spin-1/2 impurity which are solved by the bethe ansatz . in the xxz impurity model
we investigate exactly the effect of an impurity embedded in the interacting quantum spin chain similarly as the kondo model .
schlottmann studied the integrable spin-@xmath7 xxz spin chain with one spin-@xmath8 impurity @xcite in association with the multi - channel kondo effect @xcite .
the spin-1/2 impurity in the open spin-1/2 xxx chain was studied by frahm and zvyagin @xcite and it was shown that the kondo - like temperature exists , while the periodic xxz chain with an impurity was studied by eckle et al . @xcite . however , the analytic expressions of the impurity magnetization and the impurity specific heat at low temperature in ref .
@xcite are not accurate enough to evaluate the wilson ratio correctly .
they are not consistent with the numerical estimates evaluated by solving the bethe ansatz equations , in particular , when the impurity parameter is small .
moreover , the crossover temperature has not been evaluated numerically , while its analytic expression @xcite is not appropriate .
the thermodynamic behavior of the spin-1/2 impurities in the spin-1/2 xxz chain was shown in ref .
@xcite for several distributions of the impurity parameters , and in the homogeneous case the results are consistent with ours . in the present paper
we show how the thermodynamic behavior of the single spin-1/2 impurity depends on the impurity parameter . although it looks simpler than that of many impurities , it is still nontrivial and interesting from the viewpoint of universality and the crossover from the @xmath0-site to the @xmath1-site chain .
we consider the integrable model of the spin-1/2 xxz periodic chain with a spin-1/2 impurity , i.e. , the case of @xmath9 in the xxz impurity model of ref .
we investigate the finite - temperature behavior of the integrable xxz impurity model by numerically solving the truncated integral equations of the thermal bethe ansatz @xcite , and evaluate the specific heat and entropy numerically . by plotting graphs of the impurity specific heat and entropy versus temperature we show that the impurity spin gradually becomes a free spin through pseudo - decoupling from other spins , while in low temperature it couples strongly to them such as the kondo effect . here , we call it _ pseudo - decoupling
_ , if the exchange coupling between the impurity spin and other spins is very small but nonzero .
it occurs in the present model when the absolute value of the impurity parameter is very large .
we then show that the estimate of the wilson ratio at a given low temperature is independent of the impurity parameter if the absolute value of the impurity parameter is small with respect to the given temperature .
we evaluate the local magnetization at the impurity site under a given small magnetic field at zero temperature and derive its analytic expression by the two methods : by identifying the impurity contribution in the bethe ansatz equations physically ; by calculating it directly through the algebraic bethe ansatz method @xcite .
we show that they are consistent : the impurity contribution identified with the bethe ansatz equations is valid .
let us denote the hamiltonian of the xxz spin chain with the spin-1/2 impurity consisting of @xmath0 sites by @xmath10 , which depends on the impurity parameter , @xmath11 .
if the absolute value of the impurity parameter , @xmath12 , becomes large , the exchange coupling of the impurity spin to neighboring spins becomes small , and hence the hamiltonian of the @xmath0-site chain is decomposed into that of the @xmath1- site chain , @xmath13 , and that of the impurity site , @xmath14 , as follows .
@xmath15 as we shall see in 2 the term @xmath16 gives the hamiltonian of the spin-1/2 periodic xxz chain with no impurity . applying a small magnetic field , @xmath17 , on the whole system of @xmath18 down spins we have @xmath19 . here
the hamiltonian is given by @xmath20 , where @xmath21 denotes the eigenvalue of the @xmath22-component of the total spin operator .
we also assume that the xxz spin chain , @xmath16 , has the anti - ferromagnetic interactions . then , any down spin on the impurity site is attracted to the the @xmath1-site chain due to the anti - ferromagnetic interactions , since the ground - state energy of @xmath16 with @xmath18 down spins decreases as the number of down spins , @xmath18 , increases .
we thus suggest that the impurity spin tends to be given only by an up - spin and all the @xmath18 down - spins on the whole chain tend to be located on the @xmath1-site chain under non - zero magnetic field if the absolute value of the impurity parameter is large , i.e. , in the case of pseudo - decoupling . moreover ,
if the impurity spin is completely decoupled from other spins , it becomes a free spin .
we thus have the crossover from the @xmath0-site chain to the @xmath1-site chain with a free impurity spin . in the xxx case of the integrable impurity model
the wilson ratio has been evaluated numerically and systematically by applying the quantum transfer - matrix method @xcite , in which it is necessary to consider logarithmic corrections for the magnetization in low temperature due to the @xmath23 symmetry @xcite . in the xxz impurity model , however , it is not necessary to consider logarithmic corrections , and hence it is easier to derive the susceptibility at a very low temperature .
the contents of the present paper consist of the following . in section 2 ,
we introduce the spin-1/2 xxz chain with a spin-1/2 impurity and give an explicit expression of the hamiltonian in terms of the spin operators .
we show that the impurity spin operator becomes decoupled from the other spin operators if we send the absolute value of the impurity parameter to infinity . in section 3
we derive an analytic solution to the ground state under a weak magnetic field by the wiener - hopf method and present an analytic expression of the impurity susceptibility at zero temperature .
we then calculate the local magnetization on the impurity site systematically by the algebraic bethe ansatz .
we confirm the analytical result of the impurity magnetization with numerical plots .
furthermore , we evaluate the impurity susceptibility at zero temperature numerically by solving the bethe ansatz equations , and show that the numerical estimates are consistent with the analytic expression if we take the system size large enough .
in section [ spe ] we evaluate the impurity specific heat numerically by solving the truncated integral equations of the thermal bethe ansatz . in the plots of the impurity specific heat we observe that the impurity spin tends to become a free spin as the absolute value of the impurity parameter becomes very large , which is clearly seen in high temperature , while it couples strongly to other spins in low temperature ,
that is , the kondo effect appears .
we give an analytic expression of the impurity specific heat at a low temperature , which is valid if the absolute value of the impurity parameter is small with respect to the given low temperature .
we numerically determine the crossover temperature from the high- to low - temperature regimes as a function of the impurity parameter and give a good fitting formula to the data . in the last section
, we derive the universal value of the wilson ratio at low temperature and express it in terms of the dressed charge of the xxz spin chain @xcite .
we show that the estimate of the wilson ratio gives the universal value if the absolute value of the impurity parameter is small enough with respect to the temperature .
we now introduce one of the integrable xxz models with a spin-1/2 impurity through the algebraic bethe ansatz .
let @xmath24 be two - dimensional vector spaces over complex numbers @xmath25 .
here we denote by @xmath0 the number of sites on the xxz spin chain .
we call @xmath26 the auxiliary space and the tensor product @xmath27 the quantum space , where @xmath28 corresponds to the @xmath29th site of the xxz spin chain for @xmath30 . on the tensor product @xmath31
we define the @xmath32-matrix by @xmath33 } .\ ] ] here the suffix @xmath34 $ ] means that the matrix is defined for the tensor product @xmath31 .
we define the transfer matrix of the spin-1/2 xxz spin chain by the trace of the @xmath0th product of the @xmath32 matrices @xmath35 for @xmath30 , over the auxiliary space 0 : @xmath36 where @xmath37 ( @xmath38 are arbitrary .
we call them the inhomogeneity parameters .
the functions @xmath39 and @xmath40 are given by @xmath41 where @xmath42 and @xmath43 with @xmath44 in the gapless regime of the xxz spin chain .
we consider only the gapless regime throughout the paper , .
we now define the hamiltonian of the impurity model , @xmath45 .
we assume that only the 1st site has a different value for the inhomogeneity parameter @xmath46 : @xmath47 while the inhomogeneous parameters of the other sites @xmath37 for @xmath48 , have the same value @xmath49 . here
we assume that the impurity parameter @xmath11 is real .
we define the hamiltonian of the impurity model , @xmath45 , by the logarithmic derivative of the transfer matrix of the spin-1/2 xxz chain as follws .
@xmath50 by putting @xmath51 ( i.e. , @xmath52 ) , the hamiltonian @xmath45 reduces to the standard homogeneous spin-1/2 xxz hamiltonian , which has no impurity .
it is clear that the hamiltonian ( [ h ] ) is integrable .
it is a consequence of the construction in terms of the algebraic bethe ansatz .
let @xmath18 be the number of down spins .
the bethe - ansatz equations ( bae ) for the impurity hamiltonian ( [ h ] ) are given by @xmath53 where @xmath54 are called the bethe quantum numbers , and they are integers or half - integers according to the following rule : @xmath55 here the functions @xmath56 are given by @xmath57 \ , .\ ] ] we remark that in the special case , i.e. for the xxx model with @xmath58 , the baes ( [ bae ] ) are similar to the kondo model . by solving the bethe - ansatz equations ( [ bae ] )
, we can evaluate the energy of the bethe ansatz eigenvector with @xmath18 down spins .
let @xmath59 be a solution of the bethe ansatz equations ( [ bae ] ) .
the energy for the eigenstate with bethe roots @xmath60 is expressed in terms of the set of bethe roots @xmath61 as @xmath62 we remark that the energy depends on the impurity parameter @xmath11 , since the solutions @xmath63 contain the information of @xmath11 , though it does not appear explicitly .
we now present an explicit expression of the xxz hamiltonian with a spin-1/2 impurity , @xmath45 , defined by eq . in terms of the local spin operators .
the expression is useful for confirming analytic results . making use of the explicit expression of @xmath45 we can perform the exact diagonalization of the hamiltonian , and compare numerical results with analytic or numerical results which are obtained by solving the bethe ansatz equations . through a direct but straightforward calculation
we obtain the following compact expression of @xmath45 in terms of local spin operators @xmath64 and @xmath65 .
@xmath66 \nn \\ & + & c^+c^- \left [ \frac{\vp'(x)}2(s_n^+s_{1}^-+s_n^-s_{1}^+ ) + \delta\left(s_n^zs_{1}^z-\frac14\right)\right ] \nn \\ & + & c^+c^-\left [ \frac{\vp'(x)}2(s_1^+s_2 ^ -+s_1 ^ -s_2^+)+\delta s_1^zs_2^z\right ] \nn \\ & + & b^+b^-\left [ \frac{\d}2(s_n^+s_2 ^ -+s_n^-s_2^+)+\delta s_n^zs_2^z\right ] -\fr{\delta}{4 } \nn \\ & + & b^+c^-\biggl [ \vp'(x ) ( s_n^+s_2 ^ --s_n^-s_2^+)s_1^z \nn \\ & & \ \ \ \ \ \ \ \
-\d(s_n^+s_1 ^ --s_n^-s_1^+)s_2^z \nn \\ & & \ \ \ \ \ \ \ \
-\d(s_1^+s_2 ^ --s_1 ^ -s_2^+)s_n^z \biggr ] , \end{aligned}\ ] ] where the symbols @xmath67 and @xmath68 are given by @xmath69 here , the spin operators @xmath64 and @xmath70 are expressed in terms of the pauli matrices as follows .
@xmath71 some details of the derivation of are given in appendix [ apph ] .
we remark that the xxz hamiltonian with the spin-1/2 impurity , @xmath45 , given in eq .
is hermitian when the impurity parameter @xmath11 is real . in the cases of @xmath72
we have confirmed explicitly that the lowest energy level of the xxz hamiltonian with the spin-1/2 impurity is completely consistent with the eigenvalue evaluated through eq . from the ground - state solution of the bethe ansatz equations .
we therefore assume that the quantum numbers @xmath73 of the ground state do not change under the existence of the impurity .
they are given by @xmath74 we can confirm this assumption by directly diagonalizing the xxz hamiltonian with an impurity for a larger number of sites .
we should remark that the integrable model of an impurity given by eq .
corresponds to the spin-@xmath75 case of the integrable xxz model with the spin-1/2 impurity studied by schlottmann @xcite .
here we recall that the spin part of the kondo model @xcite is similar to the spin-1/2 xxx chain .
if one takes the limit @xmath76 , the hamiltonian reduces to @xmath77 \nn \\ & & + \frac\d2\(s^+_ns^-_2+s^-_ns^+_2\)+\d\(s_n^zs_2^z-\frac14\ ) \nn \\ & & + 2\sqrt{1-\d^2}\(s^x_ns^y_2-s^y_ns^x_2\)s_1^z .
\end{aligned}\ ] ] the impurity spin , i.e. , the spin operators defined on the 1st site such as @xmath78 and @xmath79 , appears only in the last term of the reduced hamiltonian in eq .
( [ hinf ] ) .
the exchange coupling of the impurity spin to the neighboring spins vanishes in @xmath80 . in terms of eq .
the bulk part @xmath16 and the impurity part @xmath81 corresponds to the first two lines and the third line of the right - hand side of , respectively . if one takes the xxx limit ( @xmath82 ) , the terms associated with the impurity site vanish .
the hamiltonian now reduces to the homogeneous spin-1/2 xxx hamiltonian defined on the @xmath83 sites from site 2 to site @xmath0 : @xmath84.\end{aligned}\ ] ] under a magnetic field , @xmath85 , the number of down spins , @xmath18 , becomes smaller than that of the half - filling case , i.e. , @xmath19 .
let us assume that when the impurity spin is pseudo - decoupled from other spins , we can effectively define the energy of the bulk part @xmath16 for the eigenstates with @xmath18 down spins in the bulk chain and denote it by @xmath86 .
we then suggest that any down spin on the impurity site will be attracted to the bulk chain since the bulk energy decreases as the number of down spins in the bulk increases : @xmath87 due to the anti - ferromagnetic interactions . here
the total energy reduction including the zeeman term is of the order of @xmath88 .
we also remark that the magnetic energy term , @xmath89 , in the total hamiltonian @xmath90 does not change whether the impurity spin has an up spin or down spin in the sector of @xmath18 down spins .
we thus suggest that the impurity spin tends to be given by an up spin and all the @xmath18 down - spins on the whole chain tend to be located on the @xmath1-site chain for any number @xmath18 of down spins if pseudo - decoupling occurs under a magnetic field .
furthermore , it is consistent with another viewpoint that the impurity spin becomes a free spin if it is decoupled with other spins completely .
we therefore have the crossover from the @xmath0-site chain to the @xmath1-site chain with the free impurity spin as the exchange coupling between the impurity spin and other spins decreases to zero . in the graphs of the impurity specific heat in 4 .
we show how the impurity spin becomes free from the other spins in the case of large absolute values of @xmath11
we shall calculate the magnetic susceptibility by adding a small magnetic field @xmath17 to the hamiltonian defined by : @xmath91 we recall that the hamiltonian @xmath45 is expressed explicitly in terms of spin operators in .
let us assume that the bethe quantum numbers @xmath73 of the ground state do not change under the existence of the impurity . by taking the thermodynamic limit of the bae through the euler - maclaurin formula , we derive the following inregral equation : @xmath92 here parameter @xmath93 denotes the fermi point determined by the magnetic field @xmath17 , and @xmath94 the density of the bethe roots in the ground - state solution of the bae
( [ bae ] ) , where functions @xmath95 for @xmath96 are defined by @xmath97 we remark that we have the infinite fermi point , @xmath98 , for @xmath99 .
we define the fourier transform of a given function @xmath100 by @xmath101 let us denote by @xmath102 the root density for the ground - state solution of the bae with no magnetic field for .
we now derive the analytic expression of @xmath102 . by taking the fourier transform of the following integral equation @xmath103 we have @xmath104 here @xmath105 and @xmath106
are given by @xmath107 taking the inverse transform of , one finds @xmath108 in the case of @xmath51 , it reduces to the bulk density , @xmath109 , which is given by @xmath110 , width=264 ] hereafter we shall sometimes abbreviate @xmath111 and @xmath102 as @xmath112 and @xmath113 , respectively .
let us evaluate the largest value of the impurity parameter @xmath11 for which thermodynamic expressions are valid .
we shall compare several numerical data obtained for the system of a finite number of @xmath0 with such analytic expressions that are derived in the thermodynamic limit .
we approximately estimate it by setting @xmath114 .
if the root density is smaller than @xmath88 , then the continuous approximation to the discrete distribution of the bethe roots should be not good . from fig .
[ frho ] we have @xmath115 for @xmath116 .
let us introduce a function @xmath117 by @xmath118 or equivalently by @xmath119 multiplying @xmath120 , the integral equation and are merged into the following : @xmath121 in the wiener - hopf method the integral equation plays a fundamental role .
we denote by @xmath122 the magnetization per site under magnetic field @xmath17
. we shall give the relation between the fermi point @xmath93 and the magnetic field @xmath17 by the wiener - hopf method in 3.2 .
let us recall that the energy eigenvalue of the hamiltonian @xmath45 is given by @xmath123 we consider the difference of the ground - state energy per site under magnetic field @xmath17 from that of no magnetic field , and we denote it by @xmath124 .
then , physical quantities such as @xmath122 and @xmath124 are written in terms of the inetgrals of the root density @xmath112 as follows .
@xmath125 hereafter , we shall often abbreviate the magnetization per site and the energy per site simply as the magnetization and the energy , respectively .
let us define function @xmath126 by @xmath127 by making use of the invariance of eq . under the replacement of @xmath22 and @xmath11 by @xmath128 and @xmath129 ,
respectively , we show the following relation : @xmath130 we define functions @xmath131 by @xmath132 here @xmath133 is the heaviside step function : @xmath134 for @xmath135 and @xmath136 for @xmath137 . in terms of functions
@xmath138 the integral equation is now expressed as follows .
@xmath139 let us now assume that the fermi point @xmath93 is very large .
the magnetization and the energy are expressed in terms of functions @xmath131 as follows .
@xmath140 here we recall that eq .
( [ eq : energy - yp ] ) is valid only for large values of @xmath93 : @xmath141 .
it is known that integral equations such as eqs . and can be solved by the wiener - hopf method ( see @xcite , @xcite , @xcite ) . in the following analysis
we neglect the terms with @xmath142 in eqs . since the corrections from them are negligible for @xmath143 . expressing
@xmath144 for @xmath145 , we thus have @xmath146 where functions @xmath147 are given by @xmath148 with @xmath149 , and by the following relation : @xmath150 then we have @xmath151 however , the fermi point @xmath93 remains unknown , yet .
it is determined as a function of @xmath17 by the following condition : @xmath152 we therefore have @xmath153 , width=264 ] the relation between the fermi point @xmath93 and the magnetic field @xmath17 is shown explicitly in fig .
[ fb ] . the fermi point @xmath93 plays the role of the cut - off in the integral of eq . .
the cut - off @xmath93 goes to infinity in the limit of @xmath154 .
however , it dramatically decreases to values such as 2 or 3 , which are not very large , when @xmath17 becomes nonzero such as @xmath155 . by comparing figure [ fb ] with the profile of the root density @xmath113 shown in fig .
[ frho ] , we suggest that the wiener - hopf method is valid only when the impurity parameter @xmath11 is small enough such as @xmath156 so that the small peak at @xmath157 in the root density @xmath113 is located in the interval between the fermi point @xmath158 and the origin @xmath159 .
recall that the symbol @xmath122 denotes the magnetization per site in the xxz spin chain which has the spin-1/2 impurity with impurity parameter @xmath11 . through the wiener - hopf method
the magnetization per site @xmath122 is given by @xmath160 here , the magnetic field @xmath17 is very small , so that only linear terms of @xmath17 may appear .
let us define the magnetic susceptibility per site , @xmath161 , by @xmath162 consequently , we have @xmath163 we now extract the impurity susceptibility @xmath3 induced by the spin-1/2 impurity of parameter @xmath11 .
we first define the impurity magnetization @xmath164 by @xmath165 here , @xmath166 gives the total magnetization of the xxz spin chain with the impurity parameter @xmath11 .
we also express it by @xmath167 .
we define the impurity susceptibility @xmath168 by @xmath169 the impurity susceptibility is therefore given by @xmath170 let us now derive the local magnetization at the impurity site , @xmath171 , under a small magnetic field @xmath17 at zero temperature .
we take the expectation value of the local operator @xmath172 for the ground state under the magnetic field @xmath17 . here
we remark that if the magnetic field is zero , the local magnetization at the impurity site vanishes .
we now define the monodromy matrix of the spin-1/2 xxz spin chain by the @xmath0th product of the @xmath32 matrices @xmath35 for @xmath30 : @xmath173 where @xmath37 ( @xmath38 are the inhomogeneity parameters given by @xmath174 and @xmath175 .
we define the operator - valued matrix elements of the monodromy matrix by @xmath176 let us denote the ground state under a small magnetic field @xmath17 by @xmath177 .
we express the ground - state solution of the bethe - ansatz equations with @xmath18 down spins under a magnetic field @xmath17 as @xmath178 for @xmath179 .
we have @xmath180 where @xmath181 denotes the vacuum state , which has no down spin .
the hermitian conjugate vector is given by @xmath182 we therefore define the impurity local magnetization @xmath183 by @xmath184 we now evaluate it through the expectation value of the local operator @xmath185 as follows .
@xmath186 where @xmath187 denote the two - by - two matrices with only nonzero matrix element 1 at the entry of @xmath188 for @xmath189 .
hereafter , we call @xmath190 the local magnetization at the impurity site , not @xmath191 , for simplicity .
we introduce the gaudin matrix @xmath192 with real parameters @xmath193 defined by @xmath194 for @xmath195 , as follows .
@xmath196 where function @xmath197 is defined by @xmath198 putting the ground - state solution under a magnetic field , @xmath199 , with @xmath200 and @xmath201 , we have @xmath202 here we recall that @xmath203 denotes the density of the bethe roots for the ground - state solution under magnetic field @xmath17 satisfying ( [ baeh ] ) .
making use of the formula of the quantum inverse scattering problem @xcite @xmath204 and slavnov s scalar product formula we show @xmath205 here @xmath206 denotes the solution of the integral equation @xmath207 and the matrix elements of the @xmath208 matrix @xmath209 are given by @xmath210 let us consider the case when the absolute value of the impurity parameter @xmath11 is small enough with respect to a given value of magnetic field @xmath17 .
we assume that it is much smaller than the fermi point @xmath93 : @xmath211 , where @xmath93 is related to the magnetic field @xmath17 by ( [ eq : h - b ] ) .
in this case we solve eq . by the wiener - hopf method .
we define @xmath212 by the solution of the integral equation with @xmath98 . explicitly , we have @xmath213 .
we define function @xmath214 by @xmath215 since the equation does not change by replacing @xmath22 and @xmath11 by @xmath128 and @xmath129 , respectively , we have @xmath216 .
we now define functions @xmath217 by @xmath218 by solving the integral equation @xmath219 we evaluate @xmath220 through the following : @xmath221 we thus obtain the analytic expression of the impurity local magnetization @xmath222 here we recall that the magnetic field @xmath17 is related to the fermi point @xmath93 by ( [ eq : h - b ] ) .
the analytic expression of the impurity local magnetization is consistent with the impurity susceptibility ( [ eq : sc - imp ] ) .
therefore , it follows that the identification of the impurity contribution to the magnetic susceptibility in terms of the bethe ansatz equations is valid . here
we remark that expression is valid only when the absolute value of impurity parameter @xmath11 is small enough such as @xmath223 so that we can apply the wiener - hopf method to solve the integral equation .
let us now consider the case when the absolute value of the impurity parameter @xmath11 is large .
for instance , we assume that it is larger than the logarithm of the system size @xmath0 : @xmath224 .
it follows that we have @xmath225 , and the energy gap due to the exchange interaction between the impurity site and the neighboring sites is smaller than the bulk excitation energy gaps , which are of the order of @xmath88 , and hence can be neglected with respect to them .
if the absolute value of the impurity parameter @xmath11 is very large such as @xmath226 and @xmath227 , we can show that the impurity local magnetization , @xmath228 , approaches the largest value 1 exponentially with respect to the impurity parameter @xmath11 .
the right hand side of the linear integral equation , which gives the source term to it , is proportional to @xmath229 if impurity parameter @xmath11 is very large such as @xmath227 .
therefore , we have @xmath230 for large values of @xmath11 , which leads to the following : @xmath231 we now show that the impurity local magnetization becomes easily saturated even with an infinitesimally small magnetic field , if pseudo - decoupling occurs .
in fact , even if the absolute value of the impurity parameter @xmath11 is very large such as @xmath226 , we can take a very small magnetic field @xmath17 such that the value of the fermi point @xmath93 is much larger than @xmath12 .
then , we can apply the wiener - hopf method to the integral equation , and we obtain the expression of the impurity local magnetization for a very small magnetic field @xmath17 with a large value of @xmath11 .
let us estimate the small magnetic field @xmath17 for which the expression is valid in the case of pseudo - decoupling .
if we determine the magnetic field by @xmath232 in the @xmath0-site chain with @xmath233 down spins , then @xmath234 is given by of the order of @xmath88 .
however , the magnetic field is much smaller than @xmath235 if pseudo - decoupling occurs .
let us now assume the following conditions : @xmath236 where @xmath237 .
for instance , @xmath238 for @xmath239 .
we suggest that the magnetic field @xmath17 which triggers the impurity local magnetization is given by @xmath240 , which is very small .
we thus expect that if we add a small magnetic field @xmath17 such as @xmath241 , then the local magnetization at the impurity site readily approaches the largest value 1 .
the result is consistent with another viewpoint that the impurity spin becomes a free spin if the exchange coupling with other spins vanishes .
it is clear that a free spin becomes saturated with an infinitesimally small magnetic field .
width=302 ] let us now confirm the analytic results of the impurity local magnetization by numerical data .
we plot the numerical estimates of the local magnetization at the impurity site , @xmath242 , against the impurity parameter @xmath11 in the case of @xmath243 in fig .
[ fig - imp ] .
the estimates are evaluated by numerically diagonalizing the hamiltonian @xmath45 with @xmath243 . for small values of the impurity parameter @xmath11 ,
the impurity local magnetization , @xmath244 , is proportional to the magnetic field @xmath17 , and it increases with respect to @xmath11 similarly as @xmath245 .
however , for large values of the impurity parameter @xmath11 , it approaches the largest value 1.0 .
moreover , under an infinitesimally small magnetic field @xmath17 , the impurity spin takes very quickly a positive large value less than 1.0 .
we also observe that it vanishes under zero magnetic field for any values of the impurity parameter @xmath11 .
we now evaluate numerically the magnetic susceptibility by calculating the energy eigenvalues through solutions of the bethe ansatz equations . for a given number @xmath18 satisfying @xmath246
we solve the bae for the ground state with @xmath18 down spins , and evaluate the ground state energy @xmath247 through the solution by making use of eq . .
we also evaluate the energy @xmath248 for the ground state with @xmath233 down spins .
then , we determine the magnetic field @xmath234 by the following relation : @xmath249 here we recall that the total magnetization @xmath250 is given by @xmath251 in the case of @xmath18 down spins .
we define the total susceptibility @xmath252 under magnetic field @xmath234 by @xmath253 we can evaluate the susceptibility at @xmath99 , i.e. @xmath252 for @xmath254 , by taking the difference through eq .
( [ eq : chim ] ) .
it is close to the definition of the susceptibility by the derivative of @xmath255 with respect to magnetic field @xmath17 .
the susceptibility per site @xmath161 is given by @xmath256 .
we then evaluate the impurity susceptibility @xmath168 from the total susceptibility @xmath257 by the following relation : @xmath258 . at @xmath99 and then the impurity susceptibility @xmath168 at @xmath99 as follows .
we make a graph of magnetization @xmath259 versus @xmath234 , and interpolate the data points of @xmath260 versus @xmath234 as a polynomial of @xmath17 of some degree .
we employ the ratio @xmath261 at @xmath254 as the susceptibility @xmath257 at @xmath99 assuming that @xmath262 and @xmath263 for @xmath254 . here
we remark that the total susceptibility @xmath257 is @xmath264 , i.e. of the order of the system size @xmath0 .
we repeat the evaluation of @xmath257 for different values of the system size @xmath0 , and plot the total susceptibility divided by @xmath0 , @xmath265 , against the inverse of the system size , @xmath88 .
then , we interpolate the data points of @xmath265 versus @xmath88 as a polynomial of @xmath266 of some degree and employ the coefficient of the @xmath267 term as the shift @xmath268 . by adding the susceptibility per site with no impurity , @xmath269 , to it
, we obtain @xmath168 . ] , width=226 ] , width=226 ] the numerical estimates of the impurity susceptibility @xmath168 at @xmath99 are plotted against impurity parameter @xmath11 in figs .
[ fchi2 ] and [ fchi1 ] .
the impurity susceptibility increases with respect to the absolute value of impurity parameter @xmath11 , which is due to the kondo effect .
the numerical estimates ( blue dots ) are consistent with the analytic formula ( red line ) in figs .
[ fchi2 ] and [ fchi1 ] . here , the typical system size @xmath0 for the data of figs .
[ fchi2 ] and [ fchi1 ] is given by @xmath270 and @xmath271 , respectively .
we suggest that if the system size @xmath0 is large enough , then the numerical estimates of the impurity susceptibility @xmath168 at @xmath99 are consistent with the analytic formula .
for instance , the numerical estimates of @xmath168 for @xmath272 shown in fig .
[ fchi2 ] only from @xmath273 to 1 are smaller than the corresponding values of the analytic formula if the absolute value of impurity parameter @xmath11 is larger than 2 : @xmath274 , while those of @xmath275 are rather close to the analytic curve even for @xmath276 as shown in fig .
[ fchi1 ] . by applying the wiener - hopf method formulated in ref .
@xcite we can show that the @xmath11-dependence of the susceptibility per site @xmath161 changes if we add higher - order terms with respect to the expansion parameter @xmath277 .
however , it seems that the same @xmath11-dependence as in eq . is reproduced for the impurity susceptibility at zero magnetic field since higher - order terms as in ref .
@xcite do not remain when @xmath17 is small . in ref.@xcite
, it is stated that the impurity susceptibility @xmath3 should depend on the parameter @xmath11 such as @xmath278 in terms of the notation of the present paper , which is different from @xmath279 . however , the exponential dependence is not exact as an analytic solution in the wiener - hopf method .
furthermore , it is not consistent with the numerical estimates of the impurity susceptibility @xmath3 even for small @xmath11 ( see also fig . [ fchi2 ] ) .
we now evaluate the specific heat by solving the truncated integral equations of the thermal bethe ansatz @xcite .
we show explicitly how the specific heat depends on the impurity parameter @xmath11 .
we restrict the anisotropy parameter @xmath280 as @xmath281 throughout this section .
let us consider the string solutions for the bae at @xmath281 .
it is known that there are only three different types of solutions @xmath282 for @xmath281 .
explicitly , we have @xmath283 where each @xmath284 gives the real part or the center of the string solution @xmath63 .
we remark that @xmath285 . in terms of string solutions
the bae is now expressed as follows .
@xmath286 here , unknown functions @xmath287 and @xmath288 denote the particle and hole densities , respectively , with respect to the real part @xmath284 of the string solutions @xmath63 , and the symbol @xmath289 denotes the convolution of given functions @xmath290 and @xmath291 .
the functions @xmath292 for @xmath293 are given by @xmath294 where @xmath295 are given by @xmath296 the matrix @xmath5 is defined by @xmath297 here we remark that @xmath298 and @xmath299 are the same functions , as defined by . by making use of the identity @xmath300 we reduce the number of equations of from 3 to 2 .
let us now express physical quantities in terms of @xmath287 and @xmath288 .
the energy of this system is described by @xmath301 where @xmath302 is given by @xmath303 and the entropy is expressed as @xmath304 + \rho_j^h(z)\log \[1+\fr{\rho_j(z)}{\rho_j^h(z)}\ ] \rt dz.\end{aligned}\ ] ] let us take the free energy @xmath305 ( @xmath5 : temperature ) with respect to @xmath306 under the conditions .
then , it is clear that the thermal equilibrium condition @xmath307 is equivalent to @xmath308 where functions @xmath309 are defined by @xmath310 the impurity parameter @xmath11 does not remain after by taking the variation with respect to @xmath306 and @xmath311 under the constraint of bae , and hence the above equations are the same as those of the homogeneous spin-1/2 xxz spin chain . by putting them back to the free energy ,
its final form is given by the following : @xmath312 equations are solved numerically @xcite . as we said before ,
@xmath313 and @xmath314 indicate the same equation .
only two of the three eta - functions are independent and the equations can be reformulated as follows
. @xmath315 \ ] \\
[ 1+\exp \[-\fr{\k}{t}b_2-b_1*h_1 - 2b_2*h_2 \ ] \ ] \end{array } \right.,\end{aligned}\ ] ] where @xmath316 and the fourier transforms of @xmath317 are given by @xmath318
explicitly we have @xmath319 here , the expressions of functions @xmath317 are specific to @xmath281 . in terms of these functions ,
the free energy is written by @xmath320 the specific heat is calculated by the same approach .
let us set the following quantities : @xmath321 then , the specific heat ( @xmath322 , @xmath323 ) becomes @xmath324 during the procedure , we also have its energy @xmath325 the equations for @xmath326 are given by @xmath327 \\
u_2=\(1-e^{-h_2}\ ) \[-\k b_2-b_1*u_1 - 2b_2*u_2 \ ] \end{array } \right . ,
\end{aligned}\ ] ] and those for @xmath328 by @xmath329 \\ v_2=\fr{1}{t^2}\fr{{u_2}^2}{e^{h_2}-1 } + \(1-e^{-h_2}\ ) \[-b_1*v_1 - 2b_2*v_2 \ ] .
\end{array } \right.\end{aligned}\ ] ] let us consider the impurity specific heat , @xmath330 , i.e. , the specific heat on the impurity site with parameter @xmath11 at temperature @xmath5 .
we define it by @xmath331 here the specific heat per site is denoted by @xmath332 .
we remark that the specific heat of the total system @xmath333 is given by @xmath334 : @xmath335 .
if temperature @xmath5 is very low and the absolute value of impurity parameter @xmath11 is small enough with respect to the temperature , we suggest the analytical expression of @xmath330 from that of the impurity magnetic susceptibility @xmath336 as @xmath337 , width=226 ] , width=226 ] in fig .
[ fcfit ] the impurity specific heat @xmath338 at low temperature @xmath339 is plotted against impurity parameter @xmath11 . here , the numerical estimates of the impurity specific heat @xmath338 in the interval from @xmath340 to @xmath341 near the origin of @xmath51 are fitted completely well as a function of @xmath11 to the analytic expression .
there is no fitting parameter .
the impurity specific heat reduces to the low - temperature result of the homogeneous xxz spin chain ( see e.g. @xcite ) , when we put @xmath51 .
we suggest that the expression of the impurity specific heat at low @xmath5 should be valid for any value of @xmath280 in the massless regime , although we evaluated it only for @xmath281 .
the impurity specific heat @xmath330 is plotted against parameter @xmath11 in a wider range of @xmath11 in fig .
[ fc1 ] . at @xmath51
the graph is convex and almost flat .
it increases with respect to @xmath12 , and reaches two peaks at @xmath342 .
we observe in fig .
[ fc1 ] that the impurity spin becomes close to a free spin at @xmath343 for large values of @xmath12 such as @xmath344 : the impurity specific heat @xmath338 approaches zero as @xmath12 increases .
here we remark that the specific heat of a free spin vanishes under zero magnetic field at any temperature .
we suggest that the peak positions @xmath345 in the graph of @xmath338 against parameter @xmath11 is approximately given by the solution of the following equation @xmath346 for @xmath343 we have @xmath347 with @xmath239 . for @xmath348
, @xmath338 has two peaks with @xmath349 , while the solution of eq .
is given by @xmath350 .
, width=226 ] , width=226 ] we shall show that the impurity spin has two aspects in finite temperatures : the kondo effect in low temperature and the tendency to become a free spin through pseudo - decoupling in high temperature .
similarly as the impurity specific heat @xmath330 , we define the impurity entropy @xmath351 by @xmath352 here , the entropy of the total system , @xmath353 , is given by @xmath354 .
we also remark that we evaluate the entropy by @xmath355 from the energy @xmath356 and the free energy @xmath290 .
the numerical estimates of the impurity specific heat @xmath338 and the impurity entropy @xmath351 are plotted in figs .
[ fcimp ] and [ fsimp ] , respectively , against temperature @xmath5 .
we observe the following : ( i ) the impurity specific heat @xmath330 has a peak approximately at @xmath357 in fig.[fcimp ] for small absolute values of parameter @xmath11 such as @xmath358 ; ( ii ) the impurity specific heat @xmath338 grows linearly with respect to the temperature @xmath5 near zero temperature , and the gradient of the linear specific heat becomes larger as the impurity parameter @xmath11 increases ; ( iii ) the peak position of the impurity specific heat @xmath338 shifts to lower temperatures as the impurity parameter @xmath11 increases ; ( iv ) the peak value of the impurity specific heat @xmath330 decreases as the peak position shifts to lower temperatures .
we suggest that the kondo effect leads to the peaks in the graphs of the impurity specific heat @xmath330 versus temperature @xmath5 in fig .
[ fcimp ] .
we recall that the bethe ansatz equations in the xxx limit with @xmath58 are similar to those of the kondo model @xcite .
furthermore , it seems that the increase of the gradient of the impurity specific heat @xmath338 at low temperature is consistent with the analytic expression . in high temperatures such as @xmath359 , the impurity entropy @xmath360
approaches the value @xmath361 for any value of @xmath11 , as shown in fig .
[ fsimp ] .
it is standard that the impurity spin becomes independent of the other spins in the chain at high temperature .
let us now determine the crossover temperature between the two regimes : the low - temperature regime where the kondo effect is dominant and the high - temperature regime where the impurity spin tends to become a free spin through pseudo - decoupling .
, width=302 ] by the difference : @xmath362 , we define the shift of the specific heat per site with impurity parameter @xmath11 from that of the homogeneous chain at temperature @xmath5 . here
we have from eq . the following expression in low temperature : @xmath363 the numerical estimates of the specific heat shift @xmath364 are plotted against temperature @xmath5 in fig . [ ftc ] .
each dotted curve shows the graph of the data of @xmath364 with a fixed value of impurity parameter @xmath11 .
, width=264 ] , width=226 ] we define the crossover temperature by the temperature at which the specific heat shift @xmath364 vanishes .
we plot the estimates of the crossover temperature @xmath365 as a function of the impurity parameter @xmath11 in fig .
furthermore , we find that the following formula gives a good fitting curve to the numerical estimates of the crossover temperature : @xmath366 where the best estimates of parameters @xmath367 and @xmath368 are given by @xmath369 and @xmath370 , respectively . it is suggested from the impurity susceptibility ( [ eq : sc - imp ] ) that the kondo temperature @xmath371 is given by @xmath372 .
however , the numerical estimate of parameter @xmath368 of crossover temperature @xmath365 in ( [ eq : tx ] ) is given by almost half the value of @xmath373 .
thus , the functional form of the crossover temperature @xmath365 is the same with the kondo temperature @xmath371 , while the estimate of the parameter @xmath368 is different .
let us recall that we have defined the wilson ratio by eq .
, which is given by the ratio of the impurity magnetic susceptibility @xmath168 to the impurity specific heat @xmath338 with some normalization factor .
we also recall that the impurity susceptibility is given by eq . as @xmath374
we have shown in 4 that when the absolute value of the impurity parameter @xmath11 is small , the impurity specific heat @xmath330 is given by eq .
for low temperature .
thus , by making use of and , we evaluate the wilson ratio in the xxz impurity model for small @xmath11 at low temperature @xmath5 by @xmath375 we observe that the wilson ratio @xmath376 does not depend on the impurity parameter @xmath11 .
therefore , the hamiltonians of the xxz impurity model with the same value of @xmath280 but different values of @xmath11 are classified in the same universality class .
furthermore , the wilson ratio is expressed in terms of the dressed charge @xmath377 as @xmath378 here the dressed charge @xmath377 denotes the dressed charge function @xmath379 evaluated at the fermi point : @xmath380 . the dressed charge @xmath381 is defined by the solution of the integral equation @xcite : @xmath382 it is shown in ref .
@xcite that by applying the wiener - hopf method to eq .
for very small magnetic field @xmath17 we have @xmath383 therefore , we obtain relation .
the wilson ratio of the xxz impurity model is expressed in terms of the dressed charge of the xxz spin chain . from the cft viewpoint
we give a physical argument for the wilson ratio expressed with the dressed charge at low temperature for small @xmath12 in appendix b. the numerical estimates of the wilson ratio are plotted in fig .
[ fr ] against impurity parameter @xmath11 .
here we have evaluated the wilson ratio numerically by making use of the estimates of the impurity susceptibility @xmath168 of fig .
[ fchi1 ] in 3 and those of the impurity specific heat @xmath338 at low temperature with @xmath343 of fig .
[ fc1 ] in 4 . for small absolute values of impurity parameter @xmath11 such as @xmath156
the estimate of the wilson ratio is almost constant with respect to @xmath11 and given by 3.0 .
we suggest that the low - temperature impurity effect with small @xmath12 is due to the kondo effect and the estimate of the wilson ratio is consistent with eq . .
thus , the universality class is described by the xxz coupling through the dressed charge as shown in eq . . for large absolute values of impurity parameter @xmath11 such as @xmath384 ,
however , the estimate of the wilson ratio shown in fig . [ fr ] increases with respect to @xmath12 .
we suggest that it reflects the crossover of the impurity spin becoming a free spin , for which the wilson ratio is given by infinity .
however , we also suggest that if the numerical estimate of the wilson ratio larger than 3.0 such as for @xmath385 is not universal and may depend on the temperature @xmath5 at which the impurity specific heat @xmath338 is evaluated .
we expect that the region of @xmath11 where the estimate of the wilson ratio is given by @xmath386 becomes wider if the temperature @xmath5 decreases . ,
the authors would like to thank f.h.l .
essler and k. sakai for helpful comments . in particular ,
we thank k. sakai for useful comments on the wiener - hopf method and the specific heat of the xxz spin chain .
we also thank s. okuda for helpful discussion .
we would also like to thank a. klmper for helpful comments during the workshop raqis 14 , dijon , france .
the present study is partially supported by grant - in - aid for scientific research no .
we now derive expression of the xxz hamiltonian with a spin-1/2 impurity which is defined by the logarithmic derivative of the transfer matrix as shown in eq . .
here we express the hamiltonian explicitly in terms of local spin operators @xmath64 and @xmath70 .
first , we remark that the @xmath32-matrix at zero rapidity @xmath387 gives the permutation operator @xmath388 acting on the tensor product space @xmath389 : @xmath390 .
we consider the two factors of the logarithmic derivative , separately .
@xmath391^{-1 } \nn \\ & & \times \left [ \dlam\tau_{1\cdots n}\left(\lam\big|\xi_1,\frac{i\zeta}{2},\cdots,\frac{i\zeta}{2}\right)\big|_{\lam\to\frac{i\zeta}{2}}\right ] \ , .
\label{eq : logdv } \end{aligned}\ ] ] for the first factor of ( [ eq : logdv ] ) we have @xmath392 where @xmath393 denotes the cyclic permutation operator , which is also expressed as the product of permutation operators : @xmath394 . for the second factor of ( [ eq : logdv ] )
we have @xmath395 \nn \\ & = & \sum_{n=2}^n \tr0 [ p_{0n}\cdots h_{0n } \cdots p_{02}r_{01}(x)]+\tr0 [ p_{0n}\cdots p_{02}h_{01}(x ) ] , \end{aligned}\ ] ] where symbols @xmath396 denote the derivatives of the @xmath32-matrices with respect to rapidity @xmath397
@xmath398 for eq .
we consider three cases in the following : @xmath399 for @xmath400 we have @xmath401 = \pi^{2\cdots n-1 } r_{n-1,1}(x ) \pi^{n-1 , n+1 , \cdots n } h_{n-1,n}. \label{eq : hn}\end{aligned}\ ] ] applying the inverse of to , we have @xmath402 here , we used the abbreviation @xmath403 .
@xmath399 for @xmath404 we have @xmath405 = \pi^{3\cdots n } h_{n,2}r_{n,1}(x ) .
\end{aligned}\ ] ] then , by applying the inverse of , we have @xmath406 @xmath399 for @xmath407 we have @xmath408 = \pi^{2\cdots n } h_{n,1}(x),\end{aligned}\ ] ] then @xmath409 taking the sum of these terms , we have @xmath410 we now express expression in the form of matrices .
the third term of is exactly the same matrix which appears in the xxz hamiltonian .
actually , we have @xmath411 } \nn \\ & = & \sm{n=3}n \ [ \frac12(s_n^+s_{n+1}^-+s_n^-s_{n+1}^+)+\d(s_n^zs_{n+1}^z-\fr14)\].\end{aligned}\ ] ] the first term of is given by @xmath412 } \nn \\ & = & c^+c^- \ [ \frac{\vp'(x)}2(s_n^+s_{1}^-+s_n^-s_{1}^+)+\d\(s_n^zs_{1}^z-\frac14\)\].\end{aligned}\ ] ] here we recall the notation given in eq . .
the second term of is given by @xmath413 } \nn \\ & & \times \begin{pmatrix } -\d(\fr12-s_2^z)&s_2 ^ -\\
s_2^+&-\d(\fr12+s_2^z)\\ \end{pmatrix}_{\!\![1 ] } \nn \\ & & \times \begin{pmatrix } ( \fr12+s_n^z)+c^+(\fr12-s_n^z)&b^+s_n^-\\ b^+s_n^+&c^+(\fr12+s_n^z)+(\fr12-s_n^z)\\ \end{pmatrix}_{\!\![1]}.\end{aligned}\ ] ] by a straightforward calculation , we derive the rest of .
let us explain the wilson ratio expressed in terms of the dressed charge through a physical argument based on the possible cft .
we assume that the low - lying excitation spectrum of the xxz spin chain with the spin-1/2 impurity is described by a cft .
we first derive the ratio of the bulk susceptibility @xmath414 to the bulk specific heat @xmath415 .
the magnetic susceptibility @xmath416 of the homogeneous xxz spin chain with no impurity is given by @xcite @xmath417 where @xmath418 is the group velocity of spin excitation at the fermi point : @xmath419 with @xmath420 being the dressed energy .
for the homogeneous xxz spin chain with no impurity the specific heat @xmath415 at low temperature is given by @xmath421 for the bulk susceptibility and the bulk specific heat we thus have @xmath422 we now extend the calculation for the homogeneous chain@xcite to the xxz chain with the spin-1/2 impurity .
we thus have @xmath423 where we have assumed the contribution of the impurity in the fermi velocity , @xmath424 .
it follows from the assumption that the low - lying excitation spectrum of the xxz spin chain with the spin-1/2 impurity is given by the cft of central charge @xmath425 that the specific heat is given by @xmath426 thus , the wilson ratio for the impurity susceptibility @xmath427 and the impurity specific heat @xmath428 is given by @xmath429 we have thus obtained the wilson ratio given by eq . .
for integrable impurity models such as the anderson and the kondo models the boundary cft is applied to characterize the impurity susceptibilities of the models @xcite .
however , it seems that it is nontrivial to show that some conformal field theory describes the low - lying excitation spectrum of the xxz spin chain with the spin-1/2 impurity .
we remark that the above argument for the wilson ratio in eq . does not depend on the ratio of the impurity group velocity to the bulk one @xmath430 .
however , it is nontrivial to evaluate the impurity contributions @xmath427 and @xmath428 so that they are consistent with numerical estimates , as we have shown in 3 and 4 of the present paper . | we study the zero- and the finite - temperature behavior of the integrable spin-1/2 xxz periodic chain with an impurity by the algebraic and thermal bethe ansatz methods .
we evaluate the impurity local magnetization at zero temperature analytically and derive the impurity susceptibility exactly from it . in the graphs of the impurity specific heat versus temperature ,
we show how the impurity spin becomes more liberated from the bulk many - body effect as the exchange coupling between the impurity spin and other spins decreases , and also that in low temperature it couples strongly to them such as the kondo effect .
thus , we observe not only the crossover behavior from the high- to the low - temperature regime but also another one from the @xmath0-site chain to the @xmath1-site chain with a free impurity spin .
we also show that the estimate of the wilson ratio at a given low temperature is independent of the impurity parameter if its absolute value is small enough with respect to the temperature , and the universality class is described by the xxz anisotropy in terms of the dressed charge .
department of physics , graduate school of humanities and sciences , + ochanomizu university + 2 - 1 - 1 ohtsuka , bunkyo - ku , tokyo 112 - 8610 , japan @xmath2 research center for advanced science and technology , university of tokyo , + 4 - 6 - 1 komaba , meguro - ku , tokyo 153 - 8904 , japan |
the pendular cavity , or fabry - perot cavity in which one of the mirrors is free to oscillate , has previously been investigated by a number of researchers , both experimentally @xcite , and theoretically @xcite .
closely related schemes have been theoretically proposed to entangle mirrors @xcite and create superposition states of a mirror @xcite .
common to almost all the theoretical treatments is a linearisation of quantum langevin equations around their classical steady - state solutions @xcite .
this then allows for the relatively simple calculation of spectral quantities which may be measured outside the cavity .
results obtained in this way have been used to analyse the sensitivity of gravity wave interferometers @xcite , predict the suppression of quantum noise @xcite , propose quantum nondemolition measurements of photon number @xcite , analyse the quantum limits to measurements with an atomic force microscope @xcite , analyse the quantum noise in position measurements of the oscillating mirror @xcite , calculate the phase noise in the cavity field @xcite , predict the entanglement of macroscopic oscillators via radiation pressure @xcite , and propose the quantum locking of interferometer mirrors @xcite . using the state - vector approach so common in quantum computing theory , it has been proposed that quantum superpositions , entanglement and near - number states of the cavity field , along with superposition states of the mirror , can be produced with this system @xcite . using a similar state - vector approach
, it has been proposed that quantum superpositions of a mirror may be created by the interaction with a single photon @xcite .
it is well - known that the linearised fluctuation analysis used in the majority of the theoretical papers cited above is limited in its applicability .
it has been shown , for example , that the mean - field equations derived in this way can give misleading results for travelling - wave second harmonic generation @xcite and for the intracavity interaction between light and condensed atoms @xcite as well as in raman photoassociation of atomic bose - einstein condensates @xcite .
the spectra calculated via this method are also known to not be accurate near any critical points of the system , as has been shown with the optical parametric oscillator @xcite .
there are two conditions which must be fulfilled for a linearised analysis to be trustworthy .
the first has to do with the sign of the real part of the eigenvalues of the drift matrix of the equations written for the fluctuations ; if these have the wrong sign the fluctuations can grow exponentially and the analysis loses its validity .
the second has to do with the size of the fluctuations themselves , in relation to the classical steady - state values .
as we will show , this second condition is not fulfilled for this system , as the quantum state of the mirror , or more accurately , of the mirror phonons , is reasonably expected to be thermal .
a characteristic of a thermal state is that the variance is larger than the mean value , which therefore makes any expression of the mirror phonons as having some well - defined classical mean value plus small fluctuations rather dubious .
even though cooling of the mirror via feedback mechanisms has been achieved @xcite , and analysed theoretically @xcite , due to the nature of the coupling between the electromagnetic field and the mirror phonons , it seems that all that can be achieved is a thermal state at a lower temperature , so that the problem remains . to treat the macroscopic mirror quantum mechanically
we will begin with the hamiltonian approach of law @xcite , in the approximation that only a single optical mode is important , and extended to include cavity pumping and damping . to treat the fluctuations of the mirror which result from its coupling to a thermal reservoir , we will use the brownian motion master equation developed by disi @xcite , which is suitable for the temperatures we will consider here . following a common procedure in quantum optics @xcite , we will develop a fokker - planck equation in the positive - p representation @xcite .
this fokker - planck equation allows us to write stochastic differential equations which are an exact mapping from the system master equation , and which can be used to calculate any desired normally - ordered operator moments .
we consider a system of a pumped fabry - perot cavity in which one of the mirrors is free to undergo oscillatory motion due to both the light pressure and thermal fluctuations .
we use the standard annihilation , @xmath1 , and creation , @xmath2 operators for the electromagnetic field , and the operators @xmath3 and @xmath4 for the displacement from the equilibrium position and the momentum of the mirror , which will be treated as a harmonic oscillator . neglecting the coupling of the mirror to its bath for the moment
, we can write the hamiltonian as @xmath5 where @xmath6 in the above , @xmath7 is the field frequency , @xmath8 represents the classical real pump , @xmath9 is the mass of the mirror , @xmath10 is the mirror oscillation frequency and @xmath11 is the coupling between the mirror and the cavity field , with @xmath12 being the length of the cavity .
the @xmath13s represent optical bath operators .
the damping of the mirror , which we will treat as markovian , will be included at the next step .
we now wish to write a master equation for the density matrix of our combined system in a frame rotating at @xmath7 . to do this
, we will make two different , but consistent approximations for the damping of the cavity and the mirror .
the cavity reservoir will be considered to be at zero temperature , which is consistent with the very high temperatures necessary to produce thermal photons at the frequencies involved .
the mirror reservoir will be treated as being at a finite temperature , which is necessary because of the number of thermal phonons which will be present in the system . as the temperatures required to create these respective excitations differ by many orders of magnitude , these approximations are not contradictory .
this process gives us @xmath14+\hat{{\cal l}}\hat{\rho}\nonumber\\ & = & \left[\frac{1}{2m}\hat{p}^{2}+\frac{1}{2}m\omega_{m}^{2}\hat{x}^{2}-\hbar g\hat{a}^{\dag}\hat{a}\hat{x}-i\hbar\epsilon(\hat{a}-\hat{a}^{\dag}),\hat{\rho}\right]\nonumber\\ & & + i\hbar\gamma\left(2\hat{a}\hat{\rho}\hat{a}^{\dag}-\hat{a}^{\dag}\hat{a}\hat{\rho } -\hat{\rho}\hat{a}^{\dag}\hat{a}\right)+d_{m}\hat{\rho},\end{aligned}\ ] ] where @xmath15 represents the loss rate through the fixed mirror and @xmath16 represents the mirror damping , using the brownian motion master equation developed by disi @xcite , @xmath17 -\frac{i\hbar\gamma_{m}}{2\lambda_{db}^{2}}\left[\hat{x},[\hat{x},\hat{\rho}]\right ] -\frac{i\kappa\gamma_{m}\lambda_{db}^{2}}{\hbar}\left[\hat{p},[\hat{p},\hat{\rho}]\right ] .
\label{eq : diosi}\ ] ] in the above , @xmath18 is the mirror damping rate , which depends on temperature through the mechanical quality factor , @xmath19 , the thermal de broglie wavelength of the mirror , with @xmath20 boltzmann s constant and @xmath21 the temperature .
@xmath22 is a numerical factor which must be greater than @xmath23 for this master equation to be of the lindblad form , but is not important here as it leads to terms of the order @xmath24 , which we will show to be insignificant at the temperatures we consider .
rather than writing the heisenberg equations of motion , which are difficult to solve , we will make use of the original definition @xcite of the annihilation and creation operators in terms of @xmath3 and @xmath4 and develop stochastic differential equations in the positive - p representation @xcite .
this allows us to use c - number equations which describe all the quantum properties of the mirror dynamics contained in the original master equation .
we will describe the operators @xmath3 and @xmath4 , in terms of the operators @xmath25 and @xmath26 , where @xmath27 with @xmath28 and @xmath29=1 $ ] .
writing the equations using these variables is advantageous because it allows us to automatically define a p - representation of the density matrix in terms of an expansion in the minimum uncertainty ( coherent ) states @xmath30 , defined as @xmath31 .
it also means that the mirror quadrature variances have a coherent state or vacuum value of @xmath23 , the same as for the electromagnetic field .
in fact , @xmath32 and @xmath33 represent the standard quantum limits ( sql ) for measurement of the mirror position and momentum , respectively .
we note that these phonon annihilation and creation operators have previously been used to describe the mirror , but not in the context of developing phase - space representation stochastic differential equations @xcite . in terms of these new variables , the master equation for the mirror damping is now written as @xmath34-\frac{i\hbar\gamma_{m}a^{2}}{2\lambda_{db}^{2}}\left[\hat{b}+\hat{b}^{\dag},\left[\hat{b}+\hat{b}^{\dag},\hat{\rho}\right]\right]\nonumber\\ & & -\frac{i\kappa\gamma_{m}\lambda_{db}^{2}}{\hbar}b^{2}[\hat{b}-\hat{b}^{\dag},[\hat{b}-\hat{b}^{\dag},\hat{\rho } ] ] .
\label{eq : betadiosi}\end{aligned}\ ] ]
using the well - known operator correspondences for the p - representation @xcite , we may map the master equation onto a partial differential equation for the p - function of the system , @xmath35 \right.\right.\nonumber\\ & & \left.\left .
+ \frac{\partial}{\partial\alpha^{\ast}}\left(\epsilon^{\ast}-\gamma\alpha^{\ast}-iga\alpha^{\ast}\left[\beta+\beta^{\ast}\right]\right ) \right.\right.\nonumber\\ & & \left.\left .
+ \frac{\partial}{\partial\beta}\left(-i\omega_{m}\beta+iga|\alpha|^{2}-\gamma_{m}\left[\beta-\beta^{\ast}\right]\right ) \right.\right.\nonumber\\ & & \left.\left .
+ \frac{\partial}{\partial\beta^{\ast}}\left(i\omega_{m}\beta^{\ast}-iga|\alpha|^{2}-\gamma_{m}\left[\beta^{\ast}-\beta\right]\right)\right ] \right.\nonumber\\ & & \left .
+ \frac{1}{2}\left[\frac{\partial^{2}}{\partial\alpha\partial\beta}\left(-iga\alpha\right)+ \frac{\partial^{2}}{\partial\beta\partial\alpha}\left(-iga\alpha\right)\right.\right.\nonumber\\ & & \left.\left .
+ \frac{\partial^{2}}{\partial\alpha^{\ast}\partial\beta^{\ast}}\left(iga\alpha^{\ast}\right ) + \frac{\partial^{2}}{\partial\beta^{\ast}\partial\alpha^{\ast}}\left(iga\alpha^{\ast}\right ) \right.\right.\nonumber\\ & & \left.\left .
+ \frac{\partial^{2}}{\partial\beta^{2}}\left(\gamma_{m}\left[1-\frac{2k_{b}t}{\hbar\omega_{m}}+\frac{\kappa\hbar\omega_{m}}{4k_{b}t}\right]\right ) + \frac{\partial^{2}}{\partial\beta^{\ast 2}}\left(\gamma_{m}\left[1-\frac{2k_{b}t}{\hbar\omega_{m}}+\frac{\kappa\hbar\omega_{m}}{4k_{b}t}\right]\right ) \right.\right.\nonumber\\ & & \left.\left .
+ \frac{\partial^{2}}{\partial\beta\partial\beta^{\ast}}\left(-\gamma_{m}\left[1-\frac{2k_{b}t}{\hbar\omega_{m}}-\frac{\kappa\hbar\omega_{m}}{4k_{b}t}\right]\right)\right.\right.\nonumber\\ & & \left.\left . + \frac{\partial^{2}}{\partial\beta^{\ast}\partial\beta}\left(-\gamma_{m}\left[1-\frac{2k_{b}t}{\hbar\omega_{m}}-\frac{\kappa\hbar\omega_{m}}{4k_{b}t}\right]\right ) \right]\right\}p(\alpha,\alpha^{\ast},\beta,\beta^{\ast},t ) .
\label{eq : betaplanck}\end{aligned}\ ] ] the diffusion matrix of the above equation is @xmath36 , \label{eq : betadiff}\ ] ] we note here that this drift matrix has diverging terms as @xmath37 , but this is not a problem as the disi master equation is valid in the limit where @xmath38 . as a physical example , in ref .
@xcite , we find @xmath39s@xmath40 , so that @xmath41j , whereas @xmath42j at @xmath43k , the temperature which we will mainly use in our investigations .
if we wish to treat eq .
[ eq : betaplanck ] as a genuine fokker - planck equation which we may map onto stochastic differential equations , the matrix @xmath44 must be positive - definite .
numerical investigations using typical parameters show that this is not the case , therefore for quantum calculations we will have to use the positive - p representation @xcite .
the positive - p representation equations in a doubled phase space can be found by the simple change of variables @xmath45 , so that , ( noting that @xmath46 only in the mean and similarly for @xmath47 ) , we now have four independent stochastic variables . ignoring the terms proportional to @xmath24 due to their small relative size , one possible factorisation of the diffusion matrix , @xmath48 , of eq .
[ eq : betadiff ] is @xmath49 , \label{eq : noisemat}\ ] ] which allows us to write a set of four stochastic differential equations ( note that the it form and the stratonovich form of these equations are identical . ) , @xmath50 in the above , the real gaussian noise terms have the correlations @xmath51 the set of coupled equations ( [ eq : ppsde ] ) may be integrated numerically , with averages taken over a large number of stochastic trajectories , which allows for the probabilistic calculation of any desired normally - ordered operator moments . as an example , with n trajectories , we have @xmath52 where @xmath53 labels the results from the @xmath53th trajectory .
before we return to the full stochastic equations , which we will solve numerically , we will investigate some of the classical properties of the system , which allow for analytical insights . from the drift part of eq .
[ eq : betaplanck ] , we can immediately write the mean - field equations using the notation @xmath54 for the classical mean - field value of @xmath55 , @xmath56 from which we may find the classical steady state solutions . solving eqs . [ eq : meanbeta ] for the steady - states , we find that @xmath57 is real , which means that the steady - state momentum is zero .
( note that this will not be the prediction of stochastic integration of the full equations . ) however , using this fact we may write the solutions as @xmath58 although the solutions above are not closed ( the solution for @xmath59 is a function of @xmath57 , etc . ) , we can make an iterative expansion , beginning with the result for fixed mirrors , @xmath60 and substitute this into the solution for @xmath57 .
this can then be substituted into the solution for @xmath59 , the process being repeated until we attain the required degree of convergence .
we note here that there are parameter regimes for which this expansion does not converge and that these are regions where we do not find classical steady - state solutions , but rather a limit - cycle , self - pulsing behaviour @xcite .
although we will demonstrate below that the classical steady - state solutions , especially for @xmath61 and @xmath62 , are not accurate in any parameter regime , they do allow for some insight into which property of the electromagnetic field outside the cavity is most likely to allow for an inference of the mirror position .
the usual candidates are the intensity , which may be measured by photodetection , and the quadratures , which may be measured by homodyne detection . defining the intracavity quadratures as @xmath63 and @xmath64 , with their classical equivalents written in terms of @xmath59 and @xmath65
, we find @xmath66 where we have set @xmath8 as real and used @xmath67 , with @xmath68 real .
as @xmath69 is typically much larger than @xmath70 , we may make a series expansion of these expressions .
we find that @xmath71 it is immediately obvious that @xmath72 depends on @xmath61 to first order , while the other two exhibit only a second order dependence .
this shows that homodyne measurements of the @xmath72 quadrature will be more sensitive to variations in the position of the mirror than will be the other two measurements , as previously noted by vitali _
et al . _
@xcite , although the @xmath73 quadrature will show a weaker dependence rather than being totally independent of @xmath61 as in the linearised analysis of ref .
@xcite .
it has been predicted that , with a non - zero detuning between the field and the cavity resonance , this system can exhibit bistability in the optical intensity @xcite . to find the condition for bistability ,
we start with the classical equations with detuning , @xmath74 , included , @xmath75 these have the steady - state solutions , @xmath76 which immediately leads to the cubic equation in @xmath77 , @xmath78 we find the condition for bistability by differentiating this expression with respect to @xmath79 , which gives @xmath80 the condition for bistability is that this quadratic equation has two positive real roots .
the roots are written as @xmath81 giving the inequality @xmath82 a necessary , but not sufficient condition is that @xmath74 be positive , as @xmath79 must be positive .
this immediately contradicts the condition given by mancini and tombesi @xcite , @xmath83 , which allows for negative intensities .
as @xmath79 must also be real , we find the condition for bistability as @xmath84 .
to numerically integrate eq .
[ eq : ppsde ] , we make use of the fact that the it and stratonovich forms are identical so that we may use a standard three - step predictor - corrector method .
the convergence of the algorithm was checked by comparison with a four - step method , and also by halving the time - step . in all the quantities
shown , the sampling errors are comparable to the thickness of the plotted lines
. we will use the published experimental parameters of ref .
@xcite , and make comparisons with the theoretical predictions reported elsewhere .
the oscillating mirror is considered as being perfectly reflecting , with a mass of @xmath85 kg , a mechanical quality factor of @xmath86 at @xmath43k , decreasing to @xmath87 at @xmath88k , and a resonance frequency @xmath89khz .
the damping rate of the mirror is @xmath90s@xmath40 at @xmath43k .
we consider a cavity length of @xmath91 cm , with cavity finesse @xmath92 , which gives @xmath93s@xmath40 .
we consider an optical wavelength of @xmath94 nm , which gives @xmath95s@xmath40 , and a coupling between the light and the mirror of @xmath96m@xmath40s@xmath40 .
the optical pumping of the cavity is @xmath97 , where @xmath98 is the laser power in watts . in stochastic integration of the equations which describe an intracavity optical system ,
the standard approach is to begin with the state inside the cavity as vacuum so that , with a continuous pump , the system enters the steady - state ( or its limit cycle behaviour in the case of self - pulsing ) after a few cavity lifetimes . in the present case
the situation is somewhat different , as not only the electromagnetic field , but also the oscillating mirror , has to reach the steady - state . as the relaxation time of the mirror
can be orders of magnitude larger than that of the intracavity field , and we need to average over a large number of trajectories to obtain reliable results , it is not practical to begin the integration with an arbitrary initial condition for the mirror .
naively beginning with @xmath99 , the ground state of the mirror , leads to extremely long lived transients , as this is far from the equilibrium state at finite temperature . to give some idea ,
even at @xmath43k , which is perhaps the lowest easily achievable temperature , the average number of mirror quanta becomes @xmath100 . for purposes of comparison , we will therefore choose the initial mirror state in two different ways and integrate the equations without any pumping of the cavity .
firstly , as a ( real ) coherent state , which has uncertainties in position and momentum at the sql , with the p - function @xmath101 and , secondly from the thermal distribution @xmath102 where @xmath103 . note that , at the beginning of each trajectory , @xmath104 and the phase is completely random for the thermal distribution .
we stress here that the variance in the number of mirror phonons for a thermal state is @xmath105 , which is very much larger than @xmath106 . using the planck distribution , @xmath107
we find that to achieve @xmath108 , we would need @xmath109k , and even then the variance would be @xmath0 , or twice the mean value .
it is important to note here that the number of phonons does not enter into the equations , but rather the quadratures @xmath110 and @xmath111 . in a linearised approach using our equations , it is the uncertainties in these which are important .
we can easily calculate these for a thermal state of the mirror with an unpumped cavity .
a simple integration gives @xmath112 equal to @xmath113 at @xmath114k for our system .
this is in stark contrast to a coherent state of the mirror , sometimes used to facilitate the mathematics of a linearised analysis , and for which @xmath115 . at @xmath114k and @xmath116k for an initial coherent state of the mirror and no optical pumping .
these results are the averages of @xmath117 and @xmath118 trajectories , respectively .
note that , unless otherwise stated , the values plotted here and in subsequent graphs are dimensionless . ] we have calculated the stochastic results for the means and standard deviations of the mirror position , without any optical pumping and for initial thermal and coherent states of the mirror at temperatures of @xmath43k and @xmath88k . in fig .
[ fig : xpumpco0 ] we show the stochastic results for the position , @xmath61 , of the mirror , with initial coherent states at these two temperatures . considering only these mean values
could give the erroneous impression that the mirror is in a non - stationary steady - state , which we can immediately see is not the case when we look at fig .
[ fig : vxpumpco0 ] , which shows the standard deviations , @xmath119 , for the same parameters .
although we have shown the standard deviations here , the variances for the initial coherent state continue to increase linearly for more than twice the time shown , which was as far as we continued the integration .
in contrast , for an initial thermal state , the mean value of the mirror displacement is , by definition , zero as can be seen from the equation for the p - function ( [ eq : pthermal ] ) .
we note here that our stochastic results over more than @xmath120 trajectories still showed oscillations of the order of @xmath121 m , but that we are confident that this small , but non - zero , value is due to the difficulty of sampling the distribution with a finite number of trajectories . by comparison with the coherent state values , for an initial thermal state @xmath122 m at @xmath114k , and is almost constant , indicating that this is a good choice of initial condition .
this value agrees well with the expression given in ref .
@xcite for the thermal noise , @xmath123 , which gives a value of @xmath124 m .
note that , over the time scales shown in fig .
[ fig : xpumpco0 ] we do not see any decay in the oscillations towards the thermal state values , as this would be expected to happen on a time scale of @xmath125 , which is approximately @xmath126s for the parameters used here .
in fact , although an initial coherent state of the mirror has been used in theoretical analyses ( see , for example , bose _
et al . _
@xcite ) , it is not at all obvious how this particular state may be constructed experimentally .
a thermal state arises naturally , and will be equal to a coherent state for @xmath127k , but absolute zero can not be reached experimentally . in optics ,
a coherent state can be described theoretically as a displacement of the vacuum by the displacement operator @xmath128 , which we can see has some relation to the optical pumping term of the hamiltonian , @xmath129 .
therefore an ideal empty cavity with this pumping term will naturally develop an intracavity coherent state .
we are not aware of any similar candidate for the mirror , even if it could begin in the @xmath127 vacuum state
. therefore we will use an initial thermal state in our investigations . at @xmath114k and @xmath116k for an initial coherent state of the mirror and no optical pumping .
note that these quantities were still increasing at twice the time shown here and would be expected to eventually attain the thermal values . ] when we examine the stochastic results for the intracavity field intensity , for an input power @xmath130mw , we find that the field exhibits a self - pulsing behaviour at approximately the resonance frequency of the mirror , as previously predicted @xcite .
however , the oscillations are of very small relative amplitude , at approximately @xmath131 of the average mean intensity . with increasing input power ,
the oscillations become larger until , for a power of @xmath132mw , for example , they are more than half the maximum intensity . at the lower power ,
the mean motion of the mirror is an oscillation between @xmath133 and @xmath134 m , while at the higher power oscillates between @xmath135 and @xmath136 m . even though these displacements are truly microscopic , they have a noticeable effect on the mean intensity , which should be easily detectable experimentally .
interestingly enough , these results are almost identical to what we find by numerical integration of the classical equations ( [ eq : meanbeta ] ) , although these can tell us nothing about the quantum correlations which we wish to investigate . among these quantum correlations
are the variances of the intracavity field and the fano factor , defined as @xmath137 .
these results , which we averaged over @xmath138 trajectories , are shown in fig .
[ fig : luzvar ] . for a coherent state ,
all three values are @xmath23 , which would be zero on the logarithmic vertical scale used here .
as all three correlations are greater than or equal to one , we do not see any squeezing of the field in the time domain , but do see excess noise in all three quantities . as is common with kerr media , there is more excess noise in the @xmath139 quadrature than in either the @xmath140 quadrature or in the intensity .
the solid line is @xmath141 , the dash - dotted line is @xmath142 , and the dotted line is the fano factor . ] much of the theoretical and experimental interest in this system has been in indirectly measuring small displacements or small forces which act on the oscillating mirror , by means of measurements on the optical field . generally theoretical results are presented in terms of output spectra which allow for an inferred value of the mirror position at various frequencies .
these spectra are simple to calculate in a linearised analysis which treats the system as an ornstein - uhlenbeck process @xcite , but we do not consider linearisation valid here , for the reasons we have stated . to calculate spectra from the results of stochastic integration
is also possible in many cases , but here is made difficult by the stiffness of our equations , where the field and the mirror oscillate on vastly different time scales .
the length of time needed to integrate a large enough number of trajectories over a sufficient time to give reliable results upon fourier transformation is prohibitive .
hence we will show results which were obtained in the time domain .
the first results we show for the mirror , in fig .
[ fig : sigxp ] , are for the uncertainties in the mirror position and momentum , defined as @xmath143 and @xmath144 .
the interaction with the field has not noticeably changed these quantities from the values we found via stochastic integration in the unpumped thermal state case , but has given the momentum a mean value which oscillates between approximately @xmath145kgms@xmath40 , whereas it was stable at zero in the unpumped case .
it is readily seen that the fluctuations in momentum are several orders of magnitude larger than the mean value .
we also note here that the uncertainties shown are certainly underestimated , due to the difficulties of simulating the large @xmath68 values with small probabilities in the initial conditions used .
this result also shows that , when the likely thermal state of the mirror is taken into account , the back - action noise of the field on the mirror has very little effect at the laser power considered here . at an increased power of @xmath132mw
the standard deviations in both position and momentum both increase due to this back - action noise , reaching oscillatory values around means of @xmath146 , while the mean momentum oscillates between @xmath147kgms@xmath40 .
. the standard s.i .
units for position and momentum are used . note that a minimum uncertainty state of the mirror has @xmath148 m and @xmath149kgms@xmath40 . ]
we next investigate the degree of correlation between the position of the mirror , the intracavity optical intensity , and the @xmath140 and @xmath139 quadratures .
this correlation function between two quantities @xmath150 and @xmath55 is defined as @xmath151 where a perfect correlation gives a value of @xmath23 , a perfect anticorrelation gives a value of @xmath135 and zero signifies no correlation at all . in fig .
[ fig : corrfun ] we see that , as predicted by eq .
[ eq : expansion ] , the strongest correlation is between @xmath139 and the mirror position , these two being almost perfectly correlated .
the correlation functions @xmath152 and @xmath153 oscillate at the frequency @xmath10 between a value of almost minus one and a value of approximately zero , ranging from being almost perfectly anticorrelated to almost perfectly uncorrelated .
this behaviour is quite different from that shown in a linearised fluctuation analysis , where we found that all three of these functions showed anticorrelation around the mirror frequency and were zero at other frequencies .
the difference between our fully quantum nonlinear results and the usual linearised predictions is dramatically demonstrated when we consider @xmath154 , used to analyse a possible quantum nondemolition measurement by jacobs _
et al . _
@xcite . their prediction ( see their eq .
31 ) is for a maximum spectral value of @xmath155 for the parameters we use in this work , whereas our stochastic prediction in the time domain oscillates between @xmath133 and @xmath156 , showing almost no correlation at all .
we think that it is unlikely that this difference can be explained by the difference between a spectral measurement and a time domain measurement , but that it is due to the inappropriateness of the linearisation procedure for this system .
( solid line ) , @xmath157 ( dash - dotted line ) , and @xmath153 ( dashed line ) , for an input power of @xmath158mw . ]
another way to infer the mirror position is by linear estimation following measurements of the optical field .
we follow a method proposed by reid @xcite in the context of a demonstration of the einstein - podolsky - rosen paradox , and also outlined in dechoum _
et al . _
we assume that a measurement of the @xmath139 quadrature allows for a linear estimate of @xmath110 , @xmath159 .
this is consistent with the expansion of @xmath72 given above , in eq .
[ eq : expansion ] .
after optimising for @xmath160 , the rms error in this estimate is given by @xmath161 which we may minimise as a function of @xmath162 , finding @xmath163 we may then write @xmath164^{2}}{v(\hat{y}_{a } ) } , \label{eq : vxinf}\ ] ] a quantity which we may calculate via stochastic integration .
the inferred uncertainty in a measurement of @xmath3 will then be @xmath165 , which will be equal to the standard quantum limit when the mirror is inferred to be in a minimum uncertainty state .
we have also calculated the inferred position uncertainty using measurements of @xmath140 and @xmath166 in the same manner . as shown in fig .
[ fig : xinfer ] , and expected from the previous discussion , using the @xmath139 quadrature gives better results than using either @xmath140 or @xmath166 , which give oscillatory inferences .
as it is , all of these give inferred uncertainties well above the sql of @xmath167 m for the parameters used .
we note here that the actual calculated value of @xmath119 , which oscillates between @xmath168 m and @xmath169 m , is greater than the value inferred through measurement of the @xmath139 quadrature , which has a steady - state value of approximately @xmath170 m .
this is also the case with the inferred measurements of ref .
@xcite , and is due to the almost perfect correlation between the position of the mirror and the @xmath139 quadrature . , calculated as in eq .
[ eq : vxinf ] .
the solid line is the estimate using @xmath171 , the dash - dotted line uses @xmath140 and the dotted line uses @xmath166 . ]
the pendular cavity and variations on the theme have been proposed as useful devices for the preparation of nonclassical states of the cavity field and also of the mirror or mirrors . in this section
we will give a brief review of some of these proposals , in chronological order , and explain why our results lead us to believe that they may not be as practical as the original authors suggest .
bose _ et al . _
@xcite suggest that this system may be used to prepare multi - component schrdinger cat states of the field , near number states , and entangled states when two or more modes interact with the mirror , as well as schrdinger cat - like states of the mirror prepared via quadrature measurements of the field .
beginning with a simple hamiltonian without any pumping or dissipation terms , they develop a time evolution operator which is seen to have a kerr - like term , which leads them to believe that this system may exhibit some of the nonclassical features of a kerr interaction .
they then assume that both the cavity field and the mirror are initially in coherent states .
while this simplifies the mathematics , and is reasonably accurate for the cavity field , our calculations have shown that a coherent state is not a reasonable initial condition for the mirror . as the nonclassical features predicted
depend on this initial condition , which at one stage of the paper is taken to be a vacuum state ( for simplicity ) , and interaction of the mirror with the environment is considered to be equivalent to that of a cavity field with a zero temperature reservoir , our results lead us to believe that these predicted nonclassical states will be , at the very least , extremely difficult to observe . in a subsequent article @xcite
, the same authors propose that the pendular cavity may be used to probe the decoherence of a macroscopic object , namely the mirror . in this work
the authors accept that the mirror will begin in a thermal state and note that a mixture of schrdinger cat states of the mirror can therefore be produced by the interaction with the field .
it is assumed that a superposition of the fock states @xmath172 and @xmath173 can be prepared inside the cavity , with the most simply prepared value of @xmath174 . using the p - representation expansion of the density matrix in terms of coherent states , an expression is found for the time development of this density matrix .
the field can then be measured via the interaction with a ground state two - level atom which is passed through the cavity .
the probability of the atom being excited can be related to the decoherence rate of the spatially separated superposed coherent states of the mirror s motion . examining this scheme with our parameters , we find that , with @xmath174 , the spatial separation between two of the superposed coherent states has a maximum of @xmath175 m , which is less than the thermal de broglie wavelength , @xmath176 m , so that it becomes difficult to claim we have a spatially separated superposition . when we take into account that the decomposition of the thermal state will include a huge number of coherent states , all with different phases
, it seems that the practical realisation of this scheme may be more demanding technologically than the authors had supposed .
another proposal uses radiation pressure to entangle two macroscopic mirrors @xcite , but relies on langevin equations which are linearised around their classical steady - state solutions . as we have shown above
, this process is of rather doubtful validity when the mirrors are coupled to thermal reservoirs .
a further idea is to produce an einstein - podolsky - rosen ( epr ) state in the position and momentum of two spatially separated oscillating mirrors using the output of a nondegenerate optical parametric oscillator ( opo ) @xcite .
we note here that , although the authors call the oscillator an optical parametric amplifier , the nonlinear crystal is inside a pumped cavity , so we will follow the usual terminology .
( see their fig.2 ) unlike most other treatments , this work uses an effective linear coupling between the light and the mirror .
like most of the others , the authors linearise langevin equations around their steady state solutions .
the opo is treated via two - mode equations which do not describe the normal well - known threshold behaviour of such a device at all , and lead to the prediction of an entangled state of two combined quadratures , which is said to demonstrate an epr correlation .
the two output fields of the opo are used to drive the two mirrors , which drives them into an epr state of position and momentum .
the mechanical damping of the mirrors introduces a noise term @xmath177 , where @xmath178 is the mean thermal phonon number .
this noise term seems to be the quadrature variance assuming a number state @xmath179 , of the mirror . for the parameters we use ,
this expression gives a value of @xmath180 , several orders of magnitude lower than our estimate in eq .
[ eq : vxybthermal ] , which gives @xmath181 .
due to the issues we have raised here , we are lead to believe that the epr state of the two mirrors may not be as easily demonstrated as indicated by the authors .
et al . _
@xcite have also put forward a proposal to entangle two movable mirrors which form part of a four - mirror ring cavity .
this work has the advantage of a description for the brownian motion that is consistent with quantum mechanics at all temperatures , but the results presented are also obtained following a linearisation process , which we have shown to not be valid for reasonable temperatures .
the final proposal we will consider here is by marshall _
et al . _
@xcite , which treats the creation of superposition states of a macroscopic mirror ( @xmath182 atoms ) via the interaction with a single photon .
this work again uses a hamiltonian approach without dissipation and assumes that the mirror can be prepared in its ground state , @xmath172 , which then allows for analytical solutions . the inclusion of decoherence and finite temperature follow the approach of ref .
@xcite and we therefore have the same doubts about the physical viability of this proposal .
it may be thought that to reduce the thermal noise of the mirror it is sufficient to cool to a lower temperature and impressive reductions in thermal noise , of the order of @xmath183 , have been achieved @xcite .
however , it seems that the method of feedback cooling leads to a thermal equilibrium of the mirror at a lower temperature . for our system , using eq .
[ eq : vxybthermal ] , a reduction by this factor leads to @xmath184 , which is still far above the coherent state level where linearisation of the equations may be expected to work .
@xcite suggests that a mirror may be cooled to as low as @xmath0mk by dilution refrigeration , which would give us @xmath185 , which , if we assume that feedback cooling from this temperature is as efficient as at room temperature , would allow the variances to be further reduced to something of the order of @xmath186 .
this is now actually smaller than the mean number of quanta , @xmath187 , equal to @xmath188 , although the variance in the number will now be of the order of @xmath189 , so that linearisation will still not be reliable .
standard cooling by feedback depends on the interaction of the electromagnetic field with the mirror phonons , which is proportional to @xmath190 , which will not drive the mirror toward a coherent state .
a coupling term proportional to @xmath191 may be expected to do this , while squeezing of the mirror position would conceivably be possible with a coupling of the type @xmath192 or @xmath193 .
it is interesting to note that an effective coupling of the type required to drive the mirror toward a coherent state was used by zhang _
et al . _
@xcite , although the physical parameters used to develop the coupling hamiltonian are far from those we have used in our analysis . as an example in that work , @xmath194 , whereas we have used the values @xmath195s@xmath40 and @xmath196s@xmath40 .
et al . _ also consider that the mirror damping is of the amplitude form , which is not what we find using the disi master equation .
et al . _
@xcite have actually presented a feedback scheme based on an effective hamiltonian which couples a light quadrature with @xmath197 , later proposing that this method , which they call stochastic cooling , could be used to beat the sql by achieving steady - state position squeezing of the mirror @xcite .
the effective coupling used again depends on the linearisation of the equations of motion . we feel that whether feedback can be used to cool a macroscopic oscillator towards , or even beyond , the sql remains an open question and subject to further research .
in conclusion , we have presented a fully quantum analysis of the pumped pendular cavity which does not depend on linearisation of the equations of motion .
we have shown that the thermal noise of the mirror is overwhelming for typical temperatures and experimental parameters .
this means that the linearisation procedure commonly used is of doubtful validity .
unless ways can be found to change the quantum state of the mirror in respect to the thermal excitations , we expect that the nonclassical states predicted in a number of theoretical analyses will not be able to be demonstrated experimentally .
feedback cooling techniques , while able to lower the effective temperature to an impressive degree , serve merely to produce another thermal state of the mirror at a lower temperature .
the noise is still overwhelming if we wish to reach the sql , or even beat it , which would be necessary for the detection of gravity waves .
this is also the case if we wish to observe some of the quantum superpositions and entanglement which have been theoretically predicted .
this research was supported by the new zealand foundation for research , science and technology ( grant no .
ufrj0001 ) , the australian research council , and the brazilian agencies cnpq ( conselho nacional de desenvolvimento cientfico e tecnolgico ) and capes ( coordenao de aperfeioamento de pessoal de nvel superior ) .
murray olsen would like to thank stojan rebi for interesting discussions and for his careful reading of the text . | we perform a quantum mechanical analysis of the pendular cavity , using the positive - p representation , showing that the quantum state of the moving mirror , a macroscopic object , has noticeable effects on the dynamics .
this system has previously been proposed as a candidate for the quantum - limited measurement of small displacements of the mirror due to radiation pressure , for the production of states with entanglement between the mirror and the field , and even for superposition states of the mirror . however , when we treat the oscillating mirror quantum mechanically , we find that it always oscillates , has no stationary steady - state , and exhibits uncertainties in position and momentum which are typically larger than the mean values .
this means that previous linearised fluctuation analyses which have been used to predict these highly quantum states are of limited use .
we find that the achievable accuracy in measurement is far worse than the standard quantum limit due to thermal noise , which , for typical experimental parameters , is overwhelming even at @xmath0mk . |
the applications of pedestrians dynamics range from the safety of large events to the planning of towns with a view to pedestrian comfort . because of the computational effort involved with an experimental analysis of the complex collective system of pedestrians behavior , computer simulations are run .
models continuous in space are one possibility to describe this complex collective system . in developing a model , we prefer to start with the simplest case : single lane movement .
if the model is able to reproduce reality quantitatively and qualitatively for that simple case , it is a good candidate for adaption to complex two - dimensional conditions . also in single file movement pedestrians interact in many ways and not all factors , which have an effect on their behavior , are known .
therefore , we follow three different modeling approaches in this work . all of them underlie diverse concepts in the simulation of human behavior .
+ this study is a continuation and enlargement of the validation introduced in @xcite . for validation , we introduce two criteria : on the one hand
, the relation between velocity and density has to be described correctly .
this requirement is fulfilled , if the modeled data reproduce the fundamental diagram . on the other hand
, we are aiming to reproduce the appearance of collective effects . a characteristic effect for the single file movement are stop - and - go waves as they are observed in experiments @xcite .
we obtained all empirical data from several experiments of the single file movement .
there a corridor with circular guiding was built , so that it possessed periodic boundary conditions .
the density was varied by increasing the number of the involved pedestrians . for more information about the experimental set - up ,
see @xcite,@xcite .
the first two models investigated are force based and the dynamics are given by the following system of coupled differential equations @xmath0 where @xmath1 is the force acting on pedestrian @xmath2 .
the mass is denoted by @xmath3 , the velocity by @xmath4 and the current position by @xmath5 .
this approach derives from sociology @xcite . here
psychological forces define the movement of humans through their living space .
this approach is transferred to pedestrians dynamics and so @xmath1 is split to a repulsive force @xmath6 and a driven force @xmath7 . in our case
the driving force is defined as @xmath8 where @xmath9 is the desired velocity of a pedestrian and @xmath10 their reaction time .
the other model is event driven .
a pedestrian can be in different states . a change between these states
is called event .
the calculation of the velocity of each pedestrian is straightforward and depends on these states .
the first spatially continuous model was developed by helbing and molnr @xcite and has often been modified . according to @xcite the repulsive force for pedestrian @xmath2 is defined by @xmath11 with @xmath12 , \tau=0.2\ , [ s]$ ] and @xmath13 is the distance between pedestrians @xmath2 and @xmath14 .
the fluctuation parameter @xmath15 represents the noise of the system . in two - dimensional scenes ,
this parameter is used to create jammed states and lane formations @xcite . in this study
, we are predominantly interested in the modeled relation between velocity and density for single file movement .
therefore the fluctuation parameter has no influence and is ignored .
first tests of this model indicated that the forces are too strong , leading to unrealistically high velocities . due to this
it is necessary to limit the velocity to @xmath16 , as it is done in @xcite @xmath17 in our simulation we set @xmath18 $ ] . in this
model pedestrians possess a degree of foresight , in addition to the current state of a pedestrian at time step @xmath19 .
this approach considers an extrapolation of the velocity to time step @xmath20 . for it
@xcite employs the linear relation between the velocity and the distance of a pedestrian @xmath2 to the one in front @xmath21 .
@xmath22 for @xmath23 $ ] and @xmath24 $ ] this reproduces the empirical data .
so with ( [ v(d ) ] ) @xmath25 can be calculated from @xmath26 which itself is a result of the extrapolation of the current state @xmath27 with @xmath28 .
finally , the repulsive force is defined as @xmath29 obviously the impact of the desired velocity @xmath9 in the driven force is negated by the one in the repulsive term . after some simulation time , the system reaches an equilibrium in which all pedestrians walk with the same velocity . in order to spread the values and keep the right relation between velocity and density , we added a fluctuation parameter @xmath30 . @xmath30
uniformly distributed in the interval @xmath31 $ ] reproduced the scatter observed in the empirical data . in this model pedestrians
are treated as hard bodies with a diameter @xmath32 @xcite .
the diameter depends linearly on the current velocity and is equal to the step length @xmath33 in addition to the safety distance @xmath34 @xmath35 based on @xcite the step length is a linear function of the current velocity with following parameters : @xmath36 + 0.302\ , [ s ] \,v_i(t ) \enspace .\ ] ] @xmath37 and @xmath38 can be specified through empirical data and the inverse relation of ( [ v(d ) ] ) . here
@xmath37 is the required space for a stationary pedestrian and @xmath38 affects the velocity term . for @xmath39 $ ] and @xmath40 $ ] the last equations ( [ d_i ] ) and ( [ st_i ] )
can be summarized to @xmath41 + 0.758\ , [ s ] \ , v_i(t)\enspace .\ ] ] by solving the differential equation @xmath42 the velocity function is obtained .
this is shown in fig .
[ fig1 ] together with the parameters of the pedestrians movement .
a pedestrian is accelerating to their desired velocity @xmath9 until the distance to the pedestrian in front is smaller than the safety distance . from this time on , he / she is decelerating until the distance is larger than the safety distance and so on . via @xmath43 and @xmath44 those events could be defined : deceleration ( [ bremsen ] ) and acceleration ( [ beschleunigen ] ) . to ensure good performance for high densities ,
no events are explicitly calculated .
but in each time step , it is checked whether an event has taken place and @xmath45 or @xmath46 are set to @xmath19 accordingly .
the time step , @xmath47 , of @xmath48 seconds is chosen , so that a reaction time is automatically included .
the discrete time step could lead to configurations where overlapping occurs . to guarantee volume exclusion , case ( [ stoppen ] ) is included , in which the pedestrians are too close to each other and have to stop
for the comparison of the modeled and experimental data , it is important to use the same method of measurement .
@xcite shows that the results from different measurement methods vary considerably .
the velocity @xmath4 is calculated by the entrance and exit times @xmath50 and @xmath51 to two meter section .
@xmath52}{(t_i^{out}-t_i^{in})[s]}.\end{aligned}\ ] ] to avoid discrete values of the density leading to a large scatter , we define the density by @xmath53 where @xmath54 gives the fraction to which the space between pedestrian @xmath2 and @xmath55 is inside the measured section , see @xcite .
@xmath56 is the mean value of all @xmath57 , where @xmath19 is in the interval @xmath58 $ ] .
we use the same method of measurement for the modeled and empirical data .
the fundamental diagrams are displayed in fig .
[ fd ] , where @xmath59 is the number of the pedestrians .
the velocities of the social force model are independent of the systems density and nearly equal to the desired velocity @xmath60 $ ] .
additionally we observe a backward movement of the pedestrians and pair formation . because of
these unrealistic phenomena are not observed in the other models , we suggest that this is caused by the combination of long - range forces and periodic boundary conditions . in contrast , the model with foresight results in a fundamental diagram in good agreement with the empirical one . through the fluctuation parameter ,
the values of the velocities and densities vary as in the experimental data .
we are satisfied with the results of the adaptive velocity model . for reducing computing time
, we also tested a linear adaptive velocity function , leading to a @xmath61 decrease in computing time for @xmath62 pedestrians .
the fundamental diagram for this linear adaptive velocity function is not shown , but also reproduces the empirical one .
during the experiments of the single file movement , we observed stop - and - go waves at densities higher than two pedestrians per meter , see fig . 5 in @xcite .
therefore , we compare the experimental trajectories with the modeled ones for global densities of one , two and three persons per meter . the results are shown in fig.[ges ] . since the social force model is not able to satisfy the criterion for the right relation between velocity and density , we omit this model in this section . figure [ ges ] shows the trajectories for global average densities of one , two and three persons per meter . from left
to right the data of the model with foresight , the adaptive velocity model and the experiment are shown . in the experimental data , it is clearly visible that the trajectories get unsteadier with increasing density . at a density of one person per meter
pedestrians stop for the first time .
so a jam is generated . at a density of two persons per meter stop - and - go waves
pass through the whole measurement range . at densities greater than three persons per meter pedestrians can hardly move forward .
for the extraction of the empirical trajectories , the pedestrians heads were marked and tracked .
sometimes , there is a backward movement in the empirical trajectories caused by self - dynamic of the pedestrians heads .
this dynamic is not modeled and so the other trajectories have no backward movement .
this has to be accounted for in the comparison . by adding the fluctuation parameter @xmath63 $ ] to the model with foresight a good agreement with experiment
is obtained for densities of one and two persons per meter .
the irregularities caused by this parameter are equal to the irregularities of the pedestrians dynamic .
nevertheless , this does not suffice for stopping so that stop - and - go waves appear , fig.[vor2 ] and fig.[vor3 ] . with the adaptive velocity model
stop - and - go waves already arise at a density of one pedestrian per meter , something that is not seen in experimental data .
however , this model characterizes higher densities well . so in comparison with fig .
[ ed3 ] and fig .
[ emp3 ] the stopping - phase of the modeled data seems to last for the same time as in the empirical data .
but there are clearly differences in the acceleration phase , the adaptive velocity models acceleration is much lower than seen in experiment .
finally other studies of stop - and - go waves have to be carried out .
the occurrence of this phenomena has to be clearly understood for further model modifications .
therefore it is necessarry to measure e. g. the size of the stop - and - go wave at a fixed position .
unfortunately it is not possible to measure over a time interval , because the empirical trajectories are only available in a specific range of 4 meters .
the well - known and often used social force model is unable to reproduce the fundamental diagram . the model with foresight provides a good quantitative reproduction of the fundamental diagram .
however , it has to be modified further , so that stop - and - go waves could be generated as well .
the model with adaptive velocities follows a simple and effective event driven approach . with the included reaction time , it is possible to create stop - and - go waves without unrealistic phenomena , like overlapping or interpenetrating pedestrians .
all models are implemented in c and run on a simple pc .
they were also tested for their computing time in case of large system with upto 10000 pedestrians .
the social force model offers a complexity level of @xmath64 , whereas the other models only have a level of @xmath65 . for this reason the social force model is not qualified for modeling such large systems .
both other models are able to do this , where the maximal computing time is one sixth of the simulated time . in the future
, we plan to include steering of pedestrians . for these models
more criteria , like the reproduction of flow characteristics at bottlenecks , are necessarry .
further we are trying to get a deeper insight into to occurrence of stop - and - go waves . | several spatially continuous pedestrian dynamics models have been validated against empirical data .
we try to reproduce the experimental fundamental diagram ( velocity versus density ) with simulations .
in addition to this quantitative criterion , we tried to reproduce stop - and - go waves as a qualitative criterion .
stop - and - go waves are a characteristic phenomenon for the single file movement .
only one of three investigated models satisfies both criteria . |
colossal magnetoresistance ( cmr ) in thin films of manganese based perovskites has been investigated intensively during the last years .
one motivation is the prospect to use these materials for magnetic devices , but a strong field in the tesla range is necessary to obtain large resistance changes .
+ another peculiarity , the highly spin polarized conduction band , allows to achieve high mr in low fields , because of the spin polarized tunneling of electrons between grains @xcite .
this can be realized by heteroepitaxial tunnel junctions with electrodes of manganites and a thin insulating barrier @xcite .
other approaches are to produce polycrystalline films @xcite , bulk materials @xcite , ramp - edge junctions @xcite , or thin films which are under compressive strain @xcite . as is well known from the high-@xmath3 superconductors , a further access is to deposit thin manganite films on bicrystal substrates to realize a single artificial grain boundary ( gb ) @xcite . + in order to understand the mechanism of the large low - field mr , we have carefully studied the resistivity of thin manganite films with and without an artificial gb , induced by a bicrystal substrate , as a function of temperature , magnetic field and current .
epitaxial thin films of la@xmath4ca@xmath5mno@xmath6 ( lcmo ) were grown from a stoichiometric target by pulsed laser deposition ( krf laser , @xmath7 nm ) . as substrates we used 45@xmath0 srtio@xmath6 bicrystals or monocrystals .
the optimized deposition conditions were a substrate temperature of 950@xmath0c in an oxygen partial pressure of 14pa .
the deposition rate was 0.3 / pulse with a pulse frequency of 3hz . with a mireau interferometer
the film thickness was determined to 190 nm .
after deposition the samples cooled in 500hpa o@xmath8 atmosphere without further annealing .
+ in x - ray diffraction in bragg - brentano geometry only film reflections corresponding to a ( 0 0 @xmath9 ) orientation of the cubic perovskite cell are visible .
the lattice parameter normal to the substrate plane ( @xmath10 ) is smaller than the bulk value and indicates a tensile strain of the @xmath11 axis due to the lattice mismatch of the substrate .
rocking angle analysis shows epitaxial @xmath12-axis oriented growth with an angular spread smaller than 0.03@xmath0 on both sides of the artificial gb .
the in - plane orientation was studied by @xmath13-scans on both sides of the bicrystal using the group of symmetry equivalent ( 2 0 2 ) reflections .
the cubic perovskite axes of the film are parallel to that of the srtio@xmath6 bicrystal substrate with an angular spread in the orientation of the individual epitaxial grains of less than 0.5@xmath0 .
thus the 45@xmath0 gb in the film induced by the bicrystal substrate is the only relevant large - angle gb in the following experiment .
+ the surface morphology of the samples was studied by scanning electron ( sem ) and by atomic force microscopy ( afm ) .
a shallow rectangular trench structure of 200 nm width and 5 nm depth along the grain boundary was visible in afm .
the gb itself was identified as a further sharp dip with characteristic dimensions close to the resolution limit of the used wide area scanner . the sem scan in fig .
1 shows that the actual distorted gb region is smaller than 50 nm . +
the films were patterned with conventional photolithography and chemically etched in an acidic solution of hydrogen peroxide .
we used a meander track which crosses the gb eleven times .
the total length is 5.6 mm and the width is 0.1 mm . for comparison
an identical second structure was prepared adjacent to the gb , but not crossing it .
+ the resistivity was measured in a superconducting magnet cryostat by standard four - point technique with a constant dc current between 10na and 100@xmath14a . for higher currents up to 10ma current pulses of 3ms length were applied in order to minimize sample heating . to see the low - field mr it is necessary to diminish demagnetization effects , which requires that the film plane is parallel to the magnetic field @xmath15 .
further the angle between the gb and the magnetic field is important for the mr ratio . in this paper we show the results for @xmath15 parallel gb , where the low - field mr is largest .
the resistances of the meander tracks as a function of temperature show the usual maximum in the resistivity and its suppression in high magnetic fields .
the conductivity above the curie - temperature @xmath16k is thermally activated , which can be explained by small polaron hopping @xcite . at low temperatures
the behavior of the two meanders is not the same .
the difference between the meander resistance on and beside the gb isolates the temperature dependence of the true gb resistance . in agreement with published data
there is a broad flat peak around 100k below @xmath3 @xcite .
+ in fig .
2 the resistivity as a function of magnetic field at 4.2k measured with a constant dc current of 1@xmath14a is shown . here
the magnetic field is parallel to the plane of the film and to the gb .
the mr defined as @xmath17/r(h=0)$ ] reaches its maximum value of 70% in a field of 120oe at 4.2k , which is approximately equal to the coercitive field . assuming spin polarized tunneling @xcite , the maximum mr is given by @xmath18 with @xmath19 the spin polarization parameter and @xmath20 and @xmath21 the resistances for the parallel and antiparallel configurations of the magnetizations , respectively . a value of 70% for @xmath22 will result in @xmath23 for lcmo .
we have to mention here that the different steps and switches in fig .
2 indicate that in the neighborhood of the gb multiple magnetic domains can exist and thus the spin polarization parameter will be underestimated .
+ with higher temperatures the maximum low - field mr decreases almost linearly down to 125k . at @xmath3 of the sample the hysteretic low - field mr vanishes , as can be seen in fig .
this temperature dependence is different to that of magnetic tunnel junctions @xcite .
the magnetic field value , where the maximum mr ratio occurs , does not shift with temperature up to 200k .
the inset of fig .
3 shows a hysteresis loop measured close to @xmath3 ( @xmath24k ) . at higher temperature
the hysteretic low - field mr becomes less pronounced and the intrinsic negative cmr effect becomes visible . above 150k the meander track not passing the gb shows comparable mr effects .
thus the artificial gb acts in this case as a magnetic field independent series resistor .
this indicates loss of magnetic order in the distorted gb region well below the bulk @xmath3 @xcite .
the low - field mr visible above 150k in both meanders shows a maximum value of 3% at 175k .
we think that this mr results from a large number of small angle grain boundaries between the individual epitaxial grains .
this effect decreases at higher temperatures due to the loss of spin orientation in the individual grains near @xmath3 of the sample .
below 175k the increasing magnetic domain size will lead to ferromagnetic coupling of a large number of grains in each magnetic domain .
thus only grains at magnetic domain boundaries contribute to the low - field mr .
if at low temperatures the magnetic domains are much larger than the crystal grain size , the number of not spin aligned grains is negligible and the low - field mr vanishes , as can be seen from the lower curve of fig .
3 . at 4.2k the noise level of the experiment
sets an upper limit for the mr of the reference meander of @xmath25% . within the above description
this requires a growth of the magnetic domain size by two orders of magnitude .
+ the influence of the gb on the transport properties is also visible in the current - voltage ( @xmath1-@xmath2 ) curves of the meanders . in the paramagnetic regime
both meanders show a nearly identical ohmic response .
below @xmath3 the one which crosses the gb is strongly nonlinear in contrast to the ohmic behavior of the reference meander as can be seen in fig . 4 for @xmath26k in zero field . with increasing bias current
the differential conductivity of the gb meander becomes asymptotic to that of the reference meander .
significant heating effects can be excluded , because the resistance decreases with increasing current , while the temperature coefficient of the resistivity is positive . + nonlinear @xmath1-@xmath2 curves are often taken as an indication for tunneling transport . in ferromagnetic materials
the inclusion of spin dependent density of states factors and hysteretic magnetizations on both sides of the tunneling barrier can explain a hysteretic mr . a description of the nonlinear @xmath1-@xmath2 curve shown in fig .
4 with the standard simmons tunneling model in the parabolic approximation @xcite is able to reproduce the data within the linewidth of the figure .
in addition to the tunneling parameters one has to take into account in a fitting procedure the resistivity drop in the non distorted regions . this can be done either as a free parameter or by using the resistivity of the reference meander .
we obtained as minimizing parameters a barrier height of @xmath27ev , a width of 2.1 nm , and a resistance of the undistorted region of 1630@xmath28 ( reference meander @xmath29@xmath28 ) .
such a small width of the barrier is not expected in view of the clear visibility of the gb in the sem scan . in spite of the nearly perfect fit possible for the nonlinear @xmath1-@xmath2 curve of fig . 4
, the investigation of direct @xmath1-@xmath2 data gives vanishing weight to the true differential conductivity at zero bias current .
therefore it is necessary to inspect the differential conductivity shown in fig .
5 . for very low bias
voltages a nearly parabolic behavior of the differential conductivity , as predicted by the simmons model , was found by steenbeck _
however , the inclusion of high bias current data in our measurements results in stronger systematic deviations of a simmons model from the experimental data at low voltages . the differential conductivity sketched in fig .
5 was calculated in the parabolic approximation . in view of the low barrier energy
this is not appropriate at very high bias currents , which correspond to high bias voltages . but
using the more general formulas given by simmons , will even enhance the deviations .
thus it is unclear if charge transport takes place by gb - tunneling .
+ also the observed anisotropy of the mr with respect to the magnetic field direction can not be easily understood in this framework .
therefore as an alternative mechanism evetts _
@xcite proposed , that the gb does not act as a tunneling barrier , but as a mesoscale region with distorted magnetic and transport properties .
the resistivity of the gb region is assumed to be a function of the gb magnetization , which depends on the effective field acting in the gb region .
this interpretation is in agreement with the observed anisotropy .
also the loss of magnetic order in the mesoscale region can naturally occur well below the bulk @xmath3 and will be responsible for the vanishing low - field mr at higher temperatures .
the up to now proposed transport mechanisms in the mesoscale region , however , are essentially thermally activated and ohmic .
therefore they can not simply account for nonlinear @xmath1-@xmath2 curves .
+ + a dependence of the hysteretic low - field mr on the current strength , which we present in the following , is in neither in a tunneling model nor in a mesoscale transport model a priori expected .
+ we investigated the dependence of the low - field mr on the transport current on a meander which shows a maximum mr peak @xmath30 of 22% at @xmath26k .
the applied current range spanned six decades , from 10na to 10ma .
6 shows that the width of the hysteresis loops is current independent , while there is a strong current dependence of the height of the mr peaks .
the suppression of the mr with current is shown in fig . 7 for bias currents from 22% for 10na to 1.4% for 10ma . + a close inspection of the data shows , that a current dependence exists also for magnetic field strengths well above the hysteretic region .
the decrease of this negative cmr of the gb meander with increasing current is shown in an enlarged scale in fig .
8 . on the reference
meander no current dependence of the resistivity is observed .
the negative linear cmr slope @xmath31 as function of the current was measured up to 10 t and is also plotted in fig .
the close correlation between the cmr suppression and the low - field mr peak suppression with current indicates the same physical origin for both effects in the gb region .
since high bias currents correspond to high bias voltages one possible explanation is , that at high bias voltage a new spin independent channel for charge transport opens .
however , due to the intimate relationship between ferromagnetism and spin polarized charge transport in the manganites a different scenario is possible : a polarized charge carrier from the undisturbed region will induce due to double exchange a spin polarization in the distorted region during its transition , i.e. the carrier leaves a trail of localized spins behind , which are more than usually aligned .
each manganese ion will encounter the passage of up to @xmath32 polarized carriers per second for the current values used in our experiment .
if the spin relaxation time is much longer than 10@xmath33s the charge transport induced magnetization will be nonzero in the time average and a second carrier can use this spin polarized path before the spin relaxation has destroyed the induced order !
the existence of electron spin resonance measurements at microwave frequencies in the manganites @xcite , shows that the relaxation time is longer than ( 10ghz)@xmath34 , and therefore this scenario is imaginable . in the limit of full magnetization of the distorted region
the gb meander resistivity will be indistinguishable from that of the reference meander .
this interpretation does not depend on the details of the transport of the charge carriers from one manganese ion to the next in the mesoscale region , but requires only a hopping probability depending on the respective spin alignment as proposed by several authors @xcite .
by pulsed laser ablation we prepared thin films of la@xmath4ca@xmath5mno@xmath6 on srtio@xmath6 bicrystal substrates .
x - ray diffraction showed full epitaxial growth on both sides of the bicrystal substrate .
their low- and high - field mr was measured as a function of magnetic field , temperature and current .
we obtained hysteretic mr values up to 70% at low temperatures in low magnetic fields .
the mr is induced by switching of magnetic domains at the coercitive field .
with increasing temperature the artificial grain boundary mr decreases almost linearly .
we observed a highly non ohmic behavior of the current - voltage relation , which leads to a nearly complete suppression of the mr effect at high bias currents .
the current dependence of low- and high - field mr is identical , which indicates the same physical origin for this suppression .
we discussed the data in view of proposed tunneling and mesoscale magnetic transport models .
an explicit dependence of the spin polarization on the applied current is necessary for a consistent interpretation . in the mesoscale transport model
the number of manganese ions in the distorted grain boundary region is comparable to the number of charge carriers passing the region within the spin relaxation time and we propose a current induced magnetization change . | we prepared epitaxial ferromagnetic manganite films on 45@xmath0 bicrystal substrates by pulsed laser ablation .
their low- and high - field magnetoresistance ( mr ) was measured as a function of magnetic field , temperature and current . at low temperatures hysteretic changes in resistivity up to 70% due to switching of magnetic domains at the coercitive field
are observed .
the strongly non - ohmic behavior of the current - voltage ( @xmath1-@xmath2 ) leads to a complete suppression of the mr effect at high bias currents with the identical current dependence at low and high magnetic fields .
we discuss the data in view of tunneling and mesoscale magnetic transport models and propose an explicit dependence of the spin polarization on the applied current in the grain boundary region . 2 |
we know that nature is governed by mathematics and symmetries .
not very long ago , it was an article of faith among most physicists that everything about physics would eventually be explained in terms of fundamental symmetries that nothing in the makeup of physical laws is accidental , that nature ultimately has no choices , and that all the properties of particles and fields are fixed by pure math . in the thirty years since
modern anthropic reasoning was introduced into cosmology@xcite , the competing idea that anthropic selection might have an indispensable role in fundamental physical theory has gradually become , if not universally accepted , at least mainstream .
there are now concrete physical models for realizing anthropic selection in nature .
cosmology has provided not only a concrete mechanism ( inflation ) for manufacturing multiple universes , but also a new phenomenon ( dark energy ) whose value is most often explained by invoking anthropic explanations .
string theory has uncovered a framework by which many different symmetries and parameters for fields can be realized in the low - energy , 3 + 1-dimensional universe , depending on the topology and size of the manifold of the other seven , truly fundamental dimensions , and on the configurations of p - branes within it , especially the local environment of the 3-brane on which our own standard model fields live .
the numbers of locally metastable configurations of manifold and branes , and therefore the number of options for low energy physics , are estimated to be so large that for all practical purposes , there is a continuum of choices for fundamental parameters that we observe@xcite . of course
, the details of how this works in the real world are still sketchy .
cosmology unfolds in a series of phase transitions and symmetry breakings .
for example , it is now part of standard inflation that the quantum wavefunction of the universe branches early into various options for the zero - point fluctuations of the inflaton field , different branches of which correspond to different distributions of galaxies .
string theory opens up a scenario in which the multiverse wavefunction may also branch very early into a variety of whole universes , each of which has different physics .
if things happen this way , it is natural for us to find ourselves in a branch with physics remarkably well suited to make the stuff of which we are made .
it then makes sense to ask new questions about the world : how would things change if this or that aspect of physics were changed ?
if a small change in a certain parameter changes the world a great deal in a way that matters to our presence here , that is a clue that that particular parameter is fixed by selection rather than by symmetry .
the following arguments along these lines are elaborated more fully in ref .
@xcite .
now we may be faced where a situation where some seemingly fundamental features of physics might not ever be derived from first principles .
even the particular gauge group in our grand unified theory ( that is , the the one in our branch of the wavefunction ) might be only one group selected out of many options provided by the theory of everything .
we may have to adjust our scientific style to this larger physical reality , which forces cosmology and fundamental physics into a new relationship .
for example , although we ca nt look inside the other universes of the multiverse ensemble and ca nt predict the branching outcome from first principles , cosmological experiments now under development might reveal relict gravitational waves from the same symmetry breaking that fixed the parameters .
evaluating changes in the world in response to changes in the fundamental physics is actually a difficult program to carry out . for the most fundamental theory we have , the standard model , the connection of many of its parameters with generally observable phenomena can only be roughly estimated .
first - principles calculations of the behavior of systems such as nuclei and molecules are possible only for the simplest examples .
the traditional minimal standard model has 19 `` adjustable '' parameters@xcite : yukawa coefficients fixing the masses of the six quark and three lepton flavors ( @xmath0 ) , the higgs mass and vacuum expectation value @xmath1 ( which multiplies the yukawa coefficients to determine the fermion masses ) , three angles and one phase of the ckm ( cabibbo - kobayashi - maskawa ) matrix ( which mixes quark weak- and strong - interaction eigenstates ) , a phase for the quantum chromodynamic ( qcd ) vacuum , and three coupling constants @xmath2 of the gauge group , @xmath3 .
we now know experimentally that the neutrinos are not massless , so there are at least seven more parameters to characterize their behavior ( three masses and another four ckm matrix elements ) .
thus 26 parameters , plus newton s constant @xmath4 and the cosmological constant @xmath5 of general relativity , are enough to describe the behavior of all the observed particles in all experiments , except those related to new dark matter particles .
if in addition the standard model is extended by supersymmetry , the number of parameters exceeds 100 .
imagine that you are sitting at a control panel of the universe .
it has a few dozen knobs one for each of the parameters .
suppose you start twiddling the knobs .
for all but a few of the knobs , you find nothing changes very much ; the mass of the top quark for example ( that is , its yukawa coupling coefficient in the standard model equations ) has little direct effect on everyday stuff . which knobs matter for the stuff we care about most atoms and molecules ?
some knobs are clearly important , but their exact value does not seem too critical . the fine structure constant @xmath6 for example controls the sizes of all the atoms and molecules , scaling like the bohr radius @xmath7 .
if you twiddle this knob , natural phenomena dominated by this physics which includes all of familiar chemistry and biology grow or shrink in size . on the other hand they all grow or shrink by almost the same fractional amount
so the structural effect of changes is hard to notice ; the miraculous fit of base pairs into the dna double helix would still work pretty well .
there are however subtle changes in structural relationships and molecular reaction rates .
your complicated biochemistry probably would not survive a sudden big change in @xmath6 , but if you turn the knob slowly enough , living things probably adapt to the changing physics .
simulations of cellular reaction networks show that their behavior is remarkably robust with respect to changes in reaction rates , and mostly depend on network topology@xcite .
it turns out that a few of the knobs have a particularly large qualitative effect with a very small amount of twiddling .
three knobs stand out for their particularly conspicuous effects : the yukawa coefficients controlling the masses of the electron , the up - quark , and the down - quark .
they are the light fermions that dominate the composition and behavior of atoms and molecules .
changing them by even a small fractional amount has a devastating effect on whether molecules can exist at all .
the most dramatic sensitivity of the world on their values seems to be in the physics of atomic nuclei .
the light fermion masses are all very small ( less than one percent ) compared with the mass of even a single proton .
( protons and neutrons , which comprise the bulk of the mass of ordinary matter , ironically have a mass dominated not by the `` real mass '' of their matter particles , the constituent quarks , but almost entirely by the kinetic energies of the quarks and the massless gluons mediating the color forces . )
the light fermion masses are critical because they determine the energy thresholds for reactions that control the stability of nucleons . in the three dimensional parameter space formed by these masses ,
the most reliable phenomemological statements can be made about changes within the two dimensional surface defined by holding the sum of @xmath9 and @xmath10 masses constant .
( that is because many complicated features of nuclear physics remain constant if the pion mass , which is proportional to @xmath11 , is constant ) . in this plane , some properties of worlds with different values of the masses are summarized in figures 1 and 2 , the latter taken from ref .
the figures also show a constraint from a particular so(10 ) grand unified scenario , to illustrate that likely unification schemes probably do not leave all these parameters independent at least one relationship between them is likely fixed by unification symmetry . in the lower part of figure 1 , towards larger up - quark mass , there are `` neutron worlds '' .
as one dials knobs in this direction , a threshold is soon crossed where it is energetically favorable for the electron in a hydrogen atom to join with its proton to make a neutron .
if you turn it farther , even a free proton ( without any nearby electron ) spontaneously decays into a neutron . in the upper part of the figure ,
there are `` proton worlds '' . moving up from our world , a threshold is soon crossed where
a deuteron in a plasma is no longer energetically favored over a pair of protons .
if you go farther , even an isolated deuteron spontaneously decays into a pair of protons . in the neutron world , there are nuclei , but not atoms with electrons around them , so chemistry does not happen . in the proton world , there are hydrogen atoms , but that is the only kind of atom , because the other nuclei do not form or are not stable .
fortunately for us there is a world in between , where a few dozen stable nuclei are both possible and are actually produced in stars , and are endowed with electron orbitals leading to chemistry with arbitrarily large and complex molecules .
this world would disappear with only a few percent fractional change in the quark mass difference in either direction .
it does not exist in some closely - related branches of the multiverse wavefunction .
= 3.5 in one can estimate roughly the effects of leaving this plane . in that case , nuclear physics is changed in new ways , since the mass of the pion changes .
it appears that if the masses are increased by more than about 40% , the range of nuclear forces is reduced to the point where the deuteron is unstable ; and if they are reduced by a similar amount , they are strengthened to the point where the diproton is stable . on the other hand
, the latter change also reduces the range of nuclear forces so there are fewer stable elements overall .
the sum of the quark masses in our world appears roughly optimized for the largest number of stable nuclei .
again , the situation would change qualitatively ( e.g. , far fewer stable elements ) with changes in summed quark masses at the ten percent level .
why is it even possible to find parameters balanced between the neutron world and the proton world ?
for example , if the so(10 ) model is right one , it seems that we are lucky that its trajectory passes through the region that allows for molecules .
the answer could be that even the gauge symmetries and particle content also have an anthropic explanation . a great variety of compact 7-manifolds and 3-brane configurations solve the fundamental m theory .
each one of them has dimensional scales corresponding to parameter values such as particle masses , as well as topological and geometrical relationships corresponding to symmetries .
many of these configurations undergo inflation and produce macroscopic universes .
in this situation it is not surprising that we find ourselves in one where atoms and nuclei can exist .
discussions of anthropic selection have sometimes differentiated between the kind that selects whole universes ( with different values of the electron mass , etc . ) , and the kind that selects a congenial environment ( why we do not live on an asteroid or a quasar , etc . ) while these seem very different , from a quantum - mechanical perspective they do not differ in kind .
both of them are selections of a congenial branch of the wavefunction of the universe . in the original formulation of quantum mechanics
, it was said that an observation collapsed a wavefunction to one of the eigenstates of the observed quantity .
the modern view is that the cosmic wavefunction never collapses , but only appears to collapse from the point of view of observers who are part of the wavefunction .
when schrdinger s cat lives or dies , the branch of the wavefunction with the dead cat also contains observers who are dealing with a dead cat , and the branch with the live cat also contains observers who are petting a live one . although this is sometimes called the `` many worlds '' interpretation of quantum mechanics , it is really about having just one world , one wavefunction , obeying the schrdinger equation : the wavefunction evolves linearly from one time to the next based on its previous state .
anthropic selection in this sense is built into physics at the most basic level of quantum mechanics .
selection of a wavefunction branch is what drives us into circumstances in which we thrive .
viewed from a disinterested perspective outside the universe , it looks like living beings swim like salmon up their favorite branches of the wavefunction , chasing their favorite places .
the selection of a planet or a galaxy is a matter of chance . in quantum mechanics
this means a branch of the wavefunction has been selected .
the binding energy of our galaxy was determined by an inflaton fluctuation during inflation ; that was when the branching occurred that selected the large scale gravitational potential that set the parameters for our local cosmic environment .
we can achieve statistical understanding about this kind of selection because we can observe other parts of the ensemble , by observing galaxy clustering , the microwave background , and so on . in this way , we understand the physics of the symmetry breaking .
we even know something about the formation of the different galaxy distributions in other universes we will never see .
these are regarded as just different by chance .
if the quark and electron masses are also matter of chance , the branching of the wavefunction occurred along with the symmetry breaking that fixed their masses .
there may be ways to observe aspects of the statistical ensemble for this event also , by studying the gravitational wave background rather than the microwave background .
we do not know when all the choices of parameters and symmetries are made .
some of these branchings may leave traces of other choices observable in our past light cone
. it could be that some parameters are spatially varying even today , in response to spatial variations in scalar or dark matter fields .
for example , one model of dark energy predicts large variations in the masses of neutrinos , depending on the local density of the neutrino component of dark matter@xcite .
( indeed the basic idea that effective neutrino masses depend on the local physical environment is now part of the standard theory of solar neutrino oscillations . )
thus the properties of stars can be spatially modulated depending on the dark matter density a quantity determined , in many theories , by a branching event that occurred recently enough to have an observable ensemble .
such ideas provide a new motivation for observational programs to quantify the extent to which the constants of nature are really constant in spacetime .
( a thriving example of this can be found in studies of varying @xmath6 . ) in some models , events connected with fixing the local quark and electron masses may have happened late enough to leave fossil traces .
this could happen during the final compactification of some of the extra dimensions , or the condensation of our own standard model 3d brane within a larger dimensional space .
if dimensional compactification happens in a sufficiently catastrophic symmetry breaking , it can lead to a background of gravitational waves .
because they are so penetrating , gravitational waves can carry information directly from almost the edge of our past light cone , well beyond recombination , even beyond weak decoupling indeed , back to the edge of 3d space as we know it .
if the extra dimensions are smaller than the hubble length at dimensional compactification or brane condensation , their collapse can appear as a first - order phase transition in our 3d space , leading to relativistic flows of mass - energy .
if the extra dimensions are larger than or comparable to the hubble scale , the 3d brane we inhabit may initially condense with warps and wiggles that lead to a gravitational wave background . either way the mesoscopic , classical motion of branes settling down to their final equilibrium configuration could lead to a strong gravitational - radiation background in a frequency range detectable by detectors now under development@xcite .
thus , instruments designed to observe the early boundary of spacetime may also explore the early boundary of physics as we know it , and directly test ideas concerning the separation of various branches of the multiverse having different fundamental parameters .
this blending of empirical cosmology and fundamental physics is reminiscent of our darwinian understanding of the tree of life .
the double helix , the four - base codon alphabet and the triplet genetic code for amino acids , any particular gene for a protein in a particular organism are all frozen accidents of evolutionary history .
it is futile to try to understand or explain these aspects of life , or indeed any relationships in biology , without referring to the way the history of life unfolded . in the same way that ( in dobzhansky s phrase ) , `` nothing in biology makes sense except in the light of evolution '' ,
physics in these models only makes sense in the light of cosmology . c. j. hogan , `` why the universe is just so , '' rev .
phys . * 72 * , 1149 ( 2000 ) [ arxiv : astro - ph/9909295 ] .
r. n. cahn , `` the eighteen arbitrary parameters of the standard model in your everyday life '' , rev .
phys . * 68 * , 951 ( 1996 ) | in a model where a multiverse wavefunction explores a multitude of vacua with different symmetries and parameters , properties of universes closely related to ours can be understood by examining the consequences of small departures of physical parameters from their observed values .
the masses of the light fermions that make up the stable matter of which we are made the up and down quarks , and the electron have values in a narrow window that both allows a variety of nuclei other than protons to exist , and at the same time allows atoms with stable shells of electrons that are not devoured by their nuclei .
these fundamental parameters of the standard model are good candidates for quantities whose values are determined through selection effects within a multiverse , since a living world of molecules needs stable nuclei other than just protons and neutrons .
if the fermion masses are fixed by brane condensation or compactification of extra dimensions , there may be observable fossils of the branching event , such as a gravitational wave background . |
cooperation in cellular networks has been recently suggested as a promising scheme to improve system performance , especially for cell - edge users @xcite , @xcite . in this work ,
we study a cooperation model where the positions of base stations ( bss ) follow a poisson point process ( p.p.p . )
distribution and where voronoi cells define the planar areas associated with them .
the approach is known to provide closed ( or integral ) form expressions for important performance metrics averaged out over such large random networks .
we consider here a specific scheme of cooperation with coherent transmission , where at most two bss can serve a single user .
the scheme is based on the ideas of conferencing and user - power splitting .
it originates from the seminal work of willems @xcite for the @xmath0 cooperative multiple access channel ( mac ) and its gaussian variant in @xcite .
an adaptation of the idea in a network - mimo setting is proposed by zakhour and gesbert for the @xmath1 model in @xcite , where the beamforming vectors take the role of signal weights and define the power - split ratios .
our problem position within the framework of stochastic geometry , extends previous work by andrews , baccelli and ganti @xcite to include cooperative schemes .
more specifically , for the service of each user , either one or two bss are involved ( section [ sii ] ) .
if two , these cooperate by exchange of information with conferencing over some backhaul link .
we assume that , in addition to the user data , only part of the total channel state is available and exchanged .
we choose to analyze schemes with partial knowledge in order to investigate possible benefits of cooperation without the excessive costs from full channel adaptation .
specifically , only the phase of the channel from the second closest transmitter to the user is known , while the first bs is informed over this and transmits coherently by appropriate choice of its own phase .
the scheme considers a fixed transmission power budget per user - which is split between the two bss and a common message is encoded .
the two transmitted common signals add up in - phase at the user receiver , resulting in an extra term for the beneficial signal ( section [ siii ] ) . later in the work , extra knowledge of the interference from the second bs neighbour to the reference user
is considered available .
the cooperative pair then transmits orthogonaly to this signal by application of dirty paper coding ( section [ sivd ] ) .
each user chooses on its own to be served with or without bs cooperation .
the decision is directed by a family of geometric policies , which depend on the ratio of the user distance to its first and second closest geographic bs neighbour and some design parameter @xmath2 , left as optimization variable ( section [ siv ] ) . in this way
the plane is divided into non - overlapping zones of cooperation ( @xmath3 ) and no cooperation ( @xmath4 ) .
an exact expression of the coverage probability in the network under study is derived ( section [ sv ] ) .
numerical evaluation allows one to analyze coverage benefits compared to the non - cooperative case .
these benefits are more important with dirty paper coding ( section [ svi ] ) .
it is concluded ( section [ svii ] ) that cooperation can significantly improve coverage without exploitation of further network resources ( frequency , time , power ) , even with schemes of reduced channel state information .
this work continues the line of research on applications of point processes to wireless networks and takes a step further to consider cooperative transmission in cellular networks . during the last years
important results have been derived by use of stochastic geometry tools .
the theoretical aspects of point processes and its applications in telecommunications can be found in the book of baccelli and baszczyszyn @xcite .
important contributions on the capacity of wireless networks appear in the book of weber and andrews in @xcite and for k - tier networks in the work of dhillon et al @xcite , as well as keeler , baszczyszyn and karray in @xcite .
haenggi and ganti have investigated the modeling and analysis of network interference by use of point processes @xcite . in the problem of cooperation , parallel to our work akoum and
heath @xcite have provided approximative performance evaluation when bss are randomly grouped into cooperative clusters . the transmission scheme they consider is intercell interference nulling .
their approach differs from our work , because we consider optimal geographic association of each user with a pair of bss and coherent transmission ( after conferencing ) from the latter .
all proofs of the analysis in this paper can be found in our extended version @xcite .
for the model under study the bss considered are equipped with a single antenna and are positioned at the locations of atoms from the realization of a planar p.p.p . with intensity @xmath5 , denoted by @xmath6 .
a planar tessellation , called the 1-voronoi diagram @xcite , partitions ( up to lebesgue measure zero ) the plane into subregions called cells .
the * 1-voronoi cell * @xmath7 associated with @xmath8 is the locus of all points in @xmath9 which are closer to @xmath8 than to any other atom of @xmath10 .
we consider euclidean distance @xmath11 .
then @xmath12 when the 1-voronoi cells of two atoms share a common edge , they constitute * delaunay neighbours * @xcite .
a dual graph of the 1-voronoi tessellation called the delaunay graph is constructed if delaunay neighbours are connected by an edge .
the 2-voronoi diagram @xcite consitutes another partition ( up to lebesgue measure zero ) of the plane . specifically , the * 2-voronoi cell * @xmath13 associated with @xmath14 , @xmath15 is the locus of all points in @xmath9 closer to @xmath16 than to any other atom of @xmath10 , i.e. @xmath17 the 1-voronoi tesselation for @xmath18 randomly positioned atoms is shown in fig . [
fig : examplecoop04 ] and [ fig : examplecoop09 ] . in
both , the coloured area depicts a 2-voronoi example region . in the present work we consider a geometric cooperation scenario based on the following assumptions : * each bs
is connected via backhaul links of infinite capacity with all its delaunay neighbours . *
exactly one user with a single antenna is initially associated with every bs .
each user is located randomly at some point within the 1-voronoi cell of its bs and we write @xmath19 . *
each user @xmath20 may be served by either one or two bss . if two , these correspond to the atoms of @xmath10 which are its * first * and * second closest geographic neighbour*. if one , it is just the first closest neighbour .
we use the notation @xmath21 and @xmath22 when referring to these neighbours . by definition
the user belongs to @xmath23 .
* from the point of view of a bs located at @xmath8 , we refer to the user in its 1-voronoi cell as the * primary user * @xmath20 and to all other users served by it but located outside the cell as the * secondary users*. these constitute a set @xmath24 , with cardinality that ranges between zero and the number of delaunay neighbours , depending on the users position relative to @xmath8 . the distance between user @xmath20 and its first ( resp .
second ) closest bs neighbour equals @xmath25 ( @xmath26 ) .
the second nearest bs @xmath27 can only be one of the delaunay neighbours of @xmath28 .
the communications scenario in this work applies the following idea . when two bss cooperatively serve a user in the downlink , its signal is split into one common part served by both , and private parts served by each one of the involved bss .
the common part contains information shared by both transmitters after communication between them over a reliable conferencing link .
more specifically , for each primary user @xmath20 located within the 1-voronoi cell of atom @xmath8 , consider a signal @xmath29 to be transmitted .
the user signals are independent realizations of some random process with power @xmath30 = p>0 $ ] and are uncorrelated from other user - signals meaning @xmath31 = 0 $ ] , @xmath32 .
the signal for user @xmath20 in the downlink is split in two parts : @xmath33 * a * private ( pr ) * part sent to @xmath20 from its first bs neighbour @xmath34 , denoted by @xmath35 .
the second neighbour does not have a private part to send . * a * common ( c ) * part served by both @xmath21 and @xmath22 , which is denoted by @xmath36 .
this part is communicated to both bss over the backhaul links . for the sake of clarity
, we will use the notation @xmath37 , @xmath38 for the common signal transmitted from @xmath21 and @xmath22 respectively , although the two signals are actually scaled versions of each other . the two parts are uncorrelated random variables ( r.v.s ) , in other words @xmath39=0 $ ] . considering power issues , we put the constraint that the total power transmitted from all bss to serve user @xmath20 sums up to @xmath40 to guarantee * power conservation*. we assume that the common part is transmitted by both bss with the same power percentage @xmath41 $ ] , named the * power - split ratio*. @xmath42 = \left(1 - 2a_i\right)p,\mathbb{e}\left[\left|s_{i}^{(c1)}\right|^2\right ] = \mathbb{e}\left[\left|s_{i}^{(c2)}\right|^2\right ] = a_ip.\nonumber\end{aligned}\ ] ] each bs @xmath8 transmits a total signal @xmath43 . by applying superposition coding ,
this signal consists of the private and common part for its primary user @xmath44 and the common parts for all its secondary users @xmath45 , @xmath46 .
@xmath47 the bs signals propagate through the wireless link to reach the users .
this process degrades the signal power of bs @xmath48 received at the location of user @xmath20 , by a factor which depends on the distance @xmath49 and by the power of a * complex valued * random fading component @xmath50 ( @xmath51 is the _ imaginary number _ , not to be confused with the _ index _
@xmath52 used always as subscript ) .
the fading power is an independent realization of a unit - mean exponential random variable @xmath53 and the phase is an independent realization of a uniform random variable on @xmath54 .
we denote the total gain from the first ( resp .
second ) neighbour @xmath21 ( @xmath22 ) to user @xmath20 by @xmath55 ( @xmath56 ) and the total gain from the first ( resp .
second ) neighbour @xmath57 ( @xmath58 ) of some other user @xmath59 , to user @xmath20 by @xmath60 ( @xmath61 ) , with @xmath62 .
the related channel phases are @xmath63 ( @xmath64 ) and @xmath65 ( @xmath66 ) .
the total signal received at user @xmath20 is @xmath67 where @xmath68 . the noise is a realization of the r.v .
@xmath69 , which follows the normal distribution . in the above the signal sum over @xmath70 is the interference received by user @xmath20 .
the beneficial signal received at the user location is equal to @xmath71 and a similar term with the adequate indexing appears for each interference term due to user @xmath72 .
the term with the @xmath73 is an extra term which is related to the phases of the channels from the first and second neighbour and can be positive or negative depending on the phase difference . by controlling this term we can maximize the received beneficial signal . _
if the phase @xmath64 is known and communicated to the first neighbour _ , the latter can choose to transmit with @xmath74 .
as a result the extra term is maximized , since @xmath75 .
the same action can not be applied to the interference terms .
the control affects only the primary user of each bs .
the emitted signal for some user is interference for some other with a random fading phase in @xmath54 .
the expected value of the interference terms is @xmath76=0 $ ] and the extra term disappears in expectation . with the above observations the @xmath77 takes the form @xmath78 @xmath79
in the above @xmath77 expressions , the vector of power - split ratios @xmath80 is not predifined and it gives a continuous range of cooperation possibilities .
however , taking the first derivative of the beneficial signal in ( [ sinrui2a2 ] ) over @xmath81 , we can find its maximizers by equating to zero , in other words @xmath82 .
the two maximizers are either @xmath83 or @xmath84 , depending on the value of the ratio @xmath85 .
we consider that these are the optimal cases for power - splitting when the choices of other users @xmath86 are held fixed . having this in mind
, we will focus in our work on policies which switch between the two extreme cases .
the relevant @xmath77 expressions below , result by appropriate substitution in ( [ sinrui2a1])-([sinrui2a4 ] ) . * * no cooperation * ( @xmath4 ) for @xmath87 @xmath88 * * full cooperation * ( @xmath3 ) for @xmath89 @xmath90 we identified above that the criterion for choosing between the two cases is the value of the channel power ratio @xmath85 .
however , it is usually better for network performance not to change very often the decision on which cooperative clusters serve users .
hence , the ratio of power path losses @xmath91 and consequently the inverse ratio of distances @xmath92 can better characterize the optimal choice .
[ fig : voronoireg ] [ policygeomdef ] we consider in this work the family of policies with * global * parameter @xmath93 $ ] such that @xmath94 the policies are * user - defined * and * geometric * because the choice to cooperate or not depends on the relative position of each user @xmath20 to its two closest bss .
the ratio @xmath93 $ ] defines the planar cooperation regions , has a global value for the network and is left as optimization variable .
interestingly , the geometric locus of planar points with @xmath95 for @xmath96 , where the policy chooses @xmath4 is a * disc*. for @xmath97 the locus degenerates to the line containing the 1-voronoi boundary of the two cells .
the two extreme values of @xmath2 give : * @xmath3 everywhere on the plane when @xmath98 .
* @xmath4 everywhere on the plane when @xmath97 .
[ fig : examplecoop04 ] and fig .
[ fig : examplecoop09 ] illustrate two examples for the geometric locus of points of @xmath3 , given a realization of bs positions .
the locus consists of all planar points for which @xmath99 , in other words the subregion of the 2-voronoi cell which lies outside the @xmath4 discs .
the larger the value of @xmath2 , the `` thinner '' the cooperation region .
we formulate the problem within the framework of stochastic geometry @xcite .
we focus on a random location on the plane and set its coordinates as the cartesian origin @xmath100 .
we assume that a user is positioned at this point , whom we denote by @xmath101 . the first ( resp .
second ) closest bs is @xmath102 ( @xmath103 ) with distance @xmath104 ( @xmath105 ) from the typical location and gains @xmath106 ( @xmath107 ) , @xmath108 ( @xmath109 ) .
furthermore , @xmath110 , @xmath111 , @xmath112 and @xmath113 are realizations of independent exponential r.v.s with mean @xmath40 ( the individual per - user power ) @xmath114 , @xmath115 , @xmath116 .
consequently , @xmath117 follows the rayleigh distribution .
we aim at deriving the * coverage probability * of the cooperating system @xmath118 , \label{covdef}\end{aligned}\ ] ] where the @xmath77 is measured at a * typical location * of the plane .
notice that it would be more natural to look at the sinr at the location of a * typical user * and that the two definitions do not coincide here because the point process of bss and that of users are not independent .
the coverage is a function of the threshold @xmath119 , as well as @xmath5 , @xmath120 , @xmath40 and the policy parameter @xmath2 . from here on , the parameter set @xmath121 , which does not influence the analysis , is omitted from the arguments .
the @xmath77 for the typical location can now be written using the policies in def .
[ policygeomdef ] and the expressions ( [ noc ] ) and ( [ fc ] ) .
it takes the form ( [ sinro2 ] ) .
the paragraphs to follow build up the necessary analytical blocks to calculate ( [ covdef ] ) . given a poisson p.p . of intensity @xmath5 , the joint p.d.f .
of the distances @xmath122 between the typical location @xmath101 and its first and second closest neighbour , equals @xmath123 the expected value of the distance @xmath124 equals @xmath125 = \frac{3}{4\sqrt{\lambda}}$ ] .
[ lemjointd ] using the p.d.f . above and
the geometric policies , it can be shown that the parameter @xmath93 $ ] fully determines the probability of a randomly positioned user within some 1-voronoi cell , to choose @xmath4 as action .
[ lemrho ] the probability of a user not to demand for bs cooperation for its service is equal to @xmath126 = \mathbb{p}\left[r_1\leq \rho r_2\right ] = \rho^2.\end{aligned}\ ] ] we derive the general expression for the r.v .
distribution of the channel fading , when @xmath3 is applied .
[ lemz ] given the r.v.s @xmath127 , @xmath116 , the laplace transform ( lt ) of the r.v .
@xmath128 is equal to the expression in ( [ ltz ] ) , where @xmath129 and @xmath130 for @xmath116 .
the p.d.f . is square integrable and its expectation is equal to + @xmath131 = \frac{1}{\sqrt{\mu_1\mu_2}}\left(\pi/2 + \frac{\mu_1+\mu_2}{\sqrt{\mu_1\mu_2 } } \right)$ ] .
we further provide an interesting property of the r.v .
@xmath132 of the fading for @xmath3 with relation to the r.v .
@xmath53 , which is the case for @xmath4 .
we use the notion of the _ laplace - stieltjes transform ordering _ based on which , the r.v .
@xmath133 dominates the r.v .
@xmath134 and we write @xmath135 , if @xmath136 , for all @xmath137 .
[ laplaceord ] given @xmath138 and the two r.v.s @xmath139 and @xmath132 from ( [ rvz ] ) , it holds @xmath140 .
@xmath141 we describe the interference @xmath142 as a shot - noise field ( * ? ? ?
* ch.2 ) generated by a point process outside a ball of radius @xmath124 .
we consider all the power splitting decisions of primary users @xmath59 ( either @xmath143 or @xmath144 ) related to bss with distance @xmath145 from the origin .
the decisions are determined by the value of the global parameter @xmath2 of the geometric policies .
the interference received at @xmath101 is equal to @xmath146 we associate each bs with a r.v .
@xmath147 such that ( from lemma [ lemrho ] ) @xmath148 this r.v .
models the randomness of user position within each 1-voronoi cell , which further determines the action chosen for this user , depending on the ratio @xmath149 . *
if @xmath150 , an independent mark @xmath151 is associated with the bs .
the mark is equal to @xmath152 .
the signal has to traverse a distance of @xmath153 from the closest neighbour of user @xmath59 . * if @xmath154 , an independent mark @xmath155 is associated with the bs .
this is the case of full cooperation , where the interfering signal due to user @xmath59 is coherently emitted from its two closest neighbours . here
, we make the * far field approximation * @xmath156 , so that the distances of the two cooperating atoms to the typical location are treated as equal .
based on this approximation , bss with primary users requiring @xmath3 , are considered to emit the entire signal @xmath157 , @xmath158 .
@xmath159 related to the mark @xmath151 follows the exponential - or equivalently @xmath160 distribution with expected value @xmath40 , whereas the r.v .
@xmath161 related to @xmath155 follows the @xmath162 , with the same expected value @xmath40 . in other words ,
the path - loss of the interfering signals is in expectation equal in both cases .
the interference r.v .
is given by @xmath163 the first two terms come from the interference created by the second neighbour lying on the boundary of the ball @xmath164 .
[ thlapli ] the lt of the interference r.v . for the model under study , with exponential fast - fading power , is equal to @xmath165 where @xmath166 the expected value for the interference r.v .
is equal to @xmath167 = \frac{p}{\left(\beta-2\right)r_2^{\beta}}\left(\beta-2 + 2\pi\lambda r_2 ^ 2\right),\end{aligned}\ ] ] and is independent of the parameter @xmath2 .
the second geographic bs neighbour is the dominant factor of interference , due to its proximity to the typical location .
a mark ( either @xmath168 or @xmath169 ) will be strong at the point @xmath101 .
in this subsection we will consider an ideal scenario , where the interference created by the second closest bs can be cancelled out perfectly in the case of full cooperation , by means of coding .
this requires precise * knowledge * by the first neighbour of the interfering signal from the primary user and all possible secondary users served by @xmath27 , which is extra information for the system .
if such information is available , the encoding procedure for the typical location can be projected on the signal space of @xmath27 , achievable by dirty paper coding ( dpc , see @xcite ) so that the effect of @xmath27 on interference is eliminated .
it the @xmath77 expression ( [ sinro2 ] ) we can then substitute the variable @xmath170 by @xmath171 , for the case of @xmath3 .
if @xmath101 chooses @xmath4 , the elimination is not possible .
the new r.v .
is derived by just omitting the interference part from the second closest bs neighbour .
its lt is equal to @xmath172 where @xmath173 is given in ( [ eachid ] ) .
the expected value of @xmath171 can be shown to be less than @xmath174 $ ] in ( [ exi ] ) .
[ covprob ] for the cooperation scenario under study and for a given set of system values @xmath175 , the coverage probability of a typical location as a function of the parameter @xmath93 $ ] is equal to ( [ tintall ] ) . in this expression , @xmath176 is given in ( [ ltinterfd ] ) , ( [ eachid ] ) and @xmath177 in ( [ ltz ] ) . for the case of dirty paper coding , @xmath176 should be replaced by @xmath178 given in ( [ ltinterfdcancel ] ) .
we have numerically evaluated the integrals in ( [ tintall ] ) for parameter density @xmath179 , path - loss exponent @xmath180 , per - user power @xmath181 and noise @xmath182 .
the two cases without and with dpc are shown in fig .
[ fig : nodpcf ] and fig .
[ fig : dpc ] respectively .
both figures also include plots from simulations , so that the validity of our approximations can be guaranteed .
the simulation results show the coverage probability , taken as an average of 10,000 realizations of different random bs topologies with expected number of atoms @xmath183=20 $ ] , uniformly positioned in an area of @xmath184 . in each figure , three curves are produced for the numerical evaluation and three for the simulations : ( a ) coverage with @xmath4 everywhere @xmath97 , ( b ) coverage with @xmath3 everywhere @xmath98 , ( c ) coverage with optimal @xmath185 . in ( c )
the parameter @xmath2 is optimally chosen , so that the coverage @xmath186 is maximized .
[ fig : nodpcf ] illustrates that cooperation with some @xmath187 can be optimal for @xmath188 , whereas @xmath189 ( @xmath4 everywhere ) is always optimal for @xmath190 .
the maximum gain in coverage with cooperation is @xmath191 between @xmath192 to @xmath193 .
the case of dpc for the interference from @xmath27 gives more substantial coverage gains , as shown in fig .
[ fig : dpc ] .
the gains appear in the entire domain of @xmath119 and reach a maximum of @xmath194 for @xmath195 .
the figures illustrate the fact that cooperation becomes more beneficial to the network when more information is exploited .
another important quantitative benefit , not visible in the total coverage probability evaluation we show here , is that the coverage area shapes change in favor of cell - edge users while @xmath2 reduces from @xmath196 to @xmath197 .
in the present work , we evaluated the coverage probability of a cellular network , when bss can cooperate pairwise for the service of each user . for this
we have applied tools from stochastic geometry .
cooperation is understood here as coherent transmission of a common message to the user from its two closest bss .
this message is exchanged between them by conferencing , together with the value of the channel phase from the second neighbour .
the user can choose between @xmath3 or @xmath4 based on the suggested geometric policies .
closed form expressions for the coverage probability have been derived , whose evaluation quantifies the benefits resulting from cooperation .
the benefits are larger when dpc , to avoid second neighbour interference , is applied . | we study a cooperation model where the positions of base stations follow a poisson point process distribution and where voronoi cells define the planar areas associated with them .
for the service of each user , either one or two base stations are involved .
if two , these cooperate by exchange of user data and reduced channel information ( channel phase , second neighbour interference ) with conferencing over some backhaul link .
the total user transmission power is split between them and a common message is encoded , which is coherently transmitted by the stations . the decision for a user to choose service with or without cooperation
is directed by a family of geometric policies .
the suggested policies further control the shape of coverage contours in favor of cell - edge areas .
analytic expressions based on stochastic geometry are derived for the coverage probability in the network .
their numerical evaluation shows benefits from cooperation , which are enhanced when dirty paper coding is applied to eliminate the second neighbour interference . |
neutron stars ( ns ) provide us with opportunities to probe the properties of matter at extremely high densities , and have proven to be fantastic test bodies for theories involving general relativity . in a broader perspective ns provide access to the phase diagram of matter at extreme densities and temperatures , which is the basic for understanding very early universe and several other astro - physical phenomena .
the observational quantities of primary astrophysical interest are the maximum mass and the typical radius of a ns .
neuron stars are detected as pulsar+ns or pulsar+white - dwarf or x - ray binaries .
recent observations of pulsars and x - ray binaries suggest that the maximum mass of a ns lies between @xmath0 @xcite , where @xmath1 is the solar mass . to understand the observables of ns , various theoretical models
have been developed and can be grouped into three broad categories @xcite : nonrelativistic potential models @xcite , relativistic field theoretical models @xcite and dirac - brueckner - hartree - fock models @xcite . in each of these models , the addition of hyperon @xcite or kaon or pion @xcite or quarks @xcite or their combinations , will soften the equation of state ( eos ) and hence lower the maximum mass of ns . in all of these models , coupling constants and unknown meson masses
are treated as effective parameters adjusted to fit the empirical quantities ( saturation density @xmath2 , binding energy @xmath3 , compression modulus @xmath4 , effective nucleon mass @xmath5 and asymmetry energy @xmath6 ) at nuclear saturation .
the compression modulus ( @xmath7 defines the curvature of the eos at saturation density and its value will be reflected in the high density behaviour ( stiffness or softness ) of the eos .
thus @xmath4 will have a direct bearing on the maximum mass .
interestingly , the study of isoscalar giant resonances can reveal rich information about nuclear compressibility @xcite .
the eos and hence the ns radius can directly be linked to the neutron skin thickness in heavy nuclei @xcite .
these strong links between the ns structure and finite nuclear properties prove that the knowledge in these two areas complement each other .
one among the very - well tested model in finite nuclei is the nonlinear @xmath8 model , widely mentioned as relativistic mean field ( rmf ) model @xcite .
the rmf model constructed with idea of renormalizability , produces accurate results for spherical and deformed nuclei but it gives a very stiff eos for infinite nuclear matter .
recently , inspired by effective field theory ( eft ) furnstahl , serot and tang @xcite abandoned the idea of renormalizability and extended the rmf theory , with the systematic inclusion of new interactions by adding new terms to the model lagrangian . this effective field theory motivated rmf ( e - rmf ) model calculations explained finite nuclei and nuclear matter with exactly same parameters and with significant accuracy in both the cases @xcite .
this approach can be considered as a salient step towards a unified and accurate theory for finite nuclei as well as for infinite nuclear matter .
extension of this model with the inclusion of the kaon ( @xmath9 ) is the central interest of present work . in ns , @xmath9-condensation is one of the several possible transitions that could exist at high density . as the density of ns increases
, the chemical potential of the negative electric charge ( @xmath10 ) also increases with the same rate of proton number density .
simultaneously , the effective mass of an in - medium @xmath9 will decrease , due to the attractive interaction between @xmath9 and nuclear matter .
therefore at a particular density ( when @xmath10 is greater than energy of @xmath9 ) , @xmath9 replaces electron as neutralizing agent in charge - neutral matter . after the demonstration of interaction of @xmath9 with the nuclear medium by kaplan and nelson @xcite , this topic acquired enormous interest . glendenning and schaffner - bielich @xcite , explained the interaction of @xmath9 by coupling them to meson fields using a minimal coupling
this approach was followed by several other theoretical groups using rmf models @xcite and has been adopted in this work as well . in the present work we investigate the condensation from non - kaonic to kaonic phase , with different lagrangians . in the presence of such a transition
, we study the effect of higher order couplings on the eos and hence on the ns properties .
this work differs from many previous works in the sense that all the parameters used in e - rmf model were fitted to reproduce the observables of finite nuclei .
also , the e - rmf model is well tested throughout the nuclear chart by explaining several nuclear properties @xcite .
we show in this paper that this model explains the recent observations of ns as well . in the section following this introduction
, we describe the lagrangian and field equations , both for non - kaonic and kaonic phases .
this is followed by the expressions for energy density and pressure , which define the eos . in section [ section : parameters ] we discuss the parameters used in our calculation .
our results and discussions are presented in section [ section : results ] which is followed by the summary along with the conclusions drawn from present work .
in this section we briefly sketch the e - rmf model by presenting the model lagrangian @xcite .
we then show how the physical quantities , that will determine the composition of the ns , can be obtained self consistently .
the effective lagrangian , obtained by curtailing terms irrelevant to nuclear matter in the e - rmf lagrangian , can be written as @xmath11\psi \nonumber\\&&+\frac{1}{2}\left(1+\eta_{1}\frac{g_\sigma\sigma}{m_n}+\frac{\eta_2}{2}\frac{g_{\sigma}^{2}\sigma^2}{m_n^2}\right)m_{\omega}^{2}v_\mu v^{\mu}\nonumber\\&&+\frac{1}{4!}\zeta_0g_{\omega}^{2}(v_\mu v^\mu)^2+\left(1+\eta_\rho\frac{g_\sigma \sigma}{m_n}\right)m_{\rho}^{2}\text{tr } ( r_\mu r^{\mu})\nonumber\\&&-m_{\sigma}^{2}\sigma^2\left(\frac{1}{2}+\frac{\kappa_3g_\sigma\sigma}{3!m_n}+\frac{\kappa_4g_{\sigma}^{2}\sigma^2}{4!m_n^2}\right),\end{aligned}\ ] ] where the scalar , vector and isovector meson fields and the nucleon field are denoted by @xmath12 , @xmath13 , @xmath14 and @xmath15 respectively . @xmath16 and @xmath17 are the corresponding meson masses and @xmath18 is the nucleon mass .
the symbols @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 and @xmath27 denote the various coupling constants .
more details of the lagrangian are explained explicitly in ref .
@xcite .
now we extend the e - rmf model by including a kaon ( @xmath9)-nucleon interaction term . for simplicity ,
we choose that the kaon is coupled to the meson field with minimal coupling @xcite . in this way ,
interactions of mesons with nucleons and kaons are treated in the same footing .
the lagrangian for the kaon part reads @xmath28 where the vector fields are coupled to kaons via the relation@xmath29 and @xmath30 is the effective mass of kaon .
the scalar field is coupled to kaons in a way analogous to the minimal coupling scheme @xcite of the vector fields : @xmath31 where @xmath32 mev .
note that in the mean field approximation , only the time components of the vector fields @xmath33 contribute and charge conservation implies that only the third component in isospin - space of the isovector meson field @xmath34 does not vanish .
the dispersion relation for @xmath35-wave condensation @xmath36 , for @xmath9 is @xmath37 with @xmath9 mesons having isospin projection @xmath38 .
@xmath39 represents the @xmath9 energy and is linear in the meson field .
one can note that with the increase in density , @xmath39 decreases .
in the presence of @xmath9 , equations of motion for the meson fields are @xmath40 @xmath41 @xmath42 where @xmath43 is the scalar density given by @xmath44and @xmath45 is the spin - isospin degeneracy factor and is equal to 2 ( for spin up and spin down ) .
the nucleon effective mass is defined as in the standard walecka model @xmath46 and the @xmath9 density is given by @xmath47 the proton and neutron chemical potentials are @xmath48 for a ns , in the absence of neutrino trapping , the conservation of baryon and electron chemical potentials leads to @xcite @xmath49 and @xmath50 where the first two constraints ensure the chemical equilibrium and the last one specifies the total charge which vanishes while imposing the charge neutrality .
the total baryon density is @xmath51 c d d d d d & & & & & + @xmath52 ( mev ) & 938 & 939 & 939 & 939 & 939 + @xmath53 ( mev ) & 492.25 & 508.194 & 507.06 & 520.206 & 491.5 + @xmath54 ( mev ) & 795.36 & 782.501 & 782 & 782 & 783 + @xmath17 ( mev ) & 763 & 763 & 770 & 770 & 763 + @xmath19 & 10.0730 & 10.1756 & 9.8749 & 10.5088 & 10.5924 + @xmath20 & 13.1917 & 12.7885 & 12.1270 & 12.7864 & 14.3020 + @xmath21 & 9.8553 & 8.9849&8.7886&9.5108&11.7673 + @xmath22 & 1.8324&1.4841&2.2075&3.2376&0.6194 + @xmath23 & -7.7099&-5.6596&-10.0808&0.6939&9.7466 + @xmath24&0&0&0.071&0.65&0 + @xmath25&0&0&-0.962&0.11&0 + @xmath26&0&0&-0.272&0.390&0 + @xmath27&0&0&3.5249&2.642&0.06 + @xmath55&0&0&0&0&0.03 + + @xmath2 ( @xmath56)&0.154&0.148&0.153&0.153&0.148 + ( e / a)(mev)&-16.43&-16.299&-16.14&-16.07&-16.3 + @xmath4(mev)&212&271.76&215&215&230 + @xmath6(mev)&43.6&37.4&38.5&36.4&32.59 + @xmath57&0.571&0.6&0.634&0.664&0.609 + the energy density can be written as @xmath58 where @xmath59 is the energy density due to nucleons given by @xmath60 and the energy density contributed by kaons is @xmath61 unlike the energy density , pressure is not directly affected by the inclusion of @xmath9 , but the inclusion of @xmath9 affects the fields and hence the pressure which reads @xmath62 here @xmath63 stands for the leptons @xmath64 . in the non - kaonic phase , vanishing charge density implies @xmath69 with @xmath70 .
we can calculate @xmath12 , @xmath33 , @xmath34 , @xmath71 , @xmath72 , @xmath73 , and @xmath74 by using eqs .
( [ eq : fields ] ) , ( [ eq : beta_eq ] ) , ( [ eq : cd ] ) and ( [ eq : den ] ) , at the chosen baryon density .
when we get the converged solution for the above - listed quantities , the energy density and pressure can be computed from eqs .
( [ eq : ed ] ) and ( [ eq : pre ] ) . with the solution of non - kaonic phase in hand , from eq .
( [ eq : ke ] ) we can calculate kaon energy which keeps decreasing as we increase density , while @xmath10 increases .
when the condition @xmath75 is first achieved , the kaon will occupy a small fraction of the total volume and the charge density corresponding to kaonic phase , @xmath76 we can calculate @xmath12 , @xmath33 , @xmath34 , @xmath71 , @xmath72 , @xmath73 , @xmath74 and @xmath77 by using eqs .
( [ eq : fields ] ) , ( [ eq : beta_eq ] ) ( [ eq : cd ] ) and ( [ eq : den ] ) with the condition @xmath75 , for any chosen baryon density . after getting these solutions
, we can calculate energy density and pressure for kaonic phase , using eqs .
( [ eq : ed ] ) and ( [ eq : pre ] ) .
in the present work we have chosen five sets of parameters namely nl1 @xcite , nl3 @xcite , g1 , g2 @xcite and fsugold @xcite .
the first two parameter sets correspond to the standard rmf model with very different compressibilities [ @xmath4(nl1 ) = 212 mev and @xmath4(nl3 ) = 271.76 mev ] . in the case of fsugold , there are two more coupling parameters @xmath78 and @xmath27 in comparison to nl1 and nl3 parameter sets .
these parameters represent the strength of the self - interaction of vector field ( @xmath27 ) and the isoscalar -isovector mixing ( @xmath55 ) . the interactions g1 and g2 correspond to the complete e - rmf lagrangian discussed in the earlier section .
one can note that in comparison to nl1 and nl3 , g1 and g2 have four more coupling constants ( @xmath79 and @xmath27 ) .
all the coupling constants are obtained by fit to several properties of finite nuclei @xcite .
it is worth to mention that in this case no parameter is treated as adjustable and the fitting does not involve any observation at densities above the saturation value . since the expectation value of the @xmath34 field is typically an order of magnitude smaller than that of the @xmath33 field , in the e - rmf model , the nonlinear @xmath34 couplings were retained only through third order @xcite .
however , it has been shown that the isoscalar - isovector mixing ( @xmath80 ) is useful to modify the neutron radius in heavy nuclei while making very small changes to the proton radius and the binding energy @xcite . in the effective lagrangian approach adopted here , knowledge of two distinct sets of coupling constants
one parametrizing the nucleon - nucleon interaction and one parametrizing the kaon - nucleon interactions is required for numerical computations .
we discuss each of these in turn .
the symbols @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 , @xmath27 and @xmath55 denote the nucleon coupling constants . at times
these constants bear different values in literature for the same parameter set .
for example , for the set nl1 different numbers are quoted in refs .
@xcite and @xcite but they yield same saturation properties . among the above - mentioned coupling constants ,
@xmath19 , @xmath20 , @xmath21 , @xmath22 and @xmath23 can be written algebraically in terms of empirical quantities : @xmath2 , @xmath3 , @xmath4 , @xmath6 , @xmath81 and vice versa @xcite . using these relations
we have calculated @xmath19 , @xmath20 , @xmath21 , @xmath22 and @xmath23 in the case of nl1 @xcite , nl3 @xcite , g1 and g2 @xcite .
these calculated parameters , and other parameters taken from corresponding references are listed in table [ nucleon ] . in case of fsugold
, all the coupling constants are taken from ref .
@xcite . in order to investigate the effect of kaons on the high density matter
, the kaon - nucleon coupling constant has to be specified .
the laboratory experiments give information only about kaon - nucleon interaction in free space . in this work we mainly focus on the ns ( densities @xmath82 ) , and therefore the kaon - nucleon interaction determined from experiment need
not be appropriate for our calculations .
the interaction of omega and rho mesons with kaon ( @xmath83 and @xmath84 ) can be determined using a simple quark and isospin counting argument @xcite given by , @xmath85 we can specify the interaction of sigma meson with kaon ( @xmath86 ) using its relation with the optical potential of a single kaon in infinite matter ( @xmath87 ) : @xmath88 where typically we have @xmath89 mev @xmath90 mev @xcite .
the quality of different interactions in explaining symmetric matter and pure neutron matter properties at higher densities than the normal nuclear densities has been discussed in refs .
@xcite , by comparing the calculated pressure with the experimental data @xcite .
this comparison clearly suggests a softer eos for symmetric matter , which could be obtained only with higher order couplings in the lagrangian .
the pure neutron matter data , though model dependent , also was found to favour a softer eos .
one important observation highlighted in ref .
@xcite , is that the softness in eos from g1 , g2 and fsugold is mainly due to the new couplings and not due to the difference in compressibility as usually perceived .
for example , the compressibility from nl1 ( @xmath4=212 ) and g1 , g2 ( @xmath4=215 ) are very similar but the eos at higher densities are completely different .
nl1 ( @xmath4=212 ) and nl3 ( @xmath4=271.76 ) yield different compressibilities but their eos are more similar .
the role of higher order couplings in softening the eos has been discussed in ref .
@xcite for the case of fsugold .
eos from fsugold is softer than that of g2 because of the large and positive @xmath23 value as well as the introduction of isoscalar - isovector coupling ( @xmath55 ) .
in fact it has been clearly demonstrated @xcite that @xmath55 softens the symmetry energy considerably .
the success of g1 and g2 in explaining the high density eos constrained by phenomenological flow analysis @xcite , has been carried forward in explaining ns properties as well @xcite .
fsugold also has been successful in explaining both these features @xcite .
henceforth we discuss the role of new parameters of extended rmf models on the onset and effect of kaon condensation in ns . among the different parameter sets considered here ( table [ nucleon ] )
g2 and fsugold have a positive quartic scalar self - coupling ( @xmath23 ) , which is more meaningful than a negative one @xcite .
for this reason we prefer the set g2 , when we have to choose between g1 and g2 .
energy ( @xmath39 ) in ns matter for different optical potentials ( @xmath87 in mev ) calculated with g2 parameter set .
the point at which electron chemical potential ( @xmath10 ) intersects @xmath39 defines the onset of @xmath9-condensation . ] , vector @xmath91 , and iso - vector @xmath92 fields in the ns matter inclusive of kaonic phase , calculated with g2 parameter set . ] in fig .
[ kaonen ] , we have presented the @xmath9 energy ( @xmath39 ) for different optical potentials , as a function of density along with the variation of electron chemical potential ( @xmath10 ) . the @xmath39 decreases with increasing density [ eq . ( [ eq : ke ] ) ] while @xmath10 increases . when @xmath39 is lower than @xmath10 , @xmath9 are favoured ( due to attraction between @xmath9 and nucleon ) to replace electrons while contributing to the charge neutrality . from fig .
[ kaonen ] , it is evident that the onset of @xmath9-condensation is strongly modified by the strength of the kaon optical potential ( @xmath87 ) .
calculations based on chiral models @xcite suggest that a value @xmath93 mev is more appropriate for ns . hence in most of the further discussions we choose @xmath93 mev and
look at the other dependencies for the eos of ns .
the @xmath12 , @xmath94 , and @xmath95 fields calculated for ns at @xmath93 mev are presented in fig .
[ fieldsg2 ] along with the @xmath10 . the density at which @xmath9 starts to contribute can be read from the point where the @xmath10 shows a sharp kink ( @xmath96 0.6 @xmath56 ) .
though the corresponding kinks in the fields are very subtle , any small change in the field results in a significant change in energy density and pressure .
the presence of @xmath9 alters the proton - neutron ratio and hence the contribution from the @xmath95 field is enhanced .
this is due to the fact that the processes like @xmath97 are more energetically favoured than processes like @xmath98 at higher densities where @xmath9 energy is decreasing and @xmath10 is increasing [ see fig .
[ kaonen ] ] . from fig .
[ fieldsg2 ] we can also see that the @xmath9 has very less influence on the @xmath12 field .
in other words , for the parameters considered here , the attraction caused by @xmath9 is weak in comparison with that of @xmath12 mesons .
the appearance of @xmath9 at higher densities and its role in population ( number density ) of different particles are presented in fig .
[ densityg2 ] .
as soon as @xmath9-condensation sets in , the population of @xmath9 rapidly increases with density and hence allows the presence of more protons .
interestingly , at very high densities the number of protons exceed that of the neutrons , which is not the case if @xmath9 were not present .
this can be seen from the corresponding deviation from the non - kaonic matter trends .
apart from the arguments given earlier , the reason for @xmath66-@xmath9 pairs being preferred to neutrons is quantitatively very well explained by glendenning and schaffner - bielich @xcite . however , the symmetry term in the energy obviously prefer a symmetric matter and hence hinders the protons to be populated much more than neutrons .
all the above - mentioned effects have a significant role to play in determining the eos which is discussed in the following text .
our results for the pressure calculated with g2 parameter set and with different optical potentials for @xmath9 are presented in fig .
[ pressu ] . on first sight
one can appreciate the strong influence of @xmath87 in softening the eos , which is similar to the results of many previous works . also for lower values of @xmath87
, the graph suggests a second order phase transition from a non - kaonic to kaonic phase . only with a very high value of @xmath99160 mev )
one can have a first order phase transition .
this is rather consistent with the observation in ref .
@xcite where higher order interactions are included using the lowest order chiral lagrangian .
however , the role of more general higher order operators was considered to be an open question which is answered in the present work .
one may expect some changes in the onset of @xmath9-condensation if we consider the presence of a mixed phase ( of kaonic and non - kaonic phases ) @xcite .
the presence of such a mixed phase is normally possible in first order @xmath9-condensation where one gets a dip in the pressure like the one for @xmath100 mev as shown in fig .
[ pressu ] .
this argument implies that the presence of mixed phase is not favoured in our case for @xmath101 mev .
the main observation from fig .
[ pressu ] is that even in the presence of higher order couplings , the presence of @xmath9 dramatically softens the eos and the softness is proportional to @xmath87 . ) and @xmath9 energy ( @xmath39 ) for different parameters .
the point at which @xmath10 intersects @xmath39 defines the onset of @xmath9-condensation . ] .
solid squares ( @xmath102 ) and open triangles ( @xmath103 ) represent the observational constraints @xcite , where @xmath104 is the photospheric radius .
the shaded region correspond to the recent observation of @xmath105 star @xcite . ] and @xmath87 are in mev .
the solid circles represent the maximum mass in every case . ]
so far we have discussed our results with the g2 parameter set and in this section we analyze how our results compare with different interactions considered in this work . in fig .
[ cp ] we show the chemical potentials which indicate the onset of @xmath9-condensation for different parameters .
the maximum value of @xmath106 denotes the point at which the transition happens and is seen to be larger for g1 and g2 .
@xmath107 is almost same for nl1 and fsugold and the least value is for nl3 .
@xmath107 also quantifies the number density of electrons which falls sharply when @xmath9 starts to appear [ see fig . [ densityg2 ] ] .
both @xmath107 and the @xmath108 values at which the transition happens @xmath109 , strongly depend on the interaction .
@xmath110 for g2 and fsugold are more closer resulting from a closer eos . a sharper kink in fig .
[ cp ] , leading to a situation where there could be more than one value of @xmath10 for a given @xmath108 , would have indicated a first order phase transition .
it is clear that for @xmath93 mev , all the interactions lead only to a second order phase transition where a mixed phase can not appear . in fig .
[ density_g1 ] we compare the number densities arising from the calculations with g1 and fsugold , which can be compared with fig . [ densityg2 ] as well .
the results for g1 are very similar to that of g2 except for the early onset of @xmath9-condensation in g1 . between figs .
[ density_g1 ] ( a ) and ( b ) , we can see a significant difference in the relative population between protons and neutrons at higher densities .
this is mainly due to the interplay between @xmath39 and @xmath10 which are plotted in fig .
[ kaon_para ] .
the decrease in @xmath39 or increase in @xmath10 favours more of @xmath9 and hence protons .
@xmath39 varies linearly with the meson fields as given in eq .
( [ eq : ke ] ) and should reflect the stiffness of the fields and hence that of the eos .
we can see in fig .
[ kaon_para ] that the @xmath39 for nl3 is quite different from others due to the fact that the eos of nl3 is quite stiffer than others @xcite .
this stiffness arises from the vector potential which grows almost as a straight line @xcite .
the quartic vector self - interaction brings down the vector potential and makes the equation of state and hence @xmath39 softer in other cases .
the other quantity determining the onset of @xmath9-condensation is @xmath111 which has terms similar to that of the symmetry energy and is dominated by the @xmath34 field [ eq . ( [ eq : cp_np ] ) ] .
thus the density dependence of symmetry energy directly affects @xmath10 and is crucial in determining the onset of @xmath9-condensation .
nl3 , g1 and g2 yield similar symmetry energy ( not shown here ) and the unique isoscalar - isovector mixing ( @xmath80 ) in fsugold suppresses the @xmath34 field which leads to a softer symmetry energy and hence lesser @xmath10 .
if we compare between g2 and fsugold , @xmath10(g2 ) @xmath112(fsugold ) and with similar @xmath39 , kaons are more favoured for larger @xmath10 .
lesser kaons in fsugold leads to lesser number of protons at higher densities in fig .
[ density_g1 ] ( b ) . in fig .
[ kaon_para ] one can see that the @xmath10 for nl3 and g1 are almost same till the transition point .
however due to the large difference in @xmath39 , which falls very sharply in the case of nl3 , @xmath9-condensation occurs at a smaller density in nl3 when compared with g1 .
these features can be seen from the eos presented in fig .
[ pd ] as well . figure [ pd ] shows the pressure calculated with the different interactions , with and without the inclusion of @xmath9 .
similar to the case of symmetric and pure neutron matter @xcite , the higher order couplings in g1 , g2 and fsugold lead to a softer eos .
all the interactions suggest a major change in pressure when we include @xmath9 .
the well known feature of @xmath9 making the eos softer , can also be seen clearly in fig .
for both kaonic and non - kaonic phases , the difference between pressure obtained from g1 and g2 changes with density . around @xmath113 @xmath56 ,
the difference is maximum and it decreases as the density increase .
this is due to the interplay between the terms with higher order couplings ( @xmath114 ) which can give negative contribution to pressure . between g1 and g2 , the sign of the coupling constants @xmath115 and
@xmath26 differ . among these three couplings ,
the term comprising @xmath26 is weaker . in the case of g1 , due to
the change in sign of @xmath23 , the combined effect of @xmath22 and @xmath23 on pressure is negligible .
hence in g1 , @xmath25 dominates at higher densities in reducing the pressure . in case of g2 ,
the contribution from terms having @xmath23 and @xmath25 is negligible and @xmath22 dominates at higher densities in reducing the pressure .
thus in g1 and g2 , @xmath25 ( quartic scalar - vector cross - interaction ) , and @xmath22 ( cubic scalar self - interaction ) respectively , rule the suppression of pressure at very high densities .
@xmath25 being associated with the vector self - interaction is much stronger at high densities and hence will be more effective than @xmath22 in reducing the pressure .
couplings defined by @xmath25 and @xmath22 are of different order and will naturally yield results of different curvature which leads to the varying difference in pressure between g1 and g2 . the difference in pressure from g2 and fsugold
is well reduced in the presence of @xmath9 , even at higher energies .
interestingly , g2 with @xmath9 yields same pressure as that of fsugold without @xmath9 and in a broad energy range these results cease to differ
. however , the corresponding energy densities are very different ( not shown here ) and hence the resulting maximum mass ns have different central densities ( represented by solid circles in fig . [ pd ] ) .
the eos around the region of central density is more dominant in determining the properties of ns .
this region is quite different for different parameters .
also one can see that the central density increases with softer eos , and the presence of @xmath9 decreases the central density and pressure .
the change in central density ( @xmath116 ) due to @xmath9 depends mostly on the following three quantities , viz .
. density at which @xmath9-condensation sets in ( @xmath117 ) with @xmath118 , 2 . the magnitude of central density with @xmath119 , and 3 .
the stiffness of eos at @xmath120 , with stiffer eos leading to larger @xmath116 .
the interplay of the above quantities gives arise to varying @xmath116 for different parameters .
for example , between g1 and g2 , quantities ( ii ) and ( iii ) dominate ( i ) , resulting in a larger @xmath116 in case of g2 .
these effects should naturally be seen in the calculation of maximum mass and radius of ns .
we can obtain mass - radius relation for ns by solving the well - known tolman
oppenheimer volkoff ( tov ) equations @xcite .
the results for mass - radius relation in ns are given in fig .
[ mr ] where the calculations are done using different interactions , without and with the @xmath9 ( @xmath93 mev ) . the rmf models nl3 and nl1 suggest very large and massive ns . in these cases
though the maximum mass is reduced considerably ( from @xmath121 2.8 m@xmath122 to @xmath121 2.4 m@xmath122 ) in the presence of kaons , these are overshooting the observational constraints for ns @xcite .
though g1 and g2 are from the same e - rmf model with same terms in the lagrangian , their results for ns are quite different with g1 suggesting a larger and heavier ns . as explained earlier , g1 and g2 yield eos of different nature and have varying role of @xmath9 .
the result for g2 and fsugold are the closest ones . very interestingly ,
though their difference is larger when @xmath9 are not considered , the presence of @xmath9 brings these two results closer .
this is the direct implication of the same feature we observed in the case pressure ( fig .
the g2 interaction allows more @xmath9 to be present ( due to early onset of @xmath9-condensation ) and hence makes the eos more softer which eventually get closer to that of fsugold .
the observed reduction of maximum mass due to kaons , in case of nl3 and fsugold is consistent with the values quoted in ref .
overall , in the presence of @xmath9 , the mass decreases and the radius corresponding to maximum mass increases .
all the observational constraints shown in fig . [ mr ] favor the models with higher order couplings but most of these constrains cover a broad range which may not be precise enough to ascertain the presence of kaons . in fig .
[ mr_u ] , we show the dependence of ns mass - radius on @xmath87 with g2 parameter set .
we can see clearly that the stronger @xmath87 increases the contribution of @xmath9 which results in softer eos which in turn yields a smaller and lighter ns .
the typical hook shape of the mass - radius curve is maintained till @xmath123 mev and beyond that the curve takes a straight dip which is a characteristic of first order @xmath9-condensation .
it is interesting to note that without @xmath9 , the maximum mass given by g2 ( @xmath124 ) is closer to the one suggested by demorest _
( @xmath105 ) @xcite with a radius of 11.03 km .
g1 with @xmath125 mev ( graph not shown here ) suggests a maximum mass of @xmath126 with a radius of 12.45 km .
figure [ dr ] shows the number density versus the radial distance from the center of ns , calculated with g2 and fsugold parameter sets , when the ns is having the maximum mass ( @xmath127 for fsugold , and @xmath128 for g2 with @xmath93 mev ) .
the behaviour of number densities shown in fig .
[ dr ] is directly reflecting the patterns shown for relative populations plotted against density in figs .
[ densityg2 ] and [ density_g1 ] ( b ) .
the difference between the radii of kaonic matter suggested by g2 and fsugold is almost one km .
another striking difference in the composition of ns is the asymmetry of the core .
g2 suggests a nearly symmetric matter at the core whereas fsugold indicates highly neutron - rich core .
we have seen that this trend remains same even at higher values of @xmath87 ( not shown here ) .
it will be interesting to see how the presence of hyperons affect the overall scenario @xcite .
in the present work , we have studied the role of higher order couplings of extended rmf models on the onset and effect of kaon ( @xmath9 ) condensation in neutron stars ( ns ) .
we have calculated the ns properties with the successful rmf parameters nl1 and
nl3 , the e - rmf parameters g1 and g2 ( four additional couplings to rmf ) , and the fsugold parameters ( two additional couplings to rmf ) . in extended rmf models , with most common values of kaon optical potential ( @xmath129 mev )
, the transition from a non - kaonic phase to kaonic phase in ns has a character of second order which rules out the possibility of having a mixed phase of kaonic and non - kaonic matter .
this is consistent with the observation in ref .
@xcite where higher order interactions are included using the lowest order chiral lagrangian .
the role of more general higher order operators was considered to be an open question and we have shown that they lead to a second order phase transition . this justifies the neglect of mixed phase in our calculations mostly done with @xmath93 mev @xcite . apart from the usual dependence on @xmath87 reported elsewhere @xcite , the onset of kaon condensation in ns strongly depends on the parameters of lagrangian especially the higher order couplings .
this is due to the strong variation in density dependence of the @xmath9 energy ( @xmath39 ) and electron chemical potential ( @xmath10 ) whose interplay determine the onset of @xmath9-condensation .
density dependence of @xmath39 is similar to that of eos and @xmath10 varies in a way similar to symmetry energy .
so , any change in the density dependence of eos or that of symmetry energy will affect the onset as well as the effect of @xmath9-condensation . without higher order couplings , nl1 and nl3 have stiffer eos than others which leads to a stiffer @xmath39 and hence a early onset of @xmath9-condensation .
fsugold comprises an unique isoscalar - isovector mixing which leads to a softer symmetry energy and hence a delayed onset of @xmath9-condensation .
the central density of ns increases with softer eos , and the presence of @xmath9 decreases the central density and pressure .
the change in central density due to the presence of @xmath9 is different for different parameters due to the interplay between ( i ) the density defining onset of @xmath9-condensation , ( ii ) the central density ( @xmath120 ) itself , and ( iii ) the stiffness of eos at @xmath120 . due to this ,
the impact of @xmath9 is more pronounced in g2 than in g1 and is weakest in fsugold .
all these effects are strongly reflected in the calculation of mass - radius relation of ns .
the ns suggested by models without higher order couplings ( nl3 and nl1 ) contradict the observational constraints @xcite even with the inclusion of @xmath130 .
the higher order couplings play a dominant role ( than kaons ) in bringing the mass and radius of ns within observed limits .
different parameter sets lead to different concentration of kaons ( which is known to affect the population of protons @xcite ) in ns , and hence lead to different asymmetries in the core of ns .
we have shown that g1 and g2 suggest a symmetric core whereas fsugold suggests a neutron rich core due to lesser amount of @xmath9 caused by an unique coupling which softens the symmetry energy .
we conclude that the extended rmf models , which are quite successful in explaining several properties of finite nuclei , suggest a strong influence of higher order couplings and kaons in ns whose properties are within the present observational constraints @xcite . | we discuss the role of higher order couplings in conjunction with kaon condensation using recent versions of relativistic mean field models .
we focus on an interaction ( g2 ) in which all parameters are obtained by fitting finite nuclear data and successfully applied to reproduce variety of nuclear properties .
our results show that the higher order couplings play a significant role at higher densities where kaons dominate the behavior of equation of state .
we compare our results with other interactions ( nll , nl3 , g1 and fsugold ) and show that the new couplings bring down the mass of neutron star ( ns ) , which is further reduced in the presence of kaons to yield results consistent with present observational constraints .
we show that the composition of ns vary with the parameter sets . |
nanowires are a natural mode of gan growth in plasma - assisted molecular beam epitaxy , when effectively n rich conditions and sufficiently high temperatures are provided @xcite .
gan nanowires then form spontaneously on various substrates without the need of any metal droplets that are required for many other semiconductor materials to induce the vapor - liquid - solid growth of one - dimensional nanostructures @xcite .
one of the distinct advantages of the spontaneous formation and subsequent uniaxial growth of gan nanowires is the possibility to realize abrupt axial heterojunctions between different iii - n compounds by simply switching the group iii supply @xcite .
spontaneously grown gan nanowires usually form dense arrays .
the high nanowire density is advantageous for light emitting and energy harvesting devices as well as for sensing applications @xcite . on the other hand ,
gan nanowires coalesce during growth because of their proximity @xcite .
recently , we have identified the origin of this coalescence in gan nanowire ensembles @xcite .
dense ensembles of gan nanowires evolve during growth to ensembles of nanowires with smaller densities and larger diameters .
however , the fraction of the surface area covered by the nanowires remains unchanged , which rules out radial growth of nanowires as the origin of coalescence .
instead , we have found that the nanowires bundle together after reaching a certain critical height to reduce the surface energy of their side facets at the expense of the elastic energy due to their bending @xcite . in the present work ,
we show that the elastic bending energy can be reduced by plastic relaxation at the coalescence joints .
dislocations at the joints @xcite form small angle boundaries and reduce the curvature of the nanowires .
since these dislocations run across the nanowires , the strain relaxation at the free side surface restricts the range of their elastic fields to distances comparable with the nanowire diameters .
the dislocation energy is comparatively small , making the introduction of dislocations energetically favorable .
the calculation of the bundling energetics shows that a major part of the bending energy is expected to be released due to formation of dislocation arrays at the joints .
we study the residual nanowire bending by a line shape analysis of x - ray diffraction profiles .
the bent segments of the nanowires constitute a significantly larger volume than the coalescence joints themselves and manifest themselves by extended tails of the diffraction lines .
the diffraction profiles from the ensemble of bent nanowires are calculated taking into account the distributions of lengths , diameters , and distances between bundled nanowires , as well as the mutual nanowire misorientation and the divergence of the x - ray beams . with this model
, we revisit our previous x - ray diffraction measurements @xcite and show that the diffraction profiles are well described by the residual bending of nanowires , and that a major part of the initial bending is released by plastic deformation .
the x - ray diffraction intensity from a bent nanowire is a result of interference of the waves scattered along its entire length . as a result ,
the common assumption that each small volume of the sample contributes to the diffraction intensity independently and only according to its local strain state @xcite is not applicable , and the dependence of the integral breadth on the reflection order does not follow a linear law .
let us consider first the energetics of a pair of bundled nanowires that are bent to bridge their mutual separation and remain elastically strained , as shown in fig .
[ fig : sketch](a ) .
the following calculation is based on our previous study @xcite and serves as a reference for the analysis of plastic deformation at the joints in the next section .
we treat nanowires as thin rods experiencing small deflection ( ref . ,
the rod is described by the deflection amplitude @xmath0 , where @xmath1 is the coordinate along the rod .
the deflection amplitude satisfies , in the absence of a shearing force , an elastic equilibrium equation @xmath2 , where the prime denotes the derivative over the coordinate @xmath1 along the rod .
the origin is taken at the substrate surface where the bottom end of the nanowire is clamped .
hence , one has @xmath3 and @xmath4 .
let the nanowires bundle at a height @xmath5 and touch each other for @xmath6 .
then , the second pair of the boundary conditions at the junction is @xmath7 and @xmath8 . here
@xmath9 is the distance between the nanowire surfaces at @xmath10 [ see fig . [
fig : sketch](a ) ] .
the solution of the elastic equilibrium equation with these boundary conditions is @xmath11 describes the residual nanowire bending . ]
the elastic energy of the bent nanowire is equal to ( ref . , 18 ) @xmath12 where @xmath13 is the young modulus and @xmath14 is the geometrical moment of inertia of the cross section of the rod .
for a circular cylinder of radius @xmath15 , the moment of inertia is equal to @xmath16 if the cross - section of the cylinder is a regular hexagon of side length @xmath17 , the moment of inertia is @xmath18 the surface energy gained by each nanowire due to the contact of nanowires at @xmath6 is equal to @xmath19 where @xmath20 is the surface energy and @xmath21 is the width of the contact area . for hexagonal nanowires , @xmath21 is equal to the side length @xmath17 , while for a circular cylinder it can be taken equal to the radius @xmath15 . the nanowires bundle if the total energy @xmath22 reduces .
our aim now is to find , for a given distance @xmath9 between nanowire sidewalls , the critical nanowire length @xmath23 such that the bundling is energetically favorable for @xmath24 .
the minimum of the total energy is given by the condition @xmath25 , and energy is gained by bundling if , at the minimum , @xmath26 .
the critical value @xmath27 is thus determined from the simultaneous solution of the equations @xmath28 and @xmath29 .
using expressions ( [ eq:2 ] ) and ( [ eq:4 ] ) for the energies and dividing one equation by the other , we find @xmath30 . hence ,
at the moment when the bundling becomes energetically favorable , one quarter of the total nanowire lengths merge together . further straightforward calculation yields @xmath31 where @xmath32 . with the young modulus of gan @xmath33 j / m@xmath34 and a surface energy of @xmath35 mev/@xmath36 j / m@xmath37 for the _ m_-plane facets that constitute the gan nanowire sidewalls @xcite , we get @xmath38 nm . for an estimate , consider two hexagonal nanowires with side facet width @xmath39 nm at a distance @xmath40 nm between their side facets . the distance between the nanowire centers is then @xmath41 nm , which corresponds to the mean distance between nanowires for a density of 4@xmath42 @xmath43 , the typical density of nucleated nanowires before the bundling starts @xcite . the critical height given by eq .
( [ eq:5 ] ) is @xmath44 nm . the total length of the nanowires is @xmath45 nm . the elastic energy ( [ eq:2 ] ) stored in each of the bent pillars of the critical height ( @xmath46 is equal to @xmath47 for the example above , this energy is @xmath48 j. bundeled nanowires may experience subsequent bundling with other nanowires in their vicinity .
the energetics of this process can , in principle , be described in the same way as above .
however , the bending energy of nanowire bundles staying on several pillars can only be evaluated numerically . in our previous work @xcite , we have approximated this situation by replacing the bundled pair by a single cylindrical nanowire with a volume and cross - sectional area equal to the respective sums of the constituents .
the energy of bending can be partially or completely released if dislocations are present at the joint and form a small angle boundary , as shown in fig .
[ fig : sketch](b ) .
the release of the elastic energy of bending costs an extra energy of dislocations .
let us estimate the dislocation energy first .
the energy of a straight dislocation is equal to @xcite @xmath49 here @xmath50 is the shear modulus , @xmath51 is the burgers vector , the length of the dislocation is taken equal to the width @xmath21 of the contact area introduced above , and @xmath52 is a cutoff distance . for a dislocation running across the nanowire , as shown in fig . [
fig : sketch](b ) , the dislocation strain field is relaxed at distances exceeding the nanowire diameter , and for this reason the cutoff distance in eq .
( [ eq:7 ] ) is taken to be equal to the nanowire diameter @xmath52 .
since the dislocation energy depends on the cutoff distance only logarithmically , this choice is not critical .
a row of dislocations with a separation @xmath53 provides a relative rotation of the two nanowires by the angle @xmath54 .
since the dislocation strain is relaxed on distances exceeding the nanowire diameter , dislocations located at distances from the joint exceeding the diameter would have only little effect on the rotation angle @xmath55 .
hence , the total length of the row can be taken equal to @xmath52 .
this situation contrasts to the common case of relative rotation of two bulk crystallites where the dislocation row runs along the whole boundary between crystallites . denoting the number of the dislocations in the row by @xmath56 , we have @xmath57 and hence @xmath58 if the dislocations in the row are equidistant , the dislocation strain is relaxed on the distance @xmath53 and the logarithmic term in eq .
( [ eq:7 ] ) is equal to @xmath59 .
however , we keep the logarithmic term as it is written in eq .
( [ eq:7 ] ) , thus accounting for a disorder in the dislocation positions .
the difference between these terms is in any case minor .
we express the shear modulus via the young modulus , @xmath60 , where @xmath61 is the poisson ratio , and write the energy of the dislocation row at the joint in fig .
[ fig : sketch](b ) as @xmath62 where @xmath63 . taking the burgers vector equal to the in - plane lattice parameter of gan , @xmath64 nm , the nanowire diameter @xmath65 nm , and the poisson ratio @xmath66
, we get @xmath67 .
we use this value in all calculations below , taking into account that a variation of the nanowire diameter changes this factor only very little because of the logarithmic dependence .
the atomic arrangement at the boundary between the nanowire and the substrate is unknown .
prior to nanowire nucleation , the si substrate is nitridated by the n plasma inducing the formation of an amorphous sin@xmath68 interlayer .
nevertheless , the in - plane orientation of the nanowires clearly follows that of the si(111 ) substrate with a deviation ( twist ) of about 3@xmath69 @xcite . to simplify the calculations below and to preserve a symmetric shape of the nanowires
, we assume that relaxation at the interface proceeds in the same way as at the joints .
hence , we assume that the interface between each nanowire in fig .
[ fig : sketch](b ) and the substrate contains @xmath70 dislocations with the burgers vectors of the same length @xmath51 ( burgers vectors are directed vertically in this case ) , so that each nanowire is inclined by the same angle @xmath55 given by eq .
( [ eq:8 ] ) .
the total energy of all dislocation arrays in fig .
[ fig : sketch](b ) is thus twice the energy given by eq .
( [ eq:9 ] ) .
finally , the total dislocation energy per nanowire is described by eq .
( [ eq:9 ] ) .
let us calculate the energy of the dislocations required to totally relax the elastic bending in the same example as considered at the end of the previous section , @xmath39 nm and @xmath40 nm . with the length @xmath71 nm obtained there , a rotation of the pillars by the angle @xmath72 requires , according to eq .
( [ eq:8 ] ) , @xmath73 dislocations , whose total energy ( [ eq:9 ] ) is equal to @xmath74 j , smaller than the elastic energy of bending ( [ eq:6 ] ) calculated above for the same parameters .
this example shows that the bending energy ( [ eq:6 ] ) and the dislocation energy ( [ eq:9 ] ) are comparable , and the relaxation by creation of dislocations at the nanowire joints may be energetically favorable . in the next section ,
we generalize our analysis of the energy balance by considering a partial relaxation by dislocations .
let us now estimate the total density of dislocations introduced into the nanowire ensemble .
for an array of parallel dislocations in a bulk crystal , the dislocation density is defined as the number of dislocations per unit area oriented normally to the dislocation lines .
a more general definition of the dislocation density is the total length of the dislocation lines per unit volume .
we calculate the dislocation density using this latter definition and ascribing @xmath73 dislocations of length @xmath75 nm to the volume of two nanowires of hexagonal shape with side lengths @xmath39 nm and heights @xmath76 nm . the dislocation density thus calculated is @xmath77 @xmath43 , comparable to the density of threading dislocations in heteroepitaxial gan layers grown by molecular beam epitaxy .
however , one has to keep in mind that these dislocations do not propagate along the nanowire axis , but are localized at the joints .
in addition , their strain fields are relaxed at distances from the joints exceeding the nanowire diameter . as a result , the impact of these dislocations on , particularly , the luminous efficiency of nanowires is certainly much less pronounced than the effect of the same density of dislocations intersecting a heteroepitaxial layer .
let us assume that @xmath56 dislocations are created at the nanowire joint in fig .
[ fig : sketch](b ) .
dislocations are distributed over an interval @xmath52 and provide an angle between the two pillars at the joint @xmath78 .
we assume , as discussed above , that at the interface between each nanowire and the substrate , @xmath70 dislocations are created and induce the inclination of the nanowires by an angle @xmath55 in direction to each other . the same elastic equilibrium equation @xmath2 as in sec .
[ sub : elasticrelax ] is to be solved now with another set of boundary conditions .
we have @xmath3 and @xmath79 at the substrate and @xmath7 , @xmath80 at the joint .
the solution is @xmath81 generalizing eq .
( [ eq:1 ] ) by replacing @xmath9 with the residual quantity @xmath82 the effective distance @xmath83 describes the residual bending of nanowire segments .
it is always smaller than the actual distance @xmath9 between the nanowires and approaches zero for complete relaxation . for the elastic energy of the residual nanowire
bending we have , instead of eq .
( [ eq:2 ] ) , @xmath84 our aim now is to find the minimum of the total energy @xmath85 with respect to the number of dislocations @xmath56 and the nanowire length @xmath5 .
since @xmath86 does not depend on @xmath56 , we find first the minimum of the sum of @xmath87 , described by eq .
( [ eq:12 ] ) , and @xmath88 , given by eq .
( [ eq:9 ] ) , over @xmath56 .
the number of dislocations is considered as a continuous variable .
the minimum over @xmath56 is reached at @xmath89 and is equal to @xmath90 where @xmath91 using eqs .
( [ eq:8 ] ) , ( [ eq:11 ] ) , and ( [ eq:13 ] ) , we can represent the effective length describing the remaining bending of nanowires as @xmath92 while the number of dislocations ( [ eq:13 ] ) providing the minimum of energy can be written as @xmath93 the condition of the energy gain by introduction of dislocations @xmath94 can be written now as @xmath95 .
the minimum of the sum of eq .
( [ eq:14 ] ) and the surface energy ( [ eq:4 ] ) over @xmath5 is @xmath96 where it is denoted @xmath97 and @xmath98 with the values of the parameters used in the example above , we find @xmath99 .
hence , @xmath100 in all practical cases .
the minimum of the total energy ( [ eq:15 ] ) is reached at @xmath101 thus , the gain in surface energy due to bundling exceeds the sum of the bending energy and the dislocation energy for @xmath102 , and the length @xmath5 of the two separate nanowires below the joint is approximately @xmath103 .
let us make an estimate for the same values of parameters as in sec .
[ sub : elasticrelax ] , namely , the side facet width @xmath39 nm and the distance @xmath40 nm . the critical length calculated by eq .
( [ eq:16 ] ) is @xmath104 nm , smaller than the the critical length for purely elastic relaxation in the absence of dislocations .
the joint is at the height @xmath105 nm . for the effective distance @xmath106 describing the residual curvature of the nanowires , eq .
( [ eq:19 ] ) gives @xmath107 nm , an order of magnitude smaller than the real distance @xmath9 between nanowires .
the number of dislocations ( [ eq:19a ] ) providing this minimum is @xmath108 .
the number of dislocations obtained in this estimate is almost by a factor of two larger than the number of dislocations calculated in the example of a complete plastic relaxation in the previous section , since the length of the bundling pillars is two times smaller now , and the angle @xmath109 is correspondingly larger .
dislocations release most of the nanowire bending , leaving the effective distance @xmath83 , which describes the residual bending , much smaller than the distance @xmath9 between nanowires .
thus , most of the elastic energy of bending can be released by creation of dislocations at the joint .
the energy minimization gives a very dense dislocation array , with the distance between dislocations of about 1 nm .
however , plastic relaxation at a joint to the energy minimum may be restricted . at the beginning of a bundling event , two nanowires first touch each other by atomically flat side facets .
an introduction of dislocations at a border between two perfect crystallites may be kinetically suppressed , despite being energetically favorable .
if , however , a sufficient number of dislocations is introduced to minimize the energy , the distances between dislocations are fairly small , since all dislocations that relax the relative misorientation of nanowires need to be located at distances from the joint not exceeding the nanowire diameter .
they have to relax rather large relative misorientations of nanowires .
the elastic energy of such a dense dislocation array may be reduced further by building a large - angle boundary with a coincidence site lattice at the interface .
transmission electron microscopy studies of gan nanowires reveal dislocations at the coalescence joints @xcite and also `` zipper '' defects @xcite whose atomic structure remains obscure . hence , further microscopic studies are needed to understand the relaxation at the nanowire joints in more detail .
the in - plane relative misorientation of the nanowires ( twist ) can be released upon bundling either elastically by a torsion deformation of the nanowires , as shown in fig .
[ fig : twist](a ) , or plastically by creation of an array of dislocations , as depicted in fig .
[ fig : twist](b ) . dislocations created in this case are threading dislocations running along the nanowires . during further growth ,
these threading dislocations are likely to emerge from the nanowire at its sidewall . in this section
, we estimate the energy of the torsion deformation .
the torsion energy of a thin rod is ( ref . , 16 )
@xmath110 where @xmath5 is , as above , the length of the nanowire up to the joint , @xmath111 is the torsional rigidity , and @xmath112 is the torsion angle defined as the angle of rotation per unit length . the torsional rigidity for a rod with the circular cross section is ( ref . , 16 , problem 1 ) @xmath113 .
the torsion angle is @xmath114 , where @xmath115 is the mutual in - plane misorientation of the two nanowires . expressing the shear modulus @xmath50 via the young modulus , @xmath60 , we rewrite eq .
( [ eq:26 ] ) as @xmath116 where @xmath14 is the moment of inertia of the circle of radius @xmath15 ( [ eq:3 ] ) . the elastic energy of bending ( [ eq:2 ] ) can be rewritten as @xmath117 where @xmath118 is the inclination angle of the nanowire due to bundling , as shown in fig .
[ fig : sketch](b ) .
this angle is denoted simply by @xmath55 in the rest of the paper . in this section , it is referred to as @xmath119 , to clearly distinguish it from the twist angle .
the ratio of the torsion to the bending energy is @xmath120 taking the tilt and the twist angles to be equal , one can see that eq .
( [ eq:29 ] ) contains a small numerical factor @xmath121^{-1}\approx0.07 $ ] which makes creation of threading dislocations to relax torsion energy less favorable compared with creation of dislocations running across the nanowires that relax the nanowire bending energy .
we do not have experimental evidences of the torsion relaxation that would stimulate further analysis .
the x - ray diffraction study presented below is not sensitive to torsion , since only symmetric bragg reflections are analyzed .
therefore , we restrict ourselves to the consequences of nanowire bending as analyzed in sec . [ sub : plasticrelax ] .
let us calculate the x - ray diffraction intensity from a single distorted nanowire whose shape is described by eq .
( [ eq:10 ] ) .
we restrict our study to symmetric bragg reflections and consider the displacement of atomic planes in @xmath1 direction , @xmath122 the scattering vector can be written as @xmath123 , where @xmath124 is the reciprocal lattice vector ( directed along the @xmath1 axis ) and @xmath125 is a small deviation from it .
we calculate here the scattering intensity @xmath126 for an arbitrary direction of @xmath125 , with the aim to perform an integration over orientations of nanowires and the x - ray beam divergence in the next section .
the scattering amplitude is an integral over the volume of the nanowire , @xmath127 the nanowire cross section is considered here , for the sake of simplicity , to be a square with sides @xmath128 . a straightforward integration with the displacement field described by eqs .
( [ eq:10 ] ) and ( [ eq:20 ] ) gives @xmath129}{\kappa\left(1-\xi^{2}\right)+q_{x}r},\label{eq:22}\end{aligned}\ ] ] where @xmath130 particularly , if the nanowire is not bent ( @xmath131 ) , eq .
( [ eq:22 ] ) reduces to the common expression for the x - ray scattering amplitude from a perfect crystal of a parallelepiped of sizes @xmath132 , @xmath133 the x - ray scattering intensity is @xmath134 a direct calculation of the scattering intensity by eqs .
( [ eq:22 ] ) or ( [ eq:24 ] ) gives rise to oscillations which are not observed in the experiment because of statistical variations of nanowire lengths , diameters , distances between bundled nanowires , as well as a distribution in the nanowire orientations and the x - ray beam divergence . in the next section
, we average the intensity ( [ eq:25 ] ) over all these distributions .
the x - ray scattering intensity @xmath126 is calculated in the previous section for a single nanowire in the coordinate system that refers to its orientation .
the x - ray intensity scattered from an ensemble of nanowires involves two different kinds of integration .
first , the nanowires possess a range of orientations , with a typical width of few degrees .
second , the x - ray beam illuminating the sample in a laboratory experiment is well collimated in the scattering plane but has a divergence , also a few degrees , normal to this plane ( the vertical divergence ) . as a result
, a nanowire with an orientation different from the reference orientation may provide a strong scattering due to an appropriate component of the incident x - ray beam .
this effect manifests itself in the asymmetric x - ray diffraction profiles from nanowire ensembles @xcite and needs to be properly taken into account in the present analysis .
the scattering amplitude ( [ eq:22 ] ) is written in the frame defined by the nanowire orientation : the @xmath1 axis is along the nanowire and the @xmath135 axis is along the line connecting two bundled nanowires , as shown in fig .
[ fig : sketch ] .
our aim now is to find the components of the vector @xmath125 in this frame .
let us consider first a distribution in orientations of the unit vectors @xmath136 along the nanowires for a given scattering vector @xmath137 .
we write identically @xmath138.\label{eq:30}\ ] ] the first term is a vector along the nanowire , so that @xmath139 .
the second term is a vector in the ( @xmath140 ) plane perpendicular to the nanowire , so that @xmath141 .
the average of the intensity ( [ eq:25 ] ) over all nanowire orientations is the average over the distribution of orientations @xmath136 . since the direction of the @xmath135 axis has been chosen above as the direction from one nanowire to the other in the pair of bundled nanowires , as it is shown in fig .
[ fig : sketch ] , the average of the x - ray intensity also implies an average over the choice of the direction of @xmath142 axis in the ( @xmath143 ) plane .
let us now describe the set of possible scattering vectors @xmath137 .
the laboratory x - ray diffraction setup is designed to provide high resolution in the scattering plane but does not inhibit a large divergence in the direction normal to that plane .
this `` vertical divergence '' gives rise to a fan of the incident beams making the same angle @xmath144 to the sample surface in projection to the reference scattering plane @xmath145 plane but different angles @xmath146 to that plane .
the wave vectors of the incident and the scattered beams in this fan are @xmath147 and the scattering vectors @xmath148 are @xmath149 here @xmath150 is the x - ray wavelength .
variation of the angle @xmath146 in a range corresponding to the vertical divergence of the incident beam provides a set of scattering vectors for a given scattering angle @xmath151 .
the set of waves present in the incident beam can be revealed , and the vertical divergence thus measured , by taking a perfect crystal as a sample and inclining it by an angle @xmath152 with respect to the vertical plane @xcite .
it is instructive to relate the inclination angle @xmath152 and the scattering angle @xmath151 using the formalism above . when a bulk single crystal , rather than a nanowire ensemble , is used as a sample , the first two terms in eq .
( [ eq:24 ] ) reduce to @xmath153 , i.e. , for a given inclination angle @xmath152 , the scattering vector @xmath137 is parallel to the surface normal of the sample @xmath154 the requirement of the vectors ( [ eq : a3 ] ) and ( [ eq : a31 ] ) to be collinear gives rise to the relation @xmath155 .
the bragg condition reads @xmath156 , where @xmath157 is the lattice spacing for the respective reflection . using the relation between the angles @xmath152 and @xmath146 ,
the bragg condition can be written , for @xmath158 , as @xmath159 where @xmath160 defined by @xmath161 is the diffraction peak position at @xmath162 .
equation ( [ eq : a4 ] ) describes the shift of the diffraction peak when a perfect crystal sample is inclined by an angle @xmath152 .
we have checked the positions of the diffraction peaks in fig . 2 of ref . and found that they follow eq .
( [ eq : a4 ] ) .
equations ( [ eq:30 ] ) and ( [ eq : a3 ] ) are used below to calculate the x - ray diffraction intensity from an ensemble of nanowires .
in the present work , we analyze anew the measurements performed in our former study @xcite .
the gan nanowires were grown on si(111 ) substrates by plasma - assisted molecular beam epitaxy .
growth conditions were chosen for the synthesis of two clearly distinct samples in terms of nanowire density , average diameter , and degree of coalescence .
sample 1 ( s1 ) was obtained at a temperature of 810@xmath69c and with a n / ga flux ratio of 2.5 , whereas for sample 2 ( s2 ) both a lower temperature of 780@xmath69c and a lower n / ga flux ratio of 1.2 were employed .
consequently , sample 1 is characterized by a lower degree of coalescence compared to sample 2 .
representative scanning electron micrographs of both samples are shown in fig .
[ fig : sems ] .
the key to the understanding of the coalescence process is gained by a close inspection of the bottom parts of the ensembles in side view @xcite as depicted in the insets of figs . [
fig : sems](a ) and [ fig : sems](c ) .
close to the substrate , the nanowires have diameters of 2030 nm that they attain shortly after nucleation @xcite .
the nanowires do not grow radially in their bottom parts in the process of axial growth . rather , they start to bundle at a height of 100200 nm , as clearly seen in figs .
[ fig : sems](a ) and [ fig : sems](c ) and the magnified insets .
these values are in a good agreement with the numerical estimates in sec .
[ sec : energetics ] . during axial growth , the density of the nanowires continuously decreases and their diameters continuously increase , while the fraction of the area covered by nanowires remains unchanged @xcite .
this behavior is a result of a continuous bundling of nanowires into larger aggregates , a process which is visible in the top - view scanning electron micrographs in figs .
[ fig : sems](b ) and [ fig : sems](d ) by the irregular shape of the coalesced aggregates @xcite . while the presence of bent nanowire segments is evident from the side - view micrographs shown in figs .
[ fig : sems](a ) and [ fig : sems](c ) , and partly also visible in the top view depicted in figs .
[ fig : sems](b ) , we can not determine from these micrographs whether the bending is purely elastic or partly accommodated by plastic relaxation . since x - ray diffraction is sensitive to the residual curvature of nanowires , we analyze x - ray diffraction profiles from both samples to obtain quantitative information on the elastic strain in bundled nanowires .
x - ray measurements were carried out with cuk@xmath163 radiation using a panalytical xpert diffractometer with two bounce ge(220 ) hybrid monochromator and three bounce ge(220 ) analyzer crystal .
the symmetric 0002 , 0004 , and 0006 reflections were measured by @xmath164-@xmath165 bragg scans . to obtain the resolution functions of the experimental setup at the respective scattering angles ,
the same reflections were also measured on a free - standing gan layer with a dislocation density as low as 6@xmath16610@xmath167 @xmath43 .
figure [ fig : xrayprofiles ] shows the measured triple - crystal x - ray diffraction profiles of samples 1 and 2 in successive orders of the symmetric bragg reflections . the profiles measured on the free - standing gan layer shown by thin red lines are used as resolution functions for the respective reflections .
these profiles are measured at the same bragg angles as used for the nanowire samples and thus include the chromatic aberration effects for reflections at different diffraction angles . in the present analysis , the vertical divergence of the x - ray beam and the resulting asymmetry of the diffraction profiles are taken into account in the formalism developed in sec .
[ sec : diffraction ] , as described below .
therefore , the profiles of the reference sample presented in fig .
[ fig : xrayprofiles ] and used in the calculations are the ones obtained without inclination of the sample ( @xmath162 ) , in contrast to our previous analysis @xcite where profiles integrated over @xmath152 had to be employed .
the calculated x - ray diffraction profiles are also shown in fig . [
fig : xrayprofiles ] .
integrations over the distributions of nanowire orientations , lengths , diameters , and distances between bundled nanowires were performed as a monte carlo average of the intensity as obtained by eq .
( [ eq:25 ] ) .
the simulations depicted in fig .
[ fig : xrayprofiles ] are obtained by convoluting the calculated profiles with the resolution functions of the respective reflections . the averages over the orientational distributions of the nanowires and these of the x - ray beams are modeled as follows .
the scattering vector @xmath137 in eq .
( [ eq : a3 ] ) is modeled by a gaussian distribution of the angle @xmath146 describing the vertical divergence of the x - ray beam .
its full width at half maximum ( fwhm ) is taken to be 2.8@xmath69 , equal to that measured by tilting the bulk gan sample @xcite .
then , the scattering vector @xmath125 is defined by eq .
( [ eq:30 ] ) , where the nanowire orientation @xmath136 is described by a gaussian distribution of the nanowire tilt angle with an fwhm of 3.5@xmath69 , and a uniform distribution of the tilt azimuth from 0 to @xmath168 .
the direction of the @xmath135 axis connecting two bundled nanowires is also taken to be uniformly distributed from 0 to @xmath168 . with the scattering vector @xmath125
thus defined , the calculation of the scattering amplitude given in eq .
( [ eq:22 ] ) requires the nanowire length @xmath5 , its diameter @xmath128 , and the effective distance @xmath106 between two bundled nanowires .
the lengths @xmath5 here are not the full lengths of the nanowires , which exceed 1 m , but the nanowire segments that bundle together , whose lengths are some hundreds of nanometers .
likewise , the diameters @xmath128 are not the diameters of the bundled nanowires seen in the top - view scanning electron micrographs in figs . [
fig : sems](b ) and [ fig : sems](d ) , but the smaller diameters of single nanowires forming these bundles .
the length , diameter , and effective distance distributions in the real nanowire ensemble are correlated .
first , thin as - nucleated nanowires experience bundling . during further growth ,
these bundles with a larger total diameter and larger separation compared with the initial nanowires experience further bundling .
however , in the present analysis we neglect these correlations for the sake of simplicity and assume the length , diameter , and effective distance distributions to be independent .
figure [ fig : distrparams ] presents the distributions of the nanowire lengths @xmath5 , diameters @xmath128 , and the effective distances @xmath106 used in the calculation of the diffraction profiles in fig .
[ fig : xrayprofiles ] .
the choice of the distribution type for the nanowire lengths was found to be essential .
symmetric distributions , such as a gaussian , do not lead to an agreement with the experiment .
asymmetric distributions , such as the log - normal distribution used here , enhance the weight of shorter nanowire segments as shown in fig .
[ fig : distrparams](a ) and result in a good agreement between experiment and theory . for simplicity ,
we use log - normal distributions for the diameters @xmath128 and the effective distances @xmath106 as well , but the choice of the distribution type for these quantities is not crucial . figures [ fig : distrparams](a ) and [ fig : distrparams](b ) show that fairly broad distributions of the lengths of the nanowire segments and their diameters are needed to obtain an agreement of the calculated x - ray diffraction profiles and the experimental ones . for sample 1 , the length distribution is characterized by @xmath169 nm , where the first number is the mean value and the second is the standard deviation . for the diameter ,
the corresponding values are @xmath170 nm for sample 1 . for sample 2 ,
we have used @xmath171 nm and @xmath172 nm , respectively .
the smaller lengths of the bundling segments and larger diameters for sample 2 agree with its larger coalescence degree evident from fig .
[ fig : sems ] . .
the length and diameter distributions are taken the same as used for sample 2 in figs .
[ fig : distrparams](a ) and [ fig : distrparams](b ) , respectively . ]
deviations from these values degrade the quality of the fit of the experimental data by the simulated profiles .
however , the most critical parameter for the shape of the calculated diffraction profiles is the effective distance @xmath106 .
the distances @xmath173 nm for sample 1 and @xmath174 nm for sample 2 are notably smaller than the average distance of about @xmath40 nm between the nucleated nanowires prior to their bundling .
already bundled nanowires that experience further bundling during their growth are separated by even larger distances @xmath9 .
this result implies that a large fraction of the bending energy due to nanowire bundling is released by plastic deformation at the joints .
on the other hand , these values of @xmath83 are notably larger than an estimate of @xmath83 from the energy minimization in sec .
[ sub : plasticrelax ] . since introduction of dislocations between two merging atomically flat side facets of nanowires
is hindered , the system may not reach its minimum of energy .
let us estimate the dislocation densities in samples 1 and 2 , using the mean values of the lengths , diameters , and effective distances obtained above .
the estimate is crude since the distributions of these quantities are broad and their maxima occur at smaller values than their means . besides , the distance between nanowires @xmath9 does not enter in the calculation of the x - ray diffraction profiles and has to be assigned separately .
we take the mean distance between uncoalesced nanowires @xmath40 nm for the estimate .
then , the average number of dislocations at a joint as calculated by eq .
( [ eq:19a ] ) is @xmath175 for sample 1 and @xmath176 for sample 2
. these numbers of dislocations per areas @xmath177 give rise to dislocation densities @xmath178 and @xmath179 @xmath43 for samples 1 and 2 , respectively .
hence , larger coalescence degree of sample 2 gives rise to a larger dislocation density . in any case
, this estimate shows that the dislocation densities are comparable with the nanowire densities .
figure [ fig : nwdist ] demonstrates the variation of the diffraction profiles when the mean effective distance @xmath106 is increased .
calculations are performed for the same distributions of the lengths of the nanowire segments and the nanowire diameters as used for sample 2 in figs .
[ fig : distrparams](a ) and [ fig : distrparams](b ) , respectively .
the solid blue line reproduces the calculated profile in fig .
[ fig : xrayprofiles ] and agrees with the experimental profile .
the dotted line , in turn , is calculated for a distribution of @xmath106 possessing the maximum at @xmath180 nm .
this distance between nanowires corresponds to a nanowire density of @xmath181 @xmath43 , which is a maximum density at the end of the nanowire nucleation process and before their massive coalescence @xcite .
hence , this distribution of distances approximately corresponds to purely elastic bending of nanowires without plastic relaxation .
evidently , the profile obtained under this assumption is in gross disagreement with the experimental one .
the comparison of the diffraction line widths in successive reflection orders as originally proposed by williamson and hall @xcite is an established tool of the size - strain analysis since decades@xcite and has been used for gan nanowire samples as well @xcite .
the williamson - hall analysis is based on the assumption that the displacements of atoms are correlated only on short range , so that every small volume of the sample diffracts independently according to its local strain @xcite
. then , the broadening due to an inhomogeneous strain is proportional to the reflection order while the broadening due to finite sizes of the crystallites does not depend on the reflection order .
however , the x - ray scattering from a nanowire ( or its segment ) bent due to bundling as described by eq .
( [ eq:22 ] ) depends on the atomic displacements along the whole nanowire , and does not satisfy the assumptions underlying the williamson - hall analysis . as a result , the size and the strain effects can not be resolved in a simple way .
figure [ fig : voigtfits ] presents the integral breadth analysis of samples 1 and 2 .
since the widths of the diffraction lines themselves and the resolution functions are comparable , a deconvolution of the resolution is essential .
we fit both the measured peaks and the resolution functions by voigt functions [ see fig . [
fig : voigtfits](a ) as an example ] , and then subtract gaussian and lorentzian widths of the resolution functions .
the integral breadths of the diffraction lines are calculated from the resolved gaussian and lorentzian widths ( see , e.g. , ref .
the same calculation is performed on the x - ray profiles calculated by the monte carlo method and presented in fig .
[ fig : xrayprofiles ] .
the integral breadths for the measured and the calculated diffraction profiles are presented in fig . [ fig : voigtfits](b )
. the larger coalescence degree of sample 2 gives rise to broader diffraction peaks .
however , non - linear dependencies of the integral breadths on the reflection order do not allow a clear distinction of size and strain effects .
in fact , we can not extract a single parameter from eq .
( [ eq:22 ] ) that would describe the slopes of the lines in fig .
[ fig : voigtfits](b ) and could serve as a mean - squared strain in the sense of the conventional williamson - hall analysis .
our study suggests that the high - density ensembles of spontaneously formed gan nanowires are not free of extended defects , but actually beset with dislocations .
these defects originate from the overwhelming disposition of gan nanowires to bundle after they have reached a certain critical length .
this process is driven by the reduction of the surface energy of the side facets of the nanowires at the expense of the bending of their bottom segments .
our energy calculations predict that the elastic energy of bending can be reduced by the creation of dislocations at the joints .
these dislocations possess comparatively small elastic energy since their strain fields are relaxed at distances exceeding the nanowire diameter due to the presence of free surfaces . to elucidate the actual strain state of gan nanowire ensembles ,
we have compared the experimental x - ray diffraction profiles from nanowire ensembles exhibiting substantial coalescence by bundling to calculated profiles assuming independent distributions of the lengths of the nanowire segments , their diameters , and the effective distances between nanowires that decribe their residual bending .
the parameters yielding a good fit between the experiment and simulated x - ray diffraction profiles indicate that the bending is indeed partly accommodated by dislocations at the joints .
the fact that the nanowires exhbit a residual bending suggests that plastic relaxation may be kinetically hindered by the difficulty to introduce dislocations when the atomically flat facets of the two nanowires meet .
however , plastic relaxation at the joints may also induce the creation of large - angle boundaries between coalesced nanowire segments .
the x - ray diffraction profiles studied in the present work are insensitive to the microstructure of the joints because of their small volume , and the available microcopic studies in the literature @xcite do not clarify this question either . clearly , dedicated microscopic studies are needed to reveal the actual microstructure of the joints .
the dislocation density in typical gan nanowire ensembles such as investigated in the present work is on the order of @xmath182 @xmath43 , i.e. , similar to the one encountered in heteroepitaxial gan films grown by molecular beam epitaxy .
however , contrary to the situation in films , the dislocations at the tilt boundary of bundled nanowires intersect the nanowire bundle in radial direction , and do not propagate along the nanowire axis .
the effect of these dislocations is thus much less detrimental than that of threading dislocations intersecting a heteroepitaxial layer .
nevertheless , it is likely that these boundary dislocations act as nonradiative centers as well , and it would certainly be desirable to avoid their formation . since the bundling of nanowires is inevitable once the nanowire lengths significantly exceed the distances between them , the only way to avoid this process is a drastic reduction of the nanowire nucleation density .
this reduction , in turn , may be achieved either by selective area growth or by employing a substrate with structural and chemical properties that reduce the nucleation rate of gan nanowires .
34ifxundefined [ 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 * * , ( ) | the coalescence in dense arrays of spontaneously formed gan nanowires proceeds by bundling : adjacent nanowires bend and merge at their top , thus reducing their surface energy at the expense of the elastic energy of bending .
we give a theoretical description of the energetics of this bundling process .
the bending energy is shown to be substantially reduced by the creation of dislocations at the coalescence joints .
a comparison of experimental and calculated x - ray diffraction profiles from ensembles of bundled nanowires demonstrates that a large part of the bending energy is indeed relaxed by plastic deformation .
the residual bending manifests itself by extended tails of the diffraction profiles . |
the field - matter interaction is commonly described by the dicke hamiltonian @xcite , which considers @xmath1 two - level atoms with energy separation equal to @xmath2 immersed in an external one - mode electromagnetic field of frequency @xmath3 inside a cavity .
this hamiltonian comes from the standard quantization of the multipolar form of many atoms interacting with classical radiation @xcite in the long wave approximation .
the dicke model was established exploiting the analogy with the spin .
it is considered a multiatom generalization of the jaynes
cummings model @xcite , a completely soluble quantum hamiltonian of a two - level atom in a one mode electromagnetic field .
the tavis
cummings model @xcite is also a many body extension of the jaynes
cummings model , which considers the rotating - wave approximation .
a very important feature of the dicke hamiltonian discovered by hepp and lieb @xcite is the presence of a phase transition from the normal to the superradiant behaviour .
the latter , introduced by dicke to describe coherent radiation , is a collective effect involving all @xmath1 atoms in the sample , where the decay rate is proportional to @xmath4 instead of @xmath1 , the expected result for independent atom emission .
while this transition to the superradiant phase has been much debated in the literature @xcite , due to the smallness of the field - matter coupling strength , recent experimental results indicate that it can actually be observed @xcite . in this contribution
we use the tensorial product of @xmath5- and @xmath6-coherent states as a trial state to calculate the expectation value of the hamiltonian and construct its energy surface .
this is a function that depends on four variables @xmath7 which define the coherent states , and two essential parameters @xmath8 which are associated to the energy separation of the two - level atoms and the coupling strength between the atoms and the field , respectively . by studying the stability properties of the critical points of this energy surface we obtain the best variational wavefunction and a separatrix which divides the parameter space in two regions : normal and superradiant behaviour , according to @xmath9 or @xmath10 , respectively , with @xmath11 .
the dicke hamiltonian is invariant under transformations of the cyclic group @xmath12 , a symmetry which is not preserved by the trial state .
the symmetry can be restored by acting with the projectors of the symmetric and antisymmetric representations of the cyclic group @xmath12 .
we determine the statistical properties of field and matter observables in the restored variational states @xmath13 in the superradiant region .
the comparison with the same observables evaluated from the exact diagonalization of the hamiltonian at finite number of atoms @xmath1 exhibits an excellent agreement between them , which improves as @xmath1 becomes larger @xcite .
the fidelity between both types of states is also evaluated ; we establish that the fidelity in the superradiant regime is very close to unity , dropping fast in a small neighbourhood of the separatrix .
the great advantage of working with the restored - symmetry states is that they allow us to write down analytic expressions for all the expectation values of field and matter observables , as well as with important mathematical relations .
these analytic expressions are exact in the thermodynamic limit , when the number of particles goes to infinity . the dicke hamiltonian has been studied thoroughly by emary and brandes @xcite with a variational procedure in terms of two @xmath6-coherent states .
one of them is related to the radiation field degrees of freedom and the other arises from a holstein - primakoff ( h - p ) realization of the @xmath5 generators @xcite .
they obtain the appropriate relation between the parameters of the dicke hamiltonian to describe the phase transition , as well as a description for the normal phase in the thermodynamic limit . however , they consider in the superradiant region , an approximated dicke hamiltonian coming from a series expansion of the h - p realization truncated to second order in terms of the ratio between the number of excited atoms over the total number of atoms , assuming that is a very small quantity . under this assumption nagy et al .
@xcite found singularities at @xmath14 in the expectation values of the number of photons and the number of excited atoms , and related them with the divergences in their corresponding fluctuations @xcite . in this contribution we extend and provide details of the calculations presented in @xcite
the expectation values of the mentioned observables are calculated by means of the symmetry adapted variational states and find that there is in fact a divergence of the expectation values , but that this remains throughout the superradiant region , contrary to the published results : there is no singularity .
the expectation value of the number of photons @xmath15 and the squared fluctuation of the first quadrature of the electromagnetic field @xmath16 are proportional to the number of atoms at the phase transition and in the superradiant regime , while the squared fluctuation @xmath17 is proportional to @xmath18 in the same region .
comparison with the exact numerical solution supports our results .
this could have consequences in studies of entanglement and chaos in the superradiant region .
the dicke hamiltonian involves the collective interaction of @xmath1 two - level atoms with energy separation @xmath2 with a one - mode radiation field of frequency @xmath3 in the long wavelenght limit .
it has the form @xmath19 where @xmath20 is given in units of the frequency of the field , and @xmath21 , is the coupling parameter between the matter and field .
the @xmath22 denotes the density of atoms in the quantization volume , @xmath23 is the excitation matrix element of the electric dipole operator of a single atom , and @xmath24 the polarization vector .
the operators @xmath25 denote the one - mode creation and annihilation photon operators ; @xmath26 the atomic relative population operator ; and @xmath27 the atomic transition operators .
we will consider completely symmetric states for which we have @xmath28 .
later we will find it convenient to divide the dicke hamiltonian by the total number of particles , having in this way an intensive hamiltonian operator . in order to study the thermodynamic limit one would take @xmath29 . to find the symmetries of the hamiltonian
, one considers the unitary transformation @xmath30 with @xmath31 denoting the excitation number operator .
one can show that @xmath32 with the corresponding hermitean conjugated relations . substituting these results into the expression for the dicke hamiltonian one
has @xmath33 if the counter - rotating term ( last term of the expression ) is neglected , one recovers the tavis - cummings model and it is immediate that it is invariant under any rotation with arbitrary @xmath34 .
their symmetrized solutions have been thoroughly studied in @xcite . for the dicke hamiltonian ,
it is necessary to restrict to rotations by an angle @xmath35 for the hamiltonian to remain invariant .
thus the invariance group for the dicke model is @xmath36 .
the projection operators for this group are @xmath37 which allow us to write the eigenfunctions of the hamiltonian displaying its symmetry explicitly . in terms of the eigenstates @xmath38 of the photon number operator , and the square of the angular momentum and its projection in the z - axis @xmath39 ,
they may be expressed as @xmath40 where @xmath41 contains only even values of @xmath42 , the eigenvalues of @xmath43 , while @xmath44 contains the odd ones .
the index @xmath45 denotes the state number of the even and odd solutions . in practice one
uses a maximum value for @xmath46 which guaranteed the convergence in the energy eigenvalues of the ground and first excited states .
the coefficients @xmath47 can be obtained from the diagonalization of the hamiltonian matrix , whose dimension is @xmath48 in the basis states without definite parity of @xmath46 .
this matrix , in the symmetry adapted basis states , breaks into two pieces of even and odd parity respectively .
for @xmath49 , the dimensions of the matrices are @xmath50 for integer
@xmath51 , with @xmath52 and @xmath53 . for @xmath54
, they are @xmath55 and @xmath56 .
similar expressions can be obtained for half integer @xmath51 .
the energy surface is found by taking the expectation value of the dicke hamiltonian with respect to the tensorial product of coherent states of heisenberg weyl and @xmath5 groups @xmath57 , given by @xcite @xmath58 where the parameters @xmath59 and @xmath60 are complex numbers .
using the ritz variational principle one may find the best variational approximation to the ground state energy of the system and its corresponding eigenstate . to calculate the expectation values of the field and matter
observables it is convenient to find the representation of the angular momentum and weyl generators with respect to the tensorial product @xmath57 @xcite . for example ,
for the annihilation photon operator we consider the matrix element @xmath61 where @xmath62 denotes an arbitrary state .
we proceed in a similar form for the other observables .
the representations of the annihilation and creation photon operators are thus given by @xmath63 and for the matter observables we have @xmath64 the energy surface can be then calculated straightforwardly @xcite @xmath65 in this expression we take the harmonic oscillator realization for the field part and the stereographic projection for the angular momentum part , @xmath66 where @xmath67 correspond to the expectation values of the quadratures of the field , and @xmath68 determine a point on the bloch sphere . the minima and degenerate critical points are obtained by means of the catastrophe formalism @xcite . by calculating the hessian of the energy surface , we see that when @xmath69 the critical points degenerate and , for that value of the field - matter coupling , the phase transition from the normal to the superradiant behaviour of the atoms takes place .
the critical points which minimize @xmath70 are given by @xmath71 the last column of this array shows the conditions in parameter space to guarantee that they constitute a minimum of the energy surface .
the first row describes the minimum critical points for the normal phase , while the second row describes those for the superradiant regime . in this last case one
has @xmath72 .
( if one were to consider a case in which @xmath73 , we would then have @xmath74 and the critical points for the normal phase would be @xmath75 , for @xmath76 ) .
it is convenient to work with the variable @xmath77 in terms of which the energy values for the minima just described are @xmath78 in a similar form , the expectation values of @xmath43 at the minima are @xmath79\ .
\label{eq007}\ ] ] the fluctuations @xmath80 are zero in the normal region , and in the superradiant phase take the form @xmath81 to study the statistical properties of the variational states in the superradiant regime we calculate the expectation values of linear matter and field observables with respect to the tensorial product of coherent states , as well as their fluctuations .
the results are given in the left side of table [ tab1 ] .
notice that we have introduced the quadratures operators @xmath82 and @xmath83 of the electromagnetic field and they can be written in terms of the creation and annihilation photon operators .
to build variational states which preserve the symmetry of the dicke hamiltonian we apply the projectors of even and odd parity ( [ eq011 ] ) to the coherent states @xmath84 , to obtain @xmath85 where the normalization factors @xmath86 are given by @xmath87 the expression ( [ eq013 ] ) is useful because it allows us to calculate the expectation values of the relevant operators in closed form , in particular the expectation value of the hamiltonian . evaluating the energy surface with respect to the symmetry
adapted states , one obtains @xmath88 in the limit @xmath89 , @xmath90 reduces to eq.([eq003 ] ) , when @xmath91 .
however , it is important to stress that the analysis can be carried out for any value of @xmath1 , and in this contribution we work at finite @xmath1 .
the traditional approach in many - body physics is to use the critical points in eq.([criticos ] ) of the original energy surface , in order to obtain the trial state which approximates the two lowest energy states , and in which to evaluate the expectation values of the observables .
formally , one should calculate the critical points of ( [ symad ] ) , instead of that , we use the critical points associated to eq.([eq003 ] ) .
however all of their critical points also are extreme values for ( [ symad ] ) , except in a small vicinity of the separatrix , as shown in fig.([de_sr ] ) . in that figure ,
the derivatives of the energy surface , given in ( [ symad ] ) , with respect to @xmath92 and @xmath93 , evaluated at the critical points @xmath94 and @xmath95 , are plotted against @xmath96 .
( the derivatives with respect to the other two variables , @xmath97 and @xmath98 , are identically zero ) .
the neighborhood around the separatrix , where they differ , diminishes as @xmath51 increases ( cf .
fig.([deq_sr_js ] ) ) . for the superradiant region
in which the points are critical , they are also minima . the stability analysis of the energy surface ( [ symad ] ) and the study of their behaviour in the near - neighborhood of the separatrix will be reported elsewhere @xcite . by substituting the critical points @xmath99 ,
@xmath100 , and @xmath101 in ( [ symad ] ) , the energy surface associated to the superradiant regime takes the form @xmath102\ , \label{symad2}\ ] ] where we have defined @xmath103 the overlap between the ordinary coherent states and our symmetry - adapted states can be written as @xmath104 in fig .
( [ energias ] ) the energy surfaces for the even and odd parities of @xmath46 are shown together with those for the exact solution , showing a remarkable agreement between them in spite of having considered a small number of atoms @xmath105 . for the odd parity case ,
we have proposed , in the normal region ( @xmath106 ) , a combination of states with @xmath107 : @xmath108 and minimized the expectation value of the hamiltonian with respect to @xmath109 .
one finds @xmath110 .
this gives , for the resonant case , @xmath111 .
one can calculate the expectation values of the matter and field main operators in the superradiant regime with respect to the symmetry adapted states .
the results are collected in table [ tab1 ] together with those associated to the coherent states .
both are evaluated at the critical points for the energy surface , and are given in terms of @xmath112 and @xmath113 . ' '' ''
@xmath114 & @xmath115 & @xmath116 + ' '' '' @xmath117 & @xmath116 & @xmath116 + ' '' '' @xmath118 & @xmath119 & @xmath116 + ' '' '' @xmath120 & @xmath116 & @xmath116 + ' '' '' @xmath121 & @xmath122 & @xmath123 + ' '' '' @xmath15 & @xmath124 & @xmath125 + ' '' '' @xmath126 & @xmath127 & @xmath128 + & & @xmath129 + ' '' '' @xmath130 & @xmath131 & @xmath132 + ' '' '' @xmath133 & @xmath131 & @xmath134 + ' '' '' @xmath135 & @xmath136 & @xmath137 + ' '' '' @xmath138 & @xmath139 & @xmath140 + ' '' '' @xmath141 & @xmath142 & @xmath143 $ ] + ' '' '' @xmath144 & @xmath124 & @xmath145 + & & @xmath146\bigg\}$ ] + ' '' '' @xmath147 & @xmath148 & @xmath149 + ' '' '' @xmath150 & @xmath151 & @xmath152 + observe that , for expectation values different from zero in the symmetry adapted states , the coherent state results can be obtained from the former by letting @xmath113 go to zero , with the exception of @xmath16 and @xmath17 .
these two exceptions arise because @xmath153 for the former but not for the latter .
that @xmath113 tends very quickly to zero as a function of @xmath96 ( or @xmath112 ) can be seen in fig .
( 4 ) , especially for large @xmath1 .
this is why coherent states have been so successful in the past as trial functions .
it can be seen that the quantities which differ the most are some fluctuations of matter and field observables .
to assess this difference we evaluated the fluctuation of the atomic transition operator @xmath154 and of the first quadrature of the electromagnetic field , @xmath82 . in fig .
( [ d2jx ] ) we show the fluctuation of the transition operator @xmath157 for the ground and first excited states of the dicke model with @xmath158 atoms and @xmath159 , calculated for the symmetry adapted states and the coherent states . while @xmath160 tends erroneously to zero for the coherent state , as @xmath112 increases , the results for the symmetry - adapted states show the appropriate quadratic dependence in @xmath1 , asymptotically reaching @xmath161 . in fig .
( [ d2q ] ) we show @xmath162 for the ground and first excited states of the dicke model with @xmath158 atoms and @xmath159 calculated for the symmetry - adapted states and the coherent states . while for the coherent state the result is a constant @xmath163 , for the symmetry adapted states one has a linear dependence in the number of atoms .
additionally , in both cases one can see that the symmetry - adapted states compare really well with the exact result obtained from the diagonalization of hamiltonian for the even and odd parity states , except in the close vicinity of the separatrix of the physical system @xcite .
these results show clearly the benefit of using the symmetry - adapted states over the coherent ones .
the joint probability of finding @xmath164 photons and @xmath165 excited atoms , for even and odd parity states , is obtained by taking the modulus square of the scalar product of the fock and angular momentum states @xmath166 with the symmetry - adapted states .
the result depends on the values of @xmath1 , @xmath167 , and @xmath112 , and it is given by @xmath168\,\frac{1}{\nu ! } \left[n\,\gamma_c^{2 } \
, x^2 \ , \left(1-x^{-4}\right ) \right]^{\nu}\nonumber\\ & & \times \binom{n}{n_{e}}\left(\frac{1-x^{-2}}{2 } \right)^{n_{e}}\left(\frac{1+x^{-2}}{2 } \right)^{n - n_{e}}\ , x^n \ , \frac{\sqrt{f}}{1\pm f}\ . \end{aligned}\ ] ] the joint probability distribution functions , for the even- and odd - parity states , are shown in figs .
( [ conjunta055 ] , [ conjunta100 ] ) for a system of @xmath158 atoms and in resonance , i.e. , @xmath159 . in fig .
( [ conjunta055 ] ) , we consider @xmath169 and in fig .
( [ conjunta100 ] ) , @xmath170 .
it is clear that there are holes in the distributions of each plot , corresponding to the forbidden values ( odd or even ) of the sum @xmath171 .
note that the poissonian distribution when near the separatrix ( @xmath169 ) turns into a quasi - normal distribution as we move away ( @xmath172 ) .
while very few photons and excited atoms contribute to the state in the first case , many more contribute in the second case , as is to be expected . by summing over the number of photons @xmath164 or the excited atoms @xmath173 ,
one determines the corresponding marginal distributions . for the distribution of the number of photons one
has @xmath174^{\nu}\ , \left[x^{n}\pm\left(-1\right)^{\nu}x^{-n}\right ] \,\frac{\sqrt{f}}{1\pm f}\ , \label{phdist}\ ] ] while for the distribution of the number of excited atoms the result is @xmath175 the behavior of these probability distribution functions for even and odd parity states is shown in figs . ( [ excited ] ) and ( [ fotones ] ) for values of @xmath96 close to ( @xmath176 ) and
far from ( @xmath177 ) the separatrix . for @xmath176 , they are different , while for @xmath178 , they can not be distinguished .
we notice that the binomial and poissonian distributions become quasi - normal as we move away from the separatrix .
this is to be expected , since , from the results given in table [ tab1 ] , it is immediate that the distribution function of photons , eq .
( [ phdist ] ) , can be rewritten as @xmath179\ , \end{aligned}\ ] ] where @xmath180 is the expectation value of the photon number operator in the coherent state . for large values of @xmath1
this distribution becomes a poisson distribution , which in this limit is equivalent to the normal distribution @xmath181\ .
\end{aligned}\ ] ] in fig .
( [ excited ] ) , the distribution of the number of excited atoms for the even- and odd - parity states are shown .
for @xmath176 , close to the separatrix , they are different , while for @xmath178 , we can not distinguish them .
for the distribution function of the excited number of atoms , in the limit when @xmath182 one has @xmath183\ , \end{aligned}\ ] ] which for large values of @xmath112 also gives a normal distribution : @xmath184\ .
\end{aligned}\ ] ]
by substituting the critical points of the superradiant phase into the symmetry adapted variational states ( [ eq013 ] , [ eq014 ] ) , we obtain @xmath185\,\sum_{\nu,\,n_{e}}\left(1\pm(-1)^{\nu+n_{e}}\right)\ , \frac{\left(\sqrt{n}\,\gamma_{c}\,x\right)^{\nu}}{\sqrt{\nu ! } } \,\nonumber\\ & \times & \binom{n}{n_{e}}^{1/2}\,\left(1-x^{-2}\right)^{(\nu+n_{e})/2}\ , \left(1+x^{-2}\right)^{(n+\nu - n_{e})/2}\ , \vert\nu\rangle\otimes\vert \tfrac{n}{2},\,n_{e}-\tfrac{n}{2}\rangle\ .
\end{aligned}\ ] ] in the limit @xmath186 we have @xmath187 , @xmath188 , and @xmath189 .
this implies that each matter coherent state goes to an eigenfunction of @xmath190 , i.e. , a rotation by @xmath191 acting on the state @xmath192 . in that limit
the dicke model becomes completely integrable @xcite .
a good measure of the distance between quantum mechanical states is given by the fidelity ; for pure quantum states it measures their distinguishability in the sense of statistical distance @xcite , but it is customary to use the fidelity as a _ transition probability _ regardless of whether the states are pure or not . in this contribution
we calculate the fidelity of the exact ground and first excited states with respect to the corresponding symmetry adapted states @xmath193 in both cases the fidelity gives a result very close to 1 , except in the vicinity of the quantum phase transition . in fig.([fidelity ] ) , the fidelity is shown as a function of @xmath96 for @xmath194 and @xmath195 . for increasing @xmath1
the fidelity rises more sharply to @xmath0 as we move away from the phase transition , located at @xmath196 .
note that ( cf .
eq.([overlapcohsym ] ) ) , had we used the coherent states as trial functions , we would at best obtain a fidelity @xmath197 close to @xmath163 , in the limit of large @xmath198
by means of the symmetry properties of the dicke hamiltonian , we found analytic expressions for the ground and first excited states , which allows also to determine in closed form the expectation values of matter and field observables .
the procedure was the following : we calculated the expectation value of the hamiltonian with respect to the tensorial product of weyl and su(2 ) coherent states .
this expectation value defines a function called _ energy surface _ depending on phase space variables and parameters .
we then determine the degenerate and minimum critical points of the energy surface .
the minimum critical points yield the expressions ( [ ecoherente ] ) which give the mimimum energy of the system together with information about the quantum phase transition present in the dicke hamiltonian .
this quantum phase transition occurs when @xmath199 .
afterwards we restore the symmetry exhibited for the dicke hamiltonian by means of the projection of the states to definite parity of the excitation number operator .
these symmetry considerations lead to determine the ground and first excited states together with important differences from the mean field results for some fluctuations of the matter and field observables .
additionally they provide us with the joint probability distribution functions of the number of photons and excited atoms .
the condition to transit from the normal to the superradiant regimes is very difficult to satisfy for optical systems because the available dipole coupling strengths are usually smaller than the transition frequency .
some proposals to overcome these problems have been discussed in @xcite , where it is stated that the quantum motion of a bose - einstein condensate trapped in an optical cavity can be used to realize the dicke model . indeed , this has been reported in @xcite , where the dicke hamiltonian has been physically realized in a superfluid gas moving in an optical cavity . however , the feasibility of reaching the superradiant phase of the field - matter interaction is still under debate @xcite .
the predicted @xmath200behaviour for the collective @xmath1atom interaction strength of the model has been observed experimentally in circuit qed for up to @xmath201 qubits @xcite . in this work
we have shown that semi - classical states adapted to the symmetry of the hamiltonian are an excellent approximation to the exact quantum solution of the ground and first excited states of the dicke model in the superradiant phase .
their overlap to the exact quantum states is very close to @xmath0 except in a close vicinity of the quantum phase transition ( cf .
fig.([fidelity ] ) ) , whereas that of the ordinary coherent states would be at best equal to @xmath163 ( cf .
eq.([overlapcohsym ] ) ) .
our projected states have analytical forms in terms of the model parameters and allow us to calculate analytically the expectation values of field and matter observables .
we have found that in the superradiant regime the fluctuation @xmath202 of the first quadrature of the electromagnetic field is different from the results obtained via the standard coherent states and its value grows as a linear function of @xmath1 .
something similar happens for the fluctuation in the dipole transition operator @xmath203 , where one finds a quadratic dependence in the number of atoms .
both these results contradict those obtained previously @xcite .
the expectation values of the number of photons and of the number of excited atoms were also studied , finding that there are no singularities at the phase transition .
those found previously @xcite are an artifact of an inappropriate truncation of the hamiltonian .
the joint probability distribution functions for the ground and first excited states were shown , which may be used to characterize the states of atoms in a cavity according to whether @xmath204 has an even or odd value . in a future contribution
, we will evaluate other properties like the entanglement entropy between field and matter , and the squeezing parameter for the electromagnetic field and atomic components , which has been the subject of much interest citevidal .
finally we want to remark that the present formalism allows for a simplification of the exact quantum calculation by considering an expansion of dicke and photon number states running from a minimum value @xmath205 to @xmath206 , where these values can be estimated from @xmath207 and its corresponding fluctuations @xmath208 as functions of @xmath96 , the coupling parameter between the field and matter .
the same simplification can immediately be done in the proposed state ( [ eq013 ] ) without changing the obtained results for all the expectation values and probability distributions .
this work was partially supported by conacyt - mxico ( project-101541 ) , foncicyt ( project-94142 ) , and dgapa - unam .
o. castaos , e. nahmad - achar , r. lpez - pea , and j.g .
hirsch in _ symmetries in nature _ aip conference proceedings * 1323 * , p. 40 - 59
l. benet , p.o .
hess , j.m .
torres , k.b .
wolf ( melville , new york , 2010 ) . | we show that semi - classical states adapted to the symmetry of the hamiltonian are an excellent approximation to the exact quantum solution of the ground and first excited states of the dicke model .
their overlap to the exact quantum states is very close to @xmath0 except in a close vicinity of the quantum phase transition .
furthermore , they have analytic forms in terms of the model parameters and allow us to calculate analytically the expectation values of field and matter observables . some of these differ considerably from results obtained via the standard coherent states , and by means of holstein - primakoff series expansion of the dicke hamiltonian .
comparison with exact solutions obtained numerically support our results .
in particular , it is shown that the expectation values of the number of photons and of the number of excited atoms have no singularities at the phase transition .
we comment on why other authors have previously found otherwise . |
as part of the explore project @xcite , we have recently begun a survey of southern open clusters ( ocs ) with the aim of detecting planetary transits around cluster member stars ( explore / oc ) . probing cluster populations provides a complement to our ongoing deep monitoring studies of rich galactic fields @xcite .
open cluster monitoring provides the following advantages and incentives : * in general , metallicity , age , distance , and foreground reddening are either known or may be determined for cluster members ( more easily than for random field stars ) .
thus , planets detected around cluster stars will readily represent data points for any statistic correlating planet frequency with age or metallicity of the parent star .
* the planet - formation processes , and hence planet frequencies , may differ between the open cluster and galactic field populations .
this study allows the explore project to compare these two different environments . *
specific masses and radii for cluster stars may be targeted in the search by the choice of cluster and by adjusting exposure times for the target . in general , smaller stars offer better chances to detect the low - amplitude transit signal .
the difficulties and challenges involving open cluster surveys are : * the number of monitored stars is typically lower than in rich galactic fields , reducing the statistical chance of detecting planets . *
determining cluster membership of stars in the open cluster fields without spectroscopic data is difficult due to the contamination by galactic field stars .
since the clusters are typically concentrated toward the galactic disk , this contamination may be significant .
* significant differential reddening across the cluster field and along the line of sight can make isochrone fitting ( and subsequent determination of physical parameters such as age , distance , and metallicity ) difficult
. note that both of the latter two difficulties may at least in part be circumvented by obtaining spectra . in this work ,
we illustrate some of the points mentioned above , describe our observing and data - reduction strategies , and show some of our preliminary results of the southern open cluster ngc 2660 .
preliminary results of the study of our second target , ngc 6208 , are presented in a companion paper @xcite in this volume .
of detecting existing transiting planets with different orbital periods .
@xmath0 is calculated with the requirement that two transits must be observed .
* panel a * : @xmath0 of detecting 2 transits of an existing transiting planet with a period between 2 and 5 days after 21 ( top curve ) , 14 ( second curve from the top ) and 7 ( bottom curve ) consecutive , uninterrupted nights of observing ( 10.8 hours per night ) .
the difficulty of detecting some phase angles is shown by the dips in the curves ( e.g. , orbital periods of an integer number of days may always feature their transits during the day and are therefore statistically harder to detect ) .
all phases are considered for each period .
the second curve from the bottom shows the real @xmath0 for our monitoring study of ngc 2660 ( 19 nights of 7 - 8 hours per night , with interruptions due to weather and telescope scheduling ; see fig .
* panel b : * the mean @xmath0 as a function of number of consecutive nights in an observing run ( 10.8 hour nights ) .
the solid line is for the requirement to detect two transits and the dashed line for one transit .
this figure indicates how much the likelihood of finding existing transits grows with an increasing number of nights of observing .
* panel c : * the efficiency of < @xmath0 > per night . for the two - transit requirement ( solid line ) and 10.8 hour long nights ,
an observing run of 21 nights is most efficient . for the single transit requirement ,
the efficiency decreases monotonically with the number of nights since additional nights have progressively lower probabilities of detecting `` new '' transits . ] our observations of ngc 2660 ( @xmath1 ) were obtained with the carnegie institution s swope 1 m telescope at the las campanas observatory in chile during the nights of 2003 feb .
1028 which were partially hampered by poor weather conditions .
our field of view was approximately 23 @xmath2 15 in size .
only @xmath3-band data with a cadence of around 7 minutes were obtained for monitoring . in that filter
, a transit will be most easily distinguishable from contaminants such as grazing binaries because color - dependent limb - darkening results in steeper ingress and egress as well as a flatter eclipse bottom in @xmath3 than in other bands @xcite .
we show in fig .
1 our observing window function which indicates the likelihood of our detecting _ existing _ transits and measuring their periods to within an aliasing factor ( we require at least two transits ) .
a necessary condition for detecting existing transits is , of course , high - precision photometry @xcite . in fig .
2 , we show our photometric precision as a function of magnitude .
around 7000 unphased photometric light curves of the stars with the best photometric precision were visually inspected to detect transit - like signals .
we show in fig .
3 three examples of light curves with the low - amplitude signature typically expected for transiting hot jupiters .
although the eclipses visible in the light curves are most likely not due to planets , we are currently analyzing spectral data on them to determine spectral types of the parent stars as well as their variations in radial velocities . regardless of whether or not we are able to find a planetary transit in ngc 2660 ,
we have illustrated the potential of our methods to detect ( unphased ) amplitude variations of less than 1% , the most fundamental requirement to detect planetary transits . | we present preliminary photometric results of a monitoring study of the open cluster ngc 2660 as part of the explore / oc project to find planetary transits in galactic open clusters . analyzing a total of 21000 stars ( 3000 stars with photometry to 1% or better ) yielded three light curves with low - amplitude signals like those typically expected for transiting hot jupiters .
although their eclipses are most likely caused by non - planetary companions , our methods and photometric precision illustrate the potential to detect planetary transits around stars in nearby open clusters .
address = department of terrestrial magnetism , carnegie institution of washington , + 5241 broad branch road , washington , dc 20015 address = dept of astronomy and astrophysics , university of toronto , + 60 st .
george st . ,
toronto , canada m5s 3h8 address = harvard - smithsonian center for astrophysics , + 60 garden street , ms-15 , cambridge , ma 02138 address = dept of astronomy and astrophysics , university of toronto , + 60 st .
george st . ,
toronto , canada m5s 3h8 address = department of terrestrial magnetism , carnegie institution of washington , + 5241 broad branch road , washington , dc 20015 address = carnegie observatories , + 813 santa barbara st . , pasadena , ca 91101 |
transition to turbulence in cylindrical pipe flow is governed by one single dimensionless parameter , the reynolds number @xmath1 where @xmath2 is the mean flow speed along the pipe , @xmath3 the pipe diameter and @xmath4 the kinematic viscosity of the fluid ( reynolds 1883 ) . despite the simplicity of the set - up
, the reason for transition remains obscure due to the linear stability of the laminar hagen - poiseuille flow ( hagen 1839 , poiseuille 1840 ) and the sensitivity of the process to the exact shape and amplitude of disturbances present .
in most experiments , transition is observed at @xmath5 ( e.g. wygnanski & champagne 1973 ) but can be triggered as low as @xmath6 ( peixinho & mullin 2006 ) or delayed to @xmath7 in very carefully controlled experiments ( pfenniger 1961 ) . until recently , the only firm theoretical result was the energy stability bound of @xmath8 ( joseph & carmi , 1969 ) below which _ all _ disturbances are guaranteed to decay monotonically .
this is , however , more than an order of magnitude below the observed value for transition . an important step forward in understanding the transition process was the discovery of disconnected solutions to the navier - stokes equations in a cylindrical pipe ( faisst & eckhardt 2003 , wedin & kerswell 2004 , kerswell 2005 , pringle & kerswell 2007 ) .
these exact solutions are travelling waves ( tws ) which appear through saddle - node bifurcations as in other shear flows ( nagata 1990 , waleffe 1997,1998,2001,2003 ) .
all these solutions are linearly unstable though they have a very low - dimensional unstable manifold .
there is much interest in these solutions , as very similar structures have been observed transiently in experiments ( hof et al .
2004 , 2005 ) and direct numerical simulations ( kerswell & tutty 2007 , schneider _ et al .
_ 2007a , willis & kerswell 2008a ) .
these states can generally be divided into ` upper - branch ' and ` lower - branch ' tws , based on whether they have high or low wall shear stress .
lower branch solutions are believed to sit on a hypersurface that divides phase space into two regions : one where initial points lead directly to the laminar state , the other where initial conditions lead to turbulent episodes ( kawahara 2005 , wang _ et .
_ 2007 , kerswell & tutty 2007 , viswanath 2008 ) .
the simple translational behaviour of these travelling wave solutions is inherent to the method used to find them , and undoubtedly masks an even larger variety of more complex exact solutions .
the boundary between laminar and turbulent trajectories - labelled @xmath0 hereafter - is formally a separatrix if the turbulent state is an attractor . at low @xmath9 , however , turbulence may ultimately decay after a long transient in which case the laminar state is the unique global attractor .
then the boundary @xmath0 is generalised to the dividing set in phase space between trajectories which smoothly relaminarise and those which do undergo a turbulent evolution .
@xmath0 is thought to be of codimension 1 in phase space but one can _ a priori _ not exclude a more complex fractal structure , as suggested by the dependence of lifetime on initial conditions in simulations of pipe flow ( faisst & eckhardt 2004 ) . trajectories which start in @xmath0 - hereafter called the ` edge ' following skufca et al .
( 2006 ) - stay in @xmath0 for later times by definition and hence the long time dynamics are of obvious interest .
the long time behaviour on @xmath0 has already been found to be a periodic solution in plane poiseuille flow in a pioneering study by toh & itano ( 1999 , see also itano & toh 2001 ) .
skufca et al .
( 2006 ) studied a 9-dimensional model of plane couette flow ( pcf ) to reveal an attracting periodic orbit at low @xmath9 and an apparently chaotic state at higher @xmath9 .
however , recent fully - resolved simulations have shown that the asymptotic behaviour is an attracting tw in pcf ( schneider et al .
2008 , viswanath 2008 ) and a chaotic attractor in a short cylindrical pipe of length @xmath10 ( schneider et al .
interestingly , this seemingly chaotic end state looks to be centered around an ` asymmetric ' tw ( pringle & kerswell 2007 , mellibovsky & meseguer 2007 ) . + the purpose of this paper is to explore the dividing hypersurface @xmath0 in pipe flow with the following objectives : 1 . to establish that the dynamics restricted to this laminar - turbulent boundary explores many different saddle points embedded in it ; 2 . to find evidence for heteroclinic or ` relative ' homoclinic connections between these saddle points ; 3 .
to explore @xmath0 restricted by a discrete rotational symmetry in order to ascertain whether the limiting behaviour remains chaotic or can be a simple attractor ; 4 . to develop a practical and general way to find tws and periodic orbits without detailed knowledge of their spatial structure .
+ the works cited above employ a reduced computational domain for their calculations , imposing short - wavelength periodicity in one direction ( along the pipe , schneider et al .
2007b ) or two ( in the spanwise and streamwise directions in plane couette flow , schneider et al . 2008 , viswanath 2008 ) .
this is very much in the spirit of the minimal flow unit pioneered by jimenez & moin ( 1991 ) and used with considerable success to identify key mechanisms underpinning turbulence ( hamilton et al . 1995 ) . in this paper
, we also adopt this same approach by concentrating on pipes up to @xmath11 long .
the ensuing reduction in the degrees of freedom of the flow allows a much more detailed exploration of the flow dynamics albeit restricted to a subspace of strict spatial periodicity .
this is a necessary preliminary to motivate a more carefully focussed study in much longer pipes which can capture spatially - localised turbulent structures ( e.g. priymak & miyazaki 2004 , willis & kerswell 2007 , 2008a ) but at considerable numerical cost still .
flow structures found in the short pipes considered here also exist , of course , in a longer pipe but will be more unstable due to the greater degrees of freedom present there .
it is also worth remarking that a recent study ( nikitin 2007 ) has shown how pipe flow is slow to lose its spatial periodicity when this is restriction is lifted .
this paper is organised as follows .
section 2 discusses the formulation and numerical methods used to simulate the flow in a pipe and to extract exact recurrent flow solutions .
section 3 presents the results obtained using these in 7 subsections .
subsection 3.1 recalls the method used to follow trajectories on @xmath0 and confirms that the limiting behaviour looks chaotic ( schneider et al .
sections 3.2 and 3.3 discuss how near - recurrent states are identified and using a newton - krylov algorithm demonstrate that a number of unstable travelling waves are approached .
subsection 3.4 offers evidence for the existence of a ` relative ' homoclinic connection between a travelling wave and the same wave rotated .
subsection 3.5 looks for recurrent flow structures in @xmath0 restricted by a discrete rotational symmetry and identifies new exact travelling wave solutions .
subsection 3.6 shows that within this subspace the limiting state of @xmath0 is a simple attractor .
subsection 3.7 confirms that all the tws found in the earlier subsections are indeed embedded in the laminar - turbulent boundary .
the paper ends with a discussion in section 4 .
we consider the incompressible flow of newtonian fluid in a cylindrical pipe and adopt the usual set of cylindrical coordinates @xmath12 and velocity components @xmath13 . the domain considered here is @xmath14\times[0\!:\!2\pi]\times[0\!:\!l]$ ] , where @xmath15 is the length of the pipe and lengths are in units of radii ( @xmath16 ) .
the flow is described by the incompressible three - dimensional navier - stokes equations @xmath17 the flow is driven by a constant mass - flux condition , as in recent experiments ( e.g. peixinho & mullin 2006 ) .
the boundary conditions are periodicity across the pipe length @xmath18 and no - slip on the walls @xmath19 .
the basic tool for the numerical determination of exact recurrent states is the accurate time - stepping code used by willis & kerswell ( 2007,2008a ) .
the velocity field is derived from two scalar potentials @xmath20 and @xmath21 : @xmath22 and the incompressible navier - stokes equations are re - written using the formulation introduced by marqus ( 1990 ) .
the two scalar potentials are discretised using high - order finite differences in the radial direction @xmath23 and spectral fourier expansions in the azimuthal direction @xmath24 and axial direction @xmath25 .
for example , the decomposition of the scalar potential @xmath21 at a radial location @xmath26 , @xmath27 , reads @xmath28 the positive integer @xmath29 refers to the discrete rotational symmetry @xmath30 of the flow ( trivially @xmath31 means no rotational symmetry is imposed ) .
the resolution of a given calculation is described by a vector @xmath32 .
it is adjusted until the energy in the spectral coefficients drops by at least 5 and usually 6 decades from lowest to highest - order modes .
the time stepping is 2nd order accurate with @xmath33 updated using an adaptive method based on a cfl condition .
the number of ( real ) degrees of freedom , which defines the dimension of phase - space , is @xmath34 .
the set of all complex coefficients @xmath35 defines our phase - space with its usual euclidean norm @xmath36 .
symmetries exhibited by tws such as the shift - and - reflect symmetry @xmath37 and the mirror symmetry ( e.g. pringle & kerswell 2007 , pringle et al .
2008 ) @xmath38 are not imposed in the simulations .
the spectral expansion defined above converts the navier - stokes equations into an autonomous dynamical system of the form @xmath39 travelling wave solutions are ` relative equilibrium ' solutions of the navier - stokes , steady in an appropriate galilean frame .
it is numerically convenient to seek and interpret them as a special case of relative periodic orbit " ( rpo ) ( viswanath 2007 , 2008 ) .
we hence developed an algorithm to look for rpos defined as zeros of the functional @xmath40 here , @xmath41 is the point at time @xmath42 on the trajectory starting at time @xmath43 from the point @xmath44 .
@xmath45 is the point in phase space corresponding to the state @xmath46 shifted back in space by the distance @xmath47 in the axial direction and by the angle @xmath48 in the azimuthal direction .
a shift by @xmath49 corresponds in phase - space to the transformation @xmath50 a zero of @xmath51 corresponds to a flow repeating itself exactly after a time t , but at a different location defined by @xmath47 and @xmath52 .
a travelling wave solution is a rpo where there is a degeneracy between the shifts @xmath47 and @xmath52 , and the period @xmath42 .
for the case of a tw propagating axially with speed @xmath53 , @xmath54 , where @xmath42 is an apparent period . to remove this degeneracy , we impose @xmath55 unless otherwise specified .
as the majority of known tws do not rotate , @xmath56 is also assumed ( those that do rotate do so rather slowly , pringle & kerswell , 2007 ) .
starting from a good initial guess for @xmath44 and an estimate of the period @xmath42 ( to be discussed in section [ sec : rec ] ) , we minimise the residual @xmath51 using a newton - krylov algorithm , based on a gmres algorithm ( saad & schultz , 1986 ) . the size of the dynamical system ( typically @xmath57 degrees of freedom ) necessitates the use of a matrix - free formulation .
the use of an _ inexact _ krylov solver also leads to a significant gain in computation time ( we used the choice 2 in eisenstat & walker , 1995 ) .
moreover , we embed the newton solver into a more globally convergent strategy in order to improve likeliness of convergence , using a double dogleg step technique ( dennis & schnabel , 1995 , brown & saad 1990 ) . for the special case of tws with a short period ,
we assume full convergence when the normalised residual @xmath58 is less than @xmath59 .
once a travelling wave solution is known , along with its axial propagation speed @xmath53 , it can be expressed as a steady solution in the frame moving at speed @xmath53 . in this galilean frame
, the stability of the solutions can be studied numerically using an eigenvalue solver based on an arnoldi algorithm .
this yields the leading eigenvalues whose real part , when positive , indicates the growth rate of infinitesimal perturbations to the exact solution in the moving frame .
we present 4 ` edge ' calculations , each motivated by a different question . in the first ( see [ sec : edge]-3 ) , the edge in a pipe length @xmath60 ( @xmath61 ) is calculated starting from a typical turbulent initial condition in order to investigate whether there are coherent structures buried in it .
we are able to confirm the presence of what looks to be a chaotic attractor as found recently by schneider et .
al . ( 2007b ) . in the second edge calculation ( [ sec : heteroc ] ) we look for numerical evidence of heteroclinic connections between saddle points by using a perturbed tw as an initial condition . in the third ( [ sec : m_02 ] ) we impose the discrete rotational symmetry @xmath62 on the flow to exploit the saddle structure of the subset of @xmath0 to reveal new tws which possess @xmath62-symmetry . in the fourth ( [ sec : attractor ] ) we look for evidence of multiple attracting tws to demonstrate that under a rotational symmetry constraint the large - time dynamics need not be chaotic . in this subsection we choose a pipe with @xmath61 ( @xmath63 ) and @xmath64 , set @xmath31 so that there is no restriction on the rotational symmetry , and take a numerical resolution of @xmath65 . for this value of @xmath9 , turbulence , when triggered , is sustained over the typical simulation time which is much longer that the time taken to relaminarise . in order to constrain a numerical trajectory to stay on the edge surface , @xmath0
, we use a shooting method analogous to that first used in plane couette flow by toh & itano ( 1999 ) .
we first produce a long turbulent trajectory and pick a state @xmath66 of relatively low three - dimensional disturbance energy @xmath67 ( defined as the kinetic energy in the streamwise - dependent fourier modes @xmath68 ) . initial states are then chosen from the line , @xmath69 parameterised by the real number @xmath70 and @xmath71 is an axial average .
the initial condition @xmath72 is two - dimensional and hence does not trigger turbulence , whereas @xmath73 is the original turbulent state .
this @xmath70-interval is then repeatedly halved with , at each stage , the new interval selected such that an initial condition corresponding to @xmath70 at the lower limit relaminarises , whereas that at the upper limit leads to large turbulent values of @xmath74 .
@xmath70 is refined close to machine accuracy using double precision arithmetic , forcing the energy of the trajectory to stay at an intermediate level corresponding to the laminar - turbulent boundary for typical times of @xmath75 . at the values of @xmath9
used here , the energy of this boundary is clearly distinct from those associated with turbulence and thus makes determination of @xmath70 unambiguous .
once machine precision has been reached , the process is restarted from nearby states towards the end of the trajectory originating from neighbouring @xmath70 values . the full trajectory , of duration @xmath76 ,
does not show any sign of convergence towards a simple state but rather displays erratic and unpredictable dynamics , as found by schneider et al ( 2007b ) .
we deliberately avoid the word ` turbulent ' as this refers to larger values of @xmath74 .
figure [ e1 ] shows the time evolution of the @xmath67 . despite the lack of regularity of the energy signals , inspection of all velocity components at several random locations in the pipe
indicate clearly that the flow on @xmath0 is sometimes nearly periodic in time on short intervals . for the case
@xmath77 , it is possible by eye to identify several temporal windows where all velocity components approximately oscillate with a given frequency ( see figure [ velz ] ) . within these windows
the dynamics on the edge appear to be temporally correlated with a short period .
a large number of snapshots were taken at times @xmath78 across the whole trajectory . to investigate the possibility of recurrence in the flow , each snapshot state @xmath79
was used as an initial condition for time - stepping . the normalised distance in phase space between this initial point and its evolution in time was then examined for local minima .
specifically , we define the scalar residual function @xmath80 where @xmath47 is a distance by which the state @xmath81 is shifted back in @xmath25 for comparison .
we chose @xmath55 so that a value of @xmath82 for some @xmath83 means that the flow in the pipe is _ exactly _ the flow at time @xmath84 . in this case
, @xmath79 lies on a periodic orbit of the system with the time interval @xmath85 being a multiple of the period ( for @xmath86 , a vanishing residual would indicate a _ relative _ periodic orbit ) .
for the ease of calculation , most of the time we chose to neglect the possibility of recurrence occurring after a shift in the azimuthal direction @xmath24 , and therefore @xmath87 was imposed .
typical plots of @xmath88 starting from different values of @xmath84 are displayed in figure [ res1 ] .
every function @xmath89 attains a smallest local minimum which is defined as @xmath90 figure [ rmin1 ] shows how @xmath91 varies with the starting point .
phases for which @xmath91 is small ( from experience below @xmath92 ) may be interpreted as approximate approaches to periodic orbits of the system by the edge trajectory .
the alternating pattern of maxima and minima in @xmath91 is then the signature of a repellor - a set of states which attract trajectories before ultimately repelling them away . for hyperbolic dynamical systems ,
trajectories are attracted towards one of these states along their stable manifold and ejected away along their unstable manifold ( see the sketch in figure [ phasespace ] ) .
for the parameters here , 6 dips in the @xmath91-profile are suggestive of 6 approaches , denoted respectively by @xmath93 .
parts of the trajectory linking one state to the next ( e.g. @xmath94 or @xmath95 ) may be located in the vicinity of heteroclinic connections between the two states , should such a connection exist , or a relative homoclinic connection for two successive states which are the same but for a shift .
the notion of ` vicinity ' depends on the choice of the norm in phase space . here
, a pragmatic approach is adopted : we use the expression ` @xmath46 is close to a periodic orbit @xmath96 ' to mean that the newton - krylov algorithm converges to the periodic orbit @xmath96 starting with the initial guess @xmath46 .
we now analyse the neighbourhood of the points @xmath79 yielding a low value of @xmath91 looking for exact recurrent states for which @xmath91 exactly vanishes . starting from such initial guesses ,
the newton - krylov algorithm defined in [ sec : newton ] was used to further minimise @xmath91 .
of the six different starting guesses a to f introduced in the previous subsection , excellent convergence was obtained in the cases @xmath97 , @xmath98 , @xmath99 and @xmath100 , with @xmath91 being reduced down to @xmath101 .
the converged states are labelled respectively @xmath102 , @xmath103 , @xmath104 and @xmath105 in order to distinguish them from the starting points @xmath106 , and to indicate the parameter @xmath107 .
the cases @xmath108 and @xmath109 displayed only partial convergence to respectively @xmath110 and @xmath111 and it is not possible to say if there really are zeros of @xmath91 in this neighbourhood . despite ` globalisation ' improvements to the newton - krylov algorithm to improve convergence , sample starting guesses away from the candidates @xmath112 failed to converge .
the converged states @xmath102 , @xmath103 , @xmath104 and @xmath105 all correspond to travelling wave solutions .
close inspection of their spatial structure and dynamics shows that they are , in fact , the _ same _ travelling wave solution modulo a shift in the azimuthal direction : see figure [ plots1 ] .
this state corresponds exactly to the ` asymmetric ' tw identified in pringle & kerswell ( 2007 ) whose @xmath25-averaged velocity field is strongly reminiscent of the chaotic state calculated by schneider et al .
( 2007b ) .
this resemblance has also been noted by mellibovsky & meseguer ( 2007 ) who used a different approach to infer the significance of this same asymmetric tw .
interestingly , the present study at these parameter values did not find evidence of states with higher rotational symmetry in spite of the fact that such tws are known to be embedded in the edge ( see fig .
4 of kerswell & tutty , 2007 ) .
the search for recurrent states described above was initially undertaken with @xmath56 disallowing azimuthal propagation of the recurrent patterns . in the case
@xmath97 , we experimented with allowing @xmath52 to be updated by the newton - raphson scheme .
this involved solving an additional equation @xmath113 ( viswanath 2007 ) . using newton - krylov with the same starting guess @xmath97 but allowing the shift @xmath52 to be updated , produced convergence to another state @xmath114 which is distinct from @xmath102 .
this state , despite a velocity profile very analogous to that of @xmath102 , rotates by an angle @xmath115 degrees after travelling one pipe length .
this situation is reminiscent of the bifurcation diagram obtained by pringle & kerswell ( 2007 ) for the asymmetric tw ( though here @xmath9 is much higher ) : for a given reynolds number , a branch of solutions with helicity is connected to the mirror - symmetric branch , and intersects the non - helical subspace twice .
the solutions at the crossing points are non - helical but nevertheless possess a rotational propagation speed @xmath116 , like our solution @xmath114 .
the fact that such a solution has been found here could be explained by the fact that the numerical code used does not allow helicity , hence the newton - krylov algorithm has no choice but to look for the intersection of the helical branch with a non - helical subspace .
however , other attempts to find exact recurrent patterns with rotation invariably converged back to a non - rotating ( @xmath56 ) tw .
the results so far indicate the possible existence of heteroclinic trajectories linking the exact states found . when the two consecutively visited states are the same modulo a shift in the azimuthal direction , relative " homoclinic connection is a more appropriate term . in an attempt to find further evidence for such connections , a series of initial conditions sampling the unstable manifold of a tw were explored . the asymmetric travelling wave of pringle & kerswell ( 2007 ) was chosen ( i.e. @xmath102 ) with the parameters @xmath77 and orientated so that it was @xmath117-symmetric about the @xmath118 plane . in the corresponding @xmath117-symmetric subspace
, the solution has exactly two unstable eigendirections , denoted by @xmath119 and @xmath120 , @xmath119 being the most unstable one ( there are no unstable directions in the dual space of @xmath117-_anti_symmetric perturbations ) .
these two vectors were normalised to have unit energy and an initial condition was defined as @xmath121 here , @xmath122 is a positive parameter chosen small enough that @xmath44 is effectively in the unstable manifold of @xmath123 ( @xmath124 suffices ) .
the angle @xmath125 defines a plane spanned by @xmath119 and @xmath120 and vanishes in the direction of @xmath119 . for some values of the angle @xmath125 ,
the trajectory starting at @xmath126 returns to the laminar state whereas for others , it becomes turbulent .
boundary values of @xmath125 between these two behaviours lead to intermediate dynamics demonstrating that @xmath127 sits on the edge @xmath0 . in all cases
the trajectory starts in the @xmath117-symmetric subspace , but after a finite time this symmetry is broken by numerical noise .
a bisection method is used to determine ( to machine precision ) an angle @xmath125 for which the trajectory neither evolves to the laminar or turbulent state ( see figure [ heteroplot ] for a sketch of the method ) . the energy contained in the axially - dependent modes
is displayed in figure [ e4 ] and the residual function @xmath128 in figure [ rmin4 ] . these show that @xmath91 increases exponentially but the energy changes little until @xmath129 .
after this linear regime , @xmath91 continues to increase , then drops to a local minimum value of @xmath130 at @xmath131 before increasing again .
when the state at the local minimum ( b4 in figure [ rmin4 ] ) was used as a starting point for the newton - krylov algorithm , @xmath91 smoothly decreased down to @xmath101 .
the exact solution found was the asymmetric travelling wave again but now shifted by an angle of 51.56 degrees .
on the basis that the newton - krylov method usually needs to be in the vicinity of an exact solution to converge to it , this suggests the existence of a relative homoclinic connection near the computed trajectory . collecting firmer evidence is difficult unless the dynamics is further restricted .
recently , significant progress has been made along these lines in plane couette flow by gibson it et al .
( 2008 ) who have demonstrated the existence of a heteroclinic connection between two steady solutions on @xmath0 for a minimal flow unit .
the technique developed above to identify recurrent states in the edge can also equally be applied to the dynamics within a symmetry subspace . restricting the flow to be @xmath132-symmetric , for example ,
improves the possibility of an edge trajectory coming close to some of the @xmath132-symmetric lower branch tws found to be in the edge by kerswell & tutty ( 2007 ) .
there is also , of course , the possibility of discovering unknown branches of tws .
attention was restricted to the @xmath62-symmetric subspace by setting @xmath133 in the flow representation ( [ representation ] ) . at @xmath64 , @xmath61 and resolution @xmath65 ,
an apparently chaotic edge trajectory was followed for @xmath134 during which 5 different phases of relative recurrence - labelled @xmath135 and @xmath136 in chronological order - were detected .
interestingly , this edge trajectory , although seemingly chaotic , is much smoother than the corresponding ( same @xmath9 and @xmath107 ) edge trajectory for @xmath31 : compare figures [ e1 ] and [ e2 ] . of the 5 initial conditions for a newton - krylov search
, @xmath91 only converged to @xmath101 for @xmath137 and @xmath136 ( @xmath91 for @xmath138 and @xmath139 stagnated at @xmath110 ) .
the exact recurrence states found , @xmath140 and @xmath141 ( see figure [ plots2 ] ) , correspond to two new types of travelling wave branches distinct from those originally reported in faisst & eckhardt ( 2003 ) and wedin & kerswell ( 2004 ) .
@xmath140 is especially interesting as it is the first tw found in pipe flow which is _ not _ @xmath117-symmetric .
it does , however , possess a mirror symmetry @xmath142 where @xmath143 ( or equivalently @xmath144 because of the imposed @xmath62-symmetry ) for the snapshot shown in figure [ plots2 ] .
in contrast , @xmath141 has three symmetries : two unimposed - @xmath117 and @xmath145 ( where @xmath146 defines the plane of shift-&-reflect symmetry ) - and one imposed - @xmath62 .
@xmath141 is actually one member of a whole new class of _
tws which have subsequently been uncovered ( the @xmath147-class , see pringle et al .
these new waves are significant because they appear much earlier in @xmath9 than the originally discovered tws ( faisst & eckhardt 2003 , wedin & kerswell 2004 ) and their upper branch solutions look more highly nonlinear . in this subsection ,
the flow is again restricted to be @xmath62-symmetric but the pipe is halved in length ( @xmath148 so @xmath149 ) and @xmath9 reduced to @xmath150 ( resolution is @xmath151 ) .
at these settings , kerswell & tutty ( 2007 ) identified a lower branch tw - @xmath152 in their nomenclature - which was on the edge with only one unstable direction normal to it .
this tw is therefore a local attractor for trajectories confined to the edge .
a leading question is whether it is also a global attractor on the edge . + a starting point was randomly chosen along a turbulent trajectory , and an edge trajectory generated for @xmath153 using the method described above .
the energy signal is displayed in figure [ e3 ] and the corresponding recurrence signal @xmath128 is displayed in figure [ rmin3 ] .
two successive dips in @xmath91 can be seen at times @xmath154 and @xmath155 later corresponding to inexact recurrences labelled respectively @xmath156 and @xmath157 . when used to initialise a newton - krylov search , both led to a value of @xmath101 for @xmath91 .
the two converged exact states correspond to one unique travelling wave solution , differing only by a small shift in the azimuthal direction of less than 4 degrees . as a result , we label only the first @xmath158 : see figure [ plots2 ] .
the tw @xmath158 has all the same symmetries as @xmath141 but possesses a distinctly different structure consisting of 4 low - speed streaks near the walls separated by 4 high - speed streaks .
four other high - speed streaks of lesser amplitude and smaller size , and with a much more steady structure , can be found closer to the pipe axis .
this new tw thus has a richer structure than the @xmath62 branch of solutions found in faisst & eckhardt ( 2003 ) and wedin & kerswell ( 2004 ) . just as for @xmath141
, subsequent investigations have revealed that @xmath158 is but one member of another highly - symmetric class of tws ( the @xmath159-class , pringle et al .
2008 ) characterised by this distinctive double - layer structure . after a time @xmath160 ,
the energy of the edge trajectory reaches a constant plateau and the value of @xmath91 decreases exponentially to @xmath161 by @xmath162 .
this indicates that the trajectory is converging to an end state .
newton - krylov was used to accelerate the convergence of the trajectory by taking the state at @xmath163 and reducing @xmath91 to @xmath101 .
the spatial structure of the initial and final state of this procedure , however , is indistinguishable .
the final state is another travelling wave solution , @xmath164 ( see figure [ plots2 ] ) , distinct from @xmath152 of kerswell & tutty ( 2007 ) .
stability analysis of @xmath164 reveals only one unstable direction which has to be normal to the edge ( see pringle et al 2008 for details ) .
hence @xmath164 is an attractor in the edge , and both @xmath164 and @xmath152 can only be local attractors there .
closer examination of @xmath141 and @xmath164 , which look very similar , reveals that they are members of the same tw family distinguished only by their different wavelengths .
there is no guarantee that starting guesses taken from the edge for the newton - krylov method will converge to states also on the edge .
the calculations of subsection 3.4 have , however , demonstrated this for @xmath102 ( the asymmetric tw ) thereby confirming earlier speculation ( schneider et al 2007b , pringle & kerswell 2007 ) . to explicitly check this for the other tws found here , each was taken as an initial condition with various very small perturbations added to start a series of runs . in all cases ,
two contrasting behaviours were evident : one perturbation could be found which led to relaminarisation while another led to a turbulent episode : see figure [ fig : edgechk ] .
this confirms that all the tws found here are on the laminar / turbulent boundary for the particular parameter settings tested ( shown in table 1 ) .
a summary of the main findings of this investigation are listed below .
+ * the laminar - turbulent boundary @xmath0 in a short pipe seems to have a chaotic attractor as observed by schneider et al .
( 2007b ) .
* although @xmath0 has many unstable lower branch tws embedded within it ( e.g. kerswell & tutty 2007 ) , only one - the asymmetric tw of pringle & kerswell ( 2007 ) ( and rotated derivatives of it ) - was approached by edge trajectories studied here . *
a new asymmetric tw which has a small but finite azimuthal phase speed has been discovered .
* some evidence was gathered for a relative " homoclinic connection between the asymmetric tw and the same wave with a different orientation . *
calculations within a @xmath62-symmetric subspace have revealed new branches of @xmath62-symmetric tw solutions .
one ( @xmath140 ) is the first tw found _ not _ to be shift-&-reflect symmetric in pipe flow , while the others are members of new classes of _ highly_-symmetric tws ( pringle et al . 2008 ) .
* the laminar - turbulent boundary restricted to @xmath62-symmetry has at least two simple local attractors in the form of tws at @xmath165 .
+ our numerical experiments have shed some light on the dynamical structure of @xmath0 in pipe flow . when no symmetry is imposed , and the pipe is approximately @xmath11 long ( @xmath61 ) , trajectories which neither relaminarise nor become turbulent all look chaotic and visit some exact recurrent states in a transient manner . in all cases
studied , these exact solutions appear to be specifically the asymmetric tws found by pringle & kerswell ( 2007 ) even though many other lower branch tws are embedded in @xmath0 ( kerswell & tutty 2007 ) .
no rpos were found even though the numerical method was sufficiently general to find them , in addition to tws . whether this is because the rpos are rarely visited or just more difficult to isolate is unclear .
it is tempting to conclude that chaos on the edge of pipe flow is structured around a set of unstable saddle points ( the tws ) linked together by heteroclinic or sometimes relative heteroclinic connections .
an approximation to such a connection has been shown in subsection [ sec : heteroc ] .
the resulting dynamical structure is likely to be a heteroclinic ( or homoclinic ) tangle , an efficient mechanism to produce chaotic trajectories in phase - space with the set of tws acting as a skeleton of the tangle . seen from this point of view , transition to turbulence from a given initial condition depends on the position of the initial state , in phase space , relative to the stable manifold of each exact state embedded in @xmath0 . when a trajectory approaches a travelling wave solution belonging to @xmath0 , its relative position to the boundary determines on which side of the edge the trajectory escapes , resulting in either relaminarisation or a turbulent transient ( see sketch in figure [ phasespace ] ) . + .
parameters , axial propagation speed @xmath166 ( in units of @xmath2 ) and azimuthal propagation speed @xmath167 in @xmath168 and status of all travelling wave solutions found in this paper ( @xmath141 is in the same tw family as @xmath164 and hence not listed separately ) . [
cols="^,^,^,^,^,^,^",options="header " , ] when the flow is artificially restricted to be @xmath62-symmetric at @xmath64 , the trajectory remains chaotic wandering sufficiently close to _ distinct _ tw states ( @xmath140 and @xmath141 ) to be captured by the newton - krylov scheme . by halving the pipe length to @xmath169 and reducing @xmath9 to @xmath150 ,
multiple local attractors appear in @xmath0 . a randomly - started trajectory , after a chaotic transient and approaches to states @xmath158 and @xmath170 , is attracted towards the tw solution @xmath164 .
the fundamental difference between @xmath164 and all the other states mentioned so far lies in the number of unstable eigenvalues . when a tw has two or more unstable eigenvalues in a given subspace , the dimension of the intersection between its unstable manifold and the hypersurface @xmath0 is reduced by one but remains at least one .
hence such a state remains a saddle point on @xmath0 : edge trajectories enter its vicinity along its stable manifold and escape along the unstable manifold .
in contrast , when a state has only one unstable direction , this is necessarily normal to @xmath0 , so that the state becomes an attractor for dynamics restricted to the edge ( a relative attractor ) .
we have shown that @xmath164 is one such tw , but there is at least one other , @xmath152 of kerswell & tutty ( 2007 ) .
therefore we expect both these two states to be only _ local _ attractors on @xmath0 rather than global ones . in the cases of plane poiseuille and couette flow , a lower - branch solution exists which has only one unstable direction and is therefore a relative attractor ( toh & itano 1999 , itano & toh 2001 , wang _ et .
_ 2007 , schneider et al . , 2008 , viswanath 2008 ) .
pipe flow is different as tws all seem to have at least two unstable directions when there is no imposed discrete rotational symmetry .
for example , the @xmath152 tw has 1 harmonic ( @xmath62-symmetric ) unstable eigenfunction and one subharmonic ( only @xmath171-symmetric ) unstable eigenfunction . +
while discrete rotational symmetry constraints might appear artificial , they have nevertheless served as a useful device for discovering new exact recurrent solutions ( see table 1 ) .
the method developed here based on edge tracking , recurrence analysis and use of a newton - krylov algorithm , is a general approach for finding new exact recurrent solutions in any flow situation possessing subcritical behaviour .
importantly , the present method naturally selects the states that are most likely to be visited and does not presuppose anything about their spatial structure ( modulo the symmetries imposed on the flow ) .
the use of the scalar function @xmath91 , coupled with a newton - krylov solver , can also be used to search for more complex solutions such as relative periodic orbits , whether located on the laminar - turbulent boundary or embedded in fully developed turbulence .
this is currently underway .
+ may open questions remain regarding @xmath0 in pipe flow .
is @xmath0 really a hypersurface or can it have a more fractal structure ? why does the flow keep approaching the asymmetric tw or rotations of it , and not any of the other lower branch tws known to exist within the same parameters range ?
finally , what are the properties of @xmath0 for longer pipes , where localised turbulent ` puffs ' structures exist ?
preliminary work for the extended pipe in a reduced model has already revealed an interesting localised structure as the attracting state ( willis & kerswell 2008b ) . + acknowledgements : we would like to thank c. pringle for helping with the velocity profiles of the tw solutions and b. eckhardt for stimulating discussions about this topic .
y.d . was supported by a marie - curie intra - european fellowship ( grant number meif - ct-2006 - 024627 ) and a.w . by the epsrc ( grant number gr / s76144/01 ) .
reynolds , o. 1883 an experimental investigation of the circumstances which determine whether the motion of water shall be direct or sinuous and of the law of resistance in parallel channels _ phil . trans .
_ , * 174 * , 935 - 982 toh , s. & itano , t. 1999 low - dimensional dynamics embedded in a plane poiseuille flow turbulence : traveling - wave solution is a saddle point ?
iutam symp . on geometry and statistics of turbulence _
kambe , t. ) , kluwer .
the thick line indicates the edge trajectory and the thinner lines nearby trajectories which either relaminarise ( energy decreases ) or becomes turbulent ( energy increases to a higher level ) .
time is in units of @xmath174 . ]
separates initial conditions which relaminarise from those which become turbulent .
an edge trajectory visiting three states a , b and c is shown schematically .
the dynamics on the manifolds transverse to @xmath0 are shown by trajectory diverging towards either the laminar state or the turbulent state.,width=453 ] , @xmath31 , @xmath61 .
the thick line indicates the edge trajectory and the thinner lines nearby trajectories which either relaminarise ( energy decreases ) or becomes turbulent ( energy increases to a higher level ) .
time is in units of @xmath174 ] .
the thick line indicates the edge trajectory and the thinner lines nearby trajectories which either relaminarise ( energy decreases ) or becomes turbulent ( energy increases to a higher level ) .
time is in units of @xmath174 ] .
the thick line indicates the edge trajectory and the thinner lines nearby trajectories which either relaminarise ( energy decreases ) or becomes turbulent ( energy increases to a higher level ) .
time is in units of @xmath174 ] ) and another increasing up to the turbulent state .
@xmath137 represents @xmath140 , @xmath156 is @xmath158 , @xmath179 is @xmath102 and @xmath180 is @xmath164 with @xmath181 used to distinguish the starting perturbation.,width=453 ] | the laminar - turbulent boundary @xmath0 is the set separating initial conditions which relaminarise uneventfully from those which become turbulent .
phase space trajectories on this hypersurface in cylindrical pipe flow look to be chaotic and show recurring evidence of coherent structures . a general numerical technique
is developed for recognising approaches to these structures and then for identifying the exact coherent solutions themselves .
numerical evidence is presented which suggests that trajectories on @xmath0 are organised around only a few travelling waves and their heteroclinic connections . if the flow is suitably constrained to a subspace with a discrete rotational symmetry , it is possible to find locally - attracting travelling waves embedded within @xmath0 .
four new types of travelling waves were found using this approach . |
the @xmath1 photoproduction cross section is dominated by the production of the @xmath2 meson . in the @xmath3-wave
the near threshold @xmath1 production is strongly influenced by the isoscalar @xmath5 and the isovector @xmath6 resonances .
the analysis of the near - threshold @xmath1 dynamics is especially important for an explanation of the nature of the @xmath5 meson .
it is still a matter of controversy whether this meson is genuine quasi bound @xmath7 state .
exploiting the large coupling of the photon to the vector mesons along with the @xmath3-@xmath4 wave interference one is able to reveal many interesting properties of the scalar mesons , for instance their photoproduction cross section and their impact on the angular distribution of the produced kaons .
the data on the near threshold @xmath1 photoproduction are , however , very scarce .
our analysis is based on the results obtained by behrend et al . at desy @xcite at the average incident photon energy of 5.65 gev and by barber et al . at daresbury @xcite at 4 gev .
the main motivation of our study was to explain a huge discrepancy in the @xmath3-wave cross sections reported by these experiments .
the results varied between 2.7 and 96 nb respectively . we have also performed a more extensive analysis of the moments of the angular distributions .
these moments are the main source of information about the partial wave interference . to this end
we have taken into account all the moments constructed from the @xmath3- and @xmath4-wave amplitudes .
it is worth of mentioning that previous analyses excluded all the moments but @xmath8 .
here we present the main ingredients of our model referring the reader to our previous papers @xcite for a more thorough study . the partial wave projected amplitudes have been constructed for the reaction @xmath0 : @xmath9 where @xmath10 are helicities of the photon , incoming and outgoing protons , @xmath11 is the @xmath12 solid angle , @xmath13 is 4-momentum transfer squared , @xmath14 is the @xmath1 effective mass and @xmath15 describe the angular momentum of the outgoing @xmath1 system and its projection . for the @xmath3-wave we have used the born amplitudes schematically depicted in fig .
altogether we have 6 diagrams of the type presented in fig.1a and 5 diagrams of the type presented in fig .
the model has also taken into account the final state interactions .
the @xmath16 and @xmath17 transitions have been included along with the @xmath1 elastic rescattering . to describe the propagation of the exchanged particles we have used either the normal @xmath18 or the regge propagators @xmath19\gamma(1-\alpha(t))(\alpha's)^{\alpha(t)}/(2s^{\alpha_0}),\ ] ] where @xmath20 is the mass of the exchanged particle and
@xmath21 is its regge trajectory .
+ in the @xmath4-wave we have assumed the reaction to proceed through the pomeron exchange .
having performed the numerical calculations we have arrived at a very good description of the mass distributions and the moments which are shown in fig . 2 for the case of @xmath22=4 gev .
= 4 gev ] our model gives the value of the integrated photoproduction cross section in the range between 4 and 7 nb for the photon energy varying from 4 gev to 5.65 gev .
using the model constructed by us we have properly reproduced the mass distributions and the moments measured in the experiments . some other observables like the differential cross section @xmath23 and the density matrix elements
can also be described in agreement with experimental data .
our calculation favors the desy experimental estimation of the s - wave integrated total cross section . | we have analyzed the @xmath0 reaction in the @xmath1 effective mass region around the mass of the @xmath2 meson .
the interference of the @xmath3-wave contribution with the @xmath4-wave has been studied . both scalar resonances @xmath5 and @xmath6 have been taken into account .
we obtained a good description of the available experimental data , in particular the mass distributions and the moments of the kaon angular distribution .
our calculations give values of the integrated @xmath3-wave total photoproduction cross section between 4 and 7 nb for the @xmath1 effective mass range around the @xmath2 mass and at the laboratory photon energy near 5 gev .
these numbers favor lower experimental estimates obtained at desy . |
the subject of atomic systems under spatial confinement is of immense interest among the researchers since the advent of quantum mechanics as the spectral characteristics of atomic systems placed under different confinements change appreciably compared with those of free atoms @xcite .
a number of phenomenological potentials have been used to model atoms within cavities @xcite , atoms under pressure @xcite , impurities in quantum dots or nano crystals @xcite , nanopores @xcite , fullerenes @xcite and foreign atoms in liquid helium environment @xcite _ etc_. the study of quantum dots ( qd ) has got considerable attention in recent times due to its fundamental importance in theoretical researches as well as in fabricating new functional devices .
the qd s ( or artificial atoms ) , in general , contains several electrons subjected to an external confining potential and they show similar structural properties as compared to pure atoms .
the structural changes of the impurity atoms inside qd s _ w.r.t . _
the parameters of confining potentials provide huge physical insight about the interaction of the atom with its surroundings .
although the bound states of confined hydrogen and helium atoms have been studied extensively by several researchers @xcite , very few attempts have so far been made towards the quasi bound or resonance states of one electron impurity atom in an isolated qd @xcite and also in case of confined two electron systems @xcite .
transformations of two - electron bound states to feshbach and then to shape resonances depending upon a parameter of model rectangular well - type potential representing the qd have been studied by bylicki _ et .
_ @xcite and genkin _ et .
@xcite showed that the singly excited bound states of a two - electron atom become resonance states for appropriately chosen parameters of an external attractive spherical gaussian type confining potential used to model the qd .
+ in the present work , we have considered a spherically symmetric finite oscillator potential @xcite of the type , @xmath16 for modeling the qd confinement . here
@xmath17 is the depth of the potential well and the cavity constant @xmath18 is defined as , @xmath19 where @xmath20 is the width of the potential . by tuning the parameters @xmath17 and @xmath20 one can change the shape of the potential given by equation ( 1 ) . such type of two - parameter ( @xmath17 and @xmath20 ) potential provides much control and flexibility in modeling the size of a qd . when @xmath21 _ i.e. _ near the center of qd , @xmath22 and thus a harmonic nature is observed in the potential for a given cavity constant @xmath18 .
but for large ` _ _ r _ _ ' , it deviates from the harmonic behavior .
in fact , the fo potential is quite similar in profile to that of gaussian potential . at the same time
, it facilitates the computation of matrix elements in a simple and efficient manner , especially when the slater - type orbitals are used in constructing the wave function with appropriate boundary conditions for a confined system .
this fo potential was used by winkler @xcite to study the two - electron bound and resonant states of helium in qd where the electron correlation was not included initially in the optimized wave function .
even the inclusion of electron correlation could not remove the uncertainties in their calculations @xcite .
later , kimani _ et .
@xcite applied the restricted hartree - fock method to estimate the ground states of many - electron close - shell quantum dots modeled by the fo potential where the electron correlations were included approximately .
chakraborty and ho @xcite made a sophisticated approach to deal with this problem by expanding the wave function in single exponent hylleraas type basis within the framework of stabilization method , but their work was restricted to only the lowest lying doubly excited resonance state @xmath23 ( @xmath24 ) of helium .
it is worthwhile to mention that an appropriate knowledge of resonance structure of few - electron qd with and without a central impurity atom will help to understand the electron transport phenomena occurring in real semiconductor qds @xcite .
+ under such circumstances , we have studied the resonance parameters of @xmath1 states originated from @xmath3 and @xmath5 [ @xmath4 configurations of qd confined helium below @xmath6 ionization threshold of @xmath7 in the framework of stabilization method @xcite by using explicitly correlated multi - exponent hylleraas type basis set .
this method was successfully employed by the present workers @xcite for calculations of resonance parameters of different resonance states of the free and confined helium - like ions . in the present study ,
the resonance parameters of the states under consideration are estimated over a wide range of width ( @xmath20 ) for a fixed depth ( @xmath17 ) of the fo potential .
the energy values of bound @xmath0 ( @xmath1 ) states [ @xmath2 have also been been reported .
moreover , the positions of @xmath25 , @xmath26 ( @xmath27 ) and @xmath28 ( @xmath29 ) states of @xmath30 have been estimated for a comprehensive understanding about the structure of qd confined helium .
the variation of ionization potential of qd confined _ he _ with respect to the width of the fo potential have been studied .
it has also been shown that the potential given by equation ( 1 ) breaks the orbital angular momentum ( @xmath31 ) degeneracy in coulomb field for the energy levels of hydrogen - like atoms .
finally , we have shown that for a fixed cavity depth ( @xmath17 ) , the widths of the resonance states show oscillatory behavior with respect to the width ( @xmath20 ) of the quantum cavity .
it has been noted that for higher excited states , such oscillations are more pronounced .
the paper is arranged as follows : a brief discussion on the present methodology is given in section ii , followed by a discussion on the results in section iii , and finally concluding in section iv with a view towards further use of the present techniques in related studies of spatially confined atomic systems _
e.g. _ qd , presuure confinement , scp confinement _
for any @xmath32 state of even parity arising from two electrons having same azimuthal quantum number , the variational equation @xcite can be written as , @xmath33 dv_{r_{1},r_{2},\theta_{12 } } = 0\end{aligned}\ ] ] subject to the normalization condition , @xmath34 where the symbols used in equation ( 3 ) and equation ( 4 ) are the same as in reference @xcite .
the effective potential is given by , @xmath35+\frac{1}{r_{12 } } \end{aligned}\ ] ] the multi - exponent correlated wavefunction @xcite considered in the present calculation is expressed as , @xmath36 { \nonumber}\\ & + & \sum_{i}\sum_{j}\left[\eta_{i}(1)\eta_{j}(2)\sum_{l\geq 0}\sum_{m \geq 0}\sum_{n\geq 0}c_{lmn}r_{1}^{l}r_{2}^{m}r_{12}^{n}+ exchange \right]\end{aligned}\ ] ] where , @xmath37 where , @xmath38 s are the non - linear parameters . here , @xmath39 and @xmath40 are the radial co - ordinates of the electrons and @xmath41 is the relative distance between them . in a multiexponent basis
set , if there are @xmath42 number of non - linear parameters , then the number of terms in the radially correlated basis is @xmath43 and , therefore , the dimension of the full basis ( @xmath44 ) including angular correlation will be @xmath45 $ ] , where @xmath46 is the number of terms involving @xmath41 @xcite .
for example , as we have used here nine non - linear parameters , the number of terms in the radially correlated basis is 45 and with 10 terms involving different powers of @xmath41 , the dimension of the full basis ( @xmath44 ) becomes 450 .
the values of the non - linear parameters are taken in a geometrical sequence : @xmath47 , @xmath48 being the geometrical ratio @xcite .
the wavefunction can be squeezed or can be made more diffuse by changing the geometrical ratio ( @xmath48 ) keeping @xmath49 constant throughout . to have a preliminary guess about the initial and final values of nonlinear parameter @xmath38 , we optimize the energy eigenvalues of @xmath1 states below @xmath50 ionization threshold of @xmath51 by using nelder - mead procedure @xcite .
the energy eigenroots are then obtained by solving the generalized eigenvalue equation , @xmath52 where , @xmath53 is the hamiltonian matrix , @xmath54 is the overlap matrix and @xmath55 s are the energy eigenroots .
the wavefunction is normalized for each width ( @xmath20 ) of the fo potential to account for the modified charge distribution inside the qd .
each energy eigenroot plotted against the geometrical ratio ( @xmath48 ) produces the stabilization diagram .
subsequently , we can calculate the density of resonance states from the inverse of tangent at different points near the stabilization plateau in the neighborhood of avoided crossings for each energy eigenroot .
the plot of calculated density of resonance states versus energy for each eigenroot is then fitted to a standard lorentzian profile .
the best fit , _
i.e. _ , with the least chi square ( @xmath56 ) and the square of correlation ( @xmath57 ) near unity yields the desired position ( @xmath58 ) and width ( @xmath59 ) of the resonance state .
+ for each width ( @xmath20 ) of the confining potential , the energy eigenvalues of @xmath60 and @xmath61 states of the confined one - electron @xmath7 ion is obtained by using ritz variational technique considering the wavefunction as , @xmath62 where , @xmath63 s are the nonlinear parameters and @xmath64 s are the linear variational coefficients . for @xmath7(@xmath65 ) states [ @xmath66 , we have considered 14-parameter basis set whereas for @xmath7(@xmath67 ) state we have taken 13 parameters in the basis . in both the cases , @xmath31 is ranging from 0 to 4 .
all calculations are carried out in quadruple precision .
atomic units have been used throughout unless otherwise specified .
to construct the stabilization diagram corresponding to each width ( @xmath20 ) of the fo potential , repeated diagonalization of the hamiltonian matrix in the hylleraas basis set of 450 parameters is performed in the present work for 400 different values of @xmath48 ranging from 0.63 @xmath68 to 0.77 @xmath68 for a depth @xmath17 = 0.2 @xmath68 and a width @xmath20 = 4.0 @xmath68 of the confining potential , a portion of the stabilization diagram for @xmath1 states of confined helium below @xmath6 ionization threshold of @xmath7 is given in figure-1 .
it is evident from figure 1 that there exist two classes of states : 1 .
first few energy eigen - roots lying below @xmath69 [ @xmath70 a.u . ] level are insensitive with the variation of @xmath48 .
this feature clearly suggests that these energy eigen - roots originating from @xmath0 configurations of qd confined helium are bound _
i.e. _ stable against auto - ionization .
energy eigen - roots lying between @xmath69 and @xmath71 [ @xmath72 a.u . ] are sensitive with the variation in @xmath48 and gives rise to flat plateau in the vicinity of avoided crossings of the energy eigenroots in the neighborhood of some particular energy values .
this is a clear signature for the presence of @xmath1 resonance states of qd confined helium .
the present calculated bound state energy eigenvalues ( @xmath73 ) of @xmath0 ( @xmath1 ) [ @xmath74 states of @xmath75 as well as the @xmath69 energies for different cavity widths ( @xmath20 ) starting from a very low value of 0.001 a.u .
( corresponds to almost a free case ) to a high value of 1000.0 a.u .
are illustrated in figure 2 .
it is to be noted that for very small cavity width @xmath76 , the @xmath0 [ @xmath2 energy eigenvalues of helium and the @xmath69 threshold energy are nearly identical to the corresponding energy eigenvalues of the free ions and they remain almost unaltered upto the cavity width @xmath77 a.u .
we can see from equation ( 1 ) that , for @xmath78 , @xmath79 and thus , @xmath80 which produce no effect of confinement . in between @xmath77
a.u . and 10.0 a.u .
, the energy eigenvalues of helium decrease monotonically and ultimately saturates at ( @xmath81 ) a.u . in a similar fashion , the threshold energy @xmath69 saturates at ( @xmath82 )
this feature is physically consistent as we can note from equation ( 1 ) that for @xmath83 , the cavity constant @xmath84 , so that @xmath85 . thus the one and two electron energy levels will undergo a downward shift by @xmath17 and 2@xmath17 respectively for @xmath86 .
the variation of the ionization potential ( in ev ) _
i.e. _ the energy required to ionize one electron from the ground state ( @xmath87 ) of helium atom is plotted against the width ( @xmath20 ) of the cavity in figure 3 . in accordance with the variation of energy eigenvalues of helium and its one - electron subsystem , it is evident from figure 3 that , the ip is identical with the vacuum ip for low values of @xmath20 while for high values of @xmath20 , it increases by an amount @xmath88 ev .
it is thus evident from figures 2 and 3 that rate of variation of energy values of the ions are significant when the size of the confining cavity is of the order of atomic dimensions .
it is also remarkable that the stability of an impurity atom can be controlled by suitably tuning the size of a qd _ i.e. _ the depth and width of the representing cavity .
+ an enlarged view of the stabilization diagram ( given in figure 1 ) for @xmath1 states of @xmath75 within the energy range @xmath89 a.u .
to @xmath90 a.u . is given in figure 4 .
the @xmath1 states of @xmath75 below @xmath6 ionization threshold of @xmath71 can arise due to @xmath3 and @xmath91 ( @xmath92 ) configurations . from a closer look at figure 4
, we can see that for a short range of @xmath48 , each eigenroot between @xmath50 and @xmath93 ionization thresholds of @xmath7 becomes almost flat in the vicinity of avoided crossings in the neighborhood of different energies . in order to calculate the exact resonance parameters ,
the density of states ( dos ) @xmath94 is calculated by evaluating the inverse of the slope at a number of points near these flat plateaus of each energy eigenroot using the formula @xcite given by : @xmath95 the estimated dos @xmath96 is then fitted to the following lorentzian form @xcite @xmath97 where , @xmath98 is the baseline background , @xmath99 is the total area under the curve from the baseline , @xmath100 gives the position of the center of the peak of the curve and represents the full width of the peak of the curve at half maxima . among different fitting curves for each eigenroot corresponding to a particular resonance state , the best fitting curve
i.e. _ with least @xmath101 and the square of correlation ( @xmath57 ) closer to unity @xcite leads to the desired resonance energy ( @xmath100 ) and width ( @xmath102 ) .
the evaluation of dos following the fitting procedure has been repeated for each width of the confining potential ( @xmath20 ) .
for example , the calculated dos and the corresponding fitted lorentzian for the @xmath23 ( @xmath1 ) resonance state of @xmath75 below @xmath69 threshold for cavity width @xmath103 a.u .
( given in figure 5 ) yields resonance position @xmath100 at @xmath104 a.u . and
width @xmath59 = @xmath105 a.u .
+ the estimated resonance energies of doubly excited @xmath3 [ @xmath4 ( @xmath1 ) and @xmath5 [ @xmath4 ( @xmath1 ) states of helium and corresponding @xmath106 and @xmath67 threshold energies for the cavity depth @xmath17 = 0.2 a.u . and cavity width ( @xmath20 ) ranging from 0.001 a.u . to 1000 a.u . are given in figure-6 , while the variations of resonance energies ( @xmath100 ) of @xmath5 [ @xmath4 ( @xmath1 ) states and corresponding @xmath106 and @xmath67 threshold energies versus @xmath20 are given in figure-7 .
we have noted the following points . 1 .
it is clear from figures-6 and 7 that for @xmath20 = 0.001 a.u .
, the @xmath7 ( @xmath106 ) and @xmath7 ( @xmath67 ) states are degenerate and coincide with the energy value of @xmath6 ionization threshold of free @xmath7 ion . as @xmath20 increases , the @xmath7 ( @xmath106 ) and @xmath7 ( @xmath67 ) states become non - degenerate .
initially , the @xmath106 level of @xmath30 lies energetically below the @xmath67 level for @xmath107 up to 0.5 a.u . at @xmath107 = 1.0 a.u .
, the @xmath106 state moves above the @xmath67 level .
these results exhibits that an ` _ _ incidental degeneracy _
_ ' takes place for @xmath106 and @xmath67 states of @xmath30 at some value of @xmath107 between 0.5 and 1.0 a.u .
and then a ` _ _ level crossing _
_ ' occurs between two states having different symmetry properties .
finally these states become degenerate again for @xmath108 a.u . the incidental degeneracy for @xmath71 and
@xmath109 states occur for @xmath20 in the range @xmath110 .
such incidental degeneracy and subsequent level crossing phenomenon are being noted earlier by sen _ et .
@xcite in case of cage confined hydrogen atom and by bhattacharyya _
@xcite in case of helium - like ions within strongly coupled plasma environment .
it is seen from both figures-6 and 7 that all the resonance energies ( @xmath100 ) are almost unaltered up to @xmath20 = 0.5 a.u .
, then decrease rapidly up to @xmath20 = 20.0 a.u . and ultimately saturates . for low values of @xmath20 ( say 0.001 a.u .
) the resonance energies are identical with those of the free _ he _ atom whereas for @xmath20 = 1000.0 a.u .
the resonance energies are equal to those of free _ he _ atom plus 0.4 a.u .
( @xmath111 2.0@xmath112 ) .
thus , for a given depth ( @xmath17 ) of the finite oscillator potential , the variations of energies of the bound states and the resonance states of helium _ w.r.t . _
the width of the cavity ( @xmath20 ) are nearly identical .
the variation of widths ( @xmath102 ) of @xmath3 and @xmath5 ( @xmath1 ) @xmath113 $ ] resonance states _ w.r.t .
_ @xmath20 are given in figures-8 and 9 respectively .
a closer look at the figures 8 and 9 leads us to the following observations . 1 .
in general , it can be argued that the variation of widths shows an oscillatory behavior which are more pronounced for the higher excited states .
it is worthwhile to mention here that recently chakraborty and ho @xcite also reported such oscillation of resonance width ( @xmath102 ) for @xmath23 ( @xmath32 ) state of qd confined helium atom .
this feature clearly indicates a possibility of controlling the autoionization lifetime of doubly excited states of two - electron ions by tuning the parameters of the confining fo potential representing the quantum dot .
the variations of widths of @xmath23 and @xmath114 ( @xmath24 ) states with respect to @xmath20 are exactly opposite in nature . for @xmath24 state originating from @xmath23 configuration ,
the autoionization width first decreases in the range @xmath115 and after reaching the minima , it shows a large bump around @xmath116 6.0 a.u .
after that it starts to decrease and finally the autoionization width saturates where it becomes equal to that of a free _ he _ atom .
in contrast , for @xmath114 state , the autoionization width first increases for @xmath117 and then shows a large dip approximately at the same value of @xmath20 for which the @xmath23 state shows the bump .
3 . the values of @xmath20 corresponding to the largest bump in the values of autoionization widths ( @xmath102 ) of @xmath3 states and the lowest dip for @xmath5 states shift towards the higher values of the cavity width ( @xmath20 ) for higher excited states . inside the qd _
i.e. _ due to the presence of the surrounding fo potential , the charge distribution of the impurity ion gets reoriented which produces the behavioral changes as compared to a free ion .
the nodes or antinodes of the resonance wavefunction lie at the boundary of the qd cavity and the interference caused inside the cavity gives rise to the oscillatory behavior of the resonance widths @xcite .
the number of nodes or antinodes of the wavefunction increases for high - lying resonance states and the oscillation becomes more prominent .
structural properties of he atom confined in a qd , efficiently modeled by a two - parameter weakly confining fo type potential , has been investigated in the framework of stabilization method using explicitly correlated hylleraas - type basis sets .
it has been observed that the structure of the impurity ion is a sensitive function of the dot size .
for very small values of the cavity width , the system behaves almost like a free ion whereas , for very high cavity widths , a constant shift equal to the depth of the potential are observed in the bound as well as resonance energies .
when the dot size becomes comparable to the dimensions of the impurity atom , the effects are more pronounced and many remarkable behaviors such as increase in ionization potential , oscillations in the widths of two - electron resonance states , incidental degeneracy and subsequent level - crossing phenomena for one - electron ions are observed .
the present work is expected to lead to future investigations on the autoionizing states of different angular momenta for qd confined two - electron systems .
+ + + * acknowledgment * + + tkm gratefully acknowledges financial support under grant number 37(3)/14/27/2014-brns from the department of atomic energy , brns , government of india .
sb acknowledges financial support under grant no .
psw-160/14 - 15(ero ) from university grants commission , government of india .
40 w. jaskolski , phys . rep .
* 1 * ( 1996 ) 271 .
j. sabin , e. brandas and s. a. cruz , adv .
quantum chem .
* 57 * ( 2009 ) 1334 ; * 58 * ( 2009 ) 1297 .
s. chakraborty and y. k. ho , phys .
a * 84 * ( 2011 ) 032515 .
s. bhattacharyya , j. k. saha , p. k. mukherjee and t. k. mukherjee , phys .
* 87 * ( 2013 ) 065305 and references therein .
d. j. norris , a. l. efros and s. c. erwin , science * 319 * ( 2008 ) 1776
. n. l. rosi , j. eckert , m. eddaoudi , d. t. vodak , j. kim , m. okeeffe and o. m. yaghi , science * 300 * ( 2003 ) 1127 . j. l. c. rowsell , e. c. spencer , j. eckert , j. a. k. howard and o. m. yaghi , science * 309 * ( 2005 ) 1350 .
j. a. ludlow , t. g. lee and m. s. pindzola , j. phys .
b * 43 * ( 2010 ) 235202
. _ solvation effects on molecules and biomolecules , computational methods and applications _ , s. canuto ( ed . ) , springer , berlin ( 2008 ) . s. sahoo , y. c. lin and y. k. ho , physica e * 40 * ( 2008 ) 3107 . s. sahoo and y. k. ho , phys .
b * 69 * ( 2004 ) 165323 .
m. bylicki , w. jasklski , a. stachw and j. diaz , phys .
b * 72 * ( 2005 ) 075434 .
y. sajeev and n. moiseyev , phys .
b * 78 * ( 2008 ) 075316 .
m. genkin and e. lindroth , phys .
b * 81 * ( 2010 ) 125315 .
p. winkler , int .
* 100 * ( 2004 ) 1122 .
p. kimani , p. jones and p. winkler , int .
* 108 * ( 2008 ) 2763 .
v. a. mandelshtam , t. r. ravuri and h. s. taylor , phys .
rev . lett .
* 70 * ( 1993 ) 1932 .
j. k. saha and t. k. mukherjee , phys .
a * 80 * ( 2009 ) 022513 j. k. saha , s. bhattacharyya and t. k. mukherjee , j. chem
* 132 * ( 2010 ) 134107 .
j. k. saha , s. bhattacharyya , t. k. mukherjee and p. k. mukherjee , int . j. quant
* 111 * ( 2011 ) 1819 .
j. k. saha , s. bhattacharyya and t. k. mukherjee , int . rev . at .
phys . * 3 *
( 2012 ) 1 .
s. kasthurirangan _
* 111 * ( 2013 ) 243201 .
t. k. mukherjee and p. k. mukherjee , phys .
a * 50 * ( 1994 ) 850 .
j. k. saha , s. bhattacharyya , t. k. mukherjee and p. k. mukherjee , j. phys .
b * 42 * ( 2009 ) 245701 .
m. bylicki , j. phys .
b * 30 * ( 1997 ) 189 .
j. a. nelder and r. mead , comput .
j. * 7 * ( 1965 ) 308 .
k. d. sen , j. chem .
* 122 * ( 2005 ) 194324 .
s. bhattacharyya , j. k. saha and t. k. mukherjee , phys .
rev . a * 91 * ( 2015 ) 042515 .
jiao and y.k . ho ,
_ electronic structure of quantum confined atoms and molecules _ , k.d .
sen ( ed . ) , springer , switzerland ( 2014 ) p145 . | bound and resonance states of helium atom have been investigated inside a quantum dot by using explicitly correlated hylleraas type basis set within the framework of stabilization method . to be specific , precise energy eigenvalues of bound @xmath0 ( @xmath1 ) [ @xmath2 states and the resonance parameters _
i.e. _ positions and widths of @xmath1 states due to @xmath3 [ @xmath4 and @xmath5 [ @xmath4 configuration of confined helium below @xmath6 ionization threshold of @xmath7 have been estimated .
the two - parameter ( depth and width ) finite oscillator potential is used to represent the confining potential representing the quantum dot .
it has been explicitly demonstrated that electronic structure properties become a sensitive function of the dot size .
it is observed from the calculations of ionization potential that the stability of an impurity ion within quantum dot may be manipulated by varying the confinement parameters .
a possibility of controlling the autoionization lifetime of doubly excited states of two - electron ions by tuning the width of the quantum cavity is also discussed here . + + * pacs numbers : * 31.15.a@xmath8 , 31.15.v@xmath8 , 32.80.zb , 37.10.gh + * key words : * helium atom , variational method , quantum dot , correlation * electronic structure of helium atom in a quantum dot * + [ 8 mm ] jayanta k. saha@xmath9 , s. bhattacharyya@xmath10 and t. k. mukherjee@xmath11 + [ 4 mm ] @xmath12__indian association for the cultivation of science , jadavpur , kolkata 700032 , india _
_ + @xmath13__acharyya prafulla chandra college , new barrackpore , kolkata 700131 , india _ _ + @xmath14__narula institute of technology , agarpara ,
kolkata 700109 , india _ _ + @xmath15 corresponding author ; e - mail : [email protected] |
ngc 6791 is one of the most intriguing open cluster ( oc ) in the galaxy .
it possesses a unique set of properties that have attracted a lot of attention in the last decade .
indeed , ngc 6791 is one of the most massive ( @xmath6 ; platais et al 2011 ) , oldest ( @xmath7 gyr ; king et al .
2005 ; grundahl et al . 2008 ; garcia - berro et al . 2010 ; brogaard et al . 2012 ) and most metal - rich ( [ fe / h ] @xmath8 + 0.40 ; carraro et al .
2006 , gratton et al .
2006 , origlia et al .
2006 ) galactic ocs .
more specifically , this combination of properties is unique for clusters at its galactocentric distance ( @xmath9 ; carraro 2014 ) .
stars with similar age and metallicity are found in the galactic bulge ( bensby et al .
2013 ) where it has been suggested ( jlkov et al .
2012 ) ngc 6791 might have formed .
given its peculiarities , and in particular its mass , ngc 6791 represents an ideal link between ocs and globular clusters .
for this reason it has been widely targeted to understand whether star - to - star variations of light element ( c , n , o , na , o , mg and al ) abundances can be found also in this system as observed in basically all globular clusters ( gratton et al .
this topic is still matter of debate and no consensus has been reached so far on whether the cluster hosts more than one generation of stars .
geisler et al . (
2012 ) presented evidence of na - o anti - correlation among giant stars in ngc 6791 , while recently bragaglia et al .
( 2014 ) and cunha et al .
( 2015 ) have questioned this result finding an homogeneous na abundance ( within @xmath10 dex ) .
ngc 6791 shows also an anomalous horizontal branch ( hb ) with a well populated red clump , as expected for its high metallicity , and a group of hot hb stars ( liebert et al .
1994 , tofflemire et al .
2014 ) , whose origin is still not completely understood ( carraro 2014 ) .
the combined uv flux of the few hot hb stars and its metallicity makes this cluster a fairly good proxy of standard uv - upturn elliptical galaxies ( buzzoni et al .
this result is supported also by the lick indices analysis of its spectral energy distribution .
this plethora of unique properties makes ngc 6791 an extremely interesting object to study and understand .
a number of questions about the nature of this system still arise : how and where could such a stellar system have formed ?
is ngc 6791 a genuine oc ?
did it form close to the bulge ?
how could have survived for so long in the adverse high - density environment of the inner galactic disk , where it presently is ?
one of the still missing key ingredient that could help to understand the origin of such a fascinating system is its original mass . the mass budget of any stellar system is strongly affected by two - body relaxation , stellar evolution , encounters with giant molecular clouds and spiral arms , interactions with the tidal field .
the disruption scenarios and the survival rates of galactic ocs have been studied since the sixties ( spitzer & harm 1958 ; spitzer 1958 ) .
recently , @xmath3-body simulations ( e.g. , gieles et al . 2006 ; gieles & baumgardt 2008 ) have also shown the impact of giant molecular clouds and of the initial structure of ocs on their dissolution .
the oc orbits have generally small eccentricities and they tend to be located at a small height ( @xmath11 ) on the galactic plane , as a consequence they are expected to pass many times trough the galactic disk .
each of these gravitational shocks heat up and compress the cluster which then takes an elongate shape which flattens to its maximum at @xmath12 ( leon 1998 ) . repeated disk - shocking speed - up the disruption of ocs ( combes et al . 1999 ) and after each passage cluster stars rejected in the halo of the system are stripped out by the gravitational field of the galaxy .
these ejected stars are expected to form tidal tails extending far from the cluster inner regions . indeed , davenport & sandquist ( 2010 ) and bergond et al .
( 2001 ) have found that some ocs show a quite elongated and extended structures likely resulting from the interactions with the galactic tidal field . in general
, ocs are thought to dissipate in a timescale of the order of @xmath13 yr ( binney & tremaine 1987 ) . in this context
the existence of very old ocs like ngc 6791 represents an anomaly gyr ) ocs for which such an argument applies are ngc 188 , m 67 and berkley 11 . ] .
this system may have managed to survive so long either because it moved along a preferential orbit or its original mass was much larger than observed today .
in this paper we analyse the two - dimensional ( 2d ) and projected density distributions of ngc 6791 with the aim of understanding whether the cluster suffered from strong tidal interactions that could have significantly decreased its mass .
we find that this system has an elongated structure and shows overdensities in the external regions that are aligned to the direction of the orbital motion .
these features strongly suggest that this system has likely lost part of its mass because of gravitational shocks and interaction with the galactic tidal field .
we use this argument to support the hypothesis that this system was significantly more massive than observed today and we derive an estimate for its original mass by means of a simplified analytic approach .
the paper is organised as follows : in section 2 data - set and data - reduction procedures are described , in section 3 the 2d density map is obtained while in section 4 both the star - count projected density and the surface brightness profiles are analysed ; finally in section 5 we estimate , the initial mass of ngc 6791 .
( x@xmath14,y@xmath14 ) .
the two circles represent the location of core ( @xmath15 ) and tidal radii ( @xmath16 ) .
the box in the upper right corner defines the `` control field '' ( cf ) area.,width=321 ] we used deep proprietary images ( prop i d : opticon 2013b/005 ; pi moitinho ) obtained with the wide field imager megacam mounted at the canada - france - hawaii telescope ( cfht ) .
this camera consists of a mosaic of 36 chips of @xmath17 pixels each , with a pixel scale of @xmath18 pixel@xmath19 providing a total field of view of @xmath20 .
we used four different pointings set up to cover a total field of view of @xmath21 centered approximately on the cluster center ( figure 1 ) . for each pointing four images
acquired both with the @xmath22 and @xmath23 bands and with exposure times @xmath24 sec and @xmath25 sec respectively have been secured .
an appropriate dither pattern of few arcesc have been adopted for each pointing .
this allowed us to cover most of the inter - chip gaps , with the exception of the most prominent and horizontal ones ( see figure 1 ) .
the data were pre - processed ( i.e. bias and flat - field corrected ) by the elixir pipeline developed by the cfht team . by means of an iterative procedure , an adequate number ( @xmath26 ) of isolated and bright stars
have been selected in each chip and band to model the point spread function ( psf ) . then the psf model was applied to all the stellar - like sources at about @xmath27 from the local background by using daophot and the psf - fitting algorithm allstar ( stetson 1987 ) . for each filter and chip
we matched the single - frame catalogues to obtain a master list .
finally , the @xmath22 and @xmath23 master lists were then geometrically matched and combined .
each master list includes the instrumental magnitude , defined as the weighted mean of the single image measurement reported to the system reference frame of the transformation , and the error , which is the standard error of the mean .
the instrumental coordinates have been reported independently for each chip to the absolute astrometric system ( @xmath28 , @xmath29 ) using a large number of stars in common with the dr7 release of the sloan digital sky survey ( abazajian et al .
2007 ) and following the procedure described by dalessandro et al .
( 2009 ) . using the same stars
, we derived for each chip and band the zero - points to report the instrumental magnitudes to the sloan photometric system . at this stage
all the master lists are in the same photometric and astrometric system .
we then merged them to create the final catalogue , which counts about 270,000 stars .
the ( @xmath22 , @xmath30 ) colour - magnitude diagrams ( cmds ) of ngc 6791 and of the surrounding field are shown in figure 2 . in the cmds , it is possible to clearly distinguish the main sequence ( ms ) of the cluster ( left panel of figure 2 ) , extending down to @xmath31 in the innermost regions , corresponding to about five magnitudes below the turn - off . also , it is possible to note that our data suffer of strong saturation starting at @xmath32 .
two almost vertical sequences located around @xmath33 and @xmath34 are clearly visible both in the innermost ( figure 2 left panel ) and external regions ( figure 2 right panel ) . they are populated by m - dwarf and main - sequence galactic disk stars . a differential reddening of about @xmath35 has been estimated for this cluster ( see for example platais et al .
indeed , at the level of the turn - off ( @xmath36 ) , the ms shows a mild broadening not explainable in terms of pure photometric errors .
such an effect disappears at fainter magnitudes ( @xmath37 ) , where the ms starts to bend diagonally in the ( @xmath22 , @xmath30 ) cmd . given its relatively small amplitude
, we will neglect the effect of differential reddening in the following analysis .
, @xmath30 ) cmds of the innermost region of ngc 6791 ( _ left panel _ ) and of the `` control field '' ( _ right panel _ ) .
the black box along the ms of ngc 6791 encloses the region in the cmd where cp stars have been selected .
the shaded area defines the region where fiducial galactic disk ms stars have been selected.,width=321 ] to probe the spatial distribution of ngc 6791 we first analysed the 2d density distribution .
the density analysis of ngc 6791 has been performed in the circular area enclosed within a distance from the cluster center ( see details in section 4 ) @xmath38 , thus covering a @xmath39 area .
we selected stars along the ms in the magnitude range @xmath40 and within three times the local color dispersion about the cluster mean ridge line ( see figure 2 ) .
stars selected in this way define the cluster population " ( cp ) sample .
these selection criteria have been chosen to ensure a good level of completeness and to minimize the contamination from galactic field stars .
variations of the adopted magnitude limits do no affect the final result .
as a first qualitative step , we obtained a rough smoothed density distribution of cp stars . as expected , at a first look , the distribution of cp stars appears to be extremely concentrated around the position of the cluster and it shows some irregularities and structures allover the field of view ( fov ) .
we also note that the north - west quadrant shows on average a smoother distribution than the other quadrants for @xmath41 .
to investigate in more detail the two - dimensional distribution of cp stars , we adopted the widely used optimal matched filter technique ( see for example kuhn et al .
1996 ; rockosi et al .
2002 ; odenkirchen et al .
2001 , 2003 ) .
it has been shown ( davenport & sandquist 2010 ) that this method is robust for identifying low - density stellar populations against a significant background , as in the case of ocs .
the optimal matched filter technique requires to use a fiducial field population as reference . for this purpose , we selected stars in the same magnitude and colour bins as the cp sample and located at @xmath42 from the cluster center in the north - west quadrant ( because of its apparent homogeneous distribution ) .
we named this sample of stars control field " ( cf ; figures 1 and 2 ) .
we computed the densities in the cmds ( hess diagrams ) of both the cp and cf samples .
then we assigned to each star in the cp sample a weight defined as the reciprocal of the fraction of cf stars lying within a fixed range of colour and magnitude from its location in the cmd .
finally , the weighted distribution of star positions was transformed into a smoothed surface density function through the use of a kernel whose width has been chosen to be large enough so as to make negligible the impact of the gaps in the fov .
this procedure yields the surface density distribution shown in figure 3 . as apparent ,
the overall density distribution is elongated and irregular .
in particular , we observe that the innermost @xmath43 are quite spherical , while at larger distance the cluster starts to be distorted mainly along the north - east / south - west direction in an almost symmetrical way respect to the center of the cluster . at @xmath41 ,
two tails running at opposite directions are clearly visible , with the one in the southern direction extending up to distances larger than @xmath44 from the cluster center .
it is important to note that both the external tidal arms and the innermost elongation of the cluster run almost parallel to the absolute proper motion vector ( bedin et al .
2006 ) and to the direction of the galactic center ( galc ; figure 3 ) .
this behaviour is in agreement with a real tidal nature for the observed distortions .
in fact , the orbital motion of ngc 6791 likely develops close to the galactic equatorial plane ( jlkov et al .
2012 ) and , since tidal tails are supposed to lie on the orbital plane ( at least in axisymmetric potentials , see montuori et al .
2007 ) , they are expected to be seen with such a projected alignment .
moreover , other distortions extending almost perpendicularly to the pm direction are clearly visible and their formation could be related to the tidal effect due to the galactic disk ( e.g. , bergond et al .
2001 ) .
the features observed in the 2d density distribution of ngc 6791 are indeed typical evidence of recent stellar mass loss due to dynamical evolution of the system under the tidal influence of the overall galactic potential .
it is interesting to note that the density map of ngc 6791 is qualitatively similar to that observed in m 67 by davenport & sandquist ( 2010 ) and to a weaker extent to other three ocs ( namely ngc 2287 , ngc 2516 and ngc 2548 ) by bergond et al .
( 2001 ) .
we also note that in the south - east quadrant for @xmath45 there is a star density increase that does not seem to be directly connected to the cluster .
interestingly enough , this overdensity appears at decreasing galactic latitude therefore we speculate that it might be due to the density gradient of galactic disk stars . to further check the significance of the anisotropic distribution of ngc 6791 we used for comparison stars selected in the cmd in the range @xmath46 and @xmath47 ( see figure 2 ) .
these are likely ms galactic disk stars and they trace the behaviour of the main contributors to the contamination of cp stars . this sample is supposed to be homogeneously distributed across the fov and actual deviations from a smooth distribution reflect variations in the detection efficiency , galactic field and/or extinction gradients .
we performed on this sub - sample the same analysis done previously for cp stars .
the 2d density map of ms galactic disk stars is shown in figure 4 .
we observe that in general they are uniformly distributed across the fov as expected , but they show a density excess in the external regions of the south - east quadrant .
this evidence further suggests that the relevant feature in the 2d density distribution of cp stars at @xmath45 ( figure 3 ) in this quadrant , is likely due to a gradient of the density of the galactic disk within the fov .
in addition , we divided the fov in ten angular sectors of @xmath48 each with vertex at the cluster center .
then for each sector with angle @xmath49 measured counter - clockwise from the west direction , and for different distances from the cluster center we estimated the ratio between cp stars ( @xmath50 ) , and likely galactic disk ms stars , selected as described before ( @xmath51 ) .
the results for @xmath52 and @xmath53 are shown in figure 5 . in both cases
the ratio shows significant variations as a function of @xmath49 , as expected for a real anisotropic density distribution .
moreover it is also worth noting that different patterns in the central and outermost regions appear .
for @xmath53 the maxima of the ratio occur at @xmath54 and @xmath55 corresponding to the two opposite main tidal arms . and @xmath47 , which are representative of the galactic field contamination ( see section 3.2 ) .
colors are in the same relative scale as in figure 3.,width=321 ] these sanity checks demonstrate that the observed 2d density distribution shown in figure 3 is a real feature of ngc 6791 and can not be ascribed to observational bias or field gradients .
star density distortions , elongation and tidal tails are expected to be detected also in the radial density profile ( see for example odenkirchen et al .
2003 ; capuzzo - dolcetta et al .
therefore as a second step of our analysis , we studied the projected density profile of ngc 6791 .
first we derived the center of gravity ( c@xmath56 ) by using an iterative procedure ( see for example dalessandro et al .
2013 ) and averaging the positions @xmath28 and @xmath29 of cp stars .
we used the center reported by the webda web page as starting guess point of our procedure . to avoid spurious effects due to incompleteness we performed several estimates by adopting different magnitude and distance selections .
in particular , we used three different magnitude bins in the range @xmath40 , with the upper limit ( fainter magnitude ) increasing by 0.5 mag at each step . for each magnitude bin
, we repeated the procedure for four different radial selections , from @xmath57 to @xmath58 with a step of @xmath59 .
we ended up with a total of 20 estimates .
the resulting c@xmath56 is the average of these measures and it is located at @xmath60 and @xmath61 ( r.a.= 290.2261395 dec=37.77366712 ) , with an uncertainty of @xmath62 .
this new determination is located at @xmath63 north - east from the center reported by the webda web page .
while it is not trivial to understand the origin of this discrepancy , we stress here that it has a negligible impact on the results presented in this paper .
( section 3 ) for stars at @xmath52 ( _ upper panel _ ) and at @xmath53 ( _ lower panel_).,width=321 ] the projected density profile of ngc 6791 has been determined using direct counts of cp stars .
we excluded the south - east quadrant from the analysis since in the most external part of this region the star distribution is dominated by ms galactic disk stars ( section 3.1 ) . using the procedure described in dalessandro et al .
( 2013 ) , we divided the considered fov into 18 concentric annuli , each centered on c@xmath56 and suitably split in an adequate number of subsectors .
number counts have been calculated in each subsector and the corresponding densities were obtained dividing them by the sampled area . incomplete area coverage , mainly due to the presence of gaps ( figure 1 ) , have been properly taken into account .
the stellar density of each annulus was then defined as the average of the subsector densities and its standard deviation computed from the variance among the subsectors .
the resulting projected surface density profile is shown in figure 6 .
we estimated the background density contribution by using only stars in the north - west quadrant .
in fact , as discussed before ( see also figure 3 ) this region is free by any evident structure caused either by the cluster elongation or from field stars .
the background contribution has been estimated as the average of the four outermost density measurements ( @xmath64 ) , which define a kind of plateau at @xmath65 stars arcsec@xmath66 .
we then subtracted this value to the observed density profile .
we have verified that this choice is appropriate .
in fact , the contribution to the background due to field stars ( see selection criteria in section 3.1 ) is constant in all quadrants but the south - east one .
the derived density profile has been reproduced by using a single - mass king model ( king 1966 ) .
the fitting procedure is fully described in miocchi et al .
the fit was limited to the innermost @xmath52 to avoid the most elongated regions of the cluster .
the best - fit is obtained for a concentration @xmath67 ( corresponding to a central dimensionless potential @xmath68 ) and core radius @xmath69 .
these parameters are in partial disagreement with those obtained by platais et al .
( 2011 ; @xmath70 , @xmath71 ) which to our knowledge is the only other density profile analysis present in the literature for ngc 6791 .
the authors obtained the density profile using stars with @xmath72 and weighted according to their proper motion .
however their data were limited to the innermost @xmath73 and therefore the determination of the tidal radius and any residual in the background contribution was rather unstable . as apparent in figure 6 , for @xmath74 ( @xmath75 ) the density profile clearly deviates from the behaviour predicted by the king model .
the use of a wilson model ( wilson 1975 ) does not provide any significant improvement on the overall quality of the fit .
we have also checked that even if we do not constrain the model to fit the innermost region , there is no way to satisfactory reproduce the density profile for @xmath75 .
the external part of the projected density profile instead follows a power - law behaviour with an exponent @xmath76 , in agreement with the values found in galactic globular clusters with observed tidal tails ( see for example sollima et al . 2011 and references therein ) and as predicted by theoretical models ( johnston et al .
kpper et al .
( 2010 ) analysed the variation of the power - law slopes of tidal tails of dissolving clusters as a function of the orbital phase , eccentricity and initial cluster densities by means of n - body simulations .
they found that the power - law slopes have typical values @xmath77 that can increase reaching values @xmath78 for systems with highly eccentric orbits and close to or at the apogalacticon , where tidal tails get strongly compressed .
this upper limit is in good agreement with what found for ngc 6791 , however we should note that the analysis by kpper et al . (
2010 ) is performed on clusters moving on orbits perpendicular to the line of sight , while this is not the case for ngc 6791 . as a consequence
the relatively shallow power - law observed in ngc 6791 can be the result also of projection effects .
to further investigate the results just described and check whether they might be caused by incompleteness or bias of our photometric catalogue , we also derived the surface brightness profile by performing the analysis directly on fits images ( see for example dalessandro et al .
we used a low - resolution single - plate eso digitalized sky survey image obtained in optical band and extending up to about @xmath79 from c@xmath56 .
we limited the analysis to the same quadrants used for deriving the star counts density profile and we used the north - west quadrant to estimate the background contribution .
we divided the fov in a similar number of concentric annuli and subsectors . the surface brightness assigned to each annulus
is given by the ratio between the mean of the fluxes measured in the subsectors and the area covered by the annulus .
the resulting surface brightness profile is shown in figure 7 .
first we observe that the king model best - fitting the star counts density profile well reproduces also the surface brightness profile .
more importantly , we note that consistently with what found in the star count projected density profile , for @xmath75 the surface brightness profile starts to deviate from the behaviour predicted by the model and it declines as a power - law .
we can therefore safely conclude that both the 2d density map and the density ( surface brightness ) profile analysis suggest that ngc 6791 is currently experiencing tidal stripping events .
it is worth noting that king et al .
( 2005 ) observed a rather flat mass function in the centre of ngc 6791 by using very deep _ hubble space telescope _
this observational evidence further supports the results obtained in this work .
in fact , for clusters that experienced episodes of mass loss , a flat mass function is expected because of the preferential depletion of low mass stars .
a detailed analysis of the radial variation of the mass function of this system will be presented in an upcoming paper ( e. dalessandro et al .
the observational evidence presented in the previous sections reveal that ngc 6791 might have lost a fraction of its original mass due to environmental effects ( like tidal shocks due to close interactions with giant molecular clouds , spiral arms , the galactic disk and , in general , to the interactions with the galactic tidal field ) or internal dynamical effects ( like two - body relaxation . in this section
we attempt to estimate the total mass lost by ngc 6791 during its evolution and derive its original mass . while many recipes can be used for this purpose ( e.g. , vesperini et al .
2013 ) , we adopted the approach described by lamers et al .
it has the advantage of describing the way the mass of a cluster decreases with time by means of relatively simple analytic expressions .
the analysis by lamers et al .
( 2005 ) accounts for the effect of both stellar and dynamical evolution .
it is important to note that the results obtained with this simplified approach are in good agreement with those obtained by detailed @xmath3-body simulations for clusters in the tidal field of the galaxy ( see also lamers & gieles 2006 ) .
lamers et al .
( 2005 ) provide an expression for the initial mass of the cluster , @xmath80 , in terms of its age ( @xmath81 ) and present mass ( @xmath82 ) : @xmath83^{1/\gamma } \left[1-q_{\mathrm{ev}}(t ) \right]^{-1}\ ] ] @xmath84 is the dissolution time scale parameter .
it is a constant depending on the strength of the tidal field and it basically describes the mass loss due to dissolution processes such as galactic tidal field interactions and shocks due to encounters with giant molecular clouds or spiral arms ( gieles et al . 2006 , 2007 ) .
the smaller values of @xmath84 are typically associated to encounters with molecular clouds and spiral arms , while the larger to the galactic tidal field . also , the longer @xmath84 , the weaker the tidal effect is .
we used @xmath85myr ( corresponding to the disruption time of @xmath86gyr for a cluster with the initial mass of @xmath87m@xmath88 ) , obtained by lamers et al .
( 2005 ) by comparing the distribution of mass and age of ocs in the solar neighbourhood with theoretical predictions .
this choice is reasonable since the current galactocentric distance of ngc 6791 roughly corresponds to that of the sun .
@xmath89 is a dimensionless index , which depends on the cluster initial density distribution and has typical values ranging from about 0.6 to 0.8 , which increases with the concentration , as constrained from theoretical studies ( gieles et al .
2004 ) and observations ( boutloukos et al . 2003 ) .
@xmath89 is usually constrained by the king dimensionless potential @xmath90 , which characterises the concentration of the king models ( king 1966 ) .
we adopted here @xmath91 corresponding to @xmath92 that is a typical value for open clusters and is in agreement with what derived by the analysis of the observed star count density and surface brightness profiles ( @xmath93 ; section 4 ) .
however , it is important to note that the star cluster density distributions change with time , therefore including the present - day value of @xmath90 in eq .
( [ eq : m_ini_approx ] ) represents a crude approximation . in particular , it is well known ( e.g. trenti et al . 2010 ) that the concentration of stellar systems tend to increase ( and @xmath90 accordingly increases ) as the dynamical evolution of star cluster proceeds . as a consequence ,
given the proportionality linking @xmath89 to @xmath90 , if the adopted value of @xmath90 is larger that its initial one , also the adopted value of @xmath89 will represent an upper limit to the correct one .
the function @xmath94 describes the mass loss due to stellar evolution and can be approximated by the following analytical formula : @xmath95 where @xmath96 , @xmath97 , and @xmath98 are coefficients that depend on the cluster metallicity .
the [ fe@xmath99h ] abundance of ngc6791 is about @xmath100 ( @xmath101 ; carraro et al .
2006 , gratton et al .
2006 , origlia et al .
2006 ) leading ( see table 1 and eq.2 by lamers et al . 2005 ) to values of @xmath102 , @xmath103 , and @xmath104 .
( [ eq : m_ini_approx ] ) depends also on the cluster current mass @xmath82 and its current age @xmath81 .
we adopted a current mass @xmath105 as derived by platais et al .
( 2011 ) and age @xmath106 gyr ( grundahl et al . 2008 , garcia - berro et al . 2010 , brogaard et al . 2012 ) .
these values posses significant errors and their exact estimate is still a matter of debate , however it is important to note that the final error in the initial mass estimate is driven mostly by the uncertainty in the dissolution time scale parameter @xmath84 .
the dependence of the cluster initial mass on @xmath84 within the range of @xmath107myr is plotted in fig .
[ fig : mini_t0 ] , together with the spread caused by the uncertainty of the current cluster mass ( lower and upper limit of 4000 and 6000m@xmath88 considered ) and its current age ( lower and upper limit of 7 and 9gyr considered ) . the initial mass of ngc6791 obtained by eq .
( [ eq : m_ini_approx ] ) turns out to be @xmath4@xmath108 , i.e. more than 50 times larger than its present - day mass . for reasonable smaller values of @xmath90 ) , are expected to be completely destroyed in a short timescale ( @xmath109 yr ) . ] the initial mass derived by means of eq .
( [ eq : m_ini_approx ] ) will be slightly larger , but still compatible within the errors with the value quoted above .
it is important to stress that this estimate is based on a simplified approach and on parameters derived by the average behaviours of ocs .
different recipes and assumptions may likely lead to slightly different results . ) as a function of the dissolution time scale parameter @xmath84 .
the red curve shows the dependence derived for the current cluster mass of 5000m@xmath88 and its age of 8gyr ; the dark - grey coloured area corresponds to values derived for mass within 40006000m@xmath88 at age of 8gyr ; the light - grey area corresponds to values derived for the mass of 5000m@xmath88 at an age within 79gyr.,width=321 ]
in this study we have presented a deep , wide field , photometric data - set extending well beyond the tidal radius of the massive open cluster ngc 6791 .
this data - set allowed us to derive updated estimates of its structural parameters , by fitting a king profile over both the star counts and surface brightness profiles .
with respect to previous studies , that were limited in spatial coverage ( platais et al .
2011 ) , the king model best fitting the projected density profile turns out to have larger values of both c and @xmath16 .
more interestingly , for the first time , we detected clear signatures of tidal features in the star distribution of ngc 6791 , in the form of irregular , but evident , elongation and tidal tails .
some of them seem to closely follow the cluster motion , while others are nearly perpendicular to the galactic disk .
these are clearly visible both in the 2d surface density map and in the projected density profile and they represent typical evidence of recent stellar mass loss .
we can therefore argue that ngc 6791 might have lost quite a significant fraction of its original mass while orbiting around the galactic center . by using a simple analytic approach ( lamers et al .
2005 ) , we estimated the mass likely lost by ngc 6791 during its evolution because of the effect of both stellar evolution and dynamical interactions . from this calculation we estimated the mass at its birth to be @xmath4@xmath108 , several tens larger than its present - day mass .
this derivation lends support to a scenario in which ngc 6791 was a much more massive system .
this would qualitatively explain why the cluster could have survived for such a long time ( its age is around 7 - 8 gyr ) contrary to the expectations of current estimates for the destruction rates of ocs in the galaxy . it is interesting to note also that @xmath80 is comparable to the present - day mass of most galactic globular clusters and it is in rough agreement with the initial mass of the least massive clusters hosting multiple stellar populations ( see for example dalessandro et al .
therefore , the estimate of @xmath80 obtained in this paper might set interesting constraints on the initial conditions of stellar systems able to form multiple stellar populations . | we present the first evidence of clear signatures of tidal distortions in the density distribution of the fascinating open cluster ngc 6791 .
we used deep and wide - field data obtained with the canada - france - hawaii - telescope covering a @xmath0 area around the cluster .
the two - dimensional density map obtained with the optimal matched filter technique shows a clear elongation and an irregular distribution starting from @xmath1 from the cluster center . at larger distances , two tails extending in opposite directions beyond the tidal radius
are also visible .
these features are aligned to both the absolute proper motion and to the galactic center directions . moreover , other overdensities appear to be stretched in a direction perpendicular to the galactic plane . accordingly to the behaviour observed in the density map , we find that both the surface brightness and the star count density profiles reveal a departure from a king model starting from @xmath2 from the center .
these observational evidence suggest that ngc 6791 is currently experiencing mass loss likely due to gravitational shocking and interactions with the tidal field .
we use this evidence to argue that ngc 6791 should have lost a significant fraction of its original mass .
a larger initial mass would in fact explain why the cluster survived so long .
using available recipes based on analytic studies and @xmath3-body simulations , we derived the expected mass loss due to stellar evolution and tidal interactions and estimated the initial cluster mass to be @xmath4@xmath5 .
[ firstpage ] open clusters : individual ( ngc 6791 ) ; stars : evolution ; stars : imaging |
kernel techniques are now a standard tool of statistical practice and kernel versions of many methods of classical multivariate statistics have now been created .
a few important examples can be found in @xcite ( see the description of kernel pca , pages 4145 ) and @xcite ( for kernel ica ) , for instance .
there are several ways to describe kernel methods , but one of them is to think of them as classical multivariate techniques using generalized notions of inner - product . a basic input in these techniques is a kernel matrix , that is , an inner - product ( or gram ) matrix , for generalized inner - products . if our vectors of observations are @xmath1 , the kernel matrices studied in this paper have @xmath2 entry @xmath3 or @xmath4 , for a certain @xmath5 .
popular examples include the gaussian kernel [ entries @xmath6 , the sigmoid kernel [ entries @xmath7 and polynomial kernels [ entries @xmath8 .
we refer to @xcite for more examples .
as explained in , for instance , @xcite , kernel techniques allow practitioners to essentially do multivariate analysis in infinite - dimensional spaces , by embedding the data in a infinite - dimensional space through the use of the kernel .
a nice numerical feature is that the embedding need not be specified , and all computations can be made using the finite - dimensional kernel matrix .
kernel techniques also allow users to do certain forms of nonlinear data analysis and dimensionality reduction , which is naturally very desirable . @xcite and @xcite are two interesting relatively recent papers concerned broadly speaking with the same types of inferential questions we have in mind and investigate in this paper , though the settings of these papers is quite different from the one we will work under .
kernel matrices and the closely related laplacian matrices also play a central role in manifold learning [ see , e.g. , @xcite and @xcite for an overview of various techniques ] . in `` classical '' statistics , they have been a mainstay of spatial statistics and geostatistics in particular [ see @xcite ] . in geostatistical applications , it is clear that the dimension of the data is at most 3 . also , in applications of kernel techniques and manifold learning , it is often assumed that the data live on a low - dimensional manifold or structure , the kernel approach allowing us to somehow recover ( at least partially ) this information .
consequently , most theoretical analyses of kernel matrices and kernel or manifold learning techniques have focused on situations where the data is assumed to live on such a low - dimensional structure . in particular , it is often the case that asymptotics are studied under the assumption that the data is i.i.d . from a fixed distribution
independent of the number of points .
some remarkable results have been obtained in this setting [ see @xcite and also @xcite ] .
let us give a brief overview of such results .
in @xcite , the authors prove that if @xmath9 are i.i.d . with distribution @xmath10 , under regularity conditions on the kernel @xmath11 , the @xmath12th largest eigenvalue of the kernel matrix @xmath13 , with entries @xmath14 converges to the @xmath12th largest eigenvalue of the operator @xmath15 defined as @xmath16 in this important paper , the authors were also able to obtain fluctuation behavior for these eigenvalues , under certain technical conditions [ see theorem 5.1 in @xcite ] .
similar first - order convergence results were obtained , at a heuristic level but through interesting arguments , in @xcite .
these results gave theoretical confirmation to practitioners intuition and heuristics that the kernel matrix could be used as a good proxy for the operator @xmath15 on @xmath17 , and hence kernel techniques could be explained and justified through the spectral properties of this operator .
to statisticians well versed in the theory of random matrices , this set of results appears to be similar to results for low - dimensional covariance matrices stating that when the dimension of the data is fixed and the number of observations goes to infinity , the sample covariance matrix is a spectrally consistent estimator of the population covariance matrix [ see , e.g. , @xcite ] .
however , it is well known [ see , e.g. , @xcite , @xcite , @xcite ] that this is not the case when the dimension of the data , @xmath18 , changes with @xmath19 , the number of observations , and in particular when asymptotics are studied under the assumption that @xmath20 has a finite limit .
we refer to the asymptotic setting where @xmath18 and @xmath19 both tend to infinity as the `` high - dimensional '' setting .
we note that given that more and more datasets have observations that are high dimensional , and kernel techniques are used on some of them [ see @xcite ] , it is natural to study kernel random matrices in the high - dimensional setting .
another important reason to study this type of asymptotics is that by keeping track of the effect of the dimension of the data , @xmath18 , and of other parameters of the problem on the results , they might help us give more accurate prediction about the finite - dimensional behavior of certain statistics than the classical `` small @xmath18 , large @xmath19 '' asymptotics .
an example of this phenomenon can be found in the paper @xcite where it turned out in simulation that some of the doubly asymptotic results concerning fluctuation behavior of the largest eigenvalue of a wishart matrix with identity covariance are quite accurate for @xmath18 and @xmath19 as small as 5 or 10 , at least in the right tail of the distribution .
[ we refer the interested reader to @xcite for more details on the specific example we just described . ]
hence , it is also potentially practically important to carry out these theoretical studies for they can be informative even for finite - dimensional considerations . the properties of kernel random matrices under classical random matrix assumptions
have been studied by the author in the recent @xcite .
it was shown there that when the data is high dimensional , for instance @xmath21 , and the operator norm of @xmath22 is , for example , bounded , kernel random matrices essentially act like standard gram/``covariance matrices , '' up to recentering and rescaling , which depend only on @xmath5 .
naturally , a certain scaling is needed to make the problem nondegenerate , and the results we just stated hold , for instance , when @xmath23 , for otherwise the kernel matrix is in general degenerate .
we refer to @xcite for more details and discussions of the relevance of these results in practice . in limited simulations
, we found that the theory agreed with the numerics even when @xmath18 was of the order of several 10 s and @xmath20 was not `` too small '' ( e.g. , @xmath24 ) .
these results came as somewhat of a surprise and seemed to contradict the intuition and numerous positive practical results that have been obtained , since they suggested that the kernel matrices we considered were just a ( centered and scaled ) version of the matrix @xmath25 .
however , it should be noted that the assumptions implied that the data was truly high dimensional .
so an interesting middle ground , from modeling , theoretical and practical points of view is the following : what happens if the data does not live exactly on a fixed - dimensional manifold , but lives `` nearby ? '' in other words , the data is now sampled from a `` noisy '' version of the manifold .
this is the question we study in this paper .
we assume now that the data points @xmath26 we observe are of the form @xmath27 where @xmath28 is the `` signal '' part of the observations ( and live , for instance , on a low - dimensional manifold , e.g. , a three - dimensional sphere ) and @xmath29 is the noise part of the observations ( and is , e.g. , multivariate gaussian in dimension @xmath18 , where @xmath18 might be 100 ) .
we think this is interesting from a practical standpoint because the assumption that the data is exactly on a manifold is perhaps a bit optimistic and the `` noisy manifold '' version is perhaps more in line with what statisticians expect to encounter in practice ( there is a clear analogy with linear regression here ) . from a theoretical standpoint ,
such a model allows us to bridge the two extremes between truly low - dimensional data and fully high - dimensional data . from a modeling standpoint
, we propose to scale the noise so that its norm stays bounded ( or does not grow too fast ) in the asymptotics .
that way , the `` signal '' part of the data is likely to be affected but not totally drowned by the noise .
it is important to note , however , that the noise is not `` small '' in any sense of the word
it is of a size comparable with that of the signal . in the case of spherical noise
( see below for details but note that the gaussian distribution falls into this category ) our results say that , to first - order , the kernel matrix computed from information@xmath0noise data behaves like a kernel matrix computed from the `` signal '' part of the data , but , we might have to use a different kernel than the one we started with .
this other kernel is quite explicit . in the case of dot - product kernel matrices [ i.e. , @xmath30
, the original kernel can be used ( under certain assumptions)so , to first - order , the noise part has no effect on the spectral properties of the kernel matrix .
the results are different when looking at euclidean distance kernels [ i.e. , @xmath31 where the effect of the noise is basically to change the kernel that is used .
this is in any case a quite positive result in that it says that the whole body of work concerning the behavior of kernel random matrices with low - dimensional input data can be used to also study the `` information@xmath0noise '' case the only change being a change of kernels .
the case of elliptical noise is more complicated .
the dot - product kernels results still have the same interpretation .
but the euclidean distance kernels results are not as easy to interpret .
before we start , we set some notation . we use @xmath32 to denote the frobenius norm of the matrix @xmath13 [ so @xmath33 and @xmath34 to denote its operator norm , that is , its largest singular value .
we also use @xmath35 to denote the euclidean norm of the vector @xmath36 .
@xmath37 is shorthand for @xmath38 . unless otherwise noted , functions that are said to be lipschitz are lipschitz with respect to euclidean norm .
we split our results into two parts , according to distributional assumptions on the noise .
one deals with the gaussian - like case , which allows us to give a simple proof of the results .
the second part is about the case where the noise has a distribution that satisfies certain concentration and ellipticity properties .
this is more general and brings the geometry of the problem forward .
it also allows us to study the robustness ( and lack thereof ) of the results to the sphericity of the noise , an assumption that is implicit in the high - dimensional gaussian ( and gaussian - like ) case .
we draw some practical conclusions from our results for the case of spherical noise in section [ subsec : practicalconsequences ] .
we first study a setting where the noise is drawn according to a distribution that is similar to a gaussian , but slightly more general .
[ thm : infoplusnoisegaussiancase ] suppose we observe data @xmath1 in @xmath39 , with @xmath40 where @xmath41 where the @xmath18-dimensional vector @xmath42 has i.i.d .
entries with mean 0 , variance 1 , and fourth moment @xmath43 , and @xmath44 .
we assume that there exists a deterministic vector @xmath45 and a real @xmath46 , possibly dependent on @xmath19 , such that @xmath47 . also , @xmath43 might change with @xmath19 but is assumed to remain bounded .
@xmath48 are i.i.d . , and we also assume that @xmath49 and @xmath48 are independent .
we consider the random matrices @xmath50 with @xmath2 entry @xmath51 where @xmath52 let us call @xmath53 .
let @xmath54 be the matrix with @xmath2th entry @xmath55 assuming only that @xmath43 is bounded uniformly in @xmath19 , we have , for a constant @xmath56 independent of @xmath19 , @xmath18 and @xmath22 , @xmath57 .\ ] ] we place ourselves in the high - dimensional setting where @xmath19 and @xmath18 tend to infinity .
we assume that @xmath58 , as @xmath18 tends to infinity . under these assumptions , for any fixed @xmath59 and @xmath46
, @xmath60 if we further assume that @xmath61 remains , for instance , bounded , the same result holds if we replace the diagonal of @xmath62 by @xmath63 , because @xmath64 and therefore @xmath65 .
the approximating matrix we then get is the matrix with @xmath2th entry @xmath66 , where @xmath67 , that is , a `` pure signal '' matrix involving a different kernel from the one with which we started .
we note that there is a potential measurability issue that we address in the proof .
our theorem really means that we can find a random variable that dominates the `` random element '' @xmath68 and goes to 0 in probability .
( this measurability issue could also be addressed through separability arguments but outer - probability statements suffice for our purposes in this paper . )
a subcase of our result is the case of gaussian noise : then @xmath42 is @xmath69 and our result naturally applies .
we also note that @xmath70 can change with @xmath19 .
the class of functions we consider is fixed in the last statement of the theorem but if we were to look at a sequence of kernels we could pick a different function in the class @xmath71 for each @xmath19 [ the proof also applies to matrices with entries @xmath72 , where the functions considered also depend on @xmath2 , but we present the results with a function @xmath5 common to all entries ] .
it should also be noted that the proof technique allows us to deal with classes of functions that vary with @xmath19 : we could have a varying @xmath73 .
as ( [ eq : explicitcontrolerrorgaussianlikecase ] ) makes clear , the approximation result will hold as soon as the right - hand side of ( [ eq : explicitcontrolerrorgaussianlikecase ] ) goes to 0 asymptotically , that is , @xmath74 .
finally , we work here with uniformly lipschitz functions .
the proof technique carries over to other classes , such as certain classes of hlder functions , but the bounds would be different .
proof of theorem [ thm : infoplusnoisegaussiancase ] the strategy is to use the same entry - wise expansion approach that was used in @xcite .
to do so , we remark that @xmath75 remains essentially constant [ across @xmath2 ] in the setting we are considering this is a consequence of the `` spherical '' nature of high - dimensional gaussian distributions .
we can therefore try to approximate @xmath77 by @xmath78 and all we need to do is to show that the remainder is small .
we also note that if , as we assume , @xmath79 , then @xmath80 , since @xmath81 .
- _ work conditional on _ @xmath82 , _ for _ @xmath83 .
we clearly have @xmath84 let us study the various parts of this expansion .
conditional on @xmath85 , if we call @xmath86 , we see easily that @xmath87 and @xmath88 note that @xmath89 , which we denote @xmath90 , has i.i.d .
entries , with mean 0 , variance @xmath91 and fourth moment @xmath92 .
we call @xmath93 and @xmath94 with this notation , we have @xmath95 therefore , for any function @xmath5 in @xmath96 , @xmath97 and hence , @xmath98 ^ 2 \leq2 c_0(n)^2 [ \beta_{i ,
j}^2 + 4\alpha_{i , j}^2 ] .\ ] ] we naturally also have @xmath99 ^ 2 \leq2 c_0(n)^2 [ \beta _ { i , j}^2 + 4\alpha_{i , j}^2 ] .\ ] ] so we have found a random variable @xmath100 $ ] that dominates the random element @xmath101 ^ 2 $ ] .
one might be concerned about the measurability of @xmath102but by using outer expectations [ see @xcite , page 258 ] , we can completely bypass this potential problem . in
what follows , we denote by @xmath103 an outer expectation .
( though this technical point does not shed further light on the problem , it naturally needs to be addressed . ) hence , @xmath104 let us focus on @xmath105 for a moment .
let us call @xmath106 .
we first note that @xmath107 .
in particular , @xmath108 so @xmath109 .
therefore , @xmath110 .
now recall the results found , for instance , in lemma a-1 in @xcite : if the vector @xmath111 has i.i.d .
entries with mean 0 , variance @xmath112 and fourth moment @xmath113 , and if @xmath13 is a symmetric matrix , @xmath114 where @xmath115 is the hadamard product of @xmath13 with itself , that is , the entrywise product of two matrices . applying this result in our setting [ i.e. , using the moments ( given above ) of @xmath90 , which has i.i.d .
entries , in the previous formula ] gives @xmath116 it is easy to see that @xmath117 , since @xmath118 and @xmath119
. therefore , @xmath120 we note that under our assumptions on @xmath121 and the fact that @xmath43 remains bounded in @xmath19 ( and therefore @xmath18 ) , this term will go to 0 as @xmath122 . on the other hand , because @xmath123 , and because @xmath124 and @xmath125 , we have @xmath126 hence , we have for @xmath56 a constant independent of @xmath22 , @xmath18 and @xmath19 , @xmath127 .\end{aligned}\ ] ] this inequality allows us to conclude that , for another constant @xmath56 , @xmath128 , \ ] ] since clearly , @xmath129 under the assumption that @xmath130 exists and is less than @xmath131 , we finally conclude that @xmath132,\ ] ] and ( [ eq : explicitcontrolerrorgaussianlikecase ] ) is shown . therefore , under our assumptions , @xmath133 hence , when @xmath19 and @xmath18 tend to @xmath134 , @xmath135 as announced in the theorem .
the proof of theorem [ thm : infoplusnoisegaussiancase ] makes clear that the heart of our argument is geometric : we exploit the fact that @xmath136 is essentially constant across pairs @xmath2 .
it is therefore natural to try to extend the theorem to more general assumptions about the noise distribution than the gaussian - like one we worked under previously .
it is also important to understand the impact of the implicit geometric assumptions ( i.e. , sphericity of the noise ) that are made and in particular the robustness of our results against these geometric assumptions .
we extend the results in two directions .
first , we investigate the generalization of our gaussian - like results to the setting of euclidean - distance kernel random matrices , when the noise is distributed according to a distribution satisfying a concentration inequality multiplied by a random variable , that is , a generalization of elliptical distributions .
this allows us to show that the gaussian - like results of theorem [ thm : infoplusnoisegaussiancase ] essentially hold under much weaker assumptions on the noise distribution , as long as the gaussian geometry ( i.e. , a spherical geometry ) is preserved ( see corollary [ coro : sphericalnoisekernels ] ) .
the results of theorem [ thm : infoplusnoiseconccase ] show that breaking the gaussian geometry results in quite different approximation results .
we also discuss in theorem [ thm : infoplusnoiseconccasedotproductskernels ] the situation of inner - product kernel random matrices under the same `` generalized elliptical '' assumptions on the noise .
we have the following theorem .
[ thm : infoplusnoiseconccase ] suppose we observe data @xmath137 in @xmath39 , with @xmath138 we place ourselves in the high - dimensional setting where @xmath19 and @xmath18 tend to infinity .
we assume that @xmath44 .
@xmath48 are i.i.d . with @xmath139 , and we also assume that @xmath140 and @xmath48 are independent .
@xmath141 are random variables independent of @xmath48 .
we now assume that the distribution of @xmath29 is such that , for any 1-lipschitz function @xmath142 , if @xmath143 , @xmath144 where for simplicity we assume that @xmath145 , @xmath56 and @xmath146 are independent of @xmath18 .
we call @xmath147 and assume that @xmath61 stays bounded as @xmath122 .
we assume that @xmath148 , @xmath149 $ ] , where @xmath150 and @xmath151 are deterministic sequences depending on @xmath18 .
we assume without loss of generality that @xmath152 .
calling @xmath153 , we assume that there exists @xmath154 such that @xmath155 and @xmath156 such that @xmath157 then we have @xmath158 \bigr|\rightarrow0\qquad \mbox{in probability}.\ ] ] we call @xmath159 , and suppose we pick @xmath160 such that @xmath161 .
( note that @xmath162 is always a possibility . )
we call , for @xmath163 given , @xmath164 $ ] , and @xmath165 we consider the random matrices @xmath50 with @xmath2 entry
@xmath166 let us call @xmath54 the matrix with @xmath2th entry @xmath167 we have , for any given @xmath46 and @xmath163 ,
@xmath168 we have the following corollary in the case of `` spherical '' noise , which is a generalization of the gaussian - like case considered in theorem [ thm : infoplusnoisegaussiancase ] .
[ coro : sphericalnoisekernels ] suppose we observe data @xmath1 in @xmath39 , with @xmath40 where @xmath28 and @xmath29 satisfy the same assumptions as in theorem [ thm : infoplusnoiseconccase ] [ with @xmath169
. then the results of theorem [ thm : infoplusnoiseconccase ] apply with @xmath170\ ] ] and @xmath171 as in theorem [ thm : infoplusnoisegaussiancase ] , we deal with potential measurability issues concerning the @xmath172 in the proof .
our theorem is really that we can find a random variable that goes to 0 with probability 1 and dominates the random element @xmath173an outer - probability statement .
this theorem generalizes theorem [ thm : infoplusnoisegaussiancase ] in two ways .
the `` spherical '' case , detailed in corollary [ coro : sphericalnoisekernels ] , is a more general version of theorem [ thm : infoplusnoisegaussiancase ] limited to gaussian noise .
this is because the gaussian setting corresponds to @xmath174 and @xmath175 .
however , assuming `` only '' concentration inequalities allows us to handle much more complicated structures for the noise distribution .
some examples are given below .
we also note that if the @xmath28 s ( i.e. , the signal part of the @xmath9 s ) are sampled , for instance , from a fixed manifold of finite euclidean diameter , the conditions on @xmath176 are automatically satisfied , with @xmath154 being the euclidean diameter of the corresponding manifold .
another generalization is `` geometric '' : by allowing @xmath141 to vary with @xmath177 , we move away from the spherical geometry of high - dimensional gaussian vectors ( and generalizations ) , to a more `` elliptical '' setting . hence , our results show clearly the potential limitations and the structural assumptions that are made when one assumes gaussianity of the noise .
theorem [ thm : infoplusnoiseconccase ] and corollary [ coro : sphericalnoisekernels ] show that the gaussian - like results of theorem [ thm : infoplusnoisegaussiancase ] are not robust against a change in the geometry of the noise .
we note however that if @xmath141 is independent of @xmath29 and @xmath178 , @xmath179 , so all the noise models have the same covariance but they may yield different approximating matrices and hence different spectral behavior for our information@xmath0noise models .
however , the spherical results have the advantage of having simple interpretations . in the setting of corollary [ coro : sphericalnoisekernels ] , if we assume that @xmath180 and @xmath181 are uniformly bounded ( in @xmath19 ) over the class of functions we consider , we can replace the diagonal of @xmath62 by @xmath63 and have the same approximation results . then the `` new '' @xmath62 is a kernel matrix computed from the signal part of the data with the new kernel @xmath182 . to make our result more concrete , we give a few examples of distributions for which the concentration assumptions on @xmath29 are satisfied : * gaussian random variables , for which we have @xmath183 .
we refer to ledoux [ ( @xcite ) , theorem 2.7 ] for a justification of this claim .
* vectors of the type @xmath184 where @xmath36 is uniformly distributed on the unit ( @xmath185-)sphere in dimension @xmath18 .
theorem 2.3 in @xcite shows that our assumptions are satisfied , with @xmath186 , after noticing that a 1-lipschitz function with respect to euclidean norm is also 1-lipschitz with respect to the geodesic distance on the sphere .
* vectors @xmath187 , with @xmath36 uniformly distributed on the unit ( @xmath185-)sphere in @xmath39 and with @xmath188 having bounded operator norm .
* vectors of the type @xmath189 , @xmath190 , where @xmath36 is uniformly distributed in the unit @xmath191 ball or sphere in @xmath39 .
( see ledoux [ ( @xcite ) , theorem 4.21 ] which refers to @xcite as the source of the theorem . ) in this case , @xmath145 depends only on @xmath146 .
* vectors with log - concave density of the type @xmath192 , with the hessian of @xmath193 satisfying , for all @xmath194 , @xmath195 , where @xmath196 is the real that appears in our assumptions .
see ledoux [ ( @xcite ) , theorem 2.7 ] for a justification .
* vectors @xmath36 distributed according to a ( centered ) gaussian copula , with corresponding correlation matrix , @xmath197 , having @xmath198 bounded .
we refer to @xcite for a justification of the fact that our assumptions are satisfied .
[ if @xmath199 has a gaussian copula distribution , then its @xmath177th entry satisfy @xmath200 , where @xmath201 is multivariate normal with covariance matrix @xmath197 , @xmath197 being a correlation matrix , that is , its diagonal is 1 . here
@xmath202 is the cumulative distribution function of a standard normal distribution .
taking @xmath203 gives a centered gaussian copula .
] this last example is intended to show that the result can handle quite complicated and nonlinear noise structure .
we note that to justify that the assumptions of the theorem are satisfied , it is enough to be able to show concentration around the mean or the median , as proposition 1.8 in @xcite makes clear .
the reader might feel that the assumptions concerning the boundedness of the @xmath141 s will be limiting in practice .
we note that the same proof essentially goes through if we just require that @xmath204 s belong to the interval @xmath205 $ ] with probability going to 1 , but this requires a little bit more conditioning and we leave the details , which are not difficult , to the interested reader .
so for instance , if we had a tail condition on @xmath204 , we could bound latexmath:[$\max
so this boundedness condition is here just to make the exposition simpler and is not particularly limiting in our opinion .
on the other hand , we note that our conditions allow dependence in the @xmath141 s and are therefore rather weak requirements . finally , the theorem as stated is for a fixed @xmath131 , though the class of functions we are considering might vary with @xmath19 and @xmath18 through the influence of @xmath207 .
the proof makes clear that @xmath131 could also vary with @xmath19 and @xmath18 .
we discuss in more details the necessary adjustments after the proof .
proof of theorem [ thm : infoplusnoiseconccase ] we use the notation @xmath208 and @xmath209 to denote probability conditional on @xmath85 .
we call @xmath210 .
let us also call @xmath211 ; similarly , @xmath212 denotes probability conditional on @xmath213 .
we call @xmath214
. we will start by working conditionally on @xmath213 and eventually decondition our results .
we assume from now on that the @xmath213 we work with is such that @xmath215 .
note that @xmath216 by assumption and also @xmath217 .
the main idea now is that , in a strong sense , @xmath218 where @xmath219 . to show this formally , we write @xmath220=2\alpha_{i , j}+\beta_{i , j } , \ ] ] where @xmath221 and @xmath222 our aim is to show that , as @xmath19 and @xmath18 tend to infinity , @xmath223 - _ on _ @xmath224 .
note that if @xmath225 , @xmath226 .
clearly , @xmath227 since we assumed that @xmath228 , we see that the function @xmath229 is lipschitz ( with respect to euclidean norm ) , with lipschitz constant smaller than @xmath230 , when @xmath231 is in @xmath232 . also , since @xmath139 , @xmath233 , where the expectation is conditional on @xmath213 .
hence , our concentration assumptions on @xmath29 imply that @xmath234\bigr)^b\bigr ) .\ ] ] therefore , if we use a simple union bound , we get @xmath235\bigr)^b\bigr ) .\ ] ] in particular , if we pick , for @xmath156 , @xmath236 , we see that @xmath237\bigr)^b\bigr)\\ & = & 2c \exp ( -2(\log n)^{\varepsilon})\rightarrow0.\end{aligned}\ ] ] since @xmath238 and since the latter goes to 0 , we have , unconditionally , @xmath239 - _ on _ @xmath240 .
we see that if @xmath241 and @xmath242 are vectors in @xmath39 , the map @xmath243 is @xmath244 , by the triangle inequality .
therefore , using propositions 1.11 and 1.7 in @xcite [ and using the fact that @xmath245 as @xmath246 and @xmath247 is continuous when using the latter ] , we conclude that @xmath248 if now @xmath249 , and if @xmath250 , @xmath251 where @xmath15 is a constant which does not depend on @xmath213 .
so we conclude that unconditionally , if @xmath252 note also that under our assumptions , @xmath253 .
recall that we aim to show that @xmath254 let us first work on @xmath255 using the fact that @xmath256 , and therefore , @xmath257 if we choose @xmath258 and @xmath259 , we see that the previous equation becomes @xmath260 therefore , if we can show that @xmath261 goes to 0 in probability , we will have @xmath262 in probability . using the concentration result given in ( [ eq : conditionalconcentration ] ) , in connection with proposition 1.9 in @xcite and
a slight modification explained in @xcite , we have @xmath263\\[-8pt ] & \leq & \frac{r^2_{\infty } ( p)}{p } \frac{32 c}{b(c_0)^{2/b}}\gamma(2/b)=r^2_{\infty}(p)\frac { \kappa_b}{p } .\nonumber\end{aligned}\ ] ] using our assumption that @xmath61 remains bounded , we see that @xmath264 therefore , for some @xmath15 independent of @xmath18 , @xmath265 with probability going to 1
. our assumptions also guarantee that @xmath266 , so we conclude that , for a constant @xmath15 independent of @xmath18 , @xmath267 using ( [ eq : controlgammaijsquare ] ) , we have the deterministic inequality @xmath268 so we can finally conclude that with high probability @xmath269 putting all these elements together , we see that when @xmath270 we can find a constant @xmath15 such that @xmath271 in other words , @xmath272 \bigr|>ku_p \bigr)\rightarrow0 .\ ] ] this establishes ( a strong form of ) the first part of the theorem , that is , ( [ eq : interpointdistanceellipcase ] ) .
- _ second part of the theorem _ [ _ equation _ ( [ eq : mainresconcellipcase ] ) ] . to get to the second part
, we recall that , assuming that @xmath5 is @xmath131-lipschitz on an interval containing @xmath273 , we have @xmath274 let us define , for @xmath163 given , the event @xmath275\ } , \ ] ] and the random element @xmath276 when @xmath277 is true , all the pairs @xmath278 are in @xmath207 : the part concerning @xmath279 is obvious , and the one concerning @xmath280 comes from the definition of @xmath277 .
so when @xmath277 is true , we also have @xmath281 let us now consider the random variable @xmath282 such that @xmath283 on @xmath277 and @xmath134 otherwise , so @xmath284 .
our remark above shows that @xmath285 now , we see from our assumptions about @xmath286 , ( [ eq : controlmaxerrorgeomapproxentrywise ] ) and the fact that @xmath287 , that for any @xmath163 , @xmath288 .
so we have @xmath289 also , @xmath290 with probability tending to 1 , so we can conclude that @xmath291 hence , we also have @xmath292 where this statement might have to be understood in terms of outerprobabilities hence the @xmath293 instead of @xmath10 .
[ see @xcite , page 258 . in plain english
, we have found a random variable , latexmath:[${\tau_n \max_{i\neq j } } |2\alpha_{i , j}+\beta _ { i , j } which is larger than the random element @xmath102 . ] in other respects , we have , for all @xmath295 , @xmath296 since @xmath297 therefore , @xmath298 where once again this statement may have to be understood in terms of outer probabilities .
the result stated in ( [ eq : mainresconcellipcase ] ) is proved .
we mentioned before the proof the possibility that we might let @xmath131 vary with @xmath19 and @xmath18 and still get a good approximation result .
this can be done by looking at ( [ eq : keyeqcontrolsuperrorellipconccaseproof ] ) above : @xmath102 is less than @xmath299 with high probability , so when @xmath300 the main approximation result of theorem [ thm : infoplusnoiseconccase ] holds , for a @xmath131 and therefore a class of functions , that vary with @xmath19 ( and @xmath18 ) .
we now turn our attention to kernel matrices of the form @xmath301 which are also of interest in practice . in that setting , we are able to obtain results similar in flavor to theorem [ thm : infoplusnoiseconccase ] , with slight modifications on the assumptions we make about @xmath5 .
[ thm : infoplusnoiseconccasedotproductskernels ] suppose we observe data @xmath302 in @xmath39 , with @xmath138 we place ourselves in the high - dimensional setting where @xmath19 and @xmath18 tend to infinity .
we assume that @xmath44 .
@xmath48 are i.i.d . with @xmath139 , and we also assume that @xmath303 and @xmath48 are independent . @xmath304 are assumed to be independent of @xmath48 .
we also assume that we can find a deterministic sequence @xmath305 such that @xmath306 and @xmath152 .
we assume that the distribution of @xmath29 is such for any 1-lipschitz function @xmath142 ( with respect to euclidean norm ) , if @xmath143 , @xmath144 where for simplicity we assume that @xmath145 , @xmath56 and @xmath146 are independent of @xmath18 .
we call @xmath147 and assume that @xmath61 stays bounded as @xmath122 .
we call @xmath307 , and @xmath154 a real such that @xmath308 .
we assume that there exists @xmath156 such that @xmath309 we then have @xmath310 we call @xmath311 $ ] and @xmath312 we then consider the random matrices @xmath50 with @xmath2 entry
@xmath313 let us call @xmath62 the matrix with @xmath2th entry @xmath314 we have , for any @xmath46 and @xmath163 , @xmath315 we note that under our assumptions , we also have @xmath316 , with high probability , and uniformly in @xmath5 in @xmath317 . therefore , when @xmath318 , the result is also valid if we replace the diagonal of @xmath54 by @xmath319in which case the new approximating matrix is the kernel matrix computed from the signal part of the data . furthermore , the same argument shows that we get a valid operator norm approximation of @xmath13 by this `` pure signal '' matrix as soon as @xmath320 tends to 0
. the same measurability issues as in the previous theorems might arise here and the statement should be understood as before : we can find a random variable going to 0 in probability that is larger than the random element @xmath321 . finally , let us note that once again the theorem is stated for a fixed @xmath131 [ and hence for an essentially fixed ( with @xmath19 ) class of functions , though some changes in this class might come from varying @xmath322 , but the proof allows us to deal with a varying @xmath323 .
the adjustments are very similar to the ones we discussed after the proof of theorem [ thm : infoplusnoiseconccase ] and we leave them to the interested reader .
proof of theorem [ thm : infoplusnoiseconccasedotproductskernels ] the proof is quite similar to that of theorem [ thm : infoplusnoiseconccase ] , so we mostly outline the differences and use the same notation as before .
we now have to focus on @xmath324 the analysis of @xmath325 is entirely similar to our analysis of @xmath326 in the proof of theorem [ thm : infoplusnoiseconccase ] .
the key remark now is that as function of @xmath29 , when @xmath327 , it is , with the new definition of @xmath154 , @xmath328-lipschitz with respect to euclidean norm .
so we immediately have , with the new definition of @xmath154 : if @xmath329 , and @xmath330 , for some @xmath331 which does not depend on @xmath213 , @xmath332 now , since @xmath333 , we conclude as before that @xmath334 on the other hand , using the fact that @xmath335 , and analyzing the concentration properties of @xmath336 in the same way as we did those of @xmath337 , we conclude that if @xmath338 , we can find a constant @xmath15 such that @xmath339 and @xmath340 similar arguments , relying on the fact that is obviously 1-lipschitz with respect to euclidean norm , also lead to the fact that @xmath341 therefore , we can find @xmath15 , greater than 1 without loss of generality , such that @xmath342 we can therefore conclude that @xmath343 if @xmath344 , then both @xmath345 and @xmath346 tend to 0 . therefore , under our assumptions , @xmath347 so we have shown the first assertion of the theorem .
the final step of the proof is now clear : we have , for all @xmath2 , @xmath348 when for all @xmath2 , @xmath349 and @xmath350 are in @xmath351 .
this event happens with probability going to 1 under our assumptions .
so following the same approach as before and dealing with measurability in the same way , we have , with probability going to 1 , @xmath352 so we conclude that @xmath353 from this statement , we get in the same manner as before , @xmath354 as before , the equations above show that if @xmath355 , the same approximation result holds , now with a varying @xmath323 .
our aim in giving approximation results is naturally to use existing knowledge concerning the approximating matrix to reach conclusions concerning the information@xmath0noise kernel matrices that are of interest here . in particular
, we have in mind situations where the `` signal '' part of the data , that is , what we called @xmath286 in the theorems , and @xmath5 [ or @xmath356 , with @xmath61 being as defined in theorems [ thm : infoplusnoisegaussiancase ] or [ thm : infoplusnoiseconccase ] ] are such that the assumptions of theorems 3.1 or 5.1 in @xcite are satisfied , in which case we can approximate the eigenvalues of @xmath62 by those of the corresponding operator in @xmath17 . in this setting the matrix @xmath357 , which is normalized so its entries are of order @xmath358 has a nondegenerate limit , which is why we considered for our kernel matrices the normalization @xmath359 .
[ this normalization by @xmath358 makes our proofs considerably simpler than the ones given in @xcite . ]
another potentially interesting application is the case where the signal part of the data is sampled i.i.d . from a manifold with bounded euclidean diameter , in which case our results are clearly applicable .
the practical interest of the theorems we obtained above lie in the fact that the frobenius norm is larger than the operator norm , and therefore all of our results also hold in operator norm .
now we recall the discussion in el karoui [ ( @xcite ) , section 3.3 ] , where we explained that consistency in operator norm implies consistency of eigenvalues and consistency of eigenspaces corresponding to separated eigenvalues [ as consequences of weyl s inequality and the davis kahane @xmath360 theorem
see @xcite and @xcite ] .
theorems [ thm : infoplusnoisegaussiancase ] , [ thm : infoplusnoiseconccase ] , [ thm : infoplusnoiseconccasedotproductskernels ] therefore imply that under the assumptions stated there , the spectral properties of the matrix @xmath13 can be deduced from those of the matrix @xmath62 . in particular , for techniques such as kernel pca
, we expect , when it is a reasonable idea to use that technique , that @xmath13 will have some separated eigenvalues , that is , a few will be large and there will be a gap in the spectrum . in that setting , it is enough to understand @xmath62 , which corresponds , if @xmath361 , to a pure signal matrix , with a possibly slightly different kernel , to have a theoretical understanding of the properties of the technique .
for instance , if @xmath361 , if the assumptions underlying the first - order results of @xcite are satisfied for @xmath62 , the ( first - order ) spectral properties of @xmath13 are the same as those of @xmath62 , and hence of the corresponding operator in @xmath17 .
our analysis reveals a very interesting feature of the gaussian kernel , that is , the case where @xmath362 , for some @xmath363 : when theorem [ thm : infoplusnoisegaussiancase ] or corollary [ coro : sphericalnoisekernels ] ( i.e. , theorem [ thm : infoplusnoiseconccase ] with @xmath361 ) apply , the eigenspaces corresponding to separated eigenvalues of the signal@xmath0noise kernel matrix converge to those of the pure signal matrix .
this is simply due to the fact that in that setting , if @xmath364 is the matrix such that @xmath365 a rescaled version of the `` pure signal '' matrix @xmath176 with @xmath2th entry @xmath366 , we have @xmath367 this latter statement is a simple consequence of the fact that @xmath368 is a diagonal matrix with entries @xmath369 on the diagonal , and therefore its operator norm goes to 0 .
on the other hand , @xmath364 clearly has the same eigenvectors as the pure signal matrix @xmath176 .
hence , because the eigenspaces of @xmath62 are consistent for the eigenspaces of @xmath364 corresponding to separated eigenvalues , they are also consistent for those of @xmath176 .
( we note that our results are actually stronger and allow us to deal with a collection of matrices with varying @xmath370 and not a single @xmath370 , as we just discussed .
this is because we can deal with approximations over a collection of functions in all our theorems . ) because of the practical importance of eigenspaces in techniques such as kernel pca , these remarks can be seen as giving a theoretical justification for the use of the gaussian kernel over other kernels in the situations where we think we might be in an information@xmath0noise setting , and the noise is spherical . on the other hand
, @xmath364 underestimates the large eigenvalues of @xmath176 because @xmath371 , and obviously @xmath372 . using weyl s inequality
[ see @xcite ] , we have , if we denote by @xmath373 is the @xmath177th eigenvalue of the symmetric matrix @xmath13 , @xmath374 since the right - hand side goes to 0 asymptotically , the eigenvalues of @xmath176 ( the `` pure signal '' matrix ) that stay asymptotically bounded away from 0 are underestimated by the corresponding eigenvalues of @xmath62 .
when the noise is elliptical , that is , @xmath141 s are not all equal to 1 , the `` new '' matrix @xmath364 we have to deal with has entries @xmath375 so it can be written in matrix form @xmath376 where @xmath377 is a diagonal matrix with @xmath378 . by the same arguments as above , @xmath379 in probability , but now @xmath364 does not have the same eigenvectors as the pure signal matrix @xmath176 .
so in this elliptical setting if we were to do kernel analysis on @xmath13 , we would not be recovering the eigenspaces of the pure signal matrix @xmath176 .
in various parts of statistics and machine learning , it has been argued that laplacian matrices should be used instead of kernel matrices .
see , for instance , the very interesting @xcite , where various spectral properties of laplacian matrices have been studied , under a `` pure '' signal assumption in our terminology .
for instance , it is assumed that the data is sampled from a fixed - dimensional manifold . in light of the theoretical and practical success of these methods ,
it is natural to ask what happens in the information@xmath0noise case .
there are several definitions of laplacian matrices .
a popular one [ see , e.g. , the work of @xcite , among other publications ] , is derived from kernel matrices : given @xmath13 a kernel matrix , the laplacian matrix is defined as @xmath380 when our theorems [ thm : infoplusnoiseconccase ] or [ thm : infoplusnoiseconccasedotproductskernels ] apply , we have seen that , for relevant classes of functions @xmath381 , @xmath382 in probability .
let us now focus on the case of a single function @xmath5 . if we call @xmath383 the laplacian matrix corresponding to @xmath384 , we have @xmath385 we conclude that @xmath386 in probability
; we can therefore deduce that the spectral properties of the laplacian matrix @xmath387 from those of @xmath383 , which , when @xmath388 , is a `` pure signal '' matrix , where we have slightly adjusted the kernel . here again , the gaussian kernel plays a special role , since when we use a gaussian kernel , @xmath383 is a scaled version of the laplacian matrix computed from the signal part of the data .
finally , other versions of the laplacian are also used in practice . in particular , a `` normalized '' version
is sometimes advocated , and computed as @xmath389 , if @xmath377 is the diagonal of the matrix @xmath387 defined above .
we have just seen that @xmath390 in probability and @xmath391 in probability .
therefore , if the entries of @xmath392 are bounded away from 0 with probability going to 1 , we conclude that @xmath393 stays bounded with high probability and @xmath394 so once again , understanding the spectral properties of @xmath395 essentially boils down to understanding those of @xmath396 , which is , in the spherical setting where @xmath361 , a `` pure signal '' matrix . in the case of the gaussian kernel , @xmath396 is equal to the normalized laplacian matrix computed from the `` pure signal '' data @xmath286 .
[ [ the - question - of - centering ] ] the question of centering + + + + + + + + + + + + + + + + + + + + + + + + + in practice , it is often the case that one works with centered versions of kernel matrices : either the row sums , the column sums or both are made to be equal zero .
these centering operations amount to multiplying ( resp .
, on the right , left or both ) our original kernel matrix by the matrix @xmath397 , where @xmath398 is the @xmath19-dimensional vector whose entries are all equal to 1 .
this matrix has operator norm 1 , so when @xmath62 is such that @xmath399 , the same is true for @xmath400 and @xmath401 , where @xmath45 and @xmath146 are either 0 or 1 .
this shows that our approximations are therefore also informative when working with centered kernel matrices .
our results aim to bridge the gap in the existing literature between the study of kernel random matrices in the presence of pure low - dimensional signal data [ see , e.g. , @xcite ] and the case of truly high - dimensional data [ see @xcite ] .
our study of information@xmath0noise kernel random matrices shows that , to first order , kernel random matrices are somewhat `` spectrally robust '' to the corruption of signal by additive high dimensional and spherical noise ( whose norm is controlled ) . in particular , they tend to behave much more like a kernel matrix computed from a low - dimensional signal than one computed from high - dimensional data .
some noteworthy results include the fact that dot - product kernel random matrices are , under reasonable assumptions on the kernel and the `` signal distribution '' spectrally robust for both eigenvalues and eigenvectors .
the gaussian kernel also yields spectrally robust matrices at the level of eigenvectors , when the noise is spherical .
however , it will underestimate separated eigenvalues of the gaussian kernel matrix corresponding to the signal part of the data . on the other hand ,
euclidean distance kernel random matrices are not , in general , robust to the presence of additive noise . as our results show , under reasonably minimal assumptions on both the noise , the kernel and the signal distribution , a euclidean distance kernel random matrix computed from additively corrupted data behaves like another euclidean distance kernel matrix computed from another kernel : in the case of spherical noise , it is a shifted version of @xmath5 , the shift being twice the norm of the noise . for spherical noise ,
this is bound to create ( except for the gaussian kernel ) potentially serious inconsistencies in both estimators of eigenvalues and eigenvectors , because the eigenproperties of the kernel matrix corresponding to the function @xmath402 are in general different from that of the kernel matrix corresponding to the function @xmath5 .
the same remarks apply to the case of elliptical noise , where the change of kernel is not deterministic and even more complicated to describe and interpret .
our study also highlights the importance of the implicit geometric assumptions that are made about the noise .
in particular , the results are qualitatively different if the noise is spherical ( e.g. , multivariate gaussian ) or elliptical ( e.g. , multivariate @xmath403 ) .
interpretation is more complicated in the elliptical case and a number of nice properties ( e.g. , robustness or consistency ) which hold for spherical noise do not hold for elliptical noise .
we note that our study suggests that simple practical ( and entrywise ) corrections could be used to go from the `` signal@xmath0noise '' situation to an approximation of the `` pure signal '' situation . those would naturally depend on the noise geometry and what information practitioners have about it .
our results can therefore be seen as highlighting ( from a theoretical point of view ) the strength and limitations of techniques which rely on kernel random matrices as a primary element in a data analysis .
we hope they shed light on an interesting issue and will help refine our understanding of the behavior of kernel techniques and related methodologies for high - dimensional input data .
the author would like to thank peter bickel for suggesting that he consider the problem studied here and in general for many enlightening discussions about topics in high - dimensional statistics .
he would also like to thank an anonymous referee and associate editors for raising interesting questions which contributed to a strengthening of the results presented in the paper .
belkin , m. and niyogi , p. ( 2003 ) .
laplacian eigenmaps for dimensionality reduction and data representation .
_ neural comput . _ * 15 * 13731396 .
available at http://www.mitpressjournals.org/doi/abs/10.1162/089976603321780317 .
zwald , l. , bousquet , o. and blanchard , g. ( 2004 ) .
statistical properties of kernel principal component analysis . in _ learning theory_. _ lecture notes in computer science _ * 3120 * 594608 .
springer , berlin . | kernel random matrices have attracted a lot of interest in recent years , from both practical and theoretical standpoints .
most of the theoretical work so far has focused on the case were the data is sampled from a low - dimensional structure .
very recently , the first results concerning kernel random matrices with high - dimensional input data were obtained , in a setting where the data was sampled from a genuinely high - dimensional structure similar to standard assumptions in random matrix theory . in this paper
, we consider the case where the data is of the type `` information@xmath0noise . '' in other words , each observation is the sum of two independent elements : one sampled from a `` low - dimensional '' structure , the signal part of the data , the other being high - dimensional noise , normalized to not overwhelm but still affect the signal .
we consider two types of noise , spherical and elliptical . in the spherical setting ,
we show that the spectral properties of kernel random matrices can be understood from a new kernel matrix , computed only from the signal part of the data , but using ( in general ) a slightly different kernel .
the gaussian kernel has some special properties in this setting . the elliptical setting , which is important from a robustness standpoint , is less prone to easy interpretation . . |
in this contribution we describe experimental variations on the so - called coffee - ring effect .
if you have spilled a drop of coffee and left it to dry , then you might have observed a ring - shaped stain .
specifically , the stain is darker near the drop edges compared to the middle ( fig .
[ droppin]c ) .
this phenomenon is the coffee - ring effect ; it is produced by the interplay of fluid dynamics , surface tension , evaporation , diffusion , capillarity , and more .
briefly , as a drop evaporates , its edges easily become pinned and can not recede towards the middle of a drop , i.e. , the diameter of a pinned drop does not decrease ( fig .
[ droppin]a ) .
this effect is perhaps surprising considering that fluid regions near the edges of a drop are thinner than in the middle .
thus , fluid flows from the middle of the drop to the edge of the drop to replenish evaporated water .
this flow readily carries suspended particles , moving them from the middle of the drop to its edges , thus producing a coffee - ring .
why care about the coffee - ring effect ?
a drop of evaporating water is a complex , difficult - to - control , non - equilibrium system . along with capillary flow ,
the evaporating drop features a spherical - cap - shaped air - water interface and marangoni flows induced by small temperature differences between the top of the drop and the contact line @xcite .
thus , to understand the coffee - ring effect , one must understand pinning effects , fluid dynamics , particle - substrate interactions , substrate - fluid interactions , and more .
indeed , intellectual challenges have motivated us to understand this complex , far - from - equilibrium system , and the effects of each of these parameters . of course
, if the coffee - ring effect were only present in coffee and tea , its practical importance would be minimal .
in fact , the coffee ring effect is manifest in systems with diverse constituents ranging from large colloids @xcite to nanoparticles @xcite to individual molecules ( e.g. , salt ) @xcite . due to its ubiquity
, the coffee - ring manages to cause problems in a wide range of practical applications which call for uniform coatings , such as printing @xcite , genotyping @xcite , and complex assembly @xcite .
paint is another system susceptible to the coffee - ring effect . to avoid uneven coatings ,
paints often contain a mixture of two different solvents .
one is water , which evaporates quickly and leaves the pigment carrying particles in a second , thicker solvent .
the particles are unable to rearrange in this viscous solvent and are then deposited uniformly .
unfortunately , this second solvent also evaporates relatively slowly ( one reason why it might be boring to watch paint dry ) .
while a number of schemes to avoid the coffee - ring effect have been discovered @xcite , these approaches typically involve significant modifications of the system .
thus , the discovery of relatively simple ways to avoid the coffee - ring effect and control particle deposition during evaporation could greatly benefit a wide range of applications . to this end
, we asked ( and answered ) a question : does particle _ shape _ affect particle deposition @xcite ? at first glance , it may appear that shape should not matter .
colloidal particles of any shape are susceptible to the radially outward flow of fluid that drives the coffee - ring effect . however , changing particle shape dramatically changes the behavior of particles on the air - water interface .
in fact , smooth anisotropic ellipsoids deform the air - water interface while smooth isotropic spheres do not @xcite .
deforming the air - water interface , in turn , induces a strong interparticle capillary attraction between ellipsoids .
this capillary attraction causes ellipsoids to form a loosely - packed network that can cover the entire air - water interface , leaving ellipsoids much more uniformly distributed when evaporation finishes ( fig .
[ droppin ] b ) .
conversely , spheres pack densely at the drop s edge , producing a coffee - ring when evaporation has finished ( fig .
[ droppin ] c ) .
thus , particle _ shape _ can produce uniform coatings .
the remainder of this review is organized as follows .
first , we discuss the different interfacial properties of spheres and ellipsoids , as well as the methods to make anisotropic particles .
then , we discuss our investigation of particle behavior in evaporating sessile drops and the coffee - ring effect .
much of this work is described in a recent publication @xcite .
in particular , we demonstrate that particle shape strongly affects the deposition of particles during evaporation .
next , we investigate the role of particle shape in evaporating drops in confined geometries , and we show how to extract the bending rigidity of the membranes formed by particles adsorbed on the air - water interface .
much of this work is described in another publication @xcite .
finally , we shift focus to discuss the effects of surfactants on evaporating colloidal drops .
we show that surfactants lead to a radially inward flow on the drop surface , which creates a marangoni eddy , among other effects , which leads to differences in drying dynamics .
some of this work was published recently @xcite . as a whole ,
this review attempts to present these experiments in a unified fashion .
at small packing fractions , i.e. , outside the range which would lead to formation of crystalline or liquid crystalline phases , the diffusion and hydrodynamics of spheres and ellipsoids are only modestly different @xcite .
further , both spheres and ellipsoids will adsorb onto the air - water interface ; the binding energy of micron - sized particles to the air - water interface depends primarily on the interfacial area covered by the particle and the contact angle , quantities which are similar for spheres and ellipsoids .
the binding energy for a micron - sized particle is @xmath0 , where @xmath1 is the boltzmann constant and @xmath2 is temperature @xcite .
once adsorbed onto the air - water interface , however , the behaviors of spheres and ellipsoids are dramatically different @xcite .
anisotropic particles deform interfaces significantly , which in turn produces very strong interparticle capillary interactions @xcite .
these deformations have been predicted @xcite and have been experimentally observed via techniques such as ellipsometry and video microscopy @xcite .
two particles that deform the air - water surface will move along the interface to overlap their deformations and thus minimize total system ( particles plus interface ) energy .
this preference at the interface effectively produces a strong interparticle attraction , which has been measured to be hundreds of thousands times greater than thermal energy for micron size particles @xcite .
the interfacial deformations can be understood from expanded solutions of the young - laplace equation @xcite .
the young - laplace equation minimizes the energy associated with a surface , and thus relates the pressure difference across the surface to the curvature of the surface .
specifically , the young - laplace equation is a force balance statement : @xmath3 , where @xmath4 is surface tension , @xmath5 is mean curvature of the interface , @xmath6 is the pressure in the air , and @xmath7 is the pressure in the water @xcite . for length scales smaller than the capillary length ( i.e. , the length scale at which the laplace pressure from surface tension is equal to the hydrostatic pressure due to gravity , @xmath8 mm for water ) , gravitational effects can be ignored , and the pressure drop across the surface is zero , implying that the mean curvature everywhere is zero .
the mean curvature can be expressed as @xmath9 , where @xmath10 is the laplacian and @xmath11 is the height of the surface .
thus , @xmath12 .
when a particle attaches to the air - water interface , boundary conditions are created for the air - water interface .
theoretically , one seeks to solve the young - laplace equation with these boundary conditions . in this case
, it is useful to first rewrite the young - laplace equation in polar coordinates , @xmath13 ( see fig . [ yleq ] a - c ) .
this problem is similar to potential problems in electrostatics and can be solved by separation of variables , i.e. , with the ansatz @xmath14 .
substitution for @xmath15 leads to @xmath16 . since this equation must hold as @xmath17 and @xmath18
are varied independently , each term in the equation must equal the same constant , which is leadingly termed @xmath19 .
thus , @xmath20 and @xmath21 , with solutions @xmath22 and @xmath23 , where @xmath24 , @xmath25 , and @xmath26 are determined by boundary conditions .
the monopole term @xmath27 is only non - zero when the height of the interface near the particle is uniformly lowered ( or raised ) ( fig .
[ yleq ] c ) .
the monopole term is only stable for the particle in an external field ( e.g. , gravity ) ; however , for typical colloidal particles the gravitational buoyancy forces are not significant and this term is zero .
the dipole term @xmath28 corresponds to a situation wherein the height of the interface is lower on one side of the particle compared to the opposite side ( fig .
[ yleq ] d ) .
thus , this situation can be quickly relaxed by rotating the particle , i.e. , lowering the interface on the high side and raising it on the low side .
the dipole term is only stable when an external torque is applied ; since no external torques act on the particles , this term also is zero .
therefore , the lowest allowed term is the quadrupole term ( @xmath29 ) , i.e. , @xmath30 ( fig . [ yleq ] e ) .
notice , this derivation has not mentioned anisotropic boundary conditions .
in fact , the quadruploar form for @xmath15 is applicable in general to any deformation of the air - water interface ( absent external forces and torques ) that arises at the particle surface .
the air - water interface can be deformed on a sphere , if the three - phase contact line is heterogeneously pinned ( see fig . [ yleq ] f ) @xcite .
this effect produces a quadrupolar profile of the interfacial height .
however , the linear size of deformation , i.e. , @xmath31 , the maximum value of @xmath11 minus the minimum value of @xmath11 , from contact - line - roughness is typically much smaller than the linear size of the deformation from shape - based - roughness ( for example , see reference @xcite ) . of course , if one applies young s conditions for the three - phase contact line on the solid particle , one `` ideally '' obtains a circle contact line on the sphere and a much more complicated shape on an anisotropic particle such as an ellipsoid . on the ellipsoid , this leads to height variations of @xmath15 that are of order the particle size .
the interaction potential between two particles is related to the excess surface area created by these deformations @xcite . for a single particle ( @xmath32 ) ,
the deformation energy is @xmath33 , where @xmath34 is the excess surface area due to interfacial deformation , which is proportional to the deformation size squared , i.e. , @xmath35 the interaction energy of two particles ( @xmath36 ) is @xmath37 , where @xmath38 is the excess surface area due to both particle @xmath32 and particle @xmath39 ( which is dependent on the particle positions and orientations ) , and @xmath40 and @xmath41 are the excess areas due to particle @xmath32 and @xmath39 alone . for smooth spheres @xmath42 , since the interface is not significantly deformed , so @xmath43 . for ellipsoids ( or rough spheres ) , @xmath44r^{-4}$ ] , where @xmath45 and @xmath46 are the angular orientations of ellipsoids @xmath32 and @xmath39 .
the attractive strength decays as @xmath47 and depends on the coefficient term @xmath48 , which , in turn , depends on the deformation size squared , i.e. , @xmath49 .
thus , the strength of this attraction ultimately depends strongly on the size of the deformation at the surface of the particle .
for example , @xmath50 micron diameter particles that induce interfacial deformations of @xmath51 nm and an interparticle separation of @xmath8 microns will produce an attraction with strength @xmath52 . for micron - sized ellipsoids ,
the binding energy from capillary attraction is @xmath53 @xcite .
to summarize , spheres and ellipsoids behave similarly in bulk fluid and are bound to the air water interface by similarly large binding energies . however , on the air - water interface their behavior is dramatically different .
anisotropic particles deform the air - water interface , producing a quadrupolar attraction that is energetically strong ( @xmath53 ) . to understand how particle shape impacts particle deposition
, we need particles with different shapes .
we utilize micron - sized polystyrene spheres ( invitrogen ) , similar to the particles used in previous experiments ( e.g. , @xcite ) , and we simply modify their shape by stretching them asymmetrically to different aspect ratios @xcite . the procedures to make particles have been described previously @xcite , but for completeness we briefly discuss these methodologies below . to create ellipsoidal particles , @xmath54 @xmath55 m diameter polystyrene particles
are suspended in a polyvinyl alcohol ( pva ) gel and are heated above the polystyrene melting point ( @xmath56@xmath57 @xmath58c ) , but below the pva melting point ( @xmath56@xmath59 @xmath58c ) @xcite .
polystyrene melts in the process , but the pva gel only softens .
the pva gel is then pulled so that the spherical cavities containing liquid polystyrene are stretched into ellipsoidal cavities .
when the pva gel cools , polystyrene solidifies in the distorted cavities and becomes frozen into an ellipsoidal shape .
the hardened gel dissolves in water , and the pva is removed via centrifugation .
each sample is centrifuged and washed with water at least @xmath60 times .
each iteration of this process creates @xmath56@xmath61 ellipsoidal particles in @xmath56@xmath62 @xmath55l suspensions .
the particles are charge - stabilized , and the resultant suspensions are surfactant - free .
snapshots of experimental particles are shown in the insets of fig .
[ droppin ] b , c. the aspect ratio polydispersity is @xmath56@xmath63 . to ensure the preparation process does not affect particle deposition , our spheres undergo the same procedure , absent stretching .
importantly , in order to ensure the pva was not affecting our results , we performed a separate set of experiments investigating the effects of pva on evaporating drops . in these experiments
the pva weight percent was carefully controlled .
we found that if a sample contains more than @xmath64 pva by weight , then the contact line of the drying drop depins very quickly after the drop is placed on a glass slide .
however , in samples with less than @xmath64 pva by weight , the contact line behavior of the drying drop is identical to the contact line behavior in drops without pva . to confirm that small amounts of pva do not affect the deposition of spheres , we added pva ( @xmath65 by weight ) to a suspension of spheres . during evaporation ,
the contact line remains pinned , and the spheres exhibit the coffee ring effect .
further , when ellipsoids are diluted by a factor of @xmath57 ( and thus the pva weight percent is decreased by a factor of @xmath57 to an absolute maximum of @xmath66 ) , the spatially uniform deposition of ellipsoids persists .
understanding why ellipsoids are deposited uniformly first requires that we characterize the evaporation process , i.e. , we quantify the spatio - temporal evaporation profile of the suspensions .
first , we are interested in the evaporation rate . to this end
, we directly measure the drop mass of different suspensions ( @xmath67 @xmath55l in volume , @xmath68 mm in radius , @xmath69 ) during evaporation ( fig .
[ figs1cr ] a ) .
( in order to improve the accuracy of the reported evaporation rate , we utilized large - volume drops . ) for all suspensions ( drops of sphere suspension , drops of ellipsoid suspension , and drops of water absent colloid ) , the mass of each drop decreases linearly in time with very similar mass rates - of - change ( @xmath70 @xmath55g / s ) .
this bulk evaporation behavior for all suspensions is consistent with steady - state vapour - diffusion - limited evaporation of spherical - cap - shaped drops with pinned contact lines @xcite .
+ next , we quantified the contact line evolution during drying , i.e. , we observed when the contact line depins .
specifically , we measured the radius of the @xmath50 @xmath55l drops ( @xmath69 ) during evaporation by video microscopy ( fig .
[ figs1cr ] b ) .
the time at which evaporation finishes , @xmath71 , is clearly indicated in fig . [ figs1cr ] b as the time when the drop radius shrinks to zero . for all samples
, we observed the radius decrease by less than @xmath63 until @xmath72 ; i.e. , the contact line remains pinned for the vast majority of the evaporation , regardless of particle shape .
note , the contact line in drops containing ellipsoids does partially depin around @xmath73 ; however , it does not completely depin until @xmath72 .
these control experiments demonstrate that contact line behavior , capillary flow , and evaporation rates are independent of suspended particle shape .
thus , to produce qualitatively different deposits , the microscopic behaviors of individual spheres and ellipsoids in the droplets must differ .
the uniform deposition of ellipsoids after evaporation ( fig .
[ droppin ] b ) is especially striking when compared to the heterogeneous `` coffee ring '' deposition of spheres ( fig .
[ droppin ] c ) in the same solvent , with the same chemical composition , and experiencing the same capillary flows ( fig .
[ fig1cr2 ] a ) . to quantify the particles deposition shown in fig .
[ droppin ] b and c , we determined the areal number fraction of particles deposited as a function of radial distance from the drop center ( fig .
[ fig1cr2 ] b ) . in detail , utilizing video microscopy and particle tracking algorithms , we counted the number of particles , @xmath74 , in an area set by the annulus bounded by radial distances @xmath17 and @xmath75 from the original drop center @xcite ; here @xmath76 is @xmath56@xmath77 @xmath55 m . the areal particle density @xmath78 , with @xmath79 .
to facilitate comparisons between different samples , and eliminate small sample - to - sample particle density differences , we normalized @xmath80 by the total number of particles in the drop , n. further , to we report @xmath81 as a function of @xmath82 , where @xmath83 is the drop radius , to eliminate small sample - to - sample differences in drop radii .
dilute suspensions ( @xmath69 ) are utilized to improve image quantification . for spheres ( @xmath84 ) ,
@xmath85 is @xmath56@xmath86 times larger at @xmath87 than in the middle of the drop . conversely , the density profile of ellipsoidal particles is fairly uniform as a function of @xmath82 ( there is a slight increase at large @xmath82 ) . as particle shape anisotropy
is increased from @xmath84 to @xmath88 , the peak in @xmath81 at large @xmath82 decreases .
the coffee - ring effect persists for particles marginally distorted from their original spherical shape ( @xmath89 and @xmath90 ) , but particles that are slightly more anisotropic ( @xmath91 ) are deposited more uniformly . to further quantify the sharply peaked coffee - ring effect of spheres and the much more uniform deposition of the ellipsoids
, we calculate and plot @xmath92 ( fig .
[ fig1cr2 ] c ) , where @xmath93 is the maximum value of @xmath80 ( typically located at @xmath94 ) and @xmath95 is the average value of @xmath80 in the middle of the drop ( @xmath96 ) . for spheres , @xmath97 .
as aspect ratio increases slightly ( @xmath89 and @xmath90 ) @xmath92 decreases to @xmath98 and @xmath99 , respectively . for ellipsoids , @xmath92 is more than ten times smaller than spheres .
as @xmath100 continues to increase above @xmath101 , @xmath92 continues to decrease , albeit at a much lower rate .
note , @xmath92 was observed to be largely independent of initial volume fraction , i.e. , @xmath92 fluctuated by approximately @xmath102@xmath63 as volume fraction changed between @xmath103 and @xmath104 .
when drops with very large packing fractions evaporate , the drop surface becomes saturated with ellipsoids .
however , deposition in this limit is difficult to quantify , as at high volume fractions it is difficult to measure the local particle density .
thus , while the particles that can not attach to the interface are likely transported to the drop edge , it is difficult to demonstrate that this effect occurs . an experimental snapshot after evaporation of a drop of ellipsoids ( @xmath105 ) initially suspended at volume fraction
@xmath106 shows that overall the coffee - ring effect is avoided , but the local density can not be extracted ( fig .
[ fig1cr2 ] d and e ) .
an image of the final distribution of spheres evaporated from a suspension with initial packing fraction @xmath106 is included for comparison .
our observations thus far imply that micron - sized grains in a cup of coffee are relatively spherical . to confirm or refute this hypothesis , we prepared a microscope slide full of diluted coffee .
this coffee came from the lab - building coffee machine ( filterfresh ) , which passes through a paper filter after a relatively short brew time ( @xmath107 seconds ) . while we did not `` fully '' characterize the shape of the grains we observed , qualitatively , they appeared spherical on the micron - size scale ( see fig .
[ coffeepic ] a ) .
thus , a suspension of polystyrene spheres really is an especially good model of a cup of coffee .
the evidence suggest that the same radially outward flows are present in drops containing either spheres or ellipsoids .
the deposition of spheres and ellipsoids after drying , however , is very different . in order to understand the origin of these differences better
, we carried out a battery of experiments focused on the behaviors of spheres and ellipsoids on the air - water surface .
snapshots from video microscopy show that both spheres ( fig .
[ fig2cr ] a - d ) and ellipsoids ( fig .
[ fig2cr ] e - h ) are carried to the drop s edges . to quantify this effect ,
the average areal particle density close to the contact line , @xmath108 , was measured as a function of time ( fig .
[ fig2cr ] i ) . for spheres
, @xmath109 increases linearly until evaporation is complete , with a slope of @xmath110 s@xmath111 .
conversely , the areal density of ellipsoids near the contact line stops growing at @xmath112 , and for @xmath113 , @xmath109 increases with a slope of @xmath114 s@xmath111 .
this slope for ellipsoids is less than @xmath115 the slope for spheres , despite similar evaporation rates , capillary flows , and contact line behaviors .
thus , ellipsoid density at the drop edge grows at a slower rate than sphere density .
next , we note that both spheres and ellipsoids strongly prefer adsorption to interface than life in the bulk drop .
further , our experiments with ellipsoids and spheres , and previous experiments with spheres @xcite , suggest that @xmath116 of the particles adsorb to the air - water interface in the `` central / middle '' regions of the drop .
thus most particles move toward the drop edges , and the relative drying behaviors of ellipsoids and spheres must be controlled by their behaviors near the drop edge .
to study this issue we first determine where ellipsoids adsorb on the air - water interface , i.e. , we measure the number of ellipsoids that adsorb on the air - water interface as a function of radial position . the areal number density ( give symbol ) of ellipsoids on the air - water interface versus radial distance , ( symbol ) at a time immediately before the drop edge depins is given in fig .
[ coffeepic ] b. the majority of particles are deposited within @xmath117 microns of the drop s edge ( at @xmath118 microns ) .
approximately @xmath119 of the ellipsoid particles adsorb on the air - water interface in this region near the drop s edge .
the properties of this interfacial region and the mechanisms by which particles attach to and move within this interfacial region play a key role in the drying process .
what actually happens at the drop s edge ?
experimental snapshots of of particles moving in the region within @xmath120 @xmath55 m of the drop contact line confirm that while spheres pack closely at the edge ( fig .
[ fig3cr ] a ) , ellipsoids form `` loosely '' packed structures ( fig .
b ) , which prevent particles from reaching the contact line . particles with @xmath91 and @xmath121 pack at higher area fractions than ellipsoids with @xmath122 , resulting in larger values of @xmath92 for @xmath91 and @xmath121 and producing the small peak in @xmath123 at @xmath124 for @xmath91 .
the ellipsoid particle structures on the air - water interface appear to be locally arrested or jammed @xcite , i.e. , particles do not rearrange .
once an ellipsoid joins the collective structure , its position relative to other ellipsoids typically changes by less than @xmath67 nm ( lower limit of our resolution ) , and the overall particle structure rearranges , for the most part , only when new particles attach to the interface .
images of particles near the drop s contact line ( fig .
[ fig3cr ] b ) reveal that unlike spheres , which are carried from the bulk all the way to the contact line ( fig .
[ fig3cr ] a ) , most ellipsoids adhere to the loosely - packed structures at the air - water interface before they reach the three - phase contact line at the drop edge .
this capillary attraction has been characterized in prior experiments as long - ranged and very strong @xcite . to understand the different behaviors of spheres and ellipsoids at the edge of drying drops
, it is instructive to observe some individual particle trajectories .
the trajectory of a single sphere is highlighted in fig .
[ spheretraj ] a - d .
spheres ( like the one highlighted in fig .
[ spheretraj ] a - d ) are pushed through the bulk fluid towards the drop s edge .
when spheres reach the drop s edge , their progress is halted by a wall of spheres already at the drop s edge .
spheres then pack densely , and can not rearrange as they jam into the ring configuration .
this behavior is demonstrated quantitatively for a few typical spheres by plotting the distance ( @xmath125 ) between the sphere and the drop s edge versus time ( fig .
[ spheretraj ] e ) .
conversely , when ellipsoids reach the drop s edge , they pack loosely on the air - water interface ( fig .
[ spheretraj ] f - i ) .
notice , ellipsoids at the drop s edge do not necessarily halt the progress of other migrating ellipsoids that arrive at later times .
this can be seen in fig .
[ spheretraj ] f - h , as an ellipsoid approaches the drop s edge ( fig .
[ spheretraj ] f ) , passes underneath a cluster of ellipsoids on the air - water interface ( fig .
[ spheretraj ] g ) , and eventually adsorbs on the air - water interface near the drop s edge ( fig .
[ spheretraj ] h ) .
as evaporation continues , ellipsoids can move along the surface of the drop towards the drop s center ( fig .
[ spheretraj ] i ) .
this behavior is demonstrated quantitatively for a few typical ellipsoids by plotting @xmath125 versus time ( fig . [ spheretraj ] j ) .
if the air - water interface is not saturated with ellipsoids when the drop s edge depins , then the networks of ellipsoids are compressed as they are pushed towards the drop s center .
the loosely - packed configurations formed by ellipsoids on the interface are structurally similar to those seen in previous experiments of ellipsoids at flat air - water and water - oil interfaces @xcite .
they produce a surface viscosity that is much larger than the suspension bulk viscosity , facilitating ellipsoid resistance to radially outward flows in the bulk .
note , spheres also adsorb onto the interface during evaporation .
however , spheres do not strongly deform the interface @xcite , and they experience a much weaker interparticle attraction than ellipsoids @xcite ; therefore , the radially outward fluid flows in the bulk and interface easily push spheres to the drop s edge @xcite . in order to quantify the ability of interfacial aggregates of ellipsoids to resist bulk flow , we calculated the boussinesq number , b@xmath126 , for ellipsoids with @xmath127 .
specifically , b@xmath126 is the ratio of the surface drag to the bulk drag : b@xmath128 where @xmath129 is shear stress from bulk flow , g is the elastic modulus of the interfacial layer , and @xmath130 is the probed lengthscale @xcite .
b@xmath126 varies spatially with the average areal particle density on the air - water interface . here
, we calculate b@xmath126 in a region within @xmath120 @xmath55 m of the pinned contact line .
we first calculated b@xmath126 at an early time ( t @xmath131 t@xmath132 ) .
the shear stress can be calculated from the particle velocity and drop height via @xmath133 , where @xmath129 is shear stress , @xmath55 is viscosity , and @xmath130 is the drop height . at an early time
( t @xmath131 t@xmath132 ) @xmath134 pa .
about @xmath135 of the surface is covered with ellipsoids .
previous experimental studies measured the shear modulus , @xmath136 , of the interfacial monolayer as a function of surface coverage area fraction @xcite .
we measured the surface coverage area fraction in our system as a function of time .
this measurement enabled us to utilize the values of @xmath136 reported in @xcite ( @xmath137 n / m ) .
the probed lengthscale , @xmath130 , is at most @xmath138 m ( i.e. , the drop diameter ) .
thus , at t @xmath139t@xmath132 , b@xmath140 .
this calculation is performed at different times during evaporation , until the final stage of evaporation when the aggregate of ellipsoids begins flowing towards the drop center ( fig .
[ figs3cr ] a ) .
we found that @xmath129 grows linearly with particle velocity , which we observe to increase by a factor of @xmath141 during evaporation .
however , @xmath136 grows exponentially with the ellipsoidal area fraction@xcite , and area fraction increases by a factor of @xmath142 .
thus , the exponential growth of @xmath136 dominates this calculation , and b@xmath126 grows exponentially with time : b@xmath143 . finally , note that for spheres , b@xmath144
thus , the measured dimensionless boussinesq number clearly demonstrates that clusters of ellipsoids on the air - water interface can resist shear from radially outward fluid flows , and make sense of the fact that these clusters are not pushed to the drop s edge .
conversely , clusters of spheres on the air - water interface can not resist shear and are pushed along the air - water interface to the drop s edge where the join the coffee - ring deposit .
we have already shown that ellipsoids sit ( largely without moving ) at the air - water interface . here
we utilize confocal microscopy to directly measure the location of ellipsoids ( and spheres ) during evaporation .
confocal snapshots are shown in fig .
[ figconfocal ] . by integrating the brightness of each pixel over a period of @xmath145 seconds , only particles that are roughly stationary during this time period appear in the images .
snapshots are then projected onto a side - view of the drop .
the confocal images confirm that ellipsoids sit at the air - water interface ( fig .
[ figconfocal ] bottom ) , while spheres are carried all the way to the contact line ( fig .
[ figconfocal ] top ) . in order to assess the generality of this effect , we have analyzed three additional types of anisotropic particles .
one parameter potentially important for this process is particle hydrophobicity .
hydrophilic particles , for example , are perhaps less likely to adsorb onto the air - water interface than hydrophobic particles and might equilibrate differently on the interface than hydrophobic particles ; thus the hydrophilic ellipsoids could have different deposition during drying . to investigate the effect of hydrophilicity , we obtained suspensions of spherical and ellipsoidal polystyrene - pnipmam core - shell particles , i.e. , polystyrene particles coated with pnipmam .
we evaporated these suspensions at @xmath146 @xmath58c ; at this temperature , pnipmam is hydrophilic .
the core - shell hydrophilic spheres exhibit the coffee ring effect ( fig .
[ figs3cr ] b ) .
conversely , despite their hydrophilicity , core - shell ellipsoids are deposited uniformly .
in fact , these core - shell ellipsoids form the same type of loosely - packed ellipsoid networks on the drop surface as polystyrene ellipsoids absent pnipmam ( fig .
[ figs3cr ] b ) .
further , we have evaporated suspensions of actin filaments and pf@xmath50 viruses . in each of these suspensions
, the contact line depins at very early times . to prevent this early depinning
, we add a small amount of @xmath62 nm diameter fluorescent polystyrene spheres ( @xmath147 by weight ) ; these spheres help to pin the contact line until the final stage of evaporation ( t @xmath148 @xmath149 t@xmath132 ) via self - pinning @xcite .
the spheres in each suspension exhibit the coffee - ring effect .
both the actin filaments and pf@xmath50 viruses in suspension , however , are deposited relatively uniformly .
( note , the mean major axis length for pf@xmath50 viruses is @xmath141 @xmath55 m ; the mean minor axis length for pf@xmath50 viruses is @xmath150 nm . the mean major axis length for actin filaments is @xmath151 @xmath55 m ; the mean minor axis length for actin filaments is @xmath151 nm .
+ ) lastly , we investigate the effects of mixing ellipsoids and spheres . a small number of ellipsoids were added to suspensions of different sized spheres .
we then evaporate drops of suspensions containing both ellipsoids and spheres .
our initial hope was that a small number of ellipsoids could dramatically change the deposition behavior of spheres in suspension . to simplify this study , we concentrated on two different aspect ratios : spheres ( @xmath152 ) and ellipsoids ( @xmath127 ) .
the ellipsoids were stretched from particles of diameter @xmath153 @xmath55 m ; each suspension contains spheres suspended at a volume fraction @xmath154 .
evaporative deposits are characterized as a function of ellipsoid volume fraction @xmath155 via @xmath156 ( fig .
[ fig4cr ] a ) .
suspensions containing smaller spheres with @xmath157 @xmath55 m along with the ellipsoids at volume fractions ranging from @xmath158 to @xmath159 were evaporated .
the coffee - ring effect persists for these small spheres , regardless of how many ellipsoids are added to the initial suspension ( fig .
[ fig4cr ] a ) .
small spheres can easily navigate under or through the loosely packed ellipsoid networks , and thus reach the drop s edge ( fig .
[ fig4cr ] b - d ) . for comparison , we evaporated suspensions containing larger spheres with @xmath160 @xmath55 m , along with the same ellipsoids at the same volume fractions utilized previously . for small ellipsoid volume fraction ( @xmath161 )
, the evaporating suspensions still exhibit the coffee - ring effect .
however , for larger @xmath155 , the coffee ring is diminished ; for sufficiently large @xmath155 , i.e. , @xmath162 , the coffee - ring effect is avoided ( fig .
[ fig4cr ] a ) .
larger spheres adsorb onto the air - water interface farther from the drop edge than do the smaller ellipsoids .
absent ellipsoids , spherical particles form closely - packed aggregates . in the presence of ellipsoids , the spheres instead become entangled in the loosely - packed ellipsoid networks , thus eliminating the coffee ring effect ( fig .
[ fig4cr ] b - d ) . therefore
, large spherical particles can be deposited uniformly simply by adding ellipsoids .
the ability to deposit particles uniformly is desirable in many applications @xcite .
unfortunately , most proposed methods for avoiding the coffee ring effect require long multistage processes , which can be costly in manufacturing or require use of organic solvents which are sometimes flammable and toxic ( e.g. @xcite ) . here
we have shown that by exploiting a particle s shape , a uniform deposit can be easily derived from an evaporating aqueous solution .
the results presented here further suggest that other methods of inducing strong capillary interactions , e.g. , surface roughness @xcite , may also produce uniform deposits .
additionally , open questions about the behavior of ellipsoids in drying drops persist .
specifically , one may have thought the drop s edge would quickly saturate with ellipsoids during evaporation , and ellipsoids subsequently arriving would then be deposited in a coffee - ring stain .
however , ellipsoids ( and their collective structures ) clearly migrate towards the drop s center during evaporation , in the process creating room for more ellipsoids to adsorb on the air - water interface near the drop s edge .
it is unclear why ellipsoids move towards the drop s center .
inward fluid flows along the drop s surface push networks of ellipsoids towards the drop s center , thus making room for more ellipsoids to adsorb on the air - water interface .
alternatively , the energetic interactions of the ellipsoids on the air - water interface may play an important role in this inward migration .
however , a complete understanding of this inward motion has been elusive and will require more experimental and theoretical investigation .
the mechanism that produces a uniform coating from particles suspended in drying sessile drops requires the presence of an air - water interface that spans the entire area covered by the drop .
a drop confined between two glass plates is a completely different beast . in this case
, the air - water interface is only present at the drop edges fig .
[ confined]a and fig . [ figs1ce ]
c ( a sessile drop is shown for comparison in fig . [
confined]b ) .
thus , the mechanisms that produce uniform coatings in `` open '' or sessile drops are unlikely to be present in confined drops . to illustrate these spectacular differences , in fig .
[ confined ] c , d , we again see that suspended particle shape produces dramatically different depositions .
the confined drops do nt even exhibit the conventional coffee ring effect .
rather , spheres and slightly stretched spheres are deposited heterogeneously , and anisotropic ellipsoids are distributed relatively more uniformly . in this section ,
we show how one can understand these deposition effects .
important clues are revealed through consideration of the mechanical properties of the air - water interfaces , and changes thereof as a result of adsorbed particles .
recent experiments have explored evaporation of confined drops containing spheres @xcite , and their behaviors differ dramatically from sessile drops containing spheres . in the confined case
, particles are pushed to the ribbon - like air - fluid interface , and , as evaporation proceeds , the particle - covered air - water interface often deforms and crumples ( fig .
[ confined ] e and f ) .
the buckling behaviors exhibited by these ribbon - like colloidal monolayer membranes ( cmms ) in confined geometries are strongly dependent on the geometric shape of the adsorbed particles , and the buckling events appear similar to those observed in spherical - shell elastic membranes @xcite . before buckling
events occur , particles are densely packed near the three - phase contact line , regardless of particle shape .
further , because the particle volume fraction in the drop is relatively low , these membranes essentially contain a monolayer of particles , i.e. , buckling events occur before multilayer - particle membranes form .
these experiments utilize the same micron - sized polystyrene ellipsoids described above in section 2.2 and in @xcite .
drops of suspension are confined between two glass slides separated by @xmath163 @xmath55 m spacers ( fisher scientific ) and allowed to evaporate ; qualitatively similar results are found for chambers made from slightly hydrophobic cover slips .
we primarily study the drops with initial particle volume fraction @xmath164 .
( qualitatively similar results are found for volume fractions ranging from @xmath103 to @xmath145 . )
the confinement chambers are placed within an optical microscope wherein evaporation is observed at video rates at a variety of different magnifications .
this approach also enables measurement of the surface coverage , i.e. , the fraction of the air - water interface coated with particles , prior to buckling events .
we find that for spheres and ellipsoids the surface coverage areal packing fraction is @xmath165 .
to understand this buckling phenomenon , the elastic properties of the air - water interface with adsorbed particles , i.e. , the elastic properties of the cmms , must be quantified . to this end
, the analytical descriptions of elastic membranes are extended to our quasi-2d geometry wherein observations about bending and buckling geometry are unambiguous .
this theoretical extension has been described previously ( see @xcite and its associated supplemental online material ) , but for completeness and clarity of presentation we discuss it more completely below . following the same procedure as @xcite
, we first describe the stretching and bending energy associated with membrane buckling events .
membrane stretching energy can be written as @xmath166 , where @xmath167 is the total membrane stretching energy , @xmath168 is the 2d young s modulus , @xmath169 is the strain , and the integrand is integrated over the membrane volume . for a thin , linearly elastic material , @xmath169 does not change much in the direction perpendicular to the surface , so @xmath170 , where the integral is calculated over the membrane surface area .
the unstretched region has @xmath171 .
further , even in the stretched / buckled membrane , most of the deflected region has @xmath171 , since its configuration is identical to the undeflected membrane except that its curvature is inverted ( fig .
[ figs1ce ] a , b ) .
thus , the only region under strain is the `` rim '' of the deformation ( fig .
[ figs1ce ] a , b ) .
if the entire membrane had experienced a constant radial displacement of @xmath172 , its radius would change from @xmath17 to @xmath173 , and the circumference would change from @xmath174 to @xmath175 . then the membrane strain would be @xmath176 . on the other hand , if ( as is the case for our samples ) the displacement is confined to a small region subtended by angle @xmath177 , then the in - plane length of this region changes from @xmath178 to @xmath179 , and the total strain in the membrane is @xmath180 .
again , this estimate assumes that the interfacial deflection does not change in the @xmath181-direction ( out - of - plane ) , i.e. , @xmath182 . within these approximations , @xmath183 .
the integral is readily performed over an area normal to the glass plates described by @xmath184 , where @xmath185 is the in - plane length of the deflected region , and @xmath11 is the chamber height .
thus , @xmath186 .
the membrane bending energy can be written as @xmath187 , where @xmath188 is the total bending energy , @xmath189 is the bending rigidity , and @xmath190 is the membrane curvature . here , the curvature is @xmath191 , where @xmath125 is the coordinate in - plane along the membrane ( see fig .
[ figs1ce ] a , b ) .
the first derivative can be written as @xmath192 , as @xmath172 is the change in the membrane position over a distance of approximately @xmath185 in the @xmath125 direction .
the second derivative can then be estimated as @xmath193 , as the first derivative changes from @xmath194 in the undeflected region to @xmath195 in the deflected region of approximate length @xmath185 .
therefore , @xmath196 .
( this approach again assumes that the second derivative of the deflection in the @xmath181-direction is small , i.e. , @xmath197 . )
the integral is readily performed over an area described by @xmath184 , and @xmath198 .
the total energy from the deflection is @xmath199 .
this energy is concentrated within the deflected rim ( i.e. , with width @xmath185 ) .
membranes will buckle in the way that minimizes their energy . to derive this condition ,
we minimize the total deflection energy with respect to @xmath185 , i.e. , @xmath200 . minimizing the total bending and stretching energy
gives the relation , @xmath201 .
thus , by measuring @xmath185 and @xmath17 in a series of drops with the same particles and membrane characteristics , we can experimentally determine @xmath202 .
( interestingly , @xmath172 drops out of the calculation , i.e. , a precise determination of @xmath172 is not necessary for this calculation within the assumptions listed above .
also , note that this calculation is independent of the depth of the invagination ; the only requirement is that the deflection minimizes total membrane energy .
finally , note that this derivation assumes that the interfacial displacement varies little in the @xmath181-direction , i.e. , the air - water interface deflects the same distance at the top , middle , and bottom of the chamber . ) in practice we measure @xmath185 as the rim full - width located @xmath203 @xmath55 m from the rim vertex ( see fig .
[ confined ] g and fig .
[ figs1ce ] a , b and d ) .
the exact value of @xmath185 , however , is not very sensitive to measurement protocol .
for example , defining @xmath185 as the full - width at @xmath67 @xmath55 m or @xmath204 @xmath55 m from the rim vertex changes @xmath185 by approximately @xmath67 percent .
this simple experimental approach enables us to extract the ratio of cmm bending rigidity , @xmath189 , to its young s modulus , @xmath168 , from measurements of @xmath185 and @xmath17 across a series of drops from the relation @xmath205 . with all other parameters constant , e.g. , particle anisotropy , particle surface coverage , etc .
, this formula predicts that @xmath206 . in fig .
[ figs2ce ] b we show results from evaporated drops of particles with anisotropy @xmath91 and with different initial values of @xmath17 , plotting @xmath185 versus @xmath207 .
a good linear relationship is observed ( coefficient of determination , r@xmath208 ) , implying that our analysis is self - consistent .
similar high quality linear results were found for other values of @xmath100 . in principle , the air - water interface can be distorted in the @xmath181-direction as well as in - plane .
the analysis thus far has assumed these distortions are small , and it is possible to check that these corrections are small . using bright field microscopy , we can identify the inner and outer position of the air - water interface and thus estimate the radius of curvature in the @xmath181-direction @xcite ( fig .
[ figs2ce ] a ) .
we find that the radius of curvature is approximately equal to the chamber thickness ( @xmath209 m @xmath210 m ) both before and after buckling events .
the relevant partial derivatives are then @xmath211 and @xmath212 ; therefore the corrections to theory are indeed small .
we extract and plot @xmath213 for evaporating drops of particles with different @xmath100 ( fig .
[ figs2ce ] c ) .
notice , @xmath213 increases with increasing @xmath100 , implying that as particle shape becomes more anisotropic , @xmath189 increases faster than @xmath168 , i.e. , @xmath213 is larger for ellipsoids ( @xmath214 and @xmath88 ) than for spheres ( @xmath152 ) .
since we measure the ratio @xmath213 , in order to isolate the bending rigidity we require knowledge of the young s modulus of the membrane .
previous experiments have observed that the cmm young s modulus increases with @xmath100 @xcite . for particles with @xmath152 and @xmath215 , we use previously reported values of the bulk modulus @xcite , @xmath216 , the shear modulus @xcite , @xmath136 , and the relationship @xmath217 in order to extract the cmm young s modulus .
we were unable to find data for @xmath218 or @xmath88 , so we linearly interpolated from reported values of @xmath216 and @xmath136 .
using these previously reported values , we obtained @xmath219 and @xmath220 n / m for @xmath221 and @xmath88 , respectively .
( a - e , respectively ) . f. the area fraction covered by particles after evaporation is complete , _ f _ , for suspensions of particles as a function of their aspect ratio @xmath100 . ]
( a - e , respectively ) . f. the area fraction covered by particles after evaporation is complete , _ f _ , for suspensions of particles as a function of their aspect ratio @xmath100 . ] utilizing these previously reported measurements and calculations of @xmath168 we are able to plot @xmath189 versus @xmath168 ( fig .
[ figs2ce ] d ) .
the best power - law fit finds that @xmath222 .
interestingly , this observation is consistent with theoretical models which predict @xmath223 @xcite .
however , the full physical origin of this connection is unclear .
further , while at first glance it may seem contradictory to claim that @xmath205 and @xmath223 , these formulae are consistent .
a simple elastic model assumes that @xmath224 and @xmath225 , where @xmath226 is the 3d young s modulus and @xmath227 is the membrane thickness @xcite .
based on this model , @xmath228 , so @xmath229 .
thus , @xmath230 . to test this prediction , we plot @xmath231 versus @xmath168 ( fig . [ figs3ce ] ) .
the best power law fit is @xmath232 , implying that these two seemingly contradictory equations are in fact consistent .
note , this simple elastic model suggests that @xmath233 kpa for all @xmath100 , which is similar to stiff jello .
finally , our estimates of cmm bending rigidity are given in ( fig .
[ figs2ce ] e ) . clearly , membrane bending becomes much more energetically costly with increasing particle shape anisotropy .
finally , we turn our attention to the problem we initially hoped to understand : the consequences of increased bending rigidity on particle deposition during evaporation processes in confined geometries . as should be evident from our discussion in sections 1 - 3 , substantial effort
has now yielded understanding of the so - called coffee - ring effect and some ability to control particle deposition from sessile drops @xcite .
much less is known , however , about particle deposition in confined geometries , despite the fact that many real systems @xcite and applications @xcite feature evaporation in geometries wherein the air - water interface is present only at the system edges .
recent experiments have explored evaporation of confined drops containing spheres @xcite , and their behaviors differ dramatically from sessile drops containing spheres . in the confined case ,
as noted previously , particles are pushed to the ribbon - like air - fluid interface , and , as evaporation proceeds , the particle - covered air - water interface often undergoes the buckling events which we have quantified in sections 4.1 and 4.2 .
we find that deposition depends dramatically on suspended particle shape .
the final deposition of particles is shown for @xmath234 , in fig .
[ fig2ce ] a - e , respectively .
spheres and slightly stretched spheres are deposited unevenly , while anisotropic ellipsoids are distributed much more homogeneously . to quantitatively describe the final deposition of particles , we plot the fraction of initial droplet area covered by deposited particles after evaporations , _ f _ ( as introduced in @xcite ) , as a function of particle anisotropy @xmath100 ( fig .
[ fig2ce ] f ) .
specifically , we divide the area into a grid of ( @xmath77 @xmath235 x @xmath77 @xmath235 ) squares .
a region is considered to be covered if its area fraction within the square is greater than @xmath236 .
( note , for uniformly deposited particles , the area fraction ( based on the initial volume fraction , initial volume , chamber height , and particle size ) would be @xmath56@xmath237 .
thus , the threshold we utilize is @xmath238 of this uniformly deposited area fraction ) .
the number of covered regions is then normalized by the total number of squares in the grid , thus producing _ f_. the fraction of area covered with particles is observed to increase with @xmath100 . for @xmath91 and @xmath121 ,
_ f _ increases modestly . for @xmath214 ,
the deposition is very uniform , and for @xmath127 , virtually the entire area is covered uniformly .
the mechanisms that produce the uneven deposition of spheres and slightly stretched particles and the uniform deposition of ellipsoids are revealed by high magnification images ( fig .
[ fig3ce ] a - e ) .
colloidal particles locally pin the contact line and thereby locally prevent its motion .
so - called self - pinning of the air - water interface can occur even in very dilute suspensions , i.e. , @xmath239 @xcite . as evaporation continues in suspensions of spheres or slightly anisotropic particles , the cmm interface bends around the pinning site ( fig .
[ fig3ce ] a - c ) . then , either it pinches off , leaving particles behind , or it remains connected to the pinned site , leading to fluid flow into the narrow channel that has formed .
the latter flow carries particles towards the pinning site ( fig .
[ fig3ce ] b and c ) , thus producing streaks of deposited particles ( see fig .
[ fig3ce ] a - c ) .
temporal and spatial variations along the interface due to these described effects lead to heterogeneous deposition of spherical particles during evaporation .
conversely , when ellipsoids adsorb onto the air - water interface ( forming ribbon - like cmms , see fig . [ fig3ce ]
d ) , they create an elastic membrane with a high bending rigidity .
the bending rigidity of ellipsoid - populated cmms can be approximately two - orders of magnitude larger than sphere - populated cmms ( see fig . [ figs2ce ] ) .
thus , while ellipsoids may also pin the contact line , bending of the cmm interface around a pinned contact line is energetically costly .
microscopically , bending requires the energetically expensive rearrangement of ellipsoids aggregated on the cmm ; attractive particle - particle capillary interactions on the air - water interface must be overcome for bending , even at very small @xmath18 .
conversely , bending a sphere coated cmm costs relatively little energy , as sphere - sphere capillary interactions on the interface are relatively weak @xcite .
thus , as the confined drop continues to evaporate , the ellipsoid coated cmm does not bend .
it recedes radially , depositing ellipsoids near the contact line during this drying process .
as we already demonstrated mixing spheres and ellipsoids in sessile drops presents qualitatively new scenarios .
it is natural to investigate the deposition of mixtures of spheres and ellipsoids in confined geometries . to this
end suspensions of @xmath240 nm spheres ( @xmath100@xmath241@xmath242 ) with @xmath18@xmath241@xmath243 were combined with suspensions containing micron - sized ellipsoids ( @xmath100@xmath241@xmath88 ) at lower volume fractions , @xmath18@xmath244 to @xmath245 .
the resulting colloidal drops were evaporated in the same confined geometries already utilized .
the addition of a very small number of ellipsoids has no effect on the deposition of spheres ( @xmath246 ) .
however , the addition of a larger , but still small number of ellipsoids produces a uniform deposition of both ellipsoids and spheres , i.e. , @xmath247 , despite the fact that spheres significantly outnumber ellipsoids ( @xmath248-@xmath249 ) ( fig .
[ fig3ce ] e ) .
again , the high bending modulus produced by ellipsoids on the cmm helps explain the observations . both spheres and ellipsoids attach to the air - water interface .
ellipsoids deform the air - water interface , creating an effective elastic membrane with a high bending rigidity .
when enough ellipsoids are present , pinning and bending the interface becomes energetically costly and the spheres ( and ellipsoids ) are deposited as the interface recedes .
further , this behavior in confined geometries is different than that of sessile drops ( see section 3.8 and @xcite ) .
from this perspective , it is somewhat surprising that small spheres are deposited uniformly from droplets doped with small numbers of ellipsoids and confined between glass plates .
interestingly , this method of producing a uniform deposition is similar to convective assembly techniques wherein the substrate , or a blade over the substrate , is pulled away from the contact line in a colloidal suspension ; a thin film is thus formed that leads to the creation of a monolayer of particles ( e.g. , @xcite ) . unlike many other convective assembly techniques , the present experimental system has neither moving nor mechanical parts .
uniform coatings are created essentially as a result of shape - induced capillary attractions which produce cmms that are hard to bend .
colloidal drops evaporating in confined geometries behave quite differently the evaporating sessile drops .
ellipsoids adsorbed on the air - water interface create an effective elastic membrane , and , as particle anisotropy aspect ratio increases , the membrane s bending rigidity increases faster than its young modulus . as a result ,
when a drop of a colloidal suspension evaporates in a confined geometry , the different interfacial elastic properties produce particle depositions that are highly dependent on particle shape .
the ability to increase cmm bending rigidity by increasing particle shape anisotropy holds potentially important consequences for applications of cmms . for example , increased bending rigidity may help stabilize interfaces ( e.g. , pickering emulsions @xcite ) and thus could be useful for many industrial applications , e.g. , food processing @xcite . in a different vein , the observations presented here suggest the buckling behavior of cmms in confined geometries may be a convenient model system to investigate buckling processes relevant for other systems , e.g. , polymeric membranes @xcite , biological membranes @xcite , and nanoparticle membranes @xcite .
in the previous sections , we showed how particle shape influences the behaviors of drying drops containing colloidal particles . for sessile drops we found that particle anisotropy could be employed to overcome the coffee ring effect ; for drops in confinement
, we found that particle anisotropy dramatically affected the bending rigidity of the air - water interfaces which in turn modified particle deposition during drying . besides particle shape
, many other ideas have been observed , developed , and utilized over the years to manipulate the drying behaviors of colloidal drops @xcite . in the final section of this review paper ,
we describe our foray into the effects of added surfactants .
a surfactant is a surface - active molecule that consists of a hydrophobic and a hydrophilic part . in water ,
surfactant molecules populate the air - water interface with their hydrophobic parts `` sticking out of the water , '' thereby reducing the water s surface tension ( which is paramount for the cleaning effects of soaps or dish washers ) . in an immiscible mixture of water and oil ,
surfactants populate the interfaces between components , thus stabilizing the emulsion . in an evaporating drop of an aqueous colloidal suspension , surfactants give rise to other effects .
herein we describe video microscopy experiments which investigate how a small ionic surfactant ( mostly ) affects particle deposition in drying drops ; these surfactants induce a concentration - driven marangoni flow on the air - water interface and a strong `` eddy''-like flow in the bulk that prevents particles from depositing in the coffee ring and thus suppresses the coffee ring effect for spheres .
although we focus here primarily on small ionic surfactants , we have explored the effects of a variety of surfactants . in general ,
common types are small , ionic surfactants , e.g. , sodium dodecyl sulfate ( sds ) , or large , polymeric ones , e.g. , pluronics ; the chemical structures of both examples are shown in fig .
[ fig : structures ] .
accordingly , surfactants can affect deposition phenomena in a variety of ways .
for example , it was found that sds can change the deposition patterns from aqueous colloidal drops @xcite . in different experiments
, surfactant is sprayed onto the drop @xcite , leading to complex patterns as a result of thermodynamic transitions between different phases formed by the surfactant . if the surface tension is heterogeneous on a liquid surface ( e.g. , the air - water interface of a drop ) a flow is induced from regions of low to high surface tension .
this effect is the so - called marangoni flow .
such marangoni flows can result from different temperatures at drop edge and center , e.g. , because of different evaporation rates and slow diffusive heat transfer ; thus , in principle such a flow should be present at the air - water interfaces of drying liquid drops @xcite .
indeed , marangoni radial flows towards the center of a drop have been found in small drops of octane @xcite . in water , however , such temperature - dependent marangoni flows are suppressed @xcite .
in addition to temperature - driven changes of the surface tension , surfactant - driven marangoni flows have been suggested to explain the relatively uniform deposition of dissolved polymer from droplets of organic solvent containing surfactant @xcite .
when the local concentration of surfactant molecules at the pinned contact line increases due to the coffee - ring effect , then the surface tension of the drop decreases locally , and a gradient in surface tension arises .
this gradient has been suggested as the source of continuous marangoni flow towards the center of the drop @xcite .
herein , we first investigate the mechanism of a small ionic surfactant , sds , on the evaporation of aqueous colloidal systems and their resulting particle coatings @xcite .
the experiments demonstrate that such small ionic surfactants do indeed produce marangoni flows in colloidal droplets , not only in agreement with the model suggested for polymer solutions @xcite , but also providing a first direct visualization .
we further demonstrate how the `` marangoni eddy '' can lead to uniform particle deposition during drying , thereby undermining the coffee ring effect . at the end of this section on surfactants
we show preliminary experiments which demonstrate that large polymeric surfactants like pluronic f-127 influence the evaporation of drops in a strikingly different way than small ionic surfactants . in this case ,
contact line pinning is prevented , leading to a uniform particle deposition .
we suggest an explanation of this behavior as due mainly to an increase of viscosity near the contact line , which is a result of high polymer concentration because the dissolved polymeric surfactant is transported to the contact line by the coffee - ring flow . the procedure for these experiments has been described previously @xcite , but for completeness and presentation clarity we briefly discuss these methodologies below .
we focus on a few representative systems of evaporating drops with the small ionic surfactant sds or the large polymeric surfactant pluronic f-127 , respectively , and we attempt to elucidate rules governing their behavior .
we employed aqueous suspensions of colloidal polystyrene ( ps ) particles ( diameter @xmath250 nm , synthesized by surfactant free radical emulsion polymerization , and stabilized by sulfate groups @xcite ) .
suspensions were prepared with deionized water , filtered by a millipore column , and then the suspensions of ps spheres and sds ( sigma - aldrich ) or pluronic f-127 ( basf ) ( in different compositions ) were thoroughly mixed by a vortexer and ultrasonicated for five minutes .
evaporation experiments were observed using a brightfield microscope with air objectives ( magnification 5x to 100x ) .
clean hydrophilic glass substrates ( fisher scientific ) were used as evaporation substrates .
( note : qualitatively similar results were found on hydrophobic cover slips . )
the drop volume was about 0.05
l , leading to deposition coatings with diameters of 1 to 3 mm .
the evaporation process was recorded by a video microscopy ( camera resolution 658x494 pixel , 60 frames per second ) with total evaporation times between 2 and 4 minutes .
all experiments were repeated several times in order to identify a consistent concentration - dependent behavior .
photographs of the entire deposit , obtained after evaporation , were taken by combining up to four high - resolution photographs when the deposition area was larger than the microscope field of view @xcite .
[ fig : sds1]a - d shows top views of the deposition pattern of an aqueous suspension of ps spheres ( 0.5 wt% ) ( * a * ) and similar suspensions but with different concentrations of sds ranging from 0.05 wt% to 1.0 wt% ( * b - d * ) . the coffee - ring effect is observed in sample * a * , i.e. , the vast majority of spheres are deposited in a thin ring located at the initial pinned contact line , and very few particles are deposited in the center of the drop .
the deposition changes slightly upon adding a small amount of sds ( 0.05 wt% , * b * ) .
specifically , the coffee - ring broadens and more particles are deposited in the center of the drop . at higher sds concentrations ( 0.5 wt% ( *
c * ) and 1.0 wt% ( * d * ) ) , however , the deposition pattern changes drastically . instead of a single ring at the initial pinned contact line , tree - ring like structures
are observed with several distinct deposition lines .
these tree - ring deposition structures can be explained by stick - slip dynamics of the drop s contact line along a large part of the perimeter ; after the edge depins and the drop shrinks , a new contact line stabilizes very quickly via self - pinning by other ps particles in suspension @xcite .
mutiple depinning and repinning events produce the observed pattern . inside these tree ring - like structures , both systems exhibit relatively uniform depositions of spheres about their centers , surrounded by dark `` flares '' .
clearly , the addition of surfactant has large influence on how the particles are deposited . but how can the effects we observe be explained ? to answer this question , we studied the temporal evolution of drops by video microscopy during evaporation .
[ fig : sds2 ] shows snapshots of the drying drop in fig .
[ fig : sds1]c ( 0.5 wt% sds ) at three different stages : * d * at the begining of evaporation , * e * after about 50% of the total evaporation duration , @xmath251 , has passed , and * f * after evaporation is complete .
high magnification images taken from a similar drop with identical composition are shown as well .
the first thing we noticed is that when the evaporation starts , drop behavior appears identical to that of drops without sds , i.e. , the contact line is pinned and spheres initially pack densely at the drop s edge arrange in a densely packed structure at the edge ( see fig . [
fig : sds2]e inset ) .
the image gets progressively darker towards the drop center where the drop is thickest as particles in the bulk are evenly distributed .
however , even at this early stage of evaporation , some spheres flow towards the drop edge but do not reach it .
rather , as they approach the edge , they are repelled back towards the drop center .
as we know from the early studies of the coffee ring effect @xcite , with advancing evaporation time , the flow towards the edge becomes stronger . in our experiment ,
more and more particles approach the edge but do not reach it .
these particles appear to be captured within a certain region of the drop , which is highlighted yellow in fig .
[ fig : sds2]f .
they form a broad corona , i.e. , an outer rim distinctly different and separated from the inner part of the drop , located between the relatively uniform dark center and the coffee - ring .
we describe the dark part of the corona as a `` marangoni eddy '' or circulating region of ps spheres that are transported towards and away from the drop edge throughout the drying process ( see fig .
[ fig : sds2]f ) .
ps spheres are pulled into the eddy ( see fig . [
fig : sds2]f ) , leading to a locally reduced number of particles in the depletion zone ( that explains why this region is less dark then the other regions ) .
the trajectory of an individual sphere is marked by the three numbers ( 1 , 2 , 3 ) in fig .
[ fig : sds2]f .
initially , the sphere is approximately in the middle of the eddy ( 1 ) .
after @xmath252 seconds , the sphere is pushed radially outward , i.e. , towards the coffee ring ( 2 ) . however , after another@xmath252 seconds , the sphere is pushed radially inward , i.e. , towards the region between the eddy and the depletion zone ( 3 ) .
video microscopy shows us that the same behavior is observed for virtually all of the particles at later times during the evaporation ( see supporting online material for @xcite ) the experiments provide evidence that the observed deposition behavior is dominated by a surfactant - driven marangoni effect . as noted above , related surfactant - driven phenomena
were recently observed in drying polymer solutions containing oligomeric fluorine - based surfactants @xcite .
however , the previously studied polymer solutions differ qualitatively from the aqueous colloidal suspensions presented here .
drying polymer solutions can exhibit gelation @xcite .
further , the local surface tension in drying drops of polymer solutions depends on the local solute concentration @xcite .
observing particle motion in real - time facilitates a comprehensive understanding of this phenomenon .
a cartoon of the mechanism is shown in [ fig : sds2]g .
the `` eddy '' forms in between the yellow bars in fig . [ fig : sds2]g , which corresponds to the highlighted region in fig . [ fig : sds2]f in ( top view ) . shortly after a drop is created , some surfactant molecules ( pictured in the cartoon as hydrophilic `` heads '' with hydrophobic `` tails '' ) adsorb on the water / air interface .
note , the air - water interface is the energetically preferred location for the amphiphilic sds molecules .
however , electrostatic repulsion of anionic heads prevents them from forming maximally dense steric equilibrium packing .
additionally , at sufficiently high concentration sds molecules are also dissolved in the bulk , either freely or as micelles ( see fig .
[ fig : sds2]d ) . as was the case with no surfactant
is added , the contact line is initially pinned .
thus , the outward convective flow that is responsible for the coffee - ring effect is present and transports spheres and sds molecules to the drop edge . as a result
, the air - water interface near the contact line becomes more concentrated with sds molecules which locally decrease the interfacial surface tension @xmath4 .
this creates a surface tension gradient along the air - water interface , @xmath253 , which is resolved by a marangoni flow from low to high @xmath4 .
this strong surface flow penetrates into the bulk fluid , so that it can carry colloidal spheres near , but not on the interface , towards the drop center .
as spheres flow towards the drop center , the local sds concentration decreases ( and the local @xmath4 increases ) , and the marangoni flow weakens .
eventually , the radially outward bulk convective flow that drives the coffee - ring effect dominates the radially inward marangoni flow .
particles that travel to this point are carried towards the drop s edge once again ; the process then repeats .
although the sds molecules are too small to be observed optically , sds molecules likely participate in the eddy as well .
otherwise , the surface would become saturated with sds , which , in turn , would end or at least weaken the marangoni flow .
thus , particles are trapped in a circulating flow driven by the local surfactant - concentration , which we call the `` marangoni eddy '' .
we find that the marangoni flows become stronger at sds concentrations above its critical micelle concentration ( cmc ) in water ( 8.3 [email protected] wt%@xcite ) .
again , as was the case for drops without sds , at late times the contact line is observed to depin . however , due to the presence of the marangoni eddy , when the final depinning occurs , many particles are left in the more central regions of the bulk , because the eddy prevented them from attaching to the edge . after the contact line depins ,
the particles that remain in the bulk are deposited onto the substrate as the radially - inward - moving contact line passes them .
the radially - inward contact line thus leaves behind a relatively uniform particle deposition in the drop center .
the formation of a `` marangoni eddy '' is a prerequisite for the relatively uniform particle deposition in the drop center ; it prevents many particles from depositing at the drop s edge , and thereby delays their deposition until times when the contact line has depinned .
= 1330 nm ) and different concentrations of pluronic f-127 ( indicated under the pictures ) .
the drops were evaporated on a hydrophilic microscope slide , the initial drop volume was about 0.05 l , and the resulting deposition area is between 1 - 2 mm in diameter . * e - h * three sets of snapshots of evaporating water drops containing 1.0 wt% ps particles ( @xmath185=1330 nm ) and 2 wt% pluronic f-127 at different states of the evaporation.,scaledwidth=90.0% ]
lastly , we describe preliminary experiments with non - ionic triblock polymer surfactants such as pluronic f-127 is present ( @xmath255 , @xmath256 , cf .
[ fig : structures ] ) . fig .
[ fig : pluronic1 ] is analogous to fig .
[ fig : sds1 ] , but with 1 wt% ps particles and 0.5 , 1.0 , and 2.0 wt% pluronic . .
pluronic f-127 is a relatively large molecule ( @xmath257 g / mol ) ; in the investigated drops , the amount of pluronic is about the same ( by weight ) as colloidal spheres .
the addition of pluronic leads to a systematic change in the deposition pattern ; as more surfactant is added , the initial ring becomes broader , and eventually the entire area is coated uniformly ( on a macroscopic scale ) with ps spheres .
additionally , at lower concentrations of pluronic , complex deposition patterns appear in the center of the drop area .
video microscopy indicates that a marangoni eddy is not present during drying . instead
, pluronic induces an early depinning of the contact line and a loose packing of spheres why does nt a surfactant like pluronic f-127 produce a `` marangoni eddy ? '' like sds , the pluronic f-127 molecule is amphiphilic and the coffee ring effect should also transport it to the edge were it could , in principle , give rise to the same flow effects as sds .
cui et al .
attribute similar behaviors found in samples containing poly(ethylene oxide ) to a combination of several effects @xcite .
their most important argument is that dissolved polymer is transported to the drop edge where it leads to a dramatic increase of viscosity such that the suspended colloidal particles are immobilized before they reach the contact line
. this argument does not provide insight about why the contact line should move , but we speculate that the hydrophobicity of the deposited polymer ( in our case , surfactant ) may play a crucial role .
interestingly , some other surfactants in the pluronic family , principally the same structure ( cf . fig . [ fig : structures ] but with different block lengths @xmath258 , show a strong marangoni eddy . specifically , different pluronics ( all basf ) were explored and their deposition patterns are shown in fig .
[ fig : pluronic6 ] ; these surfactants include pluronic f-68 prill ( @xmath259 , @xmath260 , @xmath261 kda ) , p-85 ( @xmath262 , @xmath263 , @xmath264 kda ) , and p-123 ( @xmath265 , @xmath256 , @xmath266 kda ) .
for all three pluronics , fig .
[ fig : pluronic6 ] shows a snapshot of the evaporating drop at @xmath267 on the left and a dark - field microscopy photograph of the deposition after drying is completed .
interestingly , for f-68 ( a ) and p-85 ( b ) , a marangoni eddy similar to that seen for sds appears .
correspondingly , the deposition pattern of drops with these surfactants is more similar to the case of sds than to pluronic f-127 .
on the other hand , pluronic p-123 ( c ) leads to the same phenomenon as f-127 , revealing a mostly uniform , loose deposition with no evidence of a marangoni eddy .
a comparison of the molecular properties of all poloxamer surfactants shows that neither the total molecular weight nor the ratio @xmath268 is the parameter that governs the drop evaporation . in total
, the examples in this review illustrate the complexity of the coffee ring problem . in a few carefully controlled situations
is the deposition of particles from a drying drop is dominated by a single effect . in most cases , several cooperative or antagonistic effects act at the same time , preventing easy predictions and phenomenological understanding of the underlying principles needed to open new pathways towards further technological application of the coffee ring effect or its circumvention @xcite .
we thank matthew gratale , matthew a. lohr , kate stebe , and tom c. lubensky for helpful discussions .
we also gratefully acknowledge financial support from the national science foundation through dmr-0804881 , the penn mrsec dmr11 - 20901 , and nasa nnx08ao0 g . | we explore the influence of particle shape on the behavior of evaporating drops . a first set of experiments discovered that particle shape modifies particle deposition after drying . for sessile drops ,
spheres are deposited in a ring - like stain , while ellipsoids are deposited uniformly .
experiments elucidate the kinetics of ellipsoids and spheres at the drop s edge . a second set of experiments examined evaporating drops confined between glass plates .
in this case , colloidal particles coat the ribbon - like air - water interface , forming colloidal monolayer membranes ( cmms ) . as particle anisotropy increases ,
cmm bending rigidity was found to increase , which in turn introduces a new mechanism that produces a uniform deposition of ellipsoids and a heterogeneous deposition of spheres after drying .
a final set of experiments investigates the effect of surfactants in evaporating drops .
the radially outward flow that pushes particles to the drop s edge also pushes surfactants to the drop s edge , which leads to a radially inward flow on the drop surface .
the presence of radially outward flows in the bulk fluid and radially inward flows at the drop surface creates a marangoni eddy , among other effects , which also modifies deposition after drying .
[ 1999/12/01 v1.4c il nuovo cimento ] |
this paper represents a second part of our research dedicated to the theory of diffusion of hopping charge carriers biased by electric field in disordered systems with a gaussian energy distribution of the energy of the localized states : @xmath1 here @xmath2 is the spatial concentration of sites available for hopping transport and @xmath3 is the energy scale of their density of states ( dos ) .
the gaussian dos is assumed to apply for disordered organic materials , such as molecularly doped and conjugated polymers and organic glasses.@xcite while in the previous paper @xcite one - dimensional transport was considered , in the current paper we present results for two- and three - dimensional systems .
this study has been on one hand stimulated by numerous experimental studies on organic disordered materials,@xcite which claim the invalidity of the conventional form of the einstein relation between the carrier mobility @xmath4 and the diffusion coefficient @xmath5 @xmath6 on the other hand our study is stimulated by the lack of a concise theory for the diffusion biased by electric field in the hopping transport mode .
it is , however , this transport mode that dominates the electrical conduction in disordered organic materials where transport is due to incoherent tunnelling of electrons and holes between localized states randomly distributed in space , with the dos described by eq .
( [ eq - gaussian - dos]).@xcite the transition rate between an occupied state @xmath7 and an empty state @xmath8 , separated by the distance @xmath9 , is described by the miller - abrahams expression @xcite @xmath10 where @xmath11 is the attempt - to - escape frequency .
the energy difference between the sites is @xmath12 where the electric field @xmath13 is assumed to be directed along the @xmath14-direction .
the localization length of the charge carriers in the states contributing to the hopping transport is @xmath15 .
we assume the latter quantity to be independent of energy and we will neglect correlations between the energies of the localized states , following the gaussian - disorder - model of bssler .
@xcite the field - dependent diffusion in such systems in the 3d case has so far been studied by computer simulations .
richert , pautmeier , and bssler performed monte carlo simulations in 3d and showed that the longitudinal diffusion coefficient @xmath16 is strongly dependent on the electric field and that the dependence is quadratic at such low fields that the mobility of charge carriers remains field - independent.@xcite this is in contrast to the linear field dependence of the diffusion coefficient at low fields obtained by the exact analytical theory and by numerical calculations for 1d systems in the preceding paper.@xcite in order to clarify the nature of the field effect on the diffusion for 3d and 2d systems we suggest in the current paper an analytical theory for the field - dependent diffusion coefficient in the hopping regime for such systems .
this theory confirms the conclusion of richert , pautmeier , and bssler@xcite about the parabolic field dependence of @xmath16 at low fields .
furthermore , our theory gives explicit analytical expressions for the combined effects of the electric field and temperature on the hopping diffusion coefficient .
these expressions predict a violation of the traditional form of the relation between @xmath4 and @xmath5 given by eq .
( [ eq - einstein ] ) at very low electric fields .
since the temperature effect on the field - dependent diffusion has been so far left out of the scope of computer simulations,@xcite we perform here a monte carlo study of the field- and temperature - dependent diffusion in the hopping regime for 2d and 3d systems in order to check the results of the analytical theory .
our computer simulations for the 2d and 3d cases presented in sec .
[ sec - simulation ] support the analytical theory .
furthermore , numerical results evidence that while the carrier mobility is stable with respect to different realizations of disorder , the diffusion coefficient experiences significant fluctuations from one realization to another , even for systems containing millions of localized states .
the reasons for such different behavior between @xmath4 and @xmath5 is clarified and the estimates for the system size , which is necessary to obtain stable values of @xmath5 , are given in sec .
[ subsec - hopping - transport ] .
the latter estimates show that previous numerical simulations in the literature were performed on rather small systems , insufficient for obtaining reliable results for the field - dependent diffusion coefficient at low temperatures . in developing the analytical theory for the field - dependent diffusion coefficient for the hopping transport regime we rely on the analogy between the hopping transport mode and the multiple - trapping mode.@xcite in sec .
[ sec - theory ] we present a general solution of the problem . in sec .
[ sec - comparison ] the analytical expressions from sec [ sec - theory ] are compared to simulation results .
concluding remarks are gathered in sec .
[ sec - conclusions ] .
in this section we study the effects of field and temperature on the diffusion coefficient by monte carlo simulations in more detail than it has been done previously , aiming at a comparison with the analytical theory described in the following sections .
furthermore , we study in detail the role of the system size on the simulated results and show that in order to get reliable results for the field - dependent diffusion at reasonably low temperatures one needs to perform simulations on enormously large systems .
the system is modelled as a lattice of @xmath17 or @xmath18 sites with lattice constant @xmath19 and the site energies @xmath20 chosen randomly according to eq .
( [ eq - gaussian - dos ] ) .
periodic boundary conditions are applied in all directions .
hops inside a square of @xmath21 sites ( 2d case ) or a cube of @xmath22 sites ( 3d case ) , centered at the starting site are allowed .
the simulation proceeds as follows .
a packet of @xmath23 non - interacting carriers is allowed to move in the lattice until a fixed time @xmath24 has passed .
the mobility @xmath4 is calculated from the average distance that the charge carriers have moved along the direction of the field , while the longitudinal diffusion coefficient @xmath16 is calculated from the width of the carrier packet : @xmath25 further details of the simulation algorithm and our implementation of it are given in appendix [ sec - mc - details ] .
care was taken about the necessary size of the simulated system in order to avoid finite - size effects .
the corresponding number of sites was @xmath26 in the 3d case and @xmath27 in the 2d case .
simulation results for the diffusion coefficient both along and perpendicular to the electric field are shown in fig .
[ fig - d - and - mu ] , together with the mobility ( scaled with @xmath28 ) .
the localization length was @xmath29 .
a packet consisting of @xmath30 charge carriers was simulated for each data point , and the simulation results were averaged over five different realizations of disorder . at extremely low fields
it is seen that all three of the plotted quantities are equal , which means that einstein s relation , eq .
( [ eq - einstein ] ) , is valid . with rising magnitude of the electric field the longitudinal diffusion coefficient increases drastically , while the mobility and transversal diffusion coefficient remain field - independent up to much higher fields .
the solid line shows a fit to the longitudinal diffusion coefficient by a square trial function @xmath31 it is seen that the square function well fits the data in agreement with the results of previous simulations.@xcite scaled with @xmath28 as a function of the field strength , ( a ) in 3d , and ( b ) in 2d . at small fields all three quantities are equal , implying that the einstein relation is valid.,title="fig:",width=8 ] scaled with @xmath28 as a function of the field strength , ( a ) in 3d , and ( b ) in 2d . at small fields
all three quantities are equal , implying that the einstein relation is valid.,title="fig:",width=8 ] for hopping in 3d , as a function of the applied electric field @xmath13 .
the solid lines show the best fit to the square function ( [ eq - quadratic - d]).,width=8 ] in order to study the effect of temperature on the field - dependent diffusion coefficient , the simulations were repeated for different temperatures .
the results for the longitudinal diffusion coefficient in 3d are collected in fig .
[ fig-3d - fdep ] .
the data are fitted with the square trial functions ( [ eq - quadratic - d ] ) , shown in the figure by solid lines .
the fit is good for all temperatures .
the decrease of @xmath32 at very high fields for the two highest temperatures is due to the trivial saturation effect well - known for the one - dimensional random - energy model.@xcite we focus in the following on the field dependence of @xmath16 at field magnitudes lower than the one at which @xmath16 starts to decrease with increasing field . , the diffusion coefficient at @xmath33 and @xmath34 , obtained from fig .
[ fig-3d - fdep ] for the localization length @xmath29.,width=8 ] the temperature dependences of the coefficient @xmath34 in the quadratic field - dependent term for the longitudinal diffusion coefficient @xmath16 , that were obtained in the 2d and 3d cases by fitting the simulation results in fig .
[ fig-3d - fdep ] by eq .
( [ eq - quadratic - d ] ) are shown in fig .
[ fig-3d - tdep ] along with the field - independent term @xmath35 .
analytical expressions for @xmath34 will be derived in sec .
[ sec - theory ] .
when performing computer simulations for the field - dependent diffusion one should be cautious with the choice of simulation parameters .
the field - induced spatial spreading of the carrier packet is caused by trapping of some carriers onto localized states deep in energy , while the other carriers continue their motion in shallow states being biased by the electric field . in order to obtain reliable results for the field - dependent diffusion coefficient
, one should guarantee the presence of such deep - in - energy states in the simulated system . since the dos given by eq .
( [ eq - gaussian - dos ] ) rapidly decreases for the deep - in - energy states and hence the sites with deep energies are rare , one has to simulate large systems .
the simulation time @xmath24 must also be chosen so large that the charge carriers have time to visit the deep traps , which control the diffusion . in our simulations
@xmath24 was chosen for each temperature so that the typical number of hops for each carrier was at least @xmath36 .
the importance of the deep - in - energy states for the field - dependent diffusion is demonstrated in fig .
[ fig - cutdos ] , where the values of @xmath16 ( at the field @xmath37 ) and of @xmath4 ( at low fields ) for the temperature @xmath38 are given when calculated with a cut - off of the dos below some energy @xmath39 . for these calculations
the normalization of the dos was kept , while the states with energies below @xmath39 were excluded from the simulation .
the mobility is almost unaffected by the cutting as long as @xmath40 , while the diffusion coefficient drastically decreases when sites with much smaller energies , around @xmath41 , are removed .
the result for the mobility @xmath4 is not surprising .
it has been predicted in the analytical theory@xcite that the hopping mobility in the gaussian dos for the diluted set of carriers is determined by sites with energies in the close vicinity of the average carrier energy @xmath42 . for @xmath43 ,
this energy is @xmath44 , which explains the data for @xmath4 in fig .
[ fig - cutdos ] .
the result for the diffusion coefficient in fig .
[ fig - cutdos ] shows however that rare sites with even lower energies than @xmath45 cause the strong dependence of the diffusion coefficient on the electric field . in sec .
[ subsec - hopping - transport ] we show that the most important sites for the field - dependent diffusion have energies around @xmath46 , which is much deeper than @xmath45 .
this result also agrees with the data shown in fig .
[ fig - cutdos ] : for @xmath47 , @xmath48 .
the diffusion coefficient changes drastically when the few sites with energies below @xmath49 are removed , while the mobility starts changing only when @xmath39 is in the vicinity of @xmath45 . in our simulations
a lattice of @xmath26 sites has been used .
there are typically no sites with energies below @xmath50 in such a system .
therefore , the @xmath32 data for @xmath51 can not be considered as reliable in the whole simulated range of electric fields . however , for @xmath52 , the lowest temperature considered , @xmath53 .
energies around @xmath54 are present in our lattice , and thus results for @xmath55 can be considered as reliable for the lattice of @xmath26 sites .
it seems extremely lucky that in previous simulations with the lattice of just @xmath56 sites@xcite meaningful results were claimed for @xmath51 .
( at the field @xmath37 ) and mobility @xmath4 ( at low field ) in a 3d system where sites with energies below @xmath39 are absent .
the temperature is @xmath51 .
, width=8 ] the analytical results in the preceding paper@xcite were obtained for nearest - neighbor hopping , while the simulations above allowed also longer hops . to exclude this difference between the models as the cause of the different field dependencies
, the simulations for two- and three - dimensional systems were repeated with only nearest - neighbor hopping allowed .
the mobility and diffusion coefficient obtained in this case were somewhat lower than those for the variable range hopping , however the parabolic shape of the field dependence for the diffusion coefficient did not change .
our numerical results can be qualitatively summarized as follows : ( i ) the diffusion coefficient _ along _ the electric field depends parabolically on the field strength @xmath13 at low fields ; ( ii ) the diffusion coefficient _ perpendicular _ to the field is field - independent in the range of fields where ohm s law is fulfilled ; ( iii ) the field - dependent part of the diffusion coefficient rapidly decreases with increasing temperature ; ( iv ) the field - dependent part of the diffusion coefficient is very sensitive to sites which energies are lower than the mean carrier energy .
the discussion above , in particular the latter statement on the decisive role of rare sites with very deep energies that can hardly be found in the finite simulation arrays , raises the task of developing an analytical theory for the diffusion process in the hopping regime enhanced by an electric field .
there is no such theory in the literature so far .
the only relevant theory is the one developed by rudenko and arkhipov for _ band transport _ in materials with traps.@xcite although this theory predicts a parabolic field dependence of the longitudinal diffusion coefficient , it can not be directly applied to _ hopping transport _ , because it operates with quantities that are specific for band conductivity ( for example , the effective density of states in the conduction band ) .
therefore it is necessary to develop a theory for hopping transport , particularly because monte carlo simulations suffer from finite - size effects as described above .
we will give such a theory in sec .
[ sec - theory ] .
the origin of the field - dependent diffusion can be understood qualitatively with the aid of a spatio - temporal picture of the carrier distribution sketched in fig .
[ fig - sketch ] .
the small dots indicate the scatter of carriers _ after some definite number of jumps _ , assuming that all carriers start at the same time @xmath57 from the same point . to get the spatial distribution of carriers _ at some definite time _
@xmath58 , one may `` project ''
these dots from the starting point to the line @xmath59 .
the direction of `` projecting '' is determined by the drift velocity . when the drift velocity is equal to zero ( no electric field , the left part of fig .
[ fig - sketch ] ) , the spatial distribution of carriers at @xmath59 does not depend on the scatter of times spent by the carriers in order to perform a fixed number of jumps . on the contrary , in the case of drift caused by an electric field ( the right part of fig .
[ fig - sketch ] ) , this scatter in times `` projects '' into the line @xmath59 , which gives rise to the broadening of the spatial distribution at @xmath59 .
this broadening is the reason for the enhancement of the diffusion coefficient due to electric field .
the above consideration shows that fluctuations of the durations of jumps are responsible for the field - induced diffusion .
these fluctuations can be especially pronounced for systems with a broad distribution of site energies , because jumps from energetically deep sites to transport sites demand exponentially long times .
let us start from a general form of the _ quantitative _ description for the field dependence of the diffusion coefficient .
our consideration is based on the assumption that _ successive jumps are statistically independent _ , i. e. the diffusion process is markovian .
the latter means that increments in carrier coordinates at a given jump , as well as the time interval from the preceding jump till the given one , do not depend on the carrier prehistory .
this assumption definitely does not hold for the one - dimensional hopping transport considered in the previous paper.@xcite in the 1d case , the probability of returning to an already visited trap is not negligible , and hence the consequent jumps must be correlated . therefore our analytical consideration based on the assumption of the statistical independence of the successive jumps present below can be valid only for 2d and 3d cases since it is reasonable to assume that in the latter cases the carrier trajectories are non - returning .
let @xmath58 be some fixed moment of time ( @xmath58 is large compared to the mean time of a jump ) ; @xmath60 , @xmath61 and @xmath62 be the carrier s displacements along the axes @xmath63 , @xmath64 and @xmath14 during the time interval @xmath65 $ ] ; @xmath66 , @xmath67 , @xmath68 and @xmath69 be coordinate displacements and a time increment related to only one jump , respectively ; @xmath70 , @xmath71 , @xmath72 be the displacements after @xmath23 successive jumps .
angle brackets will denote averaging over jumps performed by different carriers , equivalent to the averaging over successive jumps of one carrier .
we start with the simple case of zero electric field . since there is no drift ,
the expectation values of @xmath60 , @xmath61 and @xmath62 vanish , and one obtains for the diffusion coefficients @xmath73 , @xmath74 , @xmath16 @xmath75 for large @xmath58 , the displacements @xmath60 , @xmath61 , @xmath62 are approximately equal to @xmath76 , @xmath77 , @xmath78 displacements after @xmath2 jumps , where @xmath79 is the mean number of jumps during the time @xmath58 .
consequently , @xmath80 in the limit @xmath81 this approximate equality becomes exact , and one gets @xmath82 let us now consider the case of a finite electric field along the @xmath14-axis .
since the expectation values of @xmath60 and @xmath61 are still zero , all the considerations above remain valid , and eq . ( [ eq - d_zero_field ] ) remains correct with respect to @xmath73 and @xmath74 .
one can not obtain @xmath16 by literally the same way , because @xmath83 .
instead , one may apply this argumentation to a variable @xmath84 , where @xmath85 is the drift velocity . since @xmath86 , one gets @xmath87 and finally @xmath88 since the mean values @xmath89 , @xmath90 and @xmath91 depend on the electric field , eq .
( [ eq - d_nonzero_field ] ) describes the field - dependent diffusion along the field direction . on the other hand
, there is no reason for the mean values @xmath92 , @xmath93 and @xmath94 to be field - dependent in small electric fields .
therefore , according to eq .
( [ eq - d_zero_field ] ) , the transversal diffusion coefficients @xmath73 and @xmath74 are expected to be constant inside the ohmic regime .
let us discuss the shape of the dependence @xmath32 near the point @xmath33 . since @xmath89 is proportional to the electric field ( @xmath95 , where @xmath4 is mobility ) , the _ third _ term in the r .- h .
side of eq .
( [ eq - d_nonzero_field ] ) gives a contribution to @xmath32 that is quadratic in @xmath13 .
in the _ first _ term , @xmath96,\ ] ] where @xmath97 $ ] denotes the standard deviation of a random variable @xmath15 .
since @xmath98 $ ] is not sensitive to the electric field , the first term in the r .- h .
side of eq .
( [ eq - d_nonzero_field ] ) is a sum of a constant and a term quadratic in @xmath13 .
the behavior of the _ second _ term in the r .- h
. side of eq .
( [ eq - d_nonzero_field ] ) depends on the symmetry of the system . if in the absence of electric field the directions @xmath14 and @xmath99 are equivalent , then @xmath100 at @xmath33 . in this case one
expects that @xmath101 and consequently the second term in the r .- h .
side of eq .
( [ eq - d_nonzero_field ] ) is quadratic in @xmath13 .
however if the positive and negative directions along the @xmath14-axis are non - equivalent , @xmath91 can be non - zero at @xmath33 , which gives a contribution to the diffusion coefficient linear in @xmath13 . in summary , if the directions @xmath14 and @xmath99 are equivalent , which we will assume in the following , the longitudinal diffusion coefficient @xmath16 at small fields is described by eq .
( [ eq - quadratic - d ] ) where @xmath35 and @xmath34 are field - independent coefficients .
let us first apply our approach to the multiple - trapping ( mt ) model of _ band _ conductivity .
the model includes processes of capture of free carriers by traps , emission of trapped carriers , and free motion ( brownian motion plus drift in the electric field ) of carriers in the band ( fig .
[ fig - multi - trapping ] ) .
the band is characterized by the effective density of states @xmath102 , the mobility @xmath103 and the diffusion coefficient @xmath104 of free carriers .
the traps are characterized by the density of states @xmath105 and the capture rate ( for unit free carrier concentration ) @xmath106 .
the energy @xmath107 is counted from the band edge .
the problem of the field - effect on the diffusion coefficient in the mt model has been considered by rudenko and arkhipov,@xcite who treated the evolution in time of the one - dimensional carrier distribution function . in ref . , transport was assumed to take place in quasi - equilibrium , i. e. , the distribution function was assumed to change slowly in comparison with the rate of exchange between traps and the conduction band .
our approach is free of the assumption of quasi - equilibrium transport .
besides , our consideration provides information not only about the longitudinal diffusion coefficient , but also about the transversal one . instead of examining the carrier distribution function , we will follow the motion of a _ single _ carrier and consider the statistics of this motion .
in order to use the expressions ( [ eq - d_zero_field ] ) and ( [ eq - d_nonzero_field ] ) , we represent the motion of a carrier as a sequence of `` jumps '' , each of them beginning at the moment of escape from a trap .
successive `` jumps '' are statistically independent , because the processes of capture take place in different points in space and there is no correlation between them .
each `` jump '' of a carrier consists of two contributions : free motion and sitting on a trap .
let @xmath108 be the time of free motion ( between emission and capture ) , and @xmath109 be the time of being trapped ( between capture and emission ) ; @xmath108 and @xmath109 are independent random variables .
we denote their expectation values @xmath110 and @xmath111 as @xmath112 and @xmath113 , respectively .
since @xmath112 is the lifetime of free carriers , its reciprocal value is the capture rate : @xmath114 the variable @xmath108 obeys an exponential distribution .
hence its mean square @xmath115 is equal to @xmath116 .
collecting all this information , one gets the following statistical expressions for a single `` jump '' ( electric field @xmath13 is directed along @xmath14 ) : @xmath117 using these mean values , one can evaluate the mobility : @xmath118 the transversal diffusion coefficient via eq .
( [ eq - d_zero_field ] ) @xmath119 and the longitudinal diffusion coefficient via eq .
( [ eq - d_nonzero_field ] ) @xmath120 the parameters @xmath112 , @xmath113 , and @xmath121 are governed by the capture cross - sections and emission rates of the traps . for moderate electric fields
, these cross - sections and rates can be regarded as field - independent , and consequently one can use _ equilibrium _
values for @xmath112 , @xmath113 , and @xmath121 .
this gives an opportunity to simplify eqs .
( [ eq - ra - mu1])([eq - ra - dz1 ] ) .
the ratio @xmath122 contributing to these equations is simply the fraction of free carriers in the equilibrium state .
the distribution of the dwell times @xmath123 of a carrier at some individual trap is exponential with the mean value @xmath124 , where @xmath125 is the emission rate .
consequently , the mean square of this dwell time is @xmath126 .
the mean value @xmath121 is a weighted average of values @xmath127 , where the probability of visiting a trap serves as the weight .
therefore , @xmath128 where @xmath129 is the probability of visiting ( at a given `` jump '' ) a trap with an energy in the range @xmath130 $ ] : @xmath131 expressing the escape rates through capture cross - sections : @xmath132 one obtains @xmath133 , where @xmath134 finally , eqs .
( [ eq - ra - mu1])([eq - ra - dz1 ] ) get the following form : @xmath135 where @xmath136 is the equilibrium fraction of free carriers , and @xmath137 is the integral defined by eq .
( [ eq - ra - i ] ) .
the expressions for @xmath4 and @xmath16 are the same as the ones obtained by rudenko and arkhipov.@xcite however , our consideration provides some new information about diffusion in the multiple - trapping conductivity regime .
first , eq .
( [ eq - ra - final ] ) shows that the _ transversal _ diffusion coefficient is field - independent .
second , we have proven that the equations for @xmath4 and @xmath16 are _ exact _ ( provided that emission / capture probabilities are not influenced by electric field ) , whereas in ref .
they were obtained only under so - called `` quasi - equilibrium conditions '' .
now we will apply eq .
( [ eq - d_nonzero_field ] ) to two- and three - dimensional hopping transport in a system with a gaussian dos , in a manner very similar to our consideration of the multiple - trapping conductivity .
our analysis is based on the following hypothesis : _ the field dependence of the diffusion coefficient in 2d and 3d hopping is mainly due to very rare and energetically deep sites .
_ we will call them _ `` traps '' .
_ the traps are rare in two senses : first , the typical distance between the traps is large in comparison to the inhomogeneities of the mobility ; second , the probability of being trapped is small , so that the traps do not affect the carrier mobility .
`` energetically deep '' means that the energies of the traps are far below the mean energy of the carriers
. we will show below that this hypothesis provides a reasonable description of the monte carlo simulation results on the field - dependent diffusion .
as in the previous subsection , we consider the motion of a carrier as a sequence of `` jumps '' , each beginning when the carrier enters a trap , and ending when the carrier enters _ another _ trap .
the durations of different `` jumps '' are not correlated .
the same is true for the carrier displacements .
indeed , the three - dimensional brownian motion is non - returning .
this property guarantees that the carrier always visits a _
trap , and that the trajectories of its motion between different traps do not overlap .
hence there are no reasons for correlations between successive `` jumps '' .
this is the point where dimensionality is important . in one dimension ,
the probability of visiting a previously visited trap is not negligible .
consequently the `` jumps '' must be correlated .
since the `` jumps '' in the 3d case are not correlated , one can obtain the mobility @xmath4 and the diffusion coefficient @xmath5 from the statistics of `` jumps '' using the method of sec .
[ subsec - independent - general ] .
for the 2d case we also use the assumption of independent `` jumps '' . as in the case of multiple - trapping conductivity ,
each `` jump '' consists of two contributions : `` free '' motion of a carrier and sitting on a trap .
let @xmath108 and @xmath109 be the durations of these contributions ; @xmath112 and @xmath113 denoting the mean values @xmath110 and @xmath138 , respectively ( where averaging is over successive `` jumps '' ) ; @xmath103 and @xmath104 be the mobility and the diffusion coefficient of `` free '' motion ( energetically far above the traps ) . with these notations one can follow the same derivation as in the previous subsection and see that eqs .
( [ eq - ra - mu1])([eq - ra - dz1 ] ) are still valid also for hopping transport .
it is convenient to rewrite these equations : @xmath139 where @xmath140 @xmath141 in the regime of ohmic conductivity , the values of @xmath103 , @xmath104 , @xmath112 , @xmath113 and @xmath142 can be regarded as field - independent because a sufficiently small electric field does not significantly perturb the probabilities of capture and release .
therefore one can neglect a possible dependence of @xmath143 , @xmath35 and @xmath34 on the electric field . in the following
we will use the zero - field values for these quantities .
let us calculate the coefficient @xmath34 . for convenience , we will consider the system as a large but finite one ( with periodic boundary conditions , to allow drift ) .
then one can obtain the following expressions for @xmath112 , @xmath113 and @xmath144 : @xmath145 @xmath146 @xmath147 where the index @xmath24 runs over all traps , @xmath148 is the probability that a carrier is at site @xmath24 , and @xmath149 is the rate of escaping from the trap @xmath24 .
( [ eq-3dg - t1 ] ) results from a consideration of the carrier flow from / to traps .
namely , the flow of carriers out of traps is equal to @xmath150 ; the flow into traps is @xmath151 . in equilibrium , these flows are equal to each other , which gives eq .
( [ eq-3dg - t1 ] ) . in order to obtain eqs .
( [ eq-3dg - t2 ] ) , ( [ eq-3dg - t1f2 ] ) , we introduce the probability @xmath152 that trap @xmath24 will be the next visited trap .
let us consider the balance of flows from / to trap @xmath24 .
flow from this trap is equal to @xmath153 .
flow to it is @xmath154 .
therefore @xmath155 the mean time of being captured at trap @xmath24 is @xmath156 . to obtain @xmath113 , we average these mean times with corresponding weights @xmath152 : @xmath157 substituting here eq . (
[ eq-3dg - pt ] ) , one obtains eq .
( [ eq-3dg - t2 ] ) .
analogously , the mean _ square _ of the time of being trapped at site @xmath24 is @xmath158 ( the factor of 2 comes from the exponential distribution of dwell times ) .
again , we take a weighted average to obtain @xmath142 : @xmath159 that , together with eq .
( [ eq-3dg - pt ] ) , provides eq .
( [ eq-3dg - t1f2 ] ) . the substitution of eqs .
( [ eq-3dg - t1])([eq-3dg - t1f2 ] ) into eq .
( [ eq-3dg - a1 ] ) gives the following expression for @xmath34 : @xmath160 let us try to simplify this expression .
the sum in the denominator is the probability that a carrier is at some trap . due to the rarity of traps ,
this probability is small , and it can be neglected .
the summation in the numerator will be taken over _ all _ sites : @xmath161 we will use curly braces to denote the summation in which index @xmath162 runs over all sites : @xmath163 where @xmath15 is any quantity specific for sites .
it is obvious that curly braces mean averaging over an ensemble of particles ( or time averaging , which is the same for finite system and long enough time ) .
thus , @xmath164 then , the rate of escape @xmath165 is related to the sum @xmath166 of transition rates from site @xmath162 to any other sites : @xmath167 where @xmath168 is the mean number of `` attempts '' to escape from site @xmath162 ( including the successful one ) .
these `` attempts '' are events of carrier s hopping out of site @xmath162 before the carrier moves so far from this site that it completely forgets the prehistory related to this site .
let us denote the sum @xmath166 as @xmath169 .
then one can rewrite @xmath34 as @xmath170 or , introducing the averaged number of escape attempts @xmath171 , @xmath172 calculating @xmath173 implies averaging over site energies ( @xmath174 ) and over quantities related to the neighborhood ( energies of neighbors @xmath175 and distances to them @xmath176 ) .
it is possible to treat these two kinds of averaging _ separately _ , using our assumption that `` optimal '' traps are deep in energy .
since the traps are deep , each hop from a trap is upward in energy .
therefore the miller - abrahams rates , eq . ( [ eq - miller - abrahams ] ) , for such hops can be written as @xmath177 the dependence of this expression on @xmath174 has the form of a factor @xmath178 .
hence , the quantity @xmath179 can be factorized ( only for deep sites
@xmath162 ) as @xmath180 where @xmath181 does not depend on @xmath174 : @xmath182^{-1}.\ ] ] since there are no correlations between @xmath183 and @xmath181 , one can average these quantities separately : @xmath184 here @xmath123 is the mean value of @xmath181 ( an arithmetical average over all @xmath162 ) .
now , the calculation of @xmath185 is simple .
remember that @xmath186 is an equilibrium probability of finding a carrier at site @xmath162 : @xmath187 therefore @xmath188 @xmath189 for a gaussian dos @xmath105 given by eq .
( [ eq - gaussian - dos ] ) , one obtains @xmath190 hence @xmath191 substituting this result into eq .
( [ eq-3dg - a2 ] ) , one obtains @xmath192 this equation contains @xmath123 . the proper way to calculate
@xmath123 is the numerical one : to generate at random a large enough number of neighborhoods of the given site ( i. e. sets of energies @xmath175 of neighboring sites and distances @xmath176 from the given site @xmath162 ) , to calculate @xmath181 for each neighborhood according to eq .
( [ eq-3dg - tau - s ] ) , and then to average the results for @xmath181 . in the next section
we compare eq .
( [ eq-3dg - a3 ] ) to the results of monte carlo simulations from sec .
[ sec - simulation ] .
let us now reveal which traps are `` optimal '' in the sense that they determine the diffusion coefficient .
the integrand in the numerator of eq .
( [ eq-3dg - averaging - integrals ] ) has a sharp peak of width @xmath3 at @xmath193 .
hence the `` optimal '' traps are sites with energies in the range @xmath194 .
it is now possible to justify our assumption about the possibility to distinguish between the transport sites and the traps .
we restrict ourselves to the case of low temperatures , @xmath195 .
the typical energy of the `` optimal '' traps , @xmath196 , is significantly lower that the mean energy of carriers @xmath197 ( the difference is much larger than @xmath198 ) .
herewith the assumption that traps are deep in energy is fulfilled .
the probability for a carrier to have an energy lower than @xmath196 is @xmath199 .
this value is much smaller than unity . therefore neglecting the sum in the denominator of eq .
( [ eq-3dg - a1.5 ] ) is justified .
let us now find the typical distance @xmath200 between `` optimal '' traps .
their concentration can be estimated as the concentration of sites with energies lower than @xmath196 , which is about @xmath201 , where @xmath2 is the total concentration of sites .
hence @xmath202 .
we should compare it to the characteristic size @xmath203 of inhomogeneities of the mobility .
one can estimate this size as a typical distance between sites that are important for the mobility .
these are sites with energies close to the mean energy of carriers,@xcite @xmath204 .
their concentration is about @xmath205 .
hence @xmath206 .
herewith we obtain the strong inequality @xmath207 , which implies that it is safe to describe the motion of the carriers between traps by means of a constant mobility @xmath103 and a diffusion coefficient @xmath104 .
all of the above considerations were based on the boltzmann statistics for the charge carriers .
let us discuss the effect of the carrier concentration @xmath23 on the obtained results .
the usage of boltzmann statistics is justified if the fermi level @xmath208 is far below the energies of the sites that make a major contribution to the mean value @xmath173 .
as described above , the main contribution to @xmath173 comes from sites with energies in the vicinity of @xmath196 .
a simple calculation shows that carrier concentration @xmath209 corresponding to a fermi level @xmath210 is @xmath211 where @xmath2 is the concentration of sites .
the theory considered above presumes that @xmath212 . in this limit ,
the diffusion coefficient does not depend on the carrier concentration . for larger concentrations , @xmath213 , sites with energies @xmath214
are essentially occupied . in the latter case
they can not efficiently capture the moving charge carriers and therefore the contribution of such sites to the diffusion process ( in particular to @xmath173 ) is suppressed . according to eq .
( [ eq-3dg - a2 ] ) , the field - induced part of the diffusion coefficient @xmath215 is proportional to @xmath173 .
therefore , in the case @xmath213 , @xmath215 decreases with increasing @xmath23 .
detailed consideration of this concentration - dependent diffusion is beyond the scope of the present paper .
we should also note that if the sample contains less than @xmath216 sites , then the average number of `` optimal '' traps in the sample is less than 1 .
one should therefore expect large sample - to - sample variations of @xmath215 in this case .
in this section we discuss relations between the simulation results described in sec .
[ sec - simulation ] and the analytical expression ( [ eq-3dg - a3 ] ) .
there are three quantities in eq .
( [ eq-3dg - a3 ] ) that are not input parameters of the model : the zero - field mobility @xmath143 , the averaged number of escape attempts @xmath171 , and the value of @xmath123 . for the mobility
, we will use the values taken from our simulations .
there is no need for a special discussion of the mobility , because it is a well - studied property of the transport model considered here.@xcite the value of @xmath123 was obtained numerically , as explained in the previous section with respect to eq .
( [ eq-3dg - a3 ] ) .
the calculation of @xmath123 is relatively inexpensive ( as compared to monte carlo simulations of the diffusion ) and it can be performed over a wide range of model parameters . for the lattice model used in sec .
[ sec - simulation ] , we have found that the simulated dependence of @xmath123 on model parameters is well fitted by the following expressions in two ( @xmath217 ) and three ( @xmath218 ) dimensions : @xmath219 @xmath220 where the coefficients @xmath221 and @xmath222 depend on the ratio of the localization radius @xmath15 to the lattice parameter @xmath19 : @xmath223 the error of fitting does not exceed 2.5% within the range of parameters @xmath224 and @xmath225 .
extensive computer simulations are necessary to determine the exact value of @xmath171 . without simulations one can only claim that @xmath171 is a number larger than unity , though not exponentially large as a function of model parameters @xmath226 and @xmath227 .
indeed , it follows from the concept of transport energy,@xcite that the energy of a carrier just after its hop from a trap is close to the transport energy .
it means that there is no energetic barrier for moving further away from a trap after the first hop .
therefore it is natural to suppose that , when a carrier has jumped out of a trap the probability of escaping i. e. @xmath228 is comparable to 1 . in order to learn more about @xmath171 , we used the values of @xmath143 and @xmath34 obtained by monte carlo simulations . substituting the data shown in fig .
[ fig-3d - tdep ] into eq .
( [ eq-3dg - a3 ] ) , we have extracted @xmath171 for different temperatures in the range @xmath229 . the localization radius @xmath15 was chosen equal to @xmath230 .
we have obtained that @xmath171 varies in the range from @xmath231 to @xmath232 in the 2d system , and in the range from @xmath233 to @xmath234 in the 3d system .
simulations have also shown that the values of @xmath171 do not significantly change when the localization radius is changed from @xmath230 to @xmath235 .
one can see that values of @xmath171 found by simulations are indeed reasonable : they are larger than unity but of that order . therefore , the analytical consideration presented in sec .
[ subsec - hopping - transport ] reduces the problem of predicting the field effect on diffusion to ( i ) the evaluation of the zero - field mobility @xmath143 , and ( ii ) the evaluation of a coefficient @xmath171 , which is a _ slowly varying _ function of the system parameters @xmath226 and @xmath227 .
the mobility can be obtained either from computer simulations ( which is a much easier task than simulations of the diffusion ) , or from the analytical theory@xcite and experiments . for the sake of self - consistency
, we rely here on the simulation data for the mobility . with respect to @xmath171
, the comparison between numerical results and the theory shows that one can consider this coefficient as a _ constant _ and nevertheless obtain results that agree with simulations .
the appropriate choice of this constant for @xmath236 is @xmath237 ( symbols ) with eq .
( [ eq-3dg - a3 ] ) ( lines ) for 2d and 3d transport .
localization length @xmath238 .
values of @xmath123 and @xmath171 are taken from eqs .
( [ eq - tau-2d - fit]),([eq - tau-3d - fit]),([eq - n - esc - choice ] ) . for @xmath143 monte carlo results
are used.,width=8 ] in fig .
[ fig - a_vs_t_02 ] , monte carlo results for the coefficient @xmath34 , which describes the field - induced diffusion according to eq .
( [ eq - quadratic - d ] ) , are compared to the analytical expression ( [ eq-3dg - a3 ] ) .
the coefficient @xmath171 in eq .
( [ eq-3dg - a3 ] ) is set to a constant according to eq .
( [ eq - n - esc - choice ] ) .
it is evident that in the framework of the simplifying assumption of constant @xmath171 , the analytical theory correctly reproduces the shape of the temperature dependence for the field - induced diffusion coefficient .
the main result of this paper is the development of an analytical theory for the field - induced diffusion in the hopping transport mode in 3d and 2d systems with the gaussian dos given by eq .
( [ eq - gaussian - dos ] ) . at low electric fields ,
the field dependence of the longitudinal diffusion coefficient @xmath32 is parabolic as expressed in eq .
( [ eq - quadratic - d ] ) . the analytical expression ( [ eq-3dg - a3 ] )
gives the temperature dependence of the field - induced diffusion .
accompanying monte carlo simulations confirm the analytical results and show that the shape of the field dependence is parabolic . together with the exact results of the previous paper@xcite stating the non - analytic linear field dependence of the diffusion coefficient in the 1d case ,
our result resolves the long - standing puzzle for the reason of the discrepancies between the analytic and non - analytic behavior of @xmath32 claimed in different studies@xcite : it is the space dimensionality that is responsible for non - analytic ( 1d ) and analytic ( 2d and 3d ) dependences in @xmath32 .
furthermore , our theory shows that the main contribution to the field - induced diffusion process comes from localized states with energies in the vicinity of @xmath196 .
the dos parameter @xmath3 in organic semiconductors is of the order of @xmath239 ev.@xcite therefore at room temperatures this energy @xmath196 , which is decisive for @xmath32 , is situated very deep in the tail of the dos , around @xmath240 .
this fact raises very severe demands to the size of the system in computer simulations , which aim at studying the field - induced diffusion in organic semiconductors . in order to have the decisive traps in a simulation at room temperature one
needs approximately @xmath241 sites .
therefore , such simulations can not be considered suitable for studying the field - induced diffusion .
for instance , the system size of @xmath56 used in the previous simulations@xcite is suitable only at @xmath242 .
in contrast , our simulations carried out on systems with @xmath26 sites give reliable results at @xmath55 confirming the developed analytical theory in a wide range of parameters . in all models the influence of the electric field on the mobility and on the diffusion coefficient increases with decreasing temperature . for the mobility this phenomenon has been accounted in the frame of the concept of the effective temperature.@xcite the results of this paper show that for the longitudinal diffusion coefficient the concept of the deep traps should be used instead of the effective temperature . financial support from the academy of finland project 116995 , from the deutsche forschungsgemeinschaft and that of the fonds der chemischen industrie is gratefully acknowledged .
the calculations were done at the facilities of the finnish it center for science , csc .
for a numerical study of the diffusion in hopping transport , an algorithm is needed that can efficiently simulate transport in large systems .
since the number of sites that can be treated in the simulation is limited by the available memory , we have chosen to study hopping transport on a lattice instead of a system with randomly placed sites .
when the charge carrier is located at site @xmath7 , the probability that the next jump takes it to the site @xmath8 is given by @xmath243 where @xmath244 is the total rate of hopping away from site @xmath7 : @xmath245 the time @xmath123 that the charge carrier spends on the site @xmath7 before hopping , ( the `` dwell time '' ) is calculated as @xmath246 where @xmath247 for each hop is randomly generated with an exponential distribution with unit variance . which jump to perform
is decided by picking a random number @xmath248 between 0 and 1 from a uniform distribution , and finding @xmath8 such that @xmath249 this ensures that each site @xmath250 is selected with probability @xmath251 .
so far this is the standard monte carlo algorithm for hopping transport.@xcite below an efficient implementation of this algorithm will be described .
calculating the hopping rates ( [ eq - miller - abrahams ] ) is very time consuming since the exponential function is expensive to compute . if we instead of the site energies @xmath20 store their exponentials , @xmath252 the hopping rates @xmath253 can be computed more efficiently .
we use the fact that only a small number of discrete displacements are possible for hopping in a lattice , when we restrict the length of the hops
. therefore the geometric part of the hopping rate and the energy contribution from the electric field can be calculated and stored once for each displacement . in our case
the hops are restricted to a cube of @xmath22 sites centered at the starting site . for each displacement @xmath254 , define the quantities @xmath255 and @xmath256 where @xmath257 is the @xmath258-component of @xmath254 .
the hopping rate from site @xmath7 to site @xmath8 located at the position @xmath254 relative to @xmath7 can now be evaluated using @xmath259 one could also consider the storing of all hopping rates @xmath253 , but since this greatly increases the amount of memory needed by the simulation , it would restrict the size of the systems that can be simulated .
we have found that a good balance between the simulation speed and the memory requirements is achieved by storing only the rate @xmath244 and the quantity @xmath260 for each site , and to calculate the rates @xmath253 by eq .
( [ eq - mc - fastrate ] ) during the simulation each time they are needed .
one further optimization is to consider the jumps in order of increasing lengths , since shorter jumps typically have higher probabilities than longer ones .
this ordering greatly reduces the number of hopping rates that have to be evaluated before the destination site @xmath8 that satisfies eq .
( [ eq - mc - choise ] ) is found . | in the previous paper [ nenashev et al . , arxiv:0912.3161 ] an analytical theory confirmed by numerical simulations has been developed for the field - dependent hopping diffusion coefficient @xmath0 in one - dimensional systems with gaussian disorder .
the main result of that paper is the linear , non - analytic field dependence of the diffusion coefficient at low electric fields . in the current paper , an analytical theory is developed for the field - dependent diffusion coefficient in three- and two - dimensional gaussian disordered systems in the hopping transport regime .
the theory predicts a smooth parabolic field dependence for the diffusion coefficient at low fields .
the result is supported by monte carlo computer simulations . in spite of the smooth field dependences for the mobility and for the longitudinal diffusivity , the traditional einstein form of the relation between these transport coefficients
is shown to be violated even at very low electric fields . |
study of the stability of charged metal clusters against fragmentation is important in nano - sized systems . in this work ,
we have studied the binary fragmentation of @xmath1-atom @xmath0-ply charged @xmath2 clusters in the framework of the liquid - drop @xcite model ( ldm ) with stabilized @xcite jellium model ( sjm ) coefficients for the energies . here ,
@xmath0=1 , 2 , 3 , 4 ; and @xmath3 covers the species al , ga , li , na , k , and cs . here
, in the case of al , when @xmath1 is a multiple of 3 , it corresponds to real al@xmath4 cluster with @xmath5 .
the clusters are assumed to be spherical with sizes determined by @xmath6 with @xmath7 .
the quantity @xmath8 is the wigner - seitz radius of the electrons in the bulk metal .
the possible decay channels of @xmath2 are @xmath9 here , @xmath10 is positive integer as @xmath1 , and @xmath11 $ ] .
different processes are defined by different values of @xmath12 .
the process in which @xmath13 ( one of the fragments is neutral ) , is called evaporation . the processes in which @xmath14 are called charge - symmetric fission , and the charge - asymmetric fission is defined by @xmath15 $ ] . for any given set of values @xmath1 , @xmath0 , @xmath12 , different values of @xmath10 define different decay channels . in a particular decay channel of the evaporation processes , the negativity of the difference between total energies before and after the fragmentation , @xmath16 is sufficient for the decaying in that particular channel . in the above equation ,
@xmath17 and @xmath18 are the total energies of @xmath0-ply ionized and neutral @xmath1-atom clusters , respectively .
however , in fission processes a negative value for the difference energy is not a sufficient condition for the decay of the parent cluster .
this is because , the competition between the short - range surface tension and the long - range repulsive coulomb force may give rise to a fission barrier . the heights of the fission barriers
are calculated using the two - spheres approximation@xcite . in fig .
[ fig1 ] , the fission of a @xmath0-ply charged @xmath1-atom cluster into two clusters of respective sizes @xmath19 , @xmath20 and respective charges @xmath12 , @xmath21 is schematically shown .
@xmath22 is the energy release , @xmath23 is the fusion barrier ( or coulomb barrier ) which is the maximum energy of the coulomb interaction of two positively - charged conducting spheres , taking their polarizabilities into account .
@xmath24 is the fission barrier height which is defined as @xmath25 the coulomb interaction energy , @xmath26 , as a function of their separations , @xmath27 , for two charged metallic spheres can be numerically calculated using the classical method of image charges @xcite .
the most favored decay channel in evaporation processes is defined as the channel for which the dissociation energy assumes its minimum value @xmath28 and the most favored decay channel in fission processes is defined as the channel for which the fission - barrier height assumes its minimum value , @xmath29 the energy of an @xmath1-atom @xmath0-ply charged cluster in the ldm is given @xcite by @xmath30 in which @xmath31 , @xmath32 , @xmath33 , and @xmath34 are the work function of the bulk metal , the finite - size correction to the work function , the radius of the cluster , and the electron charge , respectively . for simplicity , the position of the centroid of excess charge , @xmath35 , is neglected in our calculations .
@xmath18 is the energy of a neutral @xmath1-atom cluster in the ldm , which is given by @xmath36 in eq .
( [ eq7 ] ) , the quantities @xmath37 , @xmath38 , and @xmath39 are total energy per electron of the bulk , surface energy , and curvature energy , respectively .
we have calculated the quantities @xmath32 , @xmath31 , @xmath38 , and @xmath39 by fitting to the self - consistent kohn - sham @xcite results in the sjm for different @xmath40 values @xcite .
however , @xmath37 is calculated using the sjm energy expression for the bulk system ( eq .
( 1 ) of ref . [
[ payamijpc01 ] ] ) .
using the method of image charges , we have calculated the coulomb interaction energy of two charged metallic spheres , taking their polarizabilities into account .
the calculations show that the maximum of the interaction energy , @xmath23 , is achieved for a separation @xmath41 .
[ fig2](a ) shows the coulomb interaction energy of an @xmath19-atom cluster with another @xmath42-atom of respective excess charges @xmath43 and @xmath44=1 .
the radii of the clusters are calculated from @xmath6 , as in the sjm . in this figure , we have taken @xmath8=2.07 which correspond to bulk al . when both @xmath19 and @xmath42 are multiples of 3 , the results correspond to real al clusters .
the value at the maximum specifies the quantity @xmath23 , and the position of the maximum , @xmath45 , is the separation between the centers of the two spheres . as is seen , for fixed values of charges
, @xmath23 is the highest when the sizes are equal . in fig .
[ fig2](b ) , the situation is shown for equal charges @xmath46 . in this case , when the sizes are equal , @xmath23 is maximum as before but here , @xmath47 .
an other feature shown in fig .
[ fig2](c ) is that , when both charges and sizes are asymmetric , @xmath23 is higher if the smaller charge corresponds to the smaller cluster . in fig .
[ fig2](d ) , we have compared the coulomb interaction energies of two al metallic spheres with respective sizes @xmath48 and @xmath49 for different charges . it is seen that keeping the sizes fixed but increasing the charge on any one of them , without changing the charge on the other one , increases the height of the maximum . in our calculations
we need the values of the coulomb interaction energy at the maxima points , i.e. , the @xmath23s . in fig .
[ fig3](a ) , we have compared the @xmath23 values for equally charged @xmath50 , al metallic particles but with different radii @xmath51 and @xmath52 .
this figure shows that when both of the sizes are small , the barrier is higher than the case when at least one of them is larger . in fig .
[ fig3](b ) , we have compared the @xmath23s for singly ionized pairs al , na , and cs with different values of @xmath19 and @xmath42 . in the aluminum case , since the radii of the spheres are smaller than those of na and so on , their polarizabilities are thereby smaller and thus the al values lie above that of na , and so on .
now , if the charge on one of the clusters is increased without changing that of the other one , the arrangement of the curves would not change but they shift upward according to fig . [ fig2](d ) . in the energy calculations of eq .
( [ eq6 ] ) we use the results obtained for @xmath32 and @xmath31 in ref . [ [ payami_aw ] ] .
the values of @xmath37 are calculated using eq .
( 1 ) of ref . [
[ payamijpc01 ] ] for unpolarized case . @xmath38 and @xmath39 in eq .
( [ eq7 ] ) are obtained by fitting the results of the self - consistent solutions of the kohn - sham equations in the sjm for neutral clusters of sizes @xmath7 , and different @xmath40 values .
the values of @xmath8 for al , ga , li , na , k , and cs are 2.07 , 2.19 , 3.28 , 3.99 , 4.96 , and 5.63 , respectively . in figs .
[ fig4](a)-(c ) we have plotted the sjm values of @xmath37 , @xmath38 , and @xmath39 , as functions of @xmath40 , respectively . to show the variations of the dissociation energy , @xmath53 , for different charging and different evaporation channels , we have plotted , in figs . [ fig5](a)-(b ) , the quantities @xmath54 and @xmath55 , respectively , as functions of the neutral fragment size , @xmath10 , for li . as is seen in fig .
[ fig5](a ) , for small @xmath10 values @xmath54 has an increasing behavior , then after passing a maximum , it changes to a decreasing behavior , and finally , for large @xmath10 values it increases again .
this behavior has its roots in the trade off between the surface energy ( the volume energy term does not change in the fragmentation process ) and the last term of eq .
( [ eq6 ] ) . for small @xmath10 , the surface area difference @xmath56,\ ] ] increases , and then for intermediate @xmath10 values , it assumes a maximum , and finally decreases .
the increasing behavior at the tail is because , with increasing @xmath10 , the size of the singly ionized fragment decreases and hence , beyond a certain @xmath10 value , the last term in eq .
( [ eq6 ] ) dominates the surface energy term .
however , beyond a @xmath0 value for the parent cluster , the last term of the eq .
( [ eq6 ] ) dominates the surface energy term for smaller @xmath10 values before @xmath53 reaches its maximum as in the case of fig .
[ fig5](b ) .
in fact , we have found that for li@xmath57 this transition takes place at a point @xmath58 .
the value of @xmath59 depends on @xmath1 .
this dependence is clearly shown in fig .
[ fig5](a ) by the fact that @xmath60 and @xmath61 are increasing functions of p in the whole range .
the dissociation energy @xmath53 for small @xmath62 takes the form @xmath63 which specifies the most favored channel in an evaporation process with @xmath64 .
this means that , in evaporation processes , the parent charged cluster prefers to evaporate single atoms than larger neutral clusters .
this fact is also numerically shown in figs . [ fig5](a)-(b ) .
on the other hand , the positivity of @xmath53 in the whole range implies that the parent charged cluster is stable against any spontaneous evaporation . in fig .
[ fig6 ] , we have plotted the most favored dissociation energies , @xmath65 , of na clusters as functions of the parent size , @xmath1 , for different @xmath0 values .
the results show that for large enough clusters , the dissociation energy is independent of @xmath0 , which is consistent with eq .
( [ eq9 ] ) . however , for small clusters , the dissociation energy increases with charge .
figures [ fig7](a)-(b ) compares the most favored dissociation energies of different species for @xmath66 and @xmath67 , respectively .
it is seen that detachment of a single atom from a lower electron density cluster is easier , i.e. , it needs less energy . comparing these two figures
also shows that , although the hierarchy is the same , the dissociation energies are shifted to higher energies as we go from @xmath66 to @xmath67 .
now , we consider the fission processes .
we first consider the following two cases : @xmath68 @xmath69 processes ( [ eq10 ] ) and ( [ eq11 ] ) describe charge - symmetric and charge - asymmetric fissions , respectively . in figs .
[ fig8](a ) and [ fig8](b ) we have plotted their respective fission barriers , @xmath70 , for @xmath71 as functions of @xmath10 , for different sizes of the parent . as is seen in both of these figures
, there exists a minimum size for the parent , @xmath72 , beyond which the fission barrier height is positive in all possible channels of that particular process . that is , the parent cluster @xmath73 with @xmath74 is stable against any spontaneous fission via that particular process .
one of the significant differences in the charge - symmetric and charge - asymmetric fission is that [ as seen in figs . [ fig8](a ) and [ fig8](b ) ] , in the most favored channel , the products have more or less the same sizes in the former case ; whereas , in the latter case the less charged fragment assumes quite smaller sizes .
that is , the charge - symmetric fission proceeds mostly via mass - symmetric channel , and the charge - asymmetric fission does it via mass - asymmetric channel . in fig .
[ fig8](c ) , we compare the most favored sizes , @xmath75 , for the charge - symmetric and charge - asymmetric fission of na@xmath76 .
it is seen that in the charge - symmetric case , @xmath75 increases with @xmath1 whereas , in the charge - asymmetric case it remains more or less constant . to decide whether a cluster with a given size and charge , @xmath77 , is stable against any spontaneous decay , one should compare @xmath1 with the quantity @xmath78 if @xmath79 , then the cluster @xmath73 is stable against any spontaneous decay . to show the values of @xmath72 schematically
, we plot the most favored fission barriers , @xmath80 for all possible @xmath12 values .
the evaporation processes need not be considered because , as was shown before , in the ldm there is no spontaneous decay via evaporation .
however , in order to determine at what size which process is dominant , we include the most favored evaporation process as well .
it should be mentioned that , in the self - consistent kohn - sham results @xcite , because of the shell effects , the spontaneous evaporation is also possible . in fig .
[ fig9 ] , we have compared the most favored fission barriers , @xmath81 , of ga@xmath76 for @xmath82 and @xmath83 with the most favored dissociation energy , @xmath84 .
we have labeled the intersection points of the curves with each other and with the horizontal @xmath1 axis by a , b , c , d , and e. at point a , the charge - symmetric and charge - asymmetric fission start their competition ; thus we name the corresponding size as @xmath85 . for clusters
@xmath86 with @xmath87 , the charge - symmetric fission is dominant whereas for larger sizes charge - asymmetric one dominates .
we name the size corresponding to the point b as @xmath88 because for sizes greater than this there is no charge - symmetric spontaneous fission .
however , spontaneous fission via charge - asymmetric process still persists for @xmath89 . beyond the size @xmath90 which corresponds to the point c
, the spontaneous charge - asymmetric fission also stops .
our results show that , for small @xmath40 values , @xmath91 whereas , for @xmath92 , the reverse inequality is at work ( as in the cases of k and cs ) . according to eq .
( [ eq12 ] ) , if @xmath93 , then any spontaneous fission stops and the cluster @xmath86 would be stable against any spontaneous decay . at sizes
@xmath94 and @xmath95 which correspond , respectively , to the points d and e , competition of the evaporation with charge - symmetric and charge - asymmetric fission starts , respectively . beyond the size @xmath96 the dominant decay process is evaporation .
our results show that , at least for @xmath97 , the inequality @xmath98 is satisfied .
when @xmath99 , the decay processes consist of evaporation and charge - asymmetric fission whereas , for @xmath100 , the fission is a charge - symmetric one . in figs .
[ fig10](a ) and ( b ) , we compare the most favored fission barrier heights @xmath101 and @xmath102 for different species , respectively .
the figures indicate that , for large enough clusters , the fission barriers are higher for higher electron density metals . in fig .
[ fig11](a ) , we compare the smallest stable sizes , @xmath103 for different species .
looking at the figure , one concludes that , for fixed @xmath1 , the higher electron density metal clusters ( say al ) has less capacity for charging than the lower density metal clusters ( say cs ) .
this behavior is observed for all @xmath104 . to show the size beyond which evaporation dominates all decay processes , we plot @xmath105 with @xmath106 for all species in fig . [ fig11](b ) .
it is seen that , for fixed charge , @xmath0 , the evaporation process dominates at smaller @xmath1 values for lower electron density metal clusters . the trends in figs .
[ fig11](a ) and [ fig11](b ) are seen to be the same . finally , in fig .
[ fig11](c ) , we compare the @xmath85 , the size at which the competition between the charge - symmetric and charge - asymmetric fission starts .
this size decreases from al to li , and then increases from li to k , and finally decreases from k to cs . in the plot
, we have taken into account the decimals for the intersection points which can be rounded to the next nearest integer .
in this work , we have studied the stability of @xmath0-ply charged ( @xmath107 ) metal clusters of species al , ga , li , na , k , and cs using the liquid - drop model with stabilized jellium model energies .
fragmentation of clusters into smaller ones can proceed via evaporation or fission .
our results show that , in the ldm , any charged cluster is stable against spontaneous evaporation , and the most favored channel of induced fragmentation depends on the size and the charge of the parent cluster .
sufficiently small multiply charged clusters , however , may undergo spontaneous fragmentation via different fission processes .
we have obtained , for each species , the smallest size that a @xmath0-ply charged cluster is stable against any spontaneous fragmentation .
we have also shown that , for sufficiently large clusters the most favored decay channel is atomic evaporation . comparing
the results for different species show that , for fixed @xmath1 , the lower electron density clusters have higher capacity for charging .
results also show that , for fixed amount of charge , atomic evaporation dominates at lower @xmath1s for lower electron density metal clusters .
m. brack , rev .
phys . * 65 * , 677 ( 1993 ) .
[ brack93 ] j. p. perdew , h. q. tran , and e. d. smith , phys .
b * 42 * , 11627 ( 1990 ) .
[ pertran ] u. nher , s. bjrnholm , s. frauendorf , f. garcias , and c. guet , phys . rep . * 285 * , 245 ( 1997 ) and references therein .
[ naher ] m. seidl , j. p. perdew , m. brajczewska , and c. fiolhais , phys .
b * 55 * , 13288 ( 1997 ) .
[ seidl97 ] m. payami , arxiv.org : cond - mat/0212160 .
[ payami_aw ] w. kohn and l. j. sham , phys
. rev . * 140 * , a1133 ( 1965 ) .
[ kohnsham ] m. payami , j. phys . : condens .
matter * 13 * , 4129 ( 2001 ) .
[ payamijpc01 ] m. payami , arxiv.org : cond - mat/0305600 .
[ payami_rigrlx ] | in this work , we have used the liquid - drop model in the context of stabilized jellium model , to study the stability of @xmath0-ply charged metal clusters of different species against fragmentation .
we have shown that on the one hand , singly ionized clusters are stable against any spontaneous fragmentation , and on the other hand , the most favored decay process for them is atomic evaporation .
however , multiply charged clusters of sufficiently small sizes may undergo spontaneous decay via fission processes . comparing
the results for different species show that for fixed @xmath1 , the lower electron density metal clusters can accommodate more excess charges before their coulomb explosions .
this comparison also shows that , for fixed @xmath0 , the atomic evaporation which is the most favored decay mechanism for sufficiently large clusters , takes place at lower @xmath1s for lower electron density clusters . _
center for theoretical physics and mathematics , atomic energy organization of iran , + p. o. box 11365 - 8486 , tehran , iran _ |
in the last few years the casimir effect @xcite , which is a rare macroscopic manifestation of the boundary dependence of the quantum vacuum , has attracted much experimental and theoretical attention ( see monographs @xcite and reviews @xcite ) .
the spectrum of the electromagnetic zero - point oscillations depends on the presence of material bodies .
in particular , the tangential component of the electric field vanishes on the surfaces of two parallel plates made of ideal metal ( it is small if real metals are used ) .
this leads to changes in the zero - point oscillation spectrum compared to the case of free unbounded space and results in the attractive casimir force acting normal to the surfaces of the plates .
the casimir effect finds many applications in quantum field theory , condensed matter physics , elementary particle physics , gravitation and cosmology @xcite .
recently many measurements of the casimir force have been performed @xcite .
their results have already been applied in nanotechnology for the actuation of the novel microelectromechanical devices , based entirely on the modification of the properties of quantum vacuum @xcite , and for constraining predictions of extra - dimensional physics with low compactification scales @xcite .
most theoretical papers on the casimir effect deal with idealized boundary conditions and perfectly shaped test bodies . over the last 4 decades
only a few have considered the corrections to the casimir force such as due to the finite conductivity of the boundary metal @xcite , distortions of the surface shape @xcite and nonzero temperature @xcite .
comparison of the theory with the results of modern casimir force measurements demands careful treatment of all these corrections .
both the individual corrections and their combined effect has to be evaluated ( see ref .
@xcite for review ) .
the quantification of errors and precision in the measurements and theoretical computations of the casimir force is crucial for using the casimir effect as a new test for extra - dimensional physics and other extensions to the standard model .
nevertheless , there is no general agreement on the achieved levels of experimental precision and the extent of agreement between theory and experiment . in the literature a variety of measures to characterize the experimental precision is used and the extent of agreement between measurements and theory ranges from 1% @xcite to 15% @xcite depending on the measurement scheme and configuration .
very often , the confidence levels and numerous background effects which may contribute to the theoretical results are not considered . in the present paper
we perform a reanalysis of the experimental data on the casimir force measurements between au surfaces @xcite and make a comparison with theory . in doing
so we carefully calculate the original experimental precision without relation to the theory , including the random absolute error at a 95% confidence level , and the absolute systematic error .
the total absolute error of these casimir force measurements in the experiment of ref .
@xcite is found to be equal to @xmath0pn at 95% confidence .
this corresponds to approximately 1.75% precision at the closest separation @xmath1 nm ( the 1% precision at the closest separation indicated in ref @xcite is obtained at 60% confidence ) . as a second step
, the accuracy of the theoretical computations of the casimir force for the experimental configuration @xcite is determined .
special attention is paid to the possible sample - dependent variations of the optical tabulated data due to the presence of grains , contribution of the surface plasmons , and errors introduced by the use of the proximity force theorem .
the influence of the surface roughness is carefully investigated including the nonmultiplicative contributions and recently discussed diffraction - type effects @xcite .
the contribution of electric forces due to patch potentials resulting from the polycrystalline nature of the au film is calculated for the experimental configuration @xcite at different separations . the finite size and thermal effects
are also considered and found negligible in the experimental configuration of ref .
the conclusion reached is that at the present state of our knowledge the accuracies of theoretical computations in application to the experimental configuration of ref .
@xcite are achievable on the level of 1.69% at a separation @xmath2 nm and 1.1% at a separation @xmath3 nm .
the paper is organized as follows . in sec .
ii the experimental precision of the casimir force measurements at different confidence levels is determined .
iii is devoted to the computations of the casimir force with account of finite conductivity and grain structure of the metal layers .
the role of roughness including the nonmultiplicative and diffraction - type effects is studied in sec .
iv . in sec . v both traditional and alternative thermal corrections are discussed .
also the possible role of the electric forces due to the patch potentials and finite size effects are estimated .
vi contains the final numbers on theoretical accuracy and the comparison of theory with experiment in the casimir force measurement between two gold surfaces by means of an atomic force microscope @xcite . in sec .
vii the final conclusions and some discussion are provided .
in ref . @xcite precision measurements of the casimir force between gold coated bodies , a plane plate and a sphere , were performed using an atomic force microscope .
the casimir force was measured by averaging 30 scans over a surface separation region between 62350 nm with 2583 points each ( see ref .
@xcite for all the details of the measurement procedure ) . in the analysis below we neglect data from 3 scans due to excessive noise and
use the data from the rest @xmath4 scans to find the quantitative characteristics of the experimental precision in the casimir force measurements at different confidence levels .
we start with the random error and calculate the mean values of the measured force at different separations @xmath5 within the region from 62 nm to 350 nm @xmath6 an estimate for the variance of this mean is determined by @xcite @xmath7 ^ 2 .
\label{eq2}\ ] ] calculations using the measurement data @xmath8 show that @xmath9 do not depend sensitively on @xmath5 .
the largest value @xmath10pn is taken below as an estimate for the variance of the mean force within the whole measurement range . according to student s test for the truth of a hypothesis @xcite ,
if the inequality @xmath11 is fulfilled , the hypothesis that @xmath12 is the true value of the casimir force at a separation @xmath13 must be rejected at a given confidence level @xmath14 ( this is a two - tailed test as the deviations @xmath12 from @xmath15 in two directions are possible ) .
equivalently , if the inequality @xmath16 is fulfilled , the hypothesis that @xmath12 is the true value of the casimir force should be accepted at a confidence level @xmath17 . usually in the tables for student
s @xmath18-distribution ( see , e.g. , refs .
@xcite ) the values of @xmath19 are presented , where @xmath20 , @xmath21 is the number of degrees of freedom , and @xmath22 is the number of measurements ( @xmath4 in our case ) .
choosing @xmath23 ( hypothesis is true at 95% confidence ) we obtain @xmath24 and find from tables @xmath25 @xcite . then from eq .
( [ eq4 ] ) it follows @xmath26 where @xmath27 is the random absolute error of the casimir force measurements . if we consider @xmath28 ( hypothesis is true at 60% confidence ) , then @xmath29 and @xmath30 . in this case
the random absolute error of the casimir force measurements is @xmath31pn@xmath32pn .
note that if one would like to have @xmath33 or @xmath34 ( i.e. deviations of the true force value on either side of the mean not greater than one or two @xmath35 ) , the confidence levels of @xmath36 or @xmath37 , respectively , should be chosen for the number of measurements @xmath4 .
now let us consider the systematic error .
the main contributions to the systematic error in the experiment of ref .
@xcite are given by the error in force calibration @xmath38pn , by the noise when the calibration voltage is applied to the cantilever @xmath39pn , by the instrumental sensitivity @xmath40pn , and by the restrictions on computer resolution of data @xmath41pn .
the maximal value of the systematic error is given by @xmath42 finally , the maximum total absolute error of the casimir force measurements in the experiment of ref .
@xcite is equal to @xmath43 at 95% confidence ( to be conservative , the errors are added linearly rather than quadratically ) . at 60% confidence
the total absolute error of the casimir force measurements is @xmath44pn .
these absolute errors with their confidence levels are valid within the whole measurement range from 62 nm to 350 nm . from eq .
( [ eq5 ] ) it follows that the true value of the casimir force belongs to the confidence interval @xmath45 with a chosen confidence probability .
another important characteristic of the experimental precision is the relative error of the casimir force measurements @xmath46 which is evidently separation - dependent . at the shortest separation @xmath2 nm
the value of the mean force is @xmath47pn which leads to a relative error of @xmath48% computed at 95% confidence .
if we restrict ourselves with a 60% confidence , the relative error of the casimir force measurements at the shortest separation @xmath49% is obtained as was indicated in ref .
@xcite without the detailed analysis of the confidence levels .
if we choose 95% confidence , the relative errors of the casimir force measurements at separations 70 nm , 100 nm , and 200 nm are , respectively , 2.46% , 5.9% , and 37.3% . at 60% confidence
the relative errors of the casimir force measurements at the same separations are 1.47% , 3.5% , and 22.4% , respectively .
for the configuration of a large sphere of a radius @xmath50 above a plate the casimir force can be obtained by means of the lifshitz formula , derived originally for two parallel plates @xcite , along with use of the proximity force theorem @xcite @xmath51\right .
\nonumber \\ & & \phantom{aaaaaaaaaaa } \left .
+ \ln\left[1-r_{\bot}^2(\xi , k_{\bot})e^{-2zq}\right]\right\}. \label{eq9}\end{aligned}\ ] ] here the reflection coefficients for two independent polarizations are given by @xmath52 ^ 2 , \qquad r_{\bot}^2(\xi , k_{\bot})=\left(\frac{q - k}{q+k}\right)^2 , \nonumber \\ & & q^2\equiv k_{\bot}^2+\frac{\xi^2}{c^2 } , \qquad k^2\equiv k_{\bot}^2+\varepsilon(i\xi)\frac{\xi^2}{c^2 } , \label{eq10}\end{aligned}\ ] ] @xmath53 is the dielectric permittivity of the gold layers on the sphere and the plate , and @xmath13 is the closest separation distance between them . the thickness of gold coatings , used in ref .
@xcite ( 86.6 nm ) , is much greater than the skin depth of the electromagnetic oscillations for all frequencies which make a significant contribution to the computation of the casimir force .
this allows one to use the properties of the bulk gold in all computations of the casimir force . the accuracy of eq .
( [ eq9 ] ) is restricted by the accuracy of the proximity force theorem , which is , however , very high for the experimental parameters of ref .
the error , introduced by the proximity force theorem , is less than @xmath54 @xcite . taking into account the large value of sphere radius @xmath55 m , used in ref .
@xcite , the upper limit of this error is 0.06% at the shortest separation @xmath2 nm and 0.2% at separation @xmath3 nm ( note that in ref .
@xcite the casimir force for the configuration of a sphere above a plate was precisely computed on the basis of the first physical principles which makes it quite reliable as a test of the proximity force theorem ) . in refs . @xcite the computations of the casimir force were performed using eqs .
( [ eq9 ] ) , ( [ eq10 ] ) and optical tabulated data for gold @xcite ( note that the transition coefficient from energies to frequencies is given by 1ev@xmath56rad / s ) . the imaginary part of the dielectric permittivity , obtained using the complex refractive index from the tables @xcite , was used to compute the dielectric permittivity along the imaginary frequency axis by means of the dispersion relation . at @xmath57rad
/ s , where the tabulated data are not immediately avaliable , they were usually obtained ( see , e.g. , @xcite ) by the extension from the region of higher frequencies by means of the drude dielectric function @xmath58 where the plasma frequency for au is @xmath59rad / s , and @xmath60rad / s is the relaxation parameter describing the non - elastic electron - phonon interaction ( note that in the frequency region under consideration @xmath61 ) .
this procedure was used to calculate the casimir force including the effect of finite conductivity corrections of gold ( see a few examples of the calculations in sec .
iv , table ii and comparison between experiment and theory in sec .
vi ) . later in this section we discuss the influence of possible sample to sample variations of the optical tabulated data on the values of the casimir force and the applicability region of eqs .
( [ eq9 ] ) , ( [ eq10 ] ) involving the dielectric permittivity depending only on frequency .
first , we would like to note that in the separation region @xmath62 nm the computational results obtained by eq .
( [ eq9 ] ) combined with the optical tabulated data , are almost exactly those obtained by the substitution into eq .
( [ eq9 ] ) of the plasma dielectric function for the metal @xmath63 in fact , both computations lead to results differing by less than 0.5% within the mentioned separation region .
what this means is that the real part of @xmath64 depending on only @xmath65 determines the total value of the casimir force in this region .
the value of @xmath66 , where @xmath67 is the density of conduction electrons , @xmath68 is their effective mass , is determined by the preperties of the elementary cell
. it can not be influenced by properties of sample such as the crystallite grain size or the presence of a small concentration of impurities .
this is the reason why the sample to sample variations of the optical tabulated data can not influence the value of the casimir force ( [ eq9 ] ) at separations @xmath69 nm . in the separation region
@xmath70 nm there are significant deviations depending on whether the casimir force ( [ eq9 ] ) is calculated using the optical tabulated data or by use of the plasma dielectric function ( [ eq12 ] ) .
in fact , in this separation region the small imaginary part of @xmath64 is influential and should be taken into account .
there is enough tabulated data in the optical tables to compute the casimir force , so that it is not necessary to use any extension of data .
note that the characteristic frequency corresponding to the largest separation @xmath3 nm is @xmath71rad / s ( i.e. tabulated data for frequencies several times smaller are available ) . at the same time , the characteristic frequency corresponding to the shortest separation is @xmath72 , so that the region under consideration belongs to that of infrared optics @xcite . within the region ( 62200)nm
one may expect some small dependence of the optical tabulated data on the size of the grain , presence of impurities etc . if this is indeed the case , the use of the tabulated data , which are not relevant to the particular samples used in experiment , might lead to the errors in computation of the casimir force ( [ eq9 ] )
. to investigate this possibility , we consider the pure imaginary part of the dielectric permittivity in the region of the infrared optics given by @xcite @xmath73 where @xmath74 is the relaxation parameter at high frequencies in the region of infrared optics ( note that it does not coincide with the relaxation parameter @xmath75 of the drude model ( [ eq11 ] ) which describes the volume relaxation in the region of the normal skin effect ) .
the value of @xmath74 is determined by the processes of elastic scattering of the electrons on impurities , on the boundary surfaces of the metal and of the individual grains , and on other electrons @xcite .
the scattering of electrons on phonons also contributes to the value of @xmath74 .
however , the frequency of the electromagnetic field is so high that @xmath76 , where @xmath77 is the debye temperature , so the frequency of the electron - phonon collisions is the same as it is at @xmath78 @xcite .
it is important to note that of all the above processes , only the contribution of the electron - electron collisions to @xmath74 is frequency - dependent ( and increases as @xmath79 ) .
the main sample to sample dependence of the parameter @xmath74 is determined by the sizes of grains and the density of impurities .
to calculate this dependence we use the following formula for the relaxation parameter in the region of infrared optics @xcite @xmath80 this formula leads to an approximate representation of the dielectric permittivity of au along the imaginary frequency axis given by @xmath81 where @xmath82 , @xmath83 .
it is easily seen , that the substitution of eq .
( [ eq15 ] ) into eq .
( [ eq9 ] ) leads approximately to the same result as the use of the optical tabulated data .
the errors due to use of eq .
( [ eq15 ] ) in eq .
( [ eq9 ] ) instead of the optical tabulated data at separations 62 nm , 70 nm , 100 nm , and 150 nm , are 0.45% , 0.23% , 0.09% , and 0.04% , respectively .
( [ eq15 ] ) gives the possibility to estimate the influence of the sizes of grains in the polycrystalline metal film in the experiment of ref .
@xcite on the value of the casimir force ( [ eq9 ] ) . for this purpose , the experimental data of ref .
@xcite are used where the reflectance @xmath50 of au films is measured as a function of the characteristic sizes of the grains .
the analysis of the atomic force microscopy images ( like the one in fig . 1 but on @xmath84 area ) shows that the mean size of grains in ref .
@xcite is about 90 nm ( the sizes of the typical grains are 77 nm , 103 nm , 94 nm , 68 nm , 88 nm , 121 nm etc ) . according to ref .
@xcite , the largest deviations of the reflectance from the one given by the tabulated data @xcite , takes place at shorter wavelengths .
the shortest separation of @xmath2 nm in the experiment @xcite corresponds to the characteristic wavelength @xmath85 nm . for the films containing grains of 45 nm size ( the largest ones studied in ref .
@xcite ) the reflectance at @xmath86 nm is 0.8% less than the one calculated from the tabulated data ( notice that for smaller grains the difference of the reflectance obtained from the tabulated data is greater ) .
taking into account that the reflectance in the region of the infrared optics is given by @xcite @xmath87 we find that the new value for the coefficient @xmath88 in eqs .
( [ eq14 ] ) , ( [ eq15 ] ) due to grains of 45 nm size is @xmath89 . substituting the approximate eq .
( [ eq15 ] ) ( with @xmath90 instead of @xmath88 ) into eq .
( [ eq9 ] ) , one finds the values of the correction factor to the casimir force @xmath91 at @xmath2 nm and @xmath92 at @xmath93 nm . comparing this with the results of the same approximate computations using @xmath88 ( @xmath94 , respectively , @xmath95 )
, one can conclude that the grains of 45 nm size lead to less than 0.5% decrease of the casimir force magnitude .
note that this is in fact the upper bound for the influence of crystallite grain size on the casimir force in the experiment of ref .
@xcite , as the actual sizes of grains in @xcite were two times greater than 45 nm .
the above calculations of the casimir force including the effect of the real properties of au films were performed on the basis of the lifshitz formula ( [ eq9 ] ) , which does not take into account the effects of spatial nonlocality ( wavevector dependence of the dielectric permittivity ) .
these effects may influence the casimir force value in the region of the anomalous skin effect which is important for large separations @xmath96 m @xcite , a region not relevant to the experiment of ref .
another separation region , where nonlocality may lead to important contributions to the van der waals force , is @xmath97 nm ( @xmath98 is the plasma wavelength ) which corresponds to @xmath99 @xcite .
such high characteristic frequencies lead to the propagation of surface plasmons .
the effect of the surface plasmons , however , does not contribute in the experiment of ref .
@xcite as the largest characteristic frequency there , calculated at @xmath2 nm , is 5.7 times less than @xmath65 [ notice that the frequency region ( @xmath100 ) contributes only 0.19% of the casimir force value at separation @xmath2 nm ] .
the contribution of the surface plasmon for au of about 2% at a separation @xmath101 nm , obtained in recent ref .
@xcite , is explained by the use in @xcite of the spatially nonlocal dielectric permittivity in the frequency region of infrared optics where it is in fact local @xcite ( we would like to point out that at the separation @xmath102 the characteristic frequency of the casimir effect is equal not to @xmath65 , as one might expect , but @xmath103 . as a result
, the surface plasmons do not give any contribution to the casimir force in the experimental configuration of ref .
it is well known that surface roughness corrections may play an important role in casimir force calculations at separations less than 1@xmath104 m @xcite . at the shortest separations ,
the roughness correction contributes 20% of the measured force in experiments of refs .
@xcite . in the experiment of ref .
@xcite , however , the roughness amplitude was decreased and the roughness contribution was made less than 1% of the measured force even at shortest separations . to obtain this conclusion the simple stochastic model for the surface roughness and the multiplicative approach to take into account different corrections were used . here
we obtain more exact results for the contribution of surface roughness to the casimir force taking into account both nonmultiplicative and correlation effects .
the topography of the au coatings on the plate and sphere was investigated using an atomic force microscope .
a typical 3-dimensional image resulting from the surface scan of @xmath105 m area is shown in fig . 1 .
as seen in this figure , the roughness is mostly represented by the stochastically distributed distortions with the typical heights of about 24 nm , and rare pointlike peaks with the heights up to 16 nm . in table i the fractions @xmath106 of the surface area , shown in fig . 1 , with heights
@xmath107 are presented ( @xmath108 ) .
these data allow one to determine the zero roughness level @xmath109 relative to which the mean value of the function , describing roughness , is zero ( note that separations between different bodies in the casimir force measurements are usually measured between the zero roughness levels @xcite ) : @xmath110 solving eq .
( [ k1 ] ) , one obtains @xmath111 nm . if the roughness is described by the regular ( nonstochastic ) functions @xmath112 , where @xmath113 , for the roughness amplitude it follows @xmath114 nm . in the framework of the additive approach the values of the casimir force including the effect of finite conductivity @xmath115 , obtained in sec .
iii , eq .
( [ eq9 ] ) , can be used to calculate the effect of roughness . for this purpose
, the values of @xmath116 should be geometrically averaged over all different possible separations between the rough surfaces weighted with the probability of each separation @xcite @xmath117 note that eq .
( [ k2 ] ) is not reduced to a simple multiplication of the correction factors due to finite conductivity and surface roughness but takes into account their combined ( nonmultiplicative ) effect .
an alternative method of calculating the corrections due to the stochastic surface roughness was used in ref .
@xcite . according to the results of ref .
@xcite , the casimir force between a plate and a sphere made of ideal metal and covered by a stochastic roughness with an amplitude @xmath118 is given by @xmath119 , \label{k3}\ ] ] where @xmath120 is the casimir force between perfectly shaped plate and sphere of radius @xmath50 .
then the casimir force including both the finite conductivity of the boundary metal and surface roughness can be calculated as @xmath121 , \label{k4}\ ] ] i.e. by means of the multiplicative procedure .
the variance of the random process describing the stochastic roughness is found by the formula @xmath122 using the data from table i , one obtains the values for variance @xmath123 nm and for the amplitude of a random process @xmath124 nm . this value is slightly larger than the one obtained in ref .
@xcite on the basis of less complete data on roughness topography .
now we are in a position to compare the contribution of the surface roughness computed by eq .
( [ k2 ] ) , taking into account the combined effect of the roughness and finite conductivity , and by the multiplicative procedure of eq .
( [ k4 ] ) . in table
ii the results for the correction factors @xmath125 , @xmath126 , @xmath127 , and @xmath128 are presented at the shortest separations @xmath2 nm , 70 nm , 80 nm , and 90 nm , where the roughness corrections play some role . as is seen from table ii ,
both approaches lead to practically coincident results for the roughness correction factors due to the combined effect of finite conductivity and surface roughness .
this means that for such small roughness as in ref .
@xcite the multiplicative procedure is quite satisfactory ( for larger roughness amplitudes , however , the nonmultiplicative contributions may be essential @xcite ) .
note also that for @xmath129 nm the fourth order term in eq .
( [ k4 ] ) practically does not contribute even at shortest separations and can be neglected as was done in ref .
@xcite .
both eqs .
( [ k2 ] ) and ( [ k4 ] ) used above are based on the approximation of additive summation and do not take into account the diffraction - type effects which arise in the case of roughness described by the periodic functions with small periods @xmath130 @xcite or by the stochastic functions with small correlation length @xcite . to estimate the value of the correlation length in our case , we consider the set of cross sections of the roughness image shown in fig . 1 . in fig .
2 two typical cross sections are presented , one at fixed @xmath131 ( a ) and the other one at fixed @xmath132 ( b ) .
we have performed the fourier analysis of the functions , as in figs .
2,a , b , along the lines of ref .
it was found that the fourier harmonics , giving the major contribution to the result , are characterized by significantly greater periods than the mean distance between the neighbouring peaks in figs .
2,a , b which is equal , approximately , to 180 nm . to obtain an estimate for the upper limit of the contribution of the diffraction - type effects in the above roughness analysis
, we use the correlation length @xmath133 nm ( slightly larger than the mean distance between peaks ) and consider the periodic function with this period ( clearly , the diffraction - type effects are greater for a periodic function with a period @xmath134 than for the random function with a correlation length @xmath134 ) . with this
the diffraction - type effects can be computed in the framework of the functional approach developed in ref .
@xcite . at a shortest separation @xmath2
nm one obtains @xmath135 .
then for the coefficient @xmath136 in the expression @xmath137 taking the diffraction - type effects into account , from fig . 2 of ref .
@xcite it follows @xmath138 . as a result , using the upper limit for the contribution of the diffraction effects one obtains @xmath139 , i.e. only 0.02% difference with the value of @xmath140 in table ii obtained by neglecting the diffraction effects . at larger separations
the diffraction effects lead to larger contribution to the roughness corrections .
for example , at a separation @xmath141 nm we have @xmath142 , @xmath143 , and @xmath144 , i.e. 0.03% difference with the result of table ii . at larger separations , however , the roughness correction itself is even more negligible than at the shortest separations . to conclude , the surface roughness contribution in the experiment of ref .
@xcite does not exceed 0.24% of the casimir force at the shortest separation @xmath2 nm .
the diffraction - type effects , which were not taken into account in eqs .
( [ k2 ] ) , ( [ k4 ] ) , are shown to contribute less than one tenth of this result .
although the experiment of ref .
@xcite was performed at room temperature @xmath145k , all the above computations were done at zero temperature . the thermal casimir force @xmath146 is given by eq .
( [ eq9 ] ) where integration in continuous @xmath147 is changed to a summation over the discrete matsubara frequencies @xmath148 according to @xmath149 leading to the lifshitz formula for the thermal casimir force .
here prime refers to the addition of a multiple 1/2 near the term with @xmath150 .
when @xmath151 , @xmath152 , where @xmath115 is given by eq .
( [ eq9 ] ) .
the magnitude of the relative thermal correction to the casimir force can be computed by the formula @xmath153 recently , there has been extensive discussion in literature on the correct calculation procedure for the thermal casimir force @xmath154 @xcite . in refs .
@xcite the dielectric permittivity of the plasma model ( [ eq12 ] ) was substituted into the lifshitz formula for @xmath154 .
this approach , which was later called `` traditional '' @xcite , leads to the thermal corrections @xmath155 , @xmath156 .
it is consistent with thermodynamics and agrees with the limiting case of the ideal metal . in the region of infrared optics
the same results were obtained in the framework of the impedance approach which does not consider the fluctuating electromagnetic field inside the metal and takes into account the realistic properties of the metal by means of the leontovich boundary condition @xcite . within the separation distances of ref .
@xcite , the traditional thermal corrections are very small . as an example , at a separation @xmath157 nm and @xmath145k one has @xmath158% , and @xmath159% , 0.1% at separations
@xmath3 nm , respectively , 300 nm @xcite ( to compare , in the case of ideal metals the same corrections , found in the framework of the thermal quantum field theory , are equal to 0.003% , 0.024% , and 0.08% , respectively , i.e. the results for real metals approach the results for ideal ones with the increase of separation @xcite ) .
thus , the traditional thermal corrections are negligible in the measurement range of experiment @xcite ( the contribution of the relaxation processes to the magnitude of these corrections , which can be computed by taking into account the small real part of the surface impedance , is much less than the corrections ) . alternatively , in refs .
@xcite the dielectric permittivity of the drude model ( [ eq11 ] ) was used to calculate @xmath154 . in this approach , there is no continuous transition between the cases of real and ideal metal . at the high temperature limit
, the casimir force between real metals was found equal to one half of the result obtained for the ideal metal ( independently of how high the conductivity of real metal is ) .
the thermal corrections , computed in the framework of the alternative approach @xcite , are quite different from those obtained from the traditional approach . to find the magnitude of these corrections @xcite
, one should substitude into eq .
( [ eq23 ] ) @xmath160 , \label{eq24}\ ] ] where @xmath161 is obtained by the substitution of eq .
( [ eq12 ] ) into eq .
( [ eq10 ] ) .
after calculations , one obtains that the alternative relative thermal correction increases from @xmath162% and 1.3% at separations
@xmath2 nm , respectively , 70 nm to @xmath163% at a separation @xmath164 nm .
another alternative thermal correction suggested in literature @xcite is also based on the substitution of the drude dielectric function ( [ eq11 ] ) into the lifshitz formula for @xmath154 but with a modified zero - frequency contribution for the perpendicular mode ( in ref .
@xcite this contribution is postulated to be of the same value as for an ideal metal ) .
the alternative thermal correction of ref .
@xcite is given by @xcite @xmath165 as a result , the relative alternative thermal correction of this kind takes values @xmath166% at all separations from @xmath2 nm to @xmath164 nm , i.e. slightly larger than the experimental precision at the shortest separations . as was shown in ref .
@xcite , both alternative thermal corrections of refs .
@xcite and of ref .
@xcite are not consistent with thermodynamics leading to the violation of the nernst heat theorem .
recently they were found to be in disagreement with the precision measurement of the casimir force using a microelectromechanical torsional oscillator @xcite . in sec . vi we will discuss the influence of the alternative thermal corrections on the comparison of theory and experiment in the casimir force measurement of ref .
@xcite . in the rest of this section
we discuss the probable contribution of the residual electric forces and the finite sizes of the plate on the casimir force . as was noted in ref .
@xcite , the electrostatic force due to the residual potential difference between the plate and the sphere has been lowered to negligible levels of @xmath167% of the casimir force at the closest separations . in recent ref .
@xcite it was argued , however , that the spatial variations of the surface potentials due to the grains of polycrystalline metal ( the so called `` patch potentials '' ) may mimic the casimir force .
here we apply the general results of ref .
@xcite to the experiment of ref .
@xcite and demonstrate that the patch effect does not make significant contributions .
according to ref .
@xcite , for the configuration of a sphere above a plate the electric force due to random variations in patch potentials is given by @xmath168 where @xmath169 is the variance of the potential distribution , @xmath170 ( @xmath171 ) are the magnitudes of the extremal wavevectors corresponding to minimal ( maximal ) sizes of grains , and @xmath172 is the dielectric permittivity of free space .
the work functions of gold are @xmath173ev , @xmath174ev , and @xmath175ev for different crystallographic surface orientations ( 100 ) , ( 110 ) , and ( 111 ) , respectively .
assuming equal areas of these crystallographic planes one obtains @xmath176^{1/2 } \approx 80.8\,\mbox{mv}. \label{eq27}\ ] ] using the atomic force microscopy images discussed in sec .
iii , the extremal sizes of grains in gold layers covering the test bodies were determined @xmath177 nm , and @xmath178 nm . this leads to @xmath179 and @xmath180 .
note that these grain sizes are of the same order as the thickness of the film .
the computations by eq .
( [ eq26 ] ) using the above data lead to the `` patch effect '' electric forces @xmath181n / m and @xmath182n / m at separations @xmath2 nm and @xmath157 nm , respectively . comparing the obtained results with the values of the casimir force at the same separations ( @xmath183n / m , respectively , @xmath184n / m ) , we conclude that the electric force due to the patch potentials contributes only 0.23% and 0.008% of the casimir force at separations @xmath2 nm , respectively , @xmath157 nm ( at a separation @xmath3 nm the patch effect contributes only @xmath185% of the casimir force .
so a rapid decrease of the contribution of the electric force with an increase of a separation is explained by the exponential decrease of the integral in eq .
( [ eq26 ] ) if one substitutes the physical values of the integration limits based on the properties of gold films used in experiment of ref .
@xcite .
a more important contribution of the patch electric forces may be expected in the scanning of a sharp tip of the atomic force microscope at a height of about 1020 nm above a plate . in this case
the electric forces are comparable with the van der waals forces complicating the theoretical interpretation of force - distance relations @xcite .
let us finally estimate the theoretical error caused by finiteness of the plate used in the experiment of ref .
( [ eq9 ] ) was derived for the plate of infinite radius .
in fact , the radius of the plate used in the experiment @xcite is @xmath186 m . in this case
, the casimir force can be obtained by the formula @xcite @xmath187f_c(z ) .
\label{eq28}\ ] ] to calculate @xmath188 , we put @xmath164 nm ( to make this factor maximally distinct from unity ) and obtain @xmath189 i.e. the finiteness of the plate size is too small to give any meaningful contributions to the casimir force .
now we are in a position to list all the sources of errors in the theoretical computation of the casimir force @xmath190 given by eq .
( [ k2 ] ) , to find the final theoretical accuracy and to consider the comparison of theory and experiment . the main error , which arises from eqs . ( [ eq9 ] ) , ( [ k2 ] ) when one substitutes the experimental data , is due to the errors in determination of separation @xmath13 and sphere radius @xmath50 . in terms of dimensionless variables , @xmath191 is proportional to @xmath50 and inversely proportional to @xmath192 which results in @xmath193 in ref .
@xcite the absolute separations were determined by means of the electric measurements which allowed the determination of the average separation distance on contact @xmath194 with the absolute error @xmath195 nm .
contrary to the opinion expressed in ref .
@xcite , this error should not be transferred to all separations leading to rather large contribution to @xmath196 of about @xmath197% at the shortest separations .
note that we are comparing not one experimental point with one theoretical value , but the experimental force - distance relation with the theoretical one computed on the basis of a fundamental theory .
thus , an additional fit should be made , with @xmath194 as a fitting parameter changing within the limits ( @xmath198 ) to achieve the smallest root mean square ( r.m.s . )
deviations between experiment and theory @xmath199\right\}^{1/2 } , \label{eq30}\ ] ] where @xmath200 was defined in eq .
( [ eq1 ] ) , @xmath201 were computed by eq .
( [ k2 ] ) , and @xmath202 is the number of experimental points under consideration . if , as usual , we consider two hypotheses as equivalent when they lead to the r.m.s .
deviations differing for less than 10% , this results in decrease of the error in determination of absolute separations up to @xmath203 nm .
it is important to underline that the verification of the hypothesis is performed within different separation intervals ( i.e. the total number @xmath204 experimental points within the whole separation range from 62 nm to 350 nm , @xmath205 points belonging to the interval 62210 nm , and @xmath206 points at separations less than the plasma wavelength @xmath207 nm ) .
the above value for @xmath208 is almost one and the same in all the separation intervals .
the obtained values of the r.m.s .
deviations between theory and experiment are @xmath209pn , @xmath210pn , and @xmath211pn .
these values are rather homogeneous demonstrating good agreement between theory and experiment independently of the chosen separation region .
the radius of the sphere was measured more precisely than in ref .
@xcite with a result @xmath212 m . using this together with @xmath213 nm ,
one obtains from eq .
( [ eq29 ] ) at the shortest separations @xmath214% .
now let us list the other contributions to the theoretical error of the casimir force computations at the shortest separation @xmath2 nm and indicate their magnitude . according to the results of sec .
iii , the sample to sample variations of the optical tabulated data may lead to the decrease of the casimir force magnitude for no more than @xmath215% .
the use of the proximity force theorem at @xmath2 nm leads to very small error of about @xmath216% ( sec .
the corrections due to the surface roughness are already incorporated in the theoretical expression ( [ k2 ] ) but the diffraction - type effects may contribute up to @xmath217% ( sec .
the effect of electric forces due to the patch potentials contribute a maximum @xmath218% at the shortest separations , as was shown in sec . v. the corrections due to the surface plasmons and finite size of the plate are negligible for the separation distances and experimental configuration used in ref .
@xcite ( see secs .
iii and v ) .
special attention should be paid to the thermal corrections to the casimir force . according to the results of sec .
v , the contribution of the traditional thermal correction at the shortest separation is negligible . at larger separations
it may be incorporated into the theoretical expression for the force . as to the alternative thermal corrections of refs .
@xcite , which contribute of about ( 12)% of the casimir force at the separation @xmath2 nm , they have been already ruled out both experimentally and theoretically ( see sec .
v ) . if we would include any of these corrections into the theoretical expression for the casimir force , this results in the increase of the r.m.s .
deviation between theory and experiment which can not be compensated by shifts of the separation distance in the limit of error @xmath219 . in view of the above
, we exclude the contributions from these hypothetical corrections from our error analysis .
the upper limit for the total theoretical error at a separation @xmath2 nm can be found by the summation of the above contributions @xmath220 which is a bit more accurate than the total experimental relative error at the shortest separation equal to 1.75% at 95% confidence ( see sec .
note that with the increase of separation the experimental relative error quickly increases to 37.3% at a separation @xmath3 nm . at the same time
, the theoretical error is slowly decreasing with increasing separation .
thus , at @xmath3 nm the above contributions to the theoretical error of the casimir force computations take values @xmath221% ,
@xmath215% , @xmath222% , @xmath223% , @xmath224% . as a result ,
the total theoretical error at @xmath3 nm is @xmath225% .
the obtained results demonstrate very good agreement between theory and experiment within the limits of both experimental and theoretical errors .
in this paper we have performed a detailed comparison of experiment and theory in the casimir force measurement between the gold coated plate and sphere by means of an atomic force microscope @xcite .
the random error of the experimental values of the casimir force was found to be @xmath226pn at 95% confidence ( at 60% confidence the value @xmath227pn was obtained ) .
together with the systematic error @xmath228pn , this leads to the total absolute error of the casimir force measurements in ref .
@xcite @xmath229pn at 95% confidence . in terms of the relative errors ,
the experimental precision at the shortest reported separation is equal to 1.75% ( 1% ) at 95% ( 60% ) confidence level . in order to find the theoretical accuracy of the casimir force calculations in the experimental configuration of ref .
@xcite , many corrections to the ideal casimir force were analysed .
the correction due to the finite conductivity of gold was computed by the use of the optical tabulated data of the complex refractive index .
the results were compared with those computed by the use of the plasma dielectric function and found to coincide for the surface separation range ( 200350)nm . at shorter separations the use of the optical tabulated data
more accurately represents the dielectric properties .
a special model was presented , which allows to take into account the sample to sample variations of the optical tabulated data due to the sizes of grains and impurities .
it was shown that the error introduced by the grains of 45 nm size ( even smaller than those in the experiment of ref.@xcite ) does not exceed 0.5% of the casimir force .
the influence of the surface plasmon in the separation region of the experiment @xcite was found to be negligible .
the surface roughness of the test bodies , used to measure the casimir force , was carefully investigated by means of the atomic force microscope with a sharp tipped cantilever instead of a large sphere .
the obtained profiles of roughness topography allowed calculation of the roughness corrections to the casimir force in the framework of both multiplicative and nonmultiplicative approaches .
the minor differences in the size of the effect are found only at the shortest separation .
the correlation length of the surface roughness on the test bodies was estimated and the diffraction - type effects were computed . at the shortest separation
the roughness correction contributes 0.24% of the casimir force with account of diffraction - type effects ( and 0.22% with no account of diffraction ) .
the electric forces caused by the spatial variations of the surface potentials due to the size of grains were investigated for the experimental configuration of ref .
they were shown to contribute 0.23% of the casimir force at the shortest separation , and this contribution quickly decreases with an increase of separation .
several other effects ( such as thermal corrections , corrections due to the finiteness of the plate and due to the deviation from the proximity force theorem ) were investigated and found to make only negligible contributions . the final theoretical accuracy of the casimir force calculations in the experimental configuration of ref .
@xcite is 1.69% at the shortest separation @xmath2 nm and 1.1% at a separation @xmath3 nm . in the limits of both experimental and theoretical errors ,
very good agreement between theory and experiment was demonstrated characterized by the r.m.s .
deviation of about 3.5pn ( less than 1% of the measured force at a shortest separation ) which is almost independent of the separation region and the number of the experimental points .
the above analysis does not support the conclusion of ref .
@xcite that to achieve the experiments on the casimir effect with 1% precision it is necessary to measure the separation on contact @xmath194 with atomic precision .
the obtained results demonstrate that in fact the casimir force is more stable , than one might expect , to some delicate properties of the metallized test bodies like the variations of the optical data , patch potentials , correlation effects of roughness etc .
these properties may change from sample to sample leaving the basic character , and even the values of the casimir force within some definite separation region , almost unchanged .
the stability of the casimir force opens new opportunities to use the casimir effect as a test for long - range hypothetical interactions and for the diagnostic purposes .
for example , some kind of the inverse problem could be utilized , i.e. the measured force - distance relations be exploited to determine the fundamental characteristics of solids ( such as the plasma frequency ) .
this work was supported by the national institute for standards and technology and a university of california and los alamos national laboratory grant through the lanl - care program .
g.l.k . and v.m.m .
were also partially supported by cnpq and finep ( brazil ) .
99 h. b. g. casimir , proc .
akad . wet . *
51 * , 793 ( 1948 ) .
w. milonni , _ the quantum vacuum _ ( academic press , san diego , 1994 ) .
v. m. mostepanenko and n. n. trunov , _ the casimir effect and its applications _ ( clarendon press , oxford , 1997 ) .
k. a. milton , _ the casimir effect _ ( world scientific , singapore , 2001 )
. m. kardar and r. golestanian , rev .
phys . * 71 * , 1233 ( 1999 ) .
m. bordag , u. mohideen , and v. m. mostepanenko , phys .
* 353 * , 1 ( 2001 ) .
s. k. lamoreaux , phys .
lett . * 78 * , 5 ( 1997 ) .
u. mohideen and a. roy , phys .
lett . * 81 * , 4549 ( 1998 ) ; g. l. klimchitskaya , a. roy , u. mohideen , and v. m. mostepanenko , phys . rev .
a * 60 * , 3487 ( 1999 ) .
a. roy and u. mohideen , phys .
lett . * 82 * , 4380 ( 1999 ) .
a. roy , c .- y .
lin , and u. mohideen , phys .
d * 60 * , 111101(r ) ( 1999 ) .
b. w. harris , f. chen , and u. mohideen , phys .
a * 62 * , 052109 ( 2000 ) .
t. ederth , phys .
a * 62 * , 062104 ( 2000 ) .
g. bressi , g. carugno , r. onofrio , and g. ruoso , phys .
lett . * 88 * , 041804 ( 2002 ) .
f. chen , u. mohideen , g. l. klimchitskaya , and v. m. mostepanenko , phys .
lett . * 88 * , 101801 ( 2002 ) ; phys .
a * 66 * , 032113 ( 2002 ) .
r. s. decca , d. lpez , e. fischbach , and d. e. krause , phys .
lett . * 91 * , 050402 ( 2003 ) .
r. s. decca , e. fischbach , g. l. klimchitskaya , d. e. krause , d. lpez , and v. m. mostepanenko , phys .
d * 68 * , 116003 ( 2003 ) .
h. b. chan , v. a. aksyuk , r. n. kleiman , d. j. bishop , and f. capasso , science , * 291 * , 1941 ( 2001 ) ; phys .
* 87 * , 211801 ( 2001 ) .
m. bordag , b. geyer , g. l. klimchitskaya , and v. m. mostepanenko , phys .
d * 58 * , 075003 ( 1998 ) ; * 60 * , 055004 ( 1999 ) ; * 62 * , 011701(r ) ( 2000 ) .
j. c. long , h. w. chan , and j. c. price , nucl .
b * 539 * , 23 ( 1999 ) . v. m. mostepanenko and m. novello , phys .
d * 63 * , 115003 ( 2001 ) .
e. fischbach , d. e. krause , v. m. mostepanenko , and m. novello , phys .
d * 64 * , 075010 ( 2001 ) .
g. l. klimchitskaya and u. mohideen , int .
a * 17 * , 4143 ( 2002 ) . c. h. hargreaves , proc .
b * 68 * , 231 ( 1965 ) .
j. schwinger , l. l. deraad , jr . , and k. a. milton , ann .
( n.y . ) * 115 * , 1 ( 1978 ) . v. m. mostepanenko and n. n. trunov , yad .
42 * , 1297 ( 1985 ) [ sov
. phys . * 42 * , 818 ( 1985 ) ] .
a. a. maradudin and p. mazur , phys .
b * 22 * , 1677 ( 1980 ) .
m. bordag , g. l. klimchitskaya , and v. m. mostepanenko , int .
phys . a * 10 * , 2661 ( 1995 ) .
j. mehra , physica * 37 * , 145 ( 1967 ) .
l. s. brown and g. j. maclay , phys . rev . *
184 * , 1272 ( 1969 ) .
t. emig , a. hanke , r. golestanian , and m. kardar , phys .
lett . * 87 * , 260402 ( 2001 ) ; phys .
a * 67 * , 022114 ( 2003 ) . c. genet , a. lambrecht , p. maia neto , and s. reynaud , europhys
. lett . * 62 * , 484 ( 2003 ) .
s. brandt , _ statistical and computational methods in data analysis _ ( north - holland publishing company , amsterdam , 1976 ) .
r. c. weast and m. j. astle , ( crc press , inc .
, boca raton , 1982 ) .
i. e. dzyaloshinskii , e. m. lifshitz , and l. p. pitaevskii , usp .
nauk * 73 * , 381 ( 1961 ) [ sov .
( usa ) * 4 * , 153 ( 1961 ) ] .
j. blocki , j. randrup , w. j. swiatecki , and c. f. tsang , ann .
( n.y . ) * 105 * , 427 ( 1977 ) .
p. johansson and p. apell , phys .
b * 56 * , 4159 ( 1997 ) . m. schaden and l. spruch , phys .
a * 58 * , 935 ( 1998 ) .
a. lambrecht and s. reynaud , eur .
j. d * 8 * , 309 ( 2000 ) .
g. l. klimchitskaya , u. mohideen , and v. m. mostepanenko , phys .
a * 61 * , 062107 ( 2000 ) . ,
e. d. palik ( academic , new york , 1985 ) .
e. m. lifshitz and l. p. pitaevskii , _ physical kinetics _
( pergamon press , oxford , 1981 )
. i. m. lifshitz , m. ya .
asbel , and m. i. kaganov , _ electron theory of metals _ ( consultants bureau , new york , 1973 ) .
l. p. pitaevskii , zh .
34 * , 942 ( 1958 ) [ sov .
jetp * 34 * , 652 ( 1958 ) ] .
i. m. kaganova and m. i. kaganov , phys .
b * 63 * , 054202 ( 2001 ) . j. sotelo , j. ederth , and g. niklasson , phys .
b * 67 * , 195106 ( 2003 ) . l. d. landau , e. m. lifshitz , and l. p. pitaevskii , _ electrodynamics of continuous media _ ( pergamon press , oxford , 1984 ) .
b. geyer , g. l. klimchitskaya , and v. m. mostepanenko , phys .
rev . a * 67 * , 062102 ( 2003 )
. j. heinrichs , phys .
rev . b. * 11 * , 3625 ( 1975 ) ; * 11 * , 3637 ( 1975 ) ; * 12 * , 6006 ( 1975 ) .
r. esquivel , c. villarreal , and w. l. mochn , phys .
rev . a , * 68 * , 052103 ( 2003 ) .
m. bordag , g. l. klimchitskaya , and v. m. mostepanenko , phys .
a * 200 * , 95 ( 1995 ) . c. genet , a. lambrecht , and s. reynaud , phys .
a * 62 * , 012110 ( 2000 ) .
m. bordag , b. geyer , g. l. klimchitskaya , and v. m. mostepanenko , phys .
85 * , 503 ( 2000 ) ; * 87 * , 259102 ( 2001 ) .
v. b. bezerra , g. l. klimchitskaya , and c. romero , phys .
a * 65 * , 012111 ( 2002 ) .
g. l. klimchitskaya and v. m. mostepanenko , phys .
a * 63 * , 062108 ( 2001 ) .
m. bostrm and b. e. sernelius , phys .
. lett . * 84 * , 4757 ( 2000 ) .
j. s. hye , i. brevik , j. b. aarseth , and k. a. milton , _ phys .
_ e * 67 * , 056116 ( 2003 ) . v. b. svetovoy and m. v. lokhanin , phys .
a * 280 * , 177 ( 2001 ) .
v. b. bezerra , g. l. klimchitskaya , and v. m. mostepanenko , phys .
a * 66 * , 062112 ( 2002 ) . c. c. speake and c. trenkel , phys .
lett . * 90 * , 160403 ( 2003 ) .
n. a. burnham , r. j. colton , and h. m. pollock , phys .
* 69 * , 144 ( 1992 ) .
v. b. bezerra , g. l. klimchitskaya , and c. romero , mod .
lett . a * 12 * , 2623 ( 1997 ) .
d. iannuzzi , i. gelfand , m. lisanti , and f. capasso , quant - ph/0312043 . | we compare theory and experiment in the casimir force measurement between gold surfaces performed with the atomic force microscope .
both random and systematic experimental errors are found leading to a total absolute error equal to 8.5pn at 95% confidence . in terms of the relative errors , experimental precision of 1.75% is obtained at the shortest separation of 62 nm at 95% confidence level ( at 60% confidence the experimental precision of 1% is confirmed at the shortest separation ) .
an independent determination of the accuracy of the theoretical calculations of the casimir force and its application to the experimental configuration is carefully made .
special attention is paid to the sample - dependent variations of the optical tabulated data due to the presence of grains , contribution of surface plasmons , and errors introduced by the use of the proximity force theorem .
nonmultiplicative and diffraction - type contributions to the surface roughness corrections are examined .
the electric forces due to patch potentials resulting from the polycrystalline nature of the gold films are estimated . the finite size and thermal effects
are found to be negligible .
the theoretical accuracy of about 1.69% and 1.1% are found at a separation 62 nm and 200 nm , respectively . within the limits of experimental and theoretical errors
very good agreement between experiment and theory is confirmed characterized by the root mean square deviation of about 3.5pn within all measurement range .
the conclusion is made that the casimir force is stable relative to variations of the sample - dependent optical and electric properties , which opens new opportunities to use the casimir effect for diagnostic purposes . |
the low energy qcd dynamics in the real world will be understood by lattice qcd with three - flavors of dynamical quarks .
several efforts have been spent to develop exact numerical algorithms with an odd - numbers of the wilson type quarks @xcite .
the kogut - susskind ( ks ) fermion is an attractive formalism since the numerical simulation with much lighter quark masses are possible thanks to the remnant chiral symmetry .
although lattice qcd with the two- or single - flavor ks fermions can be defined by taking the fractional power of the ks fermion , efficient exact algorithms are not still known .
approximate algorithms such as the @xmath11-algorithm @xcite have been used in these cases .
several exact algorithms are proposed for two- or single - flavor dynamical ks fermions @xcite . in this paper
, we further study the idea by horvth _
et al . _
@xcite in the case of the polynomial hybrid monte carlo ( phmc ) algorithm .
we develop two types of the phmc algorithm depending on the choice of the effective action derived through the polynomial approximation .
we compare the computational cost of these two phmc algorithms .
we investigate the property of the algorithm on several lattice sizes in @xmath0@xmath1@xmath2 case .
we found that our algorithm shows satisfactory efficiency on a @xmath3 lattice with @xmath4@xmath1@xmath5 , @xmath6@xmath1@xmath7 @xmath12@xmath9@xmath13 .
we construct two types of the phmc algorithm , which are refereed to as case a and case b. introducing a polynomial approximation and pseudo - fermion field , the partition function can be generally rewritten in the following form : @xmath14^{n_{f}/4 } \nonumber\\ & & \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \times e^{-s_{g}[u]-s^{(x)}_{q}[u,\phi_{o}^{\dag},\phi_{o } ] } , \label{eq : effective_action1}\end{aligned}\ ] ] where @xmath15 is a lattice gauge action , @xmath16 is pseudo - fermion field living only on odd sites . @xmath17 and @xmath18 are the pseudo - fermion action and the correction matrix respectively .
the superscript @xmath19 takes @xmath20 or @xmath21 depending on the type of the phmc algorithm as follows . *
* case a:**we approximate @xmath22 by a @xmath23 order polynomial @xmath24 $ ] .
the pseudo - fermion action and the correction term become @xmath25&= & w^{(a)}[{\hat{d}_{oo}}]&=&{\hat{d}_{oo}}(p_{{n_{\mathit{poly}}}^{(a)}}[{\hat{d}_{oo}}])^{8/n_{f}}. \label{eq : casea_action}\end{aligned}\ ] ] * * case b:**we approximate @xmath26 by an even - order @xmath27 polynomial @xmath28 $ ] . the pseudo - fermion action and the correction term can be written as @xmath29&= & w^{(b)}[{\hat{d}_{oo}}]&=&{\hat{d}_{oo}}(p_{{n_{\mathit{poly}}}^{(b)}}[{\hat{d}_{oo}}])^{4/n_{f } } , \label{eq : caseb_action}\end{aligned}\ ] ] where @xmath30 is the @xmath31 order polynomial defined by @xmath28$]@xmath1@xmath32|^{2}$ ] .
the ks - fermion operator @xmath33 is even - odd preconditioned as @xmath33@xmath1@xmath34@xmath35@xmath36 with @xmath37@xmath1@xmath38@xmath39 @xmath40@xmath41@xmath42 and @xmath43@xmath1@xmath44@xmath41@xmath34 .
@xmath45 is chosen so that the all eigenvalues of @xmath43 fall into the region @xmath46 $ ] . @xmath47 ( @xmath48 )
is the usual ks hopping matrix from even ( odd ) to odd ( even ) sites . for both cases ,
the algorithm takes the following two steps ; ( i ) perform the hmc algorithm according to the effective action eq .
( [ eq : casea_action ] ) or ( [ eq : caseb_action ] ) , ( ii ) when the hmc metropolis test is accepted , apply the kennedy - kuti noisy metropolis test to incorporate the correction term @xmath49^{n_f/4}$ ] .
thus we obtain two types of the phmc algorithm depending on the choice of the effective action .
the acceptance probability of the noisy - metropolis test is defined by @xmath50=\min[1,e^{-ds}],\nonumber\\ & & ds=\zeta_{o}^{\dag}w^{(x)}[{\hat{d}_{oo}}']^{-n_f/4}\zeta_{o}-|\eta_{o}|^2 , \label{eq : ds}\end{aligned}\ ] ] where @xmath51@xmath1@xmath52^{n_f/8}\eta_{o}$ ] with a gaussian noise vector @xmath53 .
@xmath52 $ ] is calculated on an initial configuration and @xmath54 $ ] is on a trial configuration generated by the preceding hmc algorithm .
the fractional power of the correction matrix @xmath18 is taken by the lanczos based krylov subspace method proposed by borii @xcite .
we modified his algorithm suitable to our purpose .
as indicated by borii we employ cg based stopping criterion for the lanczos based method .
the computational cost is counted as the number of multiplication of the hopping matrix to evolve the algorithm unit trajectory .
we employ single leapfrog integration scheme for the molecular dynamics ( md ) step .
we roughly estimate it as @xmath55 where @xmath56 is the number of md step and @xmath57 the number of iteration of cg algorithm . the cg algorithm is used to generate @xmath16 with the global heat - bath method and in the lanczos based algorithm for the noisy metropolis test .
now we compare the costs by specifying @xmath23 and @xmath27 . for this purpose
we employ the chebyshev polynomial approximation ; @xmath58=\sum_{i=0}^{{n_{\mathit{poly } } } } c_{k}t_{k}[y],\ ] ] where @xmath59 is the @xmath60-th order chebyshev polynomial , @xmath61@xmath1@xmath33 and @xmath62@xmath1@xmath63 can be read of .
the coefficients @xmath64 are calculated as usual .
figure [ fig:1 ] shows the cost @xmath65 dependence of the integrated residual defined by @xmath66@xmath1@xmath67)^{1/s}-1 |^{2 } ) ^{1/2}$ ] for each @xmath68 .
we observe that as decreasing @xmath68 by factor @xmath69 the cost @xmath65 increases by factor @xmath2 at a constant @xmath70 .
this is nothing but @xmath71 does not depend on the choice of @xmath68 and we obtain @xmath23@xmath9@xmath27 at a constant approximation level . using this relation and assuming @xmath72@xmath1@xmath73 and @xmath74@xmath1@xmath75 , we find @xmath76@xmath9@xmath77 .
we employ the chebyshev polynomial and the plaquette gauge action , and apply the case b phmc algorithm for numerical simulations .
dependence of the integrated residual @xmath66 with @xmath78 . ] [ fig:1 ]
figure [ fig:2 ] shows the md step size dependence of the averaged plaquette on a @xmath79 lattice at @xmath4@xmath1@xmath80 and @xmath6@xmath1@xmath81 .
the results with the phmc algorithm of @xmath71@xmath1@xmath82 do not depend on @xmath83 as it should be and produce the consistent result to that in the zero md step size limit of the @xmath11-algorithm .
although we do not show the results on the @xmath71 dependence , the results are independent of the choice of @xmath71 . in table
[ tab:1 ] we show the numerical results on a @xmath3 lattice at @xmath4@xmath1@xmath5 and @xmath6@xmath1@xmath7 .
the averaged plaquette are independent of @xmath71 and consistent with each other as expected . on the other hand ,
the @xmath11-algorithm yields @xmath84@xmath1@xmath85 @xcite , which differs from ours by @xmath9@xmath86 .
this indicates a potential systematic error for the @xmath11-algorithm at finite md step size .
the computational time for @xmath71@xmath1@xmath87 with @xmath6@xmath1@xmath7 ( which corresponds to @xmath8@xmath9@xmath10 @xcite ) was measured as 112 sec . to achieve unit trajectory with 14 gflops sustained speed of sr8000 at kek .
consequently we conclude that the phmc algorithm we constructed works on a moderately large lattice size with rather heavy quark masses with reasonable computational cost .
the algorithm also works with a single - flavor fermion .
we emphasize that because the phmc algorithm is exact one , it must be a promising algorithm for future realistic simulations .
this work is supported by the supercomputer project no .
79 ( fy2002 ) of high energy accelerator research organization ( kek ) , and also in part by the grant - in - aid of the ministry of education ( nos
. 11640294 , 12304011 , 12740133 , 13135204 , 13640259 , 13640260 ) .
is supported by the jsps research fellowship .
t. takaishi and ph . de forcrand , int . j. mod
c13 ( 2002 ) 343 ; nucl .
( proc . suppl . ) 94 ( 2001 ) 818 ; hep - lat/0009024 ; jlqcd collaboration , s. aoki _ et al .
d 65 ( 2002 ) 094507 ; f. farchioni , c. gebert , i. montvay , and w. schroers , nucl.phys.proc.suppl .
106 ( 2002 ) 215 . | we present a polynomial hybrid monte carlo ( phmc ) algorithm as an exact simulation algorithm with dynamical kogut - susskind fermions .
the algorithm uses a hermitian polynomial approximation for the fractional power of the ks fermion matrix . the systematic error from the polynomial approximation
is removed by the kennedy - kuti noisy metropolis test so that the algorithm becomes exact at a finite molecular dynamics step size .
we performed numerical tests with @xmath0@xmath1@xmath2 case on several lattice sizes .
we found that the phmc algorithm works on a moderately large lattice of @xmath3 at @xmath4@xmath1@xmath5 , @xmath6@xmath1@xmath7 ( @xmath8@xmath9@xmath10 ) with a reasonable computational time . |
in the past decades , extensive experimental and theoretical efforts have been devoted to search for the nuclear liquid - gas phase transition ( lgpt ) in intermediate energy heavy - ion collisions ( hic ) @xcite .
many probes have been suggested for the onset of nuclear lgpt .
for instance , the fragment size distribution @xcite and its rank distribution @xcite , the largest fluctuation of the heaviest fragment @xcite , caloric curve @xcite , bimodality @xcite etc .
in addition , it has been observed that the ratio of shear viscosity to entropy density ( @xmath3 ) reaches its local minimum at the transition temperature for a wide class of systems .
for instance , empirical observation of the temperature or incident energy dependence of the shear viscosity to entropy density ratio for h@xmath8o , he and ne2 exhibits a minimum in the vicinity of the critical point for phase transition @xcite . and
a lower bound of @xmath9 obtained by kovtun - son - starinets ( kss ) for infinitely coupled super - symmetric yang - mills gauge theory based on the ads / cft duality conjecture , is speculated to be valid universally @xcite . in ultra - relativistic hic @xcite
, people have used the ratio of shear viscosity to entropy density to study the quark - gluon plasma phase and the extracted value of @xmath3 seems close to the kss bound ( @xmath5 ) .
so far there are many interesting investigations on the ratio of @xmath3 , but it is still rare to study the behavior of @xmath3 during the heavy - ion collision at intermediate energies @xcite .
furthermore , the influences of nucleon - nucleon cross section and nuclear symmetry energy are less discussed . in this work
we use a microscopic transport model known as the isospin - dependent quantum dynamics model @xcite to simulate au+au central collisions . in order to study the effect of nucleon - nucleon cross section , 0.5 times and 1.5 times normal nucleon - nucleon cross section
are also used in the simulation . on the other hand for the symmetry energy which is important for asymmetric nuclear matter as well as nuclear astrophysics @xcite ,
different symmetry energy parameters are employed , namely 15 mev , 25 mev and 35 mev . the generalized hot thomas fermi formalism ( ghtff ) @xcite and the transport formula @xcite
are employed , respectively , to extract thermodynamic and transport properties of the nuclear fireball which is located in the central region with a moderate volume .
then different correlations between the extracted thermal and transport properties are discussed and a good agreement with our previous calculations is found @xcite .
furthermore the multiplicity of intermediate mass fragments ( imfs ) is also checked as a signal of liquid gas phase transition @xcite to verify the calculated result .
the paper is organized as follows .
section 2 provides a brief introduction for the iqmd model , ghtff as well as transport formula for shear viscosity . in section 3
we present the calculation results and discussions , where the time evolution of thermodynamic quantities and shear viscosity over entropy density are focused .
finally a summary and outlook is given .
the quantum molecular dynamics ( qmd ) @xcite model approach is a many - body theory which describes heavy ion collisions from intermediate to relativistic energy .
the isospin - dependent quantum molecular ( iqmd ) @xcite model is based on the qmd model , including the isospin effects and pauli blocking .
each nucleon in the colliding system is described as a gaussian wave packet , i.e. @xmath10.\end{aligned}\ ] ] here @xmath11 and @xmath12 are the mean position and mean momentum , and the gaussian width has the fixed value @xmath13 for au + au system .
the centers of these gaussian wave packets propagate in coordinate ( @xmath14 ) and momentum ( @xmath15 ) space according to the classical equations of motion : @xmath16 where @xmath17 is the hamiltonian of the system .
the wigner distribution function for a single nucleon density in phase space is given by @xmath18 \nonumber \\ & & \exp \left [ \frac{-2l({\bf p}-{\bf p}_i(t))^2}{\hbar^2}\right].\end{aligned}\ ] ] the mean field in iqmd model is written as @xmath19 where @xmath20 , @xmath21 , @xmath22 , and @xmath23 represents the skyrme potential , the coulomb potential , the yukawa potential and the symmetry potential interaction , respectively @xcite .
the skyrme potential is @xmath24 where @xmath25 = 0.16 @xmath26 and @xmath27 is the nuclear density . in the present work , the parameters @xmath28 mev , @xmath29 mev , and @xmath30 ,
correspond to a soft eos , are used .
@xmath31 is a long - range interaction ( surface ) potential , and takes the following form @xmath32,\end{aligned}\ ] ] with @xmath33gev , @xmath34 = 1.25@xmath35 , @xmath36 @xmath37 , and @xmath38 is the relative distance between two nucleons .
the symmetry potential is @xmath39 , where @xmath40 , @xmath41 , and @xmath42 are the neutron , proton and nucleon densities , respectively .
@xmath7 is the symmetry energy coefficient , and three different values of 15 , 25 and 35 mev are taken in order to study its influence on the ratio of @xmath3 .
furthermore the isospin degree has entered into the cross sections , which is similar to the parametrization of verwest and arndt , see ref .
the cross section for the neutron - neutron collisions is assumed to be equal to the proton - proton cross sections . in order to study the effect of cross section on the ratio of @xmath3
, the nucleon - nucleon cross section is multiplied by a coefficient @xmath43 .
three different situations are considered , namely @xmath43 equals 0.5 , 1.0 and 1.5 respectively . in a practical viewpoint
, a smaller @xmath44 seems suitable to describe hic , which was proposed in the previous work ref .
@xcite @xmath45 from eq .
( 3 ) one obtains the matter density of coordinate space by the sum over all the nucleons , namely @xmath46 the kinetic energy density in coordinates space could also be calculated from eq.(4 ) by @xmath47 thermodynamical properties of hot nuclear matter formed in heavy ion collisions , e.g. temperature and entropy density , can be extracted by using the approach developed by faessler and collaborators @xcite . in this approach one
starts from a microscopic picture of two interpenetrating pieces of nuclear matter and deduces the thermal quantities from the matter density and kinetic energy density obtained during the collisions . in this paper
, the extraction of thermal properties of the hot nuclear matter is done in two steps .
first , based on the iqmd simulation , one could calculate the nuclear matter and kinetic energy densities at each point in coordinate space at every time step .
second , by employing the hot thomas - fermi formalism , we could obtain the corresponding thermal properties for every set of nuclear matter density and nuclear kinetic energy density@xcite . in ghtff , the momentum distribution in cylindrical coordinates @xmath48 can be written as @xmath49)^{-1 } , & k_{z } < k_{0 } \\ ( 1+exp[\hbar^{2}(k_{r}^{2}+(k_{z}-k_{r})^{2})/2mt-\mu_{2}^{'}])^{-1 } , & k_{z } >
k_{0 } \nonumber \end{array } \right.\end{gathered}\ ] ] with @xmath50 is the reduced chemical potential , @xmath51/2k_{r}$ ] , @xmath52 is the relative momentum between the projectile ( index 1 ) and target ( index 2 ) .
the local nuclear matter density @xmath53 is expressed as @xmath54,\end{aligned}\ ] ] where @xmath55 with @xmath56 , and @xmath57 is the fermi integrals , i.e. @xmath58 @xmath59 $ ] .
+ the local kinetic energy density @xmath60 , where @xmath61 reads @xmath62+\delta\tau_{i}(\mu_{i}^{'}).\end{aligned}\ ] ] and the entropy density @xmath63 is written as @xmath64.\end{aligned}\ ] ] here @xmath65 represents the projectile and target , and @xmath66 , + @xmath67\nonumber \\&&+k_{r}^2\rho_{2}(\mu_{2}^{'}),\nonumber\end{aligned}\ ] ] from the eq.([eq - rho],[eq - tao],[eq - si ] ) , one can obtain the thermal properties by inversion in principle .
but such an inversion procedure is practically not feasible due to the complexity of the equations .
therefore , a more practical way is chosen to obtain the thermal properties .
first , we generate all reasonable combinations @xmath4 , @xmath68 and @xmath69 , which ranging from 0 - 100mev , 0 - 5@xmath70 and 0 - 2 , respectively . then the corresponding @xmath71 could be obtained .
second , from the extracted @xmath72 in the central region at each time step during the evolution of collision , @xmath4 , @xmath68 and @xmath69 are obtained from the calculations in the first step .
third , the entropy density is calculated according to eq.[eq - si ] .
one should pay attention that all the values displayed in the following pictures are the average one in the central region . for largely equilibrated systems , fluxes of macro quantities , leading to dissipation , are proportional to gradients within the system .
the shear viscosity denoted as @xmath1 is the coefficient of proportionality between anisotropy of momentum - flux tensor , including dissipation and velocity gradients @xcite . in the boltzmann statistical limit the shear viscosity corresponds to the first order chapman - enskog coefficients . in ref .
@xcite the nuclear shear viscosity for normal n - n cross section , has been derived from the microscopic boltzmann - uehling - ulenbeck equation and can be parameterized as a function of density @xmath27 and temperature @xmath4 : @xmath73 where @xmath1 is in mev/@xmath74 , @xmath4 in mev , and @xmath25=0.168@xmath26 .
figure [ fig_eta - eqn ] shows @xmath1 as a function of @xmath4 and @xmath75 .
one can see that @xmath1 exhibits a very distinct minimum when nuclear matter density is less than normal nuclear density . and
as the density increases , the transition temperature also get larger , e.g. for normal density the transition temperature locates around 10 mev , but for 2.5 times normal density it is almost 50 mev .
this conclusion is coincident with macroscopic result . and in the case of scaled n - n cross section , the shear viscosity just only needs to be scaled by @xmath76 , ie .
@xmath77 from eq.[dani - eqn - cross ] we can see that the shear viscosity is very sensitive to the n - n cross section , the larger the cross section is , the smaller the shear viscosity . and
it is intuitive that large n - n cross section makes the transport of particle momentum much difficult .
since the equilibrium of the considered nuclear matter at the very starting stage is not reached , the calculated shear viscosity should be considered as the transport properties of largely equilibrated nuclear matter with the same density and kinetic energy density . and @xmath4 with the eq.([dani - eqn ] ) .
different colors represent different @xmath75 , which are illustrated in the inset .
, width=377 ]
in present work , we simulate head - on collision of au + au at different beam energies . the reason why we choose the central collision
is that the participant zone is the maximal and the matter is more nuclear liquid - like during the early time evolution of collision . fig .
[ fig_xz ] shows the evolution of density profile in x - z plane for au+au head - on collisions at 50 mev / nucleon and 200 mev / nucleon at 2 , 10 , 20 , 30 , 50 and 80 fm / c .
the zero point of time is set at the initial contact between project and target ( the first left panel ) . with
the collision goes on , the system overlaps and seems more isotropic in phase space . in order to calculate the thermodynamic quantities in different time steps , we select the central sphere with radius @xmath78=5 fm , which defines a volume of nuclear fireball in this paper . the time evolution of the average nuclear matter density ( panel ( a ) ) and kinetic energy density ( panel ( b ) ) in the central region , with @xmath79 ,
is showed in fig .
[ fig_rho - time ] .
it is interesting that the maximum density reached is about @xmath80 to @xmath81 and the maximum kinetic energy density is 10 mev @xmath26 to 25 mev @xmath26 for the energy displayed in the picture . along the time scale of the collision
one can see that both @xmath75 and @xmath82 are reaching their maxima at about 20 fm / c and at a bit earlier time for higher energy .
after the compression stage the matter starts to expand and some of them will escape from the central region , mainly in the transverse plane , the matter density drops to very small values and the central region is cooled down .
in general the warm and dense nuclear matter survives much longer when the incident energy is low . at about 80 fm / c
the hot and dense matter disappears .
time evolution of matter density and kinetic energy density are shown in fig .
[ fig_130mev - time ] , when @xmath43 and @xmath7 are set by different values .
the upper panels demonstrate the different cross section situation ( @xmath83 $ ] ) , on the other hand the bottom panels are for different symmetry energy . in panels ( a ) and ( b ) , it is easy to find that the nuclear matter density and kinetic energy density are different from each other when @xmath43 is different .
we found that there is no difference for the nuclear matter density during compression stage .
but when the system starts to expand , the larger nucleon - nucleon cross section makes the dense matter stay longer .
in contrast with the behavior of density around the maximum compression stage , more distinction for the kinetic energy density is exhibited .
it shows that the smaller the nucleon - nucleon cross section , the larger the kinetic energy density . but these curves almost overlap each other after 40 fm / c
, it may be understood that the longitudinal energy enters the central region more easily when the nucleon - nucleon cross section is small .
it should be noted that the large kinetic energy density dose not mean higher temperature , since the kinetic energy is not calculated in the center of mass frame , detailed information can be found in refs .
in addition , the extracted density and kinetic energy density show insensitivity to the symmetry energy as depicted in panel ( c ) and ( d ) .
the curves are overlapped with each other for the whole process . as has been discussed in the previous paragraph ,
this leads the nuclear matter to the same thermal properties .
so the extracted thermal properties based upon the hot thomas - fermi formulism keep exactly the same with each other .
so in this paper we just investigate the nucleon - nucleon cross section effect on thermodynamic and transport quantities in the following texts .
time evolution of temperature is plotted in fig .
[ fig_temper - merge ] .
[ fig_temper - merge](a ) shows the time evolution of temperature with @xmath43=1.0 at different incident energies .
[ fig_temper - merge](b ) shows the time evolution of temperature when the incident energy is fixed at 130 mev / nucleon but with different cross sections .
the following pictures are arranged with the same mode , i.e panel ( a ) represents a constant cross section @xmath84 but at different incident energies ; panel ( b ) means a constant incident energy at 130 mev / nucleon with different cross sections . for a given beam energy , temperature increases at first , then reaches a local maximum about 20 fm / c and decreases till a saturated value at about 80 fm / c
. the higher the incident energy , the larger the maximum value .
the corresponding time at maximum value is a little earlier than that for the density and kinetic density . in panel ( b )
, it is found that the larger cross section makes the system a little hotter .
the reason is that there are more frequent nucleon - nucleon collisions as @xmath6 becomes larger , which makes the translation from the longitudinal energy to thermal energy more efficiently .
the incident energies and n - n cross section parameters are illustrated in the inset .
, width=340 ] but for chemical potential . ,
width=340 ] fig .
[ fig_mu - merge ] shows the time evolution of chemical potential ( @xmath85 ) .
again , the left panel displays the normal cross section one , we can find that @xmath85 increases in the compression stage and decreases in the expansion stage , and the lower the incident energy , the larger the chemical potential .
this might be understood as a large compound nucleus is formed during the compression stage , and the lower the incident energy , the larger the compound nucleus is . in panel ( b )
, it shows that the chemical potential becomes generally larger when nucleon - nucleon cross section is larger .
time evolution of entropy density is plotted in fig .
[ fig_sig - merge ] .
it is found that the entropy density almost synchronically evolves with the temperature .
the higher the incident energy and nucleon - nucleon cross section , the larger the entropy density is .
but for entropy density .
, width=340 ] now we can move to the discussion on transport coefficient . since the nuclear participant in central region could be seen as nuclear fluid
, we adopt the eq .
[ dani - eqn ] to calculate the shear viscosity .
unlike the green - kubo formula @xcite , the advantage of the equation is that we can investigate the time evolution of shear viscosity in the framework of transport model .
but it should be noted that eq .
[ dani - eqn ] is principally applicable when the system is largely equilibrated .
however , a full equilibrium is hardly achieved during the whole heavy - ion collision process .
so the shear viscosity extracted here should be seen as the properties of an equilibrated nuclear fireball with the same thermodynamic state as the simulated one .
[ fig_eta - time ] displays the time evolution of shear viscosity ( @xmath1 ) , it shows an increase in earlier stage and then drops with time . as eq .
[ fig_eta - eqn ] shows , here the shear viscosity depends on both temperature and density which vary with time . roughly speaking
, the shear viscosity increases in the compression stage and decreases as the system expands .
the smaller the nucleon - nucleon cross section , the larger the viscosity in the maximum compression stage . fig .
[ fig_eta - time](b ) shows that there is a big enhancement when the n - n cross section is scaled by @xmath86 , which is demonstrated in eq.[dani - eqn - cross ] , the smaller the n - n cross section the larger the viscosity is .
but for shear viscosity @xmath1 of the central fireball .
, width=340 ] when the entropy density is taken into account , the ratio of shear viscosity to entropy density shows a minimum near maximum compression point as shown in fig .
[ fig_ets - merge ] . from the hydrodynamical point of view , the less the @xmath3 , the more perfect the matter looks like . in this sense ,
the nuclear matter becomes a more ideal - like liquid around the most compressible point in comparison with other evolution stages .
but note that this minimum @xmath3 is just a transient process .
in addition , the extent of approaching an ideal - like liquid of the nuclear matter is growing up with the increasing of beam energy . in relativistic energy domain
, the @xmath3 of quark - gluon matter becomes very small , close to 1/4@xmath87 ( kss bound ) , it is called a perfect liquid .
but for the ratio of shear viscosity to entropy density @xmath3 of the central nuclear fireball .
, width=340 ] temperature dependence of @xmath3 is an important issue to understand the transport properties of the nuclear matter in different hot and dense environment . to this end , we plot a correlation between @xmath3 and temperature in fig .
[ fig_ets - tc](a ) at different energies .
note that the density is not fixed in each curve .
it is found that there is a decrease of @xmath3 at first when the system is in the compression stage .
however , @xmath3 becomes increasing as the system begins to expand . the higher the beam energy , the hotter the nuclear matter , and the smaller the @xmath3 .
from this picture it is obvious to find the time when the @xmath3 approaches its transient minimum , essentially corresponds that the nuclear matter reaches the highest temperature . on the other hand , in the present beam energy domain below 400 mev / nucleon ,
the transient minimum of @xmath3 which corresponds to the larger compression stage is around 0.4 , which is about 5 times of kss bound ( i.e. 1/4@xmath87 ) . and
temperature at different beam energies with a normal nucleon - nucleon cross section parametrization ; ( b ) : the @xmath3 evolves versus temperature at 130 mev / nucleon with the different nucleon - nucleon cross section .
the incident energy and nucleon - nucleon cross section is illustrated in the inset of ( a ) and ( b ) , respectively .
, width=340 ] furthermore , we can also extract the correlation between @xmath3 and nuclear matter density as shown in fig .
[ fig_ets - rho ] . here
temperature is another hidden variable .
similar to fig .
[ fig_ets - tc ] , @xmath3 first drops to a minimum value as the density is compressed to a maximum point and then rises up when the system expands .
larger compressible state produces a less @xmath3 , i.e. the system is close to a more ideal - like state .
but for density dependence of @xmath3 .
, width=340 ] considering that only final reaction products can be detected in experiments , such as the multiplicity and flows of the fragments and light particles , it is therefore necessary to check the @xmath3 in the freeze - out stage and see if it is a useful probe to study the properties of nuclear matter as well as liquid gas phase transition . the freeze - out volume has been already studied in some previous works @xcite , but in our case , it is more suitable to define a freeze - out density instead .
the time average values of @xmath3 when the nuclear matter is in some given freeze - out density regions of @xmath75 in @xmath88 , [ 0.24,0.26]$ ] and @xmath89 $ ] have been extracted as a function of temperature . as a function of temperature at different fixed freeze - out densities ( a ) and different cross sections ( b ) .
, width=340 ] fig .
[ fig_ets - energyn ] shows a correlation of the above average @xmath3 versus temperature for given freeze - out densities ( a ) and with different @xmath6 ( b ) . from fig .
[ fig_ets - energyn](a ) we observe that there exhibits a local minimum of @xmath3 with a value of 0.76 to 0.84 ( about 9 - 10 times of kss bound ) , depending on the freeze - out density , in the range of 8 - 12 mev of temperature , this phenomenon shall be related to the liquid gas phase transition . with the increasing of freeze - out density
, we observe the minimal value of @xmath3 decreases and while its corresponding turning temperature increases .
the former is consistent with the results showed in fig .
12 and the latter can be understood by the transition temperature / pressure increases with the freeze - out density as expected by the pressure - density phase diagram @xcite . in contrast with the sensitivity of @xmath3 to freeze - out density , fig .
[ fig_ets - energyn](b ) demonstrates that the time averaged @xmath3 when the system is in a given @xmath75 @xmath90 $ ] is also very sensitive to the @xmath6 .
the larger the n - n cross section is , the smaller the @xmath3 is , which means the nuclear matter behaves much more similar as an ideal fluid . in order to check the result of @xmath3 , another signal of liquid gas phase transition , namely intermediate mass fragment , is also studied .
the intermediate mass fragment which is defined as charge number @xmath91 $ ] , where @xmath92 is the total charge number .
these fragments are larger than typical evaporated light particles and smaller than the residues and fission products , and they can be considered as nuclear fog .
so the multiplicity of intermediate mass fragments ( @xmath93 ) is intimately related with the occurrence of liquid gas phase transition .
usually the @xmath93 increases first as the collision system changes toward gas phase , and reaches a maximum , then decreases when the system becomes vaporized @xcite .
the result of @xmath93 as a function of temperature is showed in fig .
[ fig_imf - end ] . in panel
( a ) , it is interesting to find that the higher the density , the lower the transition temperature is , where the maximum of @xmath93 approaches .
this trend is just coincident with the result of @xmath3 , which showed in fig .
[ fig_ets - energyn](a ) except the exact value of the transition temperature ; in @xmath93 case , the transition temperature @xmath94 $ ] , a little smaller than the @xmath3 s , where @xmath95 $ ] .
the difference could be explained as the nuclear matter is hotter in the central region , so we argue that the minimum of the @xmath3 could be a probe of liquid gas phase transition . in panel
( b ) the average value of @xmath93 as a function of the average temperature when the nuclear matter density @xmath96\rho_{0}$ ] , it is found that the multiplicities of the intermediate mass fragment increase as the n - n cross section is large . this can be understood as large n - n cross section increases the probability of intermediate mass clusters formation .
furthermore , the transition temperature is also increase as the n - n cross section just like the results of fig .
[ fig_ets - energyn](b ) . as a function of temperature at different fixed freeze - out density stages ( a ) ;
the average @xmath93 on the whole freeze out stage with @xmath97 $ ] as a function of temperature at different cross sections at the ( b ) .
, width=340 ] it is interesting to note that phase transition temperature @xmath98 mev which corresponds a local minimum of @xmath3 is basically coincident with previous works @xcite .
thermodynamical and transport properties of a fireball formed in head - on au+au collisions are investigated in a framework of quantum molecular dynamics model .
the relationships between different thermodynamic quantities are explored .
the influences of nucleon - nucleon cross section and symmetry energy on the thermodynamical and transport properties are also focused .
we found that all the properties are very sensitive to the nucleon - nucleon cross section and insensitive to the symmetry energy . in our calculations ,
the shear viscosity is calculated by a parametrization formula developed by danielewicz and entropy density is obtained by a generalized hot thomas fermi formalism .
the present work gives a time evolution of shear viscosity over entropy density ratio of nuclear fireball , which shows that a transient minimal @xmath3 occurs in the largest compression stage .
the results at different beam energies show that the larger the compression , the more ideal the nuclear fireball behaves like fluid . in the present beam energy domain below 400 mev
/ nucleon , this transient @xmath3 approaches to 5 times kss bound .
in addition , temperature and density dependencies of @xmath3 are also investigated .
it is of very interesting to observe that a local @xmath3 minimum , which is about 9 - 20 times kss bound , emerges from the temperature dependence of @xmath3 at different constant freeze - out densities ( 0.2 - 0.3 @xmath42 ) , which corresponds to a liquid - gas phase transition occurring in the intermediate energy heavy - ion collisions.and the larger the n - n cross section , the smaller the @xmath3 is , which means the nuclear matter behaves more like the ideal fluid .
from the temperature dependence of @xmath3 , we learn that the phase transition temperature rises up with the freeze - out density . in order to check the result of @xmath3 , another liquid gas phase transition signal ,
the multiplicity of the intermediate mass fragment is also checked , and a very nice coincidence is found . finally , we like to point out that the present work is still in a phenomenological level for investigating @xmath3 of hot nuclear matter which is formed in intermediate energy heavy - ion collisions , experimental measurements of @xmath3 are still not available so far .
therefore , proposals for direct probes of shear viscosity and entropy density in intermediate energy hic are very crucial and welcome for constraining the transport properties of nuclear matter around the liquid - gas phase transition .
this work is partially supported by the nsfc under contracts no . 11035009 , 11220101005 , 10979074 , 11175231 , the major state basic research development program in china under contract no .
2013cb834405 , and the knowledge innovation project of chinese academy of sciences under grant no .
kjcx2-ew - n01 .
n. demir , s. a. bass , phys .
lett . * 102 * , 172302 ( 2009 ) .
et al . _ ,
lett . * 98 * , 092301 ( 2007 ) .
jiunn - wei chen , eiji nakano , phys . lett .
b * 647 * , 371 ( 2007 ) .
j. i. kapusta and t. springer , phys .
d * 78 * , 066017 ( 2008 ) .
a. majumder , b. mller , and xin - nian wang , phys .
lett . * 99 * , 192301 ( 2007 ) . l. shi and p. danielewicz , phys
c * 68 * , 064604 ( 2003 ) .
s. pal , phys . lett .
b * 684 * , 211 ( 2010 ) ; phys .
c * 81 * , 051601(r ) ( 2010 ) .
n. auerbach , s. shlomo , phys .
103 * , 172501 ( 2009 ) .
d. t. khoa , n. ohtsuka , a. faessler , m. a. matin , s. w. huang , e. lehmann and y. lotfy , nucl . phys . * a 542 * , 671 ( 1992 ) .
d. t. khoa , n. ohtsuka , a. faessler , m. a. matin , s. w. huang , e. lehmann and p. k. puri , nucl . phys . * a 548 * , 102 ( 1992 ) .
y. g. ma , w. q. shen , phys .
c * 51 * , 710 ( 1995 ) .
g. f. peaslee _
c * 49 * , r2271 ( 1994 ) .
c. a. ogilvie _ et al .
67 * , 1214 ( 1991 ) .
m. b. tsang _
et al . _ ,
lett . * 71 * , 1502 ( 1993 ) . | thermodynamic and transport properties of nuclear fireball created in the central region of heavy - ion collisions below 400 mev / nucleon are investigated within the isospin - dependent quantum molecular dynamic ( iqmd ) model .
these properties including the density , temperature , chemical potential , entropy density ( @xmath0 ) and shear viscosity ( @xmath1 ) , are calculated by a generalized hot thomas fermi formulism and a parameterized function , which was developed by danielewicz . as the collision goes on ,
a transient minimal @xmath2 occurs in the largest compression stage .
besides , the relationship of @xmath3 to temperature ( @xmath4 ) in the freeze - out stage displays a local minimum which is about 9 - 20 times @xmath5 around @xmath4 = 8 - 12 mev , which can be argued as indicative of a liquid gas phase transition .
in addition , the influences of nucleon - nucleon ( nn ) cross section ( @xmath6 ) and symmetry energy coefficient ( @xmath7 ) are also discussed , and it is found that the results are sensitive to @xmath6 but not to @xmath7 . |
a pseudoscalar gluonium candidate , the so - called @xmath15 , was observed in @xmath16 annihilation in 1967 @xcite and in @xmath3 radiative decays in the 1980 s @xcite .
after 1990 , more and more observations revealed the existence of two resonant structures around 1.44 gev/@xmath7 in @xmath17 , @xmath8 , and @xmath18 spectra @xcite .
they showed that the lower state , @xmath19 , has large couplings to @xmath17 and @xmath8 , while the high mass state , @xmath20 , favors @xmath18 .
the @xmath19 was also confirmed by markiii @xcite , crystal barrel @xcite , and dm2 @xcite in decays into @xmath21 in @xmath3 radiative decays and @xmath22 annihilations .
in contrast , although @xmath20 was observed in @xmath8 ( @xmath18 ) @xcite , it was not seen in @xmath21 . nonetheless , the study of @xmath8 and @xmath21 channels in @xmath23 collisions @xcite showed that @xmath20 appeared in @xmath8 , but not in @xmath21 , while @xmath19 appeared in neither channel .
the study of the decays @xmath24 is a useful tool in the investigation of quark and possible gluonium content of the states around 1.44 gev/@xmath7 . in this paper
, we investigate the possible structure in the @xmath8 final state in @xmath3 hadronic decays at around @xmath6 gev/@xmath7 , and measure the branching fraction of @xmath25 for the first time , based on @xmath2 @xmath3 events collected with the beijing spectrometer at the beijing electron - positron collider ( bepc ) .
besii is a large solid - angle magnetic spectrometer that is described in detail elsewhere @xcite .
charged particle momenta are determined with a resolution of @xmath26 ( @xmath27 in gev/@xmath7 ) in a 40-layer cylindrical main drift chamber ( mdc ) .
particle identification ( pid ) is accomplished using specific ionization ( @xmath28 ) measurement in the mdc and time - of - flight ( tof ) information in a barrel - like array of 48 scintillation counters .
the @xmath28 resolution is @xmath29 ; the tof resolution is @xmath30 ps for bhabha events .
outside of the time - of - flight counter is a 12-radiation - length barrel shower counter ( bsc ) comprised of gas proportional tubes interleaved with lead sheets .
the bsc measures the energy and direction of photons with resolutions of @xmath31 ( @xmath32 in gev ) , @xmath33 mrad , and @xmath34 cm .
the iron flux return of the magnet is instrumented with three double layers of counters that are used to identify mouns .
a geant3 based monte carlo ( mc ) package ( simbes ) with detailed consideration of the detector performance is used .
the consistency between data and mc has been carefully checked in many high purity physics channels , and the agreement is reasonable @xcite .
the detection efficiencies and mass resolutions for each decay mode presented in this paper are obtained with uniform phase space mc generators .
in this analysis , @xmath35 mesons are observed in the @xmath36 decay , @xmath37 mesons in the @xmath38 decay , and other mesons are detected in the decays : @xmath39 , @xmath40 , @xmath41 .
the final states of the analyzed decays @xmath42 , @xmath43 , @xmath44 , and @xmath45 are @xmath46 , @xmath47 , @xmath48 , and @xmath49 , respectively .
+ candidate events are required to satisfy the following common selection criteria : 1 .
the correct number of charged tracks with net charge zero is required for each event .
each charged track should have a good helix fit in the mdc , and the polar angle @xmath50 of each track in the mdc must satisfy @xmath51 .
the event must originate from the collision point ; tracks except @xmath52 from @xmath53 must satisfy @xmath54 cm , @xmath55 cm , where @xmath56 , @xmath57 , and @xmath58 are the space coordinates of the point of closest approach of tracks to the beam axis .
2 . candidate events should have at least the minimum number of isolated photons associated with the different final states .
each photon should have an energy deposit in the bsc greater than @xmath59 mev , the angle between the shower development direction and the photon emission direction less than @xmath60 , and the angle between the photon and any charged track larger than @xmath61 .
each charged track in an event , @xmath62 is determined using both @xmath28 and tof information : @xmath63 where @xmath64 corresponds to the particle hypothesis .
a charged track is identified as a @xmath65 ( @xmath66 ) if @xmath67 for the @xmath65 ( @xmath66 ) hypothesis is less than those for the @xmath66 and @xmath27 ( @xmath65 and @xmath27 ) hypotheses .
the selected events are subjected to four constraint kinematic fits ( 4c - fit ) , unless otherwise specified . when there are more than the minimum number of photons in an event , all combinations are tried , and the combination with the smallest @xmath68 is retained .
the branching fraction is calculated using @xmath69 and the upper limit for a branching fraction is determined using @xmath70 where , @xmath71 is the number of events observed , @xmath72 is the upper limit on the number of the observed events at the @xmath73 c.l . calculated using a bayesian method @xcite , @xmath74 is the detection efficiency obtained from mc simulation , @xmath75 is the number of @xmath3 events , @xmath76 @xcite , @xmath77 is the corresponding systematic error , and @xmath78 is the branching fraction , taken from the particle data group ( pdg ) @xcite , of the @xmath79 intermediate state to the @xmath80 final state .
at least one charged track must be identified as a kaon using tof and @xmath28 information .
if there is more than one kaon candidate , the assigned kaon is the one with the largest kaon weight .
candidate events are fitted kinematically using energy momentum conservation ( 4c - fit ) under the @xmath82 hypothesis , and @xmath83 is required .
each event is required to contain one @xmath53 meson with six possible @xmath84 combinations to test for consistency with the @xmath53 .
looping over all combinations , we select the one closest to the @xmath53 mass , denoted as @xmath85 , provided it is within @xmath86 mev/@xmath7 of the @xmath53 mass .
after @xmath53 selection , the two remaining oppositely - charged pion combinations along with the two gammas are used to calculate @xmath87 .
figure [ fig : w - pi0 ] ( a ) shows the scatter plot of @xmath88 versus @xmath87 with two possible entries per event , where clear @xmath89 and @xmath35 signals are seen . versus @xmath87 , and ( b ) the @xmath90 invariant mass for @xmath91 candidate events with two possible entries per event .
the curves in ( b ) are the results of the fit described in the text , and the shaded histogram in ( b ) shows the normalized background estimated from the @xmath53-sideband region ( @xmath92 gev/@xmath93 gev/@xmath7).,title="fig:",scaledwidth=35.0% ] versus @xmath87 , and ( b ) the @xmath90 invariant mass for @xmath91 candidate events with two possible entries per event .
the curves in ( b ) are the results of the fit described in the text , and the shaded histogram in ( b ) shows the normalized background estimated from the @xmath53-sideband region ( @xmath92 gev/@xmath93 gev/@xmath7).,title="fig:",scaledwidth=35.0% ] the @xmath94 invariant mass distribution with two possible entries per event is shown in fig .
[ fig : w - pi0 ] ( b ) , where @xmath89 and @xmath35 signals are apparent .
the branching fractions of @xmath95 and @xmath96 are obtained by fitting this distribution .
the backgrounds for @xmath97 which contribute to the peak in the @xmath35 signal region mainly come from non-@xmath53 events and events from @xmath98 that survive selection criteria .
the number of background events from @xmath98 is estimated from monte - carlo simulation to be less than 2 .
backgrounds for @xmath99 contributing to the peak in the @xmath89 signal region mainly come from non-@xmath53 events and events from @xmath3 decays into @xmath100 .
the latter contribution is estimated to be less than one event from mc simulation .
non-@xmath53 events from the @xmath53 sideband region ( 0.025 gev / c@xmath101 @xmath102 gev / c@xmath103 ) are shown in fig .
[ fig : w - pi0 ] ( b ) as the shaded histogram , the background events are @xmath104 @xmath35 and @xmath105 @xmath89 by fitting the distribution with possible signals and polynomial background .
a fit to the @xmath106 distribution is performed by using @xmath35 and @xmath89 breit - wigner ( bw ) functions folded with gaussian resolution functions plus a quadratic polynomial , shown as the curve in fig .
[ fig : w - pi0 ] ( b ) .
the numbers of events in the @xmath35 and @xmath89 peaks are @xmath107 and @xmath108 , respectively . here ,
the background events in the decays of @xmath95 and @xmath109 estimated above are not subtracted but are included in the background systematic error .
the @xmath110 and @xmath111 detection efficiencies are obtained from mc simulation to be @xmath112 and @xmath113 , respectively .
the branching fractions are then determined as : @xmath114 @xmath115 here the errors are statistical only . to select the @xmath35 signal , the mass combination with @xmath117 closest to the @xmath35 mass is required to satisfy @xmath118 gev/@xmath7 .
figure [ fig : ksp - kp - wkksp ] shows the scatter plot of @xmath119 versus @xmath120 for @xmath110 candidate events , where the events in the cross bands correspond to @xmath121 .
versus @xmath122 for @xmath110 candidate events.,scaledwidth=35.0% ] figure [ fig : k892-wkksp ] ( a ) shows the scatter plot of @xmath87 versus @xmath85 , and there is an accumulation of events in the @xmath35 and @xmath53 cross bands . the combined mass spectrum of @xmath123 and @xmath124 in the signal region ( box 1 in fig .
[ fig : k892-wkksp ] ( a ) ) , which is defined as @xmath125 gev/@xmath7 and @xmath118 gev/@xmath7 , is shown in fig .
[ fig : k892-wkksp ] ( b ) , where a clear @xmath126 signal is observed .
the @xmath126 signal is fitted with a bw function folded with a gaussian resolution function plus a third - order polynomial , and @xmath127 @xmath126 events are obtained .
non-@xmath35 and non-@xmath53 backgrounds are studied using @xmath35 and @xmath53 sideband events .
figure [ fig : k892-wkksp ] ( c ) is the fitted @xmath128 mass spectrum in the @xmath35 sideband region ( @xmath125 gev/@xmath7 , @xmath129 gev@xmath130 gev/@xmath7 , shown as horizontal sideband boxes @xmath131 in fig .
[ fig : k892-wkksp ] ( a ) ) and @xmath53 sideband region ( @xmath132 gev/@xmath133 gev/@xmath7 , @xmath118 gev/@xmath7 , shown as vertical sideband boxes @xmath134 ) , and the number of @xmath126 sideband events @xmath135 is obtained .
figure [ fig : k892-wkksp ] ( d ) is background from the corner region ( @xmath132 gev/@xmath136 gev/@xmath7 , @xmath129 gev/@xmath137 gev/@xmath7 , shown as diagonal boxes @xmath138 ) , and the number of @xmath126 events @xmath139 is equal to @xmath140 .
the number of background events in the signal region is half of the sum of @xmath126 events in the @xmath35 sideband and @xmath53 sideband regions ( @xmath141 ) minus a quarter of the @xmath126 events in the corner regions ( @xmath139 ) .
so @xmath142 . versus @xmath85 , and the combined mass spectrum of @xmath143 and @xmath124 with two entries per event @xmath121 candidate events for ( b ) the signal region ( the central box 1 ) ; ( c ) the @xmath35 and @xmath53 sideband regions ( two horizontal boxes 2 and two vertical sideband boxes 3 ) ; and for ( d ) the corner region ( four diagonal boxes 4 ) .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=35.0% ] versus @xmath85 , and the combined mass spectrum of @xmath143 and @xmath124 with two entries per event @xmath121 candidate events for ( b ) the signal region ( the central box 1 ) ; ( c ) the @xmath35 and @xmath53 sideband regions ( two horizontal boxes 2 and two vertical sideband boxes 3 ) ; and for ( d ) the corner region ( four diagonal boxes 4 ) .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=35.0% ] versus @xmath85 , and the combined mass spectrum of @xmath143 and @xmath124 with two entries per event @xmath121 candidate events for ( b ) the signal region ( the central box 1 ) ; ( c ) the @xmath35 and @xmath53 sideband regions ( two horizontal boxes 2 and two vertical sideband boxes 3 ) ; and for ( d ) the corner region ( four diagonal boxes 4 ) .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=35.0% ] versus @xmath85 , and the combined mass spectrum of @xmath143 and @xmath124 with two entries per event @xmath121 candidate events for ( b ) the signal region ( the central box 1 ) ; ( c ) the @xmath35 and @xmath53 sideband regions ( two horizontal boxes 2 and two vertical sideband boxes 3 ) ; and for ( d ) the corner region ( four diagonal boxes 4 ) .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=35.0% ] the detection efficiency is estimated to be @xmath144 from mc simulation .
after background subtraction , the branching fraction is determined to be @xmath145 where the error is statistical only .
figure [ fig : w - x1440-recoiling ] ( a ) shows the scatter plot of @xmath147 versus @xmath87 , and fig .
[ fig : w - x1440-recoiling ] ( b ) is the @xmath4 invariant mass spectrum after @xmath35 selection ( @xmath118 gev / c@xmath103 ) . figs .
[ fig : w - x1440-recoiling ] ( a ) and ( b ) show a resonance near @xmath6 gev/@xmath7 , denoted as @xmath148 . to ensure that this peak is not due to background , we have made studies of potential background processes using both data and mc simulations . non-@xmath35 and
non-@xmath53 processes are studied with @xmath35 and @xmath149 mass sideband events , respectively .
the main background channel @xmath150 and other background processes with 6-prong events are studied by mc simulation .
in addition , we also checked for possible backgrounds with a mc sample of @xmath151 decays generated by the lund - charm model @xcite .
none of these background sources produces a peak around @xmath6 gev/@xmath7 in the @xmath4 invariant mass spectrum . versus @xmath87 and ( b ) the @xmath152 invariant mass distribution for @xmath153 candidate events .
the curves in ( b ) are the results of the fit described in the text .
, title="fig:",scaledwidth=35.0% ] versus @xmath87 and ( b ) the @xmath152 invariant mass distribution for @xmath153 candidate events .
the curves in ( b ) are the results of the fit described in the text .
, title="fig:",scaledwidth=35.0% ] the @xmath4 invariant mass distribution is fitted with a bw function convoluted with a gaussian mass resolution function ( @xmath154 mev/@xmath7 ) to represent the @xmath148 signal and a third - order polynomial background function .
the mass and width obtained from the fit are @xmath155 mev/@xmath7 and @xmath156 mev/@xmath7 , and the fit yields @xmath157 events . using the efficiency of @xmath158 determined from a uniform phase space mc simulation
, we obtain the branching fraction to be @xmath159 where the error is only the statistical error .
at least one charged track is required to be a kaon and the combined pid probability for @xmath161 is required to be greater than those for the @xmath162 and @xmath163 hypotheses .
a 4c kinematic fit is made under the @xmath164 hypothesis .
there are three combinations to form two @xmath165 s , and further a six - constraint kinematic fit ( 6c - fit ) with the smallest @xmath166 is made requiring two @xmath165 s from four photons .
events with @xmath167 and @xmath168 are selected . to reject the possible multiple photon background events , @xmath169
is required to be less than those for the @xmath170 , @xmath171 , and @xmath172 hypotheses .
background events with @xmath53 decays , such as @xmath173 , and @xmath174 , are eliminated by requiring @xmath175 gev/@xmath7 in the @xmath84 invariant mass .
there are two @xmath176 mass combinations , and the one closest to the @xmath35 mass , denoted as @xmath177 , is selected .
the scatter plot of @xmath178 versus @xmath177 is shown in fig .
[ fig : w - x1440-wkkp0 ] ( a ) , where the circle indicates some enhancement from @xmath179 events in the @xmath43 decay . versus @xmath177 , ( b ) the @xmath180 invariant mass distribution with two possible entries per event , and ( c ) the @xmath181 invariant mass distribution for @xmath182 candidate events .
the curves are the results of the fit described in the text , and the shaded histogram ( b ) shows the normalized background estimated from the @xmath35-sideband region.,title="fig:",scaledwidth=30.0% ] versus @xmath177 , ( b ) the @xmath180 invariant mass distribution with two possible entries per event , and ( c ) the @xmath181 invariant mass distribution for @xmath182 candidate events .
the curves are the results of the fit described in the text , and the shaded histogram ( b ) shows the normalized background estimated from the @xmath35-sideband region.,title="fig:",scaledwidth=30.0% ] versus @xmath177 , ( b ) the @xmath180 invariant mass distribution with two possible entries per event , and ( c ) the @xmath181 invariant mass distribution for @xmath182 candidate events .
the curves are the results of the fit described in the text , and the shaded histogram ( b ) shows the normalized background estimated from the @xmath35-sideband region.,title="fig:",scaledwidth=30.0% ] to suppress the main @xmath184 backgrounds , @xmath185 gev/@xmath7 is required .
in addition to the above selection , the further requirement of @xmath186 gev/@xmath7 is imposed .
the combined mass spectrum of @xmath187 and @xmath188 is shown in fig .
[ fig : w - x1440-wkkp0 ] ( b ) , where the @xmath189 signal is seen , and is fitted to obtain the branching fraction of @xmath190
. background events for @xmath191 which could contribute to the peak in the @xmath189 signal region mainly come from events with @xmath126 decays , such as @xmath192 into 4-prong plus multiple photons , @xmath193 , and @xmath194 , but their contributions can be ignored according to mc studies . it is further confirmed that the @xmath195 background is negligible using @xmath35 and @xmath165 sideband events . the @xmath196 invariant mass distribution in fig .
[ fig : w - x1440-wkkp0 ] ( b ) ( 2 entries / event ) is fitted with a @xmath197 bw function with the mass and width fixed to pdg values @xcite plus a third - order polynomial .
the number of @xmath189 events obtained is @xmath198 .
the detection efficiency is @xmath199 , and the branching fraction of @xmath200 is determined to be @xmath201 where the error is statistical only .
figure [ fig : w - x1440-wkkp0 ] ( c ) shows the @xmath5 invariant mass recoiling against the @xmath35 , where a @xmath148 signal is observed .
we have also studied potential background processes using both data and mc simulations .
non-@xmath35 processes are studied with the @xmath35 mass sideband events ( @xmath129 gev/@xmath203 gev/@xmath7 ) .
background with @xmath35 decays is studied by mc simulations , similar to those of @xmath204 .
in addition , we also checked for possible backgrounds using a mc sample of @xmath151 decays generated by the lund - charm model . in each case , the @xmath5 mass distribution shows no evidence of an enhancement near @xmath205 mev/@xmath7 . by fitting the @xmath206 mass spectrum in fig .
[ fig : w - x1440-wkkp0 ] ( c ) with a bw function convoluted with a gaussian mass resolution function ( @xmath207 mev/@xmath7 ) plus a third - order polynomial background function , the mass and width of @xmath208 mev/@xmath7 and @xmath209 mev/@xmath7 are obtained , and the number of events from the fit is @xmath210 .
a fit without a bw signal function returns a value of @xmath211 larger than the nominal fit by 31.7 with three degrees of freedom ( d.o.f . ) , corresponding to a statistical significance of @xmath212 for the signal .
the efficiency is determined to be @xmath213 from a phase space mc simulation , and the branching fraction is + @xmath214 where the error is statistical .
events with six charged tracks are selected , and at least two charged tracks must be identified as kaons . if there are more than two kaons , the two kaons with the largest kaon pid probabilities are regarded as the real kaons .
the other charged tracks are assumed , one at a time , to be a kaon , while the other three to be pions , and these combinations of three kaons and three pions are kinematically fitted .
the hypothesis with the smallest @xmath68 is considered as the right combination , and @xmath216 is required .
two combinations of oppositely charged pions are used to reconstruct the @xmath53 signal , and the one closest to the @xmath53 mass is required to be within @xmath86 mev/@xmath7 .
the invariant mass of the two mass combinations formed with oppositely charged kaons are shown in fig .
[ fig : phi - phikksp ] , where a clear @xmath37 signal is observed . a fit to the @xmath217 mass distribution in fig [ fig : phi - phikksp ] is performed to obtain the number of @xmath218 events .
backgrounds contributing to the @xmath37 signal peak mainly come from @xmath3 into @xmath219 , @xmath220 , @xmath221 , and @xmath222 ( excluding @xmath221 ) . from mc simulations of these background channels ,
the number of background @xmath37 events in the signal region is less than one , and @xmath53-sideband events also show that the background is negligible .
invariant mass distribution for @xmath223 candidate events with two possible entries per event .
the curves are the results of the fit described in the text.,scaledwidth=35.0% ] the @xmath224 mass distribution in fig .
[ fig : phi - phikksp ] is fitted with a bw function convoluted with a gaussian mass resolution function ( @xmath225 mev/@xmath7 ) plus a third - order polynomial background function .
the number of @xmath37 events from the fit is @xmath226 . using the detection efficiency of @xmath227 ,
the corresponding branching fraction is @xmath228 where the error is statistical only . to remove most non-@xmath37 background events , the @xmath217 combination closest to the @xmath37 mass is required to satisfy @xmath230 gev / c@xmath103 . the scatter plot of @xmath119 versus @xmath120 for candidate events is shown in fig .
[ fig : k892-phik892k - kksp ] ( a ) , where the events in the cross band correspond to the @xmath126 signal .
the scatter plot of @xmath85 versus @xmath231 is shown in figure [ fig : k892-phik892k - kksp ] ( b ) , and there is an accumulation of events in the @xmath37 and @xmath53 cross bands .
figure [ fig : k892-phik892k - kksp ] ( c ) shows the combined @xmath123 and @xmath124 mass spectrum for events in the signal region ( box 1 in fig .
[ fig : k892-phik892k - kksp ] ( b ) ) , which is defined as @xmath125 gev/@xmath7 and @xmath230 gev/@xmath7 .
a fit yields @xmath232 @xmath126 events .
the same method as used for the @xmath233 analysis is used to estimate the number of background events in the signal region , and @xmath234 , which is neglected in the branching fraction determination . versus @xmath120 , ( b ) the scatter plot of @xmath85 versus @xmath231 , and ( c ) the combined @xmath123 and @xmath124 invariant mass distributions for events in the signal region ( box 1 ) for @xmath215 candidate events .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=30.0% ] versus @xmath120 , ( b ) the scatter plot of @xmath85 versus @xmath231 , and ( c ) the combined @xmath123 and @xmath124 invariant mass distributions for events in the signal region ( box 1 ) for @xmath215 candidate events .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=30.0% ] versus @xmath120 , ( b ) the scatter plot of @xmath85 versus @xmath231 , and ( c ) the combined @xmath123 and @xmath124 invariant mass distributions for events in the signal region ( box 1 ) for @xmath215 candidate events .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=30.0% ] the detection efficiency of @xmath235 in this decay is @xmath236 , and its branching fraction is determined to be @xmath237 where the error is statistical only . the distribution of @xmath239 invariant mass recoiling against the @xmath37 signal is shown in fig .
[ fig : x1440-phikksp ] ( a ) , and there is no evidence for @xmath148 . the upper limit on the number of the observed events at the @xmath73 c.l .
is @xmath240 @xcite .
the likelihood distribution and the @xmath73 c.l .
limit are shown in fig .
[ fig : x1440-phikksp ] ( b ) .
the likelihood values for the number of events are obtained by fitting the @xmath239 distributions with a x(1440 ) signal plus a third - order polynomial background .
its mass and width are fixed to those in the decay @xmath241 .
invariant mass recoiling against the @xmath37 , and ( b ) the number of events of x(1440 ) .
the curve in ( a ) is a third order polynomial to describe the background , and the observed number of events at the @xmath73 confidence level using a bayesian method is indicated by the arrow in ( b).,title="fig:",scaledwidth=35.0% ] invariant mass recoiling against the @xmath37 , and ( b ) the number of events of x(1440 ) .
the curve in ( a ) is a third order polynomial to describe the background , and the observed number of events at the @xmath73 confidence level using a bayesian method is indicated by the arrow in ( b).,title="fig:",scaledwidth=35.0% ] the detection efficiency is @xmath242 , and the upper limit on the branching fraction at the @xmath73 c.l . is : @xmath243 at least three charged tracks must be identified as kaons .
a 4c - fit is applied under the hypothesis @xmath245 , and @xmath246 is required . to reject possible background events from @xmath247 , the @xmath68 of the 4c fit for @xmath248 is required to be less than that for the @xmath249 hypothesis .
there are four possible ways to combine the oppositely charged kaons in forming the @xmath37 , and the @xmath217 combination closest to the @xmath37 mass is chosen .
figure [ fig : scatter - phik892k - phi - pi0 ] ( a ) shows the scatter plot of @xmath88 versus @xmath231 , and clear @xmath37 and @xmath165 signals are seen . versus @xmath231 , and ( b ) the @xmath180 invariant mass distribution for events in the signal region ( box 1 ) for @xmath250 candidate events with two entries per event .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=35.0% ] versus @xmath231 , and ( b ) the @xmath180 invariant mass distribution for events in the signal region ( box 1 ) for @xmath250 candidate events with two entries per event .
the curves are the results of the fit described in the text.,title="fig:",scaledwidth=35.0% ] figure [ fig : scatter - phik892k - phi - pi0 ] ( b ) shows the @xmath180 combined mass spectrum for events in the signal region ( box 1 in fig .
[ fig : scatter - phik892k - phi - pi0 ] ( a ) ) , which is defined as @xmath230 gev/@xmath7 and @xmath252 gev/@xmath7 , and a clear @xmath189 signal is seen .
it is fitted with a bw , whose mass and width are fixed to those of @xmath189 in the pdg , plus a third - order polynomial .
the number of @xmath189 events from the fit is @xmath253 .
the sidebands are used as before to estimate the number of the corresponding background events in the signal region , and the result is @xmath254 . after subtracting the above background and incorporating the efficiency of @xmath255 from mc simulation
, the branching fraction is determined to be @xmath256 where the error is statistical only . the distribution of @xmath258 invariant mass recoiling against the @xmath37 is shown in fig .
[ fig : x1440-phikkp0 ] ( a ) .
no evidence for the @xmath148 is observed near @xmath205 mev/@xmath7 .
the upper limit on the number of the observed events at the @xmath73 c.l . is @xmath259 @xcite .
the likelihood distribution and the @xmath73 c.l .
limit are shown in fig .
[ fig : x1440-phikkp0 ] ( b ) .
the likelihood values for the number of events are obtained by fitting the @xmath258 distributions with a x(1440 ) signal , whose mass and width are fixed to those of the decay @xmath260 , plus a third - order background polynomial .
the detection efficiency is @xmath261 , and the upper limit on the branching fraction at the @xmath73 c.l .
is : @xmath262 invariant mass recoiling against the @xmath37 , and ( b ) the number of events of x(1440 ) .
the curve in ( a ) is the third - order polynomial to describe the background , and the observed number of events at the @xmath73 confidence level using a bayesian method is indicated by arrow in ( b).,title="fig:",scaledwidth=35.0% ] invariant mass recoiling against the @xmath37 , and ( b ) the number of events of x(1440 ) .
the curve in ( a ) is the third - order polynomial to describe the background , and the observed number of events at the @xmath73 confidence level using a bayesian method is indicated by arrow in ( b).,title="fig:",scaledwidth=35.0% ]
in this analysis , the systematic errors on the branching fractions mainly come from following sources : * mdc tracking efficiency + the mdc tracking efficiency is measured from clean channels like @xmath263 and @xmath264 with @xmath265 .
it is found that the mc simulation agrees with data within @xmath266 for each charged track .
therefore , @xmath267 is taken as the systematic error on the tracking efficiency for the channels with six charged tracks and @xmath268 for the channels with four charged tracks in the final states . + * photon detection efficiency + the photon detection efficiency is studied from @xmath269 events @xcite .
the results indicate that the difference between data and mc simulation is less than @xmath270 for each photon .
therefore , @xmath271 is taken to be the systematic error from the photon efficiency for the channels with two photons and @xmath268 for the channels with four photons in the final states . + * pid + the pid efficiency of the kaon is studied with @xmath272 events .
the average pid efficiency difference between data and mc is found to be less than @xmath270 . in this paper , @xmath270 , @xmath271 , and @xmath273
are conservatively taken as the systematic errors on pid efficiency for the channels with one , two , and three identified kaons , respectively . +
* @xmath53 reconstruction + the @xmath53 secondary vertex reconstruction is checked using @xmath274 events .
it is found that the difference of the efficiency between data and mc simulation is @xmath275 , which is taken to be the systematic error from the @xmath53 secondary vertex reconstruction . +
* intermediate decay branching fractions + the branching fractions for @xmath276 , @xmath36 , and @xmath277 are taken from the pdg @xcite , and the errors on these branching fractions are included as systematic errors in our measurements .
the error on the @xmath278 branching fraction is neglected in this analysis . +
* kinematic fit + kinematic fits are used to reduce backgrounds . using the same method as in ref .
@xcite , the decay modes @xmath279 , @xmath280 , and @xmath281 are studied @xcite in order to estimate the corresponding systematic error .
the kinematic fit efficiency differences between data and mc are @xmath282 , @xmath283 , and @xmath284 , respectively .
the efficiency difference between data and mc for the 6c - kinematic fit to @xmath285 is about @xmath286 @xcite .
since the decays in this analysis are similar to the above decays , these systematic errors are taken as the corresponding systematic errors .
+ * background uncertainty + the background uncertainties come from the uncertainties associated with the estimation of the sideband backgrounds , the events from other background channels , as well as the uncertainties of background shapes , different fit ranges , and different binning .
therefore , the statistical error in the estimated number of background events , the largest difference from changing background shape , the difference from changing the fit range , the difference of changing fit binning , and some ignored backgrounds events are taken as the systematic errors due to the background uncertainty . + * mc generator + there may be interference between charged and neutral @xmath126 modes , which is not included in the mc generator . in the sample of @xmath287
decays , the @xmath288 mass distribution in the @xmath189 sideband and signal regions and the @xmath289 mass distributions in the @xmath184 sideband and signal regions from the scatter plot of @xmath122 versus @xmath290 , are studied in real data and mc simulation .
it is found that the difference between data and mc sample is @xmath291 , so @xmath291 is taken as the systematic error from the mc model . +
* number of @xmath3 events + the number of @xmath3 is @xmath292 , determined from @xmath3 inclusive four - prong events @xcite .
the uncertainty is taken as a systematic error in the branching ratio measurement .
the bes collaboration thanks the staff of bepc and computing center for their hard efforts .
this work is supported in part by the national natural science foundation of china under contracts nos .
10491300 , 10225524 , 10225525 , 10425523 , 10625524 , 10521003 , the chinese academy of sciences under contract no .
kj 95t-03 , the 100 talents program of cas under contract nos .
u-11 , u-24 , u-25 , and the knowledge innovation project of cas under contract nos .
u-602 , u-34 ( ihep ) , the national natural science foundation of china under contract no .
10225522 ( tsinghua university ) , and the department of energy under contract no .
de - fg02 - 04er41291 ( u. hawaii ) .
99 p. h. baillon _
_ , nc 50a , 393 ( 1967 ) .
d. l. scharre _
et al . _ ,
b * 97 * , 329 ( 1980 ) . c. edwards _ et al .
lett . * 49 * , 259 ( 1982 ) ; phys . rev . lett .
* 50 * , 219 ( 1982 ) .
j. e. augustin _ et al . _
( dm2 collaboration ) , phys .
d * 42 * , 10 ( 1990 ) .
m. g. rath _
d * 40 * , 693 ( 1989 ) .
t. adams _ et al .
lett . * 87 * , 041801 ( 2001 ) .
( markiii collaboration ) , phys .
lett . * 65 * , 2507 ( 1990 ) .
j. e. augustin _ et al . _
( dm2 collaboration ) , phys .
d * 46 * , 1951 ( 1992 ) .
a bertin _ et al . _
( obelix collaboration ) , phys .
b * 361 * , 187 ( 1997 ) ; phys .
b * 400 * , 226 ( 1997 ) . c. cicalo _ et al . _
( obelix collaboration ) , phys .
b * 462 * , 453 ( 1999 ) .
f. nichitiu _
( obelix collaboration ) , phys .
b * 545 * , 261 ( 2002 ) .
t. bolton _
* 69 * , 1328 ( 1992 ) . c. amsler
_ et al . _
( crystal barrel collaboration ) , phys .
b * 358 * , 389 ( 1995 ) .
m. acciarri _
( l3 collaboration ) , phys .
b * 501 * , 1 ( 2001 ) .
j. z. bai _ et al .
_ ( bes collaboration ) , nucl .
instrum . and meth . a * 458 * , 627 ( 2001 ) .
m. ablikim _
( bes collaboration ) , nucl .
instrum . and meth . a * 552 * , 344 ( 2005 )
. w. m. yao _
( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) . s. s. fang _
_ , high energy phys .
* 27 * , 277 ( 2003 ) .
j. c. chen _ et al .
d * 62 * , 034003 ( 2000 ) .
_ , high energy phys
. nucl . phys .
* 28 * , 859 ( 2004 ) .
j. z. bai _ et al . _
( bes collaboration ) , phys .
d * 70 * , 012005 ( 2004 ) .
m. ablikim _
et al . _ ( bes collaboration ) , phys .
b * 610 * , 192 ( 2005 ) .
m. ablikim _
( bes collaboration ) , phys .
d * 72 * , 072005 ( 2005 ) .
j. becker _
( markiii collaboration ) , phys .
lett . * 59 * , 186 ( 1987 ) .
a. falvard _
( dm2 collaboration ) , phys .
d * 38 * , 2706 ( 1988 ) . | the decays of @xmath0 and @xmath1 are studied using @xmath2 @xmath3 events collected with the beijing spectrometer ( besii ) at the beijing electron - positron collider ( bepc ) .
the @xmath4 and @xmath5 systems , produced in @xmath0 , have enhancements in the invariant mass distributions at around @xmath6 gev/@xmath7 . however , there is no evidence for mass enhancements in the @xmath8 system in @xmath9 .
the branching fractions of @xmath10 , @xmath11 , @xmath12 , and @xmath13 are obtained , and the @xmath14 branching fraction is measured for the first time . plus 1pt minus 1pt |
quasipositive braids are products of conjugates of a positive standard generater of the braid group ; they generalize the class of positive braids .
a quasipositive knot is a knot which is isotopic to the closure of a quasipositive braid .
quasipositive knots emerged from the theory of complex plane curves ( @xcite ) . from certain viewpoints ,
quasipositive knots are ` dense ' in the set of knots .
for instance , every alexander polynomial can be realized by a quasipositive knot ( @xcite ) .
a similar statement is true for vassiliev invariants ( @xcite ) . in this paper
, we prove another density result for quasipositive knots , involving unoriented cobordisms .
a cobordism between two disjoint links @xmath0 , @xmath1 is a smooth embedded surface @xmath2 whose boundary is precisely @xmath3 .
we distinguish between oriented and unoriented cobordisms , depending on whether the surface and its boundary components are oriented or not . in this paper
, we consider particular unoriented cobordisms , namely punctured mbius bands embedded in @xmath4 . for every knot @xmath5
there exists a quasipositive fibred knot @xmath6 and an embedded punctured mbius band @xmath7 whose boundary components are precisely @xmath8 and @xmath9 .
let @xmath10 be an element of the braid group on @xmath11 strings .
we may assume that the first letter of @xmath12 is a negative standard generator of @xmath13 , since we can always insert a canceling pair @xmath14 in front of @xmath12 .
therefore , there exist a natural number @xmath15 , natural numbers @xmath16 , and positive ( possibly empty ) words @xmath17 , such that @xmath18 let @xmath8 be an arbitrary knot .
@xmath8 is isotopic to the closure of a braid @xmath19 , written as in ( [ beta ] ) .
the closure of @xmath12 is depicted in figure 1 .
@xmath22 is clearly a quasipositive braid , since all its negative letters appear in conjugating pairs , together with their positive counterparts : @xmath23 let @xmath9 be the closure of @xmath22 , as depicted in figure 2 .
a quick comparison with figure 1 shows that @xmath9 is also a knot .
in particular , @xmath9 is a quasipositive knot .
next , we construct an embedded punctured mbius band whose boundary components are @xmath8 and @xmath9 : choose any embedded annulus @xmath24 , one of whose boundary components is the closure of @xmath22 , i.e. @xmath8 .
glue a mbius band to @xmath25 , woven as the additional band of @xmath12 , as shown in figure 3 .
the two boundary components of the resulting punctured mbius band are isotopic to @xmath8 and @xmath9 . at last ,
we construct a natural fibre surface @xmath26 for @xmath9 .
@xmath26 is a union of @xmath27 discs , one for each braid strand of @xmath22 , and several bands with a positive half twist .
each positive crossing corresponds to such a twisted band , except in the neighbourhood of negative crossings , where we choose twisted bands as shown in figure 4 .
the dashed line indicates the core curve of a hopf band in @xmath26 , which we may deplumb without affecting the attribute of fibredness or non - fibredness .
thus we may cut the upper left twisted band and obtain the surface @xmath28 shown on the left hand side of figure 5 .
the surface @xmath28 in turn is isotopic to the canonical seifert surface of a positive braid , as shown on the right hand side of figure 5 ( slide the long band along the upper band ) .
this concludes the proof of theorem 1 , since canonical seifert surfaces of positive braids are fibre surfaces . 1 .
the knot @xmath9 is even strongly quasipositive , in the sense of l. rudolph ( @xcite ) .
it is absolutely essential that the surface @xmath29 be orientable .
indeed , there exist strong restrictions on the euler number of oriented cobordisms between quasinegative and quasipositive knots ( e.g. the rasmussen invariant ( @xcite ) ) .
the application of one saddle point move inside the dashed circle at the bottom right of figure 2 transforms @xmath9 into @xmath8 . here
a saddle point move is a local move that acts on link diagrams , as shown in figure 6 . in @xcite ,
hoste , nakanishi and taniyama proved that a simple crossing change can be expressed as a sequence of two saddle point moves .
therefore , it is an unknotting operation . | we prove that every knot in the 3-space bounds an embedded punctured mbius band whose other boundary component is a quasipositive fibred knot . |
the magnetized liner inertial fusion ( maglif ) concept@xcite is presently being investigated experimentally@xcite using the z facility@xcite at sandia national laboratories .
maglif is part of a broader class of concepts referred to collectively as magneto - inertial fusion ( mif).@xcite these concepts seek to significantly reduce the implosion velocity and pressure requirements of traditional inertial confinement fusion ( icf)@xcite by using a magnetic field to thermally insulate the hot fuel@xcite from a cold pusher and to increase fusion product confinement .
the maglif concept at sandia uses the electromagnetic pulse supplied by the z accelerator to radially implode an initially solid cylindrical metal tube ( liner ) filled with preheated and premagnetized fusion fuel ( deuterium or deuterium - tritium ) .
the implosion is a result of the fast z - pinch process , where a large gradient in the applied magnetic field pressure operates near the liner s outer surface.@xcite one- and two - dimensional simulations of maglif using the lasnex radiation magnetohydrodynamics code @xcite
predict that if sufficient liner integrity can be maintained throughout the implosion , then significant fusion yield ( @xmath1100 kj ) can be attained on the z accelerator when deuterium - tritium fuel is used and the accelerator s marx generators are charged to 95 kv to obtain a peak drive current of about 27 ma.@xcite to maintain liner integrity throughout the implosion , liners with thick walls have been proposed@xcite and used experimentally.@xcite thick walls mitigate the deleterious effects of the magneto - rayleigh - taylor ( mrt ) instability,@xcite which first develops on the liner s outer surface , and then works its way inward , toward the liner s inner / fuel - confining surface , throughout the implosion .
the parameter that is typically used to describe the robustness of a liner to the mrt instability is the liner s initial aspect ratio @xmath2 where @xmath3 is the liner s initial outer radius and @xmath4 is the liner s initial wall thickness .
lower @xmath5 liners are more robust to the mrt instability , but their use results in slower implosion velocities ; thus , there is a tradeoff between liner robustness and implosion efficiency .
two - dimensional lasnex simulations predict that liners with @xmath6 should be robust enough to keep the mrt instability from overly disrupting the fusion burn at stagnation.@xcite these preliminary lasnex simulations predict a broad optimum in the fusion yield surrounding @xmath7 ( see fig .
10 in ref . ) .
experiments@xcite on the z accelerator with no initial axial magnetic field have found the implosion dynamics of @xmath8 be liners to be in good agreement with the 2d lasnex predictions found in the original maglif paper@xcite ( e.g. , compare fig . 2 in ref . with fig .
9 in ref . )
by contrast , @xmath8 be liner experiments on the z accelerator that included an initial axial magnetic field of about 10 t have revealed the presence of a helical instability structure that is fundamentally 3d in nature,@xcite and thus could not be captured by the 2d lasnex simulations ; regardless , the inclusion of the initial axial field seems only to have improved overall implosion stability.@xcite moreover , the first fully - integrated experimental tests of maglif have shown that a fuel implosion driven by an @xmath8 be liner , combined with preheat and premagnetization , does indeed result in fusion relevant conditions on the z accelerator.@xcite therefore , we assume that the qualitative 1d versus 2d behavior shown in fig . 10 of ref .
holds fairly true for @xmath9 , and thus that maglif can be modeled reasonably well in 1d for @xmath9 . the leading hypothesis for what seeds the mrt development is the electrothermal instability ( eti).@xcite recent experiments have shown that eti growth can be tamped by applying a thick dielectric coating to the outer surface of the metal liner , thereby delaying the onset of nonlinear mrt growth.@xcite thus , it is hopeful that the application of dielectric coatings may enable maglif s 2d and/or 3d @xmath10 1d performance to be extended to liners with aspect ratios significantly greater than 6 .
note that these assumptions do not account for possible azimuthal variations ; the effects of azimuthal variations on fuel confinement are beyond the scope of this paper , but they are presently being investigated in other studies.@xcite presumably , the deleterious effects of azimuthal variations are also mitigated by the use of low @xmath5 liners , and perhaps by dielectric coatings as well .
since maglif calls for low @xmath5 liners to mitigate the mrt instability , the resulting liner implosions are too slow to shock heat the fuel ( i.e. , the fuel heating due to compression is approximately adiabatic ) .
thus , to heat the fuel to fusion - relevant temperatures ( @xmath11 kev ) while simultaneously lowering fuel convergence requirements , the maglif concept calls for preheating the fuel just prior to the implosion .
the parameter typically used to describe fuel convergence is the convergence ratio @xmath11 where @xmath12 and @xmath13 are the initial and time - dependent radii of the fuel - liner interface , respectively .
the convergence ratio is often evaluated at particular times of interest , such as the time of peak fusion power , @xmath14 , and the time of minimum @xmath15 , i.e. , the time of bounce " , @xmath16 . in many cases , @xmath14 and @xmath16 have fairly similar values , but differences between these two parameters can indicate important phenomena such as fuel collapse or fuel ignition . to reduce @xmath14 and @xmath16 requirements to @xmath17 ,
the maglif concept calls for preheat temperatures of @xmath18 ev . at the z facility
, fuel preheating has been accomplished using the z beamlet laser ( zbl),@xcite which was originally coupled to the z pulsed - power accelerator for diagnostic purposes ( e.g. , radiography@xcite ) . in the first fully - integrated maglif experiments,@xcite zbl provided about 2 kj of 532-nm light in a 2-ns pulse .
recently , zbl has been upgraded to deliver about 4 kj in a 4-ns pulse , and plans are in place to further increase zbl s delivered pulse energy to 6 kj and beyond over the next several years . to keep the fuel hot during the relatively slow implosion of a maglif liner , the concept requires premagnetization of the fuel using an axially - aligned @xmath19 field .
about 1050 t are to be supplied by external @xmath19 coils .
this initial seed field is to be amplified by a factor of 10@xmath2010@xmath21 within the fuel - filled volume of the imploding liner via magnetic flux compression .
this large axial field is required to mitigate energy loss from the fuel due to electron and ion thermal conduction .
additionally , the axial field should enhance @xmath0-particle confinement and heating of the fuel , and thus increase the overall fusion yield . in the first fully - integrated maglif experiments,@xcite as well as in liner implosion dynamics experiments,@xcite initial seed fields of up to 10 t
have been supplied by a newly developed applied @xmath22 on z ( abz ) subsystem at the z facility.@xcite analysis of the neutron diagnostics data collected during the first fully - integrated maglif experiments indicates that significant flux compression did indeed occur.@xcite also , the abz capabilities at the z facility have recently been upgraded to 15 t , and plans are in place to further increase these fields to 30 t in the next couple of years.@xcite to elucidate some of the key physics issues relevant to maglif , we have developed a new semi - analytic model of the concept .
this model is formulated as a system of ordinary differential equations ( odes ) that are straightforward to solve , particularly with standard software tools such as matlab^^ , idl^^ , mathematica^^ , etc . this model accounts for : ( 1 ) preheat of the fuel ( optionally via laser absorption ) ; ( 2 ) pulsed - power - driven liner implosion ; ( 3 ) liner compressibility with an analytic equation of state , artificial viscosity , internal magnetic pressure , and ohmic heating ; ( 4 ) adiabatic compression and heating of the fuel ; ( 5 ) radiative losses and fuel opacity ; ( 6 ) magnetic flux compression with nernst thermoelectric losses ; ( 7 ) magnetized electron and ion thermal conduction losses ; ( 8) end losses ; ( 9 ) enhanced losses due to prescribed dopant concentrations and contaminant mix ; ( 10 ) deuterium - deuterium and deuterium - tritium primary fusion reactions for arbitrary deuterium to tritium fuel ratios ; and ( 11 ) magnetized @xmath0-particle heating .
this model has been implemented in a code called samm ( semi - analytic maglif model ) .
simulations using samm typically take about 30 seconds to run on a laptop using the ode23 solver in matlab^^. using the parallel computing cluster at sandia , parameter scans of about 2000 simulations can be completed in as little as 10 minutes .
semi - analytic models have proven themselves useful in the past.@xcite their fast run times allow the system parameter space to be explored rapidly .
also , the physics included in semi - analytic models are often transparent and thus accessible by anyone , which is not always the case for more sophisticated codes ( e.g. , the lasnex code@xcite that the original maglif paper@xcite was based on ) .
finally , one of the most important aspects of developing simplified models such as samm , is that it forces those involved to verify and understand the results generated by more sophisticated codes .
for example , when differences are observed between a simple model and a more advanced code , can those differences be explained in terms of the simplifying assumptions made in the reduced model ?
this type of questioning often leads to new physics insights that would otherwise be overlooked .
for example , during the development of samm , we found an important relationship between radiative loss rates and the initial radial fraction of the fuel that is preheated ; this finding and others are discussed further in this paper .
the remainder of this paper is organized as follows . in sec .
[ sec : model ] , we present our semi - analytic model of maglif . in sec . [
sec : results ] , we verify this model relative to the 1d results presented in the original maglif paper ( i.e. , ref . ) . in sec .
[ sec : summary ] , we summarize this work . unless otherwise specified , all units are si .
an overview of the semi - analytic maglif model is provided in fig . [
fig : schematic ] .
schematic overview of the semi - analytic maglif model .
there are three primary regions : a fuel region , a liner region , and a vacuum region .
the system height is @xmath23 ; the thermally - insulating axial magnetic field , which is initially distributed uniformly over all regions , is @xmath19 ; the radius of the fuel - liner interface is @xmath15 ; the liner s outer radius is @xmath24 ; the return current radius is @xmath25 ; the azimuthal magnetic field , which drives the cylindrical implosion , is @xmath26 .
normalized profiles are shown for @xmath26 in the vacuum region , @xmath27 ( magenta ) , and in the liner region , @xmath28 ( orange ) ; their analytic expressions are given by eqs .
[ bthetav ] and [ bthetal ] in the text ( @xmath26 is assumed to be zero in the fuel region ) .
the liner region is further divided into multiple concentric liner shells ; this discretization is necessary to avoid overdriving the fuel . within the fuel region ,
normalized profiles are shown for the gas pressure , @xmath29 ( blue ) , the gas temperature , @xmath30 ( red ) , the gas density , @xmath31 ( green ) , and the radiation temperature , @xmath32 ( cyan ) .
the pressure profile is flat throughout the fuel ( i.e. , we have made an isobaric assumption due to the subsonic nature of maglif implosions ) .
the gas temperature and density profiles thus have an inverse dependence to one another ; their analytic expressions are given by eqs .
[ tgr][rhogrshelf ] in the text .
the radiation temperature is nearly constant across the fuel region .
the fuel region is further divided into a hot spot region from @xmath33 to @xmath34 and a cold dense shelf region from @xmath34 to @xmath15 .
the gas temperature in the shelf region is equal to the radiation temperature ( i.e. , the fuel material and the radiation field are in thermodynamic equilibrium in the shelf region ) .
the shelf region erodes away throughout the implosion , until @xmath35 , due to thermal transport from the hot spot to the shelf .
the shelf region is only present if the fuel is preheated from @xmath33 to @xmath36 . ]
the liner implosion , and thus the fuel implosion , is driven by the pressure associated with the azimuthal magnetic field , which is supplied by the pulsed - power driver ( e.g. , the z accelerator ) .
because of the cylindrical symmetry , the azimuthal field in the vacuum region is @xmath37 where @xmath38 h / m is the permeability of free space and @xmath39 is the liner current .
we assume that @xmath40 is partially diffused into the liner wall , and that its distribution in the liner region can be described by @xmath41 where the constant power @xmath42 is found by forcing the amplitude of @xmath43 to drop by one @xmath44-folding within one skin depth , @xmath45 , of the liner s outer surface ; that is , we set @xmath46 where @xmath47 is the base of the natural logarithm , and solve for @xmath42 , which gives @xmath48 the skin depth is given by @xmath49 where @xmath50 is the initial electrical resistivity of the liner material , and @xmath51 is the rise time of the driving @xmath40 pulse ( e.g. , @xmath52 ns for the z accelerator ) .
some parameters of interest are provided in table [ rhoe ] .
the purpose of assuming a simple power - law dependence for @xmath43 , rather than an exponential dependence , is that it allows us to integrate both the total magnetic flux and the total magnetic field energy analytically , which enables us to obtain simple analytic expressions for the inductance , circuit response , and ohmic dissipation due to @xmath28 .
.[rhoe]parameters for @xmath43 ( where @xmath53 ns ) . [ cols="^,^,^,^",options="header " , ] the reactivity rates , along with end losses , provide the rate of change of the total number of deuterons and tritons in the fuel , respectively , as @xmath54 the reactivity rates also provide the total fusion power , as given by @xmath55 where the energy yields per dt , dd-@xmath21he , and dd - t reaction are , respectively , @xmath56 since one neutron is released per dt reaction and one neutron is released per dd-@xmath21he reaction , the primary dt and primary dd neutron yields are given by @xmath57 and thus the total primary neutron yield is @xmath58 finally , the total fusion energy yield is @xmath59 the dynamics of our model are described by eqs .
[ isdot ] , [ vcdot ] , [ ildot ] , [ rliddot ] , [ rgddot ] , [ rlddot ] , [ elsdot ] , [ egdot ] , [ nmixdot ] , [ phizgdot ] , [ fmhdot ] , [ ndtdot ] , [ ndd3hedot ] , [ nddtdot ] , [ nddot ] , and [ ntdot ] .
these @xmath60 ordinary differential equations are repeated here for emphasis and clarity : @xmath61_{r = r_g } \\ & \dot{f}_{mh } = \dot{m}_{s\to h}/m_g \label{fmhdot2}\\ & \dot{n}_{dt } = h\int_0^{r_g}{n_d n_t \langle\sigma v\rangle_{dt } \cdot 2\pi r \cdot dr } \\
& \dot{n}_{dd,^3{\rm he } } = \frac{h}{2}\int_0^{r_g}{n_d^2\langle\sigma v\rangle_{dd,^3{\rm he } } \cdot 2\pi r \cdot dr } \\ & \dot{n}_{dd , t } = \frac{h}{2}\int_0^{r_g}{n_d^2\langle\sigma v\rangle_{dd , t } \cdot 2\pi r \cdot dr } \\ & \dot{n}_d = -\dot{n}_{dt}-2\dot{n}_{dd,^3he}-2\dot{n}_{dd , t}-\dot{n}_{d , ends}\\ & \dot{n}_t = -\dot{n}_{dt}-\dot{n}_{t , ends}.\end{aligned}\ ] ] note that if the liner current is prescribed , rather than derived from the voltage - driven circuit model , then eqs .
[ isdot2][ildot2 ] are discarded , and only @xmath62 ordinary differential equations are required to describe the dynamics of our model . furthermore ,
if we ignore the nernst effect , preheat all of the fuel uniformly , use either pure d@xmath63 fuel or a 50/50% dt mix , and we ignore dopants and contaminants , then the number of required differential equations drops to only @xmath64 .
in fig . [ fig_code_verification ] , we present samm simulation results where the intent was to reproduce the 1d results presented in the original maglif paper.@xcite these samm results show that our simplified model does indeed capture the 1d behavior presented throughout ref . .
all of the simulations presented in fig .
[ fig_code_verification ] used dt fuel . in fig .
[ fig_code_verification](a ) , we have plotted the fuel radius , liner radius , drive current , and average fuel temperature as a function of time for the preliminary point design discussed in ref . .
this figure should be compared with fig . 4 in ref . .
the samm simulation produced a fusion energy yield of 970 kj , compared to about 500 kj in ref . , and resulted in a maximum convergence ratio of 25 , which matches the preliminary point design of ref . .
[ fig_code_verification](b ) , we have plotted the normalized fuel temperature , magnetic field strength , and fuel density at stagnation ( at peak burn rate ) as a function of the normalized fuel radius for two samm simulations , one with the nernst effect included ( solid lines ) , and one without the nernst effect included ( dashed lines ) .
all quantities are normalized to the simulation without the nernst effect.@xcite this figure should be compared with fig . 5 in ref . .
the yield for the simulation with nernst is 72% lower than the simulation without nernst ( compared to about 70% in ref . ) .
the simulation with nernst resulted in an axial magnetic flux loss of about 64% ( compared to about 70% in ref . ) .
figures [ fig_code_verification](c ) and [ fig_code_verification](d ) should be compared with figs .
6(a ) and 6(b ) in ref . . in ref .
, the initial fuel densities that result in solutions with convergence ratios of 10 , 20 , and 30 , were found as a function of the initial fuel temperature . using these initial conditions directly with samm , shown here in fig .
[ fig_code_verification](d ) , we generated the results shown in fig .
[ fig_code_verification](c ) .
the resulting convergence ratios were [email protected] , [email protected] , and [email protected] , and thus are only approximately equal to 10 , 20 , and 30 . in fig .
[ fig_code_verification](e ) , we have plotted the optimum fusion energy yields per unit liner length as a function of the initial axial magnetic field strength . here we have held the maximum convergence ratio fixed at either 10 , 20 , or 30 . to find these optimum constant convergence ratio solutions , we had to scan over the input parameter space that consists of the initial preheat energy and the initial fuel density .
these results should be compared with fig . 7 in ref . .
figures [ fig_code_verification](f ) and [ fig_code_verification](g ) should be compared with figs .
8(a ) and 8(b ) in ref . .
[ fig_code_verification](f ) , we have plotted the optimum fusion energy yields per unit liner length as a function of the peak drive current for two cases , one with @xmath0-deposition turned on , and one with @xmath0-deposition turned off . here , as in ref .
, for the case with @xmath0-heating , we scanned over the input parameter space that consists of the initial preheat energy and the initial fuel density to find the optimum solutions that had a maximum convergence ratio of 20 ( for the case with no @xmath0-heating , the same input parameters were used , but @xmath0-deposition was simply turned off , thus the convergence ratios varied somewhat ) . in fig .
[ fig_code_verification](g ) , we have plotted the ratio of the peak fuel temperature with @xmath0-heating turned on to the peak fuel temperature with @xmath0-heating turned off .
we have also plotted the ratio of the total fusion energy yield to the total energy absorbed by the target ( liner and fuel ) , i.e. , the `` target gain '' . in fig .
[ fig_code_verification](h ) , we have plotted the optimum fusion energy yields per unit liner length as a function of the initial liner aspect ratio ( ar ) . this figure should be compared with fig . 10 in ref . .
in the samm results , the transition to higher yields with higher aspect ratios is a bit more abrupt , and then rolls off sooner ; however , the general trend is the same as that shown in ref . .
also , although not explicitly shown in ref . , the eventual rolloff in yield with higher aspect ratio is expected .
that is , to keep the implosion time constant with higher aspect ratio , the overall liner mass must be decreased while the liner s initial outer radius is increased ( and while the liner s initial wall thickness is decreased ) .
eventually , the liner s mass decreases to a point where it can not provide sufficient tamping and inertial confinement at stagnation . in fig .
[ fig_code_verification](i ) , we have plotted the normalized yield as a function of the premixed fraction of beryllium in the fuel .
the yields are normalized to the case with no mix / dopants .
this figure should be compared with fig . 11 in ref . .
[ fig_code_verification](j ) , we consider laser propagation through the fuel using the analytic treatment discussed in sec .
[ sec : preheat ] .
here we plot the laser heating front propagation distance as a function of time for the preliminary point design of ref . .
the laser pulse duration is 10 ns . at 5 ns ( red )
, the front of the `` bleaching wave '' has reached 0.56 cm into the fuel , which is consistent with fig .
13(b ) of ref . . by 8 ns ( blue ) and 11 ns ( green ) ,
the bleaching wave has propagated completely through the 0.65-cm - long imploding fuel region , which is consistent with figs .
13(c ) and 13(d ) of ref . ,
respectively . in fig .
[ fig_code_verification](k ) , we present a color - filled contour plot of total fusion energy yield as a function of both the fuel preheat temperature and the laser beam spot size ( radius ) . overlaid on top of the color - filled contour plot are the contours of constant preheat energy required to obtain the preheat temperature at the given spot sizes .
this figure should be compared with fig .
15 in ref . .
in ref . , this figure was generated by additionally scanning over a range of initial fuel densities to obtain only the solutions where the convergence ratios were 25 .
for simplicity , we used the input densities and preheat energies found by this previous effort to generate the samm results presented here in fig .
[ fig_code_verification](k ) , and thus the convergence ratios were allowed to vary somewhat . finally , in fig .
[ fig_code_verification](l ) , we have plotted results for mass end losses out of the fuel region through the laser entrance hole using the analytic treatment discussed in sec . [
sec : endlosses ] .
here we are plotting the ratio of mass remaining after fusion burn to the mass of the fuel at the start of the simulation as a function of the axial length of the liner .
we have done this for three cases , where the ratio of the radius of the laser entrance hole to the initial radius of the fuel is either 1.0 , 0.5 , or 0.25 .
this figure should be compared with fig . 16 in ref .
in this article , we have presented a new semi - analytic model of maglif .
this model , implemented in a code called samm , accounts for : ( 1 ) preheat of the fuel ( optionally via laser absorption ) ; ( 2 ) pulsed - power - driven liner implosion ; ( 3 ) liner compressibility with an analytic equation of state , artificial viscosity , internal magnetic pressure , and ohmic heating ; ( 4 ) adiabatic compression and heating of the fuel ; ( 5 ) radiative losses and fuel opacity ; ( 6 ) magnetic flux compression with nernst thermoelectric losses ; ( 7 ) magnetized electron and ion thermal conduction losses ; ( 8) end losses ; ( 9 ) enhanced losses due to prescribed dopant concentrations and contaminant mix ; ( 10 ) deuterium - deuterium and deuterium - tritium primary fusion reactions for arbitrary deuterium to tritium fuel ratios ; and ( 11 ) magnetized @xmath0-particle heating .
we have also shown that samm is capable of reproducing the general 1d behavior presented throughout the original maglif paper ( i.e. , ref . ) . presently , we are using samm to further explore the parameter space of maglif .
these efforts are complementary to other studies presently taking place using full 1d , 2d , and 3d radiation magnetohydrodynamics codes.@xcite in particular , we are using samm to study the parameter space surrounding the experimental platform of ref .
( i.e. , an initial axial magnetic field of 10 t , a preheat energy of less than 2 kj , an initial fuel density of about 1 mg / cm@xmath21 , and a peak drive current of about 20 ma ) .
we are also studying how this parameter space could change as a series of planned upgrades are made to the z facility over the next few years .
these upgrades are intended to bring maglif s experimental platform closer to the parameters of the preliminary point design described in ref .
( i.e. , an initial axial magnetic field of 30 t , a preheat energy of 8 kj , an initial fuel density of 3 mg / cm@xmath21 , and a peak drive current of 27 ma ) .
additionally , we are studying how this parameter space could change as the peak drive current is increased to about 60 ma , in consideration of maglif on possible future pulsed - power accelerators , such as the conceptual z-300 and z-800 accelerators.@xcite the results of these parameter space studies using samm will be presented in a future publication . in closing
, we note that the development of samm has led to several physical insights that were not fully appreciated previously ( e.g. , the dependence of radiative loss rates on the radial fraction of the fuel that is preheated ) .
additionally , this model s accessible physics and fast run times ( @xmath67 30 seconds / simulation ) make it a useful pedagogical tool , especially for students , experimentalists , or any researcher interested in maglif .
we would like to thank d. b. sinars for supporting this project and reviewing this manuscript .
we also thank m. e. cuneo , m. c. herrmann , c. w. nakhleh , k. j. peterson , and m. k. matzen for supporting this project .
we thank r. a. vesey for useful discussions about @xmath68 approximations , maglif parameter scans , and for supplying the simulation input parameters needed for fig .
[ fig_code_verification](k ) in this manuscript .
we thank k. r. cochrane for supplying us with sesame equation - of - state data and for suggesting the use of the birch - murnaghan formulation as we attempted to fit the sesame cold - curve data to various analytic formulae .
we thank s. b. hansen for useful discussions about various simplified radiation loss models .
we thank p. f. schmit , p. f. knapp , and m. r. gomez for some helpful beta testing of the samm software .
we thank m. r. gomez for useful discussions about fully - integrated maglif experiments .
we thank m. geissel for useful discussions about laser transmission through laser entrance windows .
we thank the maglif / icf and dynamic material properties research groups in general for many useful technical discussions regarding various aspects of maglif ; in particular , we thank a. b. sefkow , m. r. martin , c. a. jennings , r. w. lemke , t. j. awe , d. c. rovang , d. c. lamppa , and w. a. stygar . finally , we thank the personnel of the pulsed - power sciences center ( including the z and zbl facilities ) at sandia national laboratories ; without their hard work and dedication to excellence , this work would not have been possible .
sandia is a multiprogram laboratory operated by sandia corporation , a lockheed - martin company , for the united states department of energy s national nuclear security administration , under contract no .
de - ac04 - 94al85000 .
following ref . , the coulomb logarithm is given by @xmath70 where @xmath71 is the debye length of the plasma and @xmath72 is the minimum impact parameter of the plasma .
the debye length is given by @xmath73 the minimum impact parameter is conventionally taken as @xmath74 where @xmath75 is the classical minimum impact parameter , and @xmath76 is the quantum mechanical minimum impact parameter .
here , @xmath77 @xmath78 is the reduced planck constant and @xmath79 is the electron thermal velocity . for the @xmath69 calculation in eq .
[ nueizblabs ] , we use the following assignments : @xmath80 , @xmath81 , @xmath82 , and @xmath83 . for the @xmath69 calculation in eq .
[ lalpha ] , we use the following assignments : @xmath84 , @xmath82 , and @xmath83 . for the @xmath69 calculations in eqs .
[ nueic ] and [ nuiic ] , we use the following assignments : @xmath85 , @xmath86 , and @xmath87 .
this case occurs when either all of the fuel is preheated uniformly , or when the shelf region has completed eroded away ( see sec .
[ sec : erosion ] ) . for this case , @xmath35 , and we iteratively guess for @xmath88 in our bisection algorithm . after evaluating the expressions in eqs .
[ prh1][prh2 ] , if @xmath91 , then @xmath88 is too low , and the correct value for @xmath88 will be found between this value that is too low and the average temperature of the entire fuel , @xmath92 .
[ note that we take @xmath92 as the initial upper limit for @xmath88 in the bisection algorithm because with @xmath35 and @xmath93 , the resulting fuel temperature and density profiles are flat , and @xmath94 while @xmath95 . ]
conversely , if @xmath96 , then @xmath88 is too high , and the correct value for @xmath88 will be found between this value that is too high and zero .
[ note that @xmath97 results in @xmath98 while @xmath99 , thus we take zero as the initial lower limit for @xmath88 in the bisection algorithm . ] for this case , we must ensure that @xmath34 and @xmath88 are chosen such the following mass conservation argument is met : @xmath100 here @xmath101 is the fraction of the total fuel mass that is in the hot spot region , which is the system variable defined by eqs .
[ fmhdot][fmh0 ] ( see also eq . [ fmhdot2 ] ) . also , we have used eq .
[ rhogrshelf ] for @xmath102 and , in the last equality , we have used our isobaric assumption that @xmath103 .
equation [ b3 ] provides the analytic relationship between @xmath88 and @xmath34 that ensures mass conservation in the shelf region .
however , if we were to iteratively guess for @xmath88 in our bisection algorithm , then solving eq .
[ b3 ] for @xmath34 in terms of @xmath88 would require a numerical treatment .
therefore , we instead guess iteratively for @xmath34 , which must reside in the domain @xmath104 ( thus , for our bisection algorithm , we use @xmath33 and @xmath105 as our initial lower and upper limits , respectively ) . for each @xmath34 guess , eq .
[ b3 ] provides the unique value of @xmath88 that will ensure mass conservation .
note that @xmath106 is a monotonically decreasing function ( i.e. , @xmath107 as @xmath108 , and @xmath109 as @xmath110 ) .
thus , in our bisection algorithm , if @xmath91 , then @xmath88 is too low and @xmath34 is too large ; the correct value for @xmath34 will be found between this value that is too large and zero .
conversely , if @xmath96 , then @xmath88 is too high and @xmath34 is too small ; the correct value for @xmath34 will be found between this value that is too small and @xmath15 . before we can test our iterative guesses for @xmath34 and @xmath88 in eqs .
[ prh1][prh2 ] , we first need to find @xmath89 and @xmath90 to complete the profile definitions given in eqs .
[ tgr][rhogrshelf ] . to do this
, we use our isobaric assumption to first update @xmath111 .
next , we find the value for @xmath89 that makes the total mass in the hot spot , as defined by eq .
[ rhogr ] , consistent with @xmath112 ( see appendix [ rhocnotes ] for additional details ) .
then , we again apply our isobaric assumption to get @xmath113 .
finally , with all profile parameters defined and updated , we evaluate @xmath114 and @xmath115 and test whether the results meet our desired solution accuracy .
the central mass density , @xmath116 , is updated after updating @xmath88 , @xmath34 , and @xmath117 .
the updated @xmath89 value is found by ensuring that the following mass conservation argument is met : @xmath118 where @xmath102 is given by eq .
[ rhogr ] .
if the expressions given for @xmath119 and @xmath102 in eqs . [ tgr ] and [ rhogr ] were permitted to extend to radii beyond @xmath34 , then we would have @xmath120 and @xmath121 when @xmath122 using @xmath123 to define the dimensionless radius @xmath124 , the hot spot mass density given by eq .
[ rhogr ] becomes @xmath125 evaluating the dimensionless radius @xmath126 at @xmath127 gives @xmath128 and thus the total fuel mass in the hot spot region , as given by the right hand side of eq .
[ mgh1 ] , becomes @xmath129 since the average mass density in the hot spot region is @xmath130 , we have @xmath131 now , from eqs .
[ rstar ] and [ xh ] , we have @xmath132 where we have again applied our isobaric assumption ; in this case , @xmath133 .
solving for @xmath89 in eq .
[ xhxi ] gives @xmath134 substituting eq .
[ rhocc ] into eq .
[ rhohbar ] and solving for @xmath135 gives @xmath136 where @xmath137 is a dimensionless , monotonically decreasing function of the dimensionless variable @xmath138 ; it is defined over our physically relevant domain of @xmath139 with a range of @xmath140 ( see fig . [ pxh ] ) . because @xmath137 is a dimensionless function of a dimensionless argument , we only need to evaluate it once and store the results for continuous use throughout the simulation .
with this done , we take the inverse of @xmath137 , with input argument @xmath141 , to get @xmath142 finally , from eq .
[ rhocc ] , we get the central density , @xmath89 , repeated here for completeness as @xmath143 97ifxundefined [ 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 * * , ( ) link:\doibase 10.1103/physrevlett.108.025003 [ * * , ( ) ] @noop * * , ( ) \doibase doi:10.1063/1.3560911 [ * * , ( ) ] link:\doibase 10.1109/tps.2012.2223488 [ * * , ( ) ] in @noop _ _ , vol . , ( , ) pp . in @noop _ _ , vol . , ( , ) pp
. \doibase http://dx.doi.org/10.1063/1.3694519 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.135004 [ * * , ( ) ]
link:\doibase 10.1063/1.4803079 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4803074 [ * * , ( ) ] link:\doibase 10.1063/1.4751868 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4802836 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.135002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.235005 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4872331 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4890298 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.155003 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.155004 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4920948 [ * * , ( ) ] @noop * * , ( ) in @noop _ _ ( , ) pp . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://stacks.iop.org/0029-5515/21/i=1/a=004 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.863415 [ * * , ( ) ] http://stacks.iop.org/0029-5515/23/i=3/a=001 [ * * , ( ) ] http://stacks.iop.org/0029-5515/26/i=2/a=001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.56.139 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) http://stacks.iop.org/0029-5515/40/i=1/a=305 [ * * , ( ) ] http://stacks.iop.org/0029-5515/41/i=2/a=311 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0029-5515/43/i=1/a=302 [ * * , ( ) ] @noop * * , ( ) in @noop _ _ ( ) pp .
http://dx.doi.org/10.1038/239139a0 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.871025 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.045004 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3484873 [ * * , ( ) ] http://dx.doi.org/10.1038/nature13008 [ * * , ( ) ] link:\doibase 10.1103/physrev.76.904 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.72.167 [ * * , ( ) ] @noop * * , ( ) \doibase doi:10.1063/1.1724473 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.29.1429 [ * * , ( ) ]
link:\doibase 10.1109/tps.2002.805418 [ * * , ( ) ] \doibase doi:10.1063/1.3088020 [ * * , ( ) ] link:\doibase 10.1103/physreve.83.066405 [ * * , ( ) ] link:\doibase 10.1063/1.3680646 [ * * , ( ) ] @noop ( ) @noop ( ) @noop * * , ( ) \doibase doi:10.1063/1.2956823 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1063/1.4902566 [ * * , ( ) ] http://stacks.iop.org/1402-4896/70/i=5/a=008 [ * * , ( ) ] link:\doibase 10.1103/physrevstab.13.120401 [ * * , ( ) ] @noop * * , ( ) in @noop _ _ ( , , ) @noop _ _ ( , ) pp .
@noop @noop @noop @noop _ _ ( , ) pp .
link:\doibase 10.1103/physrev.71.809 [ * * , ( ) ] http://www.jstor.org/stable/2371405 [ * * , ( ) ] @noop * * , ( ) @noop @noop _ _ , ed .
( , ) pp .
@noop @noop _ _ , ed .
( , ) p. , ed .
, @noop _ _ , ed .
( , ) @noop _ _ , ed . , edited by and ( , ) p. @noop _ _ , ed .
( , ) pp . @noop _ _ ( , ) chap . , @noop ( ) @noop ( ) @noop * * , ( ) @noop @noop
@noop \doibase http://dx.doi.org/10.1063/1.865901 [ * * , ( ) ] @noop _ _ , vol .
( , ) p. @noop @noop ( ) @noop _ _ , ed .
( , ) p. http://stacks.iop.org/0029-5515/32/i=4/a=i07 [ * * , ( ) ] @noop _ _ , ed .
( , ) p. @noop @noop @noop ( ) @noop ( ) | presented is a semi - analytic model of magnetized liner inertial fusion ( maglif ) .
this model accounts for several key aspects of maglif , including : ( 1 ) preheat of the fuel ( optionally via laser absorption ) ; ( 2 ) pulsed - power - driven liner implosion ; ( 3 ) liner compressibility with an analytic equation of state , artificial viscosity , internal magnetic pressure , and ohmic heating ; ( 4 ) adiabatic compression and heating of the fuel ; ( 5 ) radiative losses and fuel opacity ; ( 6 ) magnetic flux compression with nernst thermoelectric losses ; ( 7 ) magnetized electron and ion thermal conduction losses ; ( 8) end losses ; ( 9 ) enhanced losses due to prescribed dopant concentrations and contaminant mix ; ( 10 ) deuterium - deuterium and deuterium - tritium primary fusion reactions for arbitrary deuterium to tritium fuel ratios ; and ( 11 ) magnetized @xmath0-particle fuel heating .
we show that this simplified model , with its transparent and accessible physics , can be used to reproduce the general 1d behavior presented throughout the original maglif paper [ s. a. slutz _ et al_. , phys .
plasmas * 17 * , 056303 ( 2010 ) ] .
we also discuss some important physics insights gained as a result of developing this model , such as the dependence of radiative loss rates on the radial fraction of the fuel that is preheated . |
one of the first steps in the stability analysis of an equilibrium of a rational difference equation is to determine its local stability .
this has traditionally been done by first linearizing the rational difference equation and then evaluating the jacobian of the linearized equation at the equilibrium . if both eigenvalues of the jacobian lie inside the unit circle , then the equilibrium is said to be locally asymptotically stable .
if one eigenvalue lies inside the unit circle and the second eigenvalue lies outside the unit circle , then the equilibrium is said to be a saddle point .
if both eigenvalues lie outside the unit circle , then the equilibrium is said to be unstable . and
if an eigenvalue lies on the unit circle , then the equilibrium is said to be non - hyperbolic .
but a practical problem often arises when the linearized stability analysis method of determining local stability of equilibria is applied to systems of rational difference equations .
the equilibria of even simple first - order systems of two rational difference equations often turn out to be roots of quartic polynomials .
although it is possible to calculate the explicit formulas for such equilibria using , for example , ferrari s method , such formulas are extremely complicated . as a result , the formulas for eigenvalues of the jacobian of the linearized equation evaluated at such equilibria are extremely complicated .
in fact , one simply can not use them to check algebraically whether the eigenvalues lie inside the unit circle or not , due to computational problems .
thus in such cases , the only known tool for performing local stability analysis on systems of difference equations , namely , the linearized stability analysis method , is rendered useless . in such cases ,
the natural question to ask is : can we still say something about the stability of equilibria with extremely complicated algebraic formulas ? in this paper , we show that for certain subclasses of systems of two rational difference equations , one can actually give a complete global analysis of the equilibria by focussing exclusively on geometric properties of the critical curves of the system rather than on explicit algebraic formulas for the equilibria .
we present a brand new result that establishes the connection between the number of equilibria of a system of two rational difference equations and the stability of those equilibria .
we achieve this by establishing a relationship between the natures of the eigenvalues of the jacobian of the linearized equation evaluated at an equilibrium and the slopes of the critical curves of the system at that equilibrium .
finally , we use this result to give a qualitative analysis of the global behavior of solutions to a leslie - gower - type system of two rational difference equations .
the study of dynamics of difference equations often requires that equilibria be calculated first , followed by a local stability analysis of the equilibria .
this is then complemented by other considerations ( existence of periodic points , chaotic orbits , etc . ) .
if the analysis is applied to a class of equations dependent on one or more real parameters , the task is complicated by the fact that a formula is not always available for equilibria , and even if it is , determination of stability properties of parameter - dependent equilibria may be a daunting task . in this paper
we address the study of stability character of equilibria for a parameter dependent class of systems of difference equations . for this
we develop local geometric criteria for planar systems of difference equations , which under certain conditions determine local stability character of an equilibrium point .
this idea is used to study a class of difference equations in the plane , for which we show that the _ number of equilibria determines , essentially in a unique way , the local stability character of the equilibria , as well as the kind of global dynamics that is possible_. the class of difference equations we will focus on has its roots in biology applications .
the leslie - gower model in difference equations is the two - species competition model @xmath1 it is a modified version of the beverton - holt equation which is a well - known difference equation in population biology .
the normalized leslie - gower model @xmath2 was studied in detail by liu and elaydi @xcite , j. cushing et .
al @xcite , and kulenovi and merino @xcite .
in this paper we consider the system of equations obtained by adding positive constants to the right - hand - side of the leslie - gower equations : @xmath3 the positive constants @xmath4 and @xmath5 in ( lgi ) may account for immigration .
the change of variables @xmath6 , @xmath7 normalizes ( lgi ) as follows : @xmath8 where in ( lgin ) the tildes have been removed from the variables to simplify notation , and @xmath9 the system ( lgin ) is an example of a _ competitive system _ , which is defined next .
let @xmath10 and @xmath11 be intervals of real numbers and let @xmath12 and @xmath13 be continuous functions . consider the system @xmath14 system ( [ eq : competitive ] ) is _ competitive _ if @xmath15 is non - decreasing in @xmath16 and non - increasing in @xmath17 , and @xmath18 is non - increasing in @xmath17 and non - decreasing in @xmath16 .
the map @xmath19 associated to a competitive system ( [ eq : competitive ] ) is said to be _ competitive_. system ( [ eq : competitive ] ) is _ strongly competitive _ if it is competitive , with strict coordinate - wise monotonicity of the functions @xmath15 and @xmath18 . if @xmath20 is differentiable on an open set @xmath21 , a sufficient condition for @xmath20 to be strongly competitive on @xmath21 is that the jacobian matrix of @xmath20 at any @xmath22 has nonzero entries with sign configuration @xmath23 competitive systems of the form ( [ eq : competitive ] ) have been studied by many authors @xcite , @xcite , @xcite , @xcite and others .
the term _ competitive _ was introduced by hirsch @xcite ( see also @xcite ) for systems of autonomous differential equations @xmath24 , @xmath25 satisfying @xmath26 . the main motivation for the study of these systems is the fact that many mathematical models in biological sciences may be classified as competitive ( or cooperative ) @xcite , @xcite , @xcite .
consideration of poincar maps of these systems leads to the concept of competitive and cooperative in the discrete case
. denote with @xmath27 the _ south - east _ partial order in the plane whose nonnegative cone is the standard fourth quadrant @xmath28 , that is , @xmath29 if and only if @xmath30 and @xmath31 . competitive maps @xmath20 in the plane preserve the south - east ordering : @xmath32 whenever @xmath33 . the concept of competitive ( for maps )
may be defined in terms of the order preserving properties of the maps .
finally we note that the terms _ equilibrium _ ( of a planar system of difference equations ) and _ fixed point _ ( of a map ) are used interchangeably in this paper .
this paper is organized as follows . in section [ sec : a result ] a relation between local stability of equilibria and slopes of critical curves for a specific class of difference equations is established . in section [ sec : preliminaries ] it is shown that every solution to eq.(lgin ) converges to an equilibrium .
section [ sec : main results ] contains the main result of this paper .
it states that eq.(lgin ) has between one and three equilibria , and that the number of equilibria determines global behavior as follows : if there is only one equilibrium , then it is globally asymptotically stable .
if there are two equilibria , then one is a local attractor and the other one is nonhyperbolic .
if there are three equilibria , then they are linearly ordered in the south - east ordering of the plane , and consist of a local attractor , a saddle point , and another local attractor .
finally , in section [ sec : a global attractivity result ] we give sufficient conditions for eq.(lgin ) to have a unique equilibrium .
our first result establishes a connection between local stability of the fixed point of a planar map and the slopes of certain curves at the fixed point
. the result will be useful in the proof of the main result in section [ sec : main results ] .
[ thm : main theorem ] let @xmath21 be a subset of @xmath34 with nonempty interior , and let @xmath35 be a map of class @xmath36 for some @xmath37 .
suppose that @xmath20 has a fixed point @xmath38 such that @xmath39 satisfy @xmath40 let @xmath41 , @xmath42 be the _ critical sets _
@xmath43 then , * there exists neighborhood @xmath44 of @xmath45 and @xmath46 of @xmath47 such that the sets @xmath48 and @xmath49 are the graphs of class @xmath36 functions @xmath50 and @xmath51 for @xmath52 . *
the eigenvalues @xmath53 , @xmath54 of the jacobian matrix of @xmath20 at @xmath55 are real , distinct , and the eigenvalue with larger absolute value is positive . if @xmath53 is the larger eigenvalue , then @xmath56 * the existence of of @xmath10 and @xmath11 and of smooth functions @xmath57 and @xmath58 defined in @xmath10 as in the statement of the theorem is guaranteed by the hypotheses and the implicit function theorem .
moreover , @xmath59 * the characteristic polynomial of the jacobian of @xmath20 , @xmath60 has positive discriminant , thus its roots @xmath61 are real .
since @xmath62 , then the root @xmath63 with larger absolute value is positive , and @xmath64 .
now by the hypothesis @xmath65 it follows that @xmath66 , which implies @xmath67 .
since @xmath68 has at least one positive root and since by hypothesis @xmath69 , we have @xmath70 . to prove the second part of ( [ eq : signs of derivatives ] ) , note that from , we have @xmath71
the result is a direct consequence of , the inequality @xmath72 and hypotheses .
in this section we show that every solution to eq.(lgin ) converges to an equilibrium .
let @xmath73 then the map @xmath74 associated with ( lgin ) is @xmath75 for future reference we give the jacobian matrix of @xmath20 at @xmath76 : @xmath77 by direct inspection of ( [ eq : jacobian ] ) we obtain the following result .
[ lemma : t is competitive ] the system of difference equations ( lgin ) is strongly competitive on @xmath78 .
the following lemma has an easy proof which we skip .
[ lemma : t is bounded ] @xmath79 \times [ h_{2},\,h_{2}\,+\,b_{2 } ] $ ] .
[ lem : boundedness ] [ thm : convergence ] every solution of eq.(lgin ) converges to an equilibrium .
it is easy matter to show that the map @xmath20 is one - to - one . from ( [ eq : jacobian ] ) the determinant of @xmath80 is @xmath81 which is clearly positive for @xmath82 .
if follows that the map @xmath20 satisfies hypothesis @xmath83 in @xcite . by lemma 4.3 of @xcite and
theorem 4.2 of @xcite we have that every solution of eq.(lgin ) is eventually coordinate - wise monotone .
since by lemma [ lem : boundedness ] every solution enters the compact set @xmath84 \times [ h_{2},\,h_{2}\,+\,b_{2 } ] $ ] , the conclusion follows .
by solving for @xmath17 and @xmath16 , respectively , in the equations defining the _ critical sets _ @xmath85 and @xmath86 and taking derivatives we obtain the following result .
[ lem : hyperbolas ] all branches of the sets @xmath87 are the graphs of decreasing functions of one variable .
[ lem : hyp .
1 ] the map @xmath20 satisfies hypotheses of theorem [ thm : main theorem ] .
set @xmath88 , @xmath89 , @xmath90 , @xmath91 . implicit differentiation of the equations defining @xmath92 and @xmath93 in ( [ eq : critical curves system ] ) at @xmath94 gives @xmath95 from lemma [ lem : hyperbolas ] , @xmath96 and @xmath97 , and from ( [ eq : jacobian ] ) , @xmath98 and @xmath99 .
it follows that @xmath100 and @xmath101 in .
furthermore , since the map @xmath20 is strongly competitive , @xmath102 , @xmath103 , @xmath98 , @xmath99 .
hence , @xmath20 must satisfy inequalities @xmath104 and @xmath105 from hypotheses of theorem [ thm : main theorem ] . finally note that @xmath106 by , hence @xmath107 .
denote with @xmath108 , @xmath109 the four regions @xmath110 , @xmath111 , @xmath112 , @xmath113 .
[ lem : fixed pts . in quads .
i and iii ] each of the sets @xmath114 and @xmath115 contain at least one equilibrium of eq.(lgin ) .
the existence of an equilibrium of eq.(lgin ) in the invariant and attracting set @xmath116 \times [ h_{2},\,h_{2}\,+\,b_{2 } ] $ ] from lemma [ lem : boundedness ] is guaranteed by the schauder fixed point theorem @xcite for continuous maps on compact , convex and invariant sets . since @xmath117 ,
such equilibrium points are elements of @xmath118 . to see that @xmath115 contains an equilibrium of eq.(lgin ) , note that the critical curves @xmath92 and @xmath93 can be given explicitly as functions of @xmath16 : @xmath119 @xmath120 then , @xmath121 one can conclude from and the continuity of @xmath57 , @xmath122 that there exists @xmath123 such that @xmath124 . since @xmath125 is a vertical asymptote of @xmath92 and @xmath126
is a horizontal asymptote of @xmath93 , it follows from the decreasing characters of @xmath57 and @xmath127 that @xmath128 must lie in @xmath115 .
[ def : multiplicity ] let @xmath129 be a positive integer .
let @xmath130 be an intersection point of the graphs @xmath92 and @xmath93 of two @xmath129-times differentiable functions @xmath131 defined in a neighborhood of @xmath132 .
the point @xmath133 is a _ contact point of @xmath92 and @xmath93 of order @xmath129 _ if @xmath134 note that @xmath92 and @xmath93 intersect transversally at @xmath133 if and only if @xmath133 is a contact point of @xmath92 , @xmath93 of order one .
[ lem : contact pt . - multiplicity ] let @xmath135 be an intersection point of two non - singular algebraic curves @xmath41 and @xmath42 in the plane that have no common component through @xmath136
. if @xmath41 and @xmath42 have tangent lines at @xmath136 that are not parallel to either axis , then there exists a neighborhood @xmath137 of @xmath136 for which both @xmath138 and @xmath139 are the graphs of real analytic strictly monotonic functions @xmath50 and @xmath51 for @xmath16 in some neighborhood of @xmath140 , and the order of @xmath136 as a contact point of @xmath41 and @xmath42 equals the multiplicity of @xmath136 as an intersection point of the algebraic curves @xmath41 and @xmath42 . without loss of generality
we may assume @xmath141 .
let @xmath142 and @xmath143 be such that @xmath144 and @xmath145 are parametrized by @xmath146 and @xmath147 for @xmath148
. then @xmath142 and @xmath143 are real analytic on @xmath10 .
the rest of the statement follows .
[ lem : d_x = d_y = k ] let @xmath149 and @xmath150 be real analytic strictly monotone functions defined in neighborhoods of @xmath140 and @xmath151 respectively , such that @xmath152 , @xmath153 , and neither @xmath154 nor @xmath155 is the identity function .
let @xmath129 be the order of @xmath156 as a contact point of @xmath157 and @xmath158 , and let @xmath159 and @xmath160 be the multiplicities of @xmath151 and @xmath140 as zeros of @xmath161 and @xmath162 respectively
. then @xmath163 .
consider the function @xmath164 , where @xmath165 is the inverse of the strictly monotone function @xmath149 .
clearly , @xmath166 in a neighborhood of @xmath167 .
it follows that @xmath168 for @xmath169 .
in particular , @xmath170 from which we get @xmath171 now consider the function @xmath172 .
since @xmath167 is a root of @xmath173 of multiplicity @xmath174 by hypothesis , we must have @xmath175 for @xmath176 and @xmath177 .
in particular , @xmath178 from which we have @xmath179 from and , it follows that @xmath180 .
similarly , one can show that @xmath181 .
however , since @xmath182 , we must have @xmath183 .
hence @xmath184 it is a direct consequence of definition [ def : multiplicity ] and that @xmath185 . by a similar argument using the equation defining @xmath93
, one can show that @xmath186 .
the main result of this paper is the following .
[ th : main result 1 ] the following statements are true : * eq.(lgin ) has at least one and at most three equilibria in @xmath187 .
the set of equilibrium points in @xmath187 is linearly ordered by @xmath27 . * if eq.(lgin ) has exactly one equilibrium in @xmath187 , then it is is globally asymptotically stable . *
if eq.(lgin ) has three distinct equilibria in @xmath187 , say @xmath188 , with + @xmath189 then @xmath190 and @xmath191 are locally asymptotically stable , while @xmath192 is a saddle point .
the global stable manifold of @xmath192 is the graph of a continuous increasing function of the first variable with endpoints in the boundary of @xmath187 , which is a separatrix of the basins of attraction of @xmath190 and @xmath191 .
* if there exist exactly two equilibria in @xmath187 , then one is locally asymptotically stable and the other is a nonhyperbolic fixed point .
if @xmath193 and @xmath194 are the two equilibria with @xmath195 , then @xmath196 is a subset of the basin of attraction of @xmath193 , and @xmath197 is a subset of the basin of attraction of @xmath194 .
* it is a consequence of bzout s theorem ( theorem 3.1 , chapter iii in @xcite ) that the hyperbolas @xmath92 and @xmath93 given in ( [ eq : critical curves system ] ) must intersect in at most four points .
since intersection points of @xmath92 and @xmath93 are precisely the equilibrium points of eq.(lgin ) , it follows that eq.(lgin ) must have at most four equilibrium points .
lemmas [ lemma : t is bounded ] and [ lem : fixed pts . in quads .
i and iii ] guarantee that at most three of these equilibria can lie in the invariant attracting box @xmath198\times[h_1+b_1,h_2+b_2]$ ] and hence in @xmath114 , which proves the first part of statement ( i. ) . to see that the set of equilibrium points in @xmath187 is linearly ordered by @xmath27 , note that the equilibria are precisely the intersection points of the decreasing critical curves @xmath92 and @xmath93 given in lemma [ lem : hyperbolas ] . * let @xmath20 be the map of eq.(lgin ) , and let @xmath199 . by lemma [ lemma :
t is bounded ] and since @xmath20 is competitive , @xmath200 . by induction ,
@xmath201 for @xmath202 .
then the sequences @xmath203 and @xmath204 are respectively monotonically increasing and decreasing , they are bounded and converge ( this follows from the bounded and monotonic coordinate - wise character of the sequences ) .
since both must converge to a fixed point , the limit point is the unique fixed point of @xmath20 .
again by lemma [ lemma : t is bounded ] , for @xmath199 , @xmath205 , hence @xmath206 , and by induction , @xmath207 for @xmath202 . thus @xmath208 as @xmath209 .
this proves global attractivity of @xmath55 stability follows from the fact that @xmath210 for @xmath211 , @xmath202 , and @xmath212 .
* suppose eq.(lgin ) has three distinct equilibria in @xmath187 , say @xmath213 , with @xmath214 .
note that the equilibria are precisely the intersection points of the critical curves @xmath92 and @xmath93 given in ( [ eq : critical curves system ] ) . from bezout
s theorem , lemma [ lem : fixed pts . in quads . i and iii ] and lemma [
lem : contact pt .
- multiplicity ] , each equilibrium in @xmath187 must be a contact point of @xmath92 and @xmath93 of order one .
it follows from the remark after definition [ def : multiplicity ] that @xmath92 and @xmath93 must intersect transversally at each of the three equilibria in @xmath187 .
furthermore , solving for y and x respectively in the equations defining @xmath92 and @xmath93 in @xmath215 gives that the vertical asymptote of @xmath92 is @xmath125 and the horizontal asymptote of @xmath93 is @xmath126 .
the asymptotes guarantee that in order to have three intersection points in @xmath187 , the slopes of the functions @xmath216 and @xmath217 of @xmath92 and @xmath93 respectively , must satisfy the relations @xmath218 and @xmath219 .
theorem [ thm : main theorem ] then gives that @xmath190 and @xmath191 must be locally asymptotically stable , while @xmath192 must be a saddle point .
the rest of the proof follows from theorem 4 and corollary 1 in @xcite .
* suppose eq.(lgin ) has two distinct equilibria in @xmath187 , say , @xmath190 and @xmath220 .
note that @xmath190 and @xmath220 can not both be contact points of @xmath92 and @xmath93 of order one .
indeed if they were , then as a result of the remark after definition [ def : multiplicity ] , @xmath92 and @xmath93 would have to intersect transversally at @xmath190 and @xmath220 .
thus by theorem [ thm : main theorem ] , both equilibria would have to be locally asymptotically stable .
but this would lead to a contradiction since by theorem 4 of @xcite , at least one of the equilibria has to be a non - attracting equilibrium .
suppose @xmath190 and @xmath220 are contact points of @xmath92 and @xmath93 of orders two and one respectively .
then by the remark after definition [ def : multiplicity ] , @xmath92 and @xmath93 must intersect tangentially at @xmath190 and transversally at @xmath220 .
hence by theorem [ thm : main theorem ] , @xmath190 must be a non - hyperbolic fixed point and @xmath192 must be locally asymptotically stable .
the statement on basins of attraction is a consequence of @xmath20 being competitive and the fact that @xmath196 and @xmath197 are invariant sets for @xmath20 . indeed , if @xmath221 then @xmath222 .
since @xmath223 converges to a fixed point by theorem [ thm : convergence ] , it must converge to the only fixed point in @xmath196 , namely , @xmath193 .
a similar argument applies if @xmath224 .
a version of the following theorem was proved first in @xcite .
it is given here for easy reference .
[ thm : m - m theorem ] let @xmath225 $ ] and @xmath226 $ ] be intervals of real numbers and let @xmath227\rightarrow [ a , b]$ ] and @xmath228\rightarrow [ c , d]$ ] be continuous functions that satisfy the following properties ( a ) and ( b ) : * suppose @xmath233 from direct inspection of ( [ eq : jacobian ] ) one can see that the functions @xmath15 and @xmath18 in @xmath74 satisfy hypothesis ( a ) of theorem [ thm : m - m theorem ] . to verify hypothesis ( b ) of theorem [ thm : m - m theorem ] , note that for @xmath20 as in ( [ eq : map ] ) , the system of equations in ( [ eq : mm system ] ) is given by @xmath234 algebraic manipulation of the equations in ( [ eq : m - m theorem 1 ] ) yields the equation @xmath235 \\
\,+\,c_{1}(\overline{m}\,-\,\overline{m})\,[(\overline{m}\,-\,h_{2})\ , + \ , ( \overline{m}\,-\,h_{2})\,+\,(1\,-\,b_{2 } \,+\,h_{2}\,+\,c_{2}\,h_{1 } ) ] = 0 \end{array}\ ] ] using hypothesis ( [ eq : hypothesis ] ) and the facts @xmath236 and @xmath237 , we obtain that in equation ( [ eq : m - m theorem 2 ] ) , @xmath238 and @xmath239 which shows that hypothesis ( b ) holds .
then theorem [ thm : m - m theorem ] implies that there is a unique equilibrium .
* suppose @xmath240 .
applying the transformation @xmath241 to the equations defining the sets @xmath92 , @xmath93 in lemma [ eq : critical curves system ] gives rise to the equations @xmath242 where @xmath243 and @xmath244 are functions of @xmath245 .
we consider two cases .
+ case 1 : suppose @xmath246 .
since the leading coefficient of ( [ eq : new system]a ) is @xmath247 and the constant coefficient is @xmath248 , the product of the roots of ( [ eq : new system]a ) is @xmath249 .
similarly , the product of the roots of ( [ eq : new system]b ) is @xmath250 . from lemma [ lem : fixed pts . in quads .
i and iii ] , each equation in must have a positive root @xmath251 and a negative root @xmath252 .
furthermore , none of the two equations can have complex roots . indeed
if , say , ( [ eq : new system]a ) had a pair of complex conjugate roots @xmath253 and @xmath254 , then the product of the roots of ( [ eq : new system]a ) given by @xmath255 would be negative , which is impossible .
it follows that the set of roots of each equation in must have the sign structure @xmath256 . as a result ,
the set of equilibria of eq.(lgin ) must have one of the sign structures ( i. ) @xmath257 or ( ii . ) @xmath258 .
the sign structure given by ( i. ) is not possible since otherwise , there must exist two equilibria of eq.(lgin ) in @xmath259 and two equilibria outside @xmath259 . by the proof of theorem [ th : main result 1 ]
, one of the equilibria in @xmath259 must have multiplicity two while the other equilibrium in @xmath259 must have multiplicity one .
thus the sum total of the multiplicities of all the equilibrium points must be at least five , contradicting bzout s theorem .
hence when @xmath246 , the only possible sign structure for the set of equilibria of eq.(lgin ) is ( ii . ) , which implies that eq.([eq : new system ] ) has a unique equilibrium in @xmath260 , and hence eq.(lgin ) has a unique equilibrium in @xmath261 .
+ case 2 : suppose @xmath262 : in this case , the equations in reduce to @xmath263 where @xmath264 and @xmath265 are functions of @xmath266 .
we consider three cases : @xmath267 is negative , zero , and positive .
if @xmath268 , the product of the roots of the first equation in is @xmath269 and the product of the roots of the second equation in is @xmath270 .
note that from lemma [ lem : fixed pts . in quads .
i and iii ] , we have that each equation in must have a positive root and a negative root .
thus the equations in can not have complex roots due to their cubic nature , since complex roots of real polynomials always occur in conjugate pairs .
if @xmath271 , the sets of roots of the equations in must have sign structures @xmath272 and @xmath273 , respectively , by an argument similar to the one already used in case 1 .
hence the set of equilibria of eq.(lgin ) must have the sign structure @xmath274 .
if @xmath275 , the sets of roots of the equations in must have sign structures @xmath273 and @xmath272 , respectively .
hence the set of equilibria of eq.(lgin ) must have the sign structure @xmath276 . if @xmath277 , the equations in reduce to a pair of quadratic equations .
it follows from lemma [ lem : fixed pts . in quads .
i and iii ] that each equation must have a positive root and a negative root .
so the set of roots of each equation must have the sign structure @xmath278 .
hence the set of equilibria of eq.(lgin ) must have the sign structure @xmath279 , which implies that eq.([eq : new system ] ) has a unique equilibrium in @xmath260 , and hence eq.(lgin ) has a unique equilibrium in @xmath261 .
cushing et .
al . , some discrete competition models and the competitive exclusion principle , _ journal of difference equations and applications , vol . 10 , no . 13 - 15 , november - december 2004 _ , pp . 1139 - 1151 . | we establish the relation between local stability of equilibria and slopes of critical curves for a specific class of difference equations .
we then use this result to give global behavior results for nonnegative solutions of the system of difference equations @xmath0 with positive parameters .
in particular , we show that the system has between one and three equilibria , and that the number of equilibria determines global behavior as follows : if there is only one equilibrium , then it is globally asymptotically stable .
if there are two equilibria , then one is a local attractor and the other one is nonhyperbolic .
if there are three equilibria , then they are linearly ordered in the south - east ordering of the plane , and consist of a local attractor , a saddle point , and another local attractor . finally , we give sufficient conditions for having a unique equilibrium .
+ * key words : * difference equation , rational , global behavior , global attractivity , model , competitive system , leslie - gower model . + * ams subject classification : * 39a05 , 39a11 + |
vanderveld _ et al .
_ @xcite ( abbreviated as vfw ) claimed that there was a contradiction between two results concerning the putative accelerated expansion of the universe .
their reasoning was , in brief , this ( italics mark quotations ) : on the one hand , hirata and seljak @xcite ( abbreviated as hs ) claimed to have proved that in a perfect fluid cosmological model that is geodesic , rotation - free and obeys the strong energy condition @xmath5 , a certain generalisation of the deceleration parameter @xmath6 must be non - negative .
but on the other hand , iguchi _ et al . _
@xcite ( hereafter inn ) did obtain simulated acceleration in lematre
tolman ( l t ) models @xcite with @xmath7 that obey hs s conditions .
this _ contradiction _ is resolved by showing that l
t models that simulate accelerated expansion also contain a _ weak singularity _ , and in this case the derivation of hs breaks down .
in addition to this , there are _ other singularities that tend to arise _ in l
t models , and vfw _ have failed to find any singularity - free models that agree with observations .
_ this leaves the impression that physically acceptable inhomogeneous models are unable to account for observations .
it is shown here that this reasoning is not correct . in brief , our theses are the following : \1 .
of the five definitions of the deceleration parameter given by hs , only @xmath0 is a correct invariant measure of deceleration of fluid expansion .
their @xmath3 derives from a taylor expansion of the luminosity distance
redshift relation , @xmath8 , in powers of @xmath9 , done separately for each direction .
their @xmath2 is based on the angular average of a taylor expansion of @xmath10 , and is conceptually different from what is done in observational practice .
although all these definitions reduce to the familiar @xmath6 in flrw models , they have different values and distinct meanings in an inhomogeneous model .
in particular @xmath3 and @xmath2 are not measures of deceleration in an inhomogeneous model .
hs s @xmath2 does not correctly represent the observers deceleration parameter @xmath6 .
when observers assume that our real universe is in the flrw class , @xmath6 is the same in all directions .
but when they consider inhomogeneous and anisotropic models , they have to measure the distance
redshift relation for each direction separately rather than average the measured results over directions , as this would mean destroying useful information .
thus , hs s @xmath2 that represents the result of such averaging does not correspond to the observations actually done in relation to this quantity .
hs s reasoning is correct ( in the approximate sense ) up to their eq .
( 31 ) , which allows one to calculate @xmath3 at the centre of symmetry in a lematre tolman ( l t ) model .
but , at a spherical centre , the final result is the opposite to what they intended no limitation on the sign of @xmath11 follows ( see our secs .
[ hsproblems ] and [ fproblems ] ) .
thus , there is no contradiction involved in a decelerating inhomogeneous model imitating observational relations of an accelerating flrw model .
even so , it is incorrect to use an approximate equation to decide whether some quantity is positive or negative .
what vfw call a _ weak singularity _ is not a singulalrity .
it is also not true that _ other singularities _ invalidate the l
the differential equations for the _ inverse problem _ given observational data functions , calculate the lt model that would give them are indeed singular at the apparent horizon .
vfw refer to this location as a _ critical point _ , and to this phenomenon as a _ pathology_. in truth , this is simply the reconvergence of the observer s past null cone towards the big bang , that is a consequence of the decelerating cosmic expansion when lambda is zero , long known in the flrw case , e.g @xcite . though some authors ( inn and vfw among them ) were unable to propagate their solutions through the ` critical point ' , this difficulty was overcome by lu and hellaby @xcite and in fact the ah relation was used to provide extra information by mcclure and hellaby @xcite .
since these misunderstandings have propagated into the literature , it is essential to resolve these issues .
but the most important point to be stressed is this : what one wants to reproduce , in connection with the distant type ia supernovae , is not the accelerated expansion of the universe , but the observed luminosity distance
redshift relation .
the apparent dimming of supernovae ( first reported by riess _
_ @xcite and perlmutter _ et al . _
@xcite ) stems from a comparison between observations and the einstein - de sitter model , the ` old ' standard model of the universe .
the first proposed explanation , originating from the assumption that the universe should be friedmannian , is an accelerated expansion .
however , this acceleration is not the phenomenon to be explained , but a component of a particular explanation one that assumes homogeneity on the scale in question .
since in the following we often refer to the l t model , we summarise here the basic facts about it in the case @xmath7 . for more extended expositions see refs .
the metric of the l
t model is : @xmath12 where @xmath13 is an arbitrary function , and @xmath14 is determined by the integral of the einstein equations : @xmath15 @xmath16 being another arbitrary function .
equation ( [ 2.2 ] ) has the same algebraic form as one of the well - known friedmann equations , except that here it contains arbitrary functions of @xmath17 in place of arbitrary constants .
the friedmann limit follows when @xmath18 and @xmath19 where @xmath20 , @xmath21 and @xmath22 are the corresponding friedmann constants and scale factor .
the solution of ( [ 2.2 ] ) may be written as @xmath23 where @xmath24 is one more arbitrary function called the bang - time function ; in the friedmann limit it is constant .
the @xmath25 sign applies for an expanding region , @xmath26 applies for a collapsing region .
the mass density is @xmath27 the pressure is zero , and so the matter ( dust ) particles move on geodesics .
the equations determining the l t model are covariant with the transformations @xmath28 .
therefore , we may use such a transformation to give one of the functions @xmath29 a handpicked form , provided the chosen function is monotonic in the range under investigation .
an incoming radial null geodesic is given by the differential equation @xmath30 and its solution is denoted @xmath31 .
the general solution determines the past null cone ( pnc ) of the observer situated at the centre of symmetry .
we use the hat @xmath32 and the subscript @xmath33 to indicate evaluation on the pnc
. then the redshift along the pnc , @xmath34 , is given by @xcite : @xmath35_\wedge ~,\ ] ] and the area and luminosity distances are @xmath36 equations ( [ 2.5 ] ) and ( [ 2.6 ] ) can be solved numerically once @xmath16 , @xmath13 and @xmath24 are specified .
t model may possibly have a curvature singularity at the centre of symmetry .
the singularity will be absent when there is no point mass there , and the mass density is finite .
this will happen when the conditions given below are obeyed ( see ref .
@xcite for the derivation ) .
let @xmath37 be the radial coordinate of the centre of symmetry , where @xmath38 for all @xmath39 , and let @xmath40 .
we choose the @xmath17-coordinate so that @xmath41 and @xmath42 where @xmath43 constant
. then , in a neighbourhood of @xmath44 , the functions behave as follows : @xmath45 @xmath46 and @xmath47 being constants , where the symbols @xmath48 will denote quantities with the property @xmath49
hirata and seljak @xcite give in their paper five alternative definitions of the deceleration parameter , three of which we briefly recall here for reference . in all cases
it is assumed that the spacetime is filled with a perfect fluid obeying einstein s equations with zero cosmological constant .
hs assume that the fluid moves geodesically , but for the beginning we shall drop this assumption .
their first definition of the deceleration parameter is : @xmath50 where @xmath51 is the ( local ) hubble parameter and @xmath52 is the velocity field of the fluid .
then , using these definitions , the raychaudhuri equation may be rewritten as follows : @xmath53 + \frac { 4 \pi g } 3 ( \rho + 3p),\ ] ] where @xmath54 is the acceleration of the fluid , @xmath55 and @xmath56 are the shear and rotation tensors , respectively , @xmath57 is the mass density and @xmath58 is the pressure .
it follows that with @xmath59 and @xmath60 , @xmath0 must be non - negative .
this result applies in particular to the l
t models ( where @xmath61 and @xmath62 ) and to the flrw models ( where @xmath63 ) .
the definition ( [ 3.1 ] ) is invariant , applies in all geometries , is a generalization of the deceleration parameter long used in the flrw geometries , and measures the relative change of the expansion scalar along the flow lines of the perfect fluid . with @xmath64 we have @xmath65 , so the expansion slows down toward the future of the comoving observers .
this is the case in all cosmological models that obey the assumptions listed above .
the quantities @xmath3 and @xmath66 refer to a taylor expansion of the luminosity distance - redshift relation @xmath10 : @xmath67 where @xmath66 and @xmath3 are functions of the angle of observation .
the definitions of @xmath68 and @xmath2 follow by averaging ( [ 3.3 ] ) over the full solid angle @xmath69 at each fixed @xmath70 : @xmath71 these equations have the same algebraic form as the corresponding equation in an flrw model .
by considering a bundle of null geodesics converging to a point in space , hirata and seljak @xcite derived an equation relating the quantities in ( [ 3.4 ] ) to the invariants of flow of a perfect fluid filling the spacetime .
flanagan s ( hereafter f , @xcite ) equation is more general , and it reads , in hs s notation , @xmath72.\ ] ] ( hs s result is the subcase @xmath73 . ) equation ( [ 4.1 ] ) has the appearance of being exact and covariant .
unfortunately , it is neither .
as we show below , its derivations , both by hs and by flanagan , contain approximations ( some not explicitly spelled out ) and involve averaging over directions , which produces coordinate - dependent results at locations corresponding to coordinate singularities .
approximations may be good for some purposes , but not when one intends to show that some quantity is positive .
below we note instances where hs and f introduced approximations in their reasoning . we have no objections to the reasoning of hs up to their eq .
having established the relation between the luminosity distance @xmath70 and the redshift @xmath9 ( their ( 24 ) ) , and between the radiation amplitude and the expansion scalar @xmath74 of the radiation field ( their ( 25 ) ) , hs then copy a relation between @xmath74 ( @xmath75 in hs s notation ) and the affine parameter @xmath76 from ref .
@xcite ( hs s eq . ( 26 ) ) .
this requires a comment . to obtain this relation
one considers another family of rays , originating at the centre of a source ( call it a star for brevity ) .
the star is assumed spherical and having the radius @xmath77 .
one of the rays hits the observer .
the values of the affine parameter along this ray are @xmath78 at the observer , @xmath79 at the surface of the star and @xmath80 at the centre of the star .
the relation is then @xmath81.\ ] ] in ref .
@xcite this is obtained as an approximate solution of the raychaudhuri equation for null geodesics , which is @xmath82 where @xmath83 , @xmath84 and @xmath85 are , respectively , the expansion , shear and rotation of the null congruence @xmath86 and @xmath87 is the ricci tensor in spacetime . equation ( [ 4.2 ] ) without the second term on the right is the exact solution of ( [ 4.3 ] ) in the case @xmath88 .
thus the approximation involved in ( [ 4.2 ] ) is that the contributions to @xmath89 from shear , rotation and lensing by matter are negligible compared to @xmath90 .
these are additional assumptions about light propagation in spacetime ( @xmath84 and @xmath85 negligible ) and about the spacetime itself ( @xmath91 negligible ) .
however , hs s ( 38 ) is used by vfw as if it were exact . when @xmath92 hold exactly , the goldberg
sachs theorem @xcite applies , which says that the spacetime must be algebraically special , with @xmath93 being the double principal null congruence of the weyl tensor .
the assumption @xmath94 , when imposed on the l
t spacetime , reduces it to the schwarzschild solution .
therefore , hs s equation ( 38 ) , when treated as exact , involves strong assumptions that do nt apply to the l
t model , or even the flrw model .
the two errors shown below have no meaning for the final result because ( in the approximation in which hs worked ) they cancelled out in the end , but they might be confusing . hs s equation ( 27 ) should actually read @xmath95 i.e. it is the inverse of what they wrote . similarly
, their equation ( 28 ) should actually read @xmath96 ( where @xmath97 is the radiation amplitude ) , i.e. again the inverse of their result . since hs assumed @xmath98 , the above
is approximately equal to @xmath99 which results in hs s eqs .
( 29 ) ( 31 ) being right again .
since we will later refer to their ( 31 ) , we copy it here : @xmath100 the meaning of the symbols in ( [ 4.7 ] ) is as follows .
since hs work under the assumption that @xmath101 for the cosmic fluid , comoving and synchronous coordinates exist in which the metric @xmath102 has the properties @xmath103 , @xmath104 , where @xmath105 is the time coordinate and @xmath106 , @xmath107 , are the space coordinates .
then @xmath108 is the metric of a space of constant @xmath109 , and @xmath110 is the second fundamental form of this space ; in these coordinates @xmath111 . then the geodesic null vector @xmath112 can be normalised at the observer so that @xmath113 and its space components @xmath114 form a 3-dimensional unit vector , @xmath115 .
there is a problem with averaging products like @xmath116 and @xmath117 at coordinate singularities in space , and the centre of symmetry in l
t is a singularity of the spherical coordinates .
for example , a general unit vector in euclidean space attached at a point @xmath118 , when referred to cartesian coordinates and parametrised by spherical angles @xmath119 , has the components @xmath120 . in this case , the averages ] and similarly for @xmath121 .
] are @xmath122 , @xmath123 , @xmath124 .
the first and last of these are special cases ( in euclidean space ) of the covariant equations @xmath125 however , these results of averaging , change when we transform to spherical coordinates centred at @xmath118 .
when the components of the unit vector @xmath126 in euclidean space are referred to the spherical coordinates , they are @xmath127 ( with @xmath128 ) , and then the averages are @xmath129 , @xmath130 .
explicit calculation done in sec .
[ anysign ] for the l
t model will show that the general formulae ( [ 4.8 ] ) ( [ 4.10 ] ) , in particular ( [ 4.9 ] ) , lead to incorrect results when applied at the centre of symmetry . since ( [ 4.1 ] ) is supposed to follow from ( [ 4.7 ] ) by this kind of averaging , the argument above invalidates ( [ 4.1 ] ) at the centre of symmetry of the l
t model even as an approximate equation .
we will now use ( [ 4.7 ] ) to estimate hs s deceleration parameter @xmath3 at the centre of the l
t model without averaging over directions . identifying the coefficient of @xmath70 in ( [ 4.7 ] ) with the @xmath66 of ( [ 3.3 ] ) we find from ( [ 3.3 ] ) : @xmath131,\ ] ] in the coordinates of ( [ 2.1 ] ) we have , with @xmath132 : @xmath133 the components not listed are zero . at a general point of an l
t manifold , the unit spatial vector @xmath126 must obey @xmath115 , and so , in consequence of ( [ 5.2 ] ) , its components are constrained by @xmath134}.\ ] ] at the centre of symmetry each vector @xmath126 is necessarily radial , so @xmath135 , and ( [ 5.6 ] ) gives @xmath136 which is finite by ( [ 2.9 ] ) and ( [ a.4 ] ) . however , ( [ 5.6 ] ) and ( [ 5.7 ] ) imply that at the centre of symmetry the angle - average of ( [ 4.7 ] ) is discontinuous .
namely , as long as the vector @xmath126 is attached off the centre , its @xmath137 component may point in the direction of increasing @xmath17 ( in which case we take ( [ 5.6 ] ) with the @xmath25 sign ) or toward decreasing @xmath17 ( in which case we take @xmath26 in ( [ 5.6 ] ) ) .
then the average comes out zero .
but when the vector @xmath126 is attached at the centre of symmetry , @xmath138 may point only toward increasing @xmath17 , so @xmath139 on the whole sphere , while @xmath140 .
thus , @xmath141 everywhere on the sphere , while all other components of @xmath117 are zero .
consequently , @xmath142 , while all the remaining averages are zero .
thus , in the spherical coordinates , @xmath143 is discontinuous at the centre of symmetry .
let us apply ( [ 5.1 ] ) at the centre of symmetry in the l t model . using ( [ 5.2 ] ) ( [ 5.7 ] ) and @xmath144 in ( [ 5.1 ] ) we find @xmath145 in order to clearly see the contribution of the inhomogeneous part of the geometry , we will separate @xmath77 and @xmath146 into the part that survives in the friedmann limit and the part that disappears in that limit .
note that the whole calculation will be exact , and our definitions for @xmath147 , @xmath146 and @xmath77 will ensure they automatically obey the conditions of regularity at the centre ( [ 2.9 ] ) .
we define : @xmath148 where @xmath20 is a constant ( the familiar flrw curvature index , so it may be of any sign , also zero ) and @xmath22 , @xmath149 and @xmath150 are functions ( @xmath151 being the flrw scale factor ) .
the terms @xmath152 and @xmath153 represent non - friedmannian contributions to @xmath77 and @xmath146 , respectively ; with @xmath154 the l t model reduces to that of friedmann . in calculating the limits at @xmath155
we will assume that the derivatives of @xmath156 and @xmath157 do not diverge too fast , that is @xmath158 = ( 0 , 0 , 0 , 0 , 0).\ ] ] since some of the functions will be complicated , we introduce the following abbreviations : @xmath159 @xmath160 \label{5.14}\ ] ] ( it can be verified that @xmath161 ) , and the subscript @xmath162 will denote the limit at @xmath155 , thus @xmath163 .
\label{5.16}\end{aligned}\ ] ] the formulae for the derivatives of @xmath77 and their limits at @xmath155 are given in appendix [ rorig ] .
we see from there that all the terms entering @xmath3 in ( [ 5.1 ] ) have nonzero and nondivergent limits at @xmath155 . for @xmath66
we get from ( [ 5.8 ] ) : @xmath164 using ( [ 5.2 ] ) ( [ 5.17 ] ) and ( [ a.1 ] ) ( [ a.8 ] ) in ( [ 5.9 ] ) we now obtain : @xmath165\right\}.\ ] ] the terms proportional to @xmath166 , @xmath167 and @xmath168 can each have any sign , and so each one can make @xmath4 negative
. note that @xmath166 ( which comes from the non - friedmannian contribution to the energy function ) alone can cause @xmath169 ( when @xmath170 ) , and so can @xmath167 alone ( when @xmath171 ) . at the centre of spherical symmetry ,
this confirms that hs s result is not correct at @xmath44 , and a negative @xmath2 , as found by inn , is perfectly possible .
the reason why hs concluded that @xmath173 is that they found @xmath174 while averaging ( [ 5.1 ] ) which is not true at a spherical centre in spherical coordinates . in consequence , they dismissed all terms in ( [ 5.18 ] ) which come from @xmath175 , i.e. all except the first one . with only the first term , one obviously gets @xmath176 .
since nothing can be concluded in general about the sign of @xmath4 , there would be no contradiction involved in certain l
t models imitating a negative flrw deceleration parameter even if eq .
( [ 5.1 ] ) were exact .
but we recall : this equation being approximate , using it to evaluate the sign of @xmath3 can not lead to an unambiguous conclusion , so the alleged contradiction was a nonexistent problem .
there is one more piece of evidence that hs s method of averaging is not universally correct .
suppose ( [ 5.1 ] ) is applied at the centre of symmetry of an l
. then each component of the sum in ( [ 5.1 ] ) is invariant under the group @xmath177 .
thus , averaging over directions should be an identity operation and change nothing .
this is indeed the case when one uses ( [ 5.3 ] ) ( [ 5.5 ] ) for @xmath110 and @xmath178 , applied at @xmath44 , and ( [ 5.7 ] ) for @xmath179 ; actually one gets ( [ 5.17 ] ) ( [ 5.18 ] ) again . however , if one uses hs s prescription for averaging , then the whole term @xmath180 still drops out in spite of being spherically symmetric and nonzero initially .
flanagan s paper @xcite addressed the question of whether superhorizon perturbations can cause apparent acceleration .
he concluded that if their effect on @xmath6 is negative , it must be too small to _ be responsible for the acceleration of the universe . _
the paper was not intended to be applied to exact inhomogeneous cosmological models , so the author should not be blamed for the incorrect use made later of his equation .
but the hs subcase of that approximate equation was used by vfw as if it were exact . in this section
we follow f s reasoning in order to identify the approximations assumed along the way . the correct form of flanagan s eq .
( 9 ) is : @xmath181 the meaning of the symbols is as follows : @xmath182 is the set of coordinates of the observer at the point @xmath183 , @xmath184 is the set of coordinates of the light source at the point @xmath185 , @xmath186 is the four - velocity of the cosmic medium at @xmath183 ( the observer is comoving ) , @xmath187 is the four - velocity of the medium parallel - transported from @xmath185 to @xmath183 along the light ray , @xmath188 is synge s @xcite world function for @xmath183 and @xmath185 , and the coefficients are @xmath189 , @xmath190 . the quantity in ( [ 6.1 ] )
is apparently only needed to calculate the redshift , which f writes as @xmath191 where @xmath192 is the tangent vector to the ray at @xmath183 .
usually the numerator is taken simply at @xmath185 , without parallel - transporting it to @xmath183 @xcite .
nevertheless , it is easy to verify that ( [ 6.2 ] ) is equal to the usual formula .
f then considers two families of light rays : one converging at @xmath183 , with the affine parameter @xmath193 and tangent vectors @xmath20 normalised so that @xmath194 at @xmath183 , and another one diverging from @xmath185 , with the affine parameter @xmath195 and tangent vectors @xmath196 normalised so that @xmath197 at @xmath185 . in consequence of this , and of ( [ 6.2 ] ) , the two affine parameters are related by @xmath198 below his ( 13 ) , f writes : _ we choose the normalization of @xmath97 so that @xmath199 for @xmath200 near @xmath185 .
_ this is an assumption equivalent to hs s ( 26 ) our ( [ 4.2 ] ) ( @xmath97 is the same quantity as in ( [ 4.5 ] ) ) , and f makes use of it further on ( see below our eq . ( [ 6.7 ] ) ) .
this _ normalization _ also defines the units for @xmath195 as the units of distance . in the paragraph of f s paper that contains eqs .
( 14 ) ( 16 ) the quantities referring to @xmath183 and to @xmath185 are mixed up .
the text below is our interpretation of this segment of f s reasoning , in which we take care about properly distinguishing between these two points . from the definition of the energy flux at @xmath183
one obtains : @xmath201 whereby a rewriting error was corrected above : f s @xmath202 should read @xmath203 .
the following information was fed into ( [ 6.4 ] ) at the consecutive equality signs : the definition of energy flux , the definition of the radiation energy - momentum tensor at @xmath185 , equation ( [ 6.2 ] ) , the normalization of @xmath112 at @xmath183 . integrating the energy flux _ at @xmath185 _ over a sphere of small radius @xmath204 , one obtains for the luminosity at @xmath185 , again from the definition : @xmath205 the simplification occurs because of the normalization assumed at @xmath185 . putting ( [ 6.2 ] ) and ( [ 6.3 ] ) into the definition of the luminosity distance ( which we denote here by @xmath70 for consistency with the previous sections ) one obtains @xmath206 in the next step f uses the assumed normalization of @xmath207 to write the above as @xmath208 , but then gives up on it and uses ( [ 6.3 ] ) to rewrite ( [ 6.6 ] ) as follows : @xmath209 ( the second equality follows by the assumed normalization of @xmath97 at @xmath185 ) .
then f cites visser @xcite for the result @xmath210 thus finally , up to @xmath211-terms : @xmath212 which is the same equation as hs s ( 29 ) . just as in the hs paper ,
this equation is approximate , where in addition to hs s ( 26 ) ( our ( [ 4.2 ] ) ) , flanagan used one more approximation our ( [ 6.8 ] ) . from ( [ 6.9 ] ) and from eqs .
( 1 ) and ( 11 ) in f s paper , f s equations ( 15 ) and ( 16 ) follow by simple substitution .
below his eq .
( 16 ) , flanagan notes that there is no unique way to choose the order in which the coefficients in the @xmath8 equation should be averaged over directions , but dismisses this problem by saying : _ thus the different averaging prescriptions give different answers .
however the fractional differences are of order @xmath213 , which we have argued above is of order @xmath214 and is small .
_ this is correct in the context of his paper .
however , this indicates that there is a further approximation involved in the reasoning , and this one is quite arbitrary and unpredictable .
namely , it depends on the will of the person doing the calculation exactly which quantity he or she wishes to average first .
f himself uses @xmath215 to obtain the @xmath216 of his eq .
( 4 ) , but @xmath217 for computing his @xmath218 .
this is perhaps the strongest indication that it makes no sense to use the resulting final equation for determining the sign of any quantity .
note that the same problem exists for the hs derivation , but it was not mentioned in the hs paper .
the averaging over directions is defined only in a certain 3-space @xmath219 , not in spacetime . without fixing @xmath219 ,
the directional angles of the light rays are not defined . when the rotation of the cosmic fluid is zero , as in hs s paper ,
@xmath219 is the 3-space orthogonal to the fluid s flow lines . however ,
when the cosmic fluid has nonzero rotation , such @xmath219 do not exist not even locally , because the 3-volume elements locally orthogonal to the flow lines , when followed around any flow - tube , refuse to connect up to a 3-space .
one could consider a 3-space of constant time - coordinate , which is not orthogonal to the flow lines , but then the metric in this space is not the @xmath220 assumed by f. thus , in the presence of rotation , the whole calculation should be reformulated . without that , for the averaging over angles considered by f there exists no space in which it occurs .
in the paragraph containing their eq . ( 2.18 ) vfw say : _ we expand the density ( 2.5 ) to second order in @xmath17 as @xmath221 the weak singularity occurs when @xmath222 is nonzero , in which case the gravitational field is singular since @xmath223 as @xmath224 , where @xmath225 is the ricci scalar .
in other words , second derivatives of the density diverge at the origin , independent of where observers may be located .
this is true both in flat spacetime and in the curved ltb metric when we have a density profile of the form ( 2.18 ) .
the singularity is weak according to the classification scheme of the literature on general relativity [ 28 ] . _
[ 28 ] in this quotation is @xcite . ) in truth , whether a curvature singularity is there is decided by curvature alone , and not by secondary constructs like the dalembertian of the curvature scalar .
we do not know about any physical or geometrical interpretation of this quantity , and vfw do not mention any , nor have they shown that their ` weak singularity ' causes any problems at the origin . as we show below the curvature of the l
t model at the centre of symmetry is nonsingular provided that @xmath57 is finite there . for the metric ( [ 2.1 ] )
, the orthonormal tetrad components of the curvature tensor , in the basis defined by @xmath226 , @xmath227 , @xmath228 , @xmath229 , are @xmath230 the quantities given above are all scalars , so any scalar polynomial in the curvature components will be a polynomial in the quantities given in ( [ 7.1 ] ) . if these are nonsingular , then there will be no scalar polynomial curvature singularity .
let @xmath37 be the radial coordinate corresponding to the centre of symmetry @xmath231 .
the requirement of no point - mass at the centre implies @xmath232 .
it can be seen that with @xmath233 all we need to make the curvature nonsingular at @xmath37 is a finite limit of @xmath234 at @xmath37 .
we have , using ( [ 2.4 ] ) @xmath235 so the curvature is finite at @xmath37 if and only if @xmath236 is .
as regards geodesic completeness , the geodesic equation is well behaved through the origin .
geodesics passing through the origin must be purely radial , and radial geodesics stay radial .
the only christoffel symbols that diverge at @xmath237 are @xmath238 , but these do not appear in the equations of radial geodesics , so can not cause any problem .
the condition @xmath223 is not a criterion for a singularity in any accepted sense .
the paper of tipler @xcite contains a proposal of a definition of a _ strong singularity _ , but does not mention a ` weak singularity ' anywhere and contains no ` classification scheme ' . for a singularity to be ` weak ' ( i.e. not strong ) , it first has to be a singularity .
an accurate description of the feature that vfw had in mind is ` the density profile is not @xmath239 through the origin ' .
therefore , vfw s _ weak singularity _ is not a singularity , so its presence is no reason to dismiss models containing it .
there is nothing wrong with a jump in the density gradient at the centre .
since our galaxy is centrally concentrated and has a central black hole , a pointed density profile is a good approximation .
the nfw density profile @xcite for galaxy clusters is divergent at the centre .
a jump in the density and its gradient exists , for example , at the surface of the earth , thus we deal with a ` weak singularity ' in everyday life .
a large part of vfw s paper is devoted to considering the _ inverse problem _ and arguing that only flrw models have rays that cross the apparent horizon what they call the _
critical point_. we here dispel the impression they convey , that _ transcritical _ solutions are well - nigh impossible to find for non - homogeneous models that account for observations .
we emphasise ( a ) that all l
t models with a decelerating expansion phase have apparent horizons that are crossed by large families of light rays , and ( b ) no cosmological model has been shown to solve the inverse problem ( as distinct from the fitting problem ) relative to real data .
vfw correctly point out that the differential equations ( des ) of the _ inverse problem _
, i.e. the des determining the l t model that reproduces given observational data , contain terms that would diverge unless a certain condition is met .
they call such loci _ critical points _ , and they correctly claim that such points are generic to these des .
this was actually pointed out by mustapha , hellaby and ellis @xcite , solved by lu and hellaby @xcite , discussed as an observational feature in @xcite , and used to advantage by mcclure and hellaby @xcite .
these papers develop a numerical procedure for extracting the metric of the cosmos from observational data . in the denominator , and on the observer s past null cone @xmath240 ,
this is @xmath241 here , so it is again zero at the apparent horizon . ] in their analysis of the inverse problem , vfw consider only flat l
t models , which are a subset of measure zero in the family of all possible l
they say that flrw models provide examples of _ transcritical _ solutions ( i.e. those that do not cause any divergence of the des ) , but they fail to find any other viable solution . and an arbitrarily chosen @xmath8 , specified via @xmath242 or @xmath243 , may well produce an l
t model with unrealistic features . ]
they then conclude that physically reasonable solutions _ must be very exceptional indeed _ , and although they have not tried the @xmath244 case , they opine that _ generic solutions with @xmath245 _ would have all the _ singularities _ and _ observational problems _ they encountered , and in this case too transcritical solutions do _ not appear to be likely . _ in assessing the likelihood of finding a ` transcritical solution ' it is important to understand the physics and geometry of the problem . for any _ given _ cosmological metric , the des that determine the path of the light ray and the variation of @xmath9 and @xmath246 along it ( for example ( [ 2.5 ] ) , ( [ 8.2 ] ) , ( [ 2.6 ] ) ) are free of critical points . however , the path followed by an incoming light ray depends on the geometry it passes through , and this affects the observations , @xmath9 , @xmath246 etc .
all known expanding cosmologies with a deceleration phase have a past apparent horizon ( ah@xmath247 ) ; relative to a given worldline ( observer ) , it is the locus where incoming light rays overcome the cosmic expansion and start to make progress inwards .
the maximum in the area distance @xmath248 occurs where an observer s pnc crosses her past apparent horizon .
the fact that @xmath249 at this maximum is the cause of the critical behaviour .
this behaviour is a generic feature of the l
t and friedmann models .
we express this in the following theorems .
we consider only models that : ( i ) have a big bang and after 14 billion years are still expanding , so any recollapse is to the future , and times well beyond the present are not considered ; ( ii ) have no shell crossings @xcite ; ( iii ) are large , e.g. if the curvature is positive , then any spatial maximum where @xmath250 is well beyond observational range , which means @xmath251 and @xmath252 ; ( iv ) if @xmath253 its effect is only discernable in fairly distant observations
. we will call such a model an rltc : a realistic l
t cosmology . in every rltc ,
every light ray arriving at the origin is a transcritical ray . * proof * the areal radius @xmath254 obeys @xmath255 where @xmath256 if @xmath77 is increasing ( expansion ) and @xmath257 if it is decreasing ( contraction ) .
worldlines with @xmath258 are ever - expanding and the rest are recollapsing .
those that recollapse reach their maximum @xmath77 where @xmath259 and this maximum is @xmath260 .
the incoming radial light rays satisfy ( [ 2.5 ] ) , so the variation of @xmath77 along a ray is @xmath261_\wedge = \left ( - \ell \frac{\sqrt{2 m / { \widehat{r}}+ 2 e + \lambda { \widehat{r}}^2 / 3}\ ; } { \sqrt{1 + 2 e}\ ; } + 1 \right ) \widehat{r_{,r } } = \frac{- \widehat{r_{,r}}}{\sqrt{1 + 2 e}\ ; } { \frac{{\rm d } { { \widehat{r}}}}{{\rm d } { t } } } \label{8.2 } \end{aligned}\ ] ] the past apparent horizon is the locus where @xmath262 for all rays , which implies @xmath263 and the past apparent horizon has @xmath256 . for a given @xmath264 , ( [ 8.3 ] )
has either two , one or zero solutions in the range @xmath265 .
the smaller @xmath77 solution corresponds to the regular ( @xmath7 ) ah , and the larger one corresponds to the de sitter horizon , but the two merge on the @xmath266 worldline . near the origin , @xmath267 , regularity @xcite requires @xmath268 and @xmath269 , so by ( [ 8.1 ] ) @xmath270 and by ( [ 8.2 ] ) @xmath271 there .
this means @xmath272 for incoming rays near the origin . following the rays outwards and back in time , ( [ 2.5 ] ) shows @xmath17 and therefore @xmath147 is increasing , and ( [ 8.2 ] ) shows @xmath273 is increasing near the origin .
provided @xmath274 stays positive as assumed , ( [ 2.5 ] ) may be integrated all the way to the big bang ; and here we have @xmath264 and @xmath275 , so @xmath276 .
therefore the rays have crossed ah@xmath247 .
this argument applies to every ray arriving at the origin before the big crunch , if there is a crunch . the des for the general l
t ` inverse problem ' are given in @xcite , and it is evident that the critical points are where @xmath277 . if a ray crosses the past ah , then by definition it ( a ) has @xmath262 there , and therefore ( b ) it is transcritical .
@xmath278 if @xmath253 then there could models , it is possible the maximum @xmath147 does not reach @xmath279 . ]
be worldlines with @xmath280 .
these worldlines never encounter a solution to ( [ 8.3 ] ) , so every incoming ray crossing them has @xmath281 and @xmath282 . in ever - expanding models , ah@xmath247 asymptotically
approaches @xmath283 ( and @xmath284 ) towards the future , and it separates incoming light rays that always have increasing @xmath77 from those that reach the origin
. therefore there are rays emerging from the bang that neither cross ah@xmath247 , nor reach the origin .
every incoming radial light ray in every @xmath7 l t model with a big bang is a transcritical ray .
* proof * when @xmath7 , ( [ 8.3 ] ) becomes the familiar @xmath285 and such apparent horizons have been well studied in @xcite . in this case , every worldline encounters ah@xmath247 , which is at @xmath265 except at an origin , and every light ray starts on the big bang .
therefore , of the rays emerging from the bang and arriving at the crunch or @xmath286 , only those starting at an origin may be said to have not crossed the past ah .
models usually have one origin , but closed models normally have two .
the rays emerging from the local origin are outgoing , not incoming , and the rays emerging from an antipodal origin may eventually be locally incoming , but in an rltc this would happen long after the present , if at all .
@xmath278 transcritical rays of inhomogeneous models that reproduce the observed @xmath246 do exist .
* proof * see e.g. @xcite .
the point here is that if a chosen model produces a @xmath246 and @xmath8 that fits the observations for a given pnc , out past a maximum in @xmath287 , then that pnc is necessarily transcritical .
@xmath278 any reasonable @xmath246 can be reproduced by a @xmath7 l t model .
* proof * this is a special case of the theorem in @xcite , which claims any pair of reasonable observational functions , @xmath246 or @xmath8 and @xmath288 can be reproduced by an l
t model . here
@xmath289 is the number density of sources ( e.g. galaxies ) in redshift space , and @xmath290 is the mean mass per source , which may also vary with @xmath9 . essentially the 2 arbitrary functions in an l
t model allow one to reproduce 2 observational relations .
in fact @xcite have demonstrated how to solve the ` inverse problem ' numerically for both observational functions , and shown that the conditions at the ah do impose a significant constraint that holds only at a single @xmath9 value .
but with only @xmath8 given , the 2 l
t functions allow the observations and the constraint to be fitted easily .
@xmath278 together these results show that there is a plentiful supply of transcritical solutions amongst the l
t models , and that there is a well defined procedure for extracting an l t model that fits the observed @xmath246 .
more important , however , is the fact that vfw do not try to find an l
t model that reproduces real observational data , but the one that reproduces a chosen idealised function for @xmath8 .
vfw write as if flrw models are known to solve the _
inverse problem_. to date the ` inverse problem ' has never even been attempted with real data .
what is usually done is to solve the fitting problem : choose your favorite type of cosmological metric and find the free functions and parameters that make it best fit the data .
even models that are arbitrarily close to reproducing the observations will fail this test .
for example , take the ` observational ' relations of a given flrw model , add small random or systematic errors , and use this ` data ' as input to the inverse problem .
the des of the inverse problem will return the given flrw model quite closely , but ( [ 8.3 ] ) will not be _ exactly _ satisfied where @xmath249 , so they will still diverge ( see @xcite ) .
the chances of an flrw light ray being exactly transcritical relative to real observational data are distinctly less than for an l t model .
the failure of a proposed solution to be transcritical does not indicate any physical problem with the underlying l t or flrw model ; at worst it indicates that a particular ray in a particular model does not ( exactly ) reproduce the given observational function(s ) . as discussed in @xcite , when attempting to reproduce _ two _ observational functions with an l
t model , the ah constraints _ do _ provide a significant challenge . in overcoming this challenge
it was shown @xcite the ah can give us useful observational information .
we have corrected a number of inaccurate statements appearing in the hs @xcite , flanagan @xcite and vfw @xcite papers .
at least two of these misunderstandings are widespread and therefore it is worth putting the record straight .
the first error - correction to be made is that a conical density profile at the origin is not a singularity in any accepted sense , and there are no physical problems associated with it . in some situations
a conical profile may be a good model of structures that are strongly centrally concentrated . secondly , critical points in the inverse problem are generic . when solving the ` inverse problem ' as defined in sec .
[ trcrgen ] for a realistic l t or flrw model , and finding that the differential equations diverge at some redshift , one must be aware that this is an inevitable consequence of the apparent horizon , where the diameter distance is maximum .
the chances that real observational data _ exactly _ satisfy the full transcriticality conditions for any given model , homogeneous or not , are essentially zero .
such divergent behaviour does not indicate any physical problem with the underlying l t or flrw model , and transcritical solutions in a variety of models that closely reproduce the observed @xmath8 do exist .
actually this point seems not to be well understood and other examples of this misunderstanding can be found in the literature ( see , e.g. @xcite ) .
our point three is that one should be careful when extending flrw definitions of some functions to other cosmological models . in most such cases ,
when inhomogeneities are present , the flrw parameters do not mean the same thing as in the homogeneous case .
it is very common to get acceleration and supernova dimming mixed up .
the interpretation of the supernova dimming as an accelerated expansion ( which is due to the use of flrw models ) has firmly taken root , which is why some have focussed on acceleration in the fluid expansion . in the l
t models , the ` acceleration ' defined from @xmath10 using ( [ 3.3 ] ) exhibits a behaviour different from the @xmath0 of the fluid flow defined by ( [ 2.1 ] ) , even though they coincide for flrw models .
we hope our paper will help establish the difference between ` accelerated expansion ' and ` supernova dimming ' in the observed magnitude
redshift relation .
we summarise and discuss different misunderstandings related to inhomogeneous cosmology in table [ tab1 ] .
ll misconceptions and misunderstandings & corrections and clarifications + + & + & + & + + & + & + & + & + & + & + & + + & + & + & + & + & + + & + & + & + & + & + * acknowledgements * mnc wants to thank the department of mathematics and applied mathematics of the university of cape town for hospitality while part of this work was carried out , and particularly charles hellaby and peter dunsby for their nice welcome .
the authors are grateful to kayll lake for his interest in their work and useful comments .
the formulae below are obtained using the coordinate @xmath17 introduced in ( [ 2.8 ] ) , the symbols introduced in ( [ 5.10 ] ) ( [ 5.11 ] ) , and eq .
( [ 2.2 ] ) .
we show the intermediate expressions in order to demonstrate that all terms that could cause divergencies cancel out before the limit is taken .
the first two equations follow trivially from ( [ 2.2 ] ) .
@xmath291 to evaluate the behaviour of @xmath274 and @xmath293 in the neighbourhood of @xmath294 we use the following equation @xcite : @xmath295 r_{,t}.\ ] ] the above applies also with @xmath296 if we neglect the @xmath297 terms ( as verified by direct calculation ) . from ( [ a.3 ] )
we obtain : @xmath298 the next equations below follow by consecutively differentiating ( [ a.1 ] ) ( [ a.3 ] ) by @xmath17 and using ( [ a.3 ] ) ( [ a.4 ] ) in the result : @xmath299 @xmath300 { \ { \underset { r \to 0 } { \longrightarrow}}\ } v_0 , \label{a.6}\ ] ] @xmath301\right\ } - \frac r { v^3}\ \left[f_{,r } - \frac { m_0 { \cal l } } { ( s + { \cal p})^2}\right]^2 \nonumber \\ & & { \ { \underset { r \to 0 } { \longrightarrow}}\ } \frac 1 { v_0 } \left[- \frac { m_0 } { s^2 } r_{,rr}(0 ) + 2 f_{,r}(0)\right],\label{a.7}\end{aligned}\ ] ] @xmath302 \left[2 f_{,r } + r \left(f_{,rr } - 2 \frac { { f_{,r}}^2 } { - k/2 + f}\right)\right ] \nonumber \\ & - & v \left[2 t_{b , r } + r \left(t_{b , rr } + \frac { f_{,r } \ , t_{b , r } } { 2 ( - k/2 + f)}\right)\right ] \nonumber \\ & + & \frac r v\ \left[\frac 3 2\ \frac { f_{,r } } { - k/2 + f } \left(t - t_b\right ) - t_{b , r}\right ] \left[f_{,r } - \frac { m_0 { \cal l } } { ( s + { \cal p})^2}\right ] \nonumber \\ { \ { \underset { r \to 0 } { \longrightarrow}}\ } & \ & - \frac { 2 f_{,r}(0 ) } k \left[- 2 s + 3 \left(t - t_b(0)\right ) v_0\right ] - 2 t_{b , r}(0 ) v_0 .
\nonumber \\\end{aligned}\ ] ] g. f. r. ellis , in _ proceedings of the international school of physics ` enrico fermi ' , course 47 : general relativity and cosmology _ , ed .
r. k. sachs . academic press , new york and london ( 1971 ) ,
pp . 104 182 ; reprinted , with historical comments , in _ gen .
_ * 41 * , 581 ( 2009 ) .
f. hoyle , cosmological tests of gravitational theories " in _ proc .
enrico fermi school of physics , course xx , varenna , evidence for gravitational theories " _
, ed . c. moller ( academic press , new york ) ( 1961 )
, p 141 - 174 .
a. g. riess , a. v. filippenko , p. challis , a. clocchiatti , a. diercks , p. m. garnavich , r. l. gilliland , c. j. hogan , s. jha , r. p. krishner , b. leibundgut , m. m. phillips , d. reiss , b. p. schmidt , r. a. schommer , r. c. smith , j. spyromilio , c. stubbs , n. b. suntzeff and j. tonry , _ astron .
j_. * 116 * , 1009 ( 1998 ) .
s. perlmutter , g. aldering , g. goldhaber , r. a. knop , p. nugent , p. g. castro , s. deustua , s. fabbro , a. goobar , d. e. groom , i. m. hook , a. g. kim , m. y. kim , l. c. lee , n. j. nunes , r. pain , c. r. pennypacker , r. quimby , c. lidman , r. s. ellis , m. irwin , r. g. mcmahon , p. ruiz - lapuente , n. walton , b. schaefer , b. j. boyle , a. v. filippenko , t. matheson , a. s. fruchter , n. panagia , h. j. m. newberg and w. j. couch , _ astrophys . j. _ * 517 * , 565 ( 1999 ) . | a number of misunderstandings about modeling the apparent accelerated expansion of the universe , and about the ` weak singularity ' are clarified : 1 . of the five definitions of the deceleration parameter given by hirata and seljak ( hs ) , only @xmath0 is a correct invariant measure of acceleration / deceleration of expansion .
the @xmath1 and @xmath2 are unrelated to acceleration in an inhomogeneous model .
2 . the averaging over directions involved in the definition of @xmath2 does not correspond to what is done in observational astronomy .
3 .
hs s equation ( 38 ) connecting @xmath2 to the flow invariants gives self - contradictory results when applied at the centre of symmetry of the lematre
tolman ( l t ) model .
the intermediate equation ( 31 ) that determines @xmath3 is correct , but approximate , so it can not be used for determining the sign of the deceleration parameter .
even so , at the centre of symmetry of the l
t model , it puts no limitation on the sign of @xmath4 .
4 .
the ` weak singularity ' of vanderveld _ et al .
_ is a conical profile of mass density at the centre a perfectly acceptable configuration .
5 . the so - called ` critical point ' in the equations of the ` inverse problem ' for a central observer in an l
t model is a manifestation of the apparent horizon a common property of the past light cones in zero - lambda l
t models , perfectly manageable if the equations are correctly integrated . |
blazars are by far the most common objects detected in the gamma - ray sky @xcite .
_ fermi _ has detected blazar variability as short as a few hours ( e.g. * ? ? ?
during these flares , the gev luminosity has been known to increase by a factor of up to several compared to its pre - flare luminosity . because blazars can not be resolved at these energies ( or at any other energy , with the possible exception of vlba observations )
, it is impossible to determine the location of these flares by direct detection . to address this issue
, we propose a diagnostic test that utilizes _ fermi _ variability data of short flares to determine the location of the gev emission in high - power blazars , namely flat spectrum radio quasars ( fsrq ) .
relativistic effects determine which photon field is dominant at varying distances from the central black hole , and as a result the location of the gev flaring site determines the dominant source of seed photons .
the co - moving ( jet frame ) energy density of a radiation field @xmath0 scales as differing factors of @xmath1 depending on the direction from which the photons enter the emitting region @xcite . if the photon field is isotropic @xmath2 for photons entering the emitting region from behind the relativistically moving blob @xmath3 in this work , all primed quantities refer to the co - moving frame of the blob and unprimed quantities refer to the galaxy frame .
if we assume a nominal fsrq accretion disk luminosity of @xmath4 erg s@xmath5 and that a fraction @xmath6 of this radiation is reprocessed by both the blr and the mt , a typical luminosity of the external radiation field is @xmath7 erg s@xmath5 @xcite .
if the emission site is located within the blr ( at @xmath8),the photon field can be considered isotropic in the galaxy frame and using eq .
[ eq : uiso ] its co - moving energy density is @xmath9 where @xmath10 @xmath11 , and @xmath12 similarly , the mt photon field is isotropic inside the blr and its co - moving seed photon energy density is @xmath13 where @xmath14 and @xmath15 .
clearly , inside the @xmath16 the co - moving blr photon field energy density dominates over that of the mt by a factor of @xmath17 .
if the emission site is located at @xmath18 ( within the mt ) then the blr uv photons enter the emitting region practically from behind , and as per eq .
[ eq : ubhd ] @xmath19 the ir photons from the mt retain the same co - moving energy density previously given by eq .
[ eq : densemt ] . in this case , therefore , it is the mt that dominates the co - moving photon energy density . these external photon field
co - moving energy densities need to be compared to the synchrotron photon field energy density .
if @xmath20 is the size of the emitting blob , the co - moving synchrotron photon energy density is @xmath21 .
the most plausible assumption for the size of the emitting region , however , is to set an upper limit to it by its variability timescale : @xmath22 .
we then obtain a lower limit for the synchrotron energy density @xmath23 where @xmath24 , @xmath25 is the observed variation time in hours , and @xmath26 is the same as previously defined .
note that a 6 hour variability scale has been observed by fermi @xcite .
comparison of the external energy density @xmath27 ( using eqs .
[ eq : uiso ] or [ eq : ubhd ] depending on if the emission is isotropic ) and synchrotron seed photon energy density @xmath28 ( eq . [ eq : densesynch ] ) shows that an isotropic external photon field dominates , i.e @xmath29 , if @xmath30 using the blr and mt descriptions adopted above and @xmath31 hours , results in a @xmath32 inside the blr and @xmath33 inside the mt , but outside the blr .
vlbi studies of superluminal speeds in powerful fsrqs ( e.g. see figure 24 of * ? ? ?
* ) show that for most fsrqs @xmath34 .
therefore , this condition is clearly satisfied inside the blr , and it is expected to be satisfied in the majority of cases in the mt .
_ the critical difference between the blr and the mt is the energy of the seed photons : photons originating from the blr are uv photons ( @xmath35 ) while photons originating from the mt are ir photons ( @xmath36 ) , where @xmath37 .
this difference by a factor of @xmath38 in typical photon energy is critical in that it affects the energy regime in which electron cooling takes place , and thus the energy dependence of the electron cooling time . _ in powerful fsrqs the ic emission in high states can dominate over the synchrotron emission by a factor of up to @xmath39 @xcite . in cases of high compton dominance , the primary electron cooling mechanism is ic scattering . for electrons cooling in the thomson regime ( @xmath40 , where both the electron lorentz factor @xmath41 and the seed photon energy @xmath42 are measured in the same frame ) the cooling rate @xmath43 . for electrons with @xmath44
cooling takes place in the klein - nishina ( kn ) regime with @xmath45 @xcite .
the electron cooling time given by @xmath46 is then @xmath47 in the thomson regime and @xmath48 in the deep kn regime . in the transition from thomson to kn regimes (
around @xmath49 ) the cooling time flattens and becomes essentially energy independent ( fig [ fig : tblrmt ] ) . .
top panel : blazar emission site located in the blr .
bottom panel : blazar emission site located in the mt .
the dotted lines represent the various cooling mechanisms ( blue = thomson cooling , red = kn cooling , green = synchrotron cooling ) , the solid black line is the total cooling time .
the dotted black line in the top panel indicates where @xmath50 plots were calculated for the following values : seed photon energies : @xmath51 and @xmath52 , energy densities @xmath53 erg @xmath54 , @xmath55 erg @xmath54 , a @xmath56 that corresponded to a compton dominance of @xmath17 , and @xmath57 .
because blr photons have energies higher than mt photons by a factor of @xmath17 , transition from the thomson to the kn regime occurs at values of @xmath41 @xmath17 times lower in the blr ( top panel ) than in the mt ( bottom panel).,width=332 ] the effects of the transition between thomson and kn regimes on the electron energy distribution ( eed ) and the resultant spectrum of the synchrotron and ic emission have been studied before ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in short , because @xmath43 in the thomson regime and @xmath58 in the kn regime , the cooling time @xmath59 scales as @xmath60 in the thomson regime and as @xmath61 in the kn regime . because the cooling time is approximately energy independent around @xmath49 , this energy - independent cooling time will be manifested at energies lower by a factor of @xmath17 for cooling taking place in the blr compared to cooling taking place in the mt , since the blr seed photons have an energy higher than that of the mt by a factor of @xmath17 .
this can be seen in fig .
[ fig : tblrmt ] where we plot the electron cooling time for a source with a ratio of external photon field energy density @xmath62 in the co - moving frame to co - moving magnetic field energy density @xmath56 , @xmath63 .
this corresponds to a factor of @xmath17 compton dominance ( ratio of inverse compton to synchrotron luminosity ) , similar to what is observed in the most compton dominated sources .
the transition from thomson cooling to kn cooling , and from kn cooling to synchrotron cooling can also be seen in fig .
[ fig : tblrmt ] .
if the electrons are cooling on photons from the blr , cooling takes place in the klein - nishina regime and the cooling time scale is approximately energy independent around @xmath49 ( fig .
[ fig : tblrmt ] , top panel ) .
if the electron population cools on photons from the molecular torus , cooling takes place in the thomson regime ( fig .
[ fig : tblrmt ] , bottom panel ) .
the cooling time is heavily energy dependent , and any variations should consequently exhibit heavy energy - dependence .
the energy dependence of the cooling time results in an energy dependence ( or lack thereof ) of variations : if the blazar emission site is located within the blr , variations should be achromatic .
the energy dependence of the variations can be used as a diagnostic test to determine if the gev emission site is located within the blr . by comparing _
fermi _ light curves of flares at different energies , we propose that the energy dependence of the light curve can be used as a diagnostic test to rule out whether the gev flare originates in the blr or mt . to demonstrate the effect of the energy independence or dependence of the electron cooling time on the variability of a flare , we utilized a one - zone numerical model to simulate a flare .
we initialized the code with values appropriate for a high power blazar with a compton dominance of @xmath17 .
for this particular simulation we assumed a source size @xmath64 , bulk lorentz factor @xmath65 , co - moving injected electron luminosity @xmath66 erg s @xmath5 , maximum electron lorentz factor @xmath67 , and electron index @xmath68 .
for the case of a flaring region located within the blr we assumed an initial photon energy @xmath69 and an energy density ( in the galaxy frame ) @xmath70 erg @xmath54 .
for the case of a flaring region located outside the blr we assumed an initial photon energy @xmath71 and an energy density @xmath72 erg @xmath54 . for each case the magnetic field * b * was fixed to assume a compton dominance @xmath73 .
the system behaves as expected , showing a noticeable difference in the decay rate as well as the amplitude of the flare depending on if the seed photons originated from within or from outside the blr ( see figs .
[ fig : blrflare ] and [ fig : mflare ] ) .
this difference in decay rate and amplitude can be used as a diagnostic test to differentiate between flares that take place inside or outside the blr .
mev , @xmath74 gev , @xmath75 gev . ,
width=328 ] mev , @xmath74 gev , @xmath75 gev .
, width=328 ] because the gev emitting region is not a point source , any change in the luminosity of the blob will not be seen instantaneously . instead , the observed decay time of the light curve is the result of the actual decay time and the light - crossing time inherent in the blob . to test whether the light - crossing time would erase any difference in the light curves ( as predicted by our diagnostic ) for the case of the emitting region
being located outside the blr , we modeled the flux of a source with light - crossing time @xmath76 that is decreasing in flux .
we assumed the flux of the source is decaying exponentially , @xmath77 , where @xmath78 is the cooling time at a specific energy .
our diagnostic predicts differences in the decay time of the light curves for the case where the emitting region is located within the mt ; for the purposes of this demonstration we assume cooling occurs in the thomson regime and as a result @xmath79 .
we plot the resultant light curves for cooling times differing by a factor of @xmath80 ( i.e. energies differing by a factor of 10 ) . as evident in fig .
[ fig : lctime ] , even with light - travel time effects convolved with exponential decay times , the differences in the decay times are still manifested .
a flare that occurs in the mt should exhibit energy dependent cooling times because cooling occurs in the thomson regime .
practical application of the diagnostic hinges on the requirement that the decay times in different energy bands should have detectable differences . in the thomon regime ,
the cooling time is given by @xmath82 , where @xmath42 is the seed photon energy and @xmath83 is the photon energy after ic scattering .
the expected difference in cooling time at two different energies is then @xmath84 where @xmath85 and @xmath86 are the photon energies at which we calculate the cooling time , and @xmath87 is the mt energy density . because of _ fermi s _ three hour time resolution , @xmath88 h and this sets an upper limit to the product @xmath89 .
the @xmath1-@xmath87 parameter space for which the difference in decay times are distinguishable is plotted in fig .
[ fig : ugam ] for three plausible values of @xmath90 , with the area below each line corresponding to the acceptable part of the parameter space .
the permitted part of this diagram comfortably accommodates bulk lorentz factors between 10 and 20 and mt photon densities @xmath91 erg @xmath54 . ) as a function of bulk lorentz factor @xmath1 plotted for @xmath90 corresponding to 5 , 10 , and 15 h. accepted values of of @xmath87 and @xmath1 values fall below the curves.,width=340 ] our discussion following equation ( [ eq : gammalimit ] ) suggests that it is possible that in the mt , for relatively low values of @xmath1 , @xmath92 and the dominant photon field is from jet synchrotron photons .
even if this is the case , cooling still occurs in the thomson regime and the energy dependent cooling time is still manifested . for fsrqs ,
the peak of the synchrotron spectral energy distribution is at @xmath93 hz @xcite , which in the comoving frame corresponds to @xmath94 .
these seed photons , therefore , have a lower energy than the energy of the mt photons in the comoving frame , and will be deeper in the thomson regime compared to the mt photons . because cooling still occurs in the thomson regime , the energy dependent cooling time , therefore , is expected to be present , even in the case that the dominant photon density is that of the synchrotron photons .
we have presented a diagnostic test that utilizes blazar variability to determine the location of the gev emitting site in blazars .
the energy difference in seed photons originating from the blr versus seed photons originating from the mt causes electrons within the emitting site to cool in different energy regimes . for the case
where the gev emitting site is located within the blr , cooling takes place at the onset of the kn regime , and the resultant electron cooling time is energy - independent .
we have demonstrated that the associated light curves exhibits decay times that are approximately energy independent .
conversely , for the case where the gev emitting site is located outside the blr , cooling takes place in the thomson regime and the electron cooling times are heavily energy dependent . in this case , the associated light curves exhibit energy dependence of their decay times .
the energy dependence of the decay time of the light curves is visible within the _ fermi _ energies ; these differences can be used as a diagnostic test to determine whether the gev emitting region is located inside or outside the blr .
these effects are observable within the time resolution of @xmath95 few hours that _ fermi _ has achieved for bright flares and are not erased due to light - travel time effects
. light curves from a sufficiently bright and rapid flare ( such as that in 3c 454.3 ; * ? ? ?
* ) should be compared at different energies .
if the gev emitting site is located within the blr , the decay times will exhibit no energy dependence , whereas if the emitting site is located within the mt , the decay times will exhibit energy dependence . | an issue currently under debate in the literature is how far from the black hole is the _
fermi_-observed gev emission of powerful blazars emitted . here
we present a diagnostic tool for testing whether the gev emission site is located within the sub - pc broad emission line ( blr ) region or further out in the pc scale molecular torus ( mt ) environment . within the blr the scattering takes place at the onset of the klein - nishina regime , causing the electron cooling time to become almost energy independent and as a result
, the variation of high - energy emission is expected to be achromatic .
contrarily , if the emission site is located outside the blr , the expected gev variability is energy - dependent and with amplitude increasing with energy .
we demonstrate this using time - dependent numerical simulations of blazar variability and discuss the applicability of our method . |
the holstein model which describes electrons coupled to some local molecular deformations is a simple model to study polaron properties@xcite .
these recent years , accurate numerical investigations of this model on finite size systems have given new insights@xcite closely linked to some current problems in condensed matter physics@xcite .
a polaron corresponds to an electron renormalized by the interactions with the lattice deformations or phonons@xcite . from a historical point of view@xcite the approach to the polaron problem
consists in computing the spectrum of the whole system containing a single electron coupled to some optical phonons .
writing @xmath0 the momentum of the electron and @xmath1 the momentum of the phonons , the problem is to compute the lowest energy of the system for a total momentum @xmath2 , since the total momentum operator commute with the hamiltonian .
this mimimum eigenenergy of the hamiltonian for a given total momentum @xmath3 gives the polaron relation dispersion , as far as we consider only the low lying excitations of the system that is as far as no phonon excitations are involved . in this paper
, we consider a different approach to the problem and try to decompose the excitations of the system in term of polaron excitations and renormalized phonons excitations . due to the difficulty of the problem ,
we address only the weak coupling limit of the problem and we use mainly second order perturbation theory in the electron phonon coupling . within this framework
, one can block diagonalize the hamiltonian .
then it is possible to label the eigenstates of the hamiltonian in term of renormalized electron and phonons . in other words ,
an eigenstate of the hamiltonian contains one renormalized electron ( a polaron ) of momentum @xmath0 and some renormalized phonons of momentum @xmath1 ; the total momentum of the state being @xmath2 .
the hamiltonian for the holstein model , in wannier representation , is given by @xmath4 where @xmath5 and @xmath6 are electron creation and annihilation operators , @xmath7 and @xmath8 are phonon creation and annihilation operators .
@xmath9 is the hopping integral , @xmath10 is the optical phonon frequency and @xmath11 is the dimensionless coupling constant . in the strong coupling limit , @xmath12 , the hamiltonian is approximatively diagonalized by the lang - firsov transformation@xcite .
the transformed hamiltonian is given by @xmath13 , with @xmath14 .
it is only when the hopping term is zero that the transformed hamiltonian is diagonal . in this case one
obtains some new fermionic operators , which describe small polarons , or localised polarons .
the small polaron creation operator is given by @xmath15 , or @xmath16 , with @xmath17 while the renormalized phonons operators are given by @xmath18 .
the action of this small polaron creation operator on the vacuum is thus to create an electron and a coherent state for the phonons where the annihilation operator takes the expectation value @xmath11 . for non zero values of the hopping integral ,
these small polarons become some quasiparticles more or less well defined depending on the value of the coupling and on the phonon frequency@xcite .
we notice that for @xmath19 , the small polaron correspond to the bare electron , since there is no coupling and thus no renormalization .
we shall use a similar unitary transformation formalism but in we weak coupling limit , @xmath20 , and work in momentum space where the hamiltonian reads @xmath21 where @xmath22 is the number of sites of the lattice , and indices of various operators correspond to bloch momenta . in the next section we consider our approach for a two site system ,
then we will generalize to a general system .
we start our perturbation theory for the two site system . in this case
, the momentum takes two values @xmath23 and the bare electron relation dispersion is given by @xmath24 and @xmath25 .
the @xmath26 mode of the phonons is coupled to the total number of electron which is one in this problem .
we drop out this mode after performing the shift corresponding to the displaced oscillator@xcite .
the reduced hamiltonian is then given by @xmath27 we have dropped the index @xmath28 for the boson operators .
the basis states are given by @xmath29 and @xmath30 is the vacuum . for @xmath19
the corresponding eigenenergies are given by @xmath31 .
using second order perturbation theory , we obtain the eigenstates and eigenenergies , @xmath32\ ] ] @xmath33 c_{\pi}^{\dagger } c_{0 } \ ; - \ ; \frac{g\omega_{0}}{\sqrt{2 } } \left [ \frac{b}{2 t + \omega_{0 } } + \frac{b^{\dagger}}{2 t - \omega_{0 } } \right ] c_{0}^{\dagger } c_{\pi } \right\ } \ ; | k ; n \rangle\ ] ] we show now that the eigenstate @xmath34 describes a state with one polaron of momentum @xmath0 and @xmath35 renormalized phonons .
the states @xmath36 are eigenstates of the hamiltonian @xmath37 , thus using @xmath38 , we obtain @xmath39 .
up to @xmath40 , the @xmath41-matrix of the unitary transformation is given by @xmath42 c_{\pi}^{\dagger } c_{0 } \ ; - \ ; \frac{g\omega_{0}}{\sqrt{2 } } \left [ \frac{b}{2 t + \omega_{0 } } + \frac{b^{\dagger}}{2 t - \omega_{0 } } \right ] c_{0}^{\dagger } c_{\pi}\ ] ] we note that if we set @xmath43 in this expression , we obtain the right transformation which diagonalize the reduced hamiltonian @xmath37 .
we can now construct the corresponding polaron operator @xmath44 through the transformation @xmath45 .
we obtain explicitly @xmath46 c _
{ \pi}\ ] ] @xmath47 c_{0}\ ] ] these operators are fermionic ( they satisfy the anticommutation relations ) , with the property @xmath48 .
this means that the eigenstate @xmath34 contains one polaron of momentum @xmath0 . as concern
the phonons , we can proceed the same way , introducing a new operator @xmath49 which describes a renormalized boson or phonon . explicitly , one obtains @xmath50 c_{0}^{\dagger } c_{\pi } \ ; + \ ; \frac{g \omega_{0}}{\sqrt{2 } } \left [ \frac{b}{2t-\omega_{0 } } - \frac{b^{\dagger}}{2t+\omega_{0 } } \right ] c_{\pi}^{\dagger } c_{0}\ ] ] again , this operator is diagonal with @xmath51 . we therefore conclude that the state @xmath34 contains one polaron of momentum @xmath0 and @xmath35 renormalized phonons .
the total momentum of the state is @xmath52 if @xmath35 is even and @xmath53 if @xmath35 is odd .
then we write the eigenenergies as @xmath54 .
one obtains @xmath55 @xmath56 for @xmath35 phonons or renormalized phonons in the system @xmath57 corresponds to the fermionic ( polaronic ) excitation energy . in the polaron problem ,
we restrict ourselves to the case @xmath58 , that is we create a polaron from the vacuum , @xmath59 .
we are left with @xmath60 which is the polaron binding energy and @xmath61 which represents the hopping integral of the polaron .
if we take @xmath62 and @xmath63 we obtains @xmath64 and @xmath65 while exact diagonalization@xcite gives @xmath66 and @xmath67 .
we notice the enhancement of the bandwidth for this two site system . in term of the renormalized operators , the hamiltonian can now be written in a block diagonal form , @xmath68 { \cal b}^{\dagger } { \cal b } p_{k}^{\dagger } p_{k}\ ] ] @xmath69 { \cal b}^{\dagger } { \cal b}^{\dagger } { \cal b } { \cal b } p_{k}^{\dagger } p_{k } \ ; + \ ; \ldots\ ] ] the original electron phonon interaction which contained off diagonal transitions between subspaces with the same number of particle ( electron or phonon ) has been eliminated : there is no more off diagonal interactions between the polaron and the renormalized phonons .
next , we consider finite size systems for which numerical results have been obtained@xcite
. let @xmath70 be the state with the configuration of phonons @xmath71 such that @xmath72 .
the basis vector are then @xmath73 .
the eigenstates are given by @xmath74 .
now we use the fact that the state @xmath70 contains no electron , so that @xmath75 . using the definition of the polaron operator @xmath45 , we obtain the desired relation , @xmath76 . in other words ,
the operator @xmath77 acting on the state @xmath70 with the configuration @xmath71 of bare phonons , gives the eigenstate @xmath78 which contains one polaron of momentum @xmath0 and the configuration @xmath71 of renormalized phonons . defining @xmath79
, we therefore obtain @xmath80 .
this is the straightforward generalization of the results obtained for the two site system . using second order perturbation theory
the polaron dispersion is given by @xmath81 this is a well known result obtained by many approaches@xcite .
this dispersion relation corresponds to a state with one polaron of momentum @xmath0 and zero renormalized phonons .
the next excited states correspond , for example , to states with one polaron of momentum @xmath0 and one renormalized phonon of momentum @xmath82 .
the total momentum of these states is therefore @xmath83 .
the energy of such a state is given by @xmath84\ ] ] as concern the corresponding unitary transformation , it has been discussed by frhlich@xcite and is given by @xmath85 we now discuss the excitation spectrum of the system in the weak coupling limit and consider a one dimensional system for simplicity with the bare electron dispersion @xmath86 .
our basic asumptions are a ) that the dispersion of the polaron @xmath87 is slightly renormalized from the bare electron dispersion relation @xmath88 and b ) the system is large enough so that the phonon frequencies are slightly renormalized , so that @xmath89 .
the ground state corresponds to the state with a single polaron of momentum @xmath90 , with energy @xmath91 .
the state of momentum @xmath3 with the lowest energy can be either the state with one polaron of momentum @xmath92 and no renormalized phonons or either the state with one polaron of momentum @xmath90 and one renormalized phonon of momentum @xmath93 such that @xmath83 .
the corresponding energies are respectively @xmath94 and @xmath95 . within our asumptions on @xmath87
, there exists a threshold @xmath96 such that @xmath97 . for @xmath98
the lowest energy is @xmath94 and the state contains only one polaron of momentum @xmath92 . for @xmath99
the lowest energy is @xmath95 and the state contains one polaron of zero momentum and one renormalized phonon of momentum @xmath93 .
as far as the others excitations are concerned , their construction is straightforward . if we compare this spectrum with the results of ref .
@xcite , where this spectrum was computed , we find a nice agreement in the case of a weak coupling .
the flat part in the spectrum is just the energy @xmath95 .
we further notice that for @xmath98 , these numerical results show that the polaron dispersion @xmath87 is slightly renormalized from the bare electron dispersion @xmath88 .
in conclusion , we have shown that for any coupling , one can define a renormalized electron of momentum @xmath0 and some renormalized bosons of momentum @xmath82 , such that the spectrum is made up of these renormalized excitations .
the renormalized operators can be obtained via an unitary transformation .
this was worked out explicitly in the weak coupling limit using perturbation theory .
recent finite size studies support our approach . in the crossover regime , where level crossings arise in the spectrum of the hamiltonian ,
it is not possible to define the unitary transformation pertubatively and it is not clear how to make some quantitative predictions .
t. holstein , ann .
( ny ) * 8 * , 325 ( 1959 ) .
a. s. alexandrov and n. f. mott , _ polarons and bipolarons _ ,
world scientific , singapor ( 1995 ) .
j. ranninger and u. thibblin , phys . rev . *
b45 * , 7730 ( 1992 ) .
w. stephan , phys . rev . * b54 * , 8981 ( 1996 ) .
g. wellein and h. fehske , phys
* b56 * , 4513 ( 1997 ) .
p. e. kornilovitch and e. r. pike , phys
. rev . * b55 * , 8634 ( 1997 ) .
e. jeckelmann and s. r. white , phys
. rev . * b57 * , 6376 ( 1998 ) . j. t. devreese . in : _
encylopedia of applied physics _ , vol .
* 14 * , ed .
g. l. trigg ( vch , 1996 ) , pp .
383 - 413 .
g. d. mahan , _ many - particle physics _ , plenum press , ( new york , london 1990 ) .
lee , f. low and d. pines , phys
. rev . * 90 * , 297 ( 1953 ) .
r. p. feynman , _ statistical mechanics _
, benjamin reading ( 1972 ) ; phys . rev . *
97 * , 660 ( 1955 ) .
i. j. lang and yu .
a. firsov , zh , eksp .
. fiz . * 43 * , 1843 ( 1962 ) [ soviet physics jetp * 16 * , 1301 ( 1962 ) ] . j. m. robin , phys . rev . *
b56 * , 13634 ( 1997 ) .
h. frhlich , proc .
a215 * , 291 ( 1952 ) . | in this paper the polaron problem for the holstein model is studied in the weak coupling limit .
we use second order perturbation theory to construct renormalized electron and phonons .
eigenstates of the hamiltonian are labelled and the excitation spectrum is constructed . |
disk accretion of gaseous material onto a massive black hole ( mbh ) is believed to be the central engine of qso and active galactic nucleus ( agn ; e.g. , * ? ? ?
. in the thin disk accretion scenario , the emergent spectrum of a qso is mainly determined by several parameters , including the mbh mass ( @xmath11 ) and spin ( @xmath12 ) , the accretion rate ( @xmath13 ) , and the inclination angle ( @xmath14 ) of the disk to the line of sight ( e.g. , * ? ? ?
* see also @xcite ) .
measuring the mbh mass ( @xmath15 ) and spin ( @xmath12 ) for each individual qsos is of fundamental importance and has been one of the major goals for the qso and mbh studies . as illustrated by a number of studies ( e.g. , * ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , obtaining the statistical distributions of @xmath15 and @xmath12 among qsos is probably crucial for our understanding of the growth history of mbhs and the cosmic evolution of qsos . in principle ,
@xmath11 , @xmath13 and @xmath12 can be estimated through the fitting of the multi - wavelength spectrum of each qso by dedicated disk model(s ) ( e.g. , * ? ? ?
* ) . however , it is practically not easy to estimate these parameters simultaneously with considerable accuracy , partly because of the complications in disk accretion model(s ) and partly because of lack of measurements at the euv to soft x - ray bands , where the disk emission is expected to be most prominent , for most qsos .
alternative methods to estimate @xmath11 and @xmath12 are necessary for understanding the evolution and assembly history of mbhs . the mass of a mbh in a qso can be estimated through the reverberation mapping ( rm ) technique @xcite , which is time consuming and currently is not possible for all qsos .
fortunately , recent developments in measuring the masses of mbhs through the rm technique have resulted in some empirical scaling relations , i.e. , the virial mass estimators , which offer simple ways to estimate the mbh mass of a qso ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . adopting these virial mass estimators ,
@xcite have obtained the mbh masses for most sdss qsos by using the qso optical luminosities and the width of some broad emission lines ( such as h@xmath16 , civ , and mg ii ) in the qso spectra .
if the mbh mass is known for a qso , the absolute accretion rate @xmath13 can be estimated based on the optical band luminosity by using the thin disk accretion model . the radiative efficiency is then obtained by @xmath17 , where @xmath1 is the bolometric luminosity .
the mbh spin @xmath12 can be inferred from @xmath0 as @xmath0 is simply determined by @xmath12 in the thin disk accretion model . with this method , it may be possible to indirectly infer the mbh spin and its distribution for a large sample of qsos .
( note here that direct measuring the spins of mbhs through features like the broad fe k@xmath18 line is currently possible only for a limited number of cases ( e.g. , mcg-6 - 30 - 15 and a dozen of other agns ; see @xcite ; and reference therein ) .
@xcite estimated @xmath0 for @xmath6 pg qsos ( hereafter dl sample or dl qsos ) by adopting the above method and found that @xmath0 is strongly correlated with @xmath11 , i.e. , @xmath19 , and they argued that such a correlation is unlikely induced by selection effects ( * ? ? ?
* see also @xcite ) .
if the above correlation is intrinsic , it might suggest that the spin of a big mbh is relatively larger than that of a small mbh and the growth histories of big mbhs are systematically different from that of small mbhs .
however , @xcite argued that such a correlation is most likely an artifact of the small parameter space in both luminosity and redshift covered by the dl sample , and additional factors , such as the host galaxy and dust contamination , uncertainties in the bolometric luminosity estimations , may also bias the estimations of @xmath0 .
@xcite concluded that the radiative efficiency itself can not be accurately measured due to the large errors in the relevant observed parameters .
therefore , it is still not clear whether @xmath0 really intrinsically correlates with @xmath11 or not . in this study , we extend the work of @xcite to the type 1 sdss qsos , in order to further study the radiative efficiency of individual qsos and its relation , if any , with mbh mass .
such a study should be helpful to identify various biases involved in the estimations of @xmath0 and address the question of whether @xmath0 correlates with @xmath11 or not , as the sdss sample has many more qsos and covers larger luminosity and redshift ranges compared to the dl sample .
this paper is organized as follows . in section [ sec : sdssqso ] , we introduce the sdss qso sample adopted in this study . in section [ sec : estepsilon ] , we first introduce the method to estimate the radiative efficiency and then obtain the radiative efficiency for individual type 1 sdss qsos .
similar to @xcite , we also find that the estimated radiative efficiency appears to be strongly correlated with the mbh virial mass at redshift @xmath20 .
such a strong correlation appears to exist for the sdss qsos at @xmath21 but not at @xmath22 . in section [ sec : bias ] , we investigate various biases that could be involved in the @xmath0 estimations and the selection effects of the qso sample , and adopt monte carlo simulations to generate mock qso samples to simulate the effects of these biases .
we find that the current data is consistent with no mass and redshift dependence of @xmath0 .
conclusions are given in section [ sec : conclusions ] .
a large number of qsos have been detected by the sdss . in the catalog of sdss qsos ( dr7 ) , @xmath23 qsos have @xmath14-band magnitude @xmath24 , at least one broad emission line broader than @xmath25 , and the estimations of the central mbh masses
. about half of these qsos ( totally @xmath26 qso at @xmath27 ) were selected out to form a homogeneous , statistical sample , which is primarily a flux limited sample with @xmath14-band magnitude @xmath28 at @xmath29 , @xmath30 at @xmath31 and an additional bright limit of @xmath32 @xcite .
estimations of the mbh masses of the qsos in this sample have been obtained by using the virial mass estimators based on scaling relationships calibrated for four different emission lines , i.e. , h@xmath18 , h@xmath16 , mg ii , and civ .
these estimations , together with the optical band luminosities , for this qso sample are given in @xcite .
we select the qsos with @xmath33 in this sample to form our qso sample ( totally @xmath34 qsos ) to be studied in this work , and these qsos are denoted as the sdss qsos hereafter if not otherwise stated . with this homogeneous , statistical qso sample , it is possible to study and model the selection effects and various biases involved in the estimations of the radiative efficiency of individual qsos as the true mbh mass function and eddington ratio distribution have also been estimated by ( * ? ? ?
* see details in section [ sec : bias ] ) .
the monochromatic luminosities are provided only at one or two wavelengths for most of the qsos in the sample of @xcite .
for example , the monochromatic luminosities at @xmath35 , @xmath36 , and @xmath37 are provided for qsos with redshift @xmath38 , @xmath39 , and @xmath40 , respectively . for those qsos with @xmath41 , there are two monochromatic luminosity measurements available at @xmath35 and @xmath42 , respectively ; and for those qsos with @xmath43 , there are also two monochromatic luminosity measurements available at @xmath36 and @xmath37 , respectively . for other qsos ,
only one monochromatic luminosity measurement is available , either at @xmath35 for those with @xmath44 or at @xmath37 for those with @xmath45 .
the radiative efficiency ( @xmath0 ) of the accretion process in a qso may be estimated by using its bolometric luminosity ( @xmath46 ) and the accretion rate ( @xmath13 ) , i.e. , @xmath47 , if these two quantities can be ( independently ) accurately determined .
given the mass of the central mbh of a qso and the qso multi - wavelength spectrum , in principle , the accretion rate can be estimated by adopting dedicated disk accretion model(s ) ; and the bolometric luminosity can also be estimated by integrating the qso multi - wavelength spectrum . in this section , we will first estimate the mass accretion rate and the bolometric luminosity for individual sdss qsos , and then estimate the radiative efficiency @xmath0 . in the standard disk accretion scenario , the multi - wavelength spectrum of a qso
is mainly determined by the mass and spin of the central mbh and the accretion rate ( e.g. , * ? ? ?
* see also @xcite ) .
the masses of the central mbhs in most sdss qsos have been estimated through the virial estimators ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , while the mbh spins are still difficult to measure .
the optical band luminosity of a qso is dominated by the emission from the outer disk , and it is mainly determined by the accretion rate of the disk and less affected by the mbh spin ( e.g. , * ? ? ?
* ; * ? ? ?
therefore , the accretion rate @xmath13 of a sdss qso , with known mbh mass , can be estimated with reasonable accuracy by only matching the observed optical band luminosity(/-ies ) @xmath48 with that predicted by an accretion disk model .
similar to @xcite , we adopt two different accretion disk models , i.e. , the standard thin disk model and the tlusty model to estimate the mass accretion rate of individual sdss qsos as follows .
* the standard thin accretion disk model ( hereafter the bb model ) .
here we adopt the relativistic standard thin disk model presented in ( * ? ? ?
* see also @xcite , @xcite ) . in this model , the inner edge of the accretion disk
is assumed to be the innermost stable circular orbit ( isco ) , which is solely determined by the mbh spin , and no emission comes from the region within isco .
the emergent spectrum is directly determined by the multi - temperature black body radiation integrated over the disk surface . * the tlusty model @xcite . in this model ,
not only the relativistic effects are included , but also the disk vertical structure and the radiative transfer in the disk are simultaneously considered ( for details , see * ? ? ?
* ) . in the tlusty model , the emergent spectrum from each annulus
is first calculated and the full spectrum is obtained by integrating over the disk radius . occasionally , the tlusty model could not converge for some specific parameter space , e.g. , at large radius or extremely large / small accretion rate , possibly because the disk material in this situation is already quite cool and/or convection becomes important . in these cases , we simply replace the radiation of these annuli by black body radiation ( see also discussions in * ? ? ?
these models have four parameters , i.e. , the mbh mass @xmath49 and spin @xmath12 , the accretion rate @xmath13 , and the inclination angle @xmath14 of the disk to the line of sight .
the model spectrum estimated from the tlusty model may also depend on the choice of the @xmath50-viscosity @xcite , which is simply set to be @xmath51 for all models as this parameter have little effect on the @xmath48 ( e.g. , see * ? ? ?
it is necessary to have some information about the other model parameters , i.e. , @xmath49 , @xmath12 and @xmath14 , in order to estimate @xmath13 with reasonable accuracy for a qso by only using its monochromatic luminosities in a limited range of wavelengths .
the mbh mass , @xmath49 , has been estimated for the majority of the sdss qsos by using the virial mass estimators but with uncertainties of @xmath52 dex ( e.g. , * ? ? ? * ; * ? ? ?
the inner radius of an accretion disk is defined by the mbh spin @xmath12 .
the model spectra in the optical bands are not significantly affected by the setting of the disk inner edge although the disk radiation at higher frequency is indeed affected by it .
if not otherwise stated , we simply adopt @xmath53 in our following calculations , which corresponds to @xmath54 in the standard disk model .
( or @xmath55 ) , i.e. , @xmath56 ( or @xmath57 ) , does not significantly affect the final results . ]
we adopt this spin value as reference since the global constraints on the mean radiative efficiency of qsos suggest @xmath58 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for the inclination angle @xmath14 , all qsos in the sample of @xcite are type 1 qsos , which presumably have small inclination angles with @xmath59 in the range of @xmath60 to @xmath61 .
the exact value of @xmath14 for each qso can not be constrained directly from the sdss observations . in this section
, we set a fixed value of @xmath62 for the fitting , with which the results obtained here can be directly compared with that found in @xcite and @xcite . with the virial mass of a central mbh and the above settings on the mbh spin and inclination angle , it is possible to estimate the accretion rate for the qso by matching the observed monochromatic luminosity in an optical band to the model ones . in the sample of @xcite ,
the monochromatic luminosities are provided at two wavelengths for those qsos with @xmath41 or @xmath43 ( i.e. , at @xmath63 and @xmath64 , or at @xmath64 and @xmath65 ; see section [ sec : sdssqso ] ) . for these cases ,
the two estimates of the accretion rate generally differ by no more than @xmath66 dex according to our calculations .
we adopt the mean value of the two estimations of the accretion rate , obtained from the two monochromatic luminosity measurements , as the qso accretion rate , and the uncertainty of this rate is @xmath67 dex , which may lead to an uncertainty of @xmath67 dex in the efficiency estimation . however , such an amount of uncertainty seem negligible compared with the biases induced by the usage of the virial mass as the true mbh mass as mainly discussed in section [ subsec : bias_m ] below . the obtained value of @xmath13 for a qso may be biased because of the usage of the virial mass as the true mbh mass and the fixed values for the mbh spin and the disk inclination angle here . the biases induced by these settings on @xmath13 and consequently the @xmath0 estimations will be further analyzed in section [ sec : bias ] .
it is necessary to estimate its bolometric luminosity with considerable accuracy to obtain the radiative efficiency @xmath0 of a qso . for a qso with multi - wavelength observations
, its bolometric luminosity can be directly obtained by integrating its spectrum over the range from radio to hard x - ray but excluding the reprocessed radiation in the infrared .
however , most sdss qsos have been observed only in a limited wavelength range , e.g. , in the optical band .
it may not be an appropriate way to directly convert the optical band luminosity to the bolometric luminosity according to a theoretical accretion disk model , as the radiation processes for high energy photons ( which contribute significantly to the total luminosity ) emitted from the inner disk region are complicated and not fully understood ( e.g. , * ? ? ?
here we adopt an empirical method to calculate the bolometric luminosity of individual qsos by constructing a spectral energy distribution ( sed ) for individual qsos based on some empirical relations obtained for small samples of qsos , which is similar to that adopted in @xcite and ( * ? ? ? * and see references therein ) . .
the solid and dashed lines show an example sed of that adopted in this study ( see section [ subsec : lbol ] ) and that adopted in ( * ? ? ?
* see case a in section 3.1 therein ) , respectively . ] at @xmath63 for the type 1 sdss qsos at redshift @xmath20 ( black dots ; totally 5,786 sdss qsos ) and the dl qsos ( black crosses ; adopted from @xcite ) , respectively .
solid and dashed lines represent the mean value of @xmath69 for the sdss qsos ( @xmath70 ) and the dl qsos ( @xmath71 ) , respectively .
] the spectra or seds of individual qsos are constructed as follows . 1 . in the optical - uv bands ( @xmath72m@xmath73 ) ,
the qso spectrum is assumed to be a power law , and we assign the power - law index to each qso according to a gaussian distribution with mean @xmath74 with scatter @xmath75 ( see * ? ? ?
the spectrum is normalized to the monochromatic luminosity at @xmath63 , i.e. , @xmath76 .
( if @xmath77 is not available for some qsos , we estimate it by extrapolation from @xmath78 or @xmath79 ) .
2 . in the uv wavelengths ( @xmath80-@xmath81 ) ,
the qso spectrum is also described by a power law , and we randomly assign the power - law index according to a gaussian distribution with mean @xmath82 with a scatter of @xmath83 @xcite .
3 . at the x - ray beyond @xmath60kev
, the spectrum is assumed to be a power law with an exponential cutoff at @xmath81kev .
we assign the photon index according to a gaussian distribution with mean @xmath84 and a scatter of @xmath85 for each qso ( * ? ? ?
* see also @xcite ) .
a reflection component is generated for each qso by using the pexrav model @xcite in the xspec package with a reflection solid angle of @xmath86 , inclination of @xmath87 and solar abundance .
the x - ray spectrum is re - normalized to a given @xmath88 $ ] , and the points at @xmath81 and @xmath89 are connected with a power - law .
the value of @xmath90 depends on luminosity , and we adopt the most recent determination by ( * ? ? ?
* see also @xcite ) , i.e. , @xmath91 with an intrinsic scatter of @xmath92 .
6 . at wavelengths
longer than @xmath72 m , the qso spectrum may be dominated by the emission from the dusty torus by re - processing the optical - uv - x - ray photons radiated from the inner disk , which should not be counted in the calculations of the bolometric luminosities of qsos .
we assume a power law continuum emission with slope @xmath93 , similar to that adopted in @xcite .
figure [ fig : f1 ] shows an example sed constructed for a qso with @xmath94 according to the above prescription .
as seen from figure [ fig : f1 ] , the general sed adopted in this study ( the solid line ) is somewhat different from that adopted in ( * ? ? ?
* the dashed line ) , and the differences include ( 1 ) the reflection component of the x - ray emission due to the disk is considered in this study but not in @xcite ; ( 2 ) our choice of the intrinsic x - ray continuum above @xmath57 kev is also slightly different from that in @xcite ; ( 3 ) the continuum slope at optical to uv band is set to be @xmath95 in our study ; but in @xcite it is set to be @xmath96 between @xmath72 m and @xmath97 , @xmath98 between @xmath99 and @xmath100 , and set to be the one measured for each qso in the wavelength range from @xmath97 to @xmath99 and from @xmath100 to @xmath57 kev , respectively .
however , these differences do not lead to much difference in the bolometric corrections as shown in figures [ fig : f2 ] and [ fig : f3 ] . by integrating over the constructed spectrum of each qso in the sample according to the above prescription
, we obtain its `` bolometric luminosity '' , and the distribution of the `` bolometric luminosity '' of these qsos should statistically reflect the underlying real distribution of the sample . given the bolometric luminosity , we also obtain the bolometric correction ( bc ) at @xmath63 for each qso with the luminosity measurement at @xmath63 , i.e. , @xmath101 .
figure [ fig : f2 ] shows the bcs at @xmath63 obtained for the sdss qsos at redshift @xmath20 .
as seen from figure [ fig : f2 ] , the obtained bcs do not depend on @xmath102 , which is consistent with the results obtained by @xcite for a sub - sample of pg qsos in a similar redshift range ( the plus symbols in figure [ fig : f2 ] ) .
the range of the bcs for the sdss qsos is also roughly consistent with that for the dl sample .
the mean logarithmic bc value of the dl qsos is @xmath103 with a scatter of @xmath104 , while that of the sdss qsos is @xmath105 with a scatter of @xmath106 ( as shown by the solid and dashed lines in figure [ fig : f2 ] , respectively ) .
the mean value of the logarithmic bcs for the sdss qsos is slightly larger than that of the dl sample , which might be due to the choice of the empirical sed .
for example , if we choose @xmath107 as given by @xcite , then we would have @xmath108 for the sdss qsos .
other estimations of the bcs for qsos at @xmath63 include @xmath109 @xcite , @xmath110 @xcite , and @xmath111 @xcite , etc . , which are roughly consistent with that shown in figure [ fig : f2 ] and suggest that the systematic uncertainties in the bcs should not exceed @xmath112 dex .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^ " , ] figure [ fig : f3 ] shows the bcs at @xmath63 obtained for the sdss qsos in different redshift bins , and these redshift bins are separated by boundaries of @xmath113 , @xmath60 , @xmath114 , @xmath115 , @xmath116 , @xmath117 , @xmath118 , @xmath119 , @xmath120 , @xmath121 , @xmath122 , @xmath123 and @xmath124 ( see * ? ? ?
as seen from figure [ fig : f3 ] , the bcs do not depend on @xmath3 in all the redshift bins ; and the bcs in different redshift bins spread in roughly the same range , though the mean value of the bcs decreases slightly from @xmath125 to @xmath126 with redshift increasing from @xmath127 to @xmath128 .
the increasing of the mean value of bcs with redshift is primarily a consequence of that @xmath90 decreases with increasing luminosity and the high redshift qsos in general have larger luminosities . in this section
, we estimate the radiative efficiency for each sdss qso in our sample ( see section [ sec : sdssqso ] ) .
we first estimate the @xmath13 for each qso by matching the observational monochromatic luminosity at @xmath63 ( or @xmath64 , or @xmath65 ) with both the bb model and the tlusty model ( see section [ subsec : mdot ] ) .
if monochromatic luminosities are available at more than one wavelength , we obtain @xmath13 for each monochromatic luminosity and take the mean value of the @xmath13 estimates as the accretion rate of the qso .
we also estimate the bolometric luminosity @xmath1 for each qso by using the empirical method described in section [ subsec : lbol ] . with the estimated @xmath13 and @xmath1
, the radiative efficiency of a qso can be then obtained through @xmath129 .
figure [ fig : f4 ] shows the distribution of the sdss qsos in the @xmath0@xmath102 plane , where @xmath0 is estimated through the bb model . from left to right and top to bottom , each panel shows the results for a given redshift bin , and these redshift bins are separated by boundaries of @xmath113 , @xmath60 , @xmath114 , @xmath115 , @xmath116 , @xmath117 , @xmath118 , @xmath119 , @xmath120 , @xmath121 , @xmath122 , @xmath123 and @xmath124 ( see * ? ? ?
as seen from figure [ fig : f4 ] , the radiative efficiencies of the sdss qsos appear to be strongly correlated with the masses of the central mbhs in each redshift bin with @xmath21 .
we perform spearman s rank order correlation analysis for the qsos in each redshift bin and the results are listed in table [ tab : t1 ] . if adopting the tlusty model , we obtain similar results , see table [ tab : t1 ] . apparently , the correlation between @xmath0 and @xmath102 is strong in the redshift bins with @xmath130 , while it becomes weak at redshift @xmath22 .
we adopt a power - law form to fit such a correlation , i.e. , @xmath131 where @xmath18 and @xmath16 are the two fitting parameters .
the parameters that best fit the estimated @xmath132 versus @xmath133 relation in each redshift bin are listed in table [ tab : t1 ] . for the lowest redshift bin ( @xmath134 ) , @xmath135 and @xmath136 if adopting the bb model and the tlusty model , respectively , which is roughly consistent with that obtained in ( * ? ? ?
* @xmath137 ) for the dl qsos in a similar redshift range
. the slight difference might be due to the sed shape adopted here is different from that in the dl sample and the selection criteria for the dl qso sample is different from that for the sdss sample . in the first panel ( @xmath138 ) , the dl qsos in @xcite are also plotted ( plus symbols ) and apparently the distribution of these dl qsos are also roughly consistent with the sdss qsos .
the fitting value of the power - law slope ( @xmath16 ) decreases from @xmath139 to @xmath140 with increasing redshift from @xmath141 to @xmath142 ( see table [ tab : t1 ] and figure [ fig : f4 ] ) . for the whole sdss qso sample ,
the best fit gives @xmath143 , which is substantially smaller than the best fit of @xmath16 ( @xmath144 ) for the lowest redshift bin ( @xmath134 ) .
note other studies , e.g. , @xcite , @xcite , @xcite , and @xcite , in addition to @xcite , also introduced the dependence of the radiative efficiency of qsos on the mbh mass ( and/or redshift ) , but based on the global integral properties of qsos through the sotan argument @xcite .
the left panel of figure [ fig : f5 ] shows the mean value of the logarithmic radiative efficiency of all the sdss qsos in each redshift bin , and the right panel of figure [ fig : f5 ] shows that of the sdss qsos within a given mass range in each redshift bin .
apparently there is no significant redshift evolution of the mean logarithmic efficiency ( e.g. , @xmath145 if adopting the bb model , and @xmath146 if adopting the tlusty model ; see table [ tab : t1 ] and figure [ fig : f5 ] ) .
the mean value of @xmath147 obtained through the tlusty model is slightly smaller than that obtained through the bb model by @xmath148 dex .
considering of the uncertainties in the bc estimations , the mean value of @xmath132 shown in the left panel of figure [ fig : f5 ] may be overestimated or underestimated at most by @xmath149 dex ( see discussions in the end of section [ subsec : lbol ] ) , i.e. , @xmath150 ( or @xmath151 ) by using the bb model ( or the tlusty model ) , which corresponds to a mean efficiency of @xmath152 ( or @xmath153 ) .
therefore , the mean efficiency of those sdss qsos should be in the range of @xmath154 , which is totally independent of but consistent with those estimations based on the sotan argument ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we obtain similar results if alternative using the mean value of @xmath0 or luminosity - weighted @xmath0 . in the right panel of figure
[ fig : f5 ] , we show the mean values of @xmath147 for those qsos in each redshift bin with mbh masses in the range of @xmath155@xmath156 , @xmath157@xmath158 and @xmath159@xmath160 , respectively , which are also roughly consistent with being a constant over the redshift range from @xmath113 to @xmath161 ( except a seemingly slight increase at @xmath162 for the qsos with mbhs in the range of @xmath155@xmath163 and @xmath157@xmath158 ) .
it appears that the mean logarithmic efficiency for the qsos with large mbh masses is larger than that with small mbh masses ( right panel of figure [ fig : f5 ] ) .
however , this tendency seems to be mainly caused by the selection effects but not intrinsic ( as seen also from figure [ fig : f4 ] and discussions in section [ sec : bias ] ) .
the apparent strong dependence of @xmath0 on @xmath102 shown in figure [ fig : f4 ] appears to suggest that more massive mbhs statistically rotate faster than less massive mbhs in qsos , as discussed in @xcite .
however , it is possible that this apparent correlation is only a result of the selection effects of the sdss qso sample and the biases induced by various assumptions in the estimations of @xmath0 , and the intrinsic radiative efficiency of individual sdss qsos may be not correlated with the mbh true mass ( * ? ? ? * but @xcite ) . to investigate whether the above @xmath164 correlation is intrinsic or
not , the key is in understanding the selection effects and the various biases involved in the @xmath0 estimations , which will be studied in details in the following section [ sec : bias ] .
figure [ fig : f6 ] shows the distribution of the sdss qsos in the eddington ratio ( @xmath165 ) versus the mbh virial mass ( @xmath166 ) plane , where @xmath167 and @xmath168 .
as seen from figure [ fig : f6 ] , the eddington ratio is strongly anti - correlated with the mbh virial mass for the sdss qsos in every redshift bin , similar to that found by ( * ? ? ?
* see figure 13 therein ) for @xmath6 pg qsos at low redshift ( @xmath20 ) . generally the larger the mbh virial mass , the smaller the eddington ratio .
however , one should note that the obtained distribution of the eddington ratios in the @xmath169 plane is also affected by the selection effects and the biases induced by the usage of the mbh virial mass ( but not the mbh true mass ) .
similar to that shown for the radiative efficiency in figures [ fig : f9 ] and [ fig : f10 ] in section [ sec : bias ] , the correlation between eddington ratio and mbh virial mass can also be explained by the selection effects and the biases induced by the usage of the mbh virial masses . in this study , we do not intend to expand the discussion on eddington ratios as that done for radiative efficiencies in section [ sec : bias ] .
the above estimations of @xmath13 may be not accurate as ( 1 ) the adopted virial masses of mbhs deviate from the true masses ; ( 2 ) the mbh spins and inclination angles are assumed to be fixed values but in reality they are probably not ; and ( 3 ) the accretion disk model(s ) may be oversimplified .
figure [ fig : f7 ] illustrates the uncertainties in the @xmath170 estimations induced by these various factors as detailed below .
first , the uncertainty due to the usage of the virial mass as the true mbh mass .
the estimated virial masses of mbhs may deviate from the true masses with a scatter of @xmath52 dex ( e.g. , * ? ? ?
* ; * ? ? ?
we perform following calculations to quantify the uncertainty of the @xmath13 estimation induced by the usage of the virial mass ( rather than the true mbh mass ) . assuming a typical qso , of which the true mbh mass is @xmath171 and spin is @xmath53 , accreting material via a true rate of @xmath172 ( corresponding to an eddington ratio of @xmath113 ) , and the inclination angle of the disk to the line of sight is @xmath173acos@xmath174 .
the optical band luminosity at @xmath63 of this system is predicted to be @xmath175 by adopting the bb model . assuming a virial mass of the mbh @xmath102 , scattered around @xmath176
, we may estimate the accretion rate @xmath13 according to @xmath177 and @xmath102 through the procedures described in section [ subsec : mdot ] .
for this specific case , @xmath63 , @xmath64and @xmath65 are all at the left side of the turnover of the disk emission spectrum ( which is also true for more than @xmath178 of the type 1 sdss qsos ) . clearly , @xmath13 can be underestimated ( or overestimated ) by adopting the bb model if @xmath3 is larger ( or smaller ) than @xmath176 ( see the blue circles in the left panel of figure [ fig : f7 ] ) . for a typical deviation of @xmath102 from @xmath176 , i.e. , @xmath113 dex
, @xmath13 may be either underestimated or overestimated by a factor of @xmath179 .
if alternatively adopting the tlusty model for the same system , we have @xmath180 . similarly we can obtain the best fit of @xmath181 for each set of ( @xmath76 , @xmath102 ) , as represented by the red triangles in the left panel of figure [ fig : f7 ] . in these cases ,
the uncertainties in the estimated @xmath13 are similar to that by the bb model above . for an observational sample of qsos , the virial masses of the qso central mbhs are more likely to be overestimated ( or underestimated ) at the high ( or low ) mass end , which may consequently leads to the overestimate ( or underestimate ) of the radiative efficiencies and
thus could contribute to the apparent correlation found in section [ sec : estepsilon ] ( see also the discussions in @xcite ) .
however , we note here that the turnover of the disk emission spectra may move to the left side of @xmath65 for some qsos with extreme large mbh masses ( @xmath182 ) and small eddington ratios ( @xmath183 ) . for these rare cases ,
if the virial masses of the qso central mbhs are overestimated , the radiative efficiencies may be consequently underestimated , which is totally different from that for qsos with smaller mbh masses and high eddington ratios .
second , the uncertainties due to the usage of a fixed value of @xmath12 and a fixed value of @xmath14 .
we first assume that the true mbh mass @xmath184 and inclination angle @xmath14 of the systems are known and arbitrarily adopt a value of @xmath12 to do the fitting in order to check the uncertainty solely due to the set of a fixed @xmath12 ( different from the intrinsic value ) .
we find that the best - fit of @xmath13 only deviates slightly , at most @xmath185 dex or @xmath186 , from its true value ( see the middle panel of figure [ fig : f7 ] ) .
the reason is that @xmath187 is insensitive to the mbh spin at @xmath188 in the thin accretion disk model(s ) for qsos .
if choosing larger @xmath189 ( e.g. , @xmath190 ) , the uncertainty introduced to the estimate of @xmath13 is no more than @xmath89% by an arbitrary set of @xmath12 ( see also * ? ? ?
most sdss qsos are type 1 qsos , their inclination angles @xmath14 may be uniformly distributed in @xmath59 between @xmath60 and @xmath61 .
we now assume that the true mbh mass @xmath184 and spin @xmath12 of the systems are known and obtain the best fit values of @xmath13 by arbitrarily setting the value of @xmath14 .
the uncertainty induced by the set of a fixed value for @xmath14 ( @xmath59 ) is shown in the right panel of figure [ fig : f7 ] , which are at most @xmath57 dex . in this section ,
we set a fixed value of @xmath173acos@xmath174 for the fitting , with which the results obtained here can be directly compared with that found in @xcite and @xcite .
third , the uncertainty induced by the choice of a specific accretion disk model .
the real accretion process may be not accurately reflected by either the bb model or the tlusty model . adopting a specific disk model
may then introduce some ( systematic ) bias to the accretion rate estimations . to illustrate this uncertainty , we first generate a number of qsos with optical luminosity predicted by the tlusty ( or bb ) model , but then adopt the bb ( or tlusty ) model to estimate @xmath181 by using @xmath191 and @xmath3 .
we find that the estimated @xmath13 could be systematically under - estimated or ( over - estimated ) due to the choice of an accretion disk model that not exactly reflect the underlying true disk accretion physics ( for example , see the magenta squares shown in the left panel of figure [ fig : f7 ] ) .
we further perform monte - carlo calculations to illustrate the uncertainties in the accretion rate estimations due to the above factors .
we first generate mock qso samples in which all qsos have the same intrinsic properties ( @xmath192 and @xmath193 ) .
the spin and inclination angle are set to be @xmath53 and @xmath173acos@xmath174 , respectively , for all the qsos in the first sample ; while @xmath173 acos@xmath174 and @xmath12 is set to be uniformly distributed in the range from @xmath194 to @xmath195 for the qsos in the second sample ; and @xmath53 and @xmath14 is assumed to be uniformly distributed in @xmath59 between @xmath60 and @xmath61 for the qsos in the third sample . we randomly assign @xmath102 to each mock qso in the first sample according to a gaussian distribution around @xmath176 with a scatter of @xmath113 dex .
we obtain @xmath191 by both the bb model and the tlusty model for each qso , respectively .
we obtain three samples of qsos with given observational properties of @xmath191 and @xmath102 ( or @xmath184 ) .
we adopt the procedures described in section [ subsec : mdot ] to estimate @xmath13 for these mock qsos .
figure [ fig : f8 ] shows the probability distribution of @xmath196 for the first ( left panel ) , second ( middle panel ) and third sample ( right panel ) , respectively . in each panel ,
the solid and dashed lines show the results obtained by adopting the bb model and the tlusty model , respectively . in the left panel , @xmath13
are obtained for each qso in the first sample by assuming @xmath53 , @xmath197 , and the virial mass as the `` true '' mbh mass ; in the middle and right panel , @xmath13 is obtained for each qso in the second and third samples by assuming @xmath198 , @xmath87 and @xmath53 .
figure [ fig : f8 ] clearly illustrates that the uncertainties of the @xmath181 estimations induced by the deviations of the virial masses of mbhs from the true masses could be as large as a factor of @xmath199 and can have significant effects on the @xmath0 estimations , while the errors due to inaccurate settings of @xmath12 , @xmath14 and the choice of the disk model are relatively less significant . if alternatively choosing other values of @xmath176 and @xmath13 , we obtain similar results as that shown in figure [ fig : f8 ] . in this section
, we investigate whether the correlation between radiative efficiency and mbh mass is intrinsic or simply due to various biases in the accretion rate estimations and/or the selection effects of the sdss qso sample ( mainly a flux limit sample ) .
first , we generate mock sdss qsos based on the true mass function of mbhs ( bhmf , @xmath200 ) and the true eddington ratio ( @xmath165 ) distribution function ( erdf , @xmath201 ) ) obtained by ( * ? ? ?
* see their tables 1 and 2 ) . according to the bhmf and the erdf at each redshift bin
, we randomly assign the mbh mass @xmath202 ( in the range from @xmath203 ) and the eddington ratio @xmath165 ( in the range from @xmath204 to @xmath205 ) to each mock qso .
@xcite argued that the eddington ratio distribution may be independent of the mbh mass and the difference in the eddington ratio distribution for mbhs with different masses at @xmath206 might be only an effect of the uncertainties in the mbh mass estimations by the usage of the civ lines .
we assume that the eddington distribution is the same for mbhs with different masses . assuming that @xmath207 is randomly distributed over the range from @xmath208 to @xmath209 , corresponding to a spin range from @xmath194 to @xmath210 , ( similar to the mean efficiency shown in figure [ fig : f3 ] ; or @xmath211 or @xmath57 ) , corresponding to a spin @xmath212 ( or @xmath213 or @xmath214 )
, we find little changes of our following conclusions . ]
the intrinsic accretion rate is then @xmath215 .
the inclination angle @xmath14 is assumed to be uniformly distributed in @xmath59 from @xmath60 to @xmath61 . given the intrinsic parameters of each mock qso , i.e. , @xmath189 , @xmath12 , @xmath216 and @xmath14 , the optical band luminosity @xmath76 can be obtained by adopting the bb ( or tlusty ) model .
a virial mass of the central mbh in each mock qso is assigned according to a gaussian distribution with a mean of @xmath217 and a scatter of @xmath113 dex . with the above procedure
, we generate @xmath218 mock qsos in each redshift bin , and finally we have a large number of mock qsos that have not only the `` observational measurements '' of their virial mass @xmath166 and the optical band luminosity(/-ies ) @xmath191 ( and/or @xmath219 , and/or @xmath79 ) but also known intrinsic properties of @xmath176 , @xmath181 and @xmath0 .
the optical band luminosities are produced at @xmath63 for those qsos with @xmath44 , at @xmath63 and @xmath64 for those qsos with @xmath41 , at @xmath64 for those qsos with @xmath220 , at @xmath64 and @xmath65 with @xmath43 , and at @xmath65 for those qsos with @xmath221 , respectively , in order to mimic the sdss qso sample .
we take these mock qsos as the parent populations of the qsos in each redshift bin that can be detected by a survey like the sdss . with the `` observational '' properties of the mock qsos , i.e. , @xmath222 and @xmath191 ( and/or @xmath223 , and/or @xmath79 )
, we can adopt the same procedure as that described in section [ sec : estepsilon ] to estimate @xmath0 . in order to single out the effects on the @xmath0 estimations due to the deviation of @xmath102 from @xmath176
, we also estimate @xmath0 for the mock qsos by using their optical luminosity @xmath191 ( and/or @xmath78 , and/or @xmath79 ) and the true mbh mass @xmath189 .
we further consider the selection criteria for those mock qsos similar to that for the sdss qsos , i.e. , only those mock qsos with apparent @xmath14-band magnitude @xmath224 at @xmath29 and @xmath225 at @xmath31 can be taken as the mock sdss qsos ( see section [ sec : sdssqso ] ) . to select mock sdss qsos , we first convert @xmath191 ( or @xmath223 , or @xmath79 ) to the monochromatic luminosity at @xmath14-band ( @xmath226 ) by assuming a power law spectrum with a canonical slope of @xmath227 , and then use the @xmath228-correction , including the effects of both the continuum and the emission lines , given by @xcite to calculate the apparent @xmath14-band magnitude of each mock qso . here
we adopt @xmath229 because the k - corrections given by @xcite are obtained by using a canonical slope of @xmath230 .
finally , we obtain the mock sdss qsos according to the above selection criteria , which are consist of a small fraction ( @xmath231 ) of the parent mock sample at each redshift bin . here
we adopt the standard @xmath232cdm cosmology with @xmath233 ) .
similar to that in section [ sec : estepsilon ] , we do the simulations by adopting both the bb model and the tlusty model , and we find that the results obtained from two different models are similar .
for simplicity , only the results obtained by adopting the bb model are presented below in this section .
figure [ fig : f9 ] shows the normalised probability density distribution of the mock qsos on the @xmath132@xmath3 ( or @xmath132@xmath176 ) plane . without considering the selection effects , the radiative efficiencies @xmath0 of the mock qsos estimated from @xmath77 and @xmath176 appear to be independent of @xmath189 ( see the red contour lines in each redshift bin ) . by performing spearman
s rank correlation analysis , we do not find any strong correlations between @xmath0 and @xmath189 for the whole mock sample ( without considering the selection effects ) .
the estimated @xmath0 scatter around the mean of the input values @xmath234 mainly because the bc is randomly assigned to each mock qso according to the empirical sed of qsos as described in section [ subsec : lbol ] . ) of the input efficiencies for some redshift bins mainly because of the following reason .
the efficiency for each qso is estimated by using the bolometric luminosity obtained empirically and the accretion rates obtained by adopting the standard thin disk model .
this method is not completely self - consistent , i.e. , the efficiency estimate could be slightly offset from the input intrinsic efficiency for the standard thin disk model if the empirical sed is not exactly the same as the sed predicted by the standard thin disk model , which is a caveat of both our model and other previous models ( e.g. , * ? ? ?
] for the mock sdss qsos , there is an obvious strong correlation between the radiative efficiencies @xmath0 ( estimated from @xmath191 and @xmath189 ) and the true mbh mass @xmath189 ( see the black contour lines in each panel of figure [ fig : f9 ] ) .
similar to that for the sdss qsos , we adopt a power law form to fit these correlations , i.e. , @xmath235 , and find @xmath236 and @xmath237 for the redshift bins from low to high , respectively . further considering the deviation of @xmath102 from @xmath189 , we find that the radiative efficiency @xmath0 ( estimated from @xmath191 and @xmath102 ) are strongly correlated with @xmath102 ( see the filled color contours in figure [ fig : f9 ] ) . adopting the same power - law form to fit these correlations , i.e. , @xmath238 , we find @xmath239 and @xmath240 for the redshift bins from low to high , respectively .
it appears that the slope @xmath16 is larger at low redshift ( @xmath241 ) and smaller at high redshift ( @xmath22 ) , compared with the cases without considering the deviation of @xmath102 from @xmath189 , which is mainly due to the biases induced by the deviations of the virial masses from the true mbh masses .
similar to that for the sdss qsos , the slope @xmath16 is relatively large at low redshift and small at high redshift .
figure [ fig : f10 ] shows the comparison of the normalised probability density distributions of the mock sdss qsos with that of the sdss qsos in the @xmath132@xmath102 plane .
apparently , the distributions of the mock sdss qsos in the @xmath132@xmath242 plane are roughly consistent with that of the sdss qsos in all the redshift bins . according to figures [ fig : f9 ] and [ fig : f10 ] ,
it is clear that the selection effects can be the dominant factors that cause the strong correlation between @xmath0 and @xmath102 , and the usage of the virial mass @xmath102 as the true mbh mass also contribute some to the correlation but it is less significant comparing with the selection effects .
as discussed in section [ subsec : bias_m ] , the uncertainties in the @xmath0 estimations induced by the usage of a constant inclination angle @xmath14 play little role in generating the @xmath0@xmath243 correlation .
the simplest explanation for the arouse of an apparent correlation between @xmath0 and @xmath166 , as an addition to the above monte - carlo analysis , is as follows . in the standard thin accretion disk scenario ,
the accretion rate is related to the optical luminosity as @xmath244 according to @xcite . since @xmath245 , @xmath246 . for a given @xmath15 , the larger the efficiency , the smaller the optical luminosity @xmath247 and thus the larger the probability that the qso is missed in a flux limited sample .
the smaller the mbh mass , the larger the probability that qsos with high radiative efficiencies can be detected by a flux limited survey like sdss . for high redshift bins ( @xmath248 ) , the general tendency of increasing efficiency with increasing
mbh mass is similar to that for low redshift bins if the mbh mass is smaller than a few times @xmath249 ; while it turns over if the mbh mass is larger .
for the qsos with mbh masses larger than a few time @xmath250 at high redshift @xmath251 , the observational measurement is only provided at @xmath65 , which is probably at the right side of the disk emission spectrum . for these cases , the empirical seds adopted in this study , a monotonically increasing function at @xmath72m@xmath252 , are somewhat different from the disk emission spectra , which might lead to underestimation of the bolometric luminosities and thus underestimation of the radiative efficiencies . this could be part of the reason that the estimated radiative efficiencies of many qsos with mbh mass @xmath253 at high redshift bins ( @xmath254 ) deviate from the correlation between efficiency and mbh mass at lower redshift .
furthermore , the accretion rate could be overestimated by using the monochromatic luminosity @xmath79 if the mbh virial mass is an overestimate of the true mbh mass ( probably true for many qsos at the high mass end at high redshift , e.g. , @xmath255 ) since the turnover of the disk emission spectra may move to the left side of @xmath65 for some qsos with extreme large mbh mass ( @xmath256 ) and small eddington ratio ( @xmath183 ; see discussion in section [ subsec : bias_m ] ) .
such a behavior is quite different from that for the cases with lower mbh masses and higher eddington ratios .
the turnover of the trend for the qso radiative efficiencies in high redshift bins ( @xmath254 ) at high mass end can also be partly explained by this .
one may note , however , that the probability distributions of the mock sdss qsos in the @xmath132@xmath242 plane seem to be offset from that of the sdss qsos .
the slopes @xmath16 of the best fits are also not exactly the same as that of the sdss qsos . especially in the high redshift bins ( @xmath22 ) , the @xmath0@xmath3 correlation obtained from the mock qsos is still significant , while that obtained from the sdss qsos is very weak . the above differences between the mock sdss qsos and the sdss qsos might be not a surprise as a number of complications are not considered in the above @xmath0 estimations for both the sdss qsos and the mock qsos .
first , we do not consider the contamination by the host galaxy light and the dust extinction for each individual sdss qso , which may introduce errors to both the @xmath181 and @xmath1 estimations and consequently the @xmath0 estimations for the sdss qsos ( see also @xcite ) .
second , the bolometric luminosity @xmath46 is obtained according to the empirical sed of qsos and its associated scatters , which are obtained from the multi - wavelength observations of samples with a small number of qsos that only cover limited ranges of luminosity and redshift ( see section [ subsec : lbol ] ) .
the usage of this sed may introduce not only scatters but also systematic biases in the @xmath0 estimations .
the small spread in the shapes of the adopted empirical seds may be also important for the resulted statistics of the @xmath0 estimations , as pointed out by @xcite .
third , the mock qsos are generated according to the true bhmf and erdf given by @xcite . in @xcite , the obtained bhmf and erdf
can best fit the qso observations , however , they are obtained by assuming a constant bolometric correction . since the bcs of the type
1 qsos may have a large scatter and may also depend on the eddington ratio and the true mbh mass ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the bhmf and erdf obtained by @xcite may still deviate somewhat from the true distributions , which may further introduce some uncertainties in the distributions of the mock sdss qsos in the @xmath132@xmath242 plane .
one needs to self - consistently consider all the factors , including the bhmf , erdf , bcs and the radiative efficiencies , etc .
, simultaneously , to fully solve this problem .. in principle , one could also check whether can an intrinsic correlation between @xmath207 and @xmath176 be ruled out ( or confirmed ) by comparing the results obtained from the sdss qsos with that obtained from mock sdss qso samples .
to check this , we re - generate the mock sdss qso sample by assuming that @xmath257 , as an alternative to the initial setting of a random @xmath207 in the above calculations . according to this new mock sdss qso sample , we find that the resulted correlation between @xmath0 and @xmath3 is generally steeper than that obtained above by assuming a constant @xmath207 in each redshift bin .
for example , @xmath258 , and @xmath259 for the first three redshift bins , respectively .
the differences in @xmath16 may help to distinguish whether there is indeed an intrinsic correlation between @xmath207 and @xmath176 . because of the various complications and large uncertainties in the @xmath0 estimations ( see discussions in the above paragraph ) , however , it is still difficult to fully rule out the existence of an intrinsic @xmath207@xmath176 correlation just by comparing the results obtained from the sdss qsos with that obtained from such mock samples . we conclude that the intrinsic correlation between the radiative efficiency and the true mbh mass , if any , must be much weaker than the apparent correlation between @xmath0 and @xmath3 found for the sdss sample in each redshift bin . to close this section
, we conclude that the strong correlations between @xmath0 and @xmath102 found in section [ sec : estepsilon ] can be produced by and mainly due to the selection effects of the sdss qso sample and the biases in the @xmath0 estimations induced by the usage of @xmath102 as the true mbh mass . by studying a number of qsos ( or agns ) with different luminosities and mbh masses and redshifts ,
@xcite find that the @xmath0@xmath102 relation found by @xcite may be an artifact of the small parameter space covered by the dl sample . with the large data set of the sdss qso catalog
, we have demonstrated that a correlation between @xmath0@xmath166 can be generated by mocking sdss qsos through modelling various selection effects and biases ( though there is no input intrinsic correlation between @xmath0@xmath189 ) , and our results are consistent with that of @xcite .
note that the empirical sed of qsos , more or less uniform , is adopted to calculate the bolometric luminosity of each sdss qso in section [ subsec : lbol ] . as argued in @xcite , the small spread in the sed shape
is an intriguing question that currently does not have a simple answer .
radiative efficiencies of the disk accretion processes in individual qsos are related to the spins of the central mbhs , which may be profoundly connected to the mbh assembly history as suggested by a number of recent studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is therefore of great importance to estimate the radiative efficiency of individual qsos and investigate its statistical distribution among qsos . in this study
, we estimate the radiative efficiency individually for a large number of sdss qsos .
we first estimate the accretion rate for each qso by matching the detected optical band luminosity(/-ies ) with that predicted by the disk model , by adopting the thin disk accretion model and assuming that the true mass of the central mbh is the same as that given by the virial mass estimator(s ) .
we also estimate the bolometric luminosity of each qso by adopting the empirical spectral energy distribution suggested by various multi - wavelength observations of small qso samples . with the estimated accretion rate and the bolometric luminosity
, we obtain the radiative efficiency for each sdss qso .
we find an apparent strong correlation between the radiative efficiency and the mbh virial mass in low redshift bins and it becomes weak in high redshift bins . in the lowest redshift bin ( @xmath260 ) , this apparent correlation ( @xmath261 , and @xmath262 ) is roughly consistent with that found by @xcite for the dl qsos in a similar redshift range .
we also find that the mean radiative efficiencies of the sdss qsos are consistent with being a constant @xmath154 ( though with large scatters ) over the redshift range from @xmath113 to @xmath161 , which does not suggest any significant evolution with redshift .
this estimate of the mean radiative efficiency of qsos is totally independent of but roughly consistent with those estimations based on the sotan argument ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . with the enormous large sample of the sdss qsos ,
it is possible to statistically model the various biases in the estimations of the radiative efficiency and the selection effects of the sdss qsos .
to do so , we generate mock sdss qso samples according to the true mbh mass function and the eddington ratio distribution obtained in @xcite , by involving the selection criteria of the sdss qsos and the uncertainties in the mbh virial mass estimations and the inclination angles .
we estimate the radiative efficiency for qsos in each mock sample by adopting the same method as that for the sdss qso and obtain the probability density distribution of those mock qsos in the radiative efficiency versus the mbh virial mass plane .
we find that the @xmath0@xmath102 correlations for the sdss qsos in different redshift bins can be well explained by the selection effects and the biases induced by the usage of @xmath102 as the true mbh mass , and the selection effects play the dominant roles in leading to the @xmath0@xmath102 correlation , as suggested by @xcite .
we conclude that the current sdss qso data is consistent with no intrinsic correlation between the qso radiative efficiency and the true mbh mass . in principle ,
the accretion rate of a qso may be better determined by fitting the qso spectrum covering a wide range of wavelengths ( e.g. , from the infrared band to the hard x - ray band ) , through an elaborate accretion disk model , and the constraints on the mbh spin and consequently the radiative efficiency may be also simultaneously obtained . in the future , if observations can obtain the spectra for a large sample of qsos , with good wavelength , luminosity and redshift coverage , it may be possible to estimate their radiative efficiencies in a more self - consistent way and then further investigate the relationship , if any , between the radiative efficiency and the mbh mass , which would shed light on not only the assembly history of mbhs ( e.g. , * ? ? ?
* ) but also the physical reasons for the small spread in the sed shapes of qsos pointed by @xcite .
we thank qingjuan yu for constructive suggestions and discussions , andreas schulze and changshuo yan for helpful discussions .
this work was supported in part by the national natural science foundation of china under nos .
10973017 and 11033001 , the national basic research program ( 973 programme ) of china ( grant 2009cb824800 ) , and the bairen program from the national astronomical observatory of china , chinese academy of sciences . | we estimate the radiative efficiency @xmath0 of individual type 1 sdss qsos by using their bolometric luminosities ( @xmath1 ) and accretion rates ( @xmath2 ) , which may be related to the assembly histories and spins of the central massive black holes ( mbhs ) .
we estimate @xmath1 by using the empirical spectral energy distributions of qsos and @xmath2 by fitting the observed optical luminosity(/-ies ) with the thin accretion disk model , assuming the mbh masses given by the virial mass estimator(s ) ( @xmath3 ) .
we find an apparent correlation between @xmath0 and @xmath3 , which is strong at redshift @xmath4 , weak at @xmath5 , and consistent with that found by @xcite for @xmath6 pg qsos at @xmath7 . to investigate whether this correlation is intrinsic or
not , we construct a mock sample of qsos according to the true mbh mass and eddington ratio distributions given in @xcite . by comparing the results obtained from the mock sample with that from the sdss sample ,
we demonstrate that the apparent @xmath8 correlation can be produced by and mainly due to the selection effects of the sdss sample and the bias induced by the usage of @xmath3 as the true mbh mass .
the mean values of @xmath0 of those sdss qsos are consistent with being a constant @xmath9 over the redshift range of @xmath10 .
we conclude that the current sdss qso data is consistent with no strong intrinsic correlation between radiative efficiency and true mbh mass and no significant redshift evolution of radiative efficiencies .
accretion , accretion discs black hole physics
galaxies : nuclei galaxies : active quasars : general |
the study of the thermodynamics of photons dates back to planck @xcite . investigating blackbody radiation
, he realized photons decay due to absorption into walls of their container , and therefore , no chemical potential appeared in his expression , in contrast to gibbs s thermodynamic expressions for other particles using the grand canonical ensemble .
later , it was understood that in the absence of absorbing walls , photon can acquire non - zero chemical potential , e.g. photon emission in semiconductors ( led ) @xcite , and thus the useful concept of chemical potential can start to be applied to these systems @xcite . moreover ,
if photons are confined in a cavity and coupled to excitons , they form polaritons which also can thermalize @xcite .
more recently , it was shown that photons can thermalize with a non - zero chemical potential and form a bose - einstein condensate @xcite when interacting with a nonlinear medium . however , finding a general solution to creating a chemical potential for light remains an open problem @xcite .
at the same time , photons provide an intriguing quantum degree of freedom for implementing quantum simulators @xcite and observing quantum phases of matter @xcite . in quantum simulation ,
one develops a quantum system with a controlled , known hamiltonian , enabling simulation of problems that are exponentially difficult on a classical computer .
this new paradigm covers a wide range of problems from chemistry @xcite and quantum field theories @xcite to strongly correlated electron systems , such as high-@xmath0 superconductors @xcite .
we apply our scheme to two platforms , circuit - qed and optomechanical systems , where recent and spectacular progress has been made in controlling and using them in a few quanta regime .
finally , we conclude by considering how a photonic lattice implementing a bose - hubbard model can be driven through the mott insulator - superfluid ( mi - sf ) transition @xcite using this approach even in the presence of finite dissipation .
we can understand thermalization via a system - bath picture , where the system of choice with hamiltonian @xmath1 is coupled via @xmath2 to a bath with hamiltonian @xmath3 and initial state @xmath4 @xcite .
h = h_s + 2(_p t)h_sb + h_b .
[ e : ham ] again with initial conditions @xmath4 .
we assume that parametric drive can be characterized by a classical field which can not be depleted .
the parametric coupling will enable up- and down - conversion of bath excitations to photons , which will lead to a controlled chemical potentials .
and response functions with a cutoff @xmath5 can be parametrically coupled to a higher frequency ( optical ) system with modes @xmath6 near the frequency @xmath7 .
additional loss via the high frequency bath can lead transport from the parametric bath through the system to the high frequency bath .
, scaledwidth=45.0% ] to see this explicitly , we will assume that @xmath8 is bi - linear , of the form h_sb = _ j ( _ j + ^_j ) b _ j [ eq : linear_coupling ] where @xmath9 is a bath operator and there exists @xmath6 , @xmath10 such that @xmath11 = \hat{a}_j$ ] , as occurs naturally for photons .
this property defines particle numbers @xmath10 and total particle number @xmath12 . .
furthermore , we will decompose @xmath1 into @xmath13 where @xmath14 includes all terms that do not commute with the total number of particles in the photonic system , given by @xmath15 . in this regime , we move to a rotating frame with the unitary transformation @xmath16 .
the transformed system hamiltonian becomes u^h_s u - iu^ h_s - _ p , where we have neglected @xmath17 by making the rotating wave approximation ( rwa ) , requiring @xmath18 .
meanwhile , the bath hamiltonian remains the same , while the system bath coupling terms become b_j _ j the key approximation is again the rwa to neglect @xmath19-type terms , consistent for a bath whose two - point bath correlation function @xmath20 has a cutoff frequency @xmath21 .
this provides our definition of a low frequency bath for this paper , with @xmath22\hat{b}_j$ ] the system - bath coupling in the rwa . through this set of transformations , and the rotating wave approximation
, we have a new system - bath hamiltonian which takes the traditional form h = h_s - + h_sb + h_b where we identity @xmath23 as the chemical potential . for weak coupling @xmath24 and an infinite bath at inverse temperature @xmath25
, we expect the system to thermalize in the long - time limit to a density matrix , [ eq : rho_thermal ] i.e. , the distribution is exactly that of the grand canonical ensemble .
now we show that such a scheme , which provides both thermalization and a finite chemical potential for photons , can be implemented in circuit - qed systems for microwave domain photons and using optomechanics for optical domain photons . following the caldeira - leggett model , in the context of circuits @xcite
, we consider the bath to be a collection of transmission lines which can be described by a quasi - continuum of harmonic oscillators .
the bath hamiltonian is given by h_b= _ _
( ^ _ _ + ) , where @xmath26 is the creation operator of an electromagnetic field quantum at mode @xmath27 with frequency @xmath28 .
we assume that the transmission lines are in thermal equilibrium , and thus , @xmath29 . we consider that each mode of the photonic system is coupled to the bath using non - degenerate parametric amplifiers , through three - wave mixing .
while many configurations can implement this concept @xcite , we focus on the conceptually cleanest case : a josephson parametric amplifier in a wheatstone bridge configuration @xcite , as depicted in fig .
[ wheatstone ] . of the coupler .
the system is connected to the mode @xmath30 .
the mode @xmath31 is driven harmonically at frequency @xmath32 and provides the up- and down - conversion necessary for particle and hole exchange with the bath .
, scaledwidth=45.0% ] examining the details of the jj - wheatstone parametric coupler , we assume that each junction has a large area , and hence , a large capacitance , so that its charging energy can be ignored . in this approximation ,
the energy @xmath33 of the jj - wheatstone bridge is @xcite @xmath34 \nonumber\end{aligned}\ ] ] where we have taken all four jj s to have the same @xmath35 , and @xmath36 , @xmath37 being the superconducting flux quantum .
setting @xmath38 by choice of flux bias , and assuming the mode intensities @xmath39 , consistent with moderate to low characteristic impedance circuits , we can expand @xmath33 in @xmath40 , @xmath41 to third order @xcite : u = -2 e_j + m ( _ x^2 + _ s^2 + _
z^2 ) + g _ x _ s _ z . here
@xmath42 and @xmath43 .
for the optical domain , we need a different parametric process .
a convenient one is the optomechanical coupling between motion of a mirror and the frequency of light in a cavity formed by the mirror .
this example case has been worked in partial detail in ref .
the key idea is for a pump field to take the radiation pressure coupling @xmath44 to a fast oscillating coupling via @xmath45 , producing a parametric coupling to the phonon `` bath '' with frequency @xmath46 .
the details and benefits of the optomechanical approach will be considered in a separate work .
note that in any experimental implementation , one needs to filter out the pump photons from the signal system photons .
this can be easily achieved by using different polarization or spatial modes of the photonic system on each site .
alternatively , in certain schemes , one can reject the pump by frequency filtering .
for example , in the mott insulator case , discussed later in this article , the pump has higher frequency than the prepared mott state , and therefore , the pump can be filtered out by frequency selection .
we now examine our assumption of a cutoff in the bath degrees of freedom , as well as a strictly parametric system - bath coupling . for simplicity
, we divide the bath modes into three , independent sets of modes , and consider coupling to a single system mode @xmath47 . given a parametric coupling at frequency @xmath27 , the low frequency modes of the bath , @xmath48 , are defined as those with natural resonance frequencies @xmath49 .
the ` natural ' modes , @xmath50 , are those with frequencies @xmath51 .
the ` doubly rotating ' modes , @xmath52 , are those with frequencies @xmath53 .
thus , the more general system - bath interaction is @xmath54 ( a + a^\dag ) \\ & \times & \sum_j f_j ( b_j + b_j^\dag ) + g_j ( c_j + c_j^\dag ) + h_j ( d_j + d_j^\dag ) \nonumber\end{aligned}\ ] ] we now move to an appropriate rotating frame , with @xmath55 , @xmath56 , @xmath57 , and @xmath58 . with the assumption of weak coupling ( @xmath59 small ) ,
we look at the rotating wave approximation for the different couplings : @xmath60 \left(a e^{-i \nu t } + a^\dag e^{i \nu t}\right)\sum_j f_j ( b_j + b_j^\dag ) \nonumber \\ & \rightarrow \frac{1}{2 } \lambda ( a + a^\dag ) \sum_j f_j ( b_j + b_j^\dag ) \\ h_{sb , c } & = \left[a + \lambda \cos(\nu t ) \right ] \nonumber \\ & \times \left(a e^{-i \nu t } + a^\dag e^{i \nu t}\right)\sum_j g_j ( c_j e^{-i \nu t } + c_j^\dag e^{i \nu t } ) \\ & \rightarrow a \sum_j g_j ( a^\dag c_j + c_j^\dag a ) \\
h_{sb , d } & = \left[a + \lambda \cos(\nu t ) \right ] \left(a e^{-i \nu t } + a^\dag e^{i \nu t}\right ) \nonumber \\ & \times \sum_j h_j ( d_j e^{-2
i \nu t } + d_j^\dag e^{2 i \nu t } ) \\ & \rightarrow \frac{1}{2 } \lambda \sum_j h_j ( a^\dag d_j + d_j^\dag a ) \end{aligned}\ ] ] by breaking up the bath into three different frequency regions , we see that the ` natural ' and ` doubly - rotating ' frequency regions both lead to a system - bath interaction of the quantum optics type , i.e. , that of a beam splitter interaction @xmath61 . for these portions of the system - bath interaction , we may then proceed in deriving the master equation in the usual way @xcite , and find , in appropriate limits , a decay of excitations of @xmath47 at a rate @xmath62 , where @xmath63 is the bath density of states near the parametric modulation frequency @xmath27 .
we now investigate the remaining portion of the system - bath interaction in a specific setting , to illustrate the emergence of a mott insulator - superfluid transition in a photonic lattice .
we consider now what happens to a lattice of coupled , interacting photonic resonators , coupled to both a parametric bath at inverse temperature @xmath25 and nominal coupling rate @xmath64 and to a high frequency ( loss ) bath with loss rate @xmath65 . for simplicity
, we consider only strong on - site repulsion @xmath33 , and have for the conservative parts of the evolution , a bose - hubbard hamiltonian @xcite in the rotating frame : @xmath66 \ \textrm{and } \\ h_j = & { } -j \sum_{{\langleij\rangle } } a_i^\dag a_j,\end{aligned}\ ] ] we explicitly derive the master equation for the system , using the usual prescription : first , move to the interaction picture with respect to @xmath67 , where @xmath3 is the bath hamiltonian and the @xmath24 prefactor in the system - bath coupling will be a perturbative parameter .
we can write the evolution equation for short times @xmath68 as @xmath69 - \lambda^2 \int_0^\tau [ h_{sb}(\tau),[h_{sb}(t),\rho_i(t ) ] ] dt\ ] ] with @xmath70 the system - bath coupling in the interaction picture , writing @xmath71 . now we make the born and markov approximations .
that is , we replace @xmath72 with @xmath73 . here
@xmath74 is the bath density matrix which will be time - translation invariant for an infinite bath , and is independent of @xmath75 with @xmath76 = 0 $ ] for all bath operators coupled to the system .
from these two approximations , we can trace over the bath and recover the master equation ( in the interaction picture ) @xmath77 \nonumber \\ & { } \
+ s_{ij}(-t ) [ \rho_s x_j(\tau - t ) x_i(\tau ) - x_j(\tau - t ) \rho_s x_i(\tau ) ] dt\end{aligned}\ ] ] with @xmath78 $ ] the bath correlation function and where , by taking the initial integration point to @xmath79 , we have assumed that bath correlations decay faster than the effective damping they induce consistent with the markov approximation .
at this point , we wish to develop a time - local master equation .
we express @xmath80 in the energy eigenbasis of @xmath1 , with states @xmath81 and energies @xmath82 and an ordering in energy such that @xmath83 .
then @xmath84 formally defines an operator that reduces or keeps constant the energy , and @xmath85 , with the last term time - independent and neglected in what follows .
thus , when the energy levels of the system are well resolved , we can derive a super operator describing both photon loss and coupling to the parametric bath . using the commutation of @xmath1 with @xmath86 ( the total photon number ) , we get transitions from @xmath87 to @xmath88 with rates that depend on whether the total photon number of the two states differs by @xmath89 or @xmath90 as : @xmath91 \sum_i \left| \bra{l } a_i \ket{k } \right|^2 \label{e : gammaminus } \end{aligned}\ ] ] where @xmath92 for the ohmic bath case , @xmath93 represents the overall strength of the coupling , and @xmath94 is the heaviside step function .
we have gone back to the physical couplings @xmath95 rather than the many - body energy lowering operator @xmath50 in order to make clear the special role loss via the high frequency bath plays in eq .
[ e : gammaminus ] .
the superoperator takes lindblad form with these rates leading to a rate equation in the energy eigenbasis . solving this numerically for a case of four coupled sites ( fig .
[ f : exact ] ) , we can immediately see an intuitive understanding of the two types of decay processes .
the first type , which increases photon number , corresponds to the decay of holes ( if the energy of the higher photon number state is lower in the rotating frame ) or the creation of particles ( if otherwise ) .
the second type decreases photon number , and includes both creation of holes via loss and via the parametric bath ; consequently , we expect a greater rate for the second process , which will lead to a particle - hole temperature asymmetry as shown below .
the simulations themselves correspond to fixing a maximum total particle number per site , finding the eigenenergies of the dissipation - free model , calculating the decay rates in eqs .
( [ e : gammaplus ] ) and ( [ e : gammaminus ] ) , determining the steady state of the master equation , and for that steady state , finding the probability of each state ( shown in the inset to fig .
[ f : exact ] ) , and estimating the mandel parameter and the average hopping @xmath96 ( shown in fig .
[ f : meanfield ] ) .
energy eigenstates plotted as a function of energy and total photon number @xmath86 for an numerical solution of a four site bose - hubbard model ( shown in the upper inset ) with @xmath97 and @xmath98 .
the opacity of the blue dots represent the probability , in steady state , of being in the associated energy eigenstate .
the lower inset shows the region near the ground state in the rotating frame ; hole - like excitations ( lower @xmath86 ) are preferentially filled due to optical loss processes @xmath65 only reducing particle number .
the relatively high temperature leads to some thermal filling of the first particle excited state.,scaledwidth=45.0% ]
we now take a simpler form of the superoperator describing both photon loss and coupling in the case of a single resonator site ( @xmath99 ) to get an analytical handle on the process .
that is , we evaluate eqs .
( [ e : gammaplus ] ) and ( [ e : gammaminus ] ) in the single site case . specifically , defining @xmath100 , the sign of @xmath101 determines both the direction of decay and the thermal bosonic enhancement factor @xmath102 .
thus @xmath103 with @xmath104\ , \\
f_-(n ) & = ( n+1 ) \left [ n_{th}(|\delta e(n)| ) + \theta(\delta e(n ) ) \right]\ .\end{aligned}\ ] ] and @xmath105 for the ohmic bath case .
the change from @xmath106 to @xmath107 that occurs in these two factors with the change in sign of @xmath108 arises from having both co- and counter - rotating terms in the system bath coupling .
one consequence of the strong interaction ( sometimes called strong coupling in the mott insulator literature ) limit ( @xmath109 ) is an analytical form for the steady state .
specifically , we recover a form of detailed balance , where the probability of a transition on a site from photon number @xmath110 to @xmath111 is given by @xmath112 while the transition from @xmath111 to @xmath110 is @xmath113 .
this gives , in steady state , a set of ratios @xmath114 where the correction from a thermal distribution arises from the term @xmath115 , which depends on the energy difference via @xmath64 .
we can characterize this for two regimes .
first , when @xmath108 is positive ( it costs energy to add a photon ) , we expect the ratio @xmath116 .
this defines the bosonic occupation as seen by particle addition as @xmath117 thus , when particles cost energy , photon loss reduces the effective temperature of the system .
similarly , when @xmath108 is negative , we expect @xmath118 , which defines the bosonic occupation as seen by hole addition : @xmath119}{1 - \kappa/\gamma}\ ] ] here , photon loss increases the energy , and thus increases the effective temperature of the system .
furthermore , any hope of a thermal description will necessarily breakdown for @xmath120 .
having established that a mott insulator - like phase emerges from the single site picture , we can ask how the asymmetry of particles and holes changes the standard picture of the edges of the mott lobes , by using a picture of free particles and holes above the @xmath121 particle - per - site mott state @xmath122 , i.e. , using small @xmath123 perturbation theory in the strong interaction limit . a crucial difference from the standard treatment @xcite
is the use of an implicit finite lifetime to such excitations due to the coupling to both parametric and high frequency baths .
when @xmath123 exceeds damping and dephasing , we can no longer use a master equation appropriate to a single site . specifically , as we want the parametric bath to resolve the kinetic terms in the hamiltonian , we require @xmath124 \ll j$ ] .
we can , however , characterize the particle or hole occupation for a wave vector @xmath87 in the dilute limit ( where particle hole collisions are neglected ) by using our @xmath125)}$ ] , and we can ask over what domain of parameter space is the combined occupation of particles and holes small compared to one per site . here
we rely upon the standard picture of particle and hole energies to order @xmath126 , neglecting loss - induced changes to the energy differences , consistent with @xmath127 .
the energy of a particle(hole ) of wave vector @xmath128 above the mott state is given by ref .
@xcite and reproduced here to order @xmath126 : @xmath129 where @xmath130 is the number of nearest neighbors .
we can then calculate the average particle and hole expectation values including both the parametric bath and the high frequency ( photon loss ) bath , and find that these lowest energy modes have just @xmath131 and @xmath132 with the above @xmath133)}$ ] .
the boundary of the phase would then correspond to this effective occupation approaching unity ( at which point we may expect a macroscopic occupation of particles and/or holes in the system , taking us far from the mott state ) .
this boundary is shown for two different values of @xmath134 in fig . [
f : meanfield ] ; as @xmath65 increases , the lobes become asymmetric , consistent with additional hole creation via particle losses . numerical results for mandel @xmath135 ( left ) and coherence @xmath96 ( right ) using the four - site ( top ) and a six - site ( bottom ) bose - hubbard model with periodic boundary conditions .
we assume an ohmic parametric bath and a flat high frequency ( loss ) bath .
the mandel @xmath136 regions ( dark blue , left plot ) are the mott insulator states ; at the same time , the finite coherence between sites on the right indicates the emergence of superfluid order ( right plot ) outside the mott lobes .
finite size effects prevent observation of sharp transitions . here
@xmath137 and @xmath138 .
overlaid are the critical values for finite occupation of particles and holes ( solid black lines ) , with the dotted line for a higher value of @xmath139 .
the asymmetry of particles and holes arises due to preferential hole creation from optical loss.,scaledwidth=50.0% ] we now consider what near equilibrium picture can emerge , and in particular focus on a picture with two reservoirs ( particles and holes ) at different temperatures due to loss into the high frequency bath . in the limit of @xmath140
, we recover the usual picture of an equilibrium system , and get a critical temperature defined as @xmath141\ ] ] however , including the non equilibrium effects , we instead have for the parametric bath temperature the requirement @xmath142\ ] ] where @xmath143}$ ] depends on @xmath144)}$ ] via @xmath145 .
further analysis of the particle - hole picture at finite temperature will no doubt elucidate additional physics for this non equilibrium system , following perhaps the efforts of refs .
in addition , an appropriate mean field theory including modifications of the system - bath coupling could provide insight into the applicability of such theories for describing non - equilibrium systems .
providing a robust chemical potential for light allows for classical and quantum systems to access a wide variety of heretofore forbidden domains .
crucially , our approach allows one to build from well established theoretical tools for non equilibrium problems with chemical potential imbalances , such as occurs in circuits and cold atom systems , rather than the thornier problems associated with driven steady - state systems more typical to the quantum optical domain . from a quantum simulation perspective , this simplification makes the state preparation problem much more straightforward than existing approaches , and yields a mechanism for robust quantum simulation of condensed matter and chemistry problems with light . in addition , our parametric coupling scheme has a wide range of potential implementations , all of which are accessible with current technology , and enables a variety of practical applications in the context of non - classical sources in the microwave and optical domain that operate more in analogy to a diode than to a pumped dissipative steady - state system .
we thank s. girvin , a. houck , b. l. hu , j. keeling , j. freericks , m. devoret , e. kapit , and p. zoller for helpful discussions .
support was provided by the nsf - funded physics frontier center at the jqi and by aro muri grant no .
as a simple test of these concepts , we implement numerically a model for the intermediate time behavior of a two - level system ( qubit ) coupled to a bosonic bath via a parametric coupling . the usual picture of quantum brownian motion @xcite has a set of bath modes coupled linearly through their position variables @xmath146 with constant @xmath147 and mass @xmath148 .
this leads to the effective spectral density @xmath149 where @xmath150 is the density of states .
our goal will be to well approximate such a bath with a discrete set of modes . before engaging in that , we mention some rescaling of the problem appropriate to simulation . first , we rewrite the system - bath coupling @xmath151 in terms of bath creation and annihilation operators with @xmath152 .
this defines @xmath153 and @xmath154 .
our goal is to approximate the bath such that the correlation function and the commutation relation of the bath are as close to the desired approximate bath as possible . specifically , for our quantum brownian motion bath with a cutoff function @xmath155 considered in this work , we assume the time - ordered spectral function used in sec . 5 : @xmath156 \nonumber \\ & \approx \sum_{j=1}^n \frac{{\bf
w}_j j(\omega_j)}{\pi f(\omega_j ) } \left [ ( n_{th}(\omega_j)+1 ) e^{-i \omega_j t } + n_{th}(\omega_j ) e^{i \omega_j t } \right]\end{aligned}\ ] ] where the approximation of the integral as a finite sum arises from gaussian quadrature over @xmath86 orthogonal polynomials under the function @xmath155 to find the set @xmath157 and the associated weights @xmath158 , and @xmath159 $ ] .
we remark that in the case of the ohmic bath with exponential cutoff function , the appropriate choice is the laguerre polynomials .
a simple reinterpretation of this formula is that of a discrete harmonic oscillator ( quantum brownian motion ) bath with frequencies @xmath160 and a coupling constants @xmath161 this approximation is immediately amenable to numerical techniques via direct integration of the schrdinger equation . in practice ,
exponential cutoffs at the relevant frequencies are unlikely as the superconductors work well into the ghz domain for our implementation .
thus , we consider a polynomial cutoff function induced by filtering the ohmic bath with a low - pass filter , such as a capacitor in parallel with the resistor forming the ohmic bath for our circuit case .
the impedance of this system becomes : @xmath162 , where @xmath163 is the characteristic time of the rc circuit .
therefore , the real part of the impedance , which appears in the spectral noise @xcite , leads to a natural modification of the effective spectral density : @xmath164 .
we neglect imaginary contributions to the circuit by assuming they are renormalized in the system hamiltonian .
we note that for an optomechanical implementation such cutoff functions arise from the cavity lorentzian and can have a similar functional form
quadratic suppression at high frequency .
we simulate the following simple case numerically to illustrate our system . working with the discrete bath approximation and a photon - blockade - regime cavity with an effective two - level system description with pauli matrices @xmath165 , etc . , we write h = _ 0 _ z / 2 + _
j g_j _ x ( b_j + b_j^ ) + _ j b_j^b_j with @xmath166 .
the parameters @xmath167 and @xmath24 represent the relative strength of the regular exponential decay bath and the additional parametric bath terms oscillating at frequency @xmath168 .
we take @xmath169 and @xmath170 , and find that decreasing or increasing @xmath171 by even a factor of two does not appreciable change the results presented below .
we also work in units of time given by @xmath172 . for improved computation speed ,
we truncate the bath hilbert space to a maximum of two bosonic excitations , and confirm post - facto that simulations produce only slightly more than one bath excitation , consistent with the truncation .
we first test the purely ohmic case , taking @xmath173 and @xmath174 ( no parametric bath ) .
we find exponential decay with a time scale @xmath175 .
furthermore , this decay is well approximated ( with around @xmath176 errors ) up to times @xmath177 for @xmath178 .
the fitted decay rate is independent of @xmath86 in this range of values , consistent with our approximation scheme .
we then consider @xmath179 with the filter on ( @xmath180 ) , leading to a slightly reduced decay rate due to the non - ohmic nature of the bath near @xmath181 from the cutoff filter .
still , exponential decay is observed over two decades with @xmath182 , and residuals are at the 3% level or less , largely due to corrections to exponential decay at long times from non - ohmic bath behavior . the decay from an initially excited state @xmath183 into the zero temperature bath is shown in red in fig .
[ f : decay ] .
after these simple tests of our model system , we move to the more complicated regime of a parametrically coupled bath . taking @xmath184 and @xmath185 , we start the system in the lower - energy spin state , @xmath186 .
we calculate over 100 time units three different values of our chemical potential parameter , @xmath187 , and plot the resulting @xmath188 as a function of time .
we expect the system to approach spin up for @xmath189 according to the chemical potential derivation given in the first part of this part , and find these expectations confirmed in this simple numerical experiment ( fig . [ f : decay ] ) . interacting normally ( red ) and parametrically ( green , yellow , blue ) with a high - frequency filtered bath . for an initial excited state @xmath190 , coupling to
the bath leads to exponential decay when the oscillating frequency of the bath is set to zero , as shown in red .
however , for an initial ground state @xmath191 , turning on the parametric coupling to the bath such that it oscillates at frequencies @xmath192 0.9 ( blue ) , 1.0 ( yellow ) , 1.1 ( green ) @xmath193 leads to inversion of the spin when @xmath194 , as predicted by our more general theoretical model .
[ f : decay],width=288 ] | while we have a concept of temperature for energy conservation , there is no equivalent chemical potential for particle number conservation .
however , the notion of a chemical potential is crucial in understanding a wide variety of single- and many - body effects , from transport in conductors and semiconductors to phase transitions in electronic and atomic systems .
here we show how a direct modification of the system - bath coupling via parametric oscillation creates an effective chemical potential for photons even in the thermodynamic limit .
specific implementations , using circuit - qed or optomechanics , are feasible using current technologies , and we show a detailed example demonstrating the emergence of mott insulator - superfluid transition in a lattice of nonlinear oscillators .
our approach paves the way for quantum simulation , quantum sources , and even electron - like circuits with light . |
often theoretical methods ( in particular computational techniques ) are developed in response to emerging scientific challenges in specific fields .
the development of the shell correction method ( scm ) by strutinsky in the late 1960 s @xcite was motivated by the observation of large nonuniformities ( oscillatory behavior ) exhibited by a number of nuclear properties as a function of the nuclear size .
these properties included : total nuclear masses , nuclear deformation energies , and large distortions and fission barriers . while it was understood already that the total energy of nuclei can be decomposed into an oscillatory part and one that shows a slow `` smooth '' variation as a function of size , strutinsky s seminal contribution was to calculate the two parts from different nuclear models : the former from the nuclear shell model and the latter from the liquid drop model .
in particular , the calculation of the oscillatory part was enabled by employing an averaging method that smeared the single particle spectrum associated with a nuclear model potential .
it is recognized that the strutinsky procedure provides a method which `` reproduces microscopic results in an optimal way using phenomenological models '' @xcite ; in the appendix we describe an adaptation of the strutinsky phenomenological procedure to metal clusters ; we term this procedure as the semi - empirical ( se)-scm .
in this chapter , we focus on our development in the early 1990 s of the microscopic density - functional - theory ( dft)-scm approach @xcite , where we have shown that the total energy of a condensed - matter finite system can be identified with the harris functional [ see eq .
( [ enhar ] ) ] , with the shell correction [ eq .
( [ densh ] ) ] being expressed through both the kinetic energy , @xmath0 , of this functional [ eq . ( [ tsh ] ) ] and the kinetic energy of an extended - thomas - fermi ( etf ) functional expanded to fourth - order density gradients [ see @xmath1 in eq .
( [ t4th ] ) ] .
it is important to note that in our procedure an optimized input density is used in the harris functional .
this optimization can be achieved through a variational procedure [ using an orbital - free ( of ) energy functional , e.g. , the etf functional with 4th - order gradients ] with a parametrized trial density profile [ see eq .
( [ rho ] ) ] , or through the use [ see eq .
( [ rhoalt ] ) ] of the variational principle applied to an orbital - free energy functional .
( for literature regarding orbital - free kinetic - energy functionals , see , e.g. , refs . @xcite . )
a similar optimization of an of/4th - order - etf density has been shown to be consistent with the strutinsky averaging approach @xcite .
such 4th - order optimization of the input density renders rather ambiguous any direct ( term - by - term ) comparison between the method proposed by us and subsequent treatments , which extend the dft - scm to include higher - order shell - correction terms without input - density optimization @xcite ( see also ref .
indeed , the input - density optimization ( in particular with the use of 4th - order gradients ) minimizes contributions from higher - order shell corrections . in light of certain existing similarities between the physics of nuclei and clusters ( despite the large disparity in spatial and energy scales and the different origins of inter - particle interactions in these systems ) , in particular the finding of electronic shell effects in clusters @xcite , it was natural to use the jellium model in the early applications of the dft - scm to clusters .
however , as noted @xcite already early on , `` the very good agreement between our results and those obtained via kohn - sham self - consistent jellium calculations suggested that it would be worthwhile to extend the application of our method to more general electronic structure calculations extending beyond the jellium model , where the trial density used for minimization of the etf functional could be taken as a superposition of site densities , as in the harris method . ''
additionally , generalization of the dft - scm method to calculations of extended ( bulk and surface ) systems appeared rather natural .
indeed , recent promising applications of dft - scm in this spirit have appeared @xcite . in this case , the term `` shell correction effects '' is also maintained , although `` quantum interference effects '' could be more appropriate for extended systems .
one of the principle themes in research on finite systems ( e.g. , nuclei , atomic and molecular clusters , and nano - structured materials ) is the search for size - evolutionary patterns ( seps ) of properties of such systems and elucidation of the physical principles underlying such patterns @xcite .
various physical and chemical properties of finite systems exhibit seps , including : .
structural characteristics pertaining to atomic arrangements and particle morphologies and shapes ; excitation spectra involving bound - bound transitions , ionization potentials ( ips ) , and electron affinities ( eas ) ; collective excitations ( electronic and vibrational ) ; magnetic properties ; abundance spectra and stability patterns , and their relation to binding and cohesion energetics , and to the pathways and rates of dissociation , fragmentation , and fission of charged clusters ; thermodynamic stability and phase changes ; chemical reactivity . clusters .
open squares : experimental measurements of refs .
@xcite . the solid line at the top panel ( a ) represents the smooth contribution to the theoretical total ips . the solid circles in the middle ( b ) and bottom ( c ) panels are the total scm ips .
the shapes of sodium clusters have been assumed spherical in the middle panel , while triaxial deformations have been considered at the bottom one . ]
the variations with size of certain properties of materials aggregates are commonly found to scale on the average with the surface to volume ratio of the cluster , i.e. , @xmath2 , where @xmath3 , @xmath4 , @xmath5 , and @xmath6 are the surface area , volume , average radius , and number of particles , respectively ( the physical origins of such scaling may vary for different properties ) . in general ,
the behavior of seps in finite systems in terms of such scaling is non - universal , in the sense that it is non - monotonic exhibiting characteristic discontinuities .
nevertheless , in many occasions , it is convenient to analyze the energetics of finite systems in terms of two contributions , namely , ( i ) a term which describes the energetics as a function of the system size in an average sense ( not including shell - closure effects ) , referrred to usually as describing the smooth " part of the size dependence , and ( ii ) an electronic shell - correction term .
the first term is the one which is expected to vary smoothly and be expressible as an expansion in @xmath7 , while the second one contains the characteristic oscillatory patterns as the size of the finite system is varied .
such a strategy has been introduced @xcite and often used in studies of nuclei @xcite , and has been adopted recently for investigations of metal clusters , fullerenes @xcite , and metal nanowires @xcite . as a motivating example , we show in the sep of the ips of na@xmath8 clusters , which illustrates odd - even oscillations in the observed spectrum , a smooth description of the pattern [ ( a ) ] , and two levels of shell - corrected descriptions one assuming spherical symmetry [ ( b ) ] , and the other allowing for triaxial shape deformations [ ( c ) ] .
the progressive improvement of the level of agreement between the experimental and theoretical patterns is evident .
the chapter is organized as follows : in , the general methodology of shell - correction methods is reviewed , and the microscopic dft - scm is introduced and presented in detail in .
applications of the dft - scm to condensed - matter finite systems are presented in for three different characteristic nanosystems , namely , metal clusters ( ) , charged fullerenes ( ) , and metallic nanowires ( ) . in the appendix
, we describe the semiempirical scm for clusters , which is closer to the spirit of strutinsky s original phenomenological approach for nuclei .
there we also briefly present applications of the se - scm to triaxial shape deformations and fission of metal clusters .
a summary is given in .
it has long been recognized in nuclear physics that the dependence of ground - state properties of nuclei on the number of particles can be viewed as the sum of two contributions : the first contribution varies smoothly with the particle number ( number of protons @xmath9 and neutrons @xmath10 ) and is referred to as the _ smooth _ part ; the second contribution gives a superimposed structure on the smooth curve and exhibits an oscillatory behavior , with extrema at the nuclear magic numbers @xcite .
nuclear masses have provided a prototype for this behavior @xcite .
indeed , the main contributions to the experimental nuclear binding energies are smooth functions of the number of protons and neutrons , and are described by the semi - empirical mass formula @xcite .
the presence of these smooth terms led to the introduction of the liquid - drop model ( ldm ) , according to which the nucleus is viewed as a drop of a nonviscous fluid whose total energy is specified by volume , surface , and curvature contributions @xcite .
the deviations of the binding energies from the smooth variation implied by the ldm have been shown @xcite to arise from the shell structure associated with the bunching of the discrete single - particle spectra of the nucleons , and are commonly referred to as the shell correction .
substantial progress in our understanding of the stability of strongly deformed open - shell nuclei and of the dynamics of nuclear fission was achieved when strutinsky proposed @xcite a physically motivated efficient way of calculating the shell corrections .
the method consists of averaging [ see the appendix , and ] the single - particle spectra of phenomenological deformed potentials and of subtracting the ensuing average from the total sum of single - particle energies .
while certain analogies , portrayed in experimental data , between properties of nuclei and elemental clusters have been recognized , the nuclear - physics approach of separating the various quantities as a function of size into a smooth part and a shell correction part has only partially been explored in the case of metal clusters . in particular ,
several investigations @xcite had used the etf method in conjunction with the jellium approximation to determine the average ( smooth , in the sense defined above ) behavior of metal clusters , but had not pursued a method for calculating the shell corrections . in the absence of a method for appropriately calculating shell - corrections for metal clusters in the context of the semiclassical etf method
, it had been presumed that the etf method was most useful for larger clusters , since the shell effects diminish with increasing size .
indeed , several studies had been carried out with this method addressing the asymptotic behavior of ground - state properties towards the behavior of a jellium sphere of infinite size @xcite .
it has been observed @xcite , however , that the single - particle potentials resulting from the semiclassical method are very close , even for small cluster sizes , to those obtained via self - consistent solution of the local density functional approximation ( lda ) using the kohn - sham ( ks ) equations @xcite .
these semiclassical potentials were used extensively to describe the optical ( linear ) response of spherical metal clusters , for small @xcite , as well as larger sizes @xcite ( for an experimental review on optical properties , cf .
refs.@xcite ) .
the results of this approach are consistent with time - dependent local density functional approximation ( tdlda ) calculations which use the ks solutions @xcite .
it is natural to explore the use of these semiclassical potentials , in the spirit of strutinsky s approach , for evaluation of shell corrections in metal clusters of arbitrary size .
below we describe a microscopic derivation of an scm in conjunction with the density functional theory @xcite , and its applications in investigations of the properties of metal clusters and fullerenes .
particularly interesting and promising is the manner by which the shell corrections are introduced by us at the microscopic level through the kinetic energy term @xcite , instead of the traditional semiempirical strutinsky averaging procedure of the single - particle spectrum @xcite . in particular , our approach leads to an energy functional that corrects many shortcomings of the orbital - free dft , and one that is competitive in numerical accuracy and largely advantageous in computational speed compared to the ks method .
underlying the development of the shell - correction method is the idea of approximating the total energy @xmath11 of a finite interacting fermion system as @xmath12 where @xmath13 is the part that varies smoothly as a function of system size , and @xmath14 is an oscillatory term .
various implementations of such a separation consist of different choices and methods for evaluating the two terms in eq .
( [ enshsm ] ) . before discussing such methods , we outline a microscopic derivation of eq .
( [ enshsm ] ) . motivated by the behavior of the empirical nuclear binding energies , strutinsky conjectured that the self - consistent hartree - fock density @xmath15 can be decomposed into a smooth density @xmath16 and a fluctuating contribution @xmath17 , namely @xmath18 .
then , he proceeded to show that , to second - order in @xmath17 , the hartree - fock energy is equal to the result that the same hartree - fock expression yields when @xmath15 is replaced by the smooth density @xmath16 and the hartree - fock single - particle energies @xmath19 are replaced by the single - particle energies corresponding to the smooth potential constructed with the smooth density @xmath16 .
namely , he showed that @xmath20 where the hartree - fock electronic energy is given by the expression @xmath21 , \label{enhf}\ ] ] with @xmath19 being the eigenvalues obtained through a self - consistent solution of the hf equation , @xmath22 where @xmath23 .
\label{pothf}\ ] ] the strutinsky approximate energy is written as follows , @xmath24 , \label{enstr}\ ] ] where the index @xmath25 in eq .
( [ enhf ] ) and eq .
( [ enstr ] ) runs only over the occupied states ( spin degeneracy is naturally implied ) .
the single - particle energies @xmath26 correspond to a smooth potential @xmath27 .
namely , they are eigenvalues of a schrdinger equation , @xmath28 where the smooth potential @xmath27 depends on the smooth density @xmath29 , i.e. , @xmath30 , \label{pothfsm}\ ] ] and @xmath31 is the nuclear two - body interaction potential . it should be noted that while eq .
( [ enstr])@xmath32 eq . ( [ pothfsm ] ) look formally similar to the hartree - fock equations ( [ enhf]-[pothf ] ) , their content is different . specifically , while in the hf equations , the density @xmath33 is self - consistent with the wavefunction solutions of eq .
( [ sphf ] ) , the density @xmath16 in eq .
( [ enstr])@xmath32 eq . ( [ pothfsm ] ) is not self - consistent with the wavefunction solutions of the corresponding single - particle eq .
( [ sphfsm ] ) , i.e. , @xmath34 .
we return to this issue below . since the second term in eq .
( [ enstr ] ) is a smooth quantity , eq .
( [ estr ] ) states that all shell corrections are , to first order in @xmath17 , contained in the sum of the single - particle energies @xmath35 .
consequently , eq . ( [ enstr ] ) can be used as a basis for a separation of the total energy into smooth and shell - correction terms as in eq .
( [ enshsm ] ) .
indeed strutinsky suggested a semi - empirical method of such separation through an averaging procedure of the single - particle energies @xmath26 in conjunction with a phenomenological ( or semi - empirical ) model [ the liquid drop model ( ldm ) ] for the smooth part ( see the appendix ) .
motivated by the above considerations , we have extended them @xcite in the context of density functional theory for electronic structure calculations .
first we review pertinent aspects of the dft theory . in dft
, the total energy is given by @xmath36=t[\rho ] + \int \left\ { \left [ \frac{1}{2 } v_h [ \rho({\bf r } ) ] + v_i({\bf r } ) \right ] \rho({\bf r } ) \right\ } d\/ { \bf r } + \int { \cal e}_{{xc } } [ \rho({\bf r})]d\/ { \bf r } + e_i , \label{enlda}\ ] ] where @xmath37 is the hartree repulsive potential among the electrons , @xmath38 is the interaction potential between the electrons and ions , @xmath39 is the exchange - correlation functional [ the corresponding xc potential is given as @xmath40 and @xmath41 $ ] is given in terms of a yet unknown functional @xmath42 $ ] as @xmath41 = \int t[\rho ( { \bf r } ) ] d{\bf r}$ ] .
@xmath43 is the interaction energy of the ions . in the kohn - sham ( ks)-dft theory ,
the electron density is evaluated from the single - particle wave functions @xmath44 as @xmath45 where @xmath44 are obtained from a self - consistent solution of the ks equations , @xmath46 \phi_{{ks , i}}({\bf r } ) = \varepsilon_{{ks , i } } \phi_{{ks , i}}({\bf r } ) \label{kseqs}\ ] ] where @xmath47 = v_h [ \rho_{ks } ( { \bf r})]+ v_{{xc}}[\rho_{ks } ( { \bf r } ) ] + v_i({\bf r } ) .
\label{vks}\ ] ] the kinetic energy term in eq .
( [ enlda ] ) is given by @xmath48 = \sum_{i=1}^\mathrm{occ } < \phi_{{ks , i } } | -\frac{\hbar^2}{2 m } \nabla^2 |\phi_{{ks , i } } > , \label{tks}\ ] ] which can also be written as @xmath48 = \sum_{i=1}^\mathrm{occ } \varepsilon_{ks , i } - \int \ ! \rho_{ks } ( { \bf r } ) v_{ks } [ \rho_{ks } ( { \bf r } ) ] d{\bf r}. \label{tks2}\ ] ] according to the hohenberg - kohn theorem , the energy functional ( [ enlda ] ) is a minimum at the true ground density @xmath49 , which in the context of the ks - dft theory corresponds the the density , @xmath50 , obtained from an iterative self - consistent solution of eq .
( [ kseqs ] ) . in other words , combining eq .
( [ enlda ] ) and eq .
( [ tks2 ] ) , and denoting by in " and out " the trial and output densities of an iteration cycle in the solution of the ks equation [ eq . ( [ kseqs ] ) ] , one obtains , @xmath51 & = & e_i + \sum_{i=1}^\mathrm{occ } \varepsilon_{ks , i}^\mathrm{out } + \nonumber \\ ~ & ~ & \int \
! \left\ { \frac{1}{2 } v_h [ \rho_{ks}^\mathrm{out } ( { \bf r } ) ] + { \cal e}_{xc } [ \rho_{ks}^\mathrm{out } ( { \bf r } ) ] + v_i ( { \bf r } ) \right\ } \rho_{ks}^\mathrm{out } ( { \bf r } ) d{\bf r } - \nonumber \\ ~ & ~ & \int \ !
\rho_{ks}^\mathrm{out } ( { \bf r } ) v_{ks } [ \rho_{ks}^\mathrm{in } ( { \bf r } ) ] d{\bf r}. \label{eniter}\end{aligned}\ ] ] note that the expression on the right involves both @xmath52 and @xmath53 .
self - consistency is achieved when @xmath54 becomes arbitrarily small ( i.e. , when @xmath52 converges to @xmath50 ) . on the other hand , it is desirable to introduce approximate energy functionals for the calculations of ground - state electronic properties , providing simplified , yet accurate , computational schemes
. it is indeed possible to construct such functionals @xcite , an example of which was introduced by j. harris @xcite , where self - consistency is circumvented and the result is accurate to second order in the difference between the trial and the self - consistent ks density * ( * see in particular eq .
( 24a ) of ref .
@xcite ; the same also holds true for the difference between the trial and the output densities of the harris functional*)*. the expression of the harris functional is obtained from eq .
( [ eniter ] ) by dropping the label @xmath55 and by replacing everywhere @xmath56 by @xmath57 , yielding [ note cancellations between the third and fourth terms on the right - hand - side of eq .
( [ eniter ] ) ] .
@xmath58 & = & e_i + \sum_{i=1}^\mathrm{occ } \varepsilon_i^\mathrm{out } - \int \ !
\left\ { \frac{1}{2 } v_h [ \rho^\mathrm{in } ( { \bf r } ) ] + v_{xc } [ \rho^\mathrm{in } ( { \bf r } ) ] \right\ } \rho^\mathrm{in } ( { \bf r } ) d{\bf r } + \nonumber \\ ~ & ~ & \int \ ! { \cal e}_{xc } [ \rho^\mathrm{in } ( { \bf r } ) ] d{\bf r}. \label{enhar}\end{aligned}\ ] ] @xmath59 are the single - particle solutions ( non - self - consistent ) of eq .
( [ kseqs ] ) , with @xmath60 [ see eq.([vks ] ) ] . as stated above this result
is accurate to second order in @xmath61 ( alternatively in @xmath62 ) , thus approximating the self - consistent total energy @xmath63 $ ] .
obviously the accuracy of the results obtained via eq .
( [ enhar ] ) depend on the choice of the input density @xmath57 . in electronic structure calculations where the corpuscular nature of the ions is included (
i.e. , all - electron or pseudo - potential calculations ) , a natural choice for @xmath57 consists of a superposition of atomic site densities , as suggested originally by harris . in the case of jellium calculations ,
we have shown @xcite that an accurate approximation to the ks - dft total energy is obtained by using the harris functional with the input density , @xmath57 , in eq .
( [ enhar ] ) evaluated from an extended - thomas - fermi ( etf)-dft calculation .
the etf - dft energy functional , @xmath64 $ ] , is obtained by replacing the kinetic energy term in eq .
( [ enlda ] ) by a kinetic energy density - functional in the spirit of the thomas - fermi approach @xcite , but comprising terms up to fourth - order in the density gradients @xcite .
the optimal etf - dft total energy is then obtained by minimization of @xmath64 $ ] with respect to the density . in our calculations
, we use for the trial densities parametrized profiles @xmath65 with @xmath66 as variational parameters ( the etf - dft optimal density is denoted as @xmath67 ) .
the single - particle eigenvalues , @xmath68 , in eq .
( [ enhar ] ) are obtained then as the solutions to a single - particle hamiltonian , @xmath69 where @xmath70 is given by eq .
( [ vks ] ) with @xmath71 replaced by @xmath72 .
these single - particle eigenvalues will be denoted by @xmath73 as is well known , the etf - dft does not contain shell effects @xcite .
consequently , the corresponding density @xmath67 can be taken la strutinsky as the smooth part , @xmath16 , of the ks density , @xmath50 .
accordingly , @xmath74 is identified with the smooth part @xmath13 in eq.([enshsm ] ) ( in the following , the etf " subscript and
@xmath75 " can be used interchangeably ) . since , as aforementioned ,
@xmath76 $ ] approximates well [ i.e. , to second order in @xmath77 the self - consistent total energy @xmath63 $ ] , it follows from eq .
( [ enshsm ] ) , with @xmath76 $ ] taken as the expression for @xmath78 , that the shell - correction , @xmath79 , is given by @xmath80 - e_{etf}[\rho_{etf } ] \equiv e_{sh}[\widetilde{\rho } ] - \widetilde{e}[\widetilde{\rho } ] .
\label{denhar}\ ] ] defining , @xmath81 and denoting the total energy @xmath82 by @xmath83 , i.e. , by identifying @xmath84 we obtain @xmath85 = \ { t_{sh } - \widetilde{t}[\widetilde{\rho}]\ } + \widetilde{e}[\widetilde{\rho } ] , \label{ensh}\ ] ] where @xmath86 $ ] is the etf kinetic energy , given to fourth - order gradients by the expression @xcite , @xmath87 & = & \frac{\hbar^2}{2 m } \int \ !
\left\ { \frac{3}{5 } ( 3\pi^2)^{2/3 } \rho^{5/3 } + \frac{1}{36 } \frac{(\nabla \rho)^2}{\rho } + \frac{1}{270 } ( 3 \pi^2)^{-2/3 } \rho^{1/3 } \right .
\nonumber \\ ~~ & ~~ & \times \left . \left
[ \frac{1}{3 } \left ( \frac{\nabla \rho } { \rho } \right)^4 - \frac{9}{8 } \left ( \frac{\nabla \rho } { \rho } \right)^2 \frac{\delta \rho } { \rho } + \left ( \frac{\delta \rho } { \rho } \right)^2 \right ] \right\ } d { \bf r } , \label{t4th}\end{aligned}\ ] ] which as noted before does not contain shell effects .
therefore , the shell correction term in eq .
( [ enshsm ] ) [ or eq.([denhar ] ) ] is given by a difference between kinetic energy terms , @xmath88 .
\label{densh}\ ] ] one should note that the above derivation of the shell correction does not involve a strutinsky averaging procedure of the kinetic energy operator .
rather it is based on using etf quantities as the smooth part for the density , @xmath16 , and energy , @xmath13 .
other descriptions of the smooth part may result in different shell - correction terms . to check the accuracy of this procedure ,
we have compared results of calculations using the functional @xmath89 [ eq . ( [ ensh ] ) ] with available kohn - sham calculations . in general , the optimized density from the minimization of the etf - dft functional can be obtained numerically as a solution of the differential equation @xmath90}{\delta{\rho}({\bf r } ) } + v_{etf}[\rho({\bf r})]=\mu , \label{rhoalt}\ ] ] where @xmath91 is the chemical potential .
as mentioned already , for the jellium dft - scm calculations , we often use a trial density profile in the etf - dft variation which is chosen as , @xmath92^{\gamma } } , \label{rho}\ ] ] with @xmath93 , @xmath94 , and @xmath95 as variational parameters that minimize the etf - dft functional ( for other closely related parametrizations , cf .
@xcite ) . ) .
solid circles : dft - scm results ( see text for details ) .
the solid line is the etf result ( smooth contribution ) . in both cases ,
a spherical jellium background was used .
open squares : kohn - sham dft results from ref .
the excellent agreement ( a discrepancy of only 1% ) between the dft - scm and the kohn - sham dft approach is to be stressed . ]
displays results of the present shell correction approach for the total energies of neutral sodium clusters .
the results of the shell correction method for ionization potentials of sodium clusters are displayed in .
the excellent agreement between the oscillating results obtained via our dft - scm theory and the kohn - sham results ( cf .
, e.g. , ref .
@xcite ) is evident . to further illustrate the two components ( smooth contribution and shell correction ) entering into our approach
, we also display the smooth parts resulting from the etf method .
( in all calculations , the gunnarsson - lundqvist exchange and correlation energy functionals were used ; see refs .
@xcite . )
[ [ charging - of - metal - clusters ] ] charging of metal clusters + + + + + + + + + + + + + + + + + + + + + + + + + + investigations of metal clusters based on dft methods and self - consistent solutions of the kohn - sham equations ( employing either a positive jellium background or maintaining the discrete ionic cores ) have contributed significantly to our understanding of these systems @xcite .
however , even for singly negatively charged metal clusters ( @xmath96 ) , difficulties may arise due to the failure of the solutions of the ks equations to converge , since the eigenvalue of the excess electron may iterate to a positive energy @xcite .
while such difficulties are alleviated for @xmath96 clusters via self - interaction corrections ( sic ) @xcite , the treatment of multiply charged clusters @xmath97 would face similar difficulties in the metastability region against electronic autodetachment through a coulombic barrier . in the following we are applying our dft - scm approach ,
described in the previous section , to these systems @xcite ( for the jellium background , we assume spherical symmetry , unless otherwise stated ; for a discussion of cluster deformations , see ) .
[ [ electron - affinities - and - borders - of - stability ] ] electron affinities and borders of stability + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the smooth multiple electron affinities @xmath98 prior to shell corrections are defined as the difference in the total energies of the clusters @xmath99 where @xmath6 is the number of atoms , @xmath100 is the valency and @xmath101 is the number of excess electrons in the cluster ( e.g. , first and second affinities correspond to @xmath102 and @xmath103 , respectively ) .
@xmath104 is the total charge of the positive background . applying the shell correction in eq .
( [ densh ] ) , we calculate the full electron affinity as @xmath105 ) and second ( @xmath106 ) electron affinities of sodium clusters as a function of the number of atoms @xmath6 .
both their smooth part ( dashed lines ) and the shell - corrected affinities ( solid circles ) are shown .
a spherical jellium background was used . ]
a positive value of the electron affinity indicates stability upon attachment of an extra electron . displays the smooth , as well as
the shell corrected , first and second electron affinities for sodium clusters with @xmath107 .
note that @xmath108 becomes positive above a certain critical size , implying that the second electron in doubly negatively charged sodium clusters with @xmath109 might not be stably attached .
the shell effects , however , create two islands of stability about the magic clusters na@xmath110 and na@xmath111 ( see @xmath112 in ) .
to predict the critical cluster size @xmath113 , which allows stable attachment of @xmath101 excess electrons , we calculated the smooth electron affinities of sodium clusters up to @xmath114 for @xmath115 , and display the results in .
we observe that @xmath116 , while @xmath117 . , @xmath102 - 4 , for sodium clusters as a function of the number of atoms @xmath6 ( @xmath101 is the number of excess electrons ) .
a spherical jellium background was used .
inset : the electron _ drip _ line for sodium clusters .
clusters stable against spontaneous electron emission are located above this line .
while for spherical symmetry , as seen from , shell effects influence the border of stability , shell - corrected calculations @xcite including deformations ( see the appendix ) yield values close to the drip line ( shown in the inset ) which was obtained from the smooth contributions . ]
the similarity of the shapes of the curves in , and the regularity of distances between them , suggest that the smooth electron affinities can be fitted by a general expression of the form : @xmath118 where the radius of the positive background is @xmath119 . from our fit , we find that the constant @xmath120 corresponds to the bulk work function . in all cases ,
we find @xmath121 , which suggests a close analogy with the classical model of the image charge @xcite .
for the spill - out parameter , we find a weak size dependence as @xmath122 .
the contribution of @xmath123 , which depends on @xmath101 , is of importance only for smaller sizes and does not affect substantially the critical sizes ( where the curve crosses the zero line ) , and consequently @xmath124 can be neglected in such estimations . using the values obtained by us for @xmath125 of sodium clusters ( namely , @xmath126 which is also the value obtained by ks - dft calculations for an infinite planar surface @xcite , @xmath127 ; with @xmath128 , and @xmath129 ) , we find for the critical sizes when the l.h.s .
( [ asman ] ) is set equal to zero , @xmath130 , @xmath131 , @xmath132 , and @xmath133 , in very good agreement with the values obtained directly from . the curve that specifies @xmath134 in the @xmath135 plane defines the border of stability for spontaneous electron decay . in nuclei ,
such borders of stability against spontaneous proton or neutron emission are known as nucleon drip lines @xcite . for the case of sodium clusters ,
the electron drip line is displayed in the inset of .
[ [ critical - sizes - for - potassium - and - aluminum ] ] critical sizes for potassium and aluminum + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + while in this investigation we have used sodium clusters as a test system , the methodology and conclusions extend to other materials as well .
thus given a calculated or measured bulk work function @xmath120 , and a spill - out parameter ( @xmath136 typically of the order of 1 - 2 @xmath137 , and neglecting @xmath124 ) , one can use eq .
( [ asman ] ) , with @xmath138 , to predict critical sizes for other materials .
for example , our calculations for potassium ( @xmath139 ) give fitted values @xmath140 ( compared to a ks - dft value of 2.54 @xmath141 for a semi - infinite planar surface with @xmath142 @xcite ) and @xmath143 for @xmath144 , yielding @xmath145 , @xmath146 , and @xmath147 . as a further example , we give our results for a trivalent metal , i.e. aluminum ( @xmath148 ) , for which our fitted values are @xmath149 ( compared to a ks - dft value of 3.78 @xmath141 for a semi - infinite plane surface , with @xmath150 @xmath137 @xcite ) and @xmath151 for @xmath144 , yielding @xmath152 ( 121 electrons ) , @xmath153 ( 626 electrons ) , and @xmath154 ( 1796 electrons ) . [ [ metastability - against - electron - autodetachment ] ] metastability against electron autodetachment + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + for the three families of sodium clusters with @xmath155 , @xmath156 , and @xmath157 atoms .
a spherical jellium background was used . as
the straight lines in the inset demonstrate , the curves are parabolic .
we find that they can be fitted by eq .
( [ etz ] ) .
see text for an explanation of how the function @xmath158 was extracted from the calculations . ]
the multiply charged anions with negative affinities do not necessarily exhibit a positive total energy . to illustrate this point , we display in the calculated total energies per atom ( @xmath159 ) as a function of excess charge ( @xmath101 ) for clusters containing 30 , 80 , and 240 sodium atoms .
these sizes allow for exothermic attachment of maximum one , two , or three excess electrons , respectively . as was the case with the electron affinities , the total - energy curves in show a remarkable regularity , suggesting a parabolic dependence on the excess charge . to test this conjecture ,
we have extracted from the calculated total energies the quantity @xmath160 where @xmath161/z + \widetilde{a}_1(n)$ ] , and have plotted it in the inset of as a function of the excess negative charge @xmath101 .
the dependence is linear to a remarkable extent ; for @xmath102 all three lines cross the energy axis at zero . combined with the results on the electron affinities , this indicates that the total energies have the following dependence on the excess number of electrons ( @xmath101 ) : @xmath162 where the dependence on the number of atoms in the cluster is not explicitly indicated .
this result is remarkable in its analogy with the classical image - charge result of van staveren _
indeed , the only difference amounts to the spill - out parameter @xmath136 and to the weak dependence on @xmath101 through @xmath124 .
this additional @xmath101-dependence becomes negligible already for the case of 30 sodium atoms . cluster .
a spherical jellium background was used .
the single - particle levels of the sic potential are also shown .
unlike the lda , this latter potential exhibits the correct asymptotic behavior .
the @xmath163 and @xmath164 electrons can be emitted spontaneously by tunneling through the coulombic barrier of the sic potential .
distances in units of the bohr radius , @xmath165 .
the specified single - particle levels are associated with the sic potential . ] for metastable multiply - charged cluster anions , electron emission ( autodetachment ) will occur via tunneling through a barrier ( shown in ) . however , to reliably estimate the electron emission , it is necessary to correct the lda effective potential for self - interaction effects .
we performed a self - interaction correction of the amaldi type @xcite for the hartree term and extended it to the exchange - correlation contribution to the total energy as follows : @xmath166= e_{\rm xc}^{\rm lda}[\rho]-n_e e_{\rm xc}^{\rm lda}[\rho / n_e]$ ] , where @xmath167 is the total number of electrons .
this self - interaction correction is akin to the orbitally - averaged - potential method @xcite . minimizing the sic energy functional for the parameters @xmath93 , @xmath94 , and @xmath95
, we obtained the effective sic potential for na@xmath168 shown in , which exhibits the physically correct asymptotic behavior @xcite . the spontaneous electron emission through the coulombic barrier is analogous to that occurring in proton radioactivity from neutron - deficient nuclei @xcite , as well as in alpha - particle decay .
the transition rate is @xmath169 , where @xmath170 is the attempt frequency and @xmath171 is the transmission coefficient calculated in the wkb method ( for details , cf . ref .
@xcite ) . for the @xmath163 electron in na@xmath168 ( cf . ) , we find @xmath172 and @xmath173 , yielding @xmath174 . for a cluster size closer to the drip line ( see ) , e.g. na@xmath175 , we find @xmath176 . finally , the exression in eq .
( [ etz ] ) for the total energy can be naturally extended to the case of multiply _ positively _ charged metal clusters by setting @xmath177 , with @xmath178 .
the ensuing equation retains the same dependence on the excess positive charge @xmath179 , but with the negative value of the first affinity , @xmath180 , replaced by the positive value of the first ionization potential , @xmath181 , a result that has been suggested from earlier measurements on multiply charged potassium cations @xcite .
naturally , the spill - out parameter @xmath182 assumes different values than in the case of the anionic clusters .
fullerenes and related carbon structures have been extensively investigated using _
ab initio _ density - functional - theory methods and self - consistent solutions of the kohn - sham ( ks ) equations @xcite . for metal clusters , replacing the ionic cores with a uniform jellium background was found to describe well their properties within the ks - dft method @xcite .
motivated by these results , several attempts to apply the jellium model in conjunction with dft to investigations of fullerenes have appeared recently @xcite .
our approach @xcite differs from the earlier ones in several aspects and , in particular , in the adaptation to the case of finite systems of the stabilized - jellium ( or structureless pseudopotential ) energy density functional ( see eq .
( [ epseu ] ) below and ref .
@xcite ) . an important shortcoming of the standard jellium approximation for fullerenes ( and other systems with high density , i.e. , small @xmath183 ) results from a well - known property of the jellium at high electronic densities , namely that the jellium is unstable and yields negative surface - energy contribution to the total energy @xcite , as well as unreliable values for the total energy .
these inadequacies of the standard jellium model can be rectified by pseudopotential corrections . a modified - jellium approach which incorporates such pseudopotential corrections and
is particularly suited for our purposes here , is the _ structureless pseudopotential _ model or _ stabilized jellium _
approximation developed in ref .
@xcite . in the stabilized jellium , the total energy @xmath184 , as a functional of the electron density @xmath185 , is given by the expression @xmath186 = e_{jell}[\rho,\rho_+ ] + \langle \delta \upsilon \rangle_{ws } \int \rho({\bf r } ) { \cal u}({\bf r } ) d{\bf r } -\widetilde{\varepsilon } \int \rho_+({\bf r } ) d{\bf r } , \label{epseu}\ ] ]
where by definition the function @xmath187 equals unity inside , but vanishes , outside the jellium volume .
@xmath188 is the density of the positive jellium background ( which for the case of c@xmath189 is taken as a spherical shell , of a certain width @xmath190 , centered at 6.7 @xmath137 ) .
@xmath184 in eq .
( [ epseu ] ) is the standard jellium - model total energy , @xmath191 , modified by two corrections .
the first correction adds the effect of an average ( i.e. , averaged over the volume of a wigner - seitz cell ) difference potential , @xmath192 , which acts on the electrons in addition to the standard jellium attraction and is due to the atomic pseudopotentials ( in this work , we use the ashcroft empty - core pseudopotential , specified by a core radius @xmath193 , as in ref .
the second correction subtracts from the jellium energy functional the spurious electrostatic self - repulsion of the positive background within each cell ; this term makes no contribution to the effective electronic potential . following ref .
@xcite , the bulk stability condition * ( * eq . ( 25 ) in ref . @xcite * ) * determines the value of the pseudopotential core radius @xmath193 , as a function of the bulk wigner - seitz radius @xmath183 .
consequently , the difference potential can be expressed solely as a function of @xmath183 as follows ( energies in @xmath194 , distances in @xmath137 ) : @xmath195 where @xmath196 is the per particle electron - gas correlation energy ( in our calculation , we use the gunnarsson - lundqvist exchange and correlation energy functionals ; see refs .
@xcite ) .
another required generalization consists in employing a parametrized electron - density profile that accounts for the hollow cage - structure of the fullerenes .
such a density profile is provided by the following adaptation of a generalization of an inverse thomas - fermi distribution , used earlier in the context of nuclear physics @xcite , i.e. , @xmath201 } { \cosh [ w_{i , o}/\alpha_{i , o } ] + \cosh [ ( r - r)/\alpha_{i , o } ] } \right)^{\gamma_{i , o } } , \label{rgram}\ ] ] where @xmath202 is the radius of the fullerene cage .
@xmath203 , @xmath94 , and @xmath95 are variables to be determined by the etf - dft minimization . for @xmath204 and large values of @xmath205 ,
expression ( [ rgram ] ) approaches the more familiar inverse thomas - fermi distribution , with @xmath203 the width , @xmath94 the diffuseness and @xmath95 the asymmetry of the profile around @xmath206 .
there are a total of six parameters to be determined , since the indices @xmath207 stand for the regions inside ( @xmath208 ) and outside ( @xmath209 ) the fullerene cage .
@xmath210 + 1)/\sinh [ w_{i , o}/\alpha_{i , o}]$ ] is a constant guaranteeing that the two parts of the curve join smoothly at @xmath211 .
the density profile in eq .
( [ rgram ] ) peaks at @xmath212 and then falls towards smaller values both inside and outside the cage ( see top panel of )
. molecule .
the wigner - seitz radius for the jellium bacground is 1.23 @xmath137 note the asymmetry of the potential about the minimum .
the associated difference potential @xmath213 .
+ top panel : solid line : radial density of the positive jellium background .
dashed line : etf electronic density .
note its asymmetry about the maximum .
thick solid line : the difference ( multiplied by 10 ) of electronic etf densities between c@xmath214 and c@xmath189 .
it illustrates that the excess charge accumulates in the outer perimeter of the total electronic density .
all densities are normalized to the density of the positive jellium background . ] to apply the scm to the present case , the potential @xmath70 in eq .
( [ tsh ] ) is replaced by the stabilized - jellium lda potential shown in .
after some rearrangements , the shell - corrected total energy @xmath215 $ ] in the stabilized - jellium case can be written in functional form as follows [ compare to eq .
( [ ensh ] ) , see also eq .
( [ enhar ] ) ] .
@xmath216= \sum_i \widetilde{\varepsilon}_i & - & \int \
! \left\ { \frac{1}{2 } \widetilde{v}_h({\bf r } ) + \widetilde{v}_{{xc}}({\bf r } ) \right\ } \widetilde{\rho}({\bf r } ) d\/{\bf r } \nonumber \\ ~ & + & \int \widetilde{{\cal e}}_{{xc } } [ \widetilde{\rho}({\bf r } ) ] d\/{\bf r } + e_i - \widetilde{\varepsilon } \int \rho_+({\bf r } ) d{\bf r } , \label{enshsj}\end{aligned}\ ] ] heretofore , the point - group icosahedral symmetry of c@xmath189 was not considered , since the molecule was treated as a spherically symmetric cage .
this is a reasonable zeroth - order approximation as noticed by several authors @xcite .
however , considerable improvement is achieved when the effects of the point - group icosahedral symmetry are considered as a next - order correction ( mainly the lifting of the angular momentum degeneracies ) .
the method of introducing the icosahedral splittings is that of the crystal field theory @xcite .
thus , we will use the fact that the bare electrostatic potential from the ionic cores , considered as point charges , acting upon an electron , obeys the well - known expansion theorem @xcite @xmath217 where the angular coefficients @xmath218 are given through the angular coordinates @xmath219 of the carbon atomic cores , namely , @xmath220 where @xmath221 denotes complex conjugation .
we take the radial parameters @xmath222 as constants , and determine their value by adjusting the icosahedral single - particle spectra @xmath223 to reproduce the pseudopotential calculation of ref .
@xcite , which are in good agreement with experimental data .
our spectra without and with icosahedral splitting are shown in ( a ) and ( b ) , respectively .
because of the spherical symmetry , they are characterized by the two principle quantum numbers @xmath224 and @xmath225 , where @xmath224 is the number of radial nodes and @xmath225 the angular momentum .
they are grouped in three bands labeled @xmath226 ( @xmath227 ) , @xmath228 ( @xmath229 ) , and @xmath182 ( @xmath230 ) .
each band starts with an @xmath231 level .
+ ( b ) the single - particle levels for c@xmath189 after the icosahedral splittings are added to the spectra of ( a ) .
the tenfold degenerate homo ( h@xmath232 ) and the sixfold degenerate lumo1 ( t@xmath233 ) and lumo2 ( t@xmath234 ) are denoted ; they originate from the spherical @xmath235 and @xmath236 ( t@xmath234 ) @xmath228 levels displayed in panel ( a ) . for the @xmath226 electrons ,
the icosahedral perturbation strongly splits the @xmath237 level of panel ( a ) .
there result five sublevels which straddle the @xmath226-electron gap as follows : two of them ( the eightfold degenerate g@xmath232 , and the tenfold degenerate h@xmath232 ) move down and are fully occupied resulting in a shell closure ( 180 @xmath226 electrons in total ) .
the remaining unoccupied levels , originating from the @xmath237 @xmath226 level , are sharply shifted upwards and acquire positive values . ]
we find that a close reproduction of the results of _ ab initio _ dft calculations @xcite is achieved when the wigner - seitz radius for the jellium background is @xmath238 @xmath137 the shell corrections , @xmath239 , including the icosahedral splittings are calculated using the icosahedral single - particle energies @xmath223 in eq .
( [ tsh ] ) .
the average quantities ( @xmath16 and @xmath240 ) are maintained as those specified through the etf variation with the spherically symmetric profile of eq .
( [ rgram ] ) .
this is because the first - order correction to the total energy ( resulting from the icosahedral perturbation ) vanishes , since the integral over the sphere of a spherical harmonic @xmath241 vanishes .
having specified the appropriate wigner - seitz radius @xmath183 and the parameters @xmath242 of the icosahedral crystal field through a comparison with the pseudopotential dft calculations for the neutral c@xmath189 , we calculate the total energies of the cationic and anionic species by allowing for a change in the total electronic charge , namely by imposing the constraint @xmath243 where @xmath244 is given by eq .
( [ rgram ] ) .
the shell - corrected and icosahedrally perturbed first and higher ionization potentials @xmath245 are defined as the difference of the ground - state shell - corrected total energies @xmath246 as follows : @xmath247 where @xmath248 is the number of electrons in the system and @xmath249 is the number of excess charges on the fullerenes ( for the excess charge , we will find convenient to use two different notations @xmath249 and @xmath179 related as @xmath250 .
a negative value of @xmath179 corresponds to positive excess charges ) .
@xmath251 denotes the total positive charge of the jellium background .
we have also calculated the corresponding average quantities @xmath254 and @xmath255 , which result from the etf variation with spherical symmetry ( that is without shell and icosahedral symmetry corrections ) .
their definition is the same as in eq .
( [ ips ] ) and eq .
( [ eas ] ) , but with the index @xmath256 replaced by a tilde and the removal of the index @xmath257 .
in our calculations of the charged fullerene molecule , the @xmath183 value and the icosahedral splitting parameters * ( * @xmath242 , see eq .
( [ cryf ] ) , and discussion below it * ) * were taken as those which were determined by our calculations of the neutral molecule , discussed in the previous section .
the parameters which specify the etf electronic density * ( * eq .
( [ rgram ] ) * ) * are optimized for the charged molecule , thus allowing for relaxation effects due to the excess charge .
this procedure is motivated by results of previous electronic structure calculations for c@xmath258 and c@xmath259 @xcite , which showed that the icosahedral spectrum of the neutral c@xmath189 shifts almost rigidly upon charging of the molecule . as a function of the excess charge @xmath179 ( @xmath261 corresponds to positive excess charge ) .
inset : the etf function @xmath262 ( open squares ) , and the shell - corrected function @xmath263 ( filled circles ) . for @xmath264 the two functions are almost identical . + ( b ) magnification of the appearance - energy curves for the region @xmath265 . filled circles : shell - corrected icosahedral values [ @xmath266 .
open squares : etf - dft values [ @xmath267 . ]
\(a ) shows that the variation of the total etf - dft energy difference ( appearance energies ) @xmath268 , as a function of excess charge @xmath179 ( @xmath269 ) , exhibits a parabolic behavior .
the inset in ( a ) exhibiting the quantity @xmath270 plotted versus @xmath179 ( open squares ) , shows a straight line which crosses the zero energy line at @xmath271 . as a result
the total etf - dft energy has the form , @xmath272 equation ( [ etfe ] ) indicates that fullerenes behave on the average like a capacitor having a capacitance @xmath273 * ( * the second term on the rhs of eq .
( [ etfe ] ) corresponds to the charging energy of a classical capacitor , corrected for the self - interaction of the excess charge @xcite*)*. we remark that regarding the system as a classical conductor , where the excess charge accumulates on the outer surface , yields a value of @xmath274 @xmath137 ( that is the outer radius of the jellium shell ) .
naturally , the etf calculated value for @xmath273 is somewhat larger because of the quantal spill - out of the electronic charge density . indeed ,
from the slope of @xmath262 we determine @xcite @xmath275 @xmath276 .
a similar plot of the shell - corrected and icosahedrally modified energy differences @xmath277 is shown in ( b ) ( in the range @xmath265 , filled circles ) . the function @xmath263 , defined as in eq .
( [ gsm ] ) but with the shell - corrected quantities ( @xmath278 and @xmath279 ) , is included in the inset to ( a ) ( filled circles ) .
the shift discernible between @xmath280 and @xmath281 is approximately 1.7 @xmath141 , and originates from the difference of shell effects on the ips and eas ( see ) .
the segments of the curve @xmath263 in the inset of ( a ) , corresponding to positively ( @xmath282 ) and negatively ( @xmath283 ) charged states , are again well approximated by straight lines , whose slope is close to that found for @xmath262 .
consequently , we may approximate the charging energy , including shell - effects , as follows , @xmath284 for _ negatively _ charged states , and @xmath285 for _ positively _ charged states .
note that without shell - corrections ( i.e. , etf only ) @xmath286 , because of the symmetry of eq .
( [ etfe ] ) with respect to @xmath179 , while the shell - corrected quantities are related as @xmath287 , where the shell correction is @xmath288 ( from , @xmath289 ) .
equations ( [ eshneg ] ) and ( [ eshpos ] ) can be used to provide simple analytical approximations for the higher ips and eas .
explicitly written , @xmath291 and @xmath292 .
such expressions have been used previously @xcite with an assumed value for @xmath293 ( i.e. , the radius of the c@xmath189 molecule , as determined by the distance of carbon nuclei from the center of the molecule ) , which is appreciably smaller than the value obtained by us ( @xmath294 see above ) via a microscopic calculation .
consequently , using the above expression with our calculated value for @xmath295 ( see ) , we obtain an approximate value of @xmath296 ( compared to the microscopically calculated value of @xmath297 given in , and @xmath298 obtained by ref .
@xcite ) indicating metastability of c@xmath299 while employing an experimental value for @xmath300 , a value of @xmath301 was calculated in ref .
@xcite . concerning the cations , our expression ( [ eshpos ] ) with a calculated @xmath302 ( see ) and
@xmath303 yields approximate values 18.5 @xmath141 and 31.5 @xmath141 for the appearance energies of c@xmath304 and c@xmath305 ( compared to the microscopic calculated values of 17.71 @xmath141 and 30.99 @xmath141 , respectively , extracted from , and 18.6 @xmath141 for the former obtained in ref.@xcite ) . employing an experimental value for @xmath306 , corresponding values of 19.20 @xmath141 and 34.96 @xmath141
were calculated in ref .
as discussed in ref.@xcite , these last values are rather high , and the origin of the discrepancy may be traced to the small value of the capacitance which was used in obtaining these estimates in ref .
@xcite .
a negative value of the second affinity indicates that c@xmath299 is unstable against electron autodetachment . in this context
, we note that the doubly negatively charged molecule c@xmath299 has been observed in the gas phase and is believed to be a long - lived metastable species @xcite . indeed , as we discuss in the next section , the small dft values of @xmath307 found by us and by ref .
@xcite yield lifetimes which are much longer than those estimated by a pseudopotential - like hartree - fock model calculation @xcite , where a value of @xmath308 1 @xmath309 was estimated .
the second and higher electron affinities of c@xmath189 were found to be negative , which implies that the anions c@xmath311 with @xmath312 are not stable species , and can lower their energy by emitting an electron . however ,
unless the number of excess electrons is large enough , the emission of an excess electron involves tunneling through a barrier .
consequently , the moderately charged anionic fullerenes can be described as metastable species possessing a decay lifetime . to calculate the lifetime for electron autodetachmant , it is necessary to determine the proper potential that the emitted electron sees as it leaves the molecule .
the process is analogous to alpha - particle radioactivity of atomic nuclei .
the emitted electron will have a final kinetic energy equal to the negative of the corresponding higher ea .
we estimate the lifetime of the decay process by using the wkb method , in the spirit of the theory of alpha - particle radioactivity , which has established that the main factor in estimating lifetimes is the relation of the kinetic energy of the emitted particle to the coulombic tail , and not the details of the many - body problem in the immediate vicinity of the parent nucleus .
essential in this approach is the determination of an appropriate single - particle potential that describes the transmission barrier .
it is well known that the ( dft ) lda potential posseses the wrong tail , since it allows for the electron to spuriously interact with itself .
a more appropriate potential would be one produced by the self - interaction correction method of ref .
this potential has the correct coulombic tail , but in the case of the fullerenes presents another drawback , namely koopman s theorem is not satisfied to an extent adequate for calculating lifetimes @xcite . in this context
, we note that koopman s theorem is known to be poorly satisfied for the case of fullerenes even in hartree - fock calculations @xcite .
therefore , the homo corresponding to the emitted electron , calculated as described above , can not be used in the wkb tunneling calculation . since the final energy of the ejected electron equals the negative of the value of the electron affinity , we seek a potential that , together with the icosahedral perturbation , yields a homo level in c@xmath311 with energy @xmath313 .
we construct this potential through a self - interaction correction to the lda potential as follows , @xmath314 - v_{h}[\frac{\widetilde{\rho}}{n_e } ] - v_{{xc}}[\xi\frac{\widetilde { \rho}}{n_e } ] , \label{sicxi}\ ] ] where the parameter @xmath315 is adjusted so that the homo level of c@xmath311 equals @xmath313 . in the above expression ,
the second term on the rhs is an average self - interaction hartree correction which ensures a proper long - range behavior of the potential ( i.e. , correct coulomb tail ) , and the third term is a correction to the short - range exchange - correlation .
( a ) and c@xmath316 ( b ) .
dashed lines correspond to barriers due solely to coulombic repulsion and solid lines to total barriers after adding the centrifugal components .
the thick horizontal solid lines correspond to the negative of the associated electron affinities @xmath307 ( a ) and @xmath317 ( b ) . in the case of c@xmath299 [ panel ( a ) ] ,
the horizontal solid line at @xmath318 crosses the total barrier at an inside point @xmath319 and again at a distance very far from the center of the fullerene molecule , namely at an outer point @xmath320 this large value of @xmath321 , combined with the large centrifugal barrier , yields a macroscopic lifetime for the metastable c@xmath299 ( see text for details ) . ] for the cases of c@xmath299 and c@xmath316 such potentials are plotted in .
we observe that they have the correct coulombic tail , namely a tail corresponding to one electron for c@xmath299 and to two electrons for c@xmath316 .
the actual barrier , however , through which the electron tunnels is the sum of the coulombic barrier plus the contribution of the centrifugal barrier . as seen from , the latter is significant , since the homo in the fullerenes possesses a rather high angular momentum ( @xmath235 ) , while being confined in a small volume . using the wkb approximation @xcite , we estimate for c@xmath299 a macroscopic half - life of @xmath322 @xmath323 , while for c@xmath316 we estimate a very short half - life of @xmath324 .
both these estimates are in correspondence with observations .
indeed , c@xmath316 has not been observed as a free molecule , while the free c@xmath299 has been observed to be long lived @xcite and was detected even 5 @xmath325 after its production through laser vaporization @xcite .
we note that the wkb lifetimes calculated for tunneling through coulombic barriers are very sensitive to the final energy of the emitted particle and can vary by many orders of magnitude as a result of small changes in this energy , a feature well known from the alpha radioctivity of nuclei @xcite .
since the second electron affinity of c@xmath189 is small , effects due to geometrical relaxation and spin polarization can influence its value and , consequently , the estimated lifetime .
nevertheless , as shown in ref .
@xcite , inclusion of such corrections yields again a negative second affinity , but of somewhat smaller magnitude , resulting in an even longer lifetime ( the sign conventions in ref .
@xcite are the opposite of ours ) . furthermore , as discussed in ref .
@xcite , the stabilization effect of the jahn - teller relaxation for the singly - charged ion is only of the order of 0.03 0.05 @xmath141 .
since this effect is expected to be largest for singly - charged species , c@xmath299 is not expected to be influenced by it @xcite . on the other hand , generalized exchange - correlation functionals with gradient corrections
yield slightly larger values for the second electron affinity .
for example , using exchange - correlation gradient corrections , ref .
@xcite found @xmath326 , which is higher ( in absolute magnitude ) than the value obtained without such corrections .
this value of @xmath327 leads to a much smaller lifetime than the several million of years that correspond to the value of @xmath328 calculated by us . indeed , using the barrier displayed in ( a ) , we estimate a lifetime for c@xmath299 of approx .
0.37 @xmath329 , when @xmath326 .
we stress , however , that even this lower - limit value still corresponds to macroscopic times and is 5 orders of magnitude larger than the estimate of ref .
@xcite , which found a lifetime of 1 @xmath309 for @xmath326 , since it omitted the large centrifugal barrier . indeed ,
when we omit the centrifugal barrier , we find a lifetime estimate of 1.4 @xmath309 , when @xmath326 . in this section ,
we show that certain aspects of the mechanical response ( i.e. , elongation force ) and electronic transport ( e.g. , quantized conductance ) in metallic nanowires can be analyzed using the dft shell correction method , developed and applied previously in studies of metal clusters ( see and ) .
specifically , we show that in a jellium - modelled , volume - conserving nanowire , variations of the total energy ( particularly terms associated with electronic subband corrections ) upon elongation of the wire lead to _ self - selection _ of a sequence of stable `` magic '' wire configurations ( mwc s , specified in our model by a sequence of the wire s radii ) , with the force required to elongate the wire from one configuration to the next exhibiting an oscillatory behavior .
moreover , we show that due to the quantized nature of electronic states in such wires , the electronic conductance varies in a quantized step - wise manner ( in units of the conductance quantum @xmath330 ) , correlated with the transitions between mwc s and the above - mentioned force oscillations . prior to introducing the model ,
it is appropriate to briefly review certain previous theoretical and experimental investigations , which form the background and motivation for this study of nanowires .
atomistic descriptions , based on realistic interatomic interactions , and/or first - principles modelling and simulations played an essential role in discovering the formation of nanowires , and in predicting and elucidating the microscopic mechanisms underlying their mechanical , spectral , electronic and transport properties .
formation and mechanical properties of interfacial junctions ( in the form of crystalline nanowires ) have been predicted through early molecular - dynamics simulations @xcite , where the materials ( gold ) were modelled using semiempirical embedded - atom potentials . in these studies
it has been shown that separation of the contact between materials leads to generation of a connective junction which elongates and narrows through a sequence of structural instabilities ; at the early stages , elongation of the junction involves multiple slip events , while at the later stages , when the lateral dimension of the wire necks down to a diameter of about 15 , further elongation involves a succession of stress accumulation and fast relief stages associated with a sequence of order - disorder structural transformations localized to the neck region @xcite .
these structural evolution patterns have been shown through the simulations to be portrayed in oscillations of the force required to elongate the wire , with a period approximately equal to the interlayer spacing .
in addition , the `` sawtoothed '' character of the predicted force oscillations * ( * see fig .
3(b ) in ref . @xcite and fig . 3 in ref .
@xcite * ) * reflects the stress accumulation and relief stages of the elongation mechanism .
moreover , the critical resolved yield stress of gold nanowires has been predicted @xcite to be @xmath308 4gpa , which is over an order of magnitude larger than that of the bulk , and is comparable to the theoretical value for au ( 1.5 gpa ) in the absence of dislocations .
these predictions , as well as anticipated electronic conductance properties @xcite , have been corroborated in a number of experiments using scanning tunneling and force microscopy @xcite , break junctions @xcite , and pin - plate techniques @xcite at ambient environments , as well as under ultrahigh vacuum and/or cryogenic conditions .
particularly , pertinent to this section are experimental observations of the oscillatory behavior of the elongation forces and the correlations between the changes in the conductance and the force oscillations ; see especially the simultaneous measurements of force and conductance in gold nanowires in ref .
@xcite , where in addition the predicted `` ideal '' value of the critical yield stress has also been measured ( see also ref .
@xcite ) .
the jellium - based model introduced in this paper , which by construction is devoid of atomic crystallographic structure , does not address issues pertaining to nanowire formation methods , atomistic configurations , and mechnanical response modes * ( * e.g . , plastic deformation mechanisms , interplanar slip , ordering and disordering mechanisms ( see detailed descriptions in refs .
@xcite and @xcite , and a discussion of conductance dips in ref .
@xcite ) , defects , mechanichal reversibility @xcite , and roughening of the wires morphology during elongation @xcite * ) * , nor does it consider the effects of the above on the electron spectrum , transport properties , and dynamics @xcite
. nevertheless , as shown below , the model offers a useful framework for linking investigations of solid - state structures of reduced dimensions ( e.g. , nanowires ) with methodologies developed in cluster physics , as well as highlighting certain nanowire phenomena of mesoscopic origins and their analogies to clusters .
[ [ the - jellium - model - for - metallic - nanowires - theoretical - method - and - results ] ] the jellium model for metallic nanowires : theoretical method and results + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + consider a cylindrical jellium wire of length @xmath331 , having a positive background with a circular cross section of constant radius @xmath332 @xcite .
for simplicity , we restrict ourselves here to this symmetry of the wire cross section . variations in the shape of the nanowire cross section serve to affect the degeneracies of the electronic spectrum @xcite without affecting our general conclusions .
we also do not include here variations of the wire s shape along its axis .
adiabatic variation of the wire s axial shape introduces a certain amount of smearing of the conductance steps through tunnelling , depending on the axial radius of curvature of the wire @xcite .
both the cross - sectional and axial shape of the wire can be included in our model in a rather straightforward manner . as elaborated in , the principal idea of
the scm is the separation of the total dft energy @xmath333 of the nanowire as @xmath334 where @xmath335 varies smoothly as a function of the radius @xmath5 of the wire ( instead of the number of electrons @xmath6 used in ) , and @xmath336 is the shell - correction term arising from the discrete quantized nature of the electronic levels . again , as elaborated in , the smooth contribution in eq .
( [ etscm ] ) is identified with @xmath337 $ ] .
the trial radial lateral density @xmath338 is given by eq .
( [ rho ] ) , and the constant @xmath339 at a given radius @xmath5 is obtained under the normalization condition ( charge neutrality ) @xmath340 , where @xmath341 is the linear positive background density . using the optimized @xmath16 ,
one solves for the eigenvalues @xmath342 of the hamiltonian @xmath343 $ ] , and the shell correction is given by @xmath344 - e_{etf}[\widetilde{\rho } ] \nonumber \\ & = & \sum_{i=1}^{\mathrm{occ } } \widetilde{\epsilon}_i - \int \widetilde{\rho } ( { \bf r } ) v_{etf}[\widetilde{\rho } ( { \bf r } ) ] d{\bf r } - t_{etf}[\widetilde{\rho } ] , \label{shcor}\end{aligned}\ ] ] where the summation extends over occupied levels . here
the dependence of all quantities on the pertinent size variable ( i.e. , the radius of the wire @xmath5 ) is not shown explicitly . additionally , the index @xmath25 can be both discrete and continuous , and in the latter case the summation is replaced by an integral ( see below ) .
potential for a sodium wire with a uniform jellium background of radius @xmath345 a.u . , plotted versus the transverse radial distance from the center of the wire , along with the locations of the bottoms of the subbands ( namely the transverse eigenvalues @xmath346 ; @xmath347 is the number of nodes in the radial direction plus one , and @xmath348 is the azimuthal quantum number of the angular momentum ) .
the fermi level is denoted by a dashed line .
top panel : the jellium background volume density ( dashed line ) and the electronic volume density @xmath338 ( solid line , exhibiting a characteristic spillout ) normalized to bulk values are shown . ] following the above procedure with a uniform background density of sodium ( @xmath349 a.u . ) , a typical potential @xmath350 for @xmath345 a.u . , where @xmath351 is the radial coordinate in the transverse plane , is shown in , along with the transverse eigenvalues @xmath346 and the fermi level ; to simplify the calculations of the electronic spectrum , we have assumed ( as noted above ) @xmath332 , which allows us to express the subband electronic spectrum as @xmath352 where @xmath353 is the electron wave number along the axis of the wire ( @xmath179 ) . [ shown in ( a ) ] and of the conductance @xmath354 ( shown in ( b ) in units of @xmath355 ) , plotted versus the radius r , for a sodium nanowire .
note the coincidence of the cusps in @xmath356 with the step - rises of the conductance .
the heights of the steps in @xmath354 reflect the subband degeneracies due to the circular shape of the wire s cross section . ]
as indicated earlier , taking the wire to be charge neutral , the electronic linear density , @xmath357(@xmath5 ) , must equal the linear positive background density , @xmath358 .
the chemical potential ( at @xmath359 the fermi energy @xmath356 ) for a wire of radius @xmath5 is determined by setting the expression for the electronic linear density derived from the subband spectra equal to @xmath358 , i.e. , @xmath360 } = \rho^{(+)}_l ( r ) , \label{linden}\ ] ] where the factor of 2 on the left is due to the spin degeneracy .
the summand defines the fermi wave vector for each subband , @xmath361 .
the resulting variation of @xmath362 versus @xmath5 is displayed in ( a ) , showing cusps for values of the radius where a new subband drops below the fermi level as @xmath5 increases ( or conversely as a subband moves above the fermi level as @xmath5 decreases upon elongation of the wire ) . using the landauer expression for the conductance @xmath354 in the limit of no mode mixing and assuming unit transmission coefficients , @xmath363 $ ] , where @xmath364 is the heaviside step function .
the conductance of the nanowire , shown in ( b ) , exhibits quantized step - wise behavior , with the step - rises coinciding with the locations of the cusps in @xmath362 , and the height sequence of the steps is 1@xmath365 , 2@xmath365 , 2@xmath365 , 1@xmath365 , ... , reflecting the circular symmetry of the cylindrical wires cross sections @xcite , as observed for sodium nanowires @xcite . solving for @xmath362
[ see eq .
( [ linden ] ) ] , the expression for the sum on the right - hand - side of eq .
( [ shcor ] ) can be written as @xmath366 \sqrt { \frac{2m}{\hbar^2 } [ \epsilon_f(r ) - \widetilde{\epsilon}_{nm}(r ) ] } , \label{sumi}\end{aligned}\ ] ] which allows one to evaluate @xmath79 [ eq .
( [ shcor ] ) ] for each wire radius @xmath5 . since the expression in eq .
( [ sumi ] ) gives the energy per unit length , we also calculate @xmath74 , @xmath1 , and the volume integral in the second line of eq .
( [ shcor ] ) for cylindrical volumes of unit height . to convert to energies per unit volume [ denoted as @xmath367 , @xmath368 , and @xmath369
all energies are further divided by the wire s cross - sectional area , @xmath370 . the smooth contribution and the shell correction to the wire s energy
are shown respectively in ( a ) and ( b ) .
the smooth contribution decreases slowly towards the bulk value ( @xmath322.25 ev per atom @xcite ) . on the other hand ,
the shell corrections are much smaller in magnitude and exhibit an oscillatory behavior .
this oscillatory behavior remains visible in the total energy [ ( c ) ] with the local energy minima occurring for values @xmath371 corresponding to conductance plateaus .
the sequence of @xmath371 values defines the mwc s , that is a sequence of wire configurations of enhanced stability .
ev / a.u.@xmath372 ) , plotted versus the radius of the wire ( in a.u . ) .
note the smaller magnitude of the shell corrections relative to the smooth contribution .
( d - e ) : the smooth contribution ( d ) to the total force and the total force ( e ) , plotted in units of nn versus the wire s radius . in ( e ) , the zeroes of the force to the left of the force maxima occur at radii corresponding to the local minima of the energy of the wire ( c ) . in ( f ) , we reproduce the conductance of the wire ( in units of @xmath330 ) , plotted versus r. interestingly , calculations of the conductance for the mwc s ( i.e. , the wire radii corresponding to the locations of the step - rises ) through the sharvin - weyl formula , @xcite corrected for the finite height of the confining potential @xcite ( see lower panel of ) , namely @xmath373 ) where @xmath3 and @xmath171 are the area and perimeter of the wire s cross section and @xmath374 is the fermi wavelength ( @xmath375 a.u . for na ) with @xmath376 ( see ref .
@xcite ) , yield results which approximate well the conductance values ( i.e. , the values at the bottom of the step - rises ) shown in ( f ) . ] from the expressions for the total energy of the wire [ i.e. , @xmath377 , where @xmath378 is the volume of the wire ] and the smooth and shell ( subband ) contributions to it , we can calculate the `` elongation force '' ( ef ) , @xmath379}{dl } = - \omega \left\ { \frac{d \widetilde{\varepsilon } ( r)}{dl } + \frac { d [ \delta \varepsilon_{sh } ( r ) ] } { dl } \right\ } \nonumber\\ & \equiv & \widetilde{f}(r ) + \delta f_{sh}(r ) .
\label{force}\end{aligned}\ ] ] using the volume conservation , i.e. , @xmath380 , these forces can be written as @xmath381 , @xmath382 , and @xmath383/dr$ ] .
@xmath384 and @xmath385 are shown in ( d , e ) .
the oscillations in the force resulting from the shell - correction contributions dominate . in all cases ,
the radii corresponding to zeroes of the force situated on the left of the force maxima coincide with the minima in the potential energy curve of the wire , corresponding to the mwc s .
consequently , these forces may be interpreted as guiding the self - evolution of the wire toward the mwc s .
also , all the local maxima in the force occur at the locations of step - rises in the conductance [ reproduced in ( f ) ] , signifying the sequential decrease in the number of subbands below the fermi level ( conducting channels ) as the wire narrows ( i.e. , as it is being elongated ) .
finally the magnitude of the total forces is comparable to the measured ones ( i.e. , in the nn range ) . while it was understood rather early that the total energy of nuclei can be decomposed into an oscillatory part and one that shows a slow `` smooth '' variation as a function of size , strutinsky s seminal contribution @xcite was to calculate the two parts from different nuclear models : the former from the nuclear shell model and the latter from the liquid drop model . in particular , the calculation of the oscillatory part ( shell correction term ) was enabled by employing an averaging method that smeared the single particle spectrum associated with a nuclear model potential .
a semiempirical shell - correction method ( referred to as se - scm ) for metal clusters , that was developed in close analogy to the original phenomenological strutinsky approach , was presented in the appendix , along with some applications to triaxial deformations and fission barriers of metal clusters .
this chapter reviewed primarily the motivation and theory of a microscopic shell correction method based on density functional theory ( often referred to as dft - scm and originally introduced in ref .
@xcite ) . in developing the dft - scm , we have used for the shell correction term ( arising from quantum interference effects ) a derivation that differs from the strutinsky methodology @xcite .
instead , we have shown @xcite that the shell correction term can be introduced through a kinetic - energy - type density functional [ see eq .
( [ tsh ] ) and eq .
( [ densh ] ) ] .
the dft - scm is computationally advantageous , since it bypasses the self - consistent iteration cycle of the more familiar ks - dft .
indeed , the dft - scm energy functional depends only on the single - particle density , and thus it belongs to the class of orbital - free dft methods .
compared to previous of - dft approaches , the dft - scm represents an improvement in accuracy .
as mentioned above already [ see , e.g. , ] , rather than proceed with the microscopic route , strutinsky proposed a method for separation of the total energy into smooth and shell - correction terms [ see eq .
( [ enshsm ] ) ] based on an averaging procedure .
accordingly , a smooth part , @xmath386 , is extracted out of the sum of the single - particle energies @xmath387 [ see eq .
( [ enstr ] ) , or equivalently eq . ( [ enhar ] ) with @xmath57 replaced by @xmath16 and @xmath59 by @xmath26 ] by averaging them through an appropriate procedure .
usually , but not necessarily , one replaces the delta functions in the single - particle density of states by gaussians or other appropriate weighting functions . as a result
, each single - particle level is assigned an averaging occupation number @xmath389 , and the smooth part @xmath386 is formally written as @xmath390 the strutinsky prescription ( [ struav ] ) has the practical advantage of using only the single - particle energies @xmath26 , and not the smooth density @xmath16 . taking advantage of this ,
the single - particle energies can be taken as those of an external potential that empirically approximates the self - consistent potential of a finite system . in the nuclear case
, an anisotropic three - dimensional harmonic oscillator has been used successfully to describe the shell - corrections in deformed nuclei .
the single - particle smooth part , @xmath386 , however , is only one component of the total smooth contribution , @xmath392 $ ] ( @xmath393 in the hartree - fock energy considered by strutinsky ) . indeed as can be seen from eq .
( [ enstr ] ) [ or equivalently eq . ( [ enhar ] ) ] , @xmath394 . \label{ehfsmsh}\ ] ]
strutinsky did not address the question of how to calculate microscopically the smooth part @xmath13 ( which necessarily entails specifying the smooth density @xmath29 ) .
instead he circumvented this question by substituting for @xmath13 the empirical energies , @xmath395 , of the nuclear liquid drop model , namely he suggested that @xmath396 in applications of eq .
( [ enhfld ] ) , the single - particle energies involved in the averaging [ see eqs .
( [ espoc ] ) and ( [ struav ] ) ] are commonly obtained as solutions of a schrdinger equation with phenomenological one - body potentials .
this last approximation has been very successful in describing fission barriers and properties of strongly deformed nuclei using harmonic oscillator or wood - saxon empirical potentials . in the following ( )
, we describe the adaptation of the se - scm approach to condensed - matter finite systems , and in particular to triaxially deformed metal clusters . moreover
( ) , we will present several figures illustrating applications of the se - scm to investigations of the effects of triaxial shape - deformations on the properties of metal clusters @xcite and to studies of large - scale deformations and barriers in fission of charged metal clusters @xcite .
we note that the se - scm has been extended to incorporate electronic entropy effects at finite temperatures .
this latter extension , referred to as finite - temperature ( ft)-se - scm is not described here , but its theory can be found in ref .
@xcite .
we mention that , in addition , strutinsky - type calculations using phenomenological potentials have been reported for the case of neutral sodium clusters assuming axial symmetry in refs .
@xcite , and for the case of fission in ref .
@xcite . for neutral clusters ,
the liquid - drop model @xcite ( ldm ) expresses the _ smooth _ part , @xmath13 , of the total energy as the sum of three contributions , namely a volume , a surface , and a curvature term , i.e. , @xmath397 where @xmath398 is the volume element and @xmath399 is the surface differential element .
the local curvature @xmath400 is defined by the expression @xmath401 , where @xmath402 and @xmath403 are the two principal radii of curvature at a local point on the surface of the jellium droplet which models the cluster .
the corresponding coefficients can be determined by fitting the extended thomas - fermi ( etf)-dft total energy @xmath404 $ ] ( see ) for spherical shapes to the following parametrized expression as a function of the number , @xmath6 , of atoms in the cluster @xcite , @xmath405 in the case of ellipsoidal shapes the areal integral and the integrated curvature can be expressed in closed analytical form with the help of the incomplete elliptic integrals @xmath409 and @xmath410 of the first and second kind @xcite , respectively . before writing the formulas , we need to introduce some notations . volume conservation must be employed , namely @xmath411 where @xmath412 is the radius of a sphere with the same volume ( @xmath413 is taken to be the radius of the positive jellium assuming spherical symmetry ) , and @xmath414 , etc ... , are the dimensionless semi - axes .
the eccentricities are defined through the dimensionless semi - axes as follows @xmath415 the semi - axes are chosen so that @xmath416 with the notation @xmath417 , @xmath418 , and @xmath419 , the relative ( with respect to the spherical shape ) surface and curvature energies are given @xcite by @xmath420 \label{esurf}\ ] ] and @xmath421 .
\label{ecurv}\ ] ] the change in the smooth part of the cluster total energy due to the excess charge @xmath422 was already discussed by us for spherical clusters in the previous section .
the result may be summarized as @xmath423 where the upper and lower signs correspond to negatively and positively charged states , respectively , @xmath120 is the work function of the metal , @xmath412 is the radius of the positive jellium assuming spherical symmetry , and @xmath182 is a spillout - type parameter . to generalize the above results to an ellipsoidal shape
, @xmath424 @xmath425 @xmath426 , which is the value of the potential on the surface of a spherical conductor , needs to be replaced by the corresponding expression for the potential on the surface of a conducting ellipsoid .
the final result , normalized to the spherical shape , is given by the expression @xmath427 where the @xmath428 sign in front of @xmath429 corresponds to negatively and positively charged clusters , respectively .
, plotted versus cluster size .
the error bars indicate the statistical uncertainty .
( b ) theoretical ft - se - scm @xcite second electron affinities @xmath106 for ag@xmath430 clusters at @xmath431 k. ldm results are depicted by the dashed line .
the figure was reproduced from ref .
, title="fig:",width=302 ] + a natural choice for an external potential to be used for calculating shell corrections with the strutinsky method is an anisotropic , three - dimensional oscillator with an @xmath432 term for lifting the harmonic oscillator degeneracies @xcite .
such an oscillator model for approximating the total energies of metal clusters , but without separating them into a smooth and a shell - correction part in the spirit of strutinsky s approach , has been used @xcite with some success for calculating relative energy surfaces and deformation shapes of metal clusters . however
, this simple harmonic oscillator model has serious limitations , since i ) the total energies are calculated by the expression @xmath433 , and thus do not compare with the total energies obtained from the ks - dft approach , ii ) the model can not be extended to the case of charged ( cationic or anionic ) clusters .
thus absolute ionization potentials , electron affinities , and fission energetics can not be calculated in this model .
alternatively , in our approach , we are making only a limited use of the external oscillator potential in calculating a modified strutinsky shell correction .
total energies are evaluated by adding this shell correction to the smooth ldm energies .
in particular , a modified nilsson hamiltonian appropriate for metal clusters @xcite is given by @xmath434 where @xmath435 is the hamiltonian for a three - dimensional anisotropic oscillator , namely @xmath436 @xmath437 in eq .
( [ hn ] ) is a dimensionless parameter , which for occupied states may depend on the principal quantum number @xmath438 of the spherical - oscillator major shell associated with a given level @xmath439 of the hamiltonian @xmath435 ( for clusters comprising up to 100 valence electrons , only a weak dependence on @xmath347 is found , see table i in ref .
@xmath437 vanishes for values of @xmath347 higher than the corresponding value of the last partially ( or fully ) filled major shell in the spherical limit .
@xmath440 is a `` stretched '' angular momentum which scales to the ellipsoidal shape and is defined as follows , @xmath441 ( with similarly obtained expressions for @xmath442 and @xmath443 via a cyclic permutation of indices ) where the stretched position and momentum coordinates are defined via the corresponding natural coordinates , @xmath444 and @xmath445 , as follows , @xmath446 @xmath447 clusters at three temperatures , @xmath448 , 300 , and 500 k. solid dots : theoretical ft - se - scm @xcite results .
open squares : experimental measurements @xcite .
the best agreement between theory and experiment happens for @xmath431 k ( room temperature ) , indicating the importance of the electronic entropy in quenching the shell effects . , title="fig:",width=302 ] + the stretched @xmath432 is not a properly defined angular - momentum operator , but has the advantageous property that it does not mix deformed states which correspond to sherical major shells with different principal quantum number @xmath438 ( see , the appendix in ref .
@xcite for the expression of the matrix elements of @xmath432 ) . the subtraction of the term @xmath449 , where @xmath450 denotes the expectation value taken over the @xmath451-major shell in spherical symmetry , guaranties that the average separation between major oscillator shells is not affected as a result of the lifting of the degeneracy .
the oscillator frequencies can be related to the principal semi - axes @xmath452 ,
@xmath453 , and @xmath454 [ see eq .
( [ volcons ] ) ] via the volume - conservation constraint and the requirement that the surface of the cluster is an equipotential one , namely @xmath455 where the frequency @xmath456 for the spherical shape ( with radius @xmath412 ) was taken according to ref .
@xcite to be @xmath457^{-2}. \label{omeg0}\ ] ] since in this paper we consider solely monovalent elements , @xmath6 in eq.([omeg0 ] ) is the number of atoms for the family of clusters m@xmath458 , @xmath183 is the wigner - seitz radius expressed in atomic units , and @xmath459 denotes the electronic spillout for the neutral cluster according to ref .
@xcite .
[ [ averaging - of - single - particle - spectra - and - semi - empirical - shell - correction ] ] averaging of single - particle spectra and semi - empirical shell correction ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ usually @xmath386 [ see eqs . ( [ espoc ] ) and ( [ struav ] ) ] is calculated numerically @xcite . however , a variation of the numerical strutinsky averaging method consists in using the semiclassical partition function and in expanding it in powers of @xmath460 . with this method , for the case of an anisotropic , fully triaxial oscillator
, one finds @xcite an analytical result , namely @xcite @xmath461 where @xmath248 denotes the number of delocalized valence electrons in the cluster . in the present work ,
expression ( [ harmav ] ) ( as modified below ) will be substituted for the average part @xmath386 in eq .
( [ struav ] ) , while the sum @xmath462 will be calculated numerically by specifying the occupied single - particle states of the modified nilsson oscillator represented by the hamiltonian ( [ hn ] ) . in the case of an isotropic oscillator , not only
the smooth contribution , @xmath463 , but also the strutinsky shell correction ( [ struav ] ) can be specified analytically , @xcite with the result @xmath464 where @xmath249 is the fractional filling of the highest partially filled harmonic oscillator shell . for a filled shell ( @xmath465 ) , @xmath466 , instead of the essentially vanishing value as in the case of the etf - dft defined shell correction ( cf .
fig . 1 of ref
. @xcite ) . to adjust for this discrepancy , we add @xmath467 to @xmath468 calculated through eq .
( [ struav ] ) for the case of open - shell , as well as closed - shell clusters .
@xmath469 na@xmath470na@xmath471 .
the final configuration of na@xmath471 is spherical .
for the heavier fragment na@xmath472 , we present results associated with three different final shape configurations , namely , oblate [ ( o , s ) ; left ] , spherical [ ( s , s ) ; middle ] , and prolate [ ( p , s ) ; right ] .
the ratio of shorter over longer axis is 0.555 for the oblate case and 0.75 for the prolate case .
+ bottom panel : ldm energy ( surface plus coulomb , dashed curve ) and total potential energy ( ldm plus shell corrections , solid curve ) as a function of fragment separation @xmath473 .
the empty vertical arrow marks the scission point .
the zero of energy is taken at @xmath474 . a number ( @xmath321.58 ev or @xmath320.98 ev ) , or a horizontal solid arrow , denotes the corresponding dissociation energy .
+ middle panel : shell - correction contribution ( solid curve ) , surface contribution ( upper dashed curve ) , and coulomb contribution ( lower dashed curve ) to the total energy , as a function of fragment separation @xmath473 .
+ top panel : single - particle spectra as a function of fragment separation @xmath473 .
the occupied ( fully or partially ) levels are denoted with solid lines .
the unoccupied levels are denoted with dashed lines .
+ on top of the figure , four snapshots of the evolving cluster shapes are displayed .
the solid vertical arrows mark the corresponding fragment separations .
observe that the doorway molecular configurations correspond to the second snapshot from the left .
notice the change in energy scale for the middle and bottom panels , as one passes from ( o , s ) to ( s , s ) and ( p , s ) final configurations . ] 1 .
parametrize results of etf - dft calculations for spherical neutral jellia according to eq .
( [ npar ] ) .
2 . use above parametrization ( assuming that parameters per differential element of volume , surface , and integrated curvature are shape independent ) in eq .
( [ lqd ] ) to calculate the liquid - drop energy associated with neutral clusters , and then add to it the charging energy according to eq .
( [ ecoul ] ) to determine the total ldm energy @xmath13 .
3 . use equations ( [ hn ] ) and ( [ h0 ] ) for a given deformation [ i.e. , @xmath452 , @xmath453 , @xmath454 , or equivalently @xmath475 , @xmath476 , @xmath477 , see eq . ( [ omeg ] ) ] to solve for the single - particle spectrum ( @xmath478 ) .
4 . evaluate the average , @xmath386 , of the single - particle spectrum according to eq .
( [ harmav ] ) and subsequent remarks . 5 .
use the results of steps 3 and 4 above to calculate the shell correction @xmath479 according to eq .
( [ struav ] ) . 6 .
finally , calculate the total energy @xmath89 as the sum of the liquid - drop contribution ( step 2 ) and the shell correction ( step 5 ) , namely @xmath480 .
the optimal ellipsoidal geometries for a given cluster m@xmath481 , neutral or charged , are determined by systematically varying the distortion ( namely , the parameters @xmath482 and @xmath483 ) in order to locate the global minimum of the total energy @xmath484 . as examples of applications of the se - scm
, we present here three cases . in ,
we show experimental electron affinities for doubly negatively charged silver clusters @xcite and compare them with theoretical calculations @xcite . in , we compare ft - se - scm calculations for the ips of neutral k@xmath8 clusters with experimental results @xcite ; such comparisons demonstrate the importance of electronic - entropy effects .
finally , in , we display se - scm calculations for the fission barriers associated with the asymmetric channel na@xmath485 @xmath469 na@xmath470na@xmath471 @xcite ; see caption for details .
the phenomenological binding potential as a function of fission - fragment separation is described via a two - center - oscillator model @xcite .
199 v. m. strutinsky , _ nucl .
a _ * 95 * , 420 ( 1967 ) ; _ nucl .
a _ * 122 * , 1 ( 1968 ) . p. ring and p. schuck , _ the nuclear many - body problem _ , ( springer , new york , 1980 ) . c. yannouleas and u. landman , _ phys .
b _ * 48 * , 8376 ( 1993 )
. l. w. wang and m. p. teter , _ phys .
b _ * 45 * , 13196 ( 1992 ) . f. perrot , _ j. phys . :
cond . matter _
* 6 * , 431 ( 1994 ) .
e. smargiassi and p. a. madden , _ phys .
b _ * 49 * , 5220 ( 1994 ) . t. j. frankcombe , g .- j .
kroes , n. i. choly , and e. kaxiras , _ j. phys
b _ * 109 * , 16554 ( 2005 )
. y. a. wang and e. a. carter , in _ theoretical methods in condensed phase chemistry _
, s. d. schwartz ( ed . ) , ( kluwer , dordrecht , 2000 ) , p. 117 . s. b. trickey , v. v. karasiev , and r. s. jones , _ int . j. quantum chem . _ * 109 * , 2943 ( 2009 ) . m. brack , in _ atomic clusters and nanoparticles : les houches session lxxiii 2 - 28 july 2000 _ , c. guest , p. hobza , f. spiegelman , and f. david ( eds . ) , ( springer , berlin , 2001 ) p. 161 .
d. ullmo , t. nagano , s. tomsovic , and h. u. baranger , _ phys .
b _ * 63 * , 125339 ( 2001 ) . ya
. i. delchev , a. i. kuleff , t. z. mineva , f. zahariev , and j. maruani , _ int . j. quantum chem . _ * 99 * , 265 ( 2004 ) . w. zhu , s. b. trickey , _ int . j. quantum chem . _ * 100 * , 245 ( 2004 ) .
this paper studied a perturbative dft approach in the context of the harris functional @xcite .
w. d. knight , k. clemenger , w. a. de heer , w. a. saunders , m. y. chou , and m. l. cohen , _ phys .
lett . _ * 52 * , 2141 ( 1984 ) .
w. a. de heer , _ rev .
phys . _ * 65 * , 611 ( 1993 ) .
m. brack , _ rev .
phys . _ * 65 * , 677 ( 1993 ) .
m. l. homer , e. c. honea , j. l. persson , and r. l. whetten , ( unpublished ) . , w. ekardt ( ed . ) , ( john - wiley , new york , 1999 ) .
b. j. zhou and y. a. wang , _ j. chem .
phys . _ * 127 * , 064101 ( 2007 ) .
b. j. zhou and y. a. wang , _ j. chem .
phys . _ * 128 * , 084101 ( 2008 ) .
u. landman , r. n. barnett , c. l. cleveland , and g. rajagopal , in _ physics and chemistry of finite systems : from clusters to crystals _
, p. jena , s. n. khanna , and b. k. rao ( eds . ) , ( kluwer academic publishers , dordrecht , 1992 ) , vol .
i , p. 165 ; j. jortner , _ z. phys .
d _ * 24 * , 247 ( 1992 ) . m. a. preston and r. k. bhaduri , _ structure of the nucleus _ , ( addison - wesley , london , 1975 ) .
c. yannouleas and u. landman , _ chem .
lett . _ * 210 * , 437 ( 1993 ) .
r. n. barnett , c. yannouleas , and u. landman , _
d _ * 26 * , 119 ( 1993 ) . c. yannouleas and u. landman , _ phys .
b _ * 51 * , 1902 ( 1995 ) . c. yannouleas , r. n. barnett , and u. landman , _ comments at . mol .
phys . _ * 31 * , 445 ( 1995 ) . c. yannouleas and u. landman , _ j. phys
. chem . _ * 99 * , 14577 ( 1995 ) .
h. koizumi , s. sugano , and y. ishii , _
* 28 * , 223 ( 1993 ) ; m. nakamura , y. ishii , a. tamura , and s. sugano , _ phys .
a _ * 42 * , 2267 ( 1990 ) .
a. bulgac and c. lewenkopf , _ phys .
lett . _ * 71 * , 4130 ( 1993 ) .
s. frauendorf and v. v. pashkevich , _ z. phys .
* 26 * , s 98 ( 1993 ) .
s. m. reimann , m. brack , and k. hansen , _ z. phys .
* 28 * , 235 ( 1993 ) . c. yannouleas and u. landman , _ j. chem
. phys . _ * 107 * , 1032 ( 1997 ) . c. yannouleas and u. landman , _ phys .
lett . _ * 78 * , 1424 ( 1997 ) . c. yannouleas , u. landman , and r. n. barnett , _ dissociation , fragmentation and fission of simple metal clusters _ ,
p. 145 in ref .
@xcite . c. yannouleas and u. landman , _ rhys .
b _ * 61 * , r10587 ( 2000 ) . c. yannouleas , u. landman , a. herlert , and l. schweikhard , _ phys . rev .
lett . _ * 86 * , 2996 ( 2001 ) . c. yannouleas , u. landman , a. herlert , and l. schweikhard , _ eur .
j. d _ * 16 * , 81 ( 2001 ) . c. yannouleas , u. landman , c. brechignac , ph .
cahuzac , b. concina , and j. leygnier , _ phys .
lett . _ * 89 * , 173403 ( 2002 ) . c. yannouleas and u. landman , _ chem .
lett . _ * 217 * , 175 ( 1994 ) . c. yannouleas and u. landman , _ j. phys
b _ * 101 * , 5780 ( 1997 ) . c. yannouleas , e. n. bogachek , and u. landman , _ phys . rev .
b _ * 57 * , 4872 ( 1998 )
. a. stafford , d. baeriswyl , and j. brki , _ phys .
lett . _ * 79 * , 2863 ( 1997 ) .
bohr and b. r. mottelson , _ nuclear structure _
( benjamin , reading , massachusetts , 1975 ) , vol .
ii . c. f. von weizscker , _ z. phys . _ * 96 * , 431 ( 1935 ) .
h. a. bethe and r. f. bacher , _ rev .
* 8 * , 82 ( 1936 ) .
w. d. myers and w. j. swiatecki , _ nucl .
phys . _ * 81 * , 1 ( 1966 ) .
d. r. snider and r. s. sorbello , _ solid state commun .
_ * 47 * , 845 ( 1983 ) .
m. brack , _ phys .
b _ * 39 * , 3533 ( 1989 )
serra , f. garcas , m. barranco , j. navarro , l. c. balbs , and a. maanes , _ phys .
b _ * 39 * , 8247 ( 1989 ) .
m. membrado , a. f. pacheco , and j. sando , _ phys .
b _ * 41 * , 5643 ( 1990 ) .
e. engel and j. p. perdew , _ phys .
b _ * 43 * , 1331 ( 1991 ) . m. seidl , k .- h .
meiwes - broer , and m. brack , _ j. chem .
phys . _ * 95 * , 1295 ( 1991 ) . c. yannouleas , r. a. broglia , m. brack , and p. f. bortignon , _ phys .
lett . _ * 63 * , 255 ( 1989 ) . c. yannouleas and r. a. broglia , _ phys .
a _ * 44 * , 5793 ( 1991 ) ; _ europhys .
lett . _ * 15 * , 843 ( 1991 ) ; c. yannouleas , p. jena , and s. n. khanna , _ phys . rev .
b _ * 46 * , 9751 ( 1992 ) . c. yannouleas , _ chem . phys .
lett . _ * 193 * , 587 ( 1992 ) . c. yannouleas and r. a. broglia , _ ann .
( n.y . ) * 217 * , 105 ( 1992 ) ; c. yannouleas , e. vigezzi , and r. a. broglia , _ phys . rev .
b _ * 47 * , 9849 ( 1993 ) ; c. yannouleas , f. catara , and n. van giai , _ phys . rev . b _ * 51 * , 4569 ( 1995 ) .
w. kohn and l. j. sham , _ phys .
rev . _ * 140 * , a1133 ( 1965 ) .
w. a. de heer , _ rev .
phys . _ * 65 * , 611 ( 1993 ) .
v. v. kresin , _ phys .
rep . _ * 220 * , 1 ( 1992 ) .
w. ekardt , w. _ phys .
b _ * 31 * , 6360 ( 1985 )
. d. e. beck , ( 1991 ) _ phys .
b _ * 43 * , 7301 ( 1991 ) .
j. harris , _ phys .
b _ * 31 * , 1770 ( 1985 ) . m. w. finnis , _ j. phys . : condens . matter _ * 2 * , 331 ( 1990 ) . h. m. polatoglou and m. methfessel , _ phys . rev .
b _ * 37 * , 10403 ( 1988 ) .
w. m. c. foulkes and r. haydock , _ phys . rev .
b _ * 39 * , 12520 ( 1989 )
. e. zaremba , _ j. phys . : condens
. matter _ * 2 * , 2479 ( 1990 ) .
l. h. thomas , _ proc .
cambridge philos .
* 23 * , 542 ( 1926 ) ; e. fermi , _ z. phys . _ * 48 * , 73 ( 1928 ) .
c. h. hodges , _ can .
j. phys . _
* 51 * , 1428 ( 1973 ) . w. ekardt , _ phys .
b _ * 29 * , 1558 ( 1894 ) .
w. a. de heer , w. d. knight , m. y. chou , and m. l. cohen , _ solid state phys .
_ * 40 * , 93 ( 1987 ) .
r. n. barnett , u. landman , and g. rajagopal , _ phys .
lett . _ * 67 * , 3058 ( 1991 ) ; see also r. n. barnett and u. landman , _ ibid .
_ * 69 * , 1472 ( 1992 ) ; r. n. barnett , u. landman , a. nitzan , and g. rajagopal , _ j. chem . phys . _ * 94 * , 608 ( 1991 ) ; h .-
cheng , r. n. barnett , and u. landman , _ phys .
b _ * 48 * , 1820 ( 1993 ) .
u. rthlisberger and w. andreoni , _ j. chem .
phys . _ * 94 * , 8129 ( 1991 ) .
j. p. perdew and a. zunger , _ phys .
b _ * 23 * , 5048 ( 1981 ) . l. c. balbs , a. rubio , and j. a. alonso , _ chemical phys . _
* 120 * , 239 ( 1988 ) .
z. penzar and w. ekardt , _ z. phys .
* 17 * , 69 ( 1990 ) .
d. m. wood , _ phys .
lett . _ * 46 * , 749 ( 1981 ) .
m. p. j. van staveren , h. b. brom , l. j. de jongh , and y. ishii , _ phys .
b _ * 35 * , 7749 ( 1987 ) .
j. p. perdew and y. wang , _ phys .
b _ * 38 * , 12228 ( 1988 ) . ph .
j. siemens and a. s. jensen , _ elements of nuclei _ ( addison - wesley , new york , 1987 ) .
we emphasize that while the effective potentials are significantly different when sic is used , other quantities , such as the total energy , ips , and eas are only slightly altered by sic as shown in ref .
@xcite , and by our own calculations .
s. hofmann , _
proton radioactivity _ in _ particle emission from nuclei _
, d. n. poenaru and m. s. ivascu ( eds . ) , ( crc press , boca raton , florida , 1989 ) vol .
ii , p. 25 .
c. brchignac , ph .
cahuzac , f. carlier , and j. leygnier , _ phys .
lett . _ * 63 * , 1368 ( 1989 ) .
n. troullier and j. l. martins , _ phys
b _ * 46 * , 1754 ( 1992 ) ; j. l. martins , n. troullier , and j. h. weaver , _ chem . phys
. lett . _ * 180 * , 457 ( 1991 ) .
j. kohanoff , w. andreoni , and m. parrinello , _ chem .
lett . _ * 198 * , 472 ( 1992 ) .
k. yabana and g. f. bertsch , _ physica scripta _ * 48 * , 633 ( 1993 ) . n. van giai and e. lipparini , _ z. phys .
* 27 * , 193 ( 1993 ) .
m. j. puska and r. m. nieminen , _ phys .
a _ * 47 * , 1181 ( 1993 ) . b. grammaticos , _ z. phys .
a _ * 305 * , 257 ( 1982 ) . g. a. gallup , _ chem .
lett . _ * 187 * , 187 ( 1991 ) .
r. c. haddon , l. e. brus , and k. raghavachari , _ chem .
lett . _ * 125 * , 459 ( 1986 ) .
m. gerloch and r. c. slade , _ ligand field parameters _ , ( cambridge univ . press ,
london , 1973 ) .
a. rosn and b. wstberg , _ j. chem .
phys . _ * 90 * , 2525 ( 1989 ) ; b. wstberg and a. rosn , _ physica scripta _ * 44 * , 276 ( 1991 ) . l. ye and a. j. freeman , _ chem
. phys . _ * 160 * , 415 ( 1992 ) . due to the changing spill - out with excess charge @xmath179 ,
the capacitance should be written as @xmath486 . for our purposes here
the small correction @xmath487 can be neglected .
m. r. pederson and a. a. quong , _ phys .
b _ * 46 * , 13584 ( 1992 ) .
y. wang , d. tomnek , g. f. bertsch , and r. s. ruoff , _ phys .
b _ * 47 * , 6711 ( 1993 ) .
m. sai baba , t. s. lakshmi narasimhan , r. balasubramanian , and c. k. mathews , _ int . j. mass spectrom .
ion processes _ * 125 * , r1 ( 1993 ) .
r. l. hettich , r. n. compton , and r. h. ritchie , _ phys .
lett . _ * 67 * , 1242 ( 1991 ) .
a. limbach , l. schweikhard , k. a. cowen , m. t. mcdermott , a. g. marshall , and j. v. coe , _ j. am . chem . soc . _ * 113 * , 6795 ( 1991 ) . for certain systems , such as for example sodium clusters ,
an orbitally - averaged - like sic treatment yielded highest - occupied - molecular - orbital ( homo ) energies for anions in adequate agreement with the calculated electron affinities ( see refs .
j. cioslowski and k. raghavachari , _ j. chem .
phys . _ * 98 * , 8734 ( 1993 ) .
a. i. baz , y. b. zeldovich , and a. m. perelomov , _ scattering , reactions , and decay in nonrelativistic quantum mechanics _ , ( israel program for scientific translations ltd . ,
jerusalem , 1969 ) . v. de coulon , j. l. martins , and f. reuse , _ phys .
b _ * 45 * , 13 671 ( 1992 ) .
u. landman , w. d. luedtke , n. burnham , and r. j. colton , _ science _ * 248 * , 454 ( 1990 ) .
u. landman , w. d. luedtke , b. e. salisbury , and r. .l .
whetten , _ phys .
lett . _ * 77 * , 1362 ( 1996 ) .
u. landman , w. d. luedtke , and j. gao , _ langmuir _ * 12 * , 4514 ( 1996 ) .
e. n. bogachek , a. m. zagoskin , and i. o. kulik , _ fiz .
* 16 * , 1404 ( 1990 )
[ _ sov . j. low temp .
* 16 * , 796 ( 1990 ) ] .
j. i. pascual , j. mendez , j. gomez - herrero , j. m. baro , n. garcia , and v. t. binh , _ phys .
lett . _ * 71 * , 1852 ( 1993 ) .
l. olesen , e. laegsgaard , i. stensgaard , f. besenbacher , j. schiotz , p. stoltze , k. w. jacobsen , and j. n. norskov , _ phys .
lett . _ * 72 * , 2251 ( 1994 ) .
j. i. pascual , j. mendez , j. gomez - herrero , j. m. baro , n. garcia , u. landman , w. d. luedtke , e. n. bogachek , and h .-
cheng , _ science _ * 267 * , 1793 ( 1995 ) .
d. p. e. smith , _ science _ * 269 * , 371 ( 1995 ) .
g. rubio , n. agrait , and s. vieira , _ phys .
lett . _ * 76 * , 2302 ( 1996 ) .
a. stalder and u. durig , _ appl .
lett . _ * 68 * , 637 ( 1996 ) .
j. m. krans , j. m. van ruitenbeek , v. v. fisun , i. k. yanson , and l. j. de jongh , _ nature _ * 375 * , 767 ( 1995 ) . j. l. costa - kramer , n. garcia , p. garcia - mochales , and p. a. serena , _ surface science _ * 342 * , 11144 ( 1995 ) .
r. n. barnett and u. landman , _ nature _ * 387 * , 788 ( 1997 ) . for an axially symmetric nanowire with variable radius ,
see ref .
a. g. scherbakov , e. n. bogachek , and u. landman , _ phys .
b _ * 53 * , 4054 ( 1996 )
. e. n. bogachek , a. g. scherbakov , and u. landman , _ phys .
b _ * 56 * , 1065 ( 1997 ) . e. n. bogachek , a. g. scherbakov , and u. landman , _ phys . rev . b _ * 53 * , r13246 ( 1996 ) . a. garcia - martin , j. a. torres , and j. j. saenz , _ phys . rev . b _ * 54 * , 13448 ( 1996 ) .
r. a. gherghescu , d. n. poenaru , a. solovyov , and w. greiner , _ int . j. mod
b _ * 22 * , 4917 ( 2008 ) .
w. a. saunders , _ phys .
a _ * 46 * , 7028 ( 1992 ) . here , we consider clusters of monovalent elements ( na , k , and cu ) . for polyvalent elements , @xmath6 in eq .
( [ npar ] ) must be replaced by @xmath488 , where @xmath100 is the valency .
i. s. gradshteyn and i. m. ryzhik , _ table of integrals , series , and products _ , ( academic , new york , 1980 ) ch . 8.11 .
r. w. hasse and w. d. myers , _ geometrical relationships of macroscopic nuclear physics _ , ( springer - verlag , berlin , 1988 ) ch .
s. g. nilsson , _
k. danske vidensk .
selsk . mat .-
* 29 * , no . 16
k. l. clemenger , _ phys .
b _ * 32 * , 1359 ( 1985 ) . w. a. saunders , _ ph.d .
, university of california , berkeley ( 1986 ) ; w. a. saunders , k. clemenger , w. a. de heer , and w. d. knight , _ phys .
b _ * 32 * , 1366 ( 1985 ) . k. l. clemenger , _ ph.d .
dissertation _ ,
university of california , berkeley ( 1985 ) .
j. r. nix , _ annu .
nucl . part .
* 22 * , 65 ( 1972 ) .
r. k. bhaduri and c. k. ross , _ phys .
* 27 * , 606 ( 1971 ) .
the perturbation @xmath489 in the hamiltonian ( [ hn ] ) influences the shell correction @xmath479 , but not the average , @xmath386 , of the single - particle spectrum , since @xmath490 for all shells with principal quantum number @xmath347 higher than the minimum number required for accomodating @xmath248 electrons ( see , ref .
@xcite , p. 598
a. herlert , l. schweikhard , and m. vogel , _ eur .
j. d _ * 16 * , 65 ( 2001 ) .
j. maruhn and w. greiner , _ z. phys . _
* 251 * , 431 ( 1972 ) . | orbital - free ( of ) methods promise significant speed - up of computations based on density functional theory ( dft ) . in this field ,
the development of accurate kinetic - energy density functionals remains an open question . in this chapter
we review the shell - correction method ( scm , commonly known as strutinsky s averaging method ) applied originally in nuclear physics and its more recent formulation in the context of dft [ yannouleas and landman , _ phys . rev .
b _ * 48 * , 8376 ( 1993 ) ] .
we demonstrate the dft - scm method through its earlier applications to condensed - matter finite systems , including metal clusters , fullerenes , and metal nanowires .
the dft - scm incorporates quantum mechanical interference effects and thus offers an improvement compared to the use of thomas - fermi - type kinetic energy density functionals in of - dft . |
the study of rich galaxy clusters at high redshift has important consequences for our understanding of structure formation in the universe and is a crucial test of cosmological models .
numerical simulations of hierarchical models such as cold dark matter ( cdm ) predict few massive clusters at large redshift ( cen & ostriker 1994 ) , and that the evolution of cluster number density as a function of x - ray luminosity and temperature depends sensitively upon @xmath15 , but only weakly upon @xmath16 and the initial power spectrum ( peebles , daly , & juszkiewicz 1989 ; evrard 1989 ; eke et al .
moderate - redshift clusters from well - defined samples such as the _ rosat _ deep cluster survey ( rdcs ; rosati et al .
1998 ) have been used to constrain @xmath17 and @xmath18 ( borgani 2001 ) .
distant x - ray luminous clusters provide the best lever arm for these studies .
however , to date , few ( @xmath19 ) @xmath2 clusters have been spectroscopically confirmed ( dickinson 1995 ; stanford et al .
1997 ; rosati et al .
1999 ; liu et al .
2000 ; rosati 2001 ; thompson et al . 2001
; stanford et al .
2002 ) .
clusters and groups of galaxies also provide a crucial tool in the study of galaxy formation and evolution .
out to at least @xmath20 clusters tend to be dominated by a population of massive ellipticals which is largely homogenous and has been quiescent since at least @xmath21 ( stanford , eisenhardt , & dickinson 1998 ) . finding high - redshift , massive elliptical systems is difficult , but the implications for the epoch of early - type galaxy formation can be provocative , as evidenced by lbds 53w091 , a galaxy at @xmath22 whose 3.5 gyr age is comparable to the hubble time for its redshift ( dunlop et al . 1996 ; spinrad et al .
an expanded census of dense environments in the early universe will provide a powerful means to test models of large scale structure formation , characterize the galaxy populations in these environments , and study the formation epoch of early - type galaxies .
most bound clusters beyond redshift unity have been identified from deep , serendipitous x - ray surveys , deep near - ir imaging surveys , and/or around powerful 3c radio sources .
radio galaxies are robust signposts of early collapse . in the local universe ,
bright radio sources are often hosted by giant elliptical and cd galaxies residing within dense environments , and this identification has been shown to continue to higher redshift 3c184 at @xmath23 ( deltorn et al .
1997 ) , 3c324 at @xmath24 ( dickinson 1995 ) , 3c294 at @xmath25 ( fabian et al .
2001 ) , and tn j1338@xmath261942 at @xmath27 ( venemans et al .
hall et al .
( 2001 ) also show an excess population of red galaxies around radio - loud quasars at @xmath28 , interpreted as being due to rich environments associated with the quasars . in terms of studying the intracluster gas , however
, clusters around the _ most luminous _ high - redshift radio galaxies ( hzrgs ) and radio - loud quasars are less than ideal , as observed x - ray emission does not necessarily derive solely from the hot gas .
hardcastle , lawrence , & worrall ( 1998 ) show that unresolved , soft x - ray flux correlates with 5 ghz core flux for hzrgs and core - dominated radio - loud quasars .
_ chandra _ has also revealed soft , extended x - ray emission coincident with the outer radio lobes in powerful 3c sources ( wilson , young , & shopbell 2000 ; harris et al .
this emission is thought to be due to inverse - compton scattering of cosmic microwave background photons by the relativistic plasma in radio jets / lobes . at larger redshifts ,
the energy density of the microwave background is higher , so that the cooling time for relativistic electrons is short .
agn - related soft x - ray emission should only be problematic in the youngest and most luminous radio sources at high redshift .
therefore , the next obvious step in identifying a sample of high - redshift clusters well - suited to x - ray follow - up is to locate clusters identified with _
less powerful _ radio galaxies . in this paper , we report the discovery of an overdensity of galaxies at @xmath29 associated with the moderate - strength ( @xmath30 mjy ; griffiths et al .
1995 ) , narrow - lined hzrg .
though the x - ray observations detailed herein do not confirm a hot intracluster medium associated with this overdensity , the technique is sound and the radio galaxy clearly marks a rich location in the early universe .
the group , cl 0442 + 0202 , was identified from a deep optical / near - ir imaging survey of radio sources at @xmath2 selected from the 5 ghz mit - greenbank ( mg ) radio catalog ( bennett et al .
1986 ; stern et al .
1999 ) . is a moderately - steep spectrum radio source ( @xmath31 , where @xmath32 ) , approximately a quarter as luminous at 1.4 ghz compared to the typical @xmath33 3c hzrg .
the 365 mhz morphology is double , extended by @xmath34 at a position angle of @xmath35 ( douglas et al .
the 5 ghz morphology , measured with the vla , reveals a triple morphology , with two nearly equal intensity lobes separated by approximately 9 at a position angle of 30 , flanking a centrally - located core with a flux 24% of the mean lobe radio flux .
initial optical identification and redshift determination of were obtained at lick observatory as part of the berkeley effort to study mg sources ( spinrad et al . 1993 ; stern et al .
1997 ) .
we assume @xmath36 , @xmath37 , and @xmath38 throughout .
for this cosmology , the luminosity distance of is @xmath39 gpc and one arcsecond subtends @xmath40 proper kpc . for @xmath41 , @xmath42 , and @xmath43 , the cosmology favored by recent high - redshift supernovae and cosmic microwave background observations , these distances change very slightly : they are smaller by 0.4% . throughout
we have corrected for foreground galactic extinction using a reddening of @xmath44 determined from the dust maps of schlegel , finkbeiner , & davis ( 1998 ) .
to search for red galaxies associated with the hzrg , we obtained deep , near - infrared imaging of the field using the gemini twin - arrays infrared camera ( mclean et al .
1994 ) on the lick 3 m shane telescope .
gemini uses a dichroic beam splitter which yields two independent beams : a short - wave channel with a rockwell nicmos 3 @xmath45 hgcdte array and a long - wave channel with a sbrc @xmath45 insb array .
gemini was used in its 07 pix@xmath46 mode , providing a 3 @xmath47 3 field of view .
our data , obtained through @xmath48 ( @xmath49 m ; @xmath50 m ) and @xmath51 ( @xmath52 m ; @xmath53 m ) filters for the two channels respectively , were obtained in non - photometric conditions on ut 1999 january 3 and ut 1999 december 2 .
the total integration time was 12 ks , and the combined @xmath51 image , reporting 23 seeing , is provided in fig .
[ fig_kimage ] . to calibrate these images
, we obtained @xmath54 snapshots of the field on ut 1999 february 26 in photometric conditions and 07 seeing with nsfcam ( shure et al .
1994 ) at the 3 m infrared telescope facility ( irtf ) .
nsfcam was used in its 03 pix@xmath46 mode , providing a 77 @xmath47 77 field .
observations of standard stars from the persson et al .
( 1998 ) list were used to calibrate the images onto the california institute of technology ( cit ) system , defined in elias et al .
the calibrated lick observatory @xmath51 image reaches a depth of @xmath55 ( 3@xmath5 in a 46 diameter aperture ) .
we obtained an optical , @xmath56-band ( @xmath57 8331 ; @xmath58 = 3131 ) image of on ut 2001 january 28 with the low resolution imaging spectrometer ( lris ; oke et al .
1995 ) on the keck i telescope .
this 600 s exposure , calibrated with observations of pg 1323@xmath26086 ( landolt 1992 ) obtained the same night , has 074 seeing and reaches a depth of 24.7 mag ( @xmath9 in 15 diameter aperture ) .
lris uses a @xmath59 backside - illuminated tek ccd with 0212 pix@xmath46 , providing a 6 @xmath47 78 field of view .
after geometrically transforming the images to a common frame , we used sextractor ( release v2.1.6 ; bertin & arnouts 1996 ) to create catalogs of sources selected from the @xmath51 images .
photometry was generated for 6 diameter apertures .
[ fig_colormag ] presents resultant color - magnitude diagrams for the lick and irtf @xmath51 images .
compared to random extragalactic fields , the field of has a high surface density of objects with red @xmath3 colors . using the irtf infrared imaging , we find six sources in the 1.6 arcmin@xmath60 field with @xmath61 and @xmath62 , corresponding to a surface density of extremely red objects ( eros ) of 3.8 eros arcmin@xmath63 .
this is 17 times the surface density of such sources found by barger et al .
( 1999 ) in a 61.8 arcmin@xmath60 survey and @xmath64 times the surface density of such sources found in the 47.2 arcmin@xmath60 herschel deep field ( mccracken et al .
2000 ) . the dotted line in fig .
[ fig_colormag ] at @xmath65 shows the red sequence characteristic of local cluster galaxies , as determined by band - shifting photometry of coma cluster galaxies ( see stanford et al .
1998 ) . particularly in the better - seeing irtf imaging where confusion is less troublesome , photometric errors less severe , and the smaller field less contaminated by field galaxies , a red sequence is evident approximately 0.5 mag blue - ward of the no - evolution prediction .
this amount of bluing in observed @xmath3 color is consistent with the color change due to passive evolution of a single - age , solar - metallicity , stellar population at @xmath0 formed in a 0.1 gyr burst at a formation redshift @xmath66 ( using the updated 2000 gissel models of bruzual & charlot 1993 ) .
we targeted this red sequence for spectroscopic investigation .
we obtained optical spectra of photometric candidates using lris in slitmask mode on the keck i telescope .
masks on ut 2001 january 30 and ut 2001 february 19 were observed for 1.5 hr , while a fainter mask on ut 2002 february 5@xmath266 was observed for 3.5 hr . on ut 2002 march 10
we observed a mask mainly targeting _
sources in the field for 1.5 hr .
a shallow , 20 min longslit observation of two bright _ chandra _ sources was also obtained on ut 2002 march 11 during twilight .
all observations used the 400 lines mm@xmath46 grating ( @xmath67 ; spectral resolution @xmath68 ) and sample the wavelength range @xmath69 to 1 @xmath70 m .
the march 2002 observations also implemented the blue side of lris , using a d560 dichroic and the 300 lines mm@xmath46 grism ( @xmath71 ; spectral resolution @xmath72 ) , providing spectral coverage down to @xmath73 .
the observations were split into 1800 s exposures with @xmath74 3 spatial offsets performed between integrations to facilitate removal of fringing at long wavelengths .
data reductions followed standard procedures and the spectra were flux - calibrated using observations of standard stars from massey & gronwall ( 1990 ) .
sample optical spectra are presented in fig .
[ fig_spec ] and results for all @xmath51-detected sources are presented in table [ tableprop ] .
redshifts were determined by visual inspection of emission and absorption features .
a histogram of the redshifts we have obtained thus far in the field is presented in fig .
[ fig_zhist ] , showing a pronounced redshift spike at @xmath75 .
besides , a total of seven color - selected targets within @xmath76 30 of the radio galaxy were attempted spectroscopically , of which five are confirmed cluster members and one has an inconclusive spectrum .
most of the confirmed cluster members have spectra typical of local early - type galaxies , showing strong 4000 breaks and weak or no emission lines .
the exception is , of course , the hzrg which shows high equivalent width , narrow , high ionization state emission features , typical of the hzrg population ( mccarthy 1993 ) .
a search of the _ rosat _ archive reveals that high resolution imager ( hri ; trmper 1983 ) observations of the @xmath77 cluster ms 0440 + 0204 cover cl 0442 + 0202 at an off - axis angle of 134 . these data , discussed in gioia et al .
( 1998 ) , were obtained in 1994 february / march and 1995 august / september .
we processed the archival data , which total 27.2 ks , and identify a 3.3@xmath5 x - ray source slightly ( 2@xmath5 of the point spread function ) offset from the radio galaxy .
the resolution of _ rosat _ at the large off - axis angle of cl 0442 + 0202 is inadequate for determining if the x - ray emission is extended or point - like . after correcting for the expected soft x - ray emission from the radio galaxy ( hardcastle et al .
1998 ) , the _ rosat _ detection suggested soft x - ray emission of @xmath78 , consistent with a high - redshift hot intracluster medium ( icm ) .
this exciting hypothesis inspired an observation with the _ chandra x - ray observatory _ ( weisskopf , odell , & van speybroeck 1996 ) .
on ut 2002 february 10 we obtained a 44.1 ks observation of with _
chandra_. compared to the _ rosat _ image , these data ( obs - id 03242 ) , taken with the advanced ccd imaging spectrometer ( acis - i ) , reach a greater depth with significantly improved ( @xmath79 ) angular resolution .
we reduced and analyzed the data following standard procedures , using the _ chandra _ interactive analysis of observations ( ciao ) software ( release v2.1.2 ) .
sources were identified independently in both soft ( 0.5@xmath262 kev ; @xmath80 ) and hard ( 2@xmath267 kev ; @xmath81 ) energy bands . following previous analyses of deep _ chandra _ data ( tozzi et al .
2001 ; stern et al .
2002b ) , we determined a hardness ratio @xmath82 , using the background - subtracted counts .
a response matrix and effective area were then determined for each detected source . assuming a galactic @xmath83 and a typical photon index , @xmath84 , counts in each band were converted into fluxes .
table 2 presents the x - ray results for those _ chandra _ sources contained within the optical image , along with optical magnitudes and spectroscopic redshifts , when available .
the most striking result of the _ chandra _ data is that the improved spatial resolution resolves the _ rosat _ detection into five point sources .
the total soft - band flux of these sources accounts for only 20% of the _ rosat _ detection , likely due to the low significance of the _ rosat _ detection with temporal variability also possible .
as expected , the radio galaxy is detected with a relatively hard x - ray spectrum ; we discuss this source further in 3.2 . at the depth of our image , @xmath85 ,
_ chandra _ observations of deep fields find a mean surface density of approximately 1000 sources deg@xmath63 ( stern et al .
the five sources we detect in our _ chandra _ image , all within 15 of the hzrg , represent a surface density approximately two orders of magnitude higher !
unfortunately , this unusual configuration of point sources rather than an extended icm is likely responsible for the _ rosat _ detection .
we find no diffuse emission associated with ( see 3.4 ) .
of the twelve _ chandra _ sources within the optical image , we have unambiguous spectroscopic redshifts for six and and tentative redshifts for an additional two .
typical of follow - up of _ chandra _ sources , the two sources with the softest ( @xmath86 ) x - ray spectra are identified with luminous , broad - lined quasars .
both have projected celestial positions close to ( see table 1 ) , but one ( cxo 6 ) is background while the other ( cxo 7 ) is foreground .
potentially they could provide useful probes of the environment of .
the two sources with ambiguous spectra both have relatively soft x - ray spectra and optical spectra that reveal blue continuum , each with the detection of a single , broad emission line .
we tentatively interpret both lines as @xmath87 2800 : at @xmath33 , the strong rest - frame uv lines characteristic of quasars have not yet entered the optical window , while the strong rest - frame optical lines have shifted to near - ir wavelengths , either beyond the optical window or at redder wavelengths where the sky is less benign .
the source cxo 10 is typical of the new population of sources being identified in deep _
chandra _ images ( hornschemeier et al .
2001 ; stern et al .
2002b ) : galaxies whose optical / near - ir properties suggest a normal galaxy with no indication of an active nucleus , but are nevertheless luminous x - ray sources .
many of these sources , including cxo 10 , have hard x - ray spectra , implying significant obscuration of a buried central engine . in the case of cxo 10 ,
the optical spectrum is dominated by an old stellar population with a strong 4000 break .
weak [ ] @xmath87 3727 is also present , possibly due to low levels of star formation , possibly due to the narrow - line region of the buried , active nucleus . finally , our _ chandra _ data reveal two obvious agn showing high equivalent width , high ionization state , narrow emission lines , but lacking the obvious broad - line signatures associated with quasars : cxo 5 and cxo 12 . following seyfert nomenclature ,
these sources are often called _ type ii quasars _ , and deep x - ray data has sparked a renewed interest in the population ( norman et al .
2002 ; stern et al .
2002a ; dawson et al .
the radio - loud end of this population has been studied for several decades : indeed , cxo 5 is identified with and we discuss it in more detail in 3.2 .
cxo 12 is the other type ii active system identified in our data .
the x - ray spectrum is relatively flat from the soft to the hard bands ( @xmath88 ) , providing a harder x - ray spectrum than the typical , bright x - ray population . the optical spectrum ( fig .
[ fig_qso2 ] ) shows narrow , rest - frame uv emission lines with very little continuum .
stern et al . ( 2002a ) discusses a similar source in detail ( see also norman et al . 2002 ; dawson et al .
both the apparently - normal galaxies and these type ii quasars are being found in large numbers in surveys with the new generation of x - ray satellites .
such studies will be essential for creating an unbiased census of agn in the universe , thereby testing models of the x - ray background and providing a history of accretion - driven energy production in the universe .
the @xmath0 radio galaxy is what initially drew our attention to this field . with @xmath89 , fits smoothly onto the surprisingly low - scatter hubble , or @xmath90 , relation of radio galaxies ( see fig . 10 in de breuck et al .
2001 ) . at each redshift , hzrgs are the most luminous galaxies known at observed 2@xmath70 m .
since rest - frame 2@xmath70 m samples stellar emission from the low - mass stars which dominate the baryon content of a galaxy , this is generally interpreted as hzrgs being the most massive systems at each cosmic epoch . indeed , our deep , optical spectrum of ( fig .
[ fig_spec ] ) shows that this @xmath91 radio source is hosted by an old , massive galaxy .
though the dominant spectral features are strong , narrow emission lines characteristic of radio galaxies ( mccarthy 1993 ) , the spectrum also clearly shows stellar light .
absorption is seen from ca h @xmath87 3933 and a conglomerate of / lines around 3840 ( pickles 1985 ) , indicative of low - mass stars .
ca k @xmath87 3968.5 has been filled in by [ neiii ] @xmath87 3967.5 emission .
the strong 4000 break seen in is another strong indicator of stellar emission in this hzrg .
the break is due to the sudden onset of stellar photospheric opacity shortward of 4000 , associated with the fraunhofer h and k lines of , as well as various ionization stages of other elements heavier than helium ( hman 1934 ) .
the soft x - ray properties of core - dominated quasars are strongly correlated with their radio properties , with a radio ( 5 ghz ) core to soft x - ray core two - point spectral index of @xmath92 ( worrall et al .
1994 ; worrall 1997 ) .
assuming an x - ray spectral index of @xmath93 ( see below ) , the observed flux density of at 1 kev is 3.3 njy . for its 5 ghz radio - core flux density of 12 mjy ,
is slightly overluminous in the soft x - ray compared to the relation plotted by worrall et al .
( 1994 ) for core - dominated quasars .
this result is consistent with other hzrgs examined by hardcastle et al .
( 1998 ) and may indicate non - jet - related x - ray emission originating close to the agn .
our _ chandra _ observation detects more than 250 counts from , providing a robust hardness ratio of @xmath94 . assuming the x - ray flux spectrum can be expressed as simple power - law function of energy @xmath95 [ @xmath96 modified by the galactic absorption column density in the direction of the source ( @xmath97 cm@xmath63 ) , this hardness ratio corresponds to a spectral index of @xmath99
slightly flatter than the steep @xmath100 spectra that are commonly observed for unabsorbed agns ( nandra & pounds 1994 ) .
given the narrow - line optical / uv spectrum of the source , it is likely that actually has an intrinsically steep x - ray spectrum that appears to be flat due to significant soft x - ray absorption by intervening material within the galaxy . to investigate this possibility , we examined the x - ray spectrum of more closely . using the ciao software ,
we extracted source counts within a 30 radius centered on ; the background was determined from a rectangular region of 100 arcseconds by 180 arcseconds .
this region was chosen to overlap with the acis - i node in which the radio galaxy falls , node two for chip three .
all detected sources in this region were removed .
as we were expecting a cluster in this observation , we observed with a @xmath101 mm sim - z offset , hence the target did not fall on the standard aimpoint .
we also included a 2 offset in the y - direction to move the potential cluster off of the chip gap in the acis - i mosaic , which moved the target 2 away from the optical axis .
we binned the spectrum into 26 channels of 10 events from 0.8 to 6.0 kev and we modeled the spectrum using the xspec v11.1.2 package ( arnaud 1996 ) with the recently - created chartas & getman acis absorption model and using the markevitch & vikhlinin ( 2001 ) correction to the effective area of acis - i .
we fit a power law with an absorption of @xmath102 and found the best fitting slope for a simple power - law x - ray spectrum to be @xmath103 .
if we add an additional absorber component at the redshift of the radio galaxy and fix the intrinsic value of @xmath104 , we find an column density of @xmath105 ( 90% confidence limits ) .
this model implies an unabsorbed luminosity of @xmath106 ( 90% confidence limits ; 2 - 10 kev ) , which is well within the quasar regime .
we also searched the extracted x - ray spectrum of for redshifted fek@xmath107 6.4 kev emission , which would appear at observed 3.03 kev .
[ fig_xspec ] presents the background - subtracted x - ray spectrum with a best - fit continuum model discussed above , assuming no intrinsic absorption .
we find no excess emission evident at 3 kev , though we note that _ chandra _ s sensitivity falls at these higher energies .
the color - selected galaxy overdensity associated with affords us the opportunity to study early - type galaxy populations at high redshift . as seen in fig .
[ fig_spec ] , many of the galaxies associated with have evolved stellar populations as evidenced , for example , by strong continuum decrements at 4000 .
this feature is of particular importance as the 4000 decrement intensifies with the age of a stellar population , potentially providing a sensitive probe of the age of a galaxy .
most of the other @xmath2 early - type galaxy overdensities known ( see 1 ) are at sufficiently high redshift ( @xmath108 ) that this feature has shifted to challenging wavelengths for silicon - based detectors , especially from the ground where telluric emission is severe . in such cases ,
astronomers are forced into the difficult situation of age - dating galaxies based on breaks at 2640 and 2900 , where the galaxy is several times fainter ( spinrad et al .
1997 ; yi et al .
the redshift of is ideal for probing the formation history of early - type galaxies : @xmath91 is sufficiently large to be cosmologically interesting , but not so large as to be challenging spectroscopically .
@xmath109 for the three galaxies presented in fig .
[ fig_spec ] . conservatively assuming that the stellar populations were formed from a single , instantaneous , solar - metallicity burst , these galaxies have minimum ages of approximately 1 gyr , where we have compared the amplitudes of the d(4000 ) break to bruzual & charlot ( 1993 ) stellar population synthesis models for both salpeter ( 1955 ) and scalo ( 1986 ) initial mass functions and a range of star formation histories ( fig . [ fig_d4000 ] ) .
the implied formation redshift is @xmath110 .
we note that several of the group members show [ ] emission at some level , suggesting some ongoing star formation .
this would imply younger , bluer stars diluting the d(4000 ) amplitude , and thus an earlier formation redshift .
for the exponentially - decreasing star formation history plotted , the minimum age is approximately 4 gyr , requiring formation redshifts @xmath111 .
we now compare the absolute magnitudes of the galaxies to local cluster luminosity functions : for a 1 gyr old single - burst stellar population , calculated using the worthey ( 1994 ) models for a salpeter ( 1955 ) initial mass function , solar metallicity , and a mass range of @xmath112 , the @xmath113-correction from observed @xmath51 to rest - frame gunn @xmath114 is 0.21 mag , implying @xmath115 for for our adopted cosmology . from a sample of 39 abell clusters , paolillo et al .
( 2001 ) find @xmath116 , using the same cosmology . during the 8.8 gyr that elapses from @xmath0 to @xmath117 ,
our model predicts @xmath118 mag fading due to luminosity evolution , suggesting that will evolve into a @xmath119 elliptical in the present epoch if there is no additional merging .
the next brightest confirmed cluster member , galaxy#2 , has @xmath120 , suggesting a @xmath121 elliptical in the present epoch for the same assumptions . including , we have identified six galaxies within an angular region of 1 diameter that have similar velocities .
their redshifts have a mean of 1.104 .
jackknife determinations of the velocity dispersion using both the gapper method and biweights yield @xmath122 .
we find similar results for the bootstrap method .
we analyzed the _
chandra _ data to study possible diffuse soft x - ray emission from a hot intracluster medium . excluding the obvious five point sources
, we implemented a simple curve of growth analysis .
we made a series of annuli of increasing width and subtracted off a background annulus from 40 to 60 in radius , all centered on . in no annulus did we achieve a @xmath9 detection of excess events in the 0.3 to 10.0 kev energy band .
in the 1.0 to 6.0 kev band , we find a 3@xmath5 upper limit of @xmath123 for a 3 kev cluster at @xmath124 within a 30 radius centered on the radio galaxy .
this radius corresponds to a 257 @xmath125 kpc at the redshift of the cluster , a typical core radius for a massive cluster of galaxies .
we choose 3 kev as the cluster temperature based on the @xmath126 velocity dispersion measured in 3.3 and equation 13 in wu , fabiand , & fang ( 1999 ) . translating the flux limit into a luminosity
, we find an upper limit of @xmath127 for the bolometric luminosity of the cluster .
using the bolometric luminosity velocity dispersion relation of equation 7 in wu et al .
( 1999 ) , we would expect a bolometric luminosity of @xmath128 .
this would translate into a @xmath129 detection within the 30 radius , assuming a core radius of 250 @xmath125 kpc and a @xmath130 model with a slope of @xmath131 .
optical / near - ir imaging of the moderate - strength , @xmath0 radio source has revealed an overdensity of sources with red optical to near - ir colors , typical of @xmath132 early - type galaxies .
spectroscopic investigations with the keck telescope have secured redshifts for five of these galaxies in the range @xmath133 within 35 of .
these data indicate the presence of an overdensity at high redshift .
the member galaxies form a homogenous population with similar colors . the colors and tightness of the color - magnitude relation support the conclusions of stanford et al .
( 1998 ) and van dokkum & franx ( 2001 ) who found that the spectrophotometric properties of early - type cluster galaxies are consistent with the passive evolution of an old stellar population formed at an early cosmic epoch .
our data ( fig . [ fig_colormag ] ) suggests a less steep color - magnitude relation than seen in the nearby coma cluster .
this effect is likely due to a blue agn contribution to the brightest galaxy in our group , .
similar flattenings of the color - magnitude relation are expected near the formation epoch of cluster ellipticals , though galactic winds in single - age monolithic models can also produce such effects ( van dokkum et al .
2001 ) .
archival _ rosat _ data found soft x - ray emission coincident with this structure with an intensity higher than expected from the radio galaxy alone .
deeper , higher - resolution _ chandra _ observations reveal no extended soft x - ray emission from a hot icm associated with the putative cluster , instead finding 5 x - ray point sources within 15 of the radio galaxy , approximately 100 times the average surface density of x - ray sources at this flux level .
the lack of extended x - ray emission coupled with the spectroscopically - confirmed @xmath134 galaxies with a tight color - magnitude relation lead us to interpret to be a protocluster in the process of formation .
based on the sparse data presented here , and the optical faintness of the associated galaxies , it is difficult to fully determine the nature of the overdensity associated with .
deeper data and more extensive spectroscopy are necessary to determine if will become a modest group or a rich cluster .
we can safely say , however , that is likely a structure which has broken away from the hubble flow , but is still a long way from virialization .
a weak icm is probably already in place , with some metal enrichment from previous generations of stars , but it remains undetected since is not virialized and the ( future ) icm s temperature and density are far too low to produce detectable x - ray emission . as noted previously ( lubin , oke , & postman 2002 ) , many optically - selected clusters at @xmath135 are also underluminous in the x - rays when compared to the x - ray luminosity velocity ( @xmath136 ) relation for local and moderate - redshift clusters .
this suggests that clusters are actively forming at these redshifts . accepting that is a cluster in the process of formation , we note that the old age of the member galaxies ( 3.3 ) implies that the stars in clusters are older than the icm .
similarly , based on _
hubble space telescope _
imaging of spectroscopically - confirmed members of the optically - selected , x - ray - underluminous cluster cl 1324 + 3011 ( @xmath137 ) , lubin et al .
( 2002 ) find a high early - type fraction ( @xmath138 ) , indicative of a substantial tenure for these constituent galaxies . since the icm is processed gas , these results : ( 1 ) mean the icm is relatively young , and ( 2 ) raise the question of which galaxies are the source of the metal enrichment .
this work shows that high - redshift radio galaxies can be beacons of rich regions in the early universe , allowing the identification of clusters and protoclusters at @xmath132
. a census of galaxy clusters at @xmath139 would provide a major breakthrough toward our understanding of early - type galaxy formation , cluster formation , as well as provide important constraints on basic cosmological parameters .
additionally , unified models of agn predict that hzrgs and radio - loud quasars differ only in the orientation of their central jets relative to the observer ; on average , they are expected to reside in similar galactic environments .
this hypothesis could be tested with a systematic study of a well - chosen sample of radio - loud galaxies and quasars .
the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community .
we are most fortunate to have the opportunity to conduct observations from this mountain .
we thank leslie maxfield for contributions to preliminary optical observations of the radio galaxy mg 0442 + 0202 from lick observatory and we gratefully acknowledge charles lawrence for enlightening discussion regarding the x - ray properties of hzrgs .
we thank megan eckart and fiona harrison for keck spectroscopy of _ chandra _ sources in this field , obtained as part of the sexsi survey ( harrison et al .
, in preparation ) , and we thank marc davis , alison coil , and ed moran for providing the keck @xmath56-band image .
we are indebted to ian mclean and elinor gates for building and supporting the gemini camera at lick observatory , and to kurt adelberger for an insightful and careful referee report .
support for this project came from _
grant go2 - 3194 .
the work of ds was carried out at the jet propulsion laboratory , california institute of technology , under a contract with nasa .
the work by sas and bh at lawrence livermore national laboratory was performed under the auspices of the department of energy under contract w-7405-eng-48 .
stern , d. , spinrad , h. , dey , a. , dickinson , m. , & schlegel , d. 1997 , in _ the hubble space telescope and the high - redshift universe _ , ed .
n. tanvir , a. aragon - salamanca , & j. wall ( singapore : world sci . ) , 413 lcccccccl mg 0442 + 0202 & & & & 20.33 & 17.11 & 3.22 & 1.110 & radio galaxy galaxy#25 & @xmath261.9 & @xmath262.1 & 2.9 & 22.79 & 18.79 & 4.00 & 1.105 & weak galaxy#2 & @xmath267.6 & 5.2 & 9.2 & 22.58 & 18.57 & 4.01 & 1.107 & early - type galaxy#29 & 3.2 & 13.7 & 14.1 & 23.42 & 21.14 & 2.28 & 1.102 & early - type galaxy#27 & 3.9 & 22.7 & 23.0 & 23.15 & 19.92 & 3.23 & 1.100 & early - type galaxy#8 & 27.4 & @xmath2614.3 & 30.9 & 22.48 & 19.06 & 3.42 & 1.10 & early - type & & & & & & & & quasar#37 & @xmath262.9 & 11.4 & 11.7 & 21.37 & @xmath14020.0&@xmath1411.37 & 2.749 & quasar quasar#43 & @xmath268.8 & @xmath2611.2 & 14.2 & 20.45 & 18.17 & 2.28 & 0.886 & quasar galaxy#3 & @xmath2622.2 & @xmath269.5 & 24.1 & 19.98 & 17.89 & 3.09 & 0.939 & early - type galaxy#104 & 24.3 & @xmath262.2 & 24.4 & 23.07 & @xmath14020.0&@xmath1413.07&1.060 & galaxy#101 & @xmath2614.5 & @xmath2634.7 & 37.6 & 23.17 & @xmath14020.0&@xmath1413.17&1.181 & galaxy#4 & 16.2 & 39.2 & 42.4 & 22.69 & 18.73 & 3.96 & 1.125 & early - type galaxy#9 & 41.7 & @xmath2615.8 & 44.6 & 22.95 & 18.25 & 4.70 & 1.02 : & early - type : galaxy#5 & 6.1 & 51.1 & 51.5 & 22.10 & 19.01 & 3.09 & 1.133 & weak galaxy#7 & 21.4 & @xmath2656.6 & 60.5 & 22.75 & 18.56 & 4.19 & 0.94 : & galaxy#14 & 48.7 & 49.6 & 69.5 & 21.96 & 19.06 & 2.90 & 0.658 & galaxy#33 & 65.3 & @xmath2633.0 & 73.2 & 21.97 & 19.08 & 2.89 & 1.031 & galaxy#22 & @xmath2645.1 & @xmath2658.6 & 74.0 & 22.79 & 18.59 & 4.20 & 1.07 : & early - type : galaxy#17 & 55.2 & 51.6 & 75.6 & 21.13 & 18.19 & 2.94 & 0.86 : & early - type : galaxy#19 & 12.6 & @xmath2674.8 & 75.8 & 21.19 & 18.46 & 2.73 & 0.62 & early - type galaxy#16 & @xmath267.4 & 79.6 & 79.9 & 18.80 & 16.77 & 2.03 & 0.194 & h@xmath107 galaxy#6 & 67.9 & @xmath2653.8 & 86.6 & 21.79 & 18.57 & 3.22 & 1.043 & early - type galaxy#105 & @xmath2662.2 & @xmath2660.6 & 86.8 & 21.79 & & & 1.120 & galaxy#13 & 39.3 & @xmath2677.9 & 87.3 & 22.08 & 18.52 & 3.56 & 1.261 & galaxy#106 & 38.8 & @xmath2681.7 & 90.4 & 20.91 & @xmath14020.0&@xmath1410.91&0.147 & h@xmath107 galaxy#20 & 69.7 & 58.7 & 91.1 & 20.92 & 18.26 & 2.66 & 0.693 & early - type galaxy#102 & @xmath2693.2 & 10.7 & 93.8 & 20.59 & & & 0.240 & , h@xmath107 galaxy#103 & @xmath26103.1 & 10.5 & 103.6 & 22.82 & & & 1.112 & galaxy#12 & 94.7 & @xmath2671.8 & 118.8 & 22.98 & 19.19 & 3.79 & 1.269 & galaxy#107 & 99.7 & @xmath2675.6 & 125.1 & 22.86 & @xmath14020.0&@xmath1412.86&0.193 & h@xmath107 rccrrrccl cxo#2 & 04@[email protected]@xmath144 & + 020007.4 & 20.39 & 8.78 & @xmath260.30 & 23.80 & & cxo#4 & 04@[email protected]@xmath144 & + 020214.2 & 5.11 & 11.91 & 0.60 & 23.5 : & & cxo#5 & 04@[email protected]@xmath144 & + 020219.8 & 115.35 & 51.39 & @xmath260.29 & 20.91 & 1.110 & mg0442
+ 0202 cxo#6 & 04@[email protected]@xmath144 & + 020231.2 & 17.90 & 3.70 & @xmath260.51 & 21.37 & 2.749 & quasar#37 cxo#7 & 04@[email protected]@xmath144 & + 020208.6 & 58.24 & 14.83 & @xmath260.55 & 20.42 & 0.886 & quasar#43 cxo#8 & 04@[email protected]@xmath144 & + 020219.8 & 44.75 & 23.08 & @xmath260.22 & 22.87 & 0.835 : & broad mgii : cxo#10 & 04@[email protected]@xmath144 & + 020311.0 & 5.91 & 10.11 & 0.37 & 22.49 & 1.133 & galaxy#5 cxo#11 & 04@[email protected]@xmath144 & + 020230.6 & 27.12 & 14.13 & @xmath260.18 & 22.72 & & cxo#12 & 04@[email protected]@xmath144 & + 020325.4 & 20.97 & 13.41 & @xmath260.03 & 23.40 & 1.863 & type ii quasar cxo#16 & 04@[email protected]@xmath144 & + 020510.8 & 16.23 & 6.62 & @xmath260.35 & 23.24 & 0.772 & cxo#17 & 04@[email protected]@xmath144 & + 020508.5 & 36.06 & 16.59 & @xmath260.24 & @xmath14024.7 & & cxo#18 & 04@[email protected]@xmath144 & + 020438.5 & 32.98 & 14.78 & @xmath260.25 & 22.84 & 1.38 : & broad mgii : | we report the discovery of a galaxy overdensity at @xmath0 associated with the @xmath1 high - redshift radio galaxy .
the group , , was found in a near - infrared survey of @xmath2 radio galaxies undertaken to identify spatially - coincident regions with a high density of objects red in @xmath3 color , typical of @xmath2 elliptical galaxies .
spectroscopic observations from the keck i telescope reveal five galaxies within 35 of at @xmath4 .
these member galaxies have broad - band colors and optical spectra consistent with passively - evolving elliptical galaxies formed at high redshift . archival
_ rosat _
observations reveal a 3@xmath5 detection of soft x - ray emission coincident with at a level five times greater than expected for the radio galaxy .
these data are suggestive of a rich galaxy cluster and inspired a 45 ks _ chandra x - ray observatory _ observation .
as expected , the radio galaxy is unresolved to _ chandra _ , but is responsible for approximately half of the observed x - ray flux .
the remaining _ rosat _ flux is resolved into four point sources within 15 of the radio galaxy , corresponding to a surface density two orders of magnitude higher than average for x - ray sources at these flux levels ( @xmath6 ) .
one of these point sources is identified with a radio - quiet , type ii quasar at @xmath7 , akin to sources recently reported in deep _ chandra _ surveys .
the limit on an extended hot intracluster medium in the _ chandra _ data is @xmath8 ( @xmath9 , 30 radius aperture ) . though the x - ray observations do not confirm the existence of a massive , bound cluster at @xmath2 , the success of the optical / near - infrared targeting of early - type systems near the radio galaxy validates searches using radio galaxies as beacons for high - redshift large - scale structure .
we interpret to be a massive cluster in the process of formation . 2 @xmath10 2 @xmath11 2s @xmath12 2 @xmath13 2hz @xmath14 # 1to 0pt#1 # 1= #
1#2=.45 = .45 # 1#2#3#4#5#6#7 to#2 ' '' '' |
the identification of confinement mechanisms is a challenge for lattice qcd @xcite .
this requires to single out observables which both capture the relevant physical content and are in the end suitable for numerical simulations , so that a non - perturbative investigation becomes possible .
the gluon and the ghost propagators with their momentum dependence are believed to encode specific features and mechanisms of a confining theory @xcite , which can be studied once a gauge is fixed .
apart from the non - unique asymptotic infrared behavior , which
if calculated in lattice simulations realizes the so - called decoupling solution instead of the scaling solution ( see refs .
@xcite ) of the schwinger - dyson and functional renormalization group @xcite approaches , in these functions the onset of non - perturbative effects is generally agreed to be associated with the following two aspects : the coupling to local condensates @xcite and to localized vacuum excitations like vortices @xcite . in order to effectively face the genuine non - perturbative effects
, it is desirable to have for reference an understanding of the perturbative behavior of these gauge - dependent functions in higher than one - loop order .
such a program can not evade a well - known practical problem : lattice perturbation theory ( lpt ) is much more involved compared to its continuum qcd counterpart .
the complexity of the diagrammatic approaches rapidly increases beyond the one - loop approximation @xcite . by now , only a limited number of results up to two - loop accuracy has been obtained . as an effective alternative to diagrammatic lpt
a numerical tool has been proposed by the parma group in @xcite .
the method is a quite direct numerical implementation of stochastic perturbation theory , which in turn comes as one of the two main applications of the ideas of stochastic quantization , the other being the use of langevin dynamics as an alternative to standard monte carlo techniques @xcite .
this numerical tool for lpt goes under the name of numerical stochastic perturbation theory ( nspt ) and has been reviewed in @xcite . in a joined effort we started a new field of application for nspt based on our respective interests in lpt ( leipzig ) , nspt ( parma ) and
monte carlo studies of gluon and ghost propagators ( berlin ) .
first results have been presented at lattice 2007 @xcite , lattice 2008 @xcite , confinement8 @xcite and at lattice 2009 @xcite .
the present paper is the first of a series of two .
while this paper is concentrated on the ghost propagator , the forthcoming one will present the results for the gluon propagator .
one of our main goals is to provide results for the comparison of higher - loop calculations of perturbative contributions with the results of lattice monte carlo calculations of the main propagators in qcd on finite size lattices .
the other goal consists in determining higher loop contributions to the logarithmically divergent perturbative ghost propagator in momentum space including non - logarithmic contributions to good precision .
the present work also aims at a further check on the precision of the method by comparison to known one - loop results from lpt . in order to proceed to two- and higher - loop results in the infinite volume limit
, peculiar emphasis will be put on the control of finite size effects .
we want to stress that our results have been obtained from two completely independent sources : while strategies have been extensively discussed , the actual implementation of parma and leipzig nspt codes are completely independent of each other , thus providing a stronger check on numerical results .
the paper is organized as follows . in sect .
[ sec : langevinequation ] we present selected elements of nspt which are relevant for the case at hand , in particular the issue of performing a nspt computation in a fixed gauge , _ i.e. _ the minimal landau gauge .
[ sec : ghostpropagator ] focuses on the observables we will be concerned with , while sect .
[ sec : standard - lpt ] summarizes known analytic results for the perturbative ghost propagator up to three - loop order and their relation to the unknown non - logarithmic coefficients we want to determine . in the following section we present the technical details of practicing nspt .
finally we collect our results in sect .
[ sec : results ] and briefly discuss a comparison to monte carlo data .
the basis of stochastic quantization is the ( stochastic ) time evolution of fields driven by the langevin equation . in the case of a @xmath2 lattice gauge theory
the langevin equation for the @xmath2 matrix @xmath3 located at the link between lattice sites @xmath4 and @xmath5 reads @xmath6- \eta_{x,\mu}(t ) \bigr ) \ ; u_{x,\mu}(t;\eta ) \ , , \label{eq : langevin_i}\ ] ]
@xmath7 is the ( stochastic ) langevin time .
the so - called drift term is given by the equations of motion derived from the euclidean gauge action @xmath8 $ ] : they are written in terms of the differential operator @xmath9 , which is a left lie derivative . the @xmath10 ( @xmath11 )
are the generators in the fundamental representation . in our case we will have @xmath12 and
the gauge action @xmath8 $ ] is the wilson plaquette action .
the key issue of stochastic quantization is that in the limit of large @xmath7 expectation values with respect to the noise converge to expectation values with respect to the qft functional integral , _
i.e. _ the distribution of the gauge fields converges to the gibbs measure @xmath13 \propto { \rm{exp}}(- s_g[u])$ ] . in any numerical approach
( and this applies to our case ) , however , the langevin equation is not solved for continuous time .
one needs to discretize time as a sequence @xmath14 , with running step number @xmath15 .
therefore , in order to extract correct physical information , one needs to perform the double extrapolation @xmath16 and @xmath17 . for the numerical solution of the langevin equation we adhere to a peculiar version of the euler scheme that guarantees all the link matrices
@xmath18 to stay in the group manifold : @xmath19 \bigr ) \ ; u_{x,\mu}(n ; \eta)\ , \label{eq : iteration } \\
f_{x,\mu}[u , \eta ] & = & \epsilon~\!\nabla_{x,\mu } s_g[u ] + \sqrt{\epsilon}~\eta_{x,\mu } \ , .
\label{eq : force}\end{aligned}\ ] ] in matrix form and at finite @xmath20 , the components of @xmath21 are subject to the mutual correlations ( @xmath22 ) @xmath23\delta_{zw } \ , .
\label{eq : matrixform}\ ] ] @xmath24 and @xmath25 ( integer ) denote here langevin time steps .
the noise is white , meaning that different time steps are uncorrelated .
the usual path from stochastic quantization to stochastic perturbation theory would go through trading the differential form of langevin equation for its integral equation counterpart in a green function formalism , which is easy to solve in a perturbative scheme .
nspt remains instead in a differential formulation .
for the case of lattice gauge theories we rewrite each link matrix as an expansion in the bare coupling constant @xmath26 . since @xmath27 ,
the expansion reads @xmath28 provided one rescales the time step to @xmath29 , the expansion converts the langevin equation ( [ eq : iteration ] ) into a system of simultaneous updates in terms of the expansion coefficients ( [ eq : expansion_of_u ] ) of @xmath3 and of similar expansion coefficients for the force @xmath30 in eq .
( [ eq : force ] ) . while the random noise @xmath21 enters only the lowest order equation , higher orders are stochastic by the noise propagating from lower order to higher order terms .
the system can be truncated according to the maximal order of the gauge field one is interested in . in a notation which is common in lattice perturbation
theory we can relate the ( antihermitean ) vector potential @xmath31 to the link matrices @xmath32 via @xmath33 an expansion similar to ( [ eq : expansion_of_u ] ) exists for that potential taking values in the algebra @xmath34 , @xmath35 we observe that unitarity of @xmath3 has a counterpart in the antihermiticity and tracelessness of all orders of @xmath31 : @xmath36 as we will see , the definition of the ghost propagator will in a sense be based on the expansion in the algebra .
we finally remark that whenever we speak about contributions of some order to an observable this has to be understood as an expansion @xmath37 and constructed by comparison to coefficients of @xmath38 .
we now come to the main point of how to perform a gauge - fixed nspt computation , in the case at hand in landau gauge . a key issue of the standard ( in other words , non - perturbative )
langevin equation for lgt is the non - necessity of fixing a gauge . in short : the drift term in langevin equation is equal to the lattice equations of motion ( put equal zero there ) . as a consequence ,
longitudinal degrees of freedom undergo an unbounded diffusion , but the group manifold is compact and hence no divergence shows up .
such a situation is no longer true in nspt .
this is quite obvious in the algebra : the diffusion of the longitudinal component of the @xmath39 fields is unbounded and hence their norms diverge in stochastic time .
we stress that gauge - invariant quantities are in principle not affected by these divergences , but this is no advantage : since norms are diverging one runs eventually into trouble by lack of accuracy .
we also remark that this is true regardless of the choice of either expansion ( [ eq : expansion_of_a ] ) or ( [ eq : expansion_of_u ] ) : perturbation theory implies decompactification in any case .
a solution comes along the lines suggested in @xcite : _ fixing _ ( to some precision ) _ the gauge is going to fix the problem _ and this goes under the name of stochastic gauge fixing .
we adhere here to the recipe of @xcite .
the langevin evolution steps are interwoven with gauge transformations @xmath40 the main requirement for @xmath41 is quite obvious : gauge transformations should be expanded in the coupling as well and the leading order should not affect the background vacuum around which the solution to langevin equation is expanded ( the trivial vacuum , in the case at hand ) . a convenient solution comes with the choice @xmath42 in general the action of the left / right lattice derivative @xmath43 on a function @xmath44 defined on site @xmath4 is given by @xmath45 the action of the left lattice derivative on the vector potential defined on midpoints of the links is defined as @xmath46 the resulting divergence is mapped to site @xmath4 .
we will come back to this choice : the iteration of this gauge transformation would drive the field configuration to the landau gauge . with respect to stochastic gauge fixing
, it suffices here to point out that one step of ( [ eq : gtrafo ] ) using ( [ eq : localstep ] ) interwoven with the langevin step ( [ eq : iteration ] ) is enough to keep fluctuations under control provided that @xmath47 is chosen of order @xmath48 .
the rationale for this is the essence of stochastic gauge fixing : any new drift term in the langevin equation which amounts to a local gauge transformation does not affect the asymptotic distribution of gauge invariant quantities , while it can provide a restoring force in the non - gauge - invariant sector .
the differential ( finite difference , actually ) operator present in ( [ eq : localstep ] ) makes it clear that zero modes are completely unaffected up to this point .
their divergences are still present .
we take the simplest prescription of subtracting zero modes at every order .
one can actually do more than this and perform a gauge - fixed nspt computation in the case of covariant gauges .
the interested reader is referred to @xcite .
the key issue is that one can rewrite the functional integral in the faddeev - popov formalism , which merely results in a different form of the gauge action ( an effective action , one would say ) . without entering into details , in this context
it is worth making two points .
first , there is no need for the introduction of ghost fields : one simply faces an extra contribution to the action in the form @xmath49 , @xmath50 being the faddeev - popov operator we will be concerned with in the following .
second , for the landau gauge the procedure is of little help , since the other extra piece in the action is the gauge fixing action @xmath51 , with @xmath52 being the covariant gauge parameter , which is zero in this case .
the failure of the general covariant gauge nspt procedure has anyway a counterpart : as a limit case @xmath53 of covariant gauges , the landau gauge @xmath54 is related to an extremum of the norm of the field , and there is a well - known iterative procedure to reach such an extremum .
the basic step is ( [ eq : localstep ] ) , _ i.e. _ the gauge transformation that we also use for stochastic gauge fixing .
the iteration stops when , for all orders @xmath55 , the summed - up quadratic violations of ( [ eq : landaucond ] ) vanish within double precision ( @xmath56 is the lattice volume ) , @xmath57 = 0 \ , .
\label{eq : condition}\ ] ] iterating ( [ eq : gtrafo ] ) with ( [ eq : localstep ] ) is actually not very efficient , since it suffers from critical slowing down due to the low modes , which becomes more and more severe as the lattice size increases .
an effective solution to this problem is also well - known and goes through fourier acceleration as introduced in @xmath58 @xcite . in nspt
one makes use of the expanded variant of the gauge transformation @xmath59(p)\right ] \right\ } \label{eq : fourieracc}\ ] ] in order to achieve the minimal landau gauge .
@xmath60 denotes the fourier transformation from discrete @xmath4 to momentum @xmath61 space and @xmath62 its inverse .
the gauge transformation ( [ eq : fourieracc ] ) applied locally is actually not local in the sense of ( [ eq : localstep ] ) .
the driving force , the divergence of the gauge potential , is non - locally smeared with a low - pass filter that attenuates the higher fourier components of @xmath63 by a factor @xmath64 . here
@xmath65 is a non - zero eigenvalue of the free field lattice laplacian , and its maximal value on a periodic lattice in all directions is @xmath66 .
while we notice that the volume - independent optimal parameter @xmath47 for the case of periodic boundary conditions in all directions is found for this perturbative landau gauge fixing to be @xmath67 , we also point out that fourier acceleration also works if one takes the plain lattice momenta ( which in the following we will reference as @xmath68 ) instead of @xmath69 .
this is a good point to pin down some notations for lattice momenta . in the following
we will reserve the latin letter @xmath70 to denote @xmath71 , the integer 4-tuples @xmath72 , while plain lattice momenta will be denoted by @xmath73 @xmath74 here @xmath75 is the lattice size in direction @xmath76 and the reference to lattice spacing @xmath77 can be made explicit : @xmath78 . in our case lattices
will be symmetric , _
i.e. _ @xmath79 and @xmath80 . according to the two most popular conventions , @xmath81 $ ] or @xmath82 $ ] .
the eigenvalues of the free laplacian are @xmath83 ^ 2 \ , , \ ] ] where for each direction separately @xmath84
the ghost propagator @xmath41 is nothing but the inverse of the faddeev - popov operator @xmath50 .
we start our discussion in the continuum , where in the case of covariant gauges the faddeev - popov ( fp ) operator reads @xmath85 @xmath86 is the covariant derivative , emerging from the response of the gauge field @xmath39 to an infinitesimal gauge transformation .
@xmath41 is diagonal in the ( adjoint ) color space ( @xmath87 ) @xmath88 as an average over field configurations @xmath89 , it is translationally invariant @xmath90 in landau gauge , because @xmath91 , we can reverse the order of @xmath92 and @xmath93 , @xmath94 to obtain the form of the fp operator on the lattice one closely mimics the steps one goes through in the continuum .
the lattice form of that operator is nevertheless much more involved , since the response of the lie - algebra @xmath39 field to a gauge transformation does not have a closed form . in perturbation theory
there is no logical obstruction , anyway : at any given order the lattice counterpart of the covariant derivative can be pinned down @xcite and all in all the fp operator reads @xmath95_{ab } \delta_{x , y } \ , .
\label{eq : fpop}\ ] ] here we denote by @xmath96 the lattice counterpart of the covariant derivative , indicating a dependence on @xmath97 , even if this dependence is through a perturbative expansion .
this results in a dependence on the adjoint representation of the lie - algebra field @xmath98= \left[1 + \frac{\rm i}{2 } \phi_\mu(x ) - \frac{1}{12 } \phi_\mu^2(x ) - \frac{1}{720 } \phi_\mu^4(x ) + \frac{1}{30240 } \phi_\mu^6(x ) + \dots\right ] \partial_\mu^r + { \rm i}\,\phi_\mu(x ) \ , .\ ] ] the @xmath99 is the adjoint representation of the lattice gauge field @xmath100 ( @xmath101 , @xmath102 antihermitean ) @xmath103_{bc } = \left[t^a\right]_{bc } \phi_\mu^a\equiv -{\rm i } \ ,
f^{abc } \phi_\mu^a \ , , \ ] ] the fundamental representation being given by ( @xmath104 ) : @xmath105_{ij } = \left[t^a\right]_{ij } \phi_\mu^a\ , , \quad \left[a_\mu(x)\right]_{ij } = \left[t^a\right]_{ij } a_\mu^a \ , .\ ] ] the ghost propagator in momentum space can in principle be obtained by fourier transformation in each lattice configuration @xmath106 , taking the translational invariance into account : @xmath107^{ab}_{xy } \ ; \exp\left({\rm i } \;p(k)\cdot(x - y)\right ) \ ; \right\rangle_{\{u\ } } = \left\langle [ m^{-1}]^{ab}(p(k ) ) \right\rangle_{\{u\ } } \label{eq : pfmom}\ , \ ] ] with the scalar product @xmath108 . given the color - diagonal form , one is interested in the scalar function defined by the adjoint trace @xmath109 due to
its eight trivial zero eigenvalues ( corresponding to global gauge transformations ) the inverse @xmath110 needs to be defined with care .
since one is interested in non - zero momenta @xmath111 , this means that one is automatically restricted to a subspace orthogonal to the space spanned by the ( space - time constant ) zero modes of the faddeev - popov operator . for an individual configuration ,
the inverse of the fp operator is not translationally invariant and hence ( [ eq : pfmom ] ) is not so useful .
the way out is quite standard and goes through the inversion of @xmath50 on a delta source in momentum space , _ i.e. _ a plane wave source @xmath112 in real space : @xmath113^{ab } = \sum_{cd , xy } \psi(a , k)^{*,c}_x \ , [ m^{-1}]_{xy}^{cd } \ , \psi(b , k)^d_y \,.\ ] ] as in standard non - perturbative applications , the issue is inverting the relevant operator on a source , but things are actually a lot easier in nspt . the matrix @xmath50 can simply be considered as expanded using the expansion of the gauge fields @xmath114 : @xmath115 then the inverse of @xmath50 will also obtained as a power series @xmath116^{(0 ) } + \sum_{l>0 } \beta^{-l/2 } [ m^{-1}]^{(l ) } \ , .\ ] ] one easily finds @xcite a recursive way to obtain the higher orders of @xmath110 : @xmath117^{(0)}&=&[m^{(0)}]^{-1 } \
, , \\ \nonumber [ m^{-1}]^{(l)}&= & - [ m^{-1}]^{(0)}\;\sum_{j=0}^{l-1 } m^{(l - j)}\;[m^{-1}]^{(j ) } \ , .\end{aligned}\ ] ] it is worth pointing out that the computation proceeds back and forth from position to momentum space , since at any step the operators at hand are diagonal ( or almost diagonal ) in one of the two spaces .
this is the time to reintroduce perturbative ( loop ) indices : we denote the @xmath118 loop order of the ghost propagator as @xmath119\right\rangle^{(l=2n ) } \ , .\ ] ] thus a real loop order @xmath15 of the ghost propagator corresponds to contributions with even powers @xmath120 , whereas contributions with odd powers @xmath121 would indicate a non - loop contribution which does not exist by construction in standard perturbation theory .
contributions related to these non - integer @xmath15 have to vanish numerically in nspt for sufficiently large statistics .
similar to the gluon propagator in the forthcoming paper , we present the various orders of the ghost dressing function ( or `` form factor '' ) in two forms : @xmath122 the dressing function has to be calculated for each 4-momentum given by four integers the momentum 4-tuple @xmath72 .
two comments are in order for ( [ eq : dressingfunction ] ) .
first of all , the @xmath123 s are dimensionless quantities .
notice that in a numerical lattice simulation one does not have access to @xmath124 , but to the dimensionless quantity @xmath125 .
multiplying by @xmath126 - or @xmath127 - one obtains the corresponding ( still dimensionless ) @xmath128 - or @xmath129 - we will be concerned with in the following .
second , the second notation in ( [ eq : dressingfunction ] ) is chosen to remind that here the quantities @xmath130 are acting as the `` physical '' momenta .
in the ri-mom scheme , the renormalized ghost dressing function @xmath131 is defined as @xmath132 with the condition @xmath133 therefore , the ghost dressing function @xmath134 is just the ghost wave function renormalization constant @xmath135 at @xmath136 . the general form of an operator / function @xmath137 in terms of the renormalized coupling @xmath138 compared to the `` standard '' definition .
] can be written as @xmath139 here @xmath140 gives the possible powers of the logarithms in loop order @xmath141 . using ( [ eq : renorm ] ) and ( [ eq : ricondition ] ) , we get the expansions @xmath142 since we measure the dressing function as an expansion in the lattice ( inverse ) coupling @xmath143 , we need the ghost wave function renormalization as an expansion in the lattice coupling @xmath144 : @xmath145 from now we restrict ourselves to three - loop expressions in landau gauge and in the quenched approximation .
the logarithmic corrections are partly known from the anomalous dimension of the ghost field and the beta function and given as follows ( compare _ e.g. _ @xcite ) @xmath146 \ , n_c^3 + \frac{91}{8 } \ , z^{\rm ri'}_{1,0}\ , n_c^2 + \frac{79}{6 } \ , z^{\rm ri'}_{2,0}\ , n_c\ , .
\nonumber\end{aligned}\ ] ] the finite one - loop constant @xmath147 has been calculated from the ghost self - energy in standard infinite volume lpt @xcite with the value @xmath148 to our knowledge , the constants @xmath149 and @xmath150 were not known so far . taking into account the relation between the lattice coupling and the renormalized coupling in two - loop accuracy @xcite ( the function @xmath151 is scheme independent ) @xmath152 with @xmath153 we get the coefficients @xmath154 of @xmath155 in ( [ zcbare ] ) up to three loops @xmath156 using these numbers and the inverse lattice coupling @xmath143 we obtain the expression for the ghost dressing function to that order @xmath157 with the one - loop coefficients @xmath158 the two - loop coefficients @xmath159 and the three - loop coefficients @xmath160 thus we conclude : the coefficients of the leading logarithmic contributions for a given order can be exclusively taken from continuum computations whereas non - leading log coefficients are already influenced by finite lattice constants from corresponding lower loop orders . in general , these logarithmic terms are the relatively easy part of a diagrammatic perturbative computation , while getting the finite constants is the big effort . in nspt
it is just the other way around . as it will be clear in section [ subsec : fitting ] , telling which is which in the measured ghost dressing function at each loop requires a fitting procedure : fitting a logarithmic term would actually require a terrific precision . as a consequence ,
we have just stated from the very beginning the ( modest ) attitude of not even trying to extract anomalous dimensions from our nspt computations . on the contrary , we have to assume that the coefficients in front of the logarithms are known . assuming that a continuum calculation has been performed up to a given loop order
defines the maximum accuracy and the maximal loop number which we can reach in a lattice nspt computation .
we have to check that nspt in the limits @xmath161 ( or @xmath162 ) and @xmath163 reproduces the constant @xmath164 from the one - loop result .
after that we may proceed to determine from the two - loop measurements the unknown @xmath165 and , therefore , @xmath149 . knowing that number , we are in the position to fix the coefficients of all non - leading logarithmic contributions in the three - loop expression and to estimate the constant @xmath166 or @xmath150 from the three - loop measurements .
choosing a maximal loop expansion order @xmath167 and solving the coupled system of equations for the gauge links @xmath168 and the gauge potentials @xmath169 simultaneously , one generates a sequence of these fields for all perturbative orders @xmath170 .
such sequences representing langevin time trajectories are generated for different finite langevin time steps @xmath171 and different lattices sizes , _ i.e. _ different @xmath172 . from the perturbative fields at finite @xmath171 we construct the observable of interest , in our case the ghost dressing function for
each chosen set of lattice momenta and average it over the used configurations within a langevin time trajectory .
it is expected that the autocorrelation time for the observable is extending over subsequent configurations and increases with decreasing @xmath171 . as a reasonable compromise between computer time and autocorrelation
we have measured the ghost propagator using a sequence of configurations separated by 50 langevin steps . before measuring
, landau gauge has to be obtained on each configuration by satisfying ( [ eq : condition ] ) to high precision . to further increase the statistics , we have generated several different trajectories for a given @xmath171 and @xmath172 .
since the ghost propagator has to be calculated for each individual momentum 4-tuple and the number of 4-tuples rapidly increases with the lattice size we have restricted ourselves to a certain set of momentum 4-tuples for our periodic symmetric lattice .
we have chosen a cylindrical cut similar to monte carlo measurements such that the individual momentum components @xmath173 do not differ too strongly from each other .
the direction of that 4-vector @xmath174 is chosen nearly along a lattice 4-volume diagonal . in a 4-tuple
we have chosen the ordered integers @xmath175 modulo lattice size not to differ by more than one unit and have assumed a maximal difference @xmath176 . during the langevin evolution we treated all 4-tuples leading to non - identical measurements as independent ones .
this led in our particular realization to @xmath177 4-tuples for which we have measured the propagator . after obtaining the langevin time averages we further averaged the propagator over 4-tuples which were equivalent due to lattice symmetries and ended up with the following set of inequivalent momentum 4-tuples for our lattices chosen with even site numbers used in the further analysis : @xmath178 choosing as example @xmath179 , out of originally chosen 108 4-tuples we get 25 inequivalent 4-tuples for which the propagator has been measured .
a final average has been performed on the results from different trajectories . in tables [ tab : statistics1 ] and [ tab : statistics2 ] .number of one - loop and two - loop ghost propagator measurements for individual momentum 4-tuples ( [ eq : tuples ] ) at different lattices sizes done with the leipzig nspt code at leipzig pc clusters . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] the average yields @xmath180 which is supported by the figure .
a major source of indetermination in our three loop analysis comes from the extrapolation @xmath181 . for three sizes ( @xmath182 )
we only have four values of @xmath171 . the number given above is obtained from a quadratic fit @xmath181 used for one- and two - loop cases as well .
computing @xmath166 from a linear @xmath171-fit and using otherwise the same strategy we obtain @xmath183 which differs about @xmath184 from the quadratic extrapolation . on the contrary ,
the one- and two - loop constants are practically not affected by the two extrapolations . for the linear extrapolation
we get @xmath185 and @xmath186 . collecting all results we can write ( [ zghostbeta3loop ] ) in a numerical form @xmath187 in ( [ j3loopnum ] ) the coefficients @xmath188 are given with their significant digits only . formulas ( [ j2loop ] ) and ( [ j3loop ] ) encode the relation to the @xmath149 and @xmath150 coefficients . due to numerical accidents ,
these are affected by relevant percental errors .
using the @xmath189 definition ( [ eq : log_mapping ] ) as in nspt in the gluon propagator and the faddeev - popov operator ( [ eq : fpop ] ) for the calculation of the ghost propagator , the berlin humboldt university group has obtained monte carlo results for the two propagators in landau gauge and the resulting running gauge coupling @xcite .
maximizing the `` linear '' gauge fixing functional @xmath190 as a preconditioner , the final gauge fixing was organized to minimize @xmath191 \label{eq : fullcondition}\ ] ] with an iterative scheme following ( [ eq : localstep ] ) , actually in a parallelized multigrid realization of fourier acceleration ( [ eq : fourieracc ] ) . since it is assumed that non - perturbative contributions dominate mainly the intermediate and infrared momentum range , it is of interest here to compare directly the perturbative ghost dressing function obtained in nspt with its monte carlo counterpart for each common momentum 4-tuple .
we calculate the perturbative dressing function at a given lattice volume summed up to loop order @xmath192 for a given lattice coupling @xmath143 as follows : @xmath193 in figure [ fig:5 ] values . ]
we compare the perturbative ghost dressing function at lattice size @xmath194 with monte carlo data at two different @xmath143 values .
we observe that at least three - loop accuracy is necessary to guarantee that the perturbative ghost propagator at larger @xmath143 ( @xmath195 ) approximately describes the full two - point function in the large momentum squared region ( @xmath196 ) .
we presented two- and three - loop results for the ghost propagator of lattice su(3 ) in landau gauge .
a careful analysis of finite volume and finite lattice size effects has been performed , whose methodology has been discussed quite in detail . the present work will be supplemented by a similar analysis for the gluon propagator ( in the same gauge ) .
this will open the way to a more careful comparison to monte carlo data for these quantities , which are supposed to encode informations on the confinement mechanism of non - abelian gauge theories not only in the asymptotic infrared region , but also in the intermediate momentum range in the form of power corrections ( condensates ) @xcite and contributions from non - perturbative excitations ( vortices etc . )
@xcite related to particular length scales .
this work is supported by dfg under contract schi 422/8 - 1 , dfg
sfb / tr 55 and by i.n.f.n . under the research project mi11 .
we acknowledge computer time made available to us by ect@xmath197 on the ben system for approximately 13000 cpuh .
j. greensite , prog .
phys . * 51 * , 1 ( 2003 ) [ http://arxiv.org/abs/hep-lat/0301023[arxiv:hep-lat/0301023 ] ] .
r. alkofer and j. greensite , j. phys .
g 34 , s3 ( 2007 ) [ http://arxiv.org/abs/hep-ph/0610365[arxiv:hep-ph/0610365 ] ] .
j. greensite , eur .
j. st * 140*,1 ( 2007 ) .
l. von smekal , [ http://arxiv.org/abs/0812.0654[arxiv:0812.0654[hep-th ] ] ] . c. s. fischer , a. maas , and j. m. pawlowski , pos * confinement8 * , 043 ( 2008 ) , [ http://arxiv.org/abs/0812.2745[arxiv:0812.2745[hep-ph ] ] ] .
i. l. bogolubsky , e .-
ilgenfritz , m. mller - preussker , and a. sternbeck , phys .
* b676 * , 69 ( 2009 ) , [ http://arxiv.org/abs/0901.0736[arxiv:0901.0736[hep-lat ] ] ] . c. s. fischer , a. maas , and j. m. pawlowski , annals phys .
* 324 * , 2408 ( 2009 ) , [ http://arxiv.org/abs/0810.1987[arxiv:0810.1987[hep-th ] ] ] .
v. n. gribov , nucl .
bf b139 , 1 ( 1978 ) .
d. zwanziger , nucl .
b364 127 ( 1991 ) .
i. ojima , nucl .
b143 340 ( 1978 ) .
t. kugo and i. ojima , prog .
suppl . * 66 * , 1 ( 1979 ) .
l. von smekal , a. jorkowski , d. mehta , and a. sternbeck , pos * confinement8 * , 048 ( 2008 ) , [ http://arxiv.org/abs/0812.2992[arxiv:0812.2992[hep-th ] ] ] .
boucaud , f. de soto , j. p. leroy , a. le yaouanc , j. micheli , o. pene and j. rodriguez - quintero , phys .
d * 79 * 014508 ( 2009 ) [ http://arxiv.org/abs/0811.2059[arxiv:0811.2059[hep-ph ] ] ] .
k. langfeld , h. reinhardt , and j. gattnar , nucl . phys .
* b 621 * , 131 ( 2002 ) [ http://arxiv.org/abs/hep-ph/0107141[arxiv:hep-ph/0107141 ] ]
. j. gattnar , k. langfeld , and h. reinhardt , phys .
lett . * 93 * , 061601 ( 2004 ) [ http://arxiv.org/abs/hep-lat/0403011[arxiv:hep-lat/0403011 ] ] . for a comprehensive review
see s. capitani , phys .
rept . * 382 * , 113 ( 2003 ) [ http://arxiv.org/abs/hep-lat/0211036[arxiv:hep-lat/0211036 ] ] .
f. di renzo , e. onofri , g. marchesini and p. marenzoni , nucl .
b * 426 * , 675 ( 1994 ) [ http://arxiv.org/abs/hep-lat/9405019[arxiv:hep-lat/9405019 ] ] .
g. parisi and y. s. wu , sci .
* 24 * , 483 ( 1981 ) .
g. g. batrouni , g. r. katz , a. s. kronfeld , g. p. lepage , b. svetitsky and k. g. wilson , phys .
d * 32 * , 2736 ( 1985 ) .
f. di renzo and l. scorzato , jhep * 0410 * , 073 ( 2004 ) [ http://arxiv.org/abs/hep-lat/0410010[arxiv:hep-lat/0410010 ] ] .
ilgenfritz , h. perlt and a. schiller , pos * lattice 2007 * , 251 ( 2007 ) [ http://arxiv.org/abs/0710.0560[arxiv:0710.0560[hep-lat ] ] ] . f. di renzo , l. scorzato and c. torrero , pos * lattice 2007 * , 240 ( 2007 ) [ http://arxiv.org/abs/0710.0552[arxiv:0710.0552[hep-lat ] ] ] .
f. di renzo , e .- m .
ilgenfritz , h. perlt , a. schiller and c. torrero , pos * lattice 2008 * , 217 ( 2008 ) [ http://arxiv.org/abs/0809.4950[arxiv:0809.4950[hep-lat ] ] ] .
f. di renzo , e .- m .
ilgenfritz , h. perlt , a. schiller and c. torrero , pos * confinement8 * 050 ( 2009 ) [ http://arxiv.org/abs/0812.3307[arxiv:0812.3307[hep-lat ] ] ] .
f. di renzo , e .- m .
ilgenfritz , h. perlt , a. schiller and c. torrero , http://arxiv.org/abs/0910.2905[arxiv:0910.2905[hep-lat ] ] .
d. zwanziger , nucl .
b * 192 * , 259 ( 1981 ) . c. torrero , m. laine , y. schroder , f. di renzo and v. miccio , pos * lat2006 * , 038 ( 2006 ) [ http://arxiv.org/abs/hep-lat/0609048[arxiv:hep-lat/0609048 ] ] .
c. t. h. davies _ et al .
_ , phys .
d * 37 * , 1581 ( 1988 ) .
h. j. rothe , lattice gauge theories - an introduction ( 2nd edition ) , world scientific , singapore ( 1997 ) .
j. a. gracey , nucl .
b * 662 * ( 2003 ) 247 [ http://arxiv.org/abs/hep-ph/0304113[arxiv:hep-ph/0304113 ] ] .
h. kawai , r. nakayama and k. seo , nucl . phys .
b * 189 * 40 ( 1981 ) .
a. hasenfratz and p. hasenfratz , phys .
b * 93 * 165 ( 1980 ) .
m. lscher and p. weisz , nucl .
b * 452 * 234 ( 1995 ) [ http://arxiv.org/abs/hep-lat/9505011[arxiv:hep-lat/9505011 ] ] . c. christou , a. feo , h. panagopoulos and e. vicari , nucl .
b * 525 * , 387 ( 1998 ) [ erratum - ibid .
b * 608 * , 479 ( 2001 ) ] [ http://arxiv.org/abs/hep-lat/9801007[arxiv:hep-lat/9801007 ] ] .
f. di renzo , v. miccio , l. scorzato and c. torrero , eur .
j. c * 51 * , 645 ( 2007 ) [ http://arxiv.org/abs/hep-lat/0611013[arxiv:hep-lat/0611013 ] ] .
m. constantinou , v. lubicz , h. panagopoulos and f. stylianou , jhep * 0910 * 064 ( 2009 ) http://arxiv.org/abs/0907.0381[arxiv:0907.0381[hep-lat ] ] . c. menz , diploma thesis , humboldt - universitt zu berlin ( 2009 ) ; we acknowledge receiving those data prior to publication . | this is the first of a series of two papers on the perturbative computation of the ghost and gluon propagators in @xmath0 lattice gauge theory .
our final aim is to eventually compare with results from lattice simulations in order to enlight the genuinely non - perturbative content of the latter . by means of numerical stochastic perturbation theory
we compute the ghost propagator in landau gauge up to three loops .
we present results in the infinite volume and @xmath1 limits , based on a general strategy that we discuss in detail . |
the pseudogap phase has been widely regarded as an essential integral part of the cuprate superconductors with extensive experimental support.timusk the underlying physics of such a phase has been a central focus in the study of high-@xmath0 problem for many years , and yet no consensus has been reached concerning its nature due to the very complexity of pseudogap phenomena observed in magnetic , transport , single - particle , and optical channels .
more and more experimental evidence in recent years further indicates the existence of two kinds of pseudogap regimes at different temperatures.@xcite among various theoretical proposals for the pseudogap phase , the rvb idea @xcite is uniquely interesting , which actually predicted @xcite the existence of a pseudogap state in doped mott insulators before experiment . in the rvb picture ,
neutral spins form the singlet pairs that are condensed as a spin liquid .
the density of states of spin excitations for such a system generally gets suppressed at low energy , exhibiting a pseudogap feature as it costs energy to break up the rvb pairs to create spin excitations . in such a scenario , there is usually no gap in the pure charge degrees of freedom , and the pseudogap phenomena observed in the charge transport , angle resolved photoemission spectroscopy ( arpes ) , and tunnelling experiments , are all indirectly attributed to the appearance of the _ spin _ gap in the spin degrees of freedom.@xcite for instance , in the charge transport the strong scattering between the charge carriers and low - lying spin fluctuations becomes weakened because of the reduction of the latter in the pseudogap regime ; the pseudogap feature exhibited in the arpes and tunnelling measurements may be also interpreted as due to the opening of a pseudogap associated with the spin degrees of freedom .
however , the original rvb description , known as the fermionic rvb ( f - rvb ) since it involves the pairing of _ fermionic _ spinons ( neutral @xmath7 object),@xcite also suffers some notable inconsistency with the experiment . note that the pseudogap phenomenon has been found in the underdoped regime of the cuprates @xcite where the antiferromagnetic ( af ) correlations are usually quite strong . but in an f - rvb description , the af correlations remain intrinsically weak , even at half - filling , where the af long range order ( aflro ) develops in the cuprates at low temperature . here the key issue is not about whether one can construct an aflro in the rvb background , which may be easy to incorporate by a simple mean - field order parameter.hsu,muthu but the crucial and highly nontrivial issue is whether the _ whole _ low - lying af spin correlations are intrinsically and sufficiently strong in an rvb state @xcite , and whether they are capable of continuously _ growing _ with reducing temperature or doping as has been clearly manifested experimentally in , say , the nmr spin - lattice relaxation rates ( see the analysis in refs .
@xcite ) .
since the pseudogap phase , which involves high - energy / temperature and short distance physics , may be properly considered as an unstable fixed point state @xcite with intrinsic instabilities towards aflro or d - wave superconductivity at low temperatures , the importance of its correct description , with regard to the latter , is like that of a fermi liquid with regard to the bcs superconducting state .
in other words , finding an accurate and correct description of the pseudogap phase will be rather important for constructing a sensible low - energy theory for describing the low - temperature af and superconducting phases in the cuprates . in this paper
, we show that there does exist a desirable candidate for characterizing the pseudogap phase based on the rvb picture , in which strong af correlations are present as an _ intrinsic _ and _ predominant _ feature .
such an rvb state , known as the bosonic rvb ( b - rvb ) state , psmfd differs from the usual f - rvb states by that it works very well at half - filling in describing the af correlations over a wide range of temperature , including zero temperature where an aflro naturally emerges .
[ fp ] schematically illustrates the global phase diagram for such a b - rvb theory,@xcite where an _ upper _ pseudogap phase ( upp ) below the characteristic temperature @xmath6 is characterized by the formation of singlet pairing of neutral spins as denoted by the b - rvb order parameter @xmath8 as illustrated by fig .
1 , the low - temperature phases , including a _ lower _ pseudogap phase below @xmath5 ( also known as the spontaneous vortex phase @xcite ) , a d - wave superconducting phase at @xmath9 beyond a critical doping concentration @xmath10 , and an aflro phase near half - filling , can all be regarded as the results of the low - temperature instabilities from such an upp.@xcite at @xmath11 whose properties are the main focus of this paper .
the antiferromagnetic ordered phase ( af ) , the lower pseudogap phase at @xmath12 , and the superconducting phase ( sc ) at @xmath13 all happen on top of this upp at low doping.,width=288 ] the main focus of the present work will be the nature of the upp itself .
we will examine the detailed behavior of uniform spin susceptibility , spin - lattice relaxation rates , and spin - echo decay rate based on the b - rvb theory , which will reveal that , as one enters the upp from above @xmath6 , the spin correlations change qualitatively . at the mean - field level
, localized spins are essentially uncorrelated at @xmath14 , resulting in a curie - weiss - like behavior in the uniform spin susceptibility .
below @xmath6 , however , finite - range spin correlations start to develop , predominantly around the af momentum @xmath15 , with the weight being transferred from the momentum @xmath16 .
the latter effect leads to the reduction of uniform susceptibility at @xmath17 .
we show that such magnetic behavior in the upp is quite consistent with the experimental measurements in the cuprates .
the above results clearly indicate that the upp corresponds to a crossover from a weakly correlated localized spin system at higher temperature into a strongly _ af _
correlated spin liquid at lower temperature .
this picture is thus in sharp contrast to the f - rvb picture of the pseudogap , where the low - energy spin excitations , either around @xmath18 or @xmath19 , _ all _ get suppressed with the opening of the pseudogap .
this latter behavior rather resembles the lower pseudogap phase of the b - rvb theory at @xmath12 ( in ref .
@xcite it is simply called the pseudogap phase ) than the upp .
but even the distinction between the pseudogap state of the f - rvb theory and such a lower pseudogap state of the b - rvb theory is very significant as has been discussed in ref .
@xcite : the former is exchange energy driven while the latter is kinetic energy driven . here
the lower pseudogap phase corresponds to the formation of cooper pair amplitude but is short of superconducting phase coherence , which exists between @xmath5 and @xmath0 and can be regarded as a vortex liquid state.@xcite it was previously pointed out in ref .
@xcite that such a spontaneous vortex phase should coincide with the experimentally discovered nernst region @xcite in the high-@xmath0 cuprates .
the quantitative phase diagram of the upp is determined in the three - dimensional parameter space of temperature , doping concentration , and external magnetic field .
the latter introduces the competition between the zeeman spin splitting and singlet spin pairing .
the mean - field theory will predict a simple proportional relation between the critical magnetic field @xmath20 , at which the upp is destroyed , and @xmath6 in zero field . a comparison with experiment will be made .
the remainder of the paper is arranged as follows . in sec .
ii a , the bosonic rvb description is briefly reviewed . in sec .
ii b , the definition of the upper pseudogap phase is given and its phase diagram is determined . in sec .
ii c and d , its magnetic properties are systematically investigated in the framework of the b - rvb theory . in sec .
ii e , we further briefly discuss the lower pseudogap phase and related magnetic behavior .
finally , sec .
iii is devoted to conclusion and discussion , where we also briefly discuss the qualitative behavior of charge channels in the pseudogap phase within the bosonic rvb description .
the so - called b - rvb state@xcite is underpinned by a _
bosonic rvb order parameter _
@xmath21 over a wide range of temperature and doping as schematically shown in fig .
1 , which defines the upp .
this upp ( before the emergence of low - temperature af and superconducting instabilities ) will be the main subject to be examined in the present work . in the following
we shall first discuss its mean - field description based on the @xmath22 model . in the phase string representation @xcite of the @xmath22 model ( see appendix a ) , a natural mean - fielddecoupling of the superexchange term @xmath23 is given as follows:@xcite @xmath24where the b - rvb order parameter @xmath21 is defined in terms of the bosonic spinon annihilation operator @xmath25 by @xmath26at half - filling , @xmath21 is equivalent to the usual schwinger - boson mean - field order parameter as @xmath27 with eq.(hs ) reducing to the schwinger - boson mean - field hamiltonian @xcite which describes the af correlations fairly well in the regime of @xmath28 at @xmath29 .
away from half - filling , a topological link field @xmath30 emerges in the hamiltonian ( [ hs ] ) as well as in eq.([parameter ] ) , which represents the influence of the nonlocal phase string effect induced by the hole hopping @xcite on the spin degrees of freedom .
it is related to the hole density by the following gauge invariant relation:@xcite @xmath31where @xmath32 denotes an arbitrary loop on the square lattice that encloses a region @xmath33 , and @xmath34 is the holon number operator at site @xmath35 .
the mean - field hamiltonian ( [ hs ] ) is by nature a gauge model , which is invariant under the @xmath36 transformation : @xmath37 and @xmath38 .
it can be easily shown that the spin rotational symmetry is respected by eq .
( [ hs ] ) by verifying @xmath39=0 $ ] where the definition of the spin operator @xmath40 in the phase string representation is given in appendix a. in the low - temperature regime , where the _ bosonic _ holons in the b - rvb theory will experience the bose condensation such that one may approximately treat @xmath30 as describing a uniform flux of the strength @xmath41 per plaquette ( @xmath42 denotes the doping concentration of holes).@xcite in the high - temperature regime , @xmath43 may be treated as describing randomly distributed static @xmath1 flux tubes of concentration @xmath42 since the holons will behave like incoherent objects there.@xcite different from a usual jordan - wigner phase , the gauge field @xmath30 , which is seen by spinons , is attached to an _ independent _ degree of freedom , holons , and therefore the above approximations are reasonable . in both limits ,
the bilinear form of eq.([hs ] ) in terms of the bosonic spinon operator @xmath25 can be diagonalized by the following bogoliubov transformation:@xcite @xmath44 \label{bogoliubov}\]]with @xmath45where @xmath46 satisfies the following eigen equation@xmath47 in which @xmath48 denotes the four nearest neighbors ( nn ) of site @xmath49 one has @xmath50 and @xmath51 where @xmath52 is the spinon spectrum .
the lagrangian multiplier @xmath53 is determined by enforcing the average constraint @xmath54 which leads to @xmath55where @xmath56 and the last equation is obtained by the self - consistent condition ( [ parameter ] ) for the rvb order parameter .
note that a bose condensed term @xmath57 related to the aflro psmfd at @xmath58 and half - filling has been dropped in eq.([lambda ] ) since we shall mainly be interested in the high - temperature behavior .
all the nontrivial effect of doping is reflected in the eigen equation ( eig ) where the phase string effect induced by hopping enters via the topological gauge field @xmath30 .
note that the spinon wave function @xmath59 should vanish at the hole sites ( where @xmath60 ) due to the no double occupancy condition .
previously,@xcite such an equation has been solved by a simple mean - field choice @xmath21 [ @xmath61@xmath62 with the relaxed constraint condition such that the spinons can go any sites . in the following
we still relax the no double occupancy constraint on average in eq.([eig ] ) , but with a replacement of @xmath63 to represent the average effect of the reduction of the superexchange coupling around holes : if holes are static , each of which will simply break two @xmath64 links in each direction such that @xmath65 in the dilute hole limit ( when the average hole - hole distance is much larger than the @xmath64 links ) .
generally @xmath66 for a moving hole since the suppression of @xmath21 around each hole extends more than the four broken @xmath64 bonds ( @xmath67 from the singular twist by @xmath30 around each hole )
. one may thus approximately rewrite eq.([eig ] ) as @xmath68 in which @xmath69and eq.([delta ] ) can be consistently rewritten as @xmath70note that the same mean - field equations have been obtained in ref .
psmfd with a slightly different definition , _
i.e. , _ with @xmath71 replaced by @xmath72 ( with @xmath73 in ref .
@xcite .
the upp is defined by the formation of the b - rvb pairing with @xmath74 its high - temperature boundary at @xmath75 is depicted by a characteristic temperature @xmath6 as illustrated in fig .
[ fp ] . in the following ,
we determine it based on the mean - field theory outlined above .
according to eq.([lambda0 ] ) , one finds @xmath76by noting that @xmath77 at @xmath78 consequently , eq.([delta1 ] ) reduces to @xmath79 , which gives rise to @xmath80by further identifying @xmath81 in terms of eq.([eig1 ] ) .
finally one obtains@xmath82it is interesting to note that the gauge field @xmath30 does not explicitly appear in @xmath6 because of the general relation ( [ eig2 ] ) , indicating that the phase string effect , which is crucial to the low temperature ( low energy ) physics , actually plays no role in determining the phase boundary of the upp .
of course , the hopping effect still enters in @xmath83 via the renormalized factor @xmath84 in @xmath85 which shall be the only adjustable parameter depending on the detailed physics of local hopping .
we find that the phase diagram and magnetic properties to be studied below are actually not sensitive to @xmath84 except for the characteristic concentration @xmath86 at which @xmath85 is extrapolated to zero at @xmath58 .
[ ft0 ] shows @xmath6 ( solid curve ) as a function of @xmath87 with @xmath88 the experimental data determined by the uniform spin susceptibility measurement in @xmath89 scaling , lsco ( see the discussion in the next section ) are shown by the full squares , which are in good agreement with the theoretical curve .
furthermore , the open circles are independently determined from the c - axis transport @xcite in the overdoped regime ( see discussion below ) .
note that the theoretical curve @xmath6 versus @xmath90 in fig .
2 is not sensitive to the choice of @xmath91 furthermore , one can use the above experimental data @xcite to fix @xmath92 at @xmath93 ( @xmath94 )
. we shall then choose @xmath95 throughout the rest of paper without any more adjustable parameter . for the b - rvb origin of the upp ,
the zeeman splitting due to the external magnetic field can effectively destroy the singlet pairing of spins in the strong field limit . since the orbit part of the neutral spins does not couple to the external field directly
, the zeeman splitting will be the only direct field effect on the rvb background .
it thus provides a direct probe of the rvb nature of the upp in , say , the overdoped regime , where the critical field strength may be within the experimental accessible range . in the following we consider the zeeman effect in the upp . of the upp versus @xmath96 .
solid line : the present theory ; full squares : determined from the uniform spin susceptibility @xmath97 in lsco compound;@xcite open circles : determined from the c - axis magneto - resistivity ( @xmath98 measurement in bi-2212 compound;@xcite the dashed line shows the scaling relation of the zero - temperature critical field @xmath99 with @xmath6 as predicted by the theory .
inset : the critical field @xmath20 as a function of temperature at @xmath100 . the experiment data from the c - axis transport in bi-2212 ( ref .
@xcite ) are also shown by the open and full squares .
, width=288 ] apply an external magnetic field @xmath101 along a spin z - axis ( which is not necessarily perpendicular to the lattice plane ) .
a spin zeeman energy term should be then added to @xmath102 in eq.([hs ] ) : @xmath103consequently the spinon excitation spectrum is modified by @xmath104which now explicitly depends on the spin index @xmath105 .
then the mean - field equation ( [ upp1 ] ) at @xmath106 is modified to : @xmath107with @xmath108 , while eq.([lambda ] ) remains the same . from these equations ,
we can easily obtain the following relation between @xmath6 at zero field and the zero - temperature critical field @xmath109 at which @xmath110 vanishes:@xmath111with the coefficient only weakly dependent on the doping concentration . in fig .
[ ft0 ] , @xmath112 is plotted as the dashed curve which scales with the zero - field @xmath6 fairly well , which predicts@xmath113 in general , the temperature dependence of the critical field @xmath114 can be obtained based on eqs .
( [ upp3 ] ) and ( [ lambda ] ) . in the inset of fig .
[ ft0 ] , @xmath115 versus @xmath116 at @xmath117 is shown together with the experimental data obtained from the c - axis magneto transport measurements.@xcite we see that the _ high - temperature _ experimental data ( open circles ) fit the theoretical curve very well without any additional adjustable parameter .
furthermore the zero - field @xmath6 determined by the _
same _ experiments is also in good agreement with the theory as shown ( open circles ) in the main panel of fig .
[ ft0 ] as mentioned above .
one may also note that the experimental @xmath114 starts to deviate from the theoretical curve in the inset ( full squares ) as the temperature is further lowered and saturated to approximately the half of the predicted number ( which implies @xmath118 .
however , we would like to point out that such a deviation occurs only for those data ( full squares ) which have been obtained by _ extrapolation _ in the experimental measurement @xcite and therefore may not be as reliable as the higher temperature ones ( open squares ) in the inset of fig .
we believe that further experiments be needed in order to convincingly verify ( falsify ) the present prediction ( [ h01 ] ) .
finally the boundaries of the upp in the three dimensional parameter space of doping concentration , temperature , and magnetic field determined based on the b - rvb mean - field theory are shown in fig .
[ f3d ] .
the rvb nature of the upp is clearly manifested in the magnetic properties .
we first consider uniform spin susceptibility @xmath119 in the following , which can be easily derived based on the above mean - field description in the presence of magnetic field . in terms of the total spin magnetic moment @xmath120\text { \ \ \ \ }
\label{sus3}\]][where @xmath121 is the bose distribution function ] , the uniform spin susceptibility @xmath119 per cite as defined by @xmath122 is found by @xmath123\text { \ \ \ \ . }
\label{sus4}\ ] ] scaled with the maximum @xmath124 at @xmath6 versus @xmath125 , which follows an approximately doping - dependent curve .
inset : the experimental data in ref . @xcite which collapse into a universal scaling curve plotted in the same fashion as in the main panel .
( b ) the theoretical @xmath97 at half - filling ( solid ) and the one obtained by the high temperature series expansion ( htse ) .
the latter fits the experimental scaling curve in the inset of ( a ) very well.@xcite , width=336 ] since we are mainly interested in the high - temperature regime well above the superconducting phase , the holes are incoherent objects such that @xmath126 can be approximately treated as describing randomly distributed @xmath1 flux tubes of concentration @xmath42 as discussed before
. then we can numerically calculate @xmath119 based on the mean - field equations given in sec .
ii a , which is averaged under different random configurations of @xmath127 a similar computation has been done before to explore the crossover from the lower to upper pseudogap phases,@xcite but not in the region up close to @xmath6 .
( in the following calculations , the largest sample size is @xmath128 lattice and the sample size is not very important as mainly the high temperature properties are concerned . ) the calculated @xmath119 is presented in the main panel of fig .
[ fsus ] ( a ) at different doping concentrations .
note that @xmath119 reaches a maximum value @xmath129 at temperature @xmath6 where the rvb order parameter @xmath71 vanishes . at @xmath14
, @xmath119 follows a curie-@xmath130 behavior as spins become free moments at the mean - field level .
the curves in fig .
[ fsus ] ( a ) are presented as @xmath131 versus @xmath125 , which approximately collapse onto a single curve independent of doping . for comparison ,
the inset shows the experimental data obtained in compounds which are plotted in the same way as in the main panel with a very good collapsing onto a universal scaling curve;@xcite and the peak positions decides the experimental pseudogap temperature @xmath6 s at different dopings , which are presented in fig .
[ ft0 ] . in fig .
[ fsus ] ( b ) , the calculated @xmath119 versus @xmath116 at @xmath132 is shown together with the high temperature series expansion ( htse ) result@xcite ( note that here the calculated @xmath119 is rescaled by a @xmath133 numerical factor as used in the schwinger - boson approach to restore the sum rule@xcite ) .
it is noted that the experimental scaling curve actually coincides with the half - filling htse very well.@xcite thus one can clearly see the overall qualitative agreement between the bosonic rvb theory and the experiment from figs .
[ fsus ] ( a ) and ( b ) .
note that the mean - field @xmath119 deviates from the htse result prominently around @xmath83 where the latter is a much smoother function of @xmath134 it reflects the fact that @xmath6 is only a crossover temperature and the vanishing @xmath135 does not represent a true phase transition .
obviously , the amplitude fluctuations beyond the mean - field @xmath71 have to be considered in order to better describe @xmath119 in this regime .
@xmath6 determined in the mean - field theory is quite close to the htse result , indicating the crossover temperature itself can still be reasonably decided by the mean - field bosonic rvb description given above .
the comparison of @xmath6 between the theory and experiment has been already presented in fig .
ft0 and discussed in the previous section .
the nmr spin - lattice relaxation rate of nuclear spins is determined bymoriya @xmath136where the form factor @xmath137 comes from the hyperfine coupling between nuclear spin and spin fluctuations . for example , for planar @xmath138 nuclear spins in the cuprates , with the applied field perpendicular to the @xmath139 plane
, the form factor @xmath140 is found to be:@xcite @xmath141^{2 } \label{factorcu1}\]]where the hyperfine couplings @xmath142 and @xmath143 are estimated as @xmath144 @xmath145 @xcite [ these coefficients may slightly vary among @xmath146 and @xmath147 compounds ] . for planar @xmath148 nuclear spins , one has @xmath149 \label{factoro1}\]]with @xmath150 . due to the fact that @xmath151 vanishes at the af wave vector @xmath15 ,
while @xmath152 is peaked at @xmath153 , a combined measurement of @xmath154 and @xmath155 will thus provide important information about the af correlations at low frequency @xmath156 . vs. temperature at different doping concentrations in the upper pseudogap phase of the b - rvb state .
the dashed line shows the result of an f - rvb state ( @xmath157 flux phase ) at half - filling .
the inset shows the non - korringa behavior of @xmath158 in the b - rvb state at various dopings ( the symbols are the same as in the main panel).,width=336 ] based on the bosonic rvb mean - field equations outlined in sec .
ii a , the spin - lattice relaxation rates , @xmath154 and @xmath159 for the planar copper and oxygen nuclear spins , can be straightforwardly computed as shown in appendix b. the results are presented in fig .
it shows that the ratio @xmath160 , which is a constant above @xmath6 , starts to increase with reducing temperature below @xmath6 . at lower temperature , @xmath161 ,
such a ratio arises sharply .
for example , @xmath162 diverges at @xmath132 as a true aflro exists at @xmath163 and it can still reaches about @xmath164 in the low temperature limit at @xmath165 , all qualitatively consistent with the experimental observation in the cuprates.@xcite as pointed out above , such behavior clearly demonstrates that strong low - lying af correlations around @xmath166 develop in the upp , leading to the simultaneous enhancement of @xmath154 and the cancellation in @xmath155 . in the inset of fig .
[ fnmr ] , @xmath158 has been plotted , which is also qualitatively consistent with the experiment , @xcite but deviates from the conventional korringa behavior @xmath167 for a fermi liquid system , thanks to the strong af fluctuations of spins in the upp below @xmath6.@xcite by contrast , the ratio @xmath162 in an f - rvb mean - field state ( the @xmath1 flux phase ) at half - filling remains flat over the whole temperature region as shown by the dashed line in fig .
[ fnmr ] , indicating the absence of any significant af correlations around @xmath153 in its pseudogap regime . in combination with the reduced uniform spin susceptibility,@xcite
one sees that in the pseudogap of the f - rvb , the low - energy spin excitations , either around @xmath153 or @xmath168 , all get suppressed with the opening of the pseudogap .
the above results clearly show that the upp in the bosonic rvb state corresponds to the crossover from a weakly correlated localized spin assembly at higher temperature into a strongly _ af _
correlated spin liquid at lower temperature .
the peculiar feature of the bosonic rvb description is that although the formation of bosonic rvb singlet pairing suppresses the spin correlations at @xmath19 below @xmath6 , it also leads to the _ enhancement _ of the low - energy spin correlations near af momentum @xmath153 .
such a feature of the upp is significantly different from the pseudogap phase in the f - rvb mean - field description , but is strongly supported by the nmr measurements.nmr1,shastry,mmp,ratio1,nmr4 finally , we further examine the spin - echo decay rate @xmath169 , which is related to the static af correlations via the real part of spin susceptibility function by:@xcite @xmath170 ^ 2\right\ } \label{t2}\ ] ] where the factor @xmath171 is @xmath172with @xmath173 , such that @xmath171 is peaked at @xmath153 and vanish at @xmath19 . vs. temperature in the upper pseudogap phase below @xmath174 at different doping concentrations.,width=336 ]
similar to @xmath175 , the detailed expression of @xmath169 in the bosonic rvb theory is given in appendix c. in fig.[3 ] , the calculated @xmath169 at different doping concentrations show that @xmath169 begins to increase with reducing temperature below @xmath6 . such behavior has been also observed in the experiment,@xcite which once again clearly illustrates the picture that the strong af correlations start to develop in the upp .
so far we have been focused on the upp , which is the high - temperature phase in the bosonic rvb description .
as stressed in the introduction , there can be several different low - temperature phases growing out of this background ( fig .
1).@xcite one particular phase we wish to discuss below is the so - called spontaneous vortex phase @xcite which can be properly classified as the lower pseudogap phase in the present approach . in this phase ,
the holon condensation occurs and cooper pair amplitude forms , but the system is still short of superconducting phase coherence which can be regarded as a vortex liquid state due to the presence of unpaired spinon - vortex composites.@xcite these spinon - vortices contribute to the nernst effect and therefore the lower pseudogap phase should coincide with the nernst region discovered @xcite experimentally in the cuprates .
recently the electromagnetic response of such a vortex liquid phase has been also discussed @xcite based on a different rvb approach . previously the magnetic properties in this phase has been discussed psk in the context of exploring the driving mechanism in comparison with the pseudogap phase in the f - rvb state . in the following , we focus on the magnetic behavior of this lower pseudogap phase and make contrast with that of the upp discussed above .
the mean - field equations are the same as in the upp , but due to the holon condensation the gauge field @xmath30 in eq .
( [ eig1 ] ) can be treated as a uniform flux of strength @xmath176 per plaquette @xcite as discussed in sec .
ii a. in the main panel of fig .
[ 6 ] , the uniform spin susceptibility shows a true spin gapbehavior , in contrast to the scalingcurve shown in the upp in fig . 4 where @xmath119 in the doped regime roughly behaves like that at half - filling in the latter case @xmath119 saturates to a constant at @xmath58 .
in the lower pseudogap phase , these @xmath177 s can drop below that at @xmath132 and vanish at @xmath58 as shown in fig . [
such a lower pseudogap behavior has been indeed observed experimentally.@xcite and the right insert @xmath169 with the same symbols as in the main panel .
, width=336 ] furthermore , in this lower pseudogap phase , @xmath154 also decreases with temperature ( see the left inset of fig .
7 ) , as opposed to the behavior in the upp , indicating the appearance of the spin gap over whole momenta . on the other hand , although the low - energy spin fluctuations are gapped , the static af spin - spin correlations as described by the real part of spin susceptibility function still remain , as reflected by @xmath169 shown in the right inset of fig . 7 , where the monotonic increase of @xmath169 in the upp ( fig .
6 ) is replaced by the saturation at lower pseudogap phase .
this feature has also been observed experimentally.@xcite
in this paper , we have systematically analyzed the magnetic characterizations of a high - temperature _ intrinsic _ phase of the bosonic rvb state , which is described by the formation of the bosonic rvb order parameter at a temperature below the characteristic @xmath4 but still higher than those for low - temperature orders , including aflro and superconductivity , to emerge .
such a phase exhibits the pseudogap features that match those of the upper pseudogap phase in the high-@xmath0 cuprate superconductors very well .
the key feature in the crossover to the upp from above @xmath6 is the onset of the development of strong af spin - spin correlations , which remain rather weak at @xmath14 where the system resembles more an ensemble of uncorrelated localized spins .
this explains why experimentally the uniform spin susceptibility shows an approximately curie - weiss behavior at @xmath178 reaches a peak at @xmath6 , and then gets reduces below @xmath6 as the weight of the spin - spin correlations at the momentum @xmath179 being transferred to @xmath180 , in contrast to an equal weight distribution above @xmath6 .
it further explains why the spin - lattice relaxation rate gets enhanced for the planar copper nuclear spins whereas reduced for the planar oxygen nuclear spins below @xmath4 and why the spin - echo decay rate increases with the decreasing temperature ; clearly the development of the af correlations is the underlying mechanism here .
we emphasize that the formation of spin singlet pairing and the onset of af correlations at the same time are not always true . in the f - rvb description of the slave - boson approach
, the formation of the f - rvb order parameter actually leads to the reduction of the spin - lattice relaxation rates for both the copper and oxygen nuclear spins at low temperatures .
this is because the spin pseudogap opens for both the ferromagnetic and af correlations .
by contrast , this case occurs in the b - rvb description only at a lower temperature when the system enters the lower pseudogap regime , characterized by the formation of cooper pair amplitude in the so - called spontaneous vortex phase @xcite which is a vortex liquid state , short of superconducting phase coherence . a comparative study of this lower pseudogap phase in the b - rvb theory and the pseudogap phase in the f - rvb theory has been given in ref .
@xcite where two opposite driving mechanisms : kinetic vs. superexchange energy driven , have been identified
. the phase diagram of the upp has been determined by a generalized mean - field description in the b - rvb theory in the parameter space of temperature , doping , and magnetic field .
since the upp essentially reflects the spin singlet pairing , the zeeman splitting competes directly , which results in a quantitative prediction for experiment as discussed in sec .
ii b. a central consequence of an rvb ( spin singlet ) description of the pseudogap phase is that the pseudogap represents the suppression of the low - lying spectral weight in spin excitations , but not in charge excitations , as pointed out in the introduction . in the following ,
we briefly outline the scenario about what happens to the charge channel when one enters the upp in the bosonic rvb theory .
we shall leave the more quantitative investigation in future work . in order to see how the charge transport is affected by the spin fluctuations , we first note that in the b - rvb theory , the charge degrees of freedom ( holons ) are described by the following effective hamiltonian : psmfd@xmath181where the bosonic holons , created by @xmath182 , interact with the gauge field @xmath183 associated with the spin degrees of freedom . similar to the definition of the gauge field @xmath30 in eq.([ah ] ) , @xmath184 , which is introduced in the phase string representation ( see appendix a ) , satisfies @xmath185 where @xmath32 is an arbitrary path .
physically @xmath183 describes @xmath186 flux tubes bound to @xmath187(@xmath188 ) spinons as seen by holons .
so at @xmath14 , uncorrelated localized spins imply a maximum scattering to the holons according to eqs.([hh ] ) and ( [ as ] ) . by forming the rvb pairing below @xmath6
, one can easily understand that the fluctuations in @xmath184 will be effectively reduced , and so does the scattering to the holons according to eqs.([hh ] ) and ( [ as ] ) , leading to a pseudogap feature in the charge transport without involving a charge gap .
note that when the temperature is further reduced to @xmath5 , where the holons gain the phase coherence and become bose condensed , the system enters the lower pseudogap phase ( spontaneous vortex phase ) in which the effect of the holon condensation will feedback to the spinon part via @xmath189 in eq.(hs ) and cause the lower pseudogap phenomenon in the spin part as discussed in sec .
ii e. finally , the quasiparticle excitation can be regarded as a recombination of holon , spinon , and phase string , in the superconducting phase.@xcite it has been argued that the deconfinement occurs above @xmath190 , and the composite structure is expected to shows up in both the lower and upper pseudogap phases , and the pseudogap feature is thus believed to be associated with that in the spinon degrees of freedom .
we thank helpful discussions with v. n. muthukumar , t. senthil , d. n. sheng , h.h .
the authors also acknowledge the support from the grants of nsfc and moe . this research was also supported in part by the nsf under grant no .
phy99 - 07949 via kitp at uc santa barbara where this work was completed .
the @xmath22 model @xmath191may be reformulated by using the phase string decomposition@xcite @xmath192where @xmath193 and @xmath25 are all _
bosonic fields_. here @xmath194 is a non - local phase factor to restore the fermionic statistics of the electron operator , and can be expressed as @xmath195 $ ] with @xmath196 , @xmath197 . here
@xmath198 is defined as an angle @xmath199 i m @xmath200
with @xmath201 representing the complex coordinate of a lattice site i. the resulting hamiltonian @xmath202 reads @xmath203under the no - double - occupancy constraint @xmath204 in the new hamiltonians , @xmath205 is a @xmath1 flux link variable , while @xmath183 and @xmath30 are constrained by the following conditions : @xmath206where @xmath32 is an arbitrary counterclockwise closed path .
finally , in the phase string representation , the spin operators can be easily reexpressed according to the decomposition ( [ cop ] ) as @xmath207
@xmath175 defined in ( [ nmrg1 ] ) can be reexpressed in terms of the real - space spin correlation function as follows @xmath208where @xmath209 is the fourier transformation of @xmath137 in real space : @xmath210where @xmath211 and @xmath212from @xmath209 we see that only up to the next - nearest - neighbor spin correlations are involved in the spin - lattice relaxation rates of @xmath213 and @xmath214 nuclear spins . in the bosonic rvb
mean - filed theory , by using eq.([bogoliubov ] ) the dynamic spin susceptibility can be expressed as @xmath215where @xmath216 in which @xmath217 denote the summation of @xmath218 with @xmath219 and @xmath220 by noting that @xmath221 at @xmath222 , we can further reexpress @xmath175 in the following form @xmath223 \label{dsus4}\]]where the density of states @xmath224 and the coefficient , @xmath225 , is defined by @xmath226with @xmath227 in eq.([dsus4 ] ) a numerical factor @xmath133 is also added just like the uniform spin susceptibility as noted in the main text as at half - filling.@xcite the final result will be an average over different random configurations of @xmath30 due to the incoherent distribution of holes in the upp ( in the lower pseudogap phase , by contrast , the holon condensation leads to a uniform flux distribution of @xmath30 and no such average is needed@xmath228 the calculation is done on a @xmath128 lattice and results are presented in fig .
the spin - echo relaxation rate @xmath169 is defined in eq.([t2 ] ) . in the b - rvb theory ,
the real part of the static susceptibility , @xmath229 , can be expressed as @xmath230with y. wang , s. ono , y. onose , g. gu , y. ando , y. tokura , s. uchida , and n. ong , science * 299 * , 86 ( 2003 ) ; y. wang , z.a .
xu , t. kakheshita , s. uchida , s. ono , y. ando , and n. ong , phys . rev .
b * 64 * , 224519 ( 2001 ) ; z.a .
ong , y. wang , t. kakheshita , and s. uchida , nature ( london ) * 406 * , 486 ( 2000 ) . note that in the inset of fig . 5 , @xmath158 is shown only below @xmath234 as @xmath158 is ill - defined at @xmath6 in the mean - field description where @xmath75 . such an ill - behavior
is also reflected in the uniform susceptibility in fig . 2 by a sharp cusp at @xmath83 as compared to the broad , smooth peak in the htse result and experimental data .
this is due to the fact that there is no true phase transition at @xmath6 and the description beyond the mean - field order parameter @xmath71 is needed in order to more accurately describe the crossover behavior in the transition regime . on the other hand ,
the ratio @xmath162 still remains well - defined around @xmath6 in the mean - field state as shown in the main panel of fig . | it is proposed that the _ upper _ pseudogap phase ( upp ) observed in the high-@xmath0 cuprates correspond to the formation of spin singlet pairing under the bosonic resonating - valence - bond ( rvb ) description .
we present a series of evidence in support of such a scenario based on the calculated magnetic properties including uniform spin susceptibility , spin - lattice relaxation and spin - echo decay rates , which consistently show that strong spin correlations start to develop upon entering the upp , being _ enhanced _ around the momentum ( @xmath1 , @xmath1 ) while _ suppressed _ around ( @xmath2 , @xmath2 ) .
the phase diagram in the parameter space of doping concentration , temperature , and external magnetic field , is obtained based on the the bosonic rvb theory . in particular
, the competition between the zeeman splitting and singlet pairing determines a simple relation between the critical magnetic field , @xmath3 and characteristic temperature scale , @xmath4 of the upp .
we also discuss the magnetic behavior in the _ lower _ pseudogap phase at a temperature @xmath5 lower than @xmath6 , which is characterized by the formation of cooper pair amplitude where the low - lying spin fluctuations get suppressed at both ( @xmath2 , @xmath2 ) and ( @xmath1 , @xmath1 ) .
properties of the upp involving charge channels will be also briefly discussed . |
blazars , consisting of flat spectrum radio quasars ( fsrqs ) and bl lacertae objects ( bl lacs ) , are a relatively rare type of active galaxy characterized by extremely variable , non - thermal spectral energy distributions ( seds ) .
the emission extends from radio to @xmath0-ray frequencies and is believed to be produced by a highly relativistic plasma jet closely orientated with the line of sight to the galaxy @xcite . in a @xmath3 representation ,
the seds display two peaks .
the lower energy peak is usually attributed to synchrotron radiation from relativistic electrons , while the higher energy peak is thought to be due to inverse compton scattering by these same electrons .
the photons involved in the scattering process may be either synchrotron photons produced in the jet , in the synchrotron self - compton ( ssc ) model , or some external population , such as photons emitted from an accretion disc @xcite or reflected from emission line clouds @xcite .
alternative explanations for the high energy component invoke pair cascades triggered via pion and pair photoproduction from high energy protons in the jet @xcite , or proton synchrotron radiation @xcite . the egret detector on board the cgro detected in excess of 66 blazars @xcite above @xmath4 ; however , only a few have been detected at higher energies ( @xmath5 ) by ground based atmospheric cherenkov telescopes .
markarian 421 @xcite and markarian 501 @xcite were the first extragalactic tev sources to be detected and the behaviour of these two nearby ( @xmath6 and @xmath7 respectively ) objects has been closely monitored and studied .
flux variability over two orders of magnitude has been observed , with doubling timescales as short as @xmath8 @xcite , placing strong constraints on the size of the emission region .
the @xmath0-ray flux is correlated with the x - ray flux @xcite and , in the case of markarian 421 , with the spectral power law index in the @xmath0-ray region @xcite .
unconfirmed detections of 1es 2344 + 514 @xcite and pks 2155 - 304 @xcite have also been reported and , most recently , the detection of very weak emission from h1426 + 428 has been confirmed by three groups @xcite .
this last object is of particular interest since it is relatively distant ( @xmath9 ) and hence the observed flux is expected to be strongly attenuated by absorption via pair production on the extragalactic infra - red background light ( ebl ) @xcite
. all of the tev blazars can be classified as high frequency peak bl lacs ( hbls ) @xcite on the basis of the location of the lower energy peak in their seds , and predictions of tev emission from other nearby sources of this type have driven observations by atmospheric cherenkov detectors .
1es1959 + 650 ( @xmath1 ) was first suggested as a good tev candidate of this type by @xcite , who used simple scaling arguments to predict it as the third strongest tev blazar , after markarian 421 and markarian 501 , with a flux prediction of @xmath10 above @xmath11 .
more recently @xcite have predicted fluxes above @xmath11 of @xmath12 , based on a simple phenomenological parameterization of the sed adapted from @xcite , and @xmath13 using a homogeneous , one - zone synchrotron self - compton model @xcite .
in fact , 1es1959 + 650 is not a particularly extreme example of an hbl .
measurements with bepposax in 1997 @xcite placed the lower energy sed peak at @xmath14 ( @xmath15 ) , as compared to typically @xmath16 for markarian 421 @xcite and @xmath17 for markarian 501 during its 1997 flare state @xcite , although for both of these sources the position of the peak is known to vary widely depending upon the flux level .
@xcite report on x - ray observations of 1es1959 + 650 obtained with the usa(1 - 16 ) and rxte ( 2 - 16 ) missions during 2000 , showing threefold increases in the x - ray flux on a timescale of a few days .
the flux increase is correlated with spectral hardening , indicative of a shift of the lower energy peak of the sed towards higher frequencies .
the source is also unusual among hbls for its strong rapid optical variability .
observations by @xcite showed rapid flickering , including a decrease of 0.28 magnitudes in four days . prior to may 2002
, only tentative evidence had been presented for tev emission from 1es1959 + 650 .
a detection with a statistical significance of @xmath18 was reported by @xcite based on @xmath19 of observations with the utah seven telescope array .
more recently , @xcite reported a preliminary detection at @xmath20 for the hegra cherenkov telescope system .
previous observations of 1es1959 + 650 with the whipple 10 m telescope produced an upper limit at a flux level of @xmath21 above @xmath22 @xcite .
observations of 1es1959 + 650 during may - july 2002 with the whipple 10 m telescope resulted in a clear detection of tev @xmath0-ray emission from this object @xcite which was rapidly confirmed by the hegra experiment @xcite .
we report here the results of the whipple 10 m observations .
the configuration of the whipple 10 m @xmath0-ray telescope during these observations is described in detail in @xcite . briefly , the telescope consists of a @xmath23 reflector and a 490 pixel photomultiplier tube ( pmt ) camera . for the purposes of this analysis only the high resolution ( @xmath24 spacing ) central 379 pmt pixels have been used .
cherenkov images are recorded and parameterized according to @xcite and then @xmath0-ray like images are selected using the `` supercuts '' criteria @xcite optimized on recent data from the crab nebula , the standard candle of tev @xmath0-ray astronomy @xcite . following a realignment of the optical system in february 2002
, observations of the crab nebula showed that the telescope was at its most sensitive since the installation of the current camera , with one hour of crab observations producing a @xmath25 detection ; however , between february and july 2002 a decrease in the relative efficiency of the telescope of @xmath26 was measured by examining the response to the cosmic ray background .
the cause of the effect is still under investigation - one explanation may be that it is due to increased atmospheric absorption caused by large forest fires in the region during this period .
observations of 1es1959 + 650 were made during moonless periods between may @xmath27 and july @xmath28 2002 .
the total dataset consists of @xmath29 of on - source data , together with @xmath30 of off - source data for background comparison . for observations from mt .
hopkins ( latitude @xmath31n ) 1es1959 + 650 culminates at a zenith angle ( @xmath32 ) of @xmath33 and so the data were necessarily taken at large @xmath32 , between @xmath34 and @xmath33 .
after accounting for the zenith angle and reduced telescope efficiency , we estimate the energy threshold ( @xmath35 ) for the majority of the observations to be @xmath36 .
@xmath35 is defined here as the energy of the peak differential @xmath0-ray flux for a source with the same spectrum as the crab nebula .
figure [ alpha ] shows the distribution of the cherenkov image orientation angle _ alpha _ for the full on - source dataset , together with the distribution for the off - source dataset scaled such that the number of events in the region @xmath37 is the same .
the average rate of the excess in the on - source @xmath0-ray region at @xmath38 is @xmath39 , corresponding to a detection of greater than @xmath40 .
figure [ daily ] shows the daily averaged rates for 1es1959 + 650 for all of the observations .
the rates are expressed in multiples of the steady crab nebula flux and have been corrected in order to account for the varying zenith angle and changes in telescope efficiency according to the procedure of @xcite .
this has been tested using crab nebula data taken over a wide range of zenith angles and atmospheric conditions .
the correction factor varies from run to run and has a mean value for this dataset of @xmath41 .
strong night - to - night variability is evident ; the largest change in rate , between mjd 52428 and mjd 52429 , corresponds to a doubling time of @xmath42 - shorter than has ever been observed in other wavebands for this source .
the mean flux over all observations was @xmath2 crab units .
figure [ nightly ] shows the rate in @xmath43 bins for two nights , may @xmath44 and june @xmath45 , during which the source was most active . the statistical evidence for variability within these nights is given by the @xmath46 probabilities of constant emission = 1% and 8% respectively .
we conclude that there is no strong evidence for flux variability on this timescale .
the detection of tev @xmath0-ray emission from 1es1959 + 650 adds another member to the class of tev blazars , all of which are close bl lac objects having a low bolometric luminosity and the peaks in their seds at high frequency . the rapid flux variability , and the fact that 1es1959 + 650 has not been detected during previous observations , indicates that the source was in an unusual flaring state during these observations .
the flux level was at times orders of magnitude above the most recent model predictions . throughout the period of the whipple observations measurements in the @xmath47 region by the all - sky monitor ( asm ) on board rxte
have shown 1es1959 + 650 to be active and variable , with daily average fluxes reaching @xmath48 in may and july .
target of opportunity observations with the rxte small field of view instruments were triggered following the @xmath0-ray detection and will be reported on elsewhere ( krawczynski , private communication ) .
the rapid flux variability observed at tev energies implies a small emission region in the jet with a high doppler factor @xcite .
the contemporaneous x - ray and tev @xmath0-ray data will allow us to constrain the jet parameters when modelling the emission processes .
analysis of the whipple data is ongoing , but attempts to reconstruct the source spectrum have been hampered by the effects of the decreased telescope efficiency during the observation period . the hegra collaboration measure a rather steep spectrum ( spectral index @xmath49 ) for observations prior to 2002 , while the spectrum during the flaring period exhibits pronounced curvature and deviates significantly from the spectrum seen during the quiescent state @xcite .
the majority of models for the ebl lead to predictions of a cut - off in the @xmath0-ray region beginning below @xmath50 for a source at @xmath1 ( @xcite ; but see @xcite for a discussion of an ebl model which does not produce a distinct feature in the observed spectrum ) . deviations from a pure power law have now been resolved in the spectra of both markarian 421 @xcite and markarian 501 @xcite and the spectrum of the most distant tev blazar , h1426 + 428 , is measured to be very steep ( @xmath51 ) @xcite , but it is not yet clear whether these features are due to absorption on the ebl or are intrinsic to the sources . clearly ,
further observations and spectral analysis of 1es1959 + 650 may help to resolve this question .
aharonian , f. a. et al . 1999 ,
a&a , 349 , 11 aharonian , f. a. 2000 , new a , 5 , 377 aharonian , f. a. et al .
2002a , a&a , 393 , 89 aharonian , f. a. et al .
2002b , a&a , 384 , l23 beckmann , v. et al .
2002 a&a , 383 , 410 blandford , r. d. & rees , m. j. 1978 , pittsburgh conference on bl lac objects , a. n. wolfe , university of pittsburgh press , 1978 , 328 buckley , j. h. et al .
1996 , , 472 , l9 catanese , m. et al .
1997a , , 487 , l143 catanese , m. et al . 1997b , proc .
4th compton symp .
, c. d. dermer , m. s. strickman , j. d. kurfess , aip conf .
410 , 1997 , 1376 catanese , m. et al .
1998 , , 501 , 616 chadwick , p. m. et al .
1999 , astropart .
phys . , 11 , 145 costamante , l. and ghisellini , g. 2002 , a&a , 384 , 56 dermer , c. d. , schlickeiser , r. and mastichiadis , a. 1992 , a&a , 256 , l27 dowdall , c. , moriarty , p. & kosack , k. 2002 , i.a.u .
circular no .
7903 djannati - atai , a. et al . 1999 ,
a&a , 350 , 17 djannati - atai , a. et al .
2002 , a&a , 391 , l25 finley , j. p. et al .
2001 , 27th icrc , m. simon , e. lorenz & m. pohl , hamburg , 2001 , 7 , 2827 fossati , g. et al .
1998 , mnras , 299 , 433 gaidos , j. a. et al .
1996 , nature , 383 , 319 ghisellini g. , celotti , a. & costamante , l. 2002 , a&a , 386 , 833 giebels , b. et al .
2002 , , 571 , 763 gould , r.j . & schrder , g. 1967 , phys .
, 155 , 1408 hartman , r. c. et al .
1999 , , 123 , 79 hillas , a. m. 1985 , 19th icrc , la jolla , nasa conf . pub . 2376 , 1985 , 3 , 445 horan , d. et al .
2002 , , 571 , 753 horns , d. et al .
2002 , proceedings for `` high energy blazar astronomy '' , turku , finland 2002 horns , d. & konopelko , a. 2002 , i.a.u .
circular no .
7907 konopelko , a. et al .
2002 , april aps / head meeting , albuquerque , session b17 , 2002 krawczynski , h. et al .
1997 , a&a , 353 , 97 krennrich , f. et al .
2001 , , 560 , l45 krennrich , f. et al .
2002 , , 575 , l9 lebohec , s. & holder , j. 2002 , astropart .
physics , in press madejski , g. et al .
1996 , , 459 , 156 mannheim , k. 1993 , a&a , 269 , 67 mannheim , k. 1998 , science , 279 , 684 maraschi , l. et al . 1999 , , 526 ,
l81 mattox , j.r .
1993 , , 410 , 609 mcke , a. & protheroe , r. j. 2001 , astropart .
phys . , 15 , 121 nishiyama , t. et al .
1999 , 26th icrc , salt lake city , utah , d. kieda , m. salamon , b. dingus , 1999 , 3 , 370 padovani , p. & giommi , p. 1995
, , 444 , 567 petry , d. et al .
2002 , , 580 , 104 pian , e. et al .
1998 , , 492 , l17 piron , f. et al .
2001 , a&a , 374 , 895 punch , m. et al .
1992 , nature , 358 , 477 primack , j. r. 2001 , high energy gamma - ray astro .
, heidelberg , f.a .
aharonian , h.j .
vlk , aip conf .
proc . 558 , 2001 , 463 quinn , j. et al .
1996 , , 456 , l83 reynolds , p.t .
1993 , , 404 , 206 samuelson , f.w .
1998 , apj , 501 , l17 sikora , m. , begelman , m. c. & rees , m. j. 1994 , , 421 , 153 stecker , f.w . , de jager , o. c. & salamon , m. h. 1992 , , 390 , l49 stecker , f.w . , de jager , o. c. & salamon , m. h. 1996 , , 473 , l75 urry , c. m. & padovani , p. 1995 , pasp , 107 , 803 vassiliev , v. v. 2000 , astropart .
, 12 , 217 villata , m. et al .
2000 , a&as , 144 , 481 weekes , t. c. et al .
1989 , , 342 , 379 | we present the first strong detection of very high energy @xmath0-rays from the close ( @xmath1 ) x - ray selected bl lacertae object 1es1959 + 650 .
observations were made with the whipple 10 m telescope on mt .
hopkins , arizona , using the atmospheric cherenkov imaging technique .
the flux between may and july 2002 was highly variable , with a mean of @xmath2 times the steady flux from the crab nebula and reaching a maximum of five crab , with variability on timescales as short as seven hours . |
the importance of supernova ( sn ) blastwaves to the interstellar medium ( ism ) has been appreciated for many decades @xcite . in recent years , it has become increasingly clear that accurate treatment of supernova remnant ( snr ) evolution and other energy inputs from massive stars is crucial not just for detailed models of the ism and star formation , but also for models of galaxy formation and evolution .
sne are believed to play the predominant role in creating the hot phase and shaping the large - scale structure of the multiphase ism , in driving turbulence in diffuse gas ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and in regulating star formation rates by maintaining the large - scale turbulent pressure that limits gravitational condensation ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
particularly important , as we shall discuss below , is the momentum injected to the ism over the lifetime of a snr .
the role of sne in dense , star - forming molecular clouds is less clear ; due to the time delay before sne occur , other forms of star formation `` feedback '' including protostellar outflows , the pressure of ionized gas in regions , forces from direct and dust - reprocessed radiation , and the pressure of shocked stellar winds may be of comparable or greater importance in driving turbulence and destroying massive molecular clouds ( see * ? ? ?
* ; * ? ? ?
* and references therein ) . through their roles in controlling star formation and launching galactic winds , sne
also regulate the internal structure of galaxies and the cosmic evolution of galactic populations , and recent numerical simulations have followed this with increasingly detailed treatments ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the expansion of snrs in a uniform medium has been extensively studied via spherically symmetric models and is well characterized by several familiar stages .
the sn explosion event ejects material into interstellar space with typical kinetic energy of @xmath6 and mass of @xmath7-@xmath8 .
the ejecta expand freely as long as the mass swept up by the forward shock is smaller than the ejecta mass ( the free expansion stage ) . as the reverse shock heats up the interior ,
the hot gas temperature and pressure become very high , and expansion into the ambient medium proceeds with negligible radiative cooling .
this is the well known phase of evolution analyzed by @xcite and @xcite ( the st stage ) . in the st stage ,
the outer shock radius varies as @xmath9 , and self - similar solutions describe the interior structure very well .
as the temperature drops , radiative cooling becomes important , and a thin and dense shell is formed at the outer edge of the snr where the temperature is lowest and the density is highest .
after shell formation , a pressure - driven snowplow ( pds ) stage occurs while the interior hot gas has non - negligible pressure ( e.g. , * ? ? ?
* ; * ? ? ?
after the interior pressure is exhausted , the shell continues to expand and sweep up ism gas as a momentum - conserving snowplow .
one - dimensional , spherical hydrodynamic simulations of snrs with radiative cooling confirm these main stages , also showing that simple treatments of the pds stage ( assuming adiabatic evolution of the hot interior ) are inadequate ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
although spherical models of snr evolution in a uniform medium are well developed , direct application to the ism is questionable since the ism in reality is highly inhomogeneous . without sne , the neutral atomic ism would consist of two distinct phases , with dense cold neutral medium ( cnm ) clouds surrounded by a diffuse warm neutral medium ( wnm ) , as a result of thermal instability @xcite .
where the cnm collects into a large enough cloud that the interior is shielded from uv , it becomes molecular ( see e.g. * ? ? ? * and references therein ) , and molecular clouds themselves are highly inhomogeneous because they are pervaded by highly supersonic turbulence .
the addition of sn - driven shocks to the ism makes its inhomogeneity even more extreme , producing a third phase of extremely tenuous gas .
although observational estimates are difficult , the hot ism is believed to fill a fraction @xmath10 of the milky way disk s volume , with the majority filled by wnm ( and embedded cold clouds ) .
since cnm clouds are two orders of magnitude denser than the intercloud wnm but the cold and warm mass fractions of the atomic ism are comparable @xcite , snr expansion will be altered compared to spherical models .
understanding the effects of cloudy structure on snr evolution requires multi - dimensional numerical simulations . as early efforts , effects of cloud compression and evaporation
have been implicitly considered in spherical numerical simulations ( e.g. , * ? ? ?
* ; * ? ? ?
however , most multi - dimensional numerical simulations have focused on the detailed evolution and fate of individual shocked clouds ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) rather than snr evolution as a whole within the cloudy ism
. the dearth of systematic studies of the snr evolution in the multiphase ism via direct numerical simulations has also had a serious impact on larger scale simulations . without a model for blastwave propagation in an inhomogeneous medium during weakly radiative stages , the momentum injection from sn feedback is sometimes severely underestimated by considering just the momentum of the immediate ejecta ( e.g. , * ? ? ?
when sn feedback is modeled by thermal energy injection , insufficient numerical resolution leads to the so - called `` overcooling - problem '' ( e.g. , * ? ? ?
* ) , which artificially suppresses the dynamical impact of sn feedback to the surrounding medium .
although it has long been considered a serious problem in galaxy formation simulations , firm resolution requirements to avoid overcooling have not yet been quantified . instead
, many studies simply adopt a prescription in which cooling is artificially turned off for a period of time ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . however , this delayed - cooling prescription has the opposite problem of extending the st phase to unrealistically long times .
the method adopted in @xcite , widely used in subsequent studies ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , disables cooling for the merging time of @xcite , which exceeds the shell formation time by an order of magnitude .
this results in the substantial overestimate of the momentum injection and hot gas production by sne . to address some of these issues
, @xcite have recently developed a ( two - phase ) subgrid model to follow feedback effects from superbubbles , in which the proportions of hot and cold subgrid phases are controlled by thermal conduction , and both phases cool radiatively .
for the purposes of building analytic models of the ism and star formation , and as inputs to subgrid feedback models in numerical simulations of galaxy formation , two important quantities are the total momentum injected by a snr , and the mass of hot gas that it produces .
the momentum , which is transferred to the warm and cold gas , determines the amplitude of turbulence in the volume - filling ism , and also enters inversely in determining the large - scale star formation rate ( e.g. * ? ? ?
* ; * ? ? ?
the hot gas produced by snrs is important because it may be able to escape the galaxy s gravitational potential and drive a wind ( e.g. * ? ?
* ; * ? ? ?
although both quantities are important , momentum is more easily characterized as it is conserved at late times , whereas the mass of hot gas steadily declines once cooling has begun . in this paper , we investigate the momentum injection to the multiphase ism by sne , using direct three dimensional hydrodynamic simulations with cooling .
we shall show that the total momentum injection to an inhomogeneous medium is similar to that in a homogeneous medium , with both comparable to the radial momentum at the end of the st stage .
this is an order of magnitude larger than the initial momentum of the sn ejecta
. we shall also provide firm numerical conditions for resolving the st stage ( thereby obtaining the correct momentum injection and early evolution of the hot medium ) using standard finite - volume methods for hydrodynamics on a grid .
the remainder of the paper is organized as follows . in section [ sec : theory ] , we first summarize theoretical models of spherical snr evolution .
we provide analytic estimates of the shell formation time and key quantities ( size , mass , velocity , momentum , temperature ) of the snr at this epoch .
we then carry out three different sets of simulations , with numerical prescriptions as described in section [ sec : method ] .
we present results from three - dimensional , hydrodynamic simulations for single sne in a uniform background in section [ sec : single ] .
we delineate the detailed evolution and compare numerical results to analytic models in section [ sec : single_ref ] , and provide a systematic parameter study to find numerical convergence conditions in section [ sec : single_conv ] . in section [ sec : single_2p ] , we perform a similar systematic study for single sne in the two - phase ism .
an initial exploration of effects of multiple correlated sne in the two - phase ism is presented in section [ sec : multi_ti ] .
finally , in section [ sec : sumndis ] , we summarize our results and discuss their implications .
here , we also relate our work to other recent studies , and discuss physical limitations of our models and their potential impact on our conclusions . in the appendix , we present results from magnetohydrodynamic models analogous to those of section [ sec : single ] , with a range of background magnetic field strengths .
in this section , we briefly review analytic theories for the dynamics of radiative snrs . here
we consider the simplest case of spherical expansion , for sn energy @xmath11 and ejecta mass @xmath12 in a uniform medium with density of @xmath13 ( see @xcite for a more comprehensive review ) . in simple analytic models , each evolutionary stage of the snr expansion
can be approximated by a power law expansion of the snr radius in time as @xmath14 .
the value of the power law exponent @xmath15 distinguishes four main stages of the expansion ( e.g. , * ? ? ?
* ; * ? ? ?
* ) : free expansion ( @xmath16 ) , sedov - taylor ( @xmath17 ) , pressure - driven snowplow ( @xmath18 ) , and momentum - conserving snowplow ( @xmath19 ) . at each successive stage ,
the exponent decreases as the available power to drive expansion declines . in the initial stage of evolution ,
the mass of the ejecta dominates the mass of material that has been swept up from the circumstellar medium .
thus , the ejecta expands ballistically into the circumstellar medium with nearly constant velocity @xmath20 so that @xmath21 ( @xmath22 ) .
the free expansion stage ends when the swept - up mass @xmath23 becomes comparable to the ejecta mass . by equating @xmath24
, we obtain the end of the free expansion stage as @xcite @xmath25 for the snr radius of @xmath26 here , @xmath27 , and @xmath28 where the hydrogen number density of the ambient medium is @xmath29 for 10% helium abundance .
after the reverse shock heats up the ejecta , the evolution of the snr for @xmath30 is well described by evolution of an idealized , point source explosion since the pressure of the snr far exceeds the ambient medium pressure . to a good approximation , in this stage
we can neglect the ejecta mass , energy losses , and the pressure of the ambient medium , and the solution depends only on @xmath11 and @xmath13 ; this defines the st stage .
simple dimensional arguments imply that any length scale in the solution at a given time @xmath31 should be proportional to @xmath32 .
the internal structure of the snr in the st stage is therefore given by a similarity solution with a similarity variable @xmath33 .
a detailed solution gives @xmath34 at the shock radius , when the specific heat ratio is @xmath35 .
we thus have the shock radius , shock velocity , and the immediate postshock temperature during the st stage given by : @xmath36 @xmath37 @xmath38 where @xmath39 is the mean mass per particles , and @xmath40 ( e.g. , * ? ? ?
when radiative energy losses are no longer negligible , the st stage ends .
cooling is strongest immediately behind the shock where the density is the highest , and a cool dense shell bounding the snr forms .
the shell formation time can be defined as the time at which the first shocked gas parcel cools ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . for a volumetric cooling rate @xmath41 , the cooling time @xmath42 , where @xmath43 is the internal energy , of the postshock gas is @xmath44
a gas parcel shocked at @xmath45 cools down during @xmath46 and forms a shell at @xmath47 .
the time for first shell formation can be obtained by substituting for @xmath48 from equation ( [ eq : tst ] ) in equation ( [ eq : tcool ] ) and then finding the time @xmath45 that minimizes @xmath49 : @xmath50 here , we have used an approximate power - law cooling function @xmath51 with @xmath52 and @xmath53 that gives a fairly good fit for our adopted cooling function in a range of @xmath54 ( see * ? ? ?
note that alternative approximations for the cooling function with @xmath55 change the exponents for energy and density only slightly , in the range 0.21 to 0.24 and -0.57 to -0.53 , respectively . at the time of shell formation
, the outer radius , velocity , postshock temperature , and total swept - up mass are given by : @xmath56 @xmath57 @xmath58 @xmath59 it is notable that the velocity , post - shock temperature , and total swept - up mass of the remnant at the time of shell formation are insensitive to the ambient density .
physically , this is simply because cooling becomes quite strong when the temperature falls below @xmath60k and c and o acquire electrons that can be collisionally excited ( e.g. * ? ? ?
* ) ; reaching this temperature ( and the corresponding shock speed ) essentially defines the radiative stage of a snr @xcite . at shell formation , the shock stalls , and the hot gas in the outer portion of the snr
is rapidly compressed into a cold thin shell .
after shell formation , the shell is still pushed outward by overpressured hot gas in the interior of the snr . under certain idealizations
, this leads to a so - called pressure - driven snowplow ( pds ) stage .
the equation of the motion for the thin shell in the idealized pds stage can be written as @xmath61 where @xmath62 is the mass of the shell and @xmath63 is the expansion velocity of the snr .
if the radiative cooling of the hot interior gas is negligible ( i.e. only the gas immediately behind the shock is assumed to cool ) and the mass of hot gas is assumed to be constant , the interior pressure would drop only from adiabatic expansion as @xmath64 .
for @xmath35 , setting @xmath65 and dropping @xmath66 on the right - hand side leads to a self - similar solution with @xmath18 ( see @xcite for possible variations of this equation with other terms ) .
for the idealized pds solution , the snr interior is treated as uniform , with a mean pressure @xmath67 with @xmath68 where @xmath69 .
note that the mean pressure is about a half of the post - shock pressure .
the final classical stage of a snr begins when the interior pressure @xmath70 has decreased sufficiently that it no longer exceeds the pressure of the ambient medium @xmath66 , making the right hand side of equation ( [ eq : shell ] ) zero .
this leads to the constant radial momentum and a self - similar expansion with @xmath19 .
we now summarize the momentum injection to the ism from each of the above snr expansion stages . in the free expansion stage ,
all the momentum is contained in the ejecta , and the magnitude of the radial momentum is @xmath71 since free expansion by definition assumes negligible interaction with the circumstellar and/or interstellar medium , the ambient medium acquires no momentum at this stage . in the st stage , the propagating strong shock heats and accelerates the ambient medium . using the st self - similar solutions ,
the total radial momentum of the shocked hot gas can be obtained as @xmath72 the momentum at the time of shell formation ( the end of the st stage ) is obtained by substituting equation ( [ eq : tsf ] ) in equation ( [ eq : pst ] ) : @xmath73 we note that this is nearly linear in the sn energy , and is quite insensitive to the ambient density .
this is because the momentum of the snr during the st stage is @xmath74 , and the expansion velocity at the time of shell formation from equation ( [ eq : vsf ] ) is very insensitive to both total energy and ambient density , @xmath75 . in the pds stage , the mass and radial momentum in the snr
are dominated by the shell . under the assumption that the interior expands adiabatically ( which is not in fact satisfied ; see below ) , the momentum that would be acquired in an idealized pds stage after @xmath49 can be obtained by integrating equation ( [ eq : shell ] ) , leading to total momentum : @xmath76 } = { p_{\rm sf}}{\left[{1 + 4.6(t_*^{1/7}-1)}\right]},\ ] ] where @xmath77 and @xmath78 .
although equation ( [ eq : ppds ] ) suggests only a very slow growth of momentum with time , the increase over @xmath79 would in principle be substantial if the interior of the remnant remained hot for @xmath80 .
however , the above idealized treatment of the post - shell - formation evolution does not agree with detailed time - dependent solutions , and momentum injection in practice increases only modestly subsequent to shell formation . as we shall show in the following sections ( see also * ? ? ? * ) , the classical pds assumption of a hot , adiabatic interior with no mass loss does not apply .
in fact , the mass of hot gas interior to the shell steadily decreases in time , because the outermost part flows into the shell and cools .
since the pressure decreases more rapidly than the simple adiabatic expectation , there is only a brief , cooling - modified pds stage . the momentum injection after shell formation is quite limited , amounting to only @xmath81 of @xmath79 .
we solve the hydrodynamics equations with cooling using the _ athena _ code , which employs an unsplit godunov algorithm @xcite . among the various solvers provided , we utilize the simple muscl - hancock type predictor - corrector scheme ( van leer integrator ; @xcite ) .
for robustness , we utilize piecewise linear spatial reconstruction and roe s riemann solver with h - correction @xcite and first order flux correction @xcite .
we apply h - correction only for zones with maximum signal speed difference ( @xmath15-coefficient in @xcite ) larger than @xmath82 , to avoid artificial diffusion at the interface of the cnm and wnm where the signal speed difference is about @xmath83 .
we solve the cooling term in an operator split manner using a fully implicit method , with newton - raphson root finding .
in addition to the courant condition , the time step is also limited by the cooling solver , which allows only a factor of two variation in the local temperature from the value at the previous time step @xcite .
the governing equations are @xmath84 @xmath85 and @xmath86 where @xmath87 is the mass density , @xmath88 is the velocity , @xmath89 is the total energy density , @xmath90 is the gas pressure , and @xmath91 is the ratio of specific heats .
the gas temperature is @xmath92 for neutral gas and @xmath93 for fully ionized gas , where the hydrogen number density is @xmath94 for @xmath95 of helium abundance .
since we did not follow the ionization and recombination in detail , we calculate the gas temperature using the equation for the neutral gas so that the temperature in our simulations are higher by factor of @xmath96 for ionized gas ( @xmath97 ) .
however , the temperature in practice is only used to obtain the cooling rate , and only indirectly affects the dynamical evolution .
the net cooling rate per unit volume is @xmath98 $ ] . for the cooling function @xmath99 at low ( @xmath100 ) and high ( @xmath101 ) temperature gas
, we respectively adopt fitting formulae from ( * ? ? ?
* see also @xcite ) and piecewise power - law fits to the cooling function of @xcite with collisional ionization equilibrium at solar metallicity .
heating is only applied at @xmath102 , to model photoelectric heating of the warm / cold ism . as the photoelectric heating rate is proportional to the star formation rate , and the pressure and star formation rate are approximately proportional to each other for self - consistent solutions @xcite , we adopt a heating rate that varies with mean density as @xmath103 , where @xmath104 .
we neglect thermal conduction in this study ( see discussion in section [ sec : sumndis ] ) .
we have run sets of simulations of three different types : ( 1 ) a single sn in a uniform background medium ( su ) , ( 2 ) a single sn in a two - phase medium ( s2p ) , and ( 3 ) multiple sne in a two - phase medium ( m2p ) .
for all three types of simulation , we vary the mean density of the background medium @xmath105 from 0.1 to 100 ( 10 for m2p ) . for our standard models ,
we fix the total energy of a single sn to be @xmath106 . only for comparison to m2p models
, we also have run a single - explosion model with @xmath107 ( e2p ) .
since our main interest is on the momentum injection to the ism at later times , we ignore the ejecta and hence the free expansion stage .
we initialize the snr in all cases within a sphere of radius @xmath3 ( @xmath3 varies ; see below ) .
the snr initialization region consists of all cells whose centers are at a distance @xmath108 from the site of sn explosion ; within the snr initialization region , we either inject thermal energy at uniform density ( most models ) , or apply a radially - dependent st solution . in order to quantify the evolution , we require definitions for the different gas components .
we define the hot gas as all zones with @xmath109 , the shell gas as zones with @xmath110 and @xmath111 , and the ambient medium as the remainder ( note that ambient gas is initially stationary for these models ) . for the ambient medium , we define the cold and warm neutral medium ( cnm and wnm ) with temperature cuts of @xmath112 and @xmath113 , respectively .
hereafter , `` snr '' refers to both hot and shell gas .
we define the radius of the snr using the mass - weighted mean radius of the shell gas as @xmath114 where @xmath115 .
this is approximately equal to the shock radius before shell formation , and slightly smaller ( due to non - negligible thickness of the shell ) than the shock radius after shell formation ( see figure [ fig : rprof ] ) .
the kinetic and thermal energies are calculated as @xmath116 and @xmath117 dv$ ] for the snr ( hot and shell gas ) , and the total energy @xmath118 .
the total radial momentum of the snr is calculated by @xmath119 .
we also measure the total mass of the shell and hot gas , @xmath120 and @xmath121 , respectively , as well as the mean pressure of the hot gas @xmath122 .
we first perform a set of numerical experiments with a single sn in a uniform , unmagnetized medium ( su models ) . ) at late times
. see appendix for details . ] for reference , we list in table [ tbl : uniform ] the theoretical estimates of the shell formation time @xmath49 as well as the theoretical snr radius , postshock temperature , swept - up mass , and total radial momentum at @xmath49 ( see equations ( [ eq : tsf ] ) , ( [ eq : rsf ] ) , ( [ eq : tsf ] ) , ( [ eq : msf ] ) , and ( [ eq : psf ] ) ) , for each choice of ambient density in the range @xmath123 0.1 to 100
. as we shall describe in section [ sec : single_conv ] , in addition to varying the physical density of the ambient medium , for our su models we also cover a wide range of parameter space for the initial radius of the snr @xmath3 and the grid spacing @xmath2 .
this latter set of tests enables us to evaluate the dependence of the momentum injection on numerical resolution , and therefore establish minimum resolution requirements for modeling sn feedback in complex numerical simulations .
table [ tbl : uniform ] also lists ( in parentheses ) the numerically - measured values ( @xmath124 , @xmath125 , @xmath126 , and @xmath127 ) obtained from high - resolution simulations , as described in section [ sec : single_ref ] .
lccccccc su - n0.1 & 0.1 & 156(150 ) & 59.5(58.5 ) & 0.31 & 3.08(2.96 ) & 2.94(2.78 ) + su - n1 & 1 & 43.7(41.9 ) & 22.6(22.5 ) & 0.57 & 1.68(1.68 ) & 2.17(2.05 ) +
su - n10 & 10 & 12.2(10.6 ) & 8.56(8.34 ) & 1.04 & 0.92(0.85 ) & 1.60(1.43 ) + su - n100 & 100 & 3.43(2.63 ) & 3.25(3.03 ) & 1.90 & 0.50(0.41 ) & 1.18(0.97 ) in this subsection , we delineate the time evolution of the snr based on high resolution simulations of a single sn in a uniform ambient medium ( su models ) .
we first present results of su - n1 with grid resolution @xmath128 .
the initial conditions are realized by uniformly applying the sn s thermal energy ( @xmath129 ) within an initial radius of @xmath130 .
figure [ fig : tevol ] plots time evolution of ( a ) radius @xmath131 , ( b ) deceleration parameter @xmath132 , where @xmath133 is calculated by finite differences of the measured @xmath131 , and @xmath15 is smoothed , ( c ) total @xmath134 , thermal @xmath135 , and kinetic @xmath136 energies , ( d ) mass of hot gas @xmath137 and shell gas @xmath62 , ( e ) total radial momentum @xmath138 , and ( f ) pressure of hot gas @xmath70 .
the dotted and dashed lines indicate theoretical predictions of each quantity for st and idealized pds solutions as described in section [ sec : theory ] , while the vertical dot - dashed line denotes the predicted shell formation time from equation ( [ eq : tsf ] ) .
we also present in figures [ fig : rprof ] and [ fig : snap ] the radial profiles and snapshots in the @xmath139-@xmath140 plane at @xmath141 ( xy - plane ) for selected times . in figure
[ fig : rprof ] , blue dashed lines show the st self - similar solutions at @xmath142 , while vertical dotted lines are @xmath131 at each epoch . at @xmath143 , the dynamical evolution of the snr is characterized by the st solutions very well .
the initial snr in the simulation has a finite size and zero kinetic energy , in contrast to an initial point source and a ratio @xmath144 of the st solution .
however , as the shock expands into the ambient medium , radial motions are rapidly generated .
after a transient of only @xmath145 , all the physical quantities follow the st solution . although the radial profiles shown in figure [ fig : rprof ] differ from the st solutions at small radii @xmath146 due to the finite initial size , they are quite close near the shock where the most of mass and momentum are concentrated . as a consequence
, there is excellent agreement with the st solutions for all integrated physical properties , as seen in figure [ fig : tevol ] .
as the shocked hot gas cools , a thin and dense shell is formed ( see green and red lines of figure [ fig : rprof ] and the 2nd and 3rd columns of figure [ fig : snap ] ) .
shell formation is clearly distinguished by sudden changes of several physical quantities . at shell formation , the expansion stalls ( sudden drop of @xmath133 and hence @xmath15 in figure [ fig : tevol](b ) ) ,
the thermal energy and the mass of the hot gas begin to decrease ( figure [ fig : tevol](c ) and ( d ) ) , and the mass of the shell gas sharply increases ( figure [ fig : tevol](d ) ) .
we define the numerical shell formation time , @xmath124 , as the time when @xmath137 attains its maximum , to identify shell formation ; our definition based on the hot gas mass is in practice quite similar . ] .
note that the theoretical prediction of @xmath49 from equation ( [ eq : tsf ] ) agrees with the maximum of @xmath137 fairly well ( see vertical line in figure [ fig : tevol](d ) and table [ tbl : uniform ] ) .
since the cooling function we use is different from the single power - law cooling function used to derive equation ( [ eq : tsf ] ) , @xmath124 is slightly different from @xmath49 , with slightly different dependence on @xmath11 and @xmath147 as well .
after shell formation , the pressure of hot interior gas decreases much faster than for the idealized adiabatic pds solution , @xmath148 ( dashed line in figure [ fig : tevol](f ) ) .
this difference was reported by @xcite , who found an additional @xmath149 dependence in thermal energy and hence the pressure of the hot gas ( after shell formation ) .
although the cooling time of the hot gas at small radii is still quite long because of its low density , the density immediately inside the shell is higher and this gas is able to cool . in addition , the velocity interior to the shell exceeds that of the shell itself ( see red , cyan , and magenta lines in figure [ fig : rprof](d ) ) , so that interior hot gas continues to accumulate , condense , and cool at the inner surface ( back ) of the shell . with a steadily decreasing mass of hot gas
, the interior pressure is ever lower than that of the idealized pds solution ( which assumes a constant interior mass ) , adding much less momentum ( compare dashed and solid lines in figure [ fig : tevol](e ) ) .
the evolution of radius with time nevertheless seems to be roughly consistent with the pds exponent @xmath150 ( or @xmath151 for the offset power - law fit in @xcite ; see figures [ fig : tevol](a ) and ( b ) ) , but this can also be explained as a transition from the st to mcs phase , i.e. from @xmath152 to @xmath153 .
since the hot gas mass drops very rapidly , the duration of the `` classical '' pds stage is negligible .
as most of the hot gas has already piled into the shell by @xmath154(2 - 3)@xmath49 , the hot gas pressure is no longer high enough to drive the shell outward and add appreciable momentum ( cyan and magenta lines in figure [ fig : rprof](b ) ; figure [ fig : tevol](e ) ) .
in addition , the shell becomes thicker at this stage so that the thin shell approximation is no longer valid .
the shell is broken up in this high resolution simulation ( fourth column of figure [ fig : snap ] ) due to the pressure - driven thin shell overstability and/or nonlinear thin - shell instability @xcite . although we do not explicitly assign perturbations to our models , at the initial interface between the snr and ism there are unavoidably grid - scale nonlinear perturbations due to the mapping of a sphere to cartesian coordinates .
this seeds instability , which develops into strong corrugations .
note that this instability is less prominent for realistic cases , in which the multiphase structure of the ambient ism dominates in shaping the shell ( see section [ sec : multi_ti ] ) . in models
su - n0.1 , su - n10 , and su - n100 , we conduct analogous simulations to model su - n1 , covering the range of ambient density of @xmath155 to @xmath156 . for the high - resolution models with @xmath155 , 1 , 10 , and 100 , we set @xmath157 , 1/2 , 1/4 , and 1/8 pc , respectively , and @xmath158 ; these choices are based on our convergence study ( see section [ sec : single_conv ] ) .
figure [ fig : tevol_norm ] plots the same quantities as in figure [ fig : tevol ] , but normalized using the corresponding numerical measures at @xmath124 . as before ,
this @xmath124 is defined by the time when the hot gas mass attains its maximum value .
note that @xmath124 , @xmath125 , @xmath126 , and @xmath127 given in table [ tbl : uniform ] for these high - resolution models can be fitted by , respectively , @xmath159 @xmath160 @xmath161 and @xmath162 figure [ fig : tevol_norm ] demonstrates that the overall ( normalized ) evolution is remarkably similar , irrespective of the ambient density .
since the normalization factors are close to the analytic estimates from the st solution and the early evolution is expected to converge to the st solution , this agreement is trivial at @xmath163 . however , even though the classical pds solution is not realized in the numerical models , the characteristic behavior after shell formation is nevertheless essentially the same over a wide range of densities .
model evolution is thus `` congruent '' for varying ambient density when rescaled relative to the snr properties at @xmath124 . in order to describe the transition between st and pds phases smoothly
, @xcite introduced an analytic description using an offset power - law for radius and a complex fitting model for the thermal energy with additional time dependence .
since energy loss by radiative cooling begins somewhat earlier than the shell formation , they also arbitrarily defined a pds time @xmath164 . here , instead of using the approach of @xcite , we introduce a simpler analytic formula to describe the post - shell formation phase separately .
as already seen in figure [ fig : tevol ] , the evolution of radius agrees reasonably well with the classical pds solution .
we thus simply adopt @xmath165 ( other exponents between 0.4 and 0.25 give similar results ) .
following @xcite , we describe the evolution of thermal energy after shell formation with an additional time dependence as @xmath166 where @xmath167 is the thermal energy at the st stage , and the factor `` 0.8 '' arises from precooling before the shell formation ( see dotted lines in figure [ fig : tevol_norm](c ) ) .
the factor @xmath168 is chosen to give a good match to the numerical solutions for the time dependence of @xmath135 .
one might expect @xmath135 to decrease after shell formation proportional to the product of @xmath137 and the interior temperature .
the mass - weighted mean temperature of hot gas is nearly constant in this phase .
thus , @xmath137 and @xmath135 show similar time dependence ( approximately @xmath169 ) until the thermal energy of shell gas begins to dominate at @xmath170 . using equation ( [ eq : eth ] ) , the effective pressure on the shell is then @xmath171 we then have an expression for the momentum similar to equation ( [ eq : ppds ] ) : @xmath172 } = { p_{\rm sf}}{\left[{1 + 0.62(1-t_*^{-6/7})}\right]}.\ ] ] note that with this modified , non - adiabatic pds stage , the total momentum is finite , @xmath173 as @xmath174 ; only 60% of @xmath79 is added during the post shell formation stages .
the black dashed lines in figure [ fig : tevol_norm ] represent the modified pds solution described above .
the final momenta ( @xmath175 ) measured at @xmath176 are in fact still smaller than @xmath177 since no momentum is acquired after @xmath178 as the hot gas pressure becomes comparable to the shell gas pressure .
we find that @xmath179 , 1.46 , 1.48 , and 1.45 for @xmath155 , 1 , 10 , and 100 , respectively . in physical units ,
the final momentum can be fitted as @xmath180 it is interesting to compare our numerical results with earlier studies based on one - dimensional spherical models . @xcite
have measured the shell formation time in terms of the `` transition '' time ( equation ( 3 ) of their paper ; simply defined by equating the cooling time with the age of the snr ) . for @xmath181 , 0.84 , and @xmath182 , their numerical measures of the shell formation time
are 1.35 , 1.23 , and 1.61 times our @xmath124 for @xmath155 , 1 , and 100 , respectively .
a wide parameter space of density and metallicity has been covered by @xcite . in their work ,
the numerical measures of the shell formation time ( @xmath183 in their notation ) , agree with our @xmath124 within 20% for corresponding density .
more interestingly , their ratios of total final momentum ( measured at @xmath184 ) to the momentum at @xmath183 are 1.5 - 1.6 for @xmath185 to @xmath186 , similar to our findings . , where @xmath187 and @xmath188 are mass and kinetic energy of the snr , respectively .
note that this definition is consistent with direct integration of the specific momentum if most of the mass is concentrated in a narrow region , which is true for both st and post - shell formation phases .
] even for more extreme parameters with @xmath189 and @xmath190 or different metallicities , the ratios still remain around 1.7 - 2 . as shown in the previous section , with sufficiently high resolution and a small enough initial size for the snr , we are able to reproduce in our 3d simulations the evolution of radiative snr as seen in previous 1d spherical models ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
our models show a clear st stage , but the post shell formation phase differs from the `` classical '' pds solution , similar to findings from previous numerical work . within @xmath191 , the total radial momentum of the snr asymptotes to a value @xmath192 times the momentum at the time of shell formation .
the above high - resolution su simulations provide a benchmark , and in this subsection , we repeat the su simulations for a wide range of numerical grid resolution ( @xmath2 ) and initial size of snr ( @xmath3 ) .
our goal is to determine minimum numerical requirements needed to reproduce the main features of expanding snrs on cartesian grids , for the purposes of tracking sn feedback effects in large - scale ism and galactic models . to test dependence on grid resolution @xmath2 , we compare results for key physical quantities in model su - n1 at @xmath193 , 1 , 2 , and 5 . figure [ fig : dxnorm ] plots ( a ) total momentum , ( b ) kinetic energy , ( c ) hot gas pressure , and ( d ) hot gas mass , for a range @xmath194 to 1.5 .
we assign @xmath195 as @xmath2 varies , such that the initial thermal energy of sn is distributed among eight zones nearest to the center of the simulation domain ( the sn explosion occurs at a grid corner ) .
all physical quantities are normalized by the values of the st solution at @xmath49 . here
, we use the analytic estimates of @xmath49 and other variables rather than numerically determined values since the numerical measures vary from model to model , whereas for a resolution study we require a fixed reference . for @xmath196 and 1/3 ,
both the momentum and kinetic energy are converged respectively within @xmath197 and @xmath198 of the highest - resolution results at every evolutionary stage . at late stages after shell formation , the hot gas mass depends on the resolution since the mass is concentrated near the boundary of the hot and shell gas ( and is sensitive to the exact definition of `` hot '' vs `` shell '' at later times when the hot gas mass is small ) , but shows very good convergence at @xmath199 for @xmath200 .
the convergence of @xmath70 , which is volume weighted , is as good as @xmath201 and @xmath202 for @xmath196 and 1/3 , respectively , omitting @xmath203 when the hot gas has negligible effect on dynamics . for @xmath204 ,
the snr cools down completely in a short time , so that there is no hot gas even at @xmath205 ; the momentum is @xmath206 and kinetic energy is @xmath207 of the converged values .
thus , the dynamical effect of sn feedback is impossible to model via injection of thermal energy if the grid resolution exceeds @xmath208 ( see also * ? ? ?
this is the origin of the `` overcooling '' problem cited in many numerical simulations of galaxy formation ( e.g. , * ? ? ?
* ) , which at best resolve scales several tens of pc and above .
in addition to grid resolution , prescriptions for sn feedback in large - scale simulations that are based on thermal energy injection must also specify an initial snr radius , @xmath3 . by varying @xmath209 ,
we also have assessed convergence with respect to this numerical parameter . for these tests
, we use the su - n1 model and adopt @xmath210 .
figure [ fig : rnorm ] plots ( a ) radial momentum , ( b ) kinetic energy , ( c ) hot gas pressure , and ( d ) hot gas mass , at several different times , as a function of initial radius ( or initial enclosed mass @xmath211 ) .
it is clear that converged results are obtained at every evolutionary stages for @xmath212 . although momentum and kinetic energy that are not too far from correct values can also be obtained with @xmath213 , the evolutionary history and internal profiles are completely different from the converged , resolved solutions if @xmath214 .
we have confirmed that the convergence criterion @xmath215 is valid for different resolutions and ambient density .
figures [ fig : rnorm_dx ] and [ fig : rnorm_n ] plot the same quantities as in figure [ fig : rnorm ] at @xmath216 , for different resolutions @xmath2 and ambient density .
similar convergence trends are shown for lower resolution simulations and for a wide range of varying ambient density .
with poorer resolution , all quantities converged to smaller values than in the high - resolution benchmark models , as seen in figure [ fig : dxnorm ] .
if , instead of initializing with pure thermal energy inside @xmath3 , the initial conditions instead employ the st profile , then the final momentum is converged out to @xmath217 . however , initializing with the st profile results in momentum exceeding the benchmark value if @xmath3 is comparable to or greater than @xmath208 .
we conclude that in order to obtain both convergence and accuracy ( consistent convergence ) , snr must be initialized with both small enough radius @xmath215 , and high enough grid resolution .
the converged values remain within roughly @xmath218 and @xmath219 of benchmarks for all physical quantities with @xmath196 and 1/3 , respectively . in terms of injected momentum , @xmath196 and 1/3 respectively give converged values within 5% and 13% of benchmarks at @xmath199 , and 18% and 25% of benchmarks at @xmath220 .
the final momentum is marginally satisfactory when @xmath213 and @xmath221 are adopted .
any value @xmath222 , initialized with either pure thermal energy or a st profile , gives poor results .
lccccccccc & whole & 0.1 & & & 156(384 ) & 59.5(83.3 ) & 0.31 & 3.08(3.44 ) & 2.94(3.21 ) + & cnm & 1.5 & 0.06 & 0.81 & 36 & 19 & 0.62 & 1.5 & 2.1 + & wnm & 0.017 & 0.92 & 0.15 & 420 & 130 & 0.19 & 4.9 & 3.7 + & whole & 1 & & & 43.7(100 ) & 22.6(31.5 ) & 0.57 & 1.68(1.44 ) & 2.17(2.13 ) + & cnm & 8.9 & 0.09 & 0.81 & 13 & 9.0 & 1.0 & 0.94 & 1.6 + & wnm & 0.14 & 0.84 & 0.12 & 130 & 52 & 0.34 & 2.8 & 2.8 + & whole & 10 & & & 12.2(24.3 ) & 8.56(10.7 ) & 1.04 & 0.92(0.70 ) & 1.60(1.46 ) + & cnm & 110 & 0.08 & 0.83 & 3.3 & 3.2 & 1.9 & 0.49 & 1.2 + & wnm & 1.5 & 0.88 & 0.13 & 35 & 19 & 0.64 & 1.5 & 2.1 + & whole & 100 & & & 3.43(5.30 ) & 3.25(3.63 ) & 1.90 & 0.50(0.36 ) & 1.18(1.00 ) + & cnm & 1300 & 0.07 & 0.82 & 0.85 & 1.1 & 3.7 & 0.26 & 0.85 + & wnm & 17 & 0.91 & 0.15 & 9.2 & 6.9 & 1.2 & 0.80 & 1.5 in the real ism , the ambient gas around sne is not uniform , but clumpy .
the immediate environment will be a photoionized region that is also impacted by the progenitor s stellar winds . at larger scales , comparable to @xmath208
, sne may propagate either into dense molecular gas of a parent cloud ( if it has not yet been dispersed ) , into the cloudy warm / cold atomic medium , or into very low density hot gas produced by previous generations of snrs .
the case of propagation into the warm / cold atomic ism is especially important to quantify , as snrs are believed to be the main source of turbulence in this phase of the ism .
the atomic ism ( @xmath102 ) is regulated by two dominant cooling processes , atomic fine structure line cooling and ly@xmath223 , yielding a two - phase medium with temperature near @xmath224 and @xmath225 , respectively ; grain - photoelectric net heating dominates nearly irrespective of the gas temperature @xcite . for a range of pressures ,
there is a third possible equilibrium temperature in between the warm and cold phases , but this has relatively low occupation due to the thermal instability , with a growth time @xmath226 that is shorter than large - scale dynamical times in the ism @xcite .
this results in an ism structure consisting of cold , dense clouds embedded in a background of warm , rarefied gas .
the warm and cold medium densities differ by two orders of magnitude . to study propagation of snrs into atomic ism gas , we first construct two - phase `` background states '' via nonlinear simulations of the thermal instability ( ti runs ) .
the initial conditions for each ti run is a thermally unstable equilibrium with @xmath95 density perturbations . for each different mean density ( @xmath227 ) , @xmath228 ( as in the su models ; this scaling is consistent with self - regulated star formation ) , which makes the initial state thermally unstable .
are more likely to be found in the molecular ism than the atomic ism . however , as the chemical state of the ambient gas is unimportant when shocks are strong , it is useful to consider this case as representative of the high - density , clumpy conditions that are present in the centers of normal disk galaxies and more pervasively in starbursts . ]
we follow each ti run long enough to obtain a saturated two - phase state with pressure balance .
we then explode a sn at the center of the simulation domain . as the details of the cloudy structure that a snr encounters affects its propagation , for each mean density in the s2p model
set we use 10 different realizations of the cloudy background state ( from ti runs with different perturbation seeds ) .
table [ tbl:2p ] lists from 3rd to 5th columns the mean densities ( @xmath229 ) , volume fractions ( @xmath230 ) , and mass fractions ( @xmath231 ) of cnm ( @xmath112 ) and wnm ( @xmath113 ) gas in the saturated state , averaged over these 10 realizations .
the temperature cuts are determined by the maximum and minimum temperatures of the stable cnm and wnm , respectively , for our choice of cooling function ( see * ? ? ?
* ; * ? ? ?
variations of the mean values in different realizations are less than 1% .
note that the sum of @xmath230 and @xmath231 is not 1 since a small fraction of gas resides in the cloud interface . as @xmath228 ,
the volume and mass fractions of the two phases are similar at all @xmath147 , and the densities in each phase are proportional to @xmath147 .
table [ tbl:2p ] also lists for reference the physical quantities at shell formation that would apply for the st solution propagating into _ uniform _ high - density ( cold cloud ) gas , _ uniform _ low density ( intercloud ) gas , and gas at the average density of the simulation .
numerical measures of the quantities averaged over 10 realizations are also shown in parentheses .
as before , we define @xmath124 as the time when the hot gas mass reaches a maximum .
power - law fits for averaged numerical measures are given by @xmath232 @xmath233 @xmath234 and @xmath235 note that the above is the mass of _ hot _ gas at @xmath124 ; there is several times as much `` shell '' gas because initially - cold , dense clouds are accelerated by the shock and by the expanding gas in the interior of the snr .
compared to the uniform - medium case ( equations ( [ eq : tsfnum])-([eq : psfnum ] ) ) , dependencies on the mean ambient density are similar .
although the shell formation epoch @xmath124 is delayed for a two - phase medium , the hot gas mass and total momentum at shell formation agree very well with the numerical measures of these quantities for a snr in a uniform background ( su models ) , as well as analytic estimates .
figure [ fig : snap_2p ] displays snapshots in the xy plane of an example run for model s2p - n1 . the time marked in the top row is the time since the explosion . for this run
, we set the grid resolution to @xmath236 and the initial size of sn to be @xmath237 , for an @xmath208 estimated using the mean density .
for this specific example , the saturated two - phase ism is characterized by @xmath238 and @xmath239 for cnm and wnm , respectively .
the simulation volume is mainly occupied by the wnm with @xmath240 , while the majority fraction @xmath241 of the mass resides in the cnm .
since the snr expands solely through the wnm in most directions , the overall size of snr is similar to expectations for a uniform medium with @xmath242 ( especially in the 2nd quadrant of the xy plane in figure [ fig : snap_2p ] ) . in some directions ,
the blastwave encounters small cnm clouds ; these may be destroyed by the passage of the shock and/or form pillars within the snr .. ] large cnm clouds are also shock compressed , and may partly block the snr expansion , but over time they are also carried outward .
at late times , the shell morphology is mainly shaped not by the shell instability observed in figure [ fig : snap ] but by the initial structure of the ambient two - phase medium . for s2p - n1 ( 10 different realizations ) ,
figure [ fig : tevol_2p ] plots time evolution of ( a ) total , thermal , and kinetic energies , ( b ) hot and shell gas masses , ( c ) radial momentum , and ( d ) hot gas pressure . also indicated with vertical dot - dashed lines are the st solutions for the shell formation time when propagating into the cnm density ( cyan ; @xmath243 ) and wnm density ( magenta ; @xmath244 ) .
dotted lines in panels ( b ) , ( c ) , ( d ) give st solutions for an ambient medium that is 100% cnm ( cyan ) and 100% wnm ( magenta ) .
beginning at @xmath245 , part of the snr s energy is radiated away as the blastwave envelopes and shocks cnm clouds ; both thermal and kinetic energy are reduced below the st values .
since the most of volume is filled with the wnm , however , all physical quantities except the mass remain close to those of the st solution in a uniform wnm until @xmath246 .
if there were no pds stage , the momentum injected to a two - phase medium would be expected to lie between @xmath79 for @xmath247 and @xmath248 ( last column of table [ tbl:2p ] ) . as @xmath79 is insensitive to the ambient medium density ( @xmath249 predicted from equation ( [ eq : tsf ] ) , or @xmath250 using the numerical results in table [ tbl : uniform ] ) , and the momentum only modestly increases during the pds stage , the total injected momentum by a single sn will always be of order @xmath79 for the mean density of the medium , @xmath147 . figure [ fig : snap_s2p ] displays example xy - slices for s2p models with different mean density at @xmath251 .
although the initial cold and warm volume fractions are similar for all models , the cloud structures are different ; there are more small clouds in higher density models . for the present models ,
the difference in structure at ti saturation is caused by differences in ti growth rates ; smaller clouds can grow faster with higher heating rates in higher density models . the effective contribution from dense clouds increases with background density because of the increasing total cloud cross - section when there are more clouds .
thus , the numerical measures of physical quantities ( in parentheses of table [ tbl:2p ] ) are closer to the analytic estimates for cnm as the mean density gets higher
. however , the total injected momentum remains similar to the estimates from the mean density .
for ten different realizations at each mean density @xmath147 , figure [ fig : rnorm_2p ] plots the radial momentum at @xmath199 ( blue ) , 5 ( green ) , and 10 ( red ) , where @xmath49 is computed using @xmath147 .
similar to figure [ fig : rnorm ] , we show the result of varying the initial radius of the snr . at each @xmath147
, we use a reference value of @xmath208 based on equation ( [ eq : rsf ] ) using @xmath147 ( note , however , that the actual mean density within this @xmath208 will differ from @xmath147 due to the particular realization of the cloudy structure ) .
the resolution @xmath2 is set to be smaller than @xmath252 , which we have found ( using tests similar to those of section [ sec : single_conv ] ) is sufficient for convergence .
the horizontal axis is normalized by radius @xmath208 using @xmath147 ( see equation ( [ eq : rsf ] ) ) .
the final momentum converges provided @xmath253 , similarly to what we found in the uniform medium case .
this `` consistent convergence '' criterion guarantees convergence at all evolutionary stages ; convergence at late stages only is possible up to slightly larger @xmath3 . for @xmath254 ,
the final momentum is significantly lower than the converged value . for mean number density of @xmath155 , 1 , 10 , and 100 ,
the mean values of the momentum at @xmath255 are @xmath256 , 1.30 , 1.29 , and 1.32 , respectively . similarly to equation ( [ eq : pfinal ] ) , the final momentum in the physical units can be fitted by @xmath257 the final momentum is always of order of @xmath79 , but slightly smaller for a clumpy medium than in the uniform - medium case . for different realizations of the model at each mean density ( i.e. different seeds for the random perturbations ) , there are only 20 - 40% variations in the final momentum .
in sections [ sec : single_ref ] , [ sec : single_conv ] , and [ sec : single_2p ] , we considered the expansion of a single snr in either a uniform or a clumpy medium for a wide range of mean densities ( @xmath258 ) , covering conditions that would apply in the atomic or molecular ism .
there , we evaluated the momentum , mass of hot gas , and other basic properties of the snr , and determined numerical requirements for consistent convergence , i.e. criteria such that simulation results agree with high - resolution benchmarks , at times before and after the snr becomes strongly radiative . in this section , we extend to the case of multiple sne that sequentially explode and expand into a surrounding two - phase medium . for multiple sne , many more defining parameters enter the problem setup than for the case of a single sn , so a comprehensive study will be deferred to future work .
however , it is useful to explore a few simple cases to assess potential differences in the momentum injection per event , compared to that from a single sn .
of course , the momentum injection from multiple spatially correlated sne is also subject to global effects , since the superbubble that forms can expand beyond the scale height of the surrounding ism and break out of the disk , venting into the galactic halo and limiting further expansion within the ism .
here , we do not consider these global effects ( but see section [ sec : sumndis ] for further discussion ) .
we consider the case in which sne occur at intervals of 0.1myr ; this could represent the effects of a moderate - mass cluster ( @xmath259 ) in which the event rate is not sufficient for energy deposition to reach the continuous limit .
we consider a range of mean densities @xmath260 in the ambient medium , and as for the s2p models we create a `` background '' two - phase state by allowing the ti to reach saturation . for the single - sn models , the initial size of the snr
was set relative to @xmath208 , the shell formation radius at the mean density of the ambient medium .
although for our tests @xmath147 has a pre - determined value and it is trivial to set @xmath3 directly , more generally the ambient density varies and it is more practical to define the initial snr region based on its total enclosed mass .
we take this approach for the m2p models . to obtain an equivalent condition to @xmath215 ,
we begin with @xmath195 ( thermal energy is distributed in eight zones closest to the center ) and calculate the enclosed mass @xmath261 and the mean density @xmath262 , where @xmath263 . if @xmath264 ( where @xmath265 is obtained with equation ( [ eq : msf ] ) for this computed @xmath13 ) we increase @xmath3 by @xmath266 and repeat the calculation in order to get the largest possible initial size of snr that is compatible with the convergence condition @xmath215 ( i.e. @xmath264 ) . for each sn event , once a sphere with appropriate @xmath3 is defined , we redistribute the enclosed mass , momentum , and internal energy uniformly within @xmath3 , and assign additional thermal energy adding up to @xmath11 .
this prescription prevents setting the temperature to extremely high values for very rarefied regions produced by previous sn explosions .
all sne explode at the center of the simulation domain . for our m2p model set
, we run each simulation until 10 sne , with @xmath106 each , have exploded . for reference
, we also run comparison simulations with a single explosion of energy @xmath107 ( e2p ) . after the last sn ( at @xmath267 ) ,
we run each simulation out until @xmath268 to measure the final injected momentum .
figure [ fig : snap_m2p ] displays xy - slices of the m2p models at final stage of evolution .
the morphology of the sn bubble is determined by the structure of the ambient medium .
due to the short cooling time in model m2p - n10 , the hot gas persists only deep inside of the sn bubble ; this is in contrast to model m2p - n0.1 , where the shell is still thin and the hot gas fills up the interior of the sn bubble .
figure [ fig : tevol_multi ] plots time evolution of ( a ) the mean radius of the combined snr @xmath131 , ( b ) the effective radius of hot gas volume @xmath269 , ( c ) momentum @xmath138 , ( d ) hot gas mass @xmath137 , ( e ) thermal energy @xmath135 , and ( f ) kinetic energy @xmath136 .
thick lines show m2p and thin lines e2p models , with mean ambient density @xmath155 ( blue ) , 1 ( green ) , and 10 ( red ) .
note that @xmath270 at early stages when the snr is filled with hot gas since the mass - weighted radius is closer to the shell radius in the cnm , and @xmath271 at later stages when the hot gas is depleted ( see right column in figure [ fig : snap_m2p ] ) . evidently , multiple sne are able to inject more momentum when the ambient density is lower , similar to the case of single sne .
however , the dependence on ambient density is even weaker than in the single sn case .
the final momenta for @xmath272 1 , and 10 are @xmath273 , 1.7 , and @xmath274 , respectively . there is negligible increase in momentum shortly after the final sn event , for all cases . comparing multiple sne to a single sn with the same total energy , the ratios of the total momentum from m2p models to e2p models are 61% , 85% , and 98% for @xmath155 , 1 , and 10 , respectively .
the final momentum per sn event fitted to these results is @xmath275 .
the differences among the m2p models can be understood partly based on the ratio of the sn time interval @xmath276 to the shell formation time @xmath49 for the ambient medium .
the interval @xmath277 corresponds to @xmath278 , @xmath279 , and @xmath280 , for @xmath155 , 1 , and 10 , respectively . for m2p - n10 ,
new sn energy is injected only after completely exhausting the previous sn , since the interval is much longer than the shell formation time @xmath281 in the ambient medium .
the injection of momentum is thus in discrete events , and the mass of hot gas never reaches values much larger than that from a single sn event with @xmath106 .
model e2p - n10 has a much higher peak @xmath137 than that of m2p - n10 , because @xmath265 is nearly linearly proportional to @xmath11 ( see equation ( [ eq : msf ] ) ) .
however , @xmath137 drops rapidly after the early shell formation for model e2p - n10 . in model m2p - n10 ,
hot gas is replenished after each new sn event , resulting in larger occupying volume than e2p - n10 . both mass and volume of hot gas decrease after the last sn due to depletion of hot gas by cooling , while the mean radius of the snr keeps increasing . for m2p - n0.1 ,
the evolution is quite different from m2p - n10 .
the mass of hot gas reaches nearly an order of magnitude higher than for m2p - n10 , and as the hot medium does not fully cool in between events , the momentum injection becomes almost continuous rather than discrete .
the peak in @xmath137 occurs at nearly the same time for m2p - n0.1 as for e2p - n0.1 , although for both the hot gas is rapidly depleted after this time .
the volume occupied by hot gas keeps increasing as @xmath131 for both models .
the evolution of model m2p - n1 is intermediate between that of m2p - n10 and m2p - n0.1 .
the mass of hot gas stays close to constant between the first and last sn , with only slight dips .
the hot gas volumes for m2p - n1 and e2p - n1 converge with each other .
the momentum acquisition for model m2p - n1 is therefore smoother than for model m2p - n10 , and more discrete than for model m2p - n0.1 .
utilizing the hydrodynamic code _ athena _ with optically thin radiative cooling , we simulate the expansion of radiative snrs on three - dimensional cartesian grids . for all models , each explosion
is numerically initiated by deposition of @xmath282 in the form of thermal energy within a volume of radius @xmath3 .
we consider three types of simulations : a single sn explosion in a uniform medium ( su models ; section [ sec : single ] ) ; a single sn explosion in a two - phase medium ( s2p models ; section [ sec : single_2p ] ) ; and multiple sn explosions in a two - phase medium ( m2p models ; section [ sec : multi_ti ] ) .
we cover a range of mean density @xmath227 . for two - phase models ,
the cloud and intercloud densities are respectively an order of magnitude above and below @xmath147 . with the su simulations , we first delineate the detailed evolution of radiative snrs , in comparison to analytic solutions and previous one - dimensional ( spherical ) numerical simulations . among other quantities
, we evaluate the time , radius , mass , and momentum of the snr at shell formation , and the final momentum of the expanding cooled shell .
we then conduct a wide , systematic parameter search to ascertain numerical conditions for `` consistent convergence '' ( convergence for all evolutionary stages ) of snr evolution .
we find that @xmath215 , where @xmath208 the size of snr at shell formation , guarantees consistent convergence , and the final momentum is within @xmath218 or @xmath283 of high - resolution benchmarks provided that the grid resolution is @xmath200 or @xmath196 , respectively .
we next conduct similar studies using more realistic ambient conditions , consisting of a two - phase ism characterized by irregular cnm clouds embedded in the wnm .
we show that the same numerical convergence conditions apply to the two - phase medium as the uniform medium , and evolution follows a similar course , with comparable hot gas mass and momentum injection .
finally , we investigate the momentum injection by multiple sne in a two - phase medium , using the same numerical criteria to set the initial snr size .
the details of evolution depend on the interval between sne compared to the shell formation time ( which decreases with increasing mean ambient density ) , but the final momenta of the expanding ( combined ) snrs are similar in all cases .
our main findings and implications are summarized below . *
_ stages of snr evolution _ + our simulations confirm the well - known evolutionary stages of a snr subsequent to early free expansion ( e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) : an energy conserving sedov - taylor ( st ) phase , shell formation when the remnant becomes radiative , a cooling - modified pressure - driven snowplow ( pds ) phase , and a final momentum - conserving expansion phase . in a uniform medium , the shell formation epoch
is well defined by the time at which the mass of hot gas attains its maximum , @xmath284 .
table [ tbl : uniform ] shows that the analytic estimate for @xmath49 from equation ( [ eq : tsf ] ) provides a reasonable match with our numerically - measured @xmath124 ; this also agrees with previous numerical findings for shell formation based on spherical one - dimensional simulations @xcite .
the radius , total snr mass , and outward radial momentum at shell formation are also close to analytic estimates .
+ in classical theory ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , an adiabatic pds stage ( with constant interior mass and @xmath285 ) follows shell formation , but this ideal situation is not realized in practice ( see also * ? ? ?
* ; * ? ? ?
instead , the hot interior expands into the back of the shell , where it condenses and cools . the mass of hot gas and interior pressure drop by an order of magnitude within a few times @xmath49 . with rapidly - falling interior pressure ,
the momentum added to the shell during the pds stage is modest ( see below ) .
figure [ fig : tevol_norm ] shows that for a uniform medium , snr evolution through st , shell formation , cooling - modified pds , and momentum - conserving stages is essentially congruent for varying ambient density , when normalized by the snr properties at @xmath124 .
+ for a two - phase ambient medium , figure [ fig : tevol_2p ] shows similar evolutionary stages to the case of a uniform medium . although energy losses commence earlier ( since dense clouds radiate after they are shock - compressed ) , the mass of hot gas reaches a maximum ( and then rapidly declines ) only when a strongly - cooling shell forms in the low - density volume - filling intercloud medium . the numerically - computed shell formation time is in between the values of @xmath49 computed using the mean ambient density and using the density of the intercloud medium ( table [ tbl:2p ] ) . a hot , low - density and low pressure bubble is left behind as dense cloudlets and an irregular shell of swept - up / cooled intercloud gas expand outward .
+ when multiple explosions occur sequentially in the same location , the evolution is either highly impulsive or relatively smooth depending on the interval between sne compared to the shell formation time in the ambient medium ( see figure [ fig : tevol_multi ] ) .
similar to the case of single sne , however , cooling of the interior implies that a constant asymptotic momentum is reached shortly after the last sn event .
+ most of our simulations consider evolution of snr in an unmagnetized medium .
however , in the appendix we include results for the evolution of magnetized models with a range of plasma beta parameter for the background medium ( @xmath286 in the range 0.1 to 10 ) . for @xmath287 and 10 ,
evolution is essentially indistinguishable from the unmagnetized case up to about @xmath288 . for the @xmath289 case , at late stages
the shell thickens and becomes anisotropic due to magnetic pressure and tension forces , and at very late stages the snr becomes elongated along the magnetic field ( e.g. , * ? ? ?
_ momentum injection by a single sn _ + our simulations show that most of the momentum of the expanding remnant is acquired during the st stage of evolution . for uniform ambient conditions ,
the final momentum is only @xmath219 larger than the momentum @xmath79 at shell formation , at all values of the ambient density ( see figure [ fig : tevol_norm ] ) . with the very weak dependence of @xmath79 on the ambient medium density ( see table [ tbl : uniform ] ) as predicted by equation ( [ eq : psf ] ) , the final momentum injected by a single sn in a uniform medium varies @xmath290 , yielding @xmath291 for @xmath292 ( see equation ( [ eq : pfinal ] ) ) . + we show in appendix that the momentum injection is the same for a magnetized medium as for the unmagnetized case , when the plasma beta parameter exceeds 0.1 .
this is because most of the momentum is acquired during stages when thermal pressure within the snr greatly exceeds magnetic forces .
+ for a two - phase ambient medium , most of the momentum is acquired during the period prior to shell formation in the low - density intercloud medium ( figure [ fig : tevol_2p ] ) . for given mean density
, there is some variation in when and how much total momentum is gained depending on the locations and sizes of dense clouds around the explosion site ( see figure [ fig : rnorm_2p ] ) , at a level of a few tens of percent . averaging over different realizations ,
however , the momentum at the shell formation time ( when the hot gas mass is maximal ) is within 20% of the value predicted by equation ( [ eq : psf ] ) using the mean density averaged over both dense clouds and low - density intercloud medium ( see `` whole '' rows in table [ tbl:2p ] ) . after shell formation , the momentum increases by at most 50% . for a cloudy medium ,
the final momentum from a single sn varies @xmath293 is also in the range @xmath294 when the @xmath292 .
( see equation ( [ eq : pfinal_2p ] ) ) .
+ our simulations show that the final momentum injected by a single sn is an order of magnitude larger than the initial momentum of the ejecta .
this conclusion is insensitive to the mean density of the ambient medium , and irrespective of strong inhomogeneity in the ism surrounding the explosion site . as a simple quantitative mnemonic
, the momentum injected by a single sn is comparable to the initial energy of the blast ( @xmath129 ) divided by the speed of the expanding shock wave when it becomes radiative ( @xmath295 ) . *
_ implications for sn momentum injection in the ism _
+ for scales larger than an individual star forming region , the amount of momentum injection to the ism by a sn dominates other possible sources of feedback associated with earlier stages of a massive star s evolution , such as stellar winds , radiation pressure , and region expansion ( see e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for analytic estimates of inputs from these sources ) . in recent work ,
a simple momentum injection model ( at a level @xmath296 comparable to that found here ) has been adopted to treat sn feedback , driving turbulence in the ism .
simulations implementing this `` pure momentum '' approach have been used to model molecule - dominated starburst regions @xcite , multiphase atomic - dominated regions @xcite , and barred spiral galaxies @xcite . among the successes of these models
are reaching realistic turbulence levels in the diffuse ism , and achieving states of self - regulated star formation with rates that are consistent with observations ( see also * ? ? ?
* ; * ? ? ?
in addition , @xcite demonstrated that synthetic 21 cm lines reconstructed from the models of @xcite are consistent with the characteristics of real 21 cm line observations by ( * ? ? ? * see also @xcite ) .
+ in many recent investigations studying individual galaxies ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) or galaxy formation in a cosmological context ( e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , it has been found that models with only sn feedback are unable to limit star formation rates to realistic values .
concerns about the inefficacy of sne has prompted increased attention to other modes of feedback , a valuable development in itself ( especially for understanding the evolution of molecular clouds at stages prior to the first sn ) . however , these concerns are to some extent misplaced , and reflect numerical rather than physical limitations .
in particular , it can be difficult to capture the crucial ( but spatially small scale )
st stage of evolution when most of the momentum is imparted in simulations that must follow extremely large spatial scales and can not afford high resolution for the ism . in item 4
below , we summarize the numerical requirements for properly resolving sn momentum injection . + one reason for the current uncertainty about feedback from sn is that the real ism is far from homogeneous , whereas previous focused studies of snr expansion have mostly considered homogeneous ( or at least isotropic ) environments .
for example , @xcite applied only the initial ejecta momentum @xmath297 rather than the momentum at the end of st phase @xmath79 or the final momentum in the momentum conserving phase @xmath175 , citing concerns regarding inhomogeneity of the site of sn explosion .
however , as seen in section [ sec : single_2p ] , inhomogeneity of the background medium does not significantly affect the total momentum injection compared to the uniform - medium case .
the effective spatial scale for momentum deposition is larger in an inhomogeneous medium than in a uniform medium , however , because the shock propagates further in low - density intercloud gas before it becomes radiative .
+ several independent studies contemporaneous with our own have also studied momentum injection by snr in the inhomogeneous ism .
although different groups have considered different environments and physical ingredients in their simulations , all have reached similar conclusions : the final injected momentum in a homogeneous and inhomogeneous medium are comparable to each other , and in good agreement with the results of our simulations .
this result is irrespective of initialization methods to impose inhomogeneity ; we begin with a two - phase cloud / intercloud medium produced by thermal instability , while @xcite adopt a lognormal density distribution , @xcite adopt a fractal density structure , and @xcite initialize based on a turbulent velocity field .
the latter two studies also consider sne within radially stratified clouds . @xcite and @xcite
find that explosion of the sn within a pre - existing low density region can enhance the momentum up to @xmath298 ( including the momentum injected by region expansion ) for solar metallicity . *
_ numerical conditions for correct snr evolution _
+ at the resolution of current galaxy formation simulations ( tens of parsecs or larger ) , and with realistic ism densities , the st phase can not be properly resolved .
this leads to the oft - cited `` overcooling problem , '' namely that the energy of a sn is radiated away without having much impact on the ism , if it is initially deposited in too large an initial volume or mass .
although this problem is well known , there has not previously been a systematic study to determine the numerical requirements needed to avoid it . + here ,
we have conducted a large set of simulations varying both the grid resolution ( @xmath2 ) and the initial size of the snr ( @xmath3 ) .
we find that @xmath215 is necessary to follow the hot gas evolution properly .
we also find that the final momentum is within 18% and 25% of high - resolution benchmarks provided @xmath299 and @xmath300 , respectively .
if the sn explosion occurs within an ambient medium density of @xmath301 or 10 , @xmath302 or @xmath303 thus gives a requirement of @xmath304 or @xmath305 , respectively , for @xmath306 + for practical purposes in an inhomogeneous medium , the resolution requirements can be expressed in terms of a maximum enclosed mass within the initial snr .
the criterion @xmath215 , equivalent to @xmath307 , corresponds to an enclosed mass of @xmath308 for @xmath309 - 0.1 .
interestingly , this is similar to the condition adopted in @xcite , who argued that @xmath310 would yield an initial temperature in the range of @xmath311-@xmath312 , for which the cooling rate would not be excessive .
@xcite and subsequent studies by their colleagues @xcite modelled local galactic disks with large vertical domains , including hot gas created by sn shocks .
the numerical resolution of @xmath313 combined with the condition @xmath310 and dispersed sn events allowed them to correctly resolve the evolution of most snrs .
however , these models did not include self - gravity , which can lead to higher densities at the sites of sn events such that @xmath313 grid resolution would be marginal or insufficient ( considering that @xmath314 for @xmath309 ) ; alternative approaches are discussed below .
* _ alternative feedback prescriptions _
+ if ambient densities are not too high , and the grid is sufficiently fine , the needed resolution requirements @xmath215 ( or @xmath307 ) and @xmath315 can be met , and both the hot gas evolution and momentum injection of a sn can be properly followed by injecting thermal energy and allowing the snr to progress through st and radiative stages .
we recommend this approach for best fidelity in modeling sn feedback .
however , this is not always possible in numerical simulations , even with adaptive mesh refinement techniques .
+ in situations where it is impossible or impractical to resolve snr evolution including the st stage , an alternative is to inject momentum directly to the gas surrounding the feedback site .
this will miss the direct effects of hot gas in the ism ( including potential losses of this gas as a galactic wind ) , but it will still drive turbulence in the warm and cold phases of the ism , which enables star formation to be correctly self - regulated @xcite . to allow for the ( modest ) dependence of momentum injection on ambient conditions , we recommend that equation ( [ eq : pfinal_2p ] ) be used to set the value of the momentum after the mean density in the numerical feedback region is measured . + in a spirit similar to the approach that we recommend , resolution - conditioned approaches to feedback in galaxy formation simulations have been implemented recently by @xcite and @xcite . in both studies ,
sn feedback is assigned either by the ejecta directly to nearby grid zones or sph particles if the local resolution was high , or as `` final '' ( post - st ) momentum otherwise . in their sph simulations , @xcite compared the smoothing length with the shell formation radius @xmath208 ( @xmath316 in their nomenclature ) derived by @xcite to determine which prescription to apply . in their amr simulations ,
@xcite apply either the initial ejecta momentum or a `` final '' ( post - st ) momentum to zones depending on whether the mass in a given feedback sector is greater or less than a snr would have at the `` transition '' time of @xcite .
the resolution requirements we recommend , combined with our calibrations of momentum injection for varying mean density , may be useful in refining and further developing conditional - feedback prescriptions similar to these .
the resolution requirements @xmath215 and @xmath315 that we have obtained through convergence studies could be immediately applied in grid - based simulations that use similar finite - volume methods to those of the _ athena _ code . for sph simulations
, it would be straightforward to conduct an analogous study to determine numerical requirements for convergence in the hot gas mass evolution and net momentum injection . * _ multiple sne and superbubbles / galactic winds _
+ when a sufficiently massive star cluster forms , one can expect not a single but multiple sn explosions to impact the surrounding ism .
many previous studies of multiple sne have focused on what requirements must be met for a superbubble to break out of the disk and drive a wind . in section
[ sec : multi_ti ] , we instead quantify momentum injection to the ism , which is key to settling the turbulence level and rate of star formation . in this paper , we consider only a highly - idealized model of the impact of a moderate - sized star cluster on the surrounding ism , in which sn blasts are intermittent .
we explode 10 sne with time interval of @xmath317 ( m2p models ) in a two - phase medium with density of @xmath155 , 1 , and 10 , following evolution for a total of @xmath318 after the first explosion ( and @xmath319 after the last explosion ) .
these specific conditions limit the sn bubble size smaller than or comparable to the disk scale height , allowing us to explore the momentum injection by a superbubble in a uniform medium .
+ when @xmath280 , as for model m2p - n10 , momentum is injected by individual events in a discrete fashion , whereas in the opposite limit of @xmath320 for model m2p - n0.1 , momentum injection is almost continuous . in spite of these differences , there is only @xmath81 variation in the final momentum measured in the three m2p models ( @xmath321 ) . as for single sne ,
the final momentum per sn event in the m2p models is comparable to the prediction @xmath79 of equation ( [ eq : psf ] ) . for comparison
, we also conduct comparison models with a single sn having @xmath107 ( e2p models ) ; the final momenta for e2p models are a factor @xmath322 larger than for the corresponding m2p models .
+ in recent one - dimensional simulations of sequential sne , @xcite have found that with a sufficiently large number of sne ( i.e. a sufficiently small mean interval @xmath276 ) , the evolution of the sn bubble approaches the limit of continuous energy injection and an adiabatic interior , as in @xcite . for continuous energy injection
, the radius evolves with @xmath323 rather than @xmath324 or @xmath325 .
substantial energy ( about 10 - 30% of total explosion energy ) can remain in this limit ( see also * ? ? ?
* ) , as is required to explain observations of systems such as m82 ( e.g. , * ? ? ?
+ although our m2p models create large , low - density hot volumes that resemble observed superbubbles , here we have not considered a stratified medium or highly - dynamic initial conditions for the ambient warm / cold ism . for a realistic cluster lifetime with @xmath317 , however , the total number of massive stars that could undergo core collapse is closer to 300 , which would drive a superbubble expansion over a much longer period , such that its radius could easily exceed the disk scale height .
more massive clusters would have even shorter mean intervals between sn events , and could easily lead to superbubble breakout . for more comprehensive studies for superbubble evolution , stratification of the disk
should thus be taken into account , in addition to the small - scale inhomogeneity that we have included .
a natural extension of the present work is to include both effects , together with multiple sne , to study the superbubble breakout and creation of galactic winds that may occur when a very massive cluster forms . finally , we note some caveats related to limitations of the models we have presented . because our resolution is insufficient to capture ablation of small clouds by the instabilities ( rayleigh - taylor and kelvin - helmholtz ) arising from shock - cloud interactions ( e.g. , * ? ? ?
* ) , our models may underestimate cooling within the snr .
similarly , as we do not include thermal conduction ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , we may be missing an increase in the mass of hot gas and enhanced cooling due to evaporation of and radiation from clouds .
compared to snr evolution within a uniform intercloud medium , cloud ablation and evaporation lead to earlier shell formation as the effective density of the ambient medium is higher than that of the intercloud medium .
the momentum injection could also be reduced somewhat due to additional radiative losses of energy .
although our simulation can not fully capture cloud ablation and evaporation , these effects have been captured to some extent : in models s2p , the shell formation radius is smaller than that in a purely warm medium ( see table [ tbl:2p ] ) . also , the final injected momentum is smaller in models s2p than su ( see equations ( [ eq : pfinal ] ) and ( [ eq : pfinal_2p ] ) ) , and among 10 different realizations a slightly smaller momentum is obtained when the snr sweeps up more cold clouds . however , it is also important to note that effects of thermal conduction and cloud ablation become more complex for a magnetized medium , as is true for the ism .
thermal conduction is anisotropic , so that the heat flux across the magnetic field is largely suppressed @xcite .
when a shock encounters a cloud , the magnetic field becomes wrapped around the cloud ( piled up in front of the cloud , and trailing downstream ) .
this geometry limits heat flux from the hot medium to the cloud .
in addition , hydrodynamic instabilities are strongly suppressed with near - equipartition magnetic fields @xmath326 ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
calculations of shock - cloud interactions including magnetic fields , anisotropic thermal conduction , and radiative cooling by @xcite have shown that magnetic fields with moderate strength ( @xmath327 ) indeed suppress heat transfer and hydrodynamic instabilities ( see also * ? ? ?
* ) . in future work
, it will be interesting to extend the snr expansion simulations of this paper to include magnetic fields and anisotropic conduction , in order to quantify possible changes in the momentum injection , maximum mass of hot gas , and other basic properties .
in addition , complete models would include cosmic ray pressure , which can potentially boost the late - time snr expansion and momentum .
the magnitude of this effect depends on the diffusivity of the cosmic ray fluid , which is becoming increasing well characterized from recent hybrid kinetic simulations ( e.g. , * ? ? ?
we are grateful to the referee for an extremely thorough report , which helped us to improve the manuscript .
this work was supported by grant no .
ast-1312006 from the national science foundation .
part of this project was conducted during a visit to the kitp at u.c .
santa barbara , which is supported by the national science foundation under grant
nsf phy11 - 25915 .
this research made use of yt visualization and data analysis package @xcite .
natexlab#1#1 , o. , kravtsov , a. v. , leitner , s. n. , & gnedin , n. y. 2013 , , 770 , 25 , m. , white , s. d. m. , naab , t. , & scannapieco , c. 2013 , , 434 , 3142 , s. a. , & mckee , c. f. 1982 , , 252 , 529 , r. , & petruk , o. 2004 , , 419 , 419 , j. m. , wright , e. b. , borkowski , k. j. , & reynolds , s. p. 1998
, , 500 , 342 , c. b. , stinson , g. , gibson , b. k. , wadsley , j. , & quinn , t. 2012 , , 424 , 1275 , d. , & spitkovsky , a. 2014 , , 794 , 47 , r. a. 1974 , , 188 , 501 , r. a. , & clegg , a. w. 1985 , , 317 , 44 , r. a. , & gardner , j. 1974 , , 192 , 457 , d. f. , mckee , c. f. , & bertschinger , e. 1988 , , 334 , 252 , l. l. , & mckee , c. f. 1977 , , 211 , 135 , l. l. , mckee , c. f. , & ostriker , j. p. 1981
, , 247 , 908 , d. p. 1972 , , 178 , 159 . 1986 , , 304 , 771 . 2005 , , 43 , 337 , d. p. , & anderson , p. r. 1982 , , 253 , 268 , d. p. , & smith , b. w. 1974 , , 189 , l105 , m. a. , & breitschwerdt , d. 2004 , , 425 , 899 , c. l. , krumholz , m. r. , ballesteros - paredes , j. , et al . 2013 , arxiv e - prints , arxiv:1312.3223 , b. t. 2011 , physics of the interstellar and intergalactic medium , k. 1998 , , 503 , 700 , k. m. 2001 , reviews of modern physics , 73 , 1031 , k. m. , mac low , m .- m . , & zweibel , e. g. 1991 , , 375 , 239 , g. b. 1965 , , 142 , 531 , g. b. , goldsmith , d. w. , & habing , h. j. 1969 , , 155 , l149 , p. c. , anninos , p. , gustafson , k. , & murray , s. d. 2005 , , 619 , 327 , s. , rosdahl , j. , blaizot , j. , devriendt , j. , & slyz , a. 2014 , arxiv e - prints , arxiv:1412.0484 , j. , callegari , s. , madau , p. , & mayer , l. 2011 , , 742 , 76 , h. , & tomisaka , k. 2006 , , 641 , 905 , c. , & troland , t. h. 2003 , , 586 , 1067 , p. , & iffrig , o. 2014 , , 570 , a81 , a. s. , joung , m. r. , mac low , m .-
, et al . 2012 , , 750 , 104 , p. f. , kere , d. , oorbe , j. , et al . 2014 , , 445 , 581 , p. f. , quataert , e. , & murray , n. 2012 , , 421 , 3522 . 2012 , , 421 , 3488 , o. , & hennebelle , p. 2014
, arxiv e - prints , arxiv:1410.7972 , e. p. g. , & ziegler , u. 2013 , , 766 , 45 , m. k. r. , & mac low , m .-
2006 , , 653 , 1266 , m. r. , mac low , m .-
, & bryan , g. l. 2009 , , 704 , 137 , n. 1992 , , 391 , 502 , b. w. , wadsley , j. , benincasa , s. m. , & couchman , h. m. p. 2014
, , 442 , 3013 , c .- g . , kim , w .- t . , & ostriker , e. c. 2008 , , 681 , 1148 . 2011 , , 743 , 25 , c .- g . , ostriker , e. c. , & kim , w .- t .
2013 , , 776 , 1 . 2014 , , 786 , 64 , t. , & cen , r. 2014 , , 788 , 121 , r. i. , mckee , c. f. , & colella , p. 1994
, , 420 , 213 , v. , kiss , c. , mor , a. , kiss , z. t. , & tth , l. v. 2007 , , 463 , 1227 , b .- c .
, & mckee , c. f. 1992 , , 388 , 93 . 1992 , , 388 , 103 , h. , & inutsuka , s .- i .
2002 , , 564 , l97 , m. r. , bate , m. r. , arce , h. g. , et al .
2014 , arxiv e - prints , arxiv:1401.2473 , m. n. , & stone , j. m. 2009 , , 691 , 1092 , m .- m . , & klessen , r. s. 2004 , reviews of modern physics , 76 , 125 , m .- m . ,
mccray , r. , & norman , m. l. 1989 , , 337 , 141 , m .- m . ,
mckee , c. f. , klein , r. i. , stone , j. m. , & norman , m. l. 1994 , , 433 , 757 , d. , faucher - gigure , c .- a . , & quataert , e. 2014 , arxiv e - prints , arxiv:1409.4425 , c. d. 2002 , , 566 , 302 , r. , & kafatos , m. 1987 , , 317 , 190 , c. f. , & ostriker , e. c. 2007 , , 45 , 565 , c. f. , & ostriker , j. p. 1977
, , 218 , 148 , n. , quataert , e. , & thompson , t. a. 2010 , , 709 , 191 , f. , mckee , c. f. , klein , r. i. , & fisher , r. t. 2006 , , 164 , 477 , c. a. , & ikeuchi , s. 1989 , , 345 , 372 , s. , bocchino , f. , reale , f. , peres , g. , & pagano , p. 2008
, , 678 , 274 , e. c. , mckee , c. f. , & leroy , a. k. 2010 , , 721 , 975 , e. c. , & shetty , r. 2011 , , 731 , 41 , j. p. , & mckee , c. f. 1988 , reviews of modern physics , 60 , 1 , r. a. , & ostriker , e. c. 2004 , , 601 , 905 , a. , & bregman , j. n. 1995 , , 440 , 634 , a. , nath , b. b. , sharma , p. , & shchekinov , y. 2013 , , 434 , 3572 , n. , kanekar , n. , braun , r. , & chengalur , j. n. 2013 , , 436 , 2352 , r. , morano , e. , & druguet , m .- c .
1998 , journal of computational physics , 145 , 511 , c. , wadepuhl , m. , parry , o. h. , et al .
2012 , , 423 , 1726 , l. i. 1959 , similarity and dimensional methods in mechanics , w .- y .
, & kim , w .- t .
2013 , , 769 , 100 . 2014 , , 792 , 47 , p. , roy , a. , nath , b. b. , & shchekinov , y. 2014 , , 443 , 3463 , r. , & ostriker , e. c. 2012 , , 754 , 2 , m .- s . , stone , j. m. , & snyder , g. f. 2008 , , 680 , 336 , j. d. , & cox , d. p. 1992
, , 392 , 131 , l. 1962 , physics of fully ionized gases . 1978 , physical processes in the interstellar medium , v. , & hernquist , l. 2003 , , 339 , 289 , s. , murray , c. e. , lee , m .- y . , heiles , c. , & miller , j. 2014 , , 793 , 132 , a. , le petit , f. , roueff , e. , & le bourlot , j. 2014 , , 790 , 10 , g. , seth , a. , katz , n. , et al . 2006 , , 373 , 1074 , g. s. , brook , c. , macci , a. v. , et al .
2013 , , 428 , 129 , j. m. , & gardiner , t. 2009 , , 14 , 139 , j. m. , gardiner , t. a. , teuben , p. , hawley , j. f. , & simon , j. b. 2008 , , 178 , 137 , d. k. , & heckman , t. m. 2009 , , 697 , 2030 , d. k. , & stevens , i. r. 2000 , , 314 , 511 , r. s. , & dopita , m. a. 1993 , , 88 , 253 , g. 1950 , royal society of london proceedings series a , 201 , 159 , g. , rozyczka , m. , & bodenheimer , p. 1990
, , 237 , 207 , r. , pontzen , a. , dubois , y. , & read , j. i. 2013 , , 429 , 3068 , r. j. , & couchman , h. m. p. 2000
, , 545 , 728 , k. , gaudlitz , m. , janka , h .- t . , & steinmetz , m. 1998 , , 500 , 95 , k. , habe , a. , & ikeuchi , s. 1981 , , 78 , 273 , k. , & ikeuchi , s. 1986 , , 38 , 697 , m. j. , smith , b. d. , oishi , j. s. , et al .
2011 , , 192 , 9 , e. o. , nath , b. b. , & shchekinov , y. 2015 , , 446 , 1703 , e. t. 1983 , , 274 , 152 . 1994 , , 428 , 186 , s. k. , & naab , t. 2014 , arxiv e - prints , arxiv:1410.0011 , r. , mccray , r. , castor , j. , shapiro , p. , & moore , r. 1977 , , 218 , 377 , m. t. , & durisen , r. h. 1987 , , 224 , 701 , m. g. , hollenbach , d. , mckee , c. f. , tielens , a. g. g. m. , & bakes , e. l. o. 1995 , , 443 , 152 , m. g. , mckee , c. f. , hollenbach , d. , & tielens , a. g. g. m. 1995 , , 453 , 673 . 2003 , , 587 , 278 , l. 1972 , , 10 , 129
the effect of magnetic fields in snr evolution was initially investigated using spherically symmetric one - dimensional simulations , in which magnetic pressure forces are included but the full mhd equations are not solved @xcite ( this can approximately represent an equatorial band tangent to a uniform external magnetic field ) .
the strong magnetic pressure within the compressed shell results in a thicker shell and recompresses hot gas in the later times .
@xcite followed the long - term evolution of a snr in a magnetized medium using axisymmetric two - dimensional simulations , and showed that the magnetic tension force also helps broaden the shell and compress the hot gas ( see also * ? ? ?
* for supperbubble expansion ) .
magnetic fields play an important role in determining the volume of the hot gas in very late stages of a snr , and therefore affects the porosity of the ism . here , we are interested in assessing potential effects of magnetic fields on momentum injection by snrs in the ism .
we have re - run the set of su models , but now include a uniform magnetic field in the @xmath328-direction @xmath329 .
the magnetic field strength is set by using the plasma beta @xmath330 where the thermal pressure of the ism in our models varies with the mean density , @xmath331 , as @xmath332 . for reference , the magnetic field strength is @xmath333 figure [ fig : tevol_mhd ] plots time evolution of physical properties of the snr in models su - n1 with @xmath289 , 1 , 10 , and @xmath334 .
all models follow almost the same evolution except at late time in the model with @xmath289 .
it is obvious that the magnetic energy in the shell becomes comparable to the thermal energy of the shell only for @xmath289 after shell formation ( see figure [ fig : tevol_mhd](c ) ) .
as reported in earlier work , a strong magnetic field within the shell recompresses the interior hot gas ( smaller @xmath335 at late time for @xmath289 ) , leading to larger @xmath137 and @xmath70 at late times ( see figure [ fig : tevol_mhd](d ) and ( f ) )
. however , the radial momentum is almost the same for all cases .
this is because the majority of the momentum is injected to the ism before the magnetic energy in the shell is very large . even for the @xmath289 case ,
full momentum is acquired by ( 2 - 3)@xmath49 , when the magnetic energy just begins to dominate the thermal energy .
figure [ fig : snap_mhd ] displays the structure of the model with @xmath301 and @xmath289 in the yz plane . early evolution is the same as for the unmagnetized model ( left three columns , see figure [ fig : snap ] ) , while non - spherical structure emerges in the latest stage ( rightmost column ) .
the outer shock can propagate farther in the direction perpendicular to the magnetic fields ( fast mode ) , while shell broadening results in compression of the interior hot gas .
we have also performed simulations analogous to models su - n0.1 , su - n10 , and su - n100 with @xmath289 , 1 , and 10 .
we find that the normalized evolution for all cases is `` congruent '' ( as in unmagnetized models ) up to 2 or 3 @xmath49 , so that the final momentum remains unchanged .
this suggests that ism magnetic fields with reasonable strengths of @xmath336 may be able to alter late time evolution and porosity of the hot gas , but would not significantly affect the injected momentum ( see also * ? ? ? | supernova ( sn ) explosions deposit prodigious energy and momentum in their environments , with the former regulating multiphase thermal structure and the latter regulating turbulence and star formation rates in the interstellar medium ( ism ) .
however , systematic studies quantifying the impact of sne in realistic inhomogeneous ism conditions have been lacking .
using three - dimensional hydrodynamic simulations , we investigate the dependence of radial momentum injection on both physical conditions ( considering a range of mean density @xmath0 ) and numerical parameters .
our inhomogeneous simulations adopt two - phase background states that result from thermal instability in atomic gas .
although the snr morphology becomes highly complex for inhomogeneous backgrounds , the radial momentum injection is remarkably insensitive to environmental details . for our two - phase simulations , the final momentum produced by a single sn
is given by @xmath1 .
this is only 5% less than the momentum injection for a homogeneous environment with the same mean density , and only 30% greater than the momentum at the time of shell formation .
the maximum mass in hot gas is also quite insensitive to environmental inhomogeneity .
strong magnetic fields alter the hot gas mass at very late times , but the momentum injection remains the same .
initial experiments with multiple spatially - correlated sne show a momentum per event nearly as large as single - sn cases .
we also present a full numerical parameter study to assess convergence requirements . for convergence in the momentum and other quantities
, we find that the numerical resolution @xmath2 and the initial size of the snr @xmath3 must satisfy @xmath4 , where the shell formation radius is given by @xmath5 for two - phase models ( or 30% smaller for a homogeneous medium ) . |
in systematic investigations of the e1 transitions @xcite the low lying 1@xmath3 have been observed in spherical nuclei states , which are characterized by strong b(e1;@xmath4 ) , of the order of several units@xmath5 .
it was demonstrated that these low lying 1@xmath3 states arise as @xmath6 due to coupling of the collective quadrupole 2@xmath7 and the collective octupole 3@xmath8 states .
similar 1@xmath8 states have been observed in the cd , sn , ba , ce , nd , and sm isotopes .
their two phonon character has been proved by the observed strong e2 and e3 transitions to the corresponding one phonon states @xcite and by the fact that the energies of these 1@xmath8 states are very close to the summed energies @xmath9 .
strong correlations between the values of b(e1;@xmath10 ) and the product of average squares of the quadrupole @xmath11 and octupole @xmath12 deformation parameters in nonmagic nuclei @xcite clearly show that away from the closed shells the large b(e1 ) values have a collective nature connected with the motion of the nuclear shape .
the ratio @xmath13/@xmath14 is amazingly constant , although the b(e1 ) strength varies by an order of magnitude in the considered nuclei . accepting a two phonon picture of the 1@xmath8 states
let us consider the other important experimental facts characterizing strong e1 transitions between low lying states .
+ analyzing the experimental data on the ratio b(e1;@xmath15/b(e1;@xmath16 it was shown in @xcite that this ratio is a constant equal to 1 within a factor of 2 although these e1 transition strengths can differ by about two orders of magnitude for different nuclei .
semimagic even@xmath17 nuclei and even even nuclei with two or four nucleons outside a closed shell were considered in this analysis .
these nuclei are rather vibrational and can be understood , at least qualitatively , in a harmonic phonon picture . in the last model b(e1;@xmath15/b(e1;@xmath18 .
this correlation of the e1 strengths can be considered as an additional support for the quadrupole octupole coupled character of the 1@xmath8 states . +
the b(e1;@xmath19 is small in sn isotopes ; however , it increase in cd , te isotopes and approach the alaga value for the ratio b(e1;@xmath19/b(e1;@xmath15 in deformed nuclei @xcite .
+ a minimum has been found in the @xmath17dependence of b(e1;@xmath15 in the nd , sm and ba isotopes when the number of neutrons @xmath20 is equal to 78 or 86 @xcite .
this characteristic feature of the behavior of b(e1;@xmath15 as a function of @xmath20 has been discovered earlier in the rpa based calculations in @xcite .
+ strong e1 transitions between the @xmath21 octupole phonon states and the 2@xmath22 mixed symmetry quadrupole one phonon states in the spherical nuclei @xmath23zr@xmath24 , @xmath25mo@xmath24 , @xmath26ru@xmath24 and @xmath27ce@xmath28 are observed @xcite . in these four nuclei
the e1 transitions to the 2@xmath22 state are stronger than to the 2@xmath29 state .
probably , this is a consequence of the isovector nature of the e1 transition operator , which enhances e1 transitions between the mixed symmetry states and the isoscalar ones . in the last case
, this is the isoscalar octupole vibrational state .
+ since b(e1;@xmath30 are strong the question arises if there should appear strong e1 transitions from the ground state to the @xmath31 state , which is expected at the excitation energy 4187 kev in @xmath32zr , 4601 kev in @xmath33mo , 4934 kev in @xmath34ru and 3658 kev in @xmath35ce .
however , anharmonic effects , which can shift the energies of these two phonon states , should not be excluded . summarizing the short review of the data we can see that the microscopical model describing strong e1 transitions at low energies basing on the two phonon quadrupole
octupole model of the low lying 1@xmath8 states should explain : + the fact that b(e1;@xmath15/b(e1;@xmath36 at least in near magic nuclei ; + an increase of b(e1;@xmath19 when going away from the closed shell ; + the existence of the minimum in the @xmath17 dependence of b(e1;@xmath37 when moving off the semimagic nuclei ; + the fact that b(e1;@xmath38 b(e1;@xmath16 .
+ proportionality of b(e1;@xmath37 to @xmath39 in collective nuclei ; +
in @xcite the @xmath1 phonon approach to the description of the positive parity collective states was developed within the interacting boson model .
it was shown that the wave vectors of the yrast , the second @xmath40 and the second @xmath41 states could be described to a high accuracy over the whole parameter space of the consistent @xmath1 hamiltonian , i.e. , far outside of the region where a picture of harmonic vibrations was correct , by simple universal expressions contaning only one or two multiple @xmath1phonon configurations .
a simple structure of the wave vectors helps derive different relations between transition matrix elements .
in contrast to @xcite , where the @xmath1phonon approach was formulated for the bosonic configurational space of the interacting boson model , in the present paper this approach is formulated for the fermionic configurational space .
in addition , we consider both positive and negative parity states . in the @xmath1phonon approach the @xmath42 state is presented by the following expression @xmath43 is the ground state vector ,
@xmath44 + = @xmath45/ @xmath46 and @xmath47 is in our case the standard shell model quadrupole moment operator @xmath48 expressed in terms of the nucleon creation @xmath49 and annihilation @xmath50 operators .
let us include into consideration the octupole mode @xmath51 = @xmath52/ @xmath53 and @xmath54 is the fermionic octupole moment operator @xmath55 as we know from the rpa type calculations @xcite both expressions ( [ 1-eq ] ) and ( [ 3-eq ] ) are very good approximations for the lowest collective @xmath40 and @xmath56 states .
in fact , this is a consequence of the well known from the experiment result that e2 and e3 transitions from the ground state to the @xmath42 and @xmath21 states , correspondingly , are much stronger than transitions to higher lying @xmath40 and @xmath56 states . continuing along this line and remembering that the 1@xmath8 state is mainly a quadrupole octupole two phonon state we suggest for the 1@xmath8 state vector that @xmath57 we obtain the following relation : @xmath58 the last three terms in the right hand side of ( [ 6-eq ] ) can give a noticeable contribution only in the case of strong mixing of the two phonon quadrupole
quadrupole and two phonon octupole
octupole states .
however , as it is known from the rpa type calculations @xcite , this mixing is insignificant : less than 1@xmath59 . neglecting these terms we obtain an approximate relation @xmath60 the wave vector ( [ 5-eq ] ) can be written in terms of the rpa collective phonons and the two
quasiparticle components which correspond to the noncollective rpa solutions . written in this way
the wave vector ( [ 5-eq ] ) has the two phonon component with one quadrupole and one octupole collective rpa phonons as the main component , the three
phonon component with two collective quadrupole and one collective octupole rpa phonons , and the two
quasiparticle @xmath61 components .
it means that the results of our @xmath1 phonon approach should be compaired not with the pure rpa calculations but with the rpa based calculations which include also anharmonic effects .
a contribution of the last two components to the norm of the wave vector ( [ 5-eq ] ) is very small for nuclei considered in the present paper .
of course , when we approach a deformed region the following @xmath1 phonon component @xmath62 becomes more and more important since in this region there are not one but two @xmath61 states with @xmath63=0 and @xmath63=1 . to construct the @xmath61 state with @xmath63 as a good quantum number ,
it is necessary to take a mixture of the states ( [ 5-eq ] ) and ( [ 8-eq ] ) .
this mixing , which can be estimated as 70@xmath59 of one component plus 30@xmath59 of the other one can be considered as an upper limit for the error connected with the approximation ( [ 5-eq ] ) .
of course , near spherical nuclei this mixing is much smaller . in transitional nuclei with soft @xmath64 mode the @xmath1 phonon state @xmath65 can also become important .
the strength of the @xmath1 phonon scheme lies in a possibility to derive relations between electromagnetic transition matrix elements also outside the analytically solvable harmonic vibrator and rotor limits .
this is possible because of the simple form of the wave vectors in this approach .
this possibility is realized in this section below and in sections iii and iv .
however , the wave vectors ( [ 1-eq ] ) , ( [ 3-eq ] ) and ( [ 5-eq ] ) only look simple .
the multipole moment operators in these expressions act on the exact ground state .
thus , for calculations we needed an expression for the ground state vector . in section
v we take as an approximation to the ground state vector the vacuum state of the rpa phonons .
let us apply the model formulated above to calculations of the matrix element of the e1 @xmath2 transition .
we obtain @xmath66 where @xmath67 is the operator of the electric dipole transition .
since multipole operators @xmath68 , @xmath69 and @xmath70 commute with each other we can rewrite the last expression as @xmath71 using expressions ( [ 1-eq ] ) , ( [ 3-eq ] ) and relation ( [ 7-eq ] ) we can rewrite ( [ 11-eq ] ) in the following way : @xmath72 > from the last expression using the wigner eckart theorem we obtain @xmath73 and from ( [ 13-eq ] ) we get @xmath74 the last value coincides with that of the harmonic phonon picture
. however , we did not use the assumption of harmonicity of quadrupole and octupole vibrations .
thus , the @xmath1phonon model explains qualitatively one of the experimental results mentioned in introduction .
however , the value of the ratio 7/3 deviates from the experimental observation 1@xmath752 . a possible reason for this deviation can be the presence of some admixtures in the @xmath76 , @xmath77 and @xmath78 states which are not taken into account by the @xmath1phonon approach .
the corrections to relation ( [ 7-eq ] ) discussed above will also influence the number in the right hand side of ( [ 14-eq ] ) .
mention , however , a very small value of @xmath79 obtain in @xcite . taken together with the known value of @xmath80 it produces the value of the ratio ( [ 14-eq ] ) which is much larger than 7/3 .
let us consider the matrix element @xmath82 .
this matrix element can be written in the following way : @xmath83 having in mind that the @xmath2 transition is the strongest one among the low lying states we can rewrite approximately ( [ 15-eq ] ) as @xmath84 and as a consequence of ( [ 16-eq ] ) we obtain @xmath85 the quadrupole moment of the @xmath0 state can be expressed approximately in terms of the quadrupole moments of the @xmath42 and @xmath21 states .
finally , we have @xmath86 > from ( [ 18-eq ] ) we can see that going from the magic or semimagic nuclei to nuclei with open shell the ratio b(e1;@xmath81)/b(e1;@xmath87 ) increases with the quadrupole moments of the first @xmath40 and @xmath56 states . in order to obtain the alaga rule for the ratio of the matrix elements @xmath88/@xmath89 , we must take into account an admixture of @xmath90 to the @xmath76 state .
for completeness of the consideration the results of the very schematic ibm based analysis of the e1 transitions between the low lying collective states are given in this section below .
a detailed analysis is presented in @xcite .
let us analyse qualitatively the quantity b(e1;@xmath2 ) in order to understand a possible reason for the appearance of the minimum in the neutron number dependence of this quantity when going away from the closed neutron shell @xmath20=82 . as it follows from relation ( [ 13-eq ] ) , instead of @xmath91 we can consider the matrix element @xmath92 . in the ibm @xmath77 and @xmath78 are the states with the maximum value of the @xmath93spin : @xmath94 .
the dipole transition operator @xmath70 is mainly the @xmath93spin vector . to simplify the discussion
, we assume below that the e1 transition operator is exactly the @xmath93spin vector .
then @xmath95 where @xmath96 and @xmath97 are the numbers of the valence protons and neutrons , respectively .
if we consider a nucleus with the closed neutron , shell , then as it is seen from ( [ 19-eq ] ) , only protons will contribute to the matrix element @xmath92 .
if we start to increase the number of valence neutrons , then a neutron contribution to @xmath92 will increase compensating , partly , a proton contribution .
this matrix element takes the minimum value when @xmath98 .
of course , this phenomenological analysis can not give us the value of @xmath99 at which @xmath92 has a minimum because the physical e1 transition operator does not reduce to the @xmath93 spin vector .
in the ibm based consideration it is assumed , in fact , that a microscopic structure of the proton and the neutron bosons does not change from nucleus to nucleus .
however , it is not evident that this assumption will be supported by the microscopic calculations .
the results of the rpa based microscopic analysis are presented below . in a similar way , transitions between the @xmath21 and @xmath100 states @xcite can be considered .
in contrast to the preceding case the @xmath101 state is a state with @xmath102 . therefore , @xmath103 where @xmath104 and @xmath105 . therefore , @xmath106 it is seen from ( [ 20-eq ] ) and ( [ 21-eq ] ) that in contrast to ( [ 19-eq ] ) there is no cancellation of the proton and neutron contributions to the matrix element .
thus , this analysis indicates a possible reason for a large value of the @xmath107 matrix element @xcite .
let us assume , by analogy with section ii , that the @xmath108 state is created by an action of the operator @xmath109 on the ground state . here
@xmath110 is the neutron ( proton ) quadrupole operator .
this state is constructed so as to be orthogonal to the @xmath42 state . then using as above a commutativity of the operators ( [ mixed - eq ] ) and @xmath70 we obtain from the inequality + @xmath111 that @xmath112 .
the last statement seems to be in disagreement with the analytical result presented in @xcite ( table 3 ) . a possible reason is connected with a different proton neutron structure of the shell model e1 transition operator and the two body term in the @xmath113 ibm2 e1 transition operator .
roughly speaking , the proton and neutron parts of the shell model e1 transition operator has opposite signs . only a small part of this operator proportinal to @xmath114 , which appears because of substraction of the center of mass motion , contains proton and neutron contributions with the same signs .
however , the two body term in the @xmath113 ibm2 e1 transition operator used in @xcite contains the proton and neutron parts with the same sign . as a consequence , the value of b(e1;@xmath115 ) , to which only a two body term contributes , becomes small and if @xmath116=1 , i.e. , a two body term becomes the @xmath93spin vector , b(e1;@xmath115)=0 . in this paper
, we did not calculate the matrix element @xmath117 .
however , if we take expression ( [ mixed - eq ] ) as an approximation for the operator , which produces the @xmath100 state acting on the ground state , then the proton and the neutron contributions to this matrix element will have the same sign and for this reason @xmath118 will be larger than @xmath119 .
the aim of this section is to calculate the dipole transitional matrix element @xmath120 , basing on expressions ( [ 1-eq ] ) , ( [ 3-eq ] ) and ( [ 5-eq ] ) for the state vectors @xmath77 , @xmath78 and @xmath76 and on the rpa approximation for the ground state .
thus , in the calculations below it is assumed that the wave vectors of the @xmath42 , @xmath21 and @xmath0 states are @xmath121 ) and ( [ 18-eq ] ) is independent of a concrete structure of the @xmath122 state and is based only on the @xmath1phonon form of the wave vectors , the results obtained below are consistent with relations ( [ 14-eq ] ) and ( [ 18-eq ] ) .
we assume below that the ground state contains only those correlations which are produced by the quadrupole
quadrupole and the octupole octupole interactions .
the quadrupole and octupole interaction constants are fixed so as to reproduce the experimental values of the b(e2;@xmath123 and the b(e3;@xmath124 , respectively .
the ground state correlations related to the dipole
dipole interaction are not included .
therefore , calculating the strength of the e1 transitions we must introduce the core polarization factor @xmath125 which takes into account a shift of a part of the e1 strength to the giant dipole resonance @xcite .
mention that the rpa type calculations of the e1 transitions were performed in @xcite . in our approach based on the @xmath1phonon description
the following expressions describe a microscopic structure of the @xmath77 and @xmath78 state vectors @xmath126 @xmath127 @xmath128 in the above expressions summation is performed over both proton and neutron single particle states . in ( [ 25-eq ] ) and ( [ 26-eq ] )
@xmath129 and @xmath130 are the creation operators of the most collective quadrupole and octupole rpa phonons corresponding to the first roots of the rpa secular equation ; @xmath131 is the quasiparticle creation operator ; @xmath132 and @xmath133 are the rpa amplitudes describing a microscopic structure of the most collective phonons @xmath129 and @xmath129 ; @xmath134 are the coefficients of the bogoliubov transformation ; @xmath135 is the normalization factor . in ( [ 25-eq ] ) and ( [ 26-eq ] ) dipole excitations are presented by the two quasiparticle and not by the phonon operators because dipole
dipole correlations in the ground state are not taken into account .
of course , summation is performed over all @xmath136 two quasiparticle states .
contributions of the second terms in ( [ 25-eq ] ) and ( [ 26-eq ] ) to the total norm are of the order of 1@xmath59 or less .
the wave vector of the @xmath0 state contains the following components : @xmath137,@xmath138 @xmath139 and @xmath140 . among them
the first one gives the main contribution to the norm , which is close to 100@xmath59 .
there are also other components in the wave vectors of the @xmath42 , @xmath21 and @xmath0 states .
however , they are small and unimportant for calculations of the e1 transitions . a similar conclusion about a microscopic structure of the @xmath0 state was obtained in @xcite .
the components of the wave vectors of the @xmath42 , @xmath21 and @xmath0 states , containing operators creating two
quasiparticle @xmath61 states , appear because the quadrupole and the octupole multipole operators representing the corresponding state vectors according to ( [ 1-eq ] ) , ( [ 3-eq ] ) and ( [ 5-eq ] ) , are not exhausted by the one boson term expressed in collective bosons only . in the calculations based on the quasiparticle
phonon model @xcite this admixture is generated by the quasiparticle
phonon coupling term which is produced by the terms in @xmath68 and @xmath69 additional to the one boson term . however , the weight of this contribution in the norm of the eigenstate is similar in both the approaches . for the reduced matrix element @xmath89 we obtain the following expression : @xmath141 where @xmath142 @xmath143 @xmath144 @xmath145 @xmath146 in ( [ 33-eq ] ) , ( [ 34-eq ] ) and ( [ 35-eq ] ) @xmath147 or @xmath148 .
the quantities @xmath149 and the matrices @xmath150 are determined by the expressions analogous to ( [ 29-eq ] ) and ( [ 27-eq ] [ 28-eq ] ) , respectively , though with summation over proton or neutron single particle states only .
the results of the calculations of the electric dipole transitional matrix element are presented in figures 1 and 2 and in tables 19 . here besides the total calculated dipole transitional matrix element @xmath151 including all contributions the results obtained without inclusion of the contribution coming from the @xmath136 two quasiparticle admixture denoted by @xmath152 are also shown .
the results presented in tables 1 and 2 and in figs . 1 and 2 show that in cd , sn , ba , ce and partly in nd and sm isotopes the experimental data are between the results of calculations obtained with and without a contribution of an admixture of a dipole two quasiparticle component to the @xmath0 state .
however , in many cases the matrix element @xmath153 , which includes only a collective contribution , is closer to the experimental data than the total matrix element , which includes both collective and two quasiparticle contributions .
nevertheless , we can see that the two quasiparticle admixture to the collective quadrupole octupole two phonon component of the @xmath0 state should be taken into account to improve agreement with the experimental data . in the present calculations ,
this contribution of the two quasiparticle component is overestimated . however , since the weight of this component is smaller than 1@xmath59 it is difficult to expect such an accuracy from the microscopic calculations .
we can conclude that the two phonon component @xmath154 , which gives the main contribution to the norm of the @xmath76 state , determines an excitation energy and the e2 and e3 decay properties of the @xmath76 state .
however , for description of the electric dipole transitions @xmath2 it is necessary to take into account also an admixture of the dipole two quasiparticle component . as it is seen from expressions ( [ 25-eq ] ) and ( [ 26-eq ] ) , the components containing two quasiparticles coupled to angular momentum @xmath155=1 contribute also to the structure of the @xmath42 and @xmath21 states .
since the microscopic e1 transitional operator contains in addition to the quadrupole
octupole phonon term a term changing the number of quasiparticles by two units , these components contribute to the @xmath156 transition matrix element .
both sets of the results obtained with and without inclusion of the contribution of the two quasiparticle component of the @xmath0 state reproduce also an experimental @xmath17dependence of the reduced matrix element . only in @xmath157nd and @xmath158sm a depth of the minimum in the @xmath17dependence of @xmath159 is significantly larger than in the calculated results . as it was noted in introduction this matrix element has a minimum when the number of neutrons is equal to the magic number plus(minus ) four neutrons . in tables
38 the calculated ratios of the matrix elements @xmath160/@xmath161 are compared with the experimental ones .
it is seen that in the case of cd , sn , ba and ce isotopes our calculations reproduce an @xmath17dependence of the reduced dipole matrix element .
however , in nd and sm isotopes experimental data have a deeper minimum than the calculated values .
calculations rather show a delay in an increase of the absolute value of the matrix element with increasing a number of the valence particles or holes
. in the case of the nd and sm isotopes the results of the microscopic calculations deviate from the picture which follows from the analysis of an @xmath17dependence of @xmath162 in the interacting boson model , where it is clearly seen that the reduced matrix element @xmath162 decreases if the number of valence protons is fixed , but the number of valence neutrons increases from zero . in the microscopic calculations , a proton contribution to the e1 transition matrix element changes when the number of valence neutrons varies
although the number of valence protons continue to be fixed .
this fact is illustrated by the results presented in table 9 where the quantities @xmath163 and @xmath164 determined by eqs .
( [ 31-eq][32-eq ] ) are shown for the nd isotopes .
the nature of the electric dipole two quasiparticle component admixed to the wave function of the @xmath0 state is an interesting question .
since this component is a mixture of many spherical two quasiparticle states , we can consider it formally as the @xmath165boson introduced in @xcite .
if we separate this component from the wave function of the @xmath0 state , change its normalization coefficient so as to obtain a wave function with the norm equal to one and calculate the e1 reduced matrix element between this artificially constructed state and the ground state , we obtain the value of the order of 0.60.7 @xmath166fm . for comparison , for a pure alpha cluster state we obtain approximately 6 @xmath166fm . for the 1p1h ( 1h@xmath167 1g@xmath168)@xmath169 proton state this matrix element is equal to 1.7 @xmath166fm .
so , it is difficult to do a definite conclusion about the nature of this two quasiparticle admixture . concluding this section we should like to stress that any microscopic consideration of the enhanced e1 transitions can not be performed in a harmonic approximation , i.e. ,
can not be carried out in the framework of the pure rpa and requires from the beginning an inclusion of the anharmonic effects .
indeed , in the pure rpa calculations only those parts of the one body fermion operators are taken into account which change the number of quasiparticles by two units .
the terms of the form @xmath170 are neglected .
the only exception is the part of the nuclear hamiltonian which describes noninteracting quasiparticles because of its leading role .
however , just the term of the form @xmath170 describes @xmath171 e1 transition .
therefore , the terms of this kind in the @xmath68 and the @xmath69 operators should be included into consideration .
however , this inclusion leads to appearance of the anharmonic terms in the hamiltonian since the latter contains quadrupole quadrupole and octupole octupole interactions .
an inclusion into the operators @xmath68 , @xmath69 and @xmath70 of the terms proportional to @xmath172 is necessary to conserve commutativity of these operators .
this conclusive remark shows that a comparison of the present microscopic consideration , which from the beginning contains anharmonic effects , with the u(5 ) limit of ibm is difficult .
in conclusion , based on the @xmath1phonon representation of the wave vectors of the low lying collective states we have derived the relations between different reduced matrix elements of the e1 transition operator .
these relations explain qualitatively the experimentally observed correlations among data on e1 transitions . using the rpa approximation for the ground state vector and the @xmath1 phonon form of the @xmath0 state we performed microscopic calculations of the reduced matrix element of the e1 transition operator between @xmath173 and @xmath76 states .
it is shown that the two quasiparticle component of the wave vector of the @xmath76 state should be taken into account to achieve an agreement with the experimental data in spite of a small contribution of this component to the norm of the @xmath76 state .
the authors are grateful to profs .
p. von brentano , j. jolie , a. richter , w. scheid and a. zilges and dr.v .
ponomarev for very useful discussions .
this work was supported in part by rfbr ( moscow ) , grant 040217376 .
99 t. guhr , k.d .
hummel , g. kilgus , d. bohle , a. richter , c.w .
de jager , h. de vries , and p.k.a .
de witt huberts , nucl.phys . *
a501 * , 95 ( 1989 ) .
a. zilges , p. von brentano , h. friedrichs , r.d .
heil , u. kneissl , s. lindenstruth , h.h.pitz , and c. wesselborg , z.phys . * a340 * , 155 ( 1991 ) .
u. kneissl , h.h.pitz , and a. zilges , prog.part.nucl.phys .
* 37 * , 349 ( 1996 ) . c. fransen , o. beck , p. von brentano , t. eckert , r .- d .
herzberg , u. kneissl , h. maser , a.nord , n. pietralla , h.h.pitz , and a. zilges , phys.rev.c * 57 * , 129 ( 1998 ) .
robinson , j. jolie , h.g .
brner , p. schillebeecks , s. ulbig , and k.p .
lieb , phys.rev.lett . * 73 * , 412 ( 1994 ) .
m. wilhelm , e. radermacher , a. zilges , and p. von brentano , phys.rev.c * 54 * , r449 ( 1996 ) .
m. wilhelm , s. kasemann , g. pascovici , e. radermacher , p. von brentano , and a. zilges , phys.rev.c * 57 * , 577 ( 1998 ) .. m. babilon , t. hartmann , p. mohr , k. vogt , s. volz , and a. zilges , phys.rev.c * 65 * , 037303 ( 2002 ) . n. pietralla , phys.rev.c * 59 * , 2941 ( 1999 ) . w. andrejtscheff , c. kohstall , p.
von brentano , c. fransen , u. kneissl , n. pietralla , and h.h .
pitz , phys.lett.*b506 * , 239 ( 2001 ) .
metzger , phys.rev.c * 14 * , 543 ( 1976 ) . f.r .
metzger , phys.rev.c * 18 * , 2138 ( 1978 ) .
t. eckert , o. beck , j. besserer , p. von brentano , r. fischer , r .- d .
herzberg , u. kneissl , j. margraf , h. maser , a.nord , n. pietralla , h.h.pitz , s.w .
yates , and a. zilges , phys.rev.c * 56 * , 1257 ( 1997 ) .
voronov , d.t .
thoa , and v.yu .
ponomarev , bull.acad.sci .
ussr , ser.phys .
* 48 * , 190 ( 1984 ) .
n. pietralla , c. fransen , a. gade , n.a .
smirnova , p. von brentano , v. werner , and s.w .
yates , phys.rev.c * 68 * , 031305(r ) ( 2003 ) . n. pietralla , p. von brentano , r.f .
casten , t.otsuka , and n.v .
zamfir , phys.rev.lett .
* 73 * , 2962 ( 1994 ) .
n. pietralla , p. von brentano , t.otsuka , and r.f .
casten , phys.lett . *
b349 * , 1 ( 1995 ) . yu.v .
palchikov , p. von brentano , and r.v .
jolos , phys.rev.c * 57 * , 3026 ( 1998 ) .
soloviev , _ theory of complex nuclei _ , oxford , pergamon press , 1976 .
m. grinberg , ch .
stoyanov , n. tsoneva , particles and nuclei , * 29 * , 1456 ( 1998 ) .
vanhoy , j.m .
anthony , b.m .
haas , b.h .
benedict , b.t .
meehau , s.f .
hicks , c.m .
davoren , and c.l .
lundstedt , phys.rev.c * 52 * , 2387 ( 1995 ) .
smirnova , n. pietralla , t. mizusaki , and p. van isacker , nucl.phys . *
a678 * , 235 ( 2000 ) .
a. bohr and b.r .
mottelson , _ nuclear structure _ ,
2 ( benjamin , 1975 ) .
ponomarev , ch .
stoyanov , n. tsoneva , and m. grinberg , nucl.phys . *
a635 * , 470 ( 1998 ) .
ponomarev , eur.phys.j . a * 6 * , 243 ( 1999 ) .
n. tsoneva , h. lenske , ch .
stoyanov , nucl.phys.*a731 * , 273 ( 2004 ) .
f. iachello , a.d .
jackson , phys.lett .
* 108b * , 151 ( 1982 ) .
.[tab : table1 ] the experimental ( exp ) and calculated electric dipole transition matrix elements for cd , sn and ba isotopes obtained including all contributions ( total ) and without contribution of the two quasiparticle admixture to the @xmath0 state ( t=0 ) ( in units @xmath174 ) . [ cols="^,^,^,^",options="header " , ] | the properties of @xmath0 two phonon states and the characteristics of e1 transition probabilities between low lying collective states in spherical nuclei are analysed within the @xmath1phonon approach to the description of collective states .
several relations between observables are obtained .
microscopic calculations of the e1 @xmath2 transition matrix elements are performed on the basis of the rpa .
a satisfactory description of the experimental data is obtained . |
as explained in detail elsewhere in this review , a precise measurement of the lifetimes of the various weakly decaying charm and beauty hadrons possesses great experimental value per se as well as in searches for @xmath0 and @xmath1 oscillations .
yet there exists also a strong theoretical interest in determining and interpreting those lifetimes ; i want to sketch that first in rather qualitative terms and give specifics later .
weak decays of hadrons depend on fundamental parameters of the standard model , in particular on the km parameters and quark masses .
it is eminently important to reliably determine their values from data .
alas this is easier said than done theoretically ( and experimentally ) ! for in such an endeavour we have to face the `` dichotomy of the two worlds '' . on the one hand
there is the `` theorists world '' where quarks and gluons are the relevant strongly interacting entities ; it is in this short - distance or femto world where theorists like to formulate their fundamental theories . on the other hand
there is the `` real world '' where hadrons constitute the relevant degrees of freedom ; it is in that world where everyone ( including theorists ) lives and measurements are performed . to formulate predictions from the theorists world in the language of the real world and to translate findings from the real world back into the idioms of the theorists world
represents the theoretical challenge one faces .
one quantitative measure for the difference between the two worlds is provided by the lifetimes of the weakly decaying hadrons carrying the same flavour . on the quark level
there is obviously only a single lifetime for a given flavour . yet in the ` real world ' hadrons carrying the same flavour quantum number possess different ( and even vastly different ) lifetimes ; e.g. , @xmath2 , @xmath3 and @xmath4 . on
the other hand deviations of the lifetime ratios from unity evidently decrease for an increasing heavy - flavour mass .
this is as expected in a simple ` two - component ' picture : in the limit of @xmath5 the mass of the heavy - flavour quark @xmath6 going to infinity the ` spectator ansatz ' should hold where the lifetimes of all hadrons @xmath7 containing @xmath6 coincide ; for finite values of @xmath5 there are _ pre - asymptotic _ corrections .
two lines of reasoning support this qualitative picture : ( @xmath8 ) the decay width of a quark @xmath6 increases very quickly with its mass @xmath5 : @xmath9 for @xmath10 changing into @xmath11 for @xmath12 .
( rather than @xmath13 ) reflects the coupling of the _ longitudinal _ @xmath14 boson the original higgs field to the top quark .
] therefore for @xmath5 sufficiently large , its decay width is bound to exceed @xmath15 , i.e. the quark @xmath6 decays before it can hadronize .
gev @xcite . ]
then there is obviously a universal lifetime for such a heavy flavour and the spectator ansatz applies trivially .
( @xmath16 ) analysing all _ non_-spectator reactions explicitely , one finds their widths to increase with a smaller power of @xmath5 than the spectator process ; thus one arrives at the spectator ansatz as the asymptotic case .
the most relevant phenomenological question then is how quickly the limit of universal lifetimes is approached .
some mechanisms had been identified very early on that generate differences in the lifetimes of hadrons @xmath7 with the same heavy flavour @xmath6 : `` weak annihilation '' ( = wa ) of @xmath6 with the light valence antiquark for mesons or `` w scattering '' ( = ws ) with the valence diquark system for baryons . such an analysis had first been undertaken for charm decays .
since wa contributes to cabibbo allowed decays of @xmath17 , but not of @xmath18 mesons ( in the valence quark description ) , it creates a difference in @xmath19 vs. @xmath20 .
however the wa rate is doubly suppressed relative to the spectator rate , namely by the helicity factor @xmath21 with @xmath22 denoting the largest mass in the final state and by the ` wavefunction overlap ' factor @xmath23 reflecting the practically zero range of the low - energy weak interactions : @xmath24 therefore it had originally been suggested that already charm hadrons should possess approximately equal lifetimes .
it then came as quite a surprise when observations showed it to be otherwise in particular since the first data suggested a considerably larger value for @xmath25 than measured today .
this near - shock caused a re - appraisal of the theoretical situation ; its results at that time can be summarized in four main points : \(i ) one source for a lifetime difference had too quickly been discarded as insignificant .
cabibbo - allowed nonleptonic decays of @xmath18 but not of @xmath17 mesons produce two antiquarks in the final state that carry the same flavour : @xmath26 { \rightarrow}(s\bar d u ) \bar d\ ] ] thus one has to allow for the _ interference _ between different quark diagrams in @xmath18 , yet not in @xmath17 decays ; the @xmath27 valence antiquark in @xmath18 mesons thus ceases to play the role of an uninvolved bystander and a difference in the @xmath18 vs. @xmath17 lifetimes will arise .
while this interference had been included in descriptions of the @xmath28 two - body modes it was ignored for total widths . for it was thought ( often without stating it explicitely ) that the required coherence would not be maintained between the two amplitudes when applied to inclusive transitions .
this assumption was challenged in ref.@xcite where it was argued that even the two inclusive amplitudes remain sufficiently coherent .
the interference turns out to be destructive , i.e. it prolongs @xmath20 over @xmath19 , but only once the qcd radiative corrections have been included .
this effect is usually referred to as ` pauli interference ' ( = pi ) although such a name would be misleading if it is interpreted as suggesting that the interference is automatically destructive .
\(ii ) it was argued @xcite that the helicity suppression of the wa contribution to @xmath29 decays can be vitiated .
the diagram in fig.1 contains gluon bremsstrahlung off the initial antiquark line in the @xmath14-exchange reaction ; evaluating this particular diagram explicitely one finds : @xmath30 with @xmath31 denoting the average energy of the initial antiquark @xmath32 .
term in the amplitude derives from the propagator in the diagram . ]
using a non - relativistic wavefunction for the decaying meson one has @xmath33 .
this contribution , although of higher order in @xmath34 , would dominate over the lowest order term @xmath35 since helicity suppression has apparently been vitiated and the decay constant @xmath36 is now calibrated by @xmath31 with @xmath37 rather than @xmath38 .
the spectator picture would still apply at asymptotic quark masses , since @xmath39 as @xmath40 due to @xmath41 . yet
if eq.(3 ) were indeed to hold , it would have a dramatic impact on the theoretical description of weak heavy - flavour decays : the impact of this particular pre - asymptotic correction , namely wa , would be enhanced considerably and actually be quite significant even in beauty decays .
alternatively it had been suggested @xcite that the wavefunction of the @xmath29 meson contains a @xmath42 component where the @xmath43 pair forms a spin - one configuration with the gluon @xmath44 balancing the spin of the @xmath45 pair .
both effects , namely pi and wa , work in the same direction , i.e. both enhance @xmath20 over @xmath19 .
\(iii ) a rich structure emerges in the decays of charm baryons @xcite : @xmath46 on the one hand ws contributes to the cabibbo allowed @xmath47 and @xmath48 decays ; one should also keep in mind that ws is _ not _ helicity suppressed in baryon decays already to lowest order in the strong coupling .
it is still reduced in size by the corresponding wavefunction overlap ; yet that is at least partially off - set by ws being described by two - body phasespace versus the three - body phase space of the spectator transition ; this relative enhancement in phase space can be estimated to be roughly of order @xmath49 .
@xmath46 pi affects the @xmath47 , @xmath50 and @xmath51 widths in various ways , generating destructive as well as constructive contributions !
this also strongly suggests that it is very hard to make reliable numerical predictions for these baryonic lifetimes ; yet the overall qualitative pattern has been predicted : @xmath52 together with @xmath53 \(iv ) pre - asymptotic corrections might be sizeable in the lifetime ratios of beauty hadrons .
reviews of these phenomenological descriptions can be found in @xcite .
it turned out , as discussed in more detail later on , that some of the phenomenological descriptions anticipated the correct results : it is pi that provides the main engine behind the @xmath18-@xmath17 lifetime ratio ; @xmath47 is considerably shorter - lived than @xmath17 ; the observed charm baryon lifetimes do obey the hierarchy stated in eq.(4 ) .
nevertheless the phenomenological treatments had significant shortcomings , both of a theoretical and of a phenomenological nature : ( i ) no agreement had emerged in the literature about how corrections in particular due to wa and ws scale with the heavy quark mass @xmath5 .
( ii ) accordingly no clear predictions could be made on the lifetime ratios among beauty hadrons , namely whether @xmath54 and @xmath55 differ by a few to several percent only , or by 20 - 30 % , or by even more !
( iii ) no unequivocal prediction on @xmath56 or @xmath57 had appeared .
( iv ) in the absence of a systematic treatment it is easy to overlook relevant contributions , and that is actually what happened ; or the absence of certain corrections had to be postulated in an ad - hoc fashion
. thus there existed an intellectual as well as practical need for a description based on a systematic theoretical framework rather than a set of phenomenological prescriptions . in the last few years
we have succeeded in showing that the non - perturbative corrections to heavy - flavour decays can be expressed through a _
systematic _ expansion in _ inverse _ powers of @xmath5 .
a simple analogy with nuclear @xmath16 decay can illustrate this point .
there are two effects distinguishing the decays of neutrons bound in a nucleus from the decay of free neutrons : \(a ) nuclear binding effects ; \(b ) the impact of pauli statistics correlating the electrons surrounding the nucleus with those emerging from @xmath16 decay .
the typical energies of the bound electrons
@xmath58 are certainly small compared to @xmath59 , the energy released in the decay ; let us assume
although this is not true in reality that also the nuclear binding energies @xmath60 were small compared to @xmath59 .
in that case nuclear @xmath16 decays would proceed , to a good approximation , like the decays of _ free _ neutrons ; corrections to this simple ` spectator ' picture could be computed via an expansion in powers of @xmath61 and @xmath62 . in practice ,
however , the corrections for nuclear @xmath16 decay are incorporated by explicitely using the wave functions of the bound nucleons and electrons ( obtained with the help of some fairly massive computer codes ) .
there arise analogous corrections to the decay rate for a quark @xmath6 inside a hadron @xmath7 : \(a ) interactions of the decaying quark with other partons in the hadron ; this includes wa of @xmath6 with the light valence antiquark for mesons or ws with the valence diquark system for baryons ; they correspond to k capture of bound electrons by a heavy nucleus in the preceding example .
\(b ) pi effects of the decay products with other partons in the hadron ; e.g. : @xmath63 or @xmath64 .
they prolong the lifetimes of @xmath18 and @xmath65 mesons .
the difference to the example of nuclear @xmath16 decay is quite obvious : even in the limit @xmath66 a non - relativistic bound - state treatment is inapplicable since the dynamical degrees of freedom of the heavy - flavour hadron @xmath7 can not fully be described by a hadronic wavefunction .
the most reliable approach is then to evaluate weak decay rates of heavy - flavour hadrons through an expansion in powers of @xmath67 where @xmath68 represents a hadronic scale @xmath69 1 gev .
the first few terms in this series should yield a good approximation for beauty decays ; the situation for charm decays is a priori unclear ( and at present remains so a posteriori as well ) ; this will be discussed in detail later on .
the vice of hadronization is then transformed into ( almost ) a virtue : the weak decays of heavy - flavour hadrons constitute an intriguing and novel laboratory for studying strong dynamics through their interplay with the weak forces and this is the secondary motivation for studying them . to be more specific : the heavy - flavour mass @xmath5 provides an expansion parameter that allows to deal with the non - perturbative dynamics of qcd in a novel way .
the formalism to be employed combines the @xmath70 expansion with other elements derived from qcd proper without having to invoke a ` deus ex machina ' in contrast to phenomenological descriptions .
there is one concept underlying , in one form or another , all efforts to deal with hadronization , namely the notion of quark - hadron duality ( henceforth referred to as _ duality _ for short ) . in its broadest formulation
it can be stated as follows : sufficiently inclusive transition rates between hadronic systems can be calculated in terms of quarks and gluons .
while the general validity of this concept has not been established in a rigorous fashion , new light has been shed on its nature , validity and applicability by @xmath70 expansions .
lifetimes obviously represent the most inclusive quantity where duality should be applicable .
nonleptonic transitions are certainly more complex than semileptonic ones ; yet i will argue later that while there probably exists a quantitative difference in the degree to which duality holds in semileptonic and in nonleptonic decays , there is _ no qualitative _ one .
deatiled measurements of lifetimes are then theoretically important not only to translate data on the semileptonic branching ratio into a determination of the semileptonic width , but also in their own right
. dedicated and comprehensive studies of both charm and beauty decays are called for .
the km parameters relevant for charm decays @xmath71 and @xmath72 are well known through unitarity constraints of the 3x3 km matrix , in contrast to the situation in beauty decays , and i consider it unlikely that charm decay studies can improve on that . on the other hand those can be used to calibrate our theoretical tools before applying them to beauty decays . before concluding this general introduction i want to point out a less straightforward aspect of accurate lifetime measurements : decay rate evolutions in proper time for neutral mesons will not follow a single exponential function when particle - antiparticle oscillations occur .
for there exist two distinct mass eigenstates with @xmath73 .
the quantity @xmath74 generates a deviation of the form @xmath75 or @xmath76 ; @xmath77 leads to the emergence of a second exponential .
the general expression reads as follows : @xmath78{\rightarrow}f)/dt \ ; \propto e^{-\gamma t}\cdot g(t)\ ] ] @xmath79 where @xmath80 \right]\ ] ] @xmath81 - re [ \frac{q}{p}\bar \rho ( f ) ] \right]\ ] ] @xmath82 \ ; , \ ; \bar \rho ( f)= \frac{\bar a(f)}{a(f ) } \eqno(5b)\ ] ] with @xmath83 and @xmath84 denoting the amplitude for @xmath85 { \rightarrow}f$ ] and @xmath86 { \rightarrow}f$ ] , respectively .
in this section i summarize the numerical results and sketch the main elements of the underlying theoretical treatment in a way that can satisfy the casual reader .
a more in - depth discussion of the theoretical concepts and tools will be presented in subsequent sections .
expanding the width for the decay of a heavy - flavour hadron @xmath7 containing @xmath6 into an inclusive final state @xmath87 through order @xmath88 one obtains @xcite @xmath89 @xmath90 , \eqno(6)\ ] ] where the dimensionless coefficients @xmath91 depend on the parton level characteristics of @xmath87 ( such as the ratios of the final - state quark masses to @xmath5 ) ; @xmath92 denotes the appropriate combination of km parameters , and @xmath93 with @xmath94 being the gluonic field strength tensor . the last term in eq.(6 )
implies also the summation over the four - fermion operators with different light flavours @xmath95 .
the expectation values of the local operators appearing on the right - hand side of eq.(6 ) contain the relativistic normalization of the state @xmath96 : @xmath97 it is through the quantities @xmath98 that the dependence on the _ decaying hadron _
@xmath7 , and on non - perturbative forces in general , enters , instead of through wavefunctions as in nuclear @xmath16 decay .
since these are matrix elements for on - shell hadrons @xmath7 , one sees that @xmath99 is indeed expanded into a power series in @xmath100 .
for @xmath66 the contribution from the lowest dimensional operator obviously dominates ; here it is the dimension three operator @xmath101 . a heavy quark expansion yields : @xmath102 i.e. @xmath103 for @xmath66 , reflecting the unit of heavy - flavour common to all hadrons @xmath7 .
eqs.(6,8 ) show that the leading nonperturbative corrections are of order @xmath104 rather than @xmath70 ; therefore they can be expected to be rather small in beauty decays .
the expectation values appearing in the @xmath104 and @xmath88 contributions can be related to other observables and their size thus be extracted in a reliable manner . applying the master formula eq.(6 ) to lifetimes one arrives at the following general results : @xmath46 the leading contribution to the total decay width is given by the first term on the right - hand - side of eq.(8 ) that is _ common _ to all hadrons of a given heavy - flavour quantum number . for @xmath66 one
has thus derived from qcd proper the spectator picture attributing equal lifetimes to all hadrons of a given heavy - flavour ! this is not a surprising result
after all without hadronization there is of course only a unique lifetime , but it is gratifying nevertheless .
@xmath46 lifetime differences first arise at order @xmath104 and are controlled by the expectation values of two dimension five operators , to be discussed later .
these terms , which had been overlooked in the original phenomenological analyses , generate a lifetime difference between heavy - flavour _ baryons _ on one side and _ mesons _ on the other .
yet apart from small isospin or @xmath105 breaking they shift the meson widths by the same amount and thus do _ not _ lead to differences among the meson lifetimes .
@xmath46 those emerge at order @xmath88 and are expressed through the expectation values of four - fermion operators which are proportional to @xmath106 with @xmath107 denoting the decay constant for the meson @xmath108 .
contributions from what is referred to as wa and pi in the original phenomenological descriptions ( see the discussion in sect.1 ) are systematically and consistently included .
further contributions to the baryon - meson lifetime difference also arise at this level due to ws .
@xmath46 since the transitions @xmath109 or @xmath110 are described by an isosinglet operator one can invoke the isospin invariance of the strong interactions to deduce for the semileptonic widths @xmath111 @xmath112 and therefore @xmath113 @xmath114 up to small corrections due to the km [ cabibbo ] suppressed transition @xmath115 [ @xmath116 which changes isospin by half a unit .
the spectator picture goes well beyond eqs.(9 ) : it assigns the same semileptonic width to all hadrons of a given heavy flavour .
yet such a property can not be deduced on _ general _ grounds : for one had to rely on @xmath117 symmetry to relate @xmath118 to @xmath119 or @xmath120 to @xmath121 and no symmetry can be invoked to relate the semileptonic widths of mesons and baryons .
there is actually a wa process that generates semileptonic decays on the cabibbo - allowed level for @xmath122 [ and also for @xmath123 , but not for @xmath17 and @xmath18 nor for @xmath124 , @xmath65 and @xmath125 mesons : the hadrons are produced by gluon emission off the @xmath126 [ or the @xmath127 line . yet since the relative weight of wa is significantly reduced in meson decays , one does not expect this mechanism to change @xmath118 significantly relative to @xmath119 .
actually there are corrections to the semileptonic widths arising already in order @xmath104 . on
rather general grounds one predicts the expectation values @xmath128 and @xmath129 to be largely independant of the flavour of the light antiquark in the meson and therefore @xmath130 @xmath131 like in the naive spectator picture , but for non - trivial reasons . on the other hand , as explained later , the values of the expectation values for these operators are different when taken between baryon states and one expects @xmath132 the remarks above can be summarized as follows : @xmath133 ^ 0 ) = \gamma ( [ q\bar q]^{\pm } ) + { \cal o}(1/m_q^2 ) \eqno(12a)\ ] ] @xmath134 ^ 0 ) = \gamma ( [ q\bar q]^{\pm } ) + { \cal o}(1/m_q^3 ) \eqno(12b)\ ] ] @xmath134 ^ 0 ) > \gamma ( [ q\bar q]^{\pm } ) + { \cal o}(1/m_q^4 ) \eqno(12c)\ ] ] quantitative predictions for the lifetime ratios of beauty hadrons through order @xmath135 are given in table [ table1 ] together with present data .
@xmath136 & @xmath137>1 $ ] & @xmath138 + & ( mainly due to _ destructive _ interference ) & + @xmath139 & @xmath140 & @xmath141 + @xmath142&@xmath143 & @xmath144 + the expectations @xcite for the lifetimes of charm hadrons are juxtaposed to the data in table [ table2 ] .
the overall agreement is rather good given the theoretical and experimental uncertainties .
it has to be kept in mind that the _ numerical _ predictions on _ baryon _ lifetimes involve quark model results for the size of the relevant expectation values .
this is indicated in tables 1 and 2 by the asterisk .
there is however one obvious discrepancy : the observed @xmath145 lifetimes is significantly shorter than predicted .
while as indicated just above predictions on _ baryon _ lifetimes have to be taken with a grain of salt , one can not change the prediction on @xmath142 with complete theoretical impunity ; this will be explained later on .
@xmath25 & @xmath146 [ for @xmath147 mev ] & @xmath148 + & ( mainly due to _ destructive _ interference ) & + @xmath149 & @xmath150 few @xmath151 & @xmath152 + @xmath153&@xmath154 & @xmath155 + @xmath156&@xmath157 & @xmath158 + @xmath159&@xmath160 & @xmath161 + @xmath162&@xmath163 & @xmath164 + these numerical predictions bear out the general expectations expressed above . this will be discussed below in more detail . for proper perspective on the role played by the heavy quark expansion
some comments should be made at this point : \(i ) as stated before , the early phenomenological treatments had already identified the relevant mechanisms by which lifetime differences arise , namely pi , wa and ws . furthermore the authors of ref.@xcite had argued
correctly that pi constitutes the dominant effect .
yet it is the heavy quark expansion that provides a firm theoretical underpinning to these expectations .
in particular it clarifies both conceptually and quantitatively the role of wa and how its weight scales with @xmath70 .
\(ii ) contributions of order @xmath70 would dominate all other effects if they were present !
the heavy quark expansion shows unequivocally that they are absent in total rates ; this has important ramifications , to be pointed out later .
\(iii ) corrections of order @xmath104 differentiate between the decays of mesons and of baryons .
they had been overlooked before .
the heavy quark expansion has identified them and basically determined their size .
the weak decay of the heavy quark @xmath6 inside the heavy - flavour hadron @xmath7 proceeds within a cloud of light degrees of freedom ( quarks , antiquarks and gluons ) with which @xmath6 and its decay products can interact strongly .
it is the challenge for theorists to treat these initial and final state hadronization effects . among the existing four post - voodoo theoretical technologies
qcd sum rules , lattice qcd , heavy quark effective theory ( hqet ) and heavy quark expansions only the last one deals with inclusive decays . on the other hand it benefits , as we will see , from the results of the other three technologies , since those can determine the size of some of the relevant expectation values .
whereas it is the special feature of hqet that @xmath5 is removed from its lagrangian . ] in analogy to the treatment of @xmath165 one describes the transition rate into an inclusive final state @xmath87 through the imaginary part of a forward scattering operator evaluated to second order in the weak interactions @xcite : @xmath166 where @xmath167 denotes the time ordered product and @xmath168 the relevant effective weak lagrangian expressed on the parton level .
if the energy released in the decay is sufficiently large one can express the _ non - local _ operator product in eq.(13 ) as an infinite sum of _ local _ operators @xmath169 of increasing dimension with coefficients @xmath170 containing higher and higher powers of @xmath70 . that controls the expansion .
] the width for @xmath171 is then obtained by taking the expectation value of @xmath172 between the state @xmath7 : @xmath173 this master formula holds for a host of different inclusive heavy - flavour decays : semileptonic , nonleptonic and radiative transitions , km favoured or suppressed etc . for semileptonic and nonleptonic decays
treated through order @xmath88 it takes the form already given in eq .
( 6 ) : @xmath174 @xmath90\ ] ] integrating out the two internal loops in figs.2 and 3 yields the operators @xmath101 and @xmath175 , respectively ; the black boxes represent @xmath176 .
likewise the diagrams in figs.4 generate four - quark operators ; fig.4a and 4b differ in how the light quark flavours are connected inside the hadron @xmath7 : cutting a quark line @xmath95 in fig.2 and connecting it to the @xmath32 constituent of the @xmath7 meson , as shown in fig.4a , one has a wa transition operator ; cutting instead the @xmath32 line and connecting it to the @xmath7 constituents , see fig.4b , leads to the four - fermion operator describing pi .
inspection of these diagrams suggests a rule - of - thumb that is borne out by explicit computations : the dimensionless coefficients @xmath177 are smaller for the two - loop diagrams of figs.2 and 3 than for the one - loop diagrams of figs.4 . integrating out the internal loops of fig.3
actually generates not only the dimension - five chromomagnetic operator , but also dimension - six quark - gluon operators , namely @xmath178 and @xmath179 . using the equations of motion
those can be reduced to four - quark operators .
yet their coefficients are suppressed by @xmath180 relative to the coefficient coming from figs.4 ; therefore one can discard their contributions at the present level of accuracy .
one important general observation should be made at this point .
the two operators @xmath101 and @xmath175 obviously do not contain any light quark or antiquark fields ; those enter first on the dimension - six level .
their expectation values are , as discussed below , practically insensitive to the light degrees of freedom up to terms of order @xmath88 .
the inclusive semileptonic and non - leptonic widths through order @xmath104 are thus controled by the _ decay _ of a single quark @xmath6 , yet occurring in a non - trivial background field ; the _ spectator _ contribution universal to the widths for all hadrons @xmath7 of a given heavy flavour constitutes its asymptotic part : @xmath181 @xmath182 however @xmath183 possesses a different value for mesons and baryons , as discussed below .
the expansion introduced above will be of practical use only if its first few terms provide a good approximation and if one can determine the corresponding expectation values in a reliable manner .
the latter can be achieved in two complementary ways : one relates the matrix element in question to other observables through a heavy quark expansion ; or one harnesses other second - generation theoretical technologies , namely qcd sum rules , qcd simulations on the lattice and heavy quark effective theory , to calculate these quantities .
the results obtained so far are listed below : \(a ) using the equations of motion one finds for the leading operator @xmath101 : @xmath184q}{2m_q^2}+ g_s^2\frac{\bar q\gamma _
0t^iq \sum _
q \bar q\gamma _ 0 t^iq}{4m_q^3 } + { \cal o}(1/m_q^4)\eqno(16)\ ] ] with the sum in the last term running over the light quarks @xmath95 ; the @xmath185 denote the colour @xmath186 generators .
i have ignored total derivatives in this expansion since they do not contribute to the @xmath7 expectation values . there
emerge two dimension - five operators , namely @xmath187 and @xmath188 with @xmath189 denoting the covariant derivative .
the first one represents the square of the spatial momentum of the heavy quark @xmath6 moving in the soft gluon background and thus describes its kinetic energy .
the second one constitutes the chromomagnetic operator that already appeared in eq.(6 ) .
the relevant dimension - six operators can be expressed as four - fermion operators .
since @xmath190 constitutes the noether current for the heavy - flavour quantum number one has @xmath191 leading to eq.(8 ) .
\(b ) the first non - perturbative corrections arise at order @xmath104 , and they can enter in two ways , namely through the chromomagnetic operator in the ope of @xmath172 or through the @xmath70 expansion of the @xmath7 expectation values of the local operator @xmath101 .
\(c ) for the chromomagnetic operator one finds @xmath192 where @xmath193 and @xmath194 denote the pseudoscalar and vector mesons , respectively . for @xmath195 and @xmath196 baryons
one has instead @xmath197 since the light diquark system inside @xmath195 and @xmath196 carries no spin .
these baryons thus represent ( though only through order @xmath104 ) a simpler system than the mesons where the light antiquark of course carries spin .
for @xmath198 baryons on the other hand the situation is different : for the light di - quarks @xmath199 form a spin - one configuration .
the @xmath198 expectation value is then given by the spin - splitting in the baryon masses : @xmath200\eqno(17c)\ ] ] \(d ) we thus have @xmath201 @xmath202 with the notation @xmath203 .
the reason for the appearance of the kinetic energy term in eqs .
( 18 ) is quite transparent : the first two terms on the r.h.s . of eqs.(18 ) represent the mean value of the factor @xmath204 reflecting the time dilation slowing down the decay of the quark @xmath6 moving inside @xmath7 .
\(e ) the value of @xmath205 has not been determined yet in an accurate way ( details will be given later ) ; we know , though , it has to obey the inequality @xcite @xmath206 which because of eq.(17 ) is useful only for @xmath207 . on the other hand ,
the difference in the kinetic energy of q inside baryons and mesons can be related to the masses of charm and beauty hadrons @xcite : @xmath208- [ \langle m_d\rangle -m_{\lambda _ c } ] \ } \eqno(20)\ ] ] with @xmath209 denoting the ` spin averaged ' meson masses : @xmath210 and likewise for @xmath211 . in deriving eq.(20 ) it was implicitely assumed that the @xmath212 quark can be treated also as heavy ; in that case @xmath213 holds .
\(f ) eqs.(17 - 18,20 ) show that the two _ dimension - five _ operators do produce differences in @xmath193 vs. @xmath214 vs. @xmath198 lifetimes ( and semileptonic widths ) of order @xmath104 ; i.e. , @xmath215 , as defined in eqs.(15 ) , depends on whether @xmath7 represents a ( pseudoscalar ) meson or a baryon . in the latter case
it is sensitive to whether the light diquark system carries spin zero ( @xmath195 , @xmath196 ) or one ( @xmath198 ) .
the origin of such a difference is quite transparent .
( i ) the heavy quark @xmath6 can be expected to possess a somewhat different kinetic energy inside a meson or a baryon ; it can also be affected by spin - spin interactions with the light di - quarks .
this difference in the heavy quark motion means that lorentz time dilation prolongs the lifetime of the quark @xmath6 to different degrees in baryons than in mesons .
( ii ) while there exists a spin interaction between @xmath6 and the antiquark @xmath32 in the meson or the @xmath199 system in the @xmath198 baryon , there is no such coupling inside @xmath195 or @xmath196 . yet
these effects do not generate a significant difference among _ meson _ lifetimes since their expectation values satisfy isospin and @xmath105 symmetry to a good degree of accuracy .
\(g ) differences in meson lifetimes are generated at order @xmath88 by _ dimension - six _ four - quark operators describing pi as well as wa ; the weight of the latter is , as explained in more detail in the next section , greatly reduced .
the expectation values of these operators look very similar to the one controlling @xmath216 oscillations : @xmath217 with @xmath218 denoting the decay constant for the meson @xmath7 ; the so - called bag factor has been set to unity , i.e. _ factorization _ has been assumed .
\(h ) the situation becomes much more complex for @xmath195 and baryon decays in general . to order @xmath88 baryons lose the relative simplicity mentioned above : there are several different ways in which the valence quarks of the baryon can be contracted with the quark fields in the four - quark operators ; furthermore ws is _ not _ helicity suppressed and thus can make a sizeable contribution to lifetime differences ; also the pi effects can now be constructive as well as destructive
. finally one can not take recourse to factorisation as a limiting case .
thus there emerge three types of numerically significant mechanisms at this order in baryon decays in contrast to meson decays where there is a _ single _ dominant source for lifetime differences and their strength can not be expressed in terms of a single observable like @xmath218 . at present
we do not know how to determine the relevant matrix elements in a model - independant way .
the best available guidance and inspiration is to be gleaned from quark model calculations with their inherent uncertainties .
this analysis had already been undertaken in the framework of phenomenological models @xcite with the following qualitative results : ws contributes to @xmath195 and @xmath219 decays ; a _ destructive _ interference affects @xmath195 , @xmath196 and @xmath220 transitions .
a _ constructive _ interference enhances @xmath221 and @xmath51 decays .
one thing should be obvious already at this point : with terms of different signs and somewhat uncertain size contributing to differences among baryon lifetimes one has to take even semi - quantitative predictions with a grain of salt ! the probability amplitude for ws or interference to occur is expressed in quark models by the wavefunction for @xmath6 and one of the light quarks at zero spatial separation .
this quantity is then related to the meson wavefunction through the hyperfine splitting in the baryon and meson masses @xcite : @xmath222 with @xmath223 denoting the _ phenomenological constituent _ mass of the light quark @xmath95 to be employed in these models . for the baryonic expectation values
one then obtains @xmath224 and likewise for @xmath196 . for the @xmath198 one
has instead @xmath225 reflecting the different spin substructure which already had caused the difference in eqs.(17a ) vs. ( 17b )
. for simplicity one has assumed in these expressions that the hyperfine splitting in the baryon masses are universal . to summarize this discussion : @xmath46 the non - perturbative corrections to total widths through order @xmath88
are expressed in terms of three non - trivial ( types of ) matrix elements : @xmath226 , @xmath227 and @xmath228 .
@xmath46 the size of @xmath226 is well - known for the mesons and @xmath195 and @xmath196 baryons ; reasonable estimates can be obtained for @xmath198 baryons .
@xmath46 a lower bound exists for the mesonic expectation value @xmath229 , but not for the baryonic one .
various arguments strongly suggest that @xmath227 is close to its lower bound .
furthermore the quantity @xmath230 , which is highly relevant for the difference in baryon vs. meson lifetimes ( and semileptonic widths ) can be extracted from mass measurements .
@xmath46 the expectation value of the four - quark operators between _ mesons _ can be expressed in terms of a single quantity , namely the decay constant @xmath231 .
@xmath46 however the _ baryonic _ expectation values of the four - quark operators constitute a veritable pandora s box , which at present is beyond theoretical control . on the other hand measurements of inclusive weak decay rates and the resulting understanding of the role played by ws and pi
present us with a novel probe of the internal structure of the heavy - flavour baryons .
before explaining these predictions in more detail and commenting on the comparison with the data in the next subsection , i want to add four general remarks and elaborate on them . while i hope they will elucidate the underlying concepts , they are not truly essential for following the subsequent discussion and the more casual reader can ignore them .
\(a ) _ on the validity of the @xmath70 expansion in the presence of gluon emission : _ the wa contribution to the decay width of pseudoscalar mesons to lowest order in the strong coupling has been sketched in eq.(2 ) .
the diagram in fig.1 contains gluon bremsstrahlung off the initial antiquark line in the @xmath14-exchange reaction ; evaluating this diagram explicitely one finds , as stated before : @xmath232 with @xmath31 denoting the average energy of the initial antiquark @xmath32 . since @xmath233
, it depends very sensitively on the low - energy quantity @xmath234 , and its magnitude is thus quite uncertain .
the presence of a @xmath235 ( or @xmath236 ) term would also mean that the nonperturbative corrections to even inclusive decay widths of heavy - flavour hadrons are not ` infrared safe ' and can not be treated consistently through an expansion in inverse powers of the heavy quark mass @xmath5 only .
fortunately the contribution stated in eq.(26 ) turns out to be spurious for _ inclusive _ transitions .
this can best be seen by studying the imaginary part of the forward scattering amplitude @xmath237 as shown in fig.5 .
there are actually three poles in this amplitude indicated by the broken lines which represent different final states : one with an on - shell gluon and the other two with an off - shell gluon materializing as a @xmath238 pair and involving interference with the spectator decay amplitude .
these processes are all of the same order in the strong coupling and therefore have to be summed over for an _ inclusive _ decay .
analysing carefully the analyticity properties of the sum of these forward scattering amplitudes one finds @xcite that it remains finite in the limit of @xmath239 , i.e. the amplitude for the _ inclusive _ width does _ not _ contain terms of order @xmath235 or even @xmath240 !
the contribution of wa can then be described in terms of the expectation value of a _ local _ four - fermion operator although the final state is dominated by low - mass hadronic systems . to summarize : the quantity @xmath218 is calibrated by the large mass @xmath5 rather than by @xmath31 ; wa is then only moderately significant even in charm decays .
the @xmath70 scaling thus persists to hold even in the presence of radiative corrections
but only for _ inclusive _ transitions ! this caveat is not of purely academic interest . for
it provides a nice toy model illustrating the _ qualitative _ difference between inclusive and exclusive transitions .
the latter are represented here by the three separate cuts in fig.5 . for charm
decays they correspond to the reactions @xmath241 { \rightarrow}s \bar d g$ ] and @xmath241 { \rightarrow}s \bar d u\bar u$ ] . the two individual transition rates quite sensitively depend on a low energy scale - @xmath31 - which considerably enhances the rate for the first transition while reducing it for the second one .
yet the dependance on @xmath240 disappears from their sum !
\(b ) _ on the fate of the corrections of order @xmath70 : _ the most important element of eq.(2 ) is the one that is _ missing _ ! namely there is no term of order @xmath70 in the total decay width whereas such a correction definitely exists for the mass formula : @xmath242 ( and likewise for differential decay distributions ) .
hadronization in the initial state does generate corrections of order @xmath70 to the total width , as does hadronization in the final state . yet
local colour symmetry enforces that they cancel against each other !
depends on the _ local _ properties of the potential , i.e. on the potential around @xmath6 in a neighourhood of size @xmath70 only ; the nature of the resulting spectrum for @xmath95 whether it is discrete or continuous , etc .
depends of course on the long range properties of the potential , namely whether it is confining or not .
] this can be understood in another more compact ( though less intuitive ) way as well : with the leading operator @xmath101 carrying dimension three only dimension - four operators can generate @xmath70 corrections ; yet there is no independent dimension - four operator @xcite _ once the equation of motion is imposed _ unless one abandons local colour symmetry thus making the operators @xmath243 and @xmath244 independent of each other ( @xmath245 denotes the gluon field ) ! the leading non - perturbative corrections to fully integrated decay widths are then of order @xmath104 and their size is controlled by two dimension - five operators , namely the chromomagnetic and the kinetic energy operators , as discussed above
. their contributions amount to no more than 10 percent for @xmath246 mesons
@xmath248 as borne out by the detailed analysis previewed in table 1 .
\(c ) _ which mass is it ?
_ for a @xmath70 expansion it is of course important to understand which kind of quark mass is to be employed there , in particular since for confined quarks there exists no a priori natural choice .
it had been claimed that the pole mass can and therefore should conveniently be used .
yet such claims turn out to be fallacious @xcite : qcd , like qed , is not borel summable ; in the high order terms of the perturbative series there arise instabilities which are customarily referred to as ( infrared ) renormalons representing poles in the borel plane ; they lead to an _ additive _ mass renormalization generating an _ irreducible uncertainty _ of order @xmath249 in the size of the pole mass : @xmath250 .
while this effect can safely be ignored in a purely perturbative treatment , it negates the inclusion of non - perturbative corrections @xmath251 , since those are then parametrically smaller than the uncertainty @xmath252 in the definition of the pole mass .
this problem can be taken care of through wilson s prescriptions for the operator product expansion : @xmath253 where a momentum scale @xmath254 has been introduced to allow a consistent separation of contributions from long distance and short distance dynamics @xmath255 with the latter contained in the coefficients @xmath256 and the former lumped into the matrix elements .
the quantity @xmath254 obviously represents an auxiliary variable which drops out from the observable , in this case the decay width . in the limit
@xmath257 infrared renormalons emerge in the coefficients ; they cancel against ultraviolet renormalons in the matrix elements .
yet that does _ not _ mean that these infrared renormalons are irrelevant and that one can conveniently set @xmath258 ! for to incorporate both perturbative as well as
non - perturbative corrections one has to steer a careful course between ` scylla ' and ` charybdis ' : while one wants to pick @xmath259 so as to make a heavy quark expansion applicable , one also has to choose @xmath260 s.t .
@xmath261 ; for otherwise the _ perturbative _ corrections become uncontrollable .
wilson s ope allows to incorporate both perturbative and non - perturbative corrections , and _ this underlies also a consistent formulation of hqet _ ; the scale @xmath254 provides an infrared cut - off that automatically freezes out infrared renormalons .
for the asymptotic difference between
the hadron and the quark mass one then has to write @xmath262 .
this nice feature does not come for free , of course : for one has to use a ` running ' mass @xmath263 evaluated at an intermediate scale @xmath254 which presents a technical complication . on the other hand
this quantity can reliably be extracted from data @xcite ; furthermore it drops out from lifetime _ ratios _ , the main subject of this discussion .
\(d ) _ on duality : _ quark - hadron duality equates transition rates calculated on the quark - gluon level with the observable ones involving the corresponding hadrons
provided one sums over a sufficient number of final states . since the early days of the quark model
this concept has been invoked in many different formulations ; among other things it has never been clearly defined what constitutes a sufficient number of final states .
this lack of a precise formulation emerged since duality had never been derived from qcd in a rigorous fashion ; furthermore a certain flexibility in an unproven , yet appealing intuitive concept can be of considerable heuristic benefit .
heavy quark expansions assume the validity of duality .
nevertheless they provide new and fruitful insights into its workings .
as already stated , heavy quark expansions are based on an ope , see eq.(26 ) , and as such are properly defined in euclidean space .
there are two possible limitations to the procedure by which duality is implemented in heavy quark expansions .
( i ) the size of the coefficients @xmath256 is controlled by short - distance dynamics . in concrete applications
they are actually obtained within perturbation theory .
those computations involve integrals over all momenta .
however for momenta below the scale @xmath254 a perturbative treatment is unreliable .
the contributions from this regime might turn out to be numerically insignificant for the problem at hand ; yet in any case one can undertake to incorporate them through the expectation values of higher - dimensional operators the so - called condensates . this procedure is the basis of what is somestimes referred to as the ` practical ' version of the ope . yet it is conceivable that there are significant short - distance contributions from non - perturbative dynamics as well ; instantons provide an illustration for such a complication , although their relevance is quite unclear at present @xcite .
( ii ) once the ope has been constructed and its coefficients @xmath91 determined there , one employs a dispersion relation to analytically continue them into minkowski space .
this means , strictly speaking , that in minskowski space only ` smeared ' transition rates can be predicted , i.e. transition rates averaged over some finite energy range .
this situation was first analyzed in evaluating the cross section for @xmath264 @xcite . through an ope qcd
allows to compute @xmath265 with @xmath266 . if the cross section happens to be a smooth function of @xmath267 as it is the case well above production thresholds
, then one can effectively take the limit @xmath268 to predict @xmath269 for a _ fixed _
energy @xmath270 .
this scenario can be referred to as _ local _ duality . yet
close to a threshold , like for charm or beauty production , with its resonance structure one has to retain @xmath271 gev .
the same considerations are applied to heavy - flavour decays . based on _ global _
duality one can predict ` smeared ' decay rates , i.e. decay rates averaged over a finite energy interval . if the energy release is large enough the decay rate will be a smooth function of it , smearing will no longer be required and _ local _ duality emerges .
how large the energy release has to be for this simplification to occur can not be predicted ( yet ) .
for the onset of local duality is determined by terms that can not be seen in any finite order of the @xmath70 expansion .
however rather general considerations lead to the following expectations : ( i ) this onset should be quite abrupt around some energy scale exceeding usual hadronic scales .
( ii ) local duality should hold to a good degree of accuracy for beauty decays , even for nonleptonic ones .
( iii ) on the other hand there is reason for concern that this onset might not have occurred for charm decays .
this is sometimes referred to by saying that inclusive charm decays might receive significant contributions from ( not - so-)`distant cuts ' .
as discussed before , see eq.(27 ) , the size of matrix elements depends in general on the scale @xmath254 at which they are evaluated .
this will be addressed separately for the different cases .
the scalar dimension - three operator @xmath101 can be expanded in terms of @xmath272 with @xmath273 and operators of dimension five and higher , see eq.(15 ) .
\(i ) dimension - five operators employing eq.(16 ) for @xmath274 and @xmath212 quarks one finds : @xmath275 @xmath276 i.e. , these two matrix elements that have to coincide in the infinite mass limit are already very close to each other .
they are already evaluated at the scale that is appropriate for beauty and charm decays . for the description of @xmath277
decays one needs the expectation value of the beauty as well as charm chromomagnetic operator taken between the @xmath277 state .
while the @xmath278 and @xmath277 masses have not been measured yet , one expects the theoretical predictions @xcite on them to be fairly reliable . with @xmath279 gev and @xmath280 gev
one obtains @xmath281 i.e. , double the value as for the mesons with light antiquarks . from eq.(20 ) one obtains for the _ difference _ in the kinetic energy of the heavy quark inside baryons and mesons @xmath282 using available data ; for charm the same value holds to this order .
we see that at present one can not tell whether the kinetic energy of the heavy quark inside baryons exceeds that inside mesons or not .
the uncertainty in the mass of @xmath145 constitutes the bottleneck ; it would be quite desirable to decrease its uncertainty down to below 10 mev or even less .
for the overall size of @xmath205 we have the following lower bounds from eq.(19 ) : @xmath283 @xmath284 to be conservative one can add a term of at most @xmath285 to the right - hand side of eqs.(31,32 ) reflecting the uncertainty in the scale at which the expectation values are to be evaluated @xcite .
there are various arguments as to why the size of @xmath286 will not exceed this lower bound significantly .
one should note that such values for @xmath286 are surprisingly large : for they state that the momenta of the heavy quark inside the hadron is typically around 600 mev .
an analysis based on qcd sum rules yields a value that is consistent with the preceding discussion @xcite : @xmath287 it can be expected that practically useful values for @xmath288 will be derived from lattice qcd in the forseeable future @xcite .
the leading non - perturbative corrections are thus controlled by the following parameters : @xmath289 @xmath290 @xmath291 the @xmath51 mass hyperfine splitting has not been measured yet . a quark model estimate of 70 mev has been used in evaluating eq.(17b ) .
\(ii ) dimension - six operators the size of the decay constants @xmath292 , @xmath293 , @xmath36 and @xmath294 has not been determined yet with good accuracy .
various theoretical technologies yield @xmath295 @xmath296 @xmath297 @xmath298 where a higher reliability is attached to the predictions for the ratios of decay constants , i.e. eqs.(38,39 ) , than for their absolute size , eqs.(36,37 ) .
recent studies by wa75 and by cleo on @xmath299 yielded values for @xmath294 that are somewhat larger , but still consistent with these predictions .
it can be expected that lattice qcd will produce more precise results on these decay constants in the foreseeable future and that those will be checked by future measurements in a significant way .
yet there are two subtleties to be kept in mind here : ( @xmath8 ) the matrix element of the four - quark operator is related to the decay constant through the assumption of ` vacuum saturation ' .
such an ansatz can not hold as an identity ; it represents an approximation the validity of which has to depend on the scale at which the matrix element is evaluated .
the quantities @xmath292 and @xmath36 are measured in @xmath300 and @xmath301 , respectively and thus probed at the heavy - flavour mass , @xmath302 and @xmath303 . yet for the strong interactions controling the size of the expectation value of the four - quark operator the heavy - flavour mass is a completely foreign parameter .
if vacuum saturation makes ( approximate ) sense anywhere , it has to be at ordinary hadronic scales @xmath304 gev .
that means the decay constant that is observed at the heavy - flavour scale has to be evaluated at the hadronic scale @xmath254 ; this is achieved by the so - called hybrid renormalization to be described later .
( @xmath16 ) the difference in decay widths generated by the four - quark operators is stated as being proportional to @xmath305 with @xmath306 $ ] ; @xmath307 represents the leading or ` static ' term which behaves like @xmath308 for @xmath66 and therefore @xmath309 .
thus asymptotically one has @xmath310 which indeed vanishes like @xmath88 .
yet the decay constant @xmath218 partly as a consequence of its usual definition via an axialvector rather than a pseudoscalar current contains large pre - asymptotic corrections which lead to @xmath311 .
then it is not clear which value to use for @xmath218 when calculating width differences , the asymptotic one @xmath307 or the one including the pre - asymptotic corrections @xmath218 where numerically @xmath312 holds . .
it can be argued that the main impact of some of the dimension - seven operators in the ope for the meson decay width is to renormalise the @xmath310 term to @xmath305 ; yet this issue can be decided only through computing the contributions from dimension - seven operators .
i will return to this issue when discussing beauty and charm decays specifically .
both these considerations also apply when determining the _ baryonic _ expectation values of four - quark operators @xcite . yet as already pointed out that is a considerably more murky affair , as will become quite apparent in the subsequent discussions of the beauty and charm lifetimes . while there are significant uncertainties and ambiguities in the values of the masses of beauty and charm
quarks their difference which is free of renormalon contributions is tightly constrained : @xmath313 this value agrees very well with the one extracted from an analysis of energy spectra in semileptonic @xmath246 decays @xcite .
lifetime ratios are hardly sensitive to this difference ; we list it here mainly for completeness .
within the standard model there is no real uncertainty about the weak forces driving beauty decays ; at the scale @xmath314 they are given by the lagrangian @xmath315\cdot [ v_{ud}^*\bar d_l\gamma _ { \mu}u_l + v_{cs}^*\bar s_l\gamma _ { \mu}c_l ]
\eqno(41)\ ] ] for non - leptonic transitions with an analogous expression for semileptonic ones . in eq.(41 ) i have ignored cabibbo - suppressed transitions and also the @xmath316 coupling since here we are not interested in penguin contributions .
radiative qcd corrections lead to a well - known renormalization at scale @xmath317 , which is often referred to as _ ultra - violet _ renormalization : @xmath318 for @xmath319 transitions and an analogous expression for @xmath320 ; the qcd corrections are lumped together into the coefficients @xmath321 and @xmath322 with @xmath323 @xmath324 ^{\gamma _ { \pm } } , \ ; \ ; \gamma _ + = \frac{6}{33 - 2n_f}=-\frac{1}{2}\gamma _ - \eqno(43b)\ ] ] in the leading - log approximation ; @xmath325 denotes the number of active flavours .
numerically this amounts to : @xmath326 next - to - leading - log corrections modify mainly @xmath322 @xcite : @xmath327 radiative qcd corrections thus lead to a mild enhancement of the original coupling
@xmath328 together with the appearance of an induced operator with a different colour flow : @xmath329 .
later on we will also include the so - called ` hybrid ' renormalization reflecting radiative corrections in the domain from @xmath317 down to @xmath68 , the scale at which the hadronic expectation values are to be evaluated .
such perturbative corrections affect the overall scale of the semileptonic and non - leptonic width ( and thus of the semileptonic branching ratio ) ; but by themselves they can not generate _ differences _ among the lifetimes of beauty hadrons .
those have to be initiated by non - perturbative contributions although their numerical size depends also on perturbative corrections .
the semileptonic and non - leptonic widths of beauty _ hadrons _ @xmath330 through order @xmath331 are given by @xmath332 \ ; , \eqno(45a)\ ] ] @xmath333 @xmath334 - \frac{4}{3 } a_2 \frac{{\langle \mu _ g^2\rangle}_{h_b}}{m_b^2 } \cdot [ i_2(x,0,0 , ) + i_2(x , x,0 ) ] \right ] \ ; , \eqno(45b)\ ] ] @xmath335 the following notation has been used here : @xmath336 and @xmath337 are phase - space factors , namely @xmath338 @xmath339 for @xmath340 and @xmath341 @xmath342 for @xmath320 transitions .
the quantities @xmath343 , @xmath344 , @xmath345 and @xmath346 represent the qcd radiative corrections .
more specifically one has @xmath347 with @xmath348 reflecting the effect of the subleading logarithms @xcite and @xmath349 the function @xmath44 can be computed numerically for arbitrary arguments @xcite and analytically for the most interesting case @xmath350 @xcite .
the allowed values for @xmath344,which reflects the fact that qcd radiative corrections are quite sensitive to the final state quark masses , can be found in ref.@xcite . with @xmath351 one
obtains @xmath352 @xmath353 since these functions are normalized to unity for @xmath354 , one notes that the final - state quark masses reduce the available phase space quite considerably
. some qualitative statements can illuminate the dynamical situation : \(i ) as indicated in eqs.(15 ) , @xmath355 differ from the naive spectator result in order @xmath331 .
\(ii ) since @xmath356 and @xmath357 are @xmath117 singlet operators and their expectation values are practically isospin and even @xmath117 invariant , one obtains , as stated before , @xmath358 through order @xmath331 and likewise @xmath359 .
yet the meson and baryon lifetimes get differentiated on this level since @xmath360 and @xmath361 , see eqs.(20,30 ) .
\(iii ) the semileptonic branching ratios of @xmath145 and @xmath362 baryons remain unaffected in order @xmath331 due to @xmath363 , whereas for @xmath246 mesons it is ( slightly ) reduced @xcite ; i.e. , one finds @xmath364 beyond order @xmath331 , however , this relation is changed , as discussed later on .
differences in @xmath365 vs. @xmath55 are generated by the local dimension - six four - quark operators @xmath366 which explicitely depend on the light quark flavours .
such corrections are of order @xmath135 . based on this scaling law one
can already infer from the observed @xmath29 meson lifetime ratios that the various @xmath246 lifetimes will differ by no more than 10% or so .
assuming factorization for the expectation values of these four - fermion operators , i.e. applying eq.(22 ) , one obtains @xmath367 as discussed above .
it turns out that wa can change the @xmath246 lifetimes by no more than1% or so ; due to interference with the spectator reaction , they could conceivably _ prolong _ the @xmath124 lifetime relative to the @xmath65 lifetime , rather than reduce it .
the dominant effect is clearly provided by pi , which produces a negative contribution to the @xmath65 width : @xmath368 @xmath369 \ ; .\eqno(48b)\ ] ] eq.(48b ) exhibits an intriguing result : @xmath370 is positive for @xmath371 , i.e. pi acts _ constructively _ and thus would shorten the @xmath65 lifetime in the _ absence _ of radiative qcd corrections . including those as evaluated on the leading - log and next - to - leading - log level @xmath372 , @xmath373 turns pi into a _ destructive _ interference prolonging the @xmath65 lifetime , albeit by a tiny amount only at this point . in eq.(48b )
only ultraviolet renormalization has been incorporated .
hybrid renormalization @xcite down to the hadronic scale @xmath68 amplifies this effect considerably ; one obtains @xmath374 \ , , \ ] ] @xmath375 ^{1/b } , \ ; b=11-\frac{2}{3 } n_f \eqno(49)\ ] ] putting everything together one finds : @xmath376 i.e. the lifetime of a charged @xmath246 meson is predicted to definitely _ exceed _ that of a neutral @xmath246 meson by typically several percent .
three comments are in order for properly evaluating eq.(49 ) : @xmath46 although wa could conceivably prolong the @xmath124 lifetime as stated before , its numerical significance pales by comparison to that of pi .
pi acts destructive in @xmath65 decays once radiative qcd corrections are included to the best of our knowledge .
the lifetime of @xmath65 mesons therefore has to exceed that of @xmath124 mesons .
@xmath46 while the sign of the effect is predicted in an unequivocal manner , its magnitude is not .
the main uncertainty in the prediction on @xmath136 is given by our present ignorence concerning the size of @xmath292 .
@xmath46 even a precise measurement of the lifetime ratio @xmath136 would not automatically result in an exact determination of @xmath292 by applying eq.(50 ) . for corrections of order
@xmath377 have not ( or only partially see the discussion in the preceding subsection ) been included . unless those have been determined
, one can not extract the size of @xmath292 even from an ideal measurement with better than a roughly 15 % uncertainty .
very little @xmath117 breaking is anticipated between the @xmath124 and @xmath125 expectation values of the two dimension - five operators .
vs. @xmath122 lifetimes . ] among the contributions from dimension - six operators wa affects @xmath124 and @xmath125 lifetimes somewhat differently due to different colour factors for these two decays .
yet it is quite irrelevant in either case .
thus one predicts the @xmath124 and the average @xmath125 lifetimes to practically coincide : @xmath378 the term ` average @xmath125 lifetime ' is used for a reason : @xmath379 oscillations generate two neutral beauty mesons carrying strangeness that differ in their mass as well as in their lifetime . due basically to @xmath380
one finds @xmath381 . while @xmath382 can be calculated in terms of the expectation value of the _ local _ four - fermion operator @xmath383 ( since @xmath384 ) , the situation is more complex for @xmath385 , since the underlying operator is nonlocal .
one can however apply a heavy quark expansion ; to lowest nontrivial order one obtains @xcite @xmath386 for @xmath293 not much larger than 200 mev . comparing eqs.(50 - 52 )
leads to the intriguing observation that the largest lifetime difference among @xmath65 , @xmath124 and @xmath125 mesons is generated by a very subtle source : @xmath379 oscillations !
there are also two different lifetimes for neutral @xmath246 mesons without strangeness ; yet @xmath387 is suppressed by @xmath388 relative to @xmath385 .
one can search for the existence of two different @xmath125 lifetimes by comparing @xmath57 as measured in @xmath389 on one hand and in @xmath390 on the other .
the former decay predominantly leads to a cp even final state and thus would , to good accuracy , reveal @xmath391 ; the latter exhibits the average lifetime @xmath392/2 $ ] ; for semileptonic @xmath125 decays involve the cp even and odd components in a nearly equal mixture .
thus @xmath393\simeq\ ] ] @xmath394
whether an effect of this size is large enough to be ever observed in a real experiment , is doubtful .
nevertheless one should search for it even if one has sensitivity only for a 50% lifetime difference or so . for while eq.(52 ) represents the best presently available estimate , it is not a ` gold - plated ' prediction .
it is conceivable that the underlying computation _ underestimates _ the actual lifetime difference !
@xmath277 mesons with their two heavy constituents @xmath395
represent a highly intriguing system that merits special efforts to observe it .
one expects a rich spectroscopy probing the inter - quark potential at distances intermediate to those that determine quarkonia spectroscopy in the charm and in the beauty system .
the isgur - wise function for the striking channel @xmath396 can be computed .
what is most relevant for our discussion here is that its overall decays , and thus its lifetime , reflect the interplay of three classes of transitions , namely the decay of the @xmath274 quark , that of the @xmath397 ( anti)quark and wa of @xmath274 and @xmath397 : @xmath398 while @xmath399 and @xmath400 transitions do not interfere with each other in any practical way and one can thus cleanly separate their widths , the situation is much more delicate concerning @xmath401 and @xmath402 , as briefly explained later . yet for the moment i ignore the latter although its helicity and wavefunction suppressions represented by the factors @xmath403 and @xmath404 are relatively mild ( @xmath405 mev @xcite ) and partially offset by the numerical factor @xmath49 reflecting the enhancement of two - body phase space
relevant for wa over three - body phase space appropriate for the spectator decay . concerning the other two transitions one would naively expect the @xmath397 decay to dominate : @xmath406 .
it had been suggested that for a tightly bound system like @xmath277 the decay width should be expressed not in terms of the usual quark masses ( whatever they are ) , but instead in terms of effective quark masses reduced by something like a binding energy @xmath407 mev . _
if _ so , then the @xmath277 width would be reduced considerably since due to @xmath408 one would find a relative reduction @xmath409 .
even more significantly , beauty decays would become more abundant than charm decays in the @xmath277 transitions since the binding energy constitutes a higher fraction of @xmath410 than of @xmath317 : @xmath411 .
however this conjecture that might look quite plausible at first sight , turns out to be fallacious ! for it
is manifestly based on the existence of nonperturbative corrections of order @xmath70 ; yet as discussed in sect.4 those are absent in fully integrated decay rates like lifetimes due to a non - trivial cancellation between @xmath70 contributions from initial - state and final - state radiation ! the leading corrections arise in order @xmath104 and they enter through the first two terms on the right - hand - side of eq.(54 ) : @xmath412 @xmath413 the quantities @xmath414 and @xmath415 are defined as in eqs.(45 ) and ( 68 ) ; i.e. , the differences between @xmath416 and @xmath414 enter through the different expectation values of the same operators @xmath101 and @xmath417 , and likewise for @xmath418 ( to be discussed in detail in sect.7 ) vs. @xmath419 . for
the @xmath277 meson provides a different environment for quark decay than either @xmath246 or @xmath29 mesons .
the relevance of that difference is illustrated by eqs.(28,29 ) .
similarly one expects @xmath420 , @xmath421 to differ from @xmath286 , @xmath422 .
while there are large corrections of order @xmath423 in @xmath424 that reduce the corresponding semileptonic branching ratio considerably , they largely cancel against each other in the total width .
therefore @xmath425 writing the @xmath426 width of the @xmath277 meson as a simple incoherent sum @xmath427 actually represents an oversimplification .
for there arises considerable interference between higher order wa and spectator processes . yet for the purposes of our discussion here such effects can be ignored ; they will be discussed in detail in a forthcoming publication .
the @xmath331 contributions to @xmath428 are small , namely around a few percent .
the numerical impact of @xmath429 , which is formally of order @xmath135 , is nevertheless sizeable due to the large decay constant , the merely mild helicity suppression , given by @xmath403 , and the enhancement of wa by a factor @xmath430 due to its two - body kinematics .
one finds : @xmath431 comparing eqs.(56,57 ) one concludes that the naive expectation turns out to be basically correct , i.e. @xmath432 is well below @xmath433 sec and charm decays dominate over beauty decays with an ensuing reduction in the ` interesting ' branching ratios like @xmath434 or @xmath435 .
it is quite natural to assume that the kinetic energy of the @xmath274 quark is practically the same inside the @xmath145 and @xmath362 baryons : @xmath436 together with eqs.(15,17a ) this yields @xmath437 the leading _ differences _ among the @xmath145 and @xmath362 lifetimes then arise in order @xmath135 : they are generated by four - quark operators analogous to those that had already been identified in the phenomenological studies of charm baryons@xcite .
some complexities arise , though , due to the presence of the two transitions @xmath438 and @xmath320 ; one finds @xmath439 @xmath440 @xmath441 where has been ignored here for simplicity . ]
@xmath442 @xmath443 @xmath444 \eqno(60c)\ ] ] using the same notation as in eqs.(45 ) ; @xmath445 @xmath446 @xmath447 @xmath448 @xmath449 is obtained from eq.(60e ) by taking the expectation value between @xmath362 rather than @xmath145 states .
those expectation values vanish for the wrong charge - flavour combination ; i.e. , @xmath450 .
@xmath451 includes the naive spectator term : @xmath452 contributions ; the latter are practically identical for @xmath145 and @xmath362 , but differ for @xmath246 mesons .
@xmath453 denotes the contribution due to ws and @xmath454 the reduction due to destructive interference in the channels @xmath319 and @xmath320 , respectively .
is used in eq.(59c ) , one has , strictly speaking , to evaluate the expectation value for the state @xmath362 . ] on general grounds one thus obtains an inequality : @xmath455 naively one might expect @xmath456 to hold , since the @xmath320 part of the width which is reduced by pi in @xmath457 decays is smaller than the @xmath438 component which suffers from pi reduction in @xmath145 decays . on the other hand
since phase space is more restricted for @xmath320 than for @xmath438 , one would likewise expect the degree of ( destructive ) interference to be higher for the former than the latter ; it is then quite conceivable that actually @xmath458 holds .
the _ sign _ of the difference in the @xmath145 and @xmath457 lifetimes therefore provides us with valuable information on the strong dynamics .
there are two complementary ways to transform these qualitative predictions into quantitative ones .
\(i ) one evaluates the required expectation values explicitely within a quark model , as expressed in eq.(17b , 23 - 25 ) .
since the model also predicts the baryon masses in terms of its parameters , one can cross check it with the observed spectroscopy .
this will become increasingly relevant in the future , yet at present provides little guidance .
the expressions given in eqs.(23 - 25 ) have to be augmented by the radiative qcd corrections : @xmath459 @xmath460 with @xmath461 as defined in eq.(49 ) .
it should be noted that here unlike in the case of meson decays the sign of the pi contribution is quite robust under radiative corrections : it is proportional to @xmath462 which is negative already for @xmath371 .
furthermore @xmath453 is proportional to @xmath463 and thus colour - enhanced , since the baryon wavefunction is purely antisymmetric in colour space .
this prescription yields lifetime differences of not more than a few percent in eq.(59 ) and @xmath464 ( ii ) as will be discussed in the next subsection , the pattern predicted for charm baryon lifetimes agrees with the observations within the experimental and theoretical uncertainties . taking this as prima facie evidence that the heavy quark expansion through order @xmath88 applies at least in a semi - quantitative fashion already to charm lifetime ratios
, one can extrapolate to the weight of these pre - asymptotic corrections in beauty decays using scaling like @xmath104 and @xmath88 ( or @xmath465 ) .
that way one again finds that the differences in eq.(56 ) amount to not more than a few to several percent @xcite : @xmath466 similarly one estimates @xmath467 through a simple @xmath88 scaling behaviour from the charm baryon lifetimes : @xmath468 @xmath469 obviously and not surprisingly there is some numerical fuzziness in these predictions ; yet they seem to be unequivocal in stating that the @xmath124 lifetime exceeds the @xmath145 lifetime and the average beauty baryon lifetime by about 10 percent .
however this prediction appears to be in serious ( though not yet quite conclusive ) disagreement with the data .
if the predictions were based exclusively on adopting the quark model results for the baryonic expectation values , one could abandon them in a relatively light - hearted way : for it should not come as a shocking surprise that quark model results for baryonic matrix elements can be off - target by a substantial amount .
yet we have encountered a more serious problem here : data seem to contradict also the prediction based on an extrapolation from the observed lifetime pattern in the charm family ; furthermore as discussed next the observed lifetime ratios of charm hadrons can be reproduced , within the expected uncertainties .
this allows only one conclusion : if @xmath55 indeed exceeds @xmath470 by 25 - 30 % , then a ` theoretical price ' has to be paid , namely that @xmath46 the charm mass represents too low of a scale for allowing to go beyond merely qualitative predictions on charm baryon ( or even meson ) lifetimes , since it appears that corrections of order @xmath471 and higher are still important ; @xmath46 that the present agreement between theoretical expectations and data on charm baryon lifetimes is largely accidental and most likely would not survive in the face of more precise measurements ! at the same time
an intriguing puzzle arises : @xmath46 why are the quark model results for the relevant expectation values so much off the mark for beauty baryons ? it is the deviation from unity in the lifetime ratios that is controlled by these matrix elements ; finding a 30 % difference rather than the expected 10 % then represents a 300 % error !
some new features emerge in @xmath220 decays : since the @xmath199-diquark system forms a spin - triplet , there are @xmath331 contributions to the semileptonic and nonleptonic @xmath220 widths from the chromomagnetic operator . in order @xmath135
there arises a destructive interference in the @xmath320 channels .
however a detailed discussion of @xmath220 decays seems academic at the moment . as briefly discussed before
, one expects the semileptonic branching ratios for @xmath124 and @xmath125 mesons to practically coincide , in particular since semileptonic decays probe @xmath472 , the average @xmath125 lifetime .
it has already been pointed out that through order @xmath331 the expected value for @xmath473 exceeds that for @xmath474 by a few percent . in order
@xmath135 the nonleptonic widths of both @xmath124 and @xmath145 states receive new contributions , with @xmath475 getting further enhanced relative to @xmath476 as expressed in eq.(63b ) .
thus one predicts @xmath477 with the inequalities indicating differences of a few to several percent .
if the present trend in the data persists , i.e. if the total @xmath145 width exceeds @xmath478 by , say , 20 -25 % , one would interprete this discrepancy as meaning that the nonleptonic but not the semileptonic
width has received a unforeseen enhancement . in that case one
expects @xmath473 to fall below @xmath479 by @xmath480 20 % .
putting all these observations together one concludes that the beauty lifetime averaged over @xmath124 , @xmath65 , @xmath125 and @xmath145 states should yield a value that is _ smaller _ by a few percent than @xmath481 \equiv { \langle br_{sl}(b)\rangle}$ ] : @xmath482
considering the wealth of rather precise experimental information available on the lifetimes of charm hadrons one feels the urge to apply heavy quark expansions to charm decays as well . yet in doing so one has to keep in mind that such a treatment might fail here for two basic reasons , one of which has just been stated : \(i ) the charm quark mass does not provide a sufficiently large scale to make the @xmath483 expansion converge quickly . to obtain an estimate of the size of the expansion parameter @xmath484 one can take the square root of the expression in eq.(35 ) representing the @xmath423 corrections : @xmath485 this is not a small number although it is at least smaller than unity .
\(ii ) as discussed in sect.4 one has to go beyond _ global _ duality and invoke _ local _ duality to predict the decay widths of real hadrons from heavy quark expansions properly defined in euclidean space . yet
a priori it is not clear at all whether contributions from ` distant cuts ' can be ignored since the charm quark mass does not exceed ordinary hadronic scales by a large amount .
this concern is a posteriori strengthened by the following observation : equating the observed semileptonic width of @xmath29 mesons with its theoretical expression through order @xmath423 ( and assuming @xmath486 ) leads to the requirement @xmath487 gev .
this is however a high value relative to what is derived from charmonium spectroscopy , namely @xmath488 gev .
a difference of @xmath489 gev in @xmath410 might appear quite innocuous till one realizes that the corresponding semileptonic width depending on @xmath490 changes by a factor of two when @xmath410 is shifted by those @xmath489 gev ! at this point one might suspect that corrections of higher order in @xmath483 contribute _ constructively _ boosting the theoretical value .
the analysis of ref.@xcite finds however that terms of order @xmath491 show a tendency to _ decrease _ @xmath492 , and their inclusion does not help at all in reproducing the measured value of @xmath492 with @xmath493 gev .
this suggests that contributions from ` distant cuts ' which can not be seen in any finite order of the @xmath483 expansion create this problem . for a different opinion on this point ,
see ref.@xcite .
i draw the following somewhat tentative conclusions from these observations : @xmath46 predictions for the absolute size of charm hadron lifetimes can not be trusted .
@xmath46 however it is quite conceivable that lifetime
_ ratios _ do not suffer from such a fundamental uncertainty due to not - so - distant cuts .
i will then explore the working hypothesis that the ratio of lifetimes can be trusted in principle though in practise only with a quite considerable grain of salt !
the quark level lagrangian is again well known . on the cabibbo - allowed level
there is now a single non - leptonic transition described by @xmath494 where the short - distance coefficients @xmath495 are now evaluated at a lowerjscale than in beauty decays yielding @xmath496 .
the expressions for the semileptonic and non - leptonic decay widths of charm hadrons @xmath497 are quite analogous to the ones for beauty hadrons ( though simpler since there is only one non - leptonic decay class rather than two ) ; through order @xmath423 they are given by @xmath498 \ ; , \eqno(68a)\ ] ] @xmath499 - \right.\ ] ] @xmath500\ ; . \eqno(68b)\ ] ] where now @xmath501 and the radiative corrections lumped into @xmath345 and @xmath346 are given by the appropriate values for @xmath502 and @xmath503 , see eq .
( 67 ) . with @xmath504 one
finds : @xmath505 i.e. there is much less phase space suppression than for @xmath399 transitions .
analogously to the case of @xmath365 vs. @xmath55 the @xmath18 and @xmath17 lifetimes get first differentiated in order @xmath491 when pi and wa intervene . both affect the lifetime ratio in the same direction , namely they enhance @xmath20 over @xmath19 with the _ destructive interference _ due to pi being the _ dominant _ effect .
quantitatively one finds through order @xmath491 by employing eqs.(48,49 ) with the appropriate substitutions and using @xmath506 mev : @xmath507 for proper perspective one has to keep four observations in mind : \(i ) while the expectation expressed in eq.(69 ) does not coincide numerically with the measured value
@xmath508 it agrees with it to within @xmath509 % .
such a deviation could be ascribed to @xmath471 contributions ignored in eq.(69 ) .
\(ii ) on the other hand , as discussed before , there is reason to doubt the reliability and thus validity of heavy quark expansions for treating nonperturbative corrections in _ charm _ decays . yet
i adopt , as already stated , the working hypothesis that a heavy quark expansion can be employed for treating the _ ratios _ of lifetimes , though not the lifetimes themselves .
such a conjecture is tested a posteriori by analysing the whole pattern of the expected and the observed lifetimes of the various charm hadrons .
\(iii ) even the perturbative corrections contain sizeable numerical uncertainties . to cite but
one glaring example : the size and the nature of the socalled ` hybrid ' renormalization reflecting dynamics between the scales @xmath410 and @xmath510 gev is quite important quantitatively ( they considerably enhance the destructive interference in @xmath18 decays . ) yet a leading - log treatment of those corrections seems woefully inadequate for dealing with such a small slice in momentum space .
\(iv ) one should note also the following : while wa plays only a relatively minor role in inclusive rates ( generating only a 10 % or so difference in @xmath18 vs. @xmath17 lifetimes ) , it is likely to play a considerably more significant role in _ exclusive _ modes . since the impact of wa is reduced relative to that of pi in meson
decays , it is natural to compare @xmath56 to @xmath19 rather than to @xmath20 .
such a comparison and of their semileptonic branching ratios touches upon several intriguing dynamical issues .
a priori @xmath56 and @xmath19 could differ substantially from each other due to @xmath117 breaking and in particular due to a different weight of wa in the two cases . yet
the heavy quark expansion strongly suggests those two lifetimes to be close to each other .
the analysis proceeds in several steps @xcite .
the two operators contributing in order @xmath423 are singlets under the light quark flavours ; yet @xmath117 breaking could enter through their expectation values . for the chromomagnetic operator one has : @xmath512 @xmath513 since the measured values for the @xmath514 and the @xmath515 mass splittings are practically identical ( @xmath516 mev and @xmath517 mev ) , the chromomagnetic operator can not induce an appreciable difference between @xmath19 and @xmath56 . through order @xmath491
there are four distinct sources for a difference in @xmath518 vs. @xmath519 exceeding the 1% level : ( a ) the decay @xmath520 ; ( b ) pi in those cabibbo suppressed @xmath122 decays that are driven by the quark level transition @xmath521 ; ( c ) the effects of @xmath117 breaking on the expectation values of the kinetic energy operator ; ( d ) wa in nonleptonic @xmath17 and in nonleptonic as well as semileptonic @xmath122 decays .
corrections listed under ( a ) , ( b ) and ( d ) are generated by dim-6 operators whereas the much less familiar correction ( c ) is derived from a dim-5 operator .
\(a ) the width for the decay @xmath520 is completely determined in terms of the axial decay constant for the @xmath122 meson : @xmath522 for @xmath523 mev one gets @xmath524 this effect necessarily _ reduces _
@xmath56 relative to @xmath19 .
\(b ) in @xmath122 decays pi appears in the cabibbo - suppressed @xmath525 channel .
its weight is expressed in terms of the matrix elements of the two four - fermion operators @xmath526 with known coefficients that are computed perturbatively ( including the ` hybrid ' renormalization of these operators down from the scale @xmath410 ) .
the effect of pi is most reliably estimated from the observed difference in the @xmath18 and @xmath17 widths .
it is easy to see that the structure of the operators responsible for pi in @xmath122 decays is exactly the same as in @xmath18 decays if one replaces the @xmath527 quark with the @xmath528 quark and adds the extra factor @xmath529 ; it is then _ destructive _ as well .
thus one arrives at : @xmath530 where the factor @xmath531 has been introduced to allow for @xmath117 breaking in the expectation values of the four - fermion operators .
the factor @xmath531 is expected to exceed unity somewhat ; in an factorization ansatz it is given by the ratio @xmath532 , see eq.(39 ) .
thus @xmath533 \(c ) the impact of the _ chromomagnetic _ operator on the @xmath534 lifetime ratio can be derived from the hyperfine splitting .
inserting the observed meson masses one obtains @xmath535 @xmath536 from which one can conclude : there is not surprisingly very little @xmath117 breaking in these matrix elements and a sizeable difference in @xmath56 vs. @xmath19 _ can not _ arise from this source .
the second dim-5 operator generating @xmath423 corrections is the kinetic energy operator @xmath537 where @xmath189 denotes the covariant derivative .
its expectation value describes the largely non - relativistic motion of the charm quark in the gluon background field inside the charm hadron .
it reflects lorentz time dilation and thus prolongs the _ hadron _ lifetime . on general grounds
one expects it to extend @xmath56 over @xmath19 , as seen as follows .
the spatial wavefunction should be more concentrated around the origin for @xmath122 than for @xmath17 mesons .
this implies , via the uncertainty principle , the mean value of @xmath538 to be larger for @xmath122 than for @xmath17 ; in other words the charm quark undergoes more fermi motion as constituent of @xmath122 than of @xmath29 mesons .
the lifetime of the charm quark is then prolonged by time dilation to a higher degree inside @xmath122 than inside @xmath17 mesons . while the trend of this effect is quite transparent , its size is not yet clear .
the relevant matrix elements @xmath539 and @xmath540 will be determined from qcd sum rules and lattice simulations in the future . yet in the meantime one can estimate their size from the meson masses in the charm and beauty family according to the following prescription : @xmath541- [ { \langle m_{b_s}\rangle}-{\langle m_b\rangle}]\ } \eqno(78)\ ] ] where as before @xmath542 denote the spin averaged masses .
accordingly one finds @xmath543 @xmath544- [ { \langle m_{b_s}\rangle}-{\langle m_b\rangle}]\ } \ , .\eqno(79)\ ] ] since the @xmath545 mass has not been measured yet , one can not give a specific numerical prediction and has to content oneself with semi - quantitative statements .
a 10 mev shift in any of the terms @xmath546 corresponds numerically to the kinetic energy operator generating approximately a 1% change in the ratio @xmath547 . invoking our present understanding of the heavy quark kinetic energy and its relationship to the hyperfine splitting one arrives at the following _ conjecture _ : @xmath548 \(d ) the mechanisms listed above in ( a ) - ( c ) taken together can be expected to extend the @xmath122 over the @xmath17 lifetime by at most a few percent .
comparing that estimate with the data allows us to conclude that wa can not change @xmath149 by more than 10 % .
it is however quite unclear how to refine this estimate at present .
for the quantitative impact of wa on charm meson lifetimes in general and on the ratio @xmath149 in particular is the most obscure theoretical element in the analysis .
the uncertainty centers mainly on the question of how much the wa amplitude suffers from helicity suppression . in the valence quark
approximation the answer is easily given to _
lowest _ order in the strong coupling : the wa amplitude is ( helicity ) suppressed by the ratio @xmath549 , where @xmath22 denotes the largest quark mass in the final state . for a proper qcd treatment
one has to use the _ current _ rather than the larger _ constituent _ mass , at least for the @xmath491 corrections .
the wa amplitude is then small in @xmath17 decays where the helicity factor reads @xmath550 and a fortiori in @xmath122 decays where only non - strange quarks are present in the final state .
the emission of semi - hard gluons that is included by summing the leading log terms in the perturbative expansion can not circumvent this suppression @xcite .
for such gluon corrections
when properly accounted for drive the hybrid renormalization of the corresponding four - fermion operators ; however they preserve the lorentz structure of the lowest order term and therefore do not eliminate the helicity suppression .
a helicity allowed amplitude arises in perturbation theory only at the subleading level of order @xmath551 and is thus expected to be numerically insignificant . on the other hand _ nonperturbative _ dynamics can quite naturally vitiate helicity suppression and thus provide the dominant source of wa .
these nonperturbative effects enter through _ non - factorizable _ contributions to the hadronic matrix elements , as analyzed in considerable detail in refs.@xcite . as such
we do not know ( yet ) , how to predict their weight from first principles .
however , as shown in ref.@xcite , a detailed experimental study of the width of semileptonic decays and their lepton spectra in particular in the endpoint region in @xmath17 vs. @xmath122 and/or in @xmath552 vs. @xmath553 decays would allow us to extract the size of the matrix elements that control the weight of wa in all inclusive @xmath246 and @xmath29 decays . before such data become available ,
we can draw only qualitative , or at best semi - quantitative conclusions : wa is not expected to affect the total lifetimes of @xmath122 and @xmath17 mesons by more than 10 - 20 % , and their ratio by less .
furthermore wa does not necessarily decrease @xmath149 ; due to its reduced amplitude its leading impact on the lifetime might be due to its interference with the spectator amplitude and thus it might even _ enhance _ @xmath149 ! to summarize our findings on the @xmath534 lifetime ratio : ( i ) @xmath117 breaking in the expectation values of the dim-5 operators generating the leading nonperturbative corrections of order @xmath423 can due to ` time dilatation ' increase @xmath56 by 3 - 5 % over @xmath19 .
( ii ) on the @xmath491 level there arise three additional effects .
destructive interference in cabibbo suppressed @xmath122 decays increases @xmath56 again by 3 - 5 % , whereas the single mode @xmath520 decreases it by 3 % .
the three phenomena listed so far combine to yield @xmath554 .
( iii ) any difference over and above that has to be attributed to wa .
therefore one has to interprete the measured @xmath534 lifetime ratio as more or less direct evidence for wa to contribute no more than 10 - 20 % of the lifetime ratio between charm mesons .
this is consistent with the indirect evidence discussed above that wa does not constitute the major effect there .
( iv ) these predictions can be refined in the future by two classes of more accurate data : analyzing the _ difference _ in the semileptonic spectra of charged and neutral mesons in the charm and beauty family allows to extract the size of the matrix elements that control the weight of wa ; measuring the masses of @xmath145 and @xmath545 to better than 10 mev allows to determine the expectation values of the kinetic energy operator .
the lifetimes of the weakly decaying charm baryons reflect a complex interplay of different dynamical mechanisms increasing or decreasing transition rates @xcite : latexmath:[\[\gamma ( \lambda _ c^+)= \gamma _ { decay}(\lambda _ c^+ ) + \gamma _ { ws}(\lambda _ c^+ ) - latexmath:[\[\gamma ( \xi _ c^0)= \gamma _ { decay}(\xi _ c^0 ) + \gamma _ { ws}(\xi _ c^0 ) + latexmath:[\[\gamma ( \xi _ c^+)= \gamma _ { decay}(\xi _ c^+ ) + expressions for @xmath183 , @xmath453 and @xmath454 are obtained from eqs.(60 ) by the obvious substitutions ; for @xmath556 one finds @xmath557 on rather general grounds one then concludes : @xmath558 to go further requires calculating the relative weight of matrix elements of the four - quark operators . at present
this can be done only by using quark wavefunctions as obtained from a potential ansatz ( or from the quark model ) . in doing so
one also has to pay proper attention to the normalization point appropriate for such an evaluation , i.e. , one has to include ultraviolet as well as hybrid renormalization .
one then arrives at the numbers quoted in table 2 .
it has been argued @xcite that in these expressions one should not use the ` real ' value for the decay constant , as expressed by @xmath36 , and the observed mass splitting , but solely the leading contribution in a @xmath88 expansion : @xmath559 mev and @xmath560 mev .
one reason for that is self - consistency since the widths have been calculated through order @xmath88 only ; the other and maybe the more telling one is based on the needs of phenomenology : for otherwise one can not reproduce the observed _ magnitude _ of the lifetime differences .
the predictions given in table 2 agree remarkably well with the data within the experimental and theoretical uncertainties . in the comparison of the @xmath47 and @xmath29 lifetimes
one should note the following : ws and pi _ counteract _ each other in @xmath561 , see eq.(81a ) ; nevertheless @xmath562 can be obtained .
this results from three effects : ( i ) wa is very much reduced in @xmath29 decays .
( ii ) the _ baryonic _ expectation values of the four - quark operators are evaluated , as stated above , with the ` static ' decay constant , which is considerably larger than @xmath36 used for the _ mesonic _ expectation values .
( iii ) the @xmath423 contributions enhance @xmath561 over @xmath563 , as briefly discussed below .
unfortunately there appears now a rather unpleasant ` fly in the ointment ' : as discussed above , extrapolating from these prima facie successful predictions to the lifetimes of beauty baryons leads to a less - than - successful prediction on @xmath142 .
the decays of @xmath51 baryons require and deserve a separate treatment : the @xmath199 diquarks carry spin one and the resulting spin - spin interactions of @xmath212 with @xmath199 lead to new effects . to be more specific : through order @xmath491 one
has latexmath:[\[\gamma ( \omega _
c ) = \gamma _ { decay}(\omega _ c ) + the right - hand - side of eq.(84 ) differing from the corresponding ones for @xmath47 or @xmath221 decays .
( i ) firstly , @xmath565 , see eqs.(17b,17c,35 ) .
secondly , it is quite conceivable that these spin - spin interactions can create an appreciable difference in the kinetic energy of the @xmath212 quark inside @xmath51 and @xmath47 baryons . thus there
arise differences of order @xmath423 in the total as well as in the semileptonic widths of @xmath51 and @xmath47 baryons : @xmath566 ( ii ) these spin - spin interactions also affect the expectation values of the dimension - six four - fermion operators that control the strength of ws and the interference effects , see eqs.(24,25 ) , in order @xmath491 . taking everything together one estimates the @xmath51 to be the shortest lived charm baryon , as shown in table 2 .
this is quite remarkable since it means that the constructive interference in @xmath51 decays outweighs the _ combined _ effect of constructive interference and ws in @xmath48 decays .
it is actually not unexpected on intuitive grounds : in @xmath51 transitions interference can happen with _ both _
constituent @xmath528 quarks whereas in @xmath221 baryons there is only one such @xmath528 quark .
this shows that inclusive weak decay rates can be harnessed to probe the internal structure of charm baryons in a novel way .
as expected , the ratio of the semileptonic @xmath18 and @xmath17 branching ratios agrees , within the errors , with the ratio of their lifetimes : @xmath567 also the _ absolute _ size of , say , @xmath568 is of interest .
if the @xmath569 lifetime difference is generated mainly by a destructive interference affecting @xmath18 transitions , then @xmath17 decays should proceed in a largely normal way and the semileptonic branching ratio of @xmath17 mesons should be at its ` normal ' value . in a parton model description one
finds @xmath570 .
if that number represented the proper yardstick for normal decays , one would actually conclude that @xmath18 decays more or less normally , whereas the nonleptonic decays of @xmath17 are strongly enhanced ;
i.e. , it would imply that it is actually wa that provides the dominant mechanism for the lifetime difference .
this long - standing ` fly in the ointment ' can be removed now . for at order @xmath104 the chromomagnetic operator @xmath175 generates an isoscalar enhancement in the nonleptonic widths of the charged and neutral mesons @xcite .
it turns out that the corresponding reduction in the semileptonic branching ratios of @xmath246 mesons is rather insignificant , namely @xmath480 2 % . yet for charm mesons this reduction is quite large , namely around 40 % , since it is scaled up by @xmath571 and less reduced by colour factors .
the semileptonic branching ratio for @xmath29 mesons is thus around 9 % through order @xmath423 before the dimension - six four - fermion operators reduce @xmath572 by a factor of almost two and enhance @xmath573 by a moderate amount . as discussed above , we have to take these numbers for charm transitions with a grain of salt ; on the other hand the emerging pattern in the ratio of semileptonic branching ratios as well as their absolute size is self - consistent !
it should be apparent from the discussion on @xmath56 vs. @xmath19 that the semileptonic widths and branching ratios of @xmath122 and @xmath17 mesons practically coincide through order @xmath423 . in order @xmath491
wa can contribute to @xmath118 ( as well as to @xmath574 and @xmath573 ) . yet from @xmath575 one
infers that the relative weight of wa is quite reduced ; therefore one expects @xmath576 to differ from @xmath568 by not more than 10% or so in _ either _ direction : @xmath577 the semileptonic width is not universal for all charm hadrons due to @xmath578 contributions even when one ignores cabibbo suppressed channels .
one actually estimates @xmath579 therefore the ratio of semileptonic branching ratios of mesons and baryons does not faithfully reproduce the ratio of their lifetimes .
instead one expects @xmath580
second generation theoretical technologies have been developed for treating heavy - flavour decays that are directly related to qcd without the need for invoking a ` deus ex machina ' .
as far as _ inclusive _ heavy - flavour decays are concerned , the relevant technology is based on an operator expansion in inverse powers of the heavy quark mass .
it allows to express the leading nonperturbative corrections through the expectation values of a small number of dimension - five and -six operators .
basically all such matrix elements relevant for _ meson _ decays can reliably be related to other observables ; this allows to extract their size in a model - independant way .
baryons , however , possess a more complex internal structure , which becomes relevant for their decays in order @xmath88 . at present
one has to rely on quark model calculations to determine the expectation values of the dimension - six operators relevant for lifetime differences .
the numerical results of such computations are of dubious reliability ; accordingly predictions for lifetime ratios involving heavy - flavour _ baryons _ suffer from larger uncertainties than those involving only mesons .
there is also a positive side to this , though : measuring the lifetimes ( and if at all possible also the semileptonic branching ratios ) of the various heavy - flavour baryons sheds new light onto their internal strong dynamics .
comprehensive studies of inclusive weak decays in the charm baryon family are particularly promising from a practical point for such studies on hadronic structure ; for the pre - asymptotic corrections are much larger here than for beauty baryons . at present there
exists one rather glaring phenomenological problem for the heavy quark expansion and not surprisingly , as just indicated it concerns baryon decays : the observed @xmath145 lifetime is clearly shorter than predicted relative to the @xmath124 lifetime .
unless future measurements move it up significantly , one has to pay a theoretical price for that failure . as stated above , for the time being one has to employ a quark model to determine the size of the baryonic expectation values of the relevelant dimenion - six operators .
it is not surprising per se that quark model computations can yield numerically incorrect results , in particular when sizeable cancellations occur between different contributions .
the more serious problem is provided by the following aspects : the analoguous treatment of charm baryons , using the same quark models , had yielded predictions that give not only the correct pattern , but also numerical values for the lifetime ratios which are surprisingly close to the data considering the uncertainties due to higher order terms in the @xmath483 expansion . furthermore scaling these charm lifetime ratios up to the beauty scale , using a @xmath104 and @xmath88 behaviour for the pre - asymptotic corrections from the dimension - five and -six operators , respectively ,
also tells us that the @xmath145 lifetime should be shorter than the @xmath124 lifetime by no more than 10 % . to the degree that the observed value for @xmath142 falls below
0.9 one has to draw the following conclusion : one can not trust the numerical results of quark model calculations for _ baryonic _ matrix elements not very surprising by itself ; yet furthermore and more seriously it would mean that @xmath581 or even higher order contributions are still relevant in charm baryon decays before fading away for beauty decays .
then one had to view the apparently successful predictions on the lifetimes of charm baryons as largely coincidental a quite sobering result !
a _ future _ discrepancy between the predictions on @xmath582 or @xmath583 and the data refers to the algebraic average of the two @xmath125 lifetimes . ] in particular an observation that the lifetime for @xmath553 mesons is definitely _ shorter _ than for @xmath124 mesons would have quite fundamental consequences . for the leading deviation of these ratios from unity
arises at order @xmath135 and should provide a good approximation since the expansion parameter is small : @xmath584 .
the size of this term is given by the expectation value of a four - fermion operator expressed in terms of @xmath292 .
a failure in this simple situation would raise very serious doubts about the validity or at least the practical relevance of the @xmath70 expansion for treating even fully inclusive nonleptonic transitions ; this would leave only semileptonic transitions in the domain of their applicability .
such a breakdown of quark - hadron duality would _ a priori _ appear as a quite conceivable and merely disappointing scenario .
however such an outcome would have to be seen as quite puzzling _ a posteriori _ ; for in our analysis we have not discerned any sign indicating the existence of such a fundamental problem or a qualitative distinction between nonleptonic and semileptonic decays @xcite .
thus even a failure would teach us a valuable , albeit sad lesson about the intricacies of the strong interactions ; for the heavy quark expansion is directly and unequivocally based on qcd with the only additional assumption concerning the workings of quark - hadron duality !
@xmath46 _ improved _ by a better understanding of quark - hadron duality , preferably by deriving it from qcd or at least by analysing how it operates in the lepton and photon spectra of semileptonic and radiative @xmath246 decays , respectively ; the extension will certainly have been made very soon ; the refinement should be obtained in the foreseeable future , probably from a lattice simulation of qcd ; when finally quark - hadron duality will be derived from qcd can not be predicted , of course . *
acknowledgements : * i am deeply indebted to my collaborators b. blok , m. shifman , n. uraltsev and a. vainshtein for sharing their insights with me . prodding from many experimental colleagues , in particular s. paul and v. sharma , has been instrumental in making me face reality ( i think ) .
i have been fortified by many excellent chinese dinners prepared late at night by bai - ju s in south bend .
r. r " uckl , in : proc .
international school of physics `` enrico fermi '' , course xcii , varenna , 1984 , ed . by n. cabibbo ,
north - holland , amsterdam , 1987 , p. 43 - 75 ; z physics at lep 1 , ed . by g. altarelli , r. kleiss , c. verzegnassi , cern yellow report , cern 89 - 08 , geneva , 1989 , vol . 1 ,
p. 311 .
m. voloshin , preprint tpi - minn-94/18-t ; this analysis improved upon an inequality previously derived in : i.i .
bigi , m. shifman , n.g .
uraltsev , a. vainshtein , _ int .
a _ * 9 * ( 1994 ) 2467 . for a discussion of these issues
see : m. shifman , talk given at the v international symposium on particles , strings and cosmology pascos , john hopkins university , baltimore , march 1995 , to appear in the proceedings , preprint tpi - minn-95/15-t [ hep - ph/9505289 ] .
m. voloshin , preprint tpi - minn-94/38-t ; this property was first recognized within a phenomenological parton model description by r. r " uckl , in : proc .
xxiith international conference on high energy physics , leipzig , 1984 , ed . by a.
meyer , e. wieczorek , akad .
der ddr , zeuthen , 1984 , vol .
ii , p. 135 .
bigi , b. blok , m. shifman , n.g .
uraltsev , a. vainshtein , in : _ the fermilab meeting _ ,
proceed . of the 1992 aps - dpf meeting , c.h .
albright et al .
( eds . ) , world scientific , singapore , 1993 , vol.1 , pg . 610 . . | over the last few years a theoretical treatment for the weak decays of heavy - flavour hadrons has been developed that is genuinely based on qcd .
its methodology is described as it applies to total lifetimes , and the underlying theoretical issues are discussed .
theoretical expectations are compared with present data .
one discrepancy emerges : the beauty baryon lifetime appears to be significantly shorter than predicted .
the ramifications of those findings are analyzed in detail , and future refinements are described . 8.5 in -1.0 in -42pt plus 2 mm minus 2 mm + und - hep-95-big02 + june 1995 + version 3.0 + the qcd perspective on lifetimes of heavy - flavour hadrons + i.i .
bigi + physics dept . ,
university of notre dame du lac + notre dame , in 46556 , u.s.a .
+ e - mail address : [email protected] |
it is well known that the `` gold - plated '' mode @xmath1 @xcite plays an outstanding role in the determination of @xmath9 , where @xmath4 is one of the three angles @xmath10 , @xmath4 and @xmath3 of the usual non - squashed unitarity triangle @xcite of the cabibbo
maskawa matrix ( ckm matrix ) @xcite .
first attempts to measure @xmath9 in this way , which is one of the major goals of several @xmath11-physics experiments starting very soon , have recently been performed by the opal and cdf collaborations @xcite . in this paper , we will have a closer look at the general structure of the @xmath1 decay amplitude arising within the standard model , and at the one of its @xmath2-spin counterpart @xmath0 .
the two decays are related to each other by interchanging all down and strange quarks , i.e. through the `` @xmath2-spin '' subgroup of the @xmath12 flavour symmetry of strong interactions . whereas the weak phase factor @xmath13 enters in @xmath1 in a strongly cabibbo - suppressed way ,
this is not the case in @xmath0 .
consequently , there may be sizeable cp - violating effects in this @xmath14 decay , which are due to the interference between current current and penguin operator contributions .
interestingly , the time evolution of the @xmath0 decay rate allows us to determine @xmath3 . to this end
, we have to employ the @xmath2-spin symmetry to fix the overall normalization of @xmath0 through the cp - averaged @xmath1 rate . this new strategy to extract @xmath3
is not affected by qcd or electroweak penguin effects it rather makes use of these topologies and does not rely on certain `` plausible '' dynamical or model - dependent assumptions .
moreover , final - state - interaction effects are taken into account _ by definition _ , and do not lead to any problems .
the theoretical accuracy is only limited by @xmath2-spin - breaking corrections .
an analogous strategy is provided by the time - dependent @xmath5 rate , if its overall normalization is fixed through the cp - averaged @xmath6 rate , and if the @xmath15@xmath16 mixing phase , i.e. @xmath17 , is determined with the help of @xmath18 .
in particular the determination of @xmath3 is an important goal for future @xmath11-physics experiments .
this angle should be measured in a variety of ways so as to check whether one consistently finds the same result .
there are several methods to accomplish this task on the market @xcite .
since the @xmath19 @xmath11-factories operating at the @xmath20 resonance will not be in a position to explore @xmath14 decays , a strong emphasis has been given to decays of non - strange @xmath11 mesons in the recent literature .
however , also the @xmath14 system provides interesting strategies to determine @xmath3 . in order to make use of these methods , dedicated @xmath11-physics experiments at hadron machines , such as lhcb , are the natural place . within the standard model , the weak @xmath21@xmath22
mixing phase is very small , and studies of @xmath14 decays involve very rapid @xmath21@xmath22 oscillations due to the large mass difference @xmath23 between the mass eigenstates @xmath24 ( `` heavy '' ) and @xmath25 ( `` light '' ) .
future @xmath11-physics experiments performed at hadron machines should be in a position to resolve these oscillations .
interestingly , in contrast to the @xmath26 case , there may be a sizeable width difference @xmath27 between the mass eigenstates of the @xmath14 system @xcite , which may allow studies of cp violation with `` untagged '' @xmath14 data samples , where one does not distinguish between initially , i.e. at time @xmath28 , present @xmath29 or @xmath22 mesons @xcite .
in such untagged rates , the rapid @xmath21@xmath22 oscillations cancel .
some of the @xmath14 strategies proposed in the literature are theoretically clean , and use pure `` tree '' decays , for example @xmath30 @xcite . since no flavour - changing neutral - current ( fcnc ) processes contribute to the corresponding decay amplitudes , it is quite unlikely that they are significantly affected by new physics .
consequently , the preferred mechanism for physics beyond the standard model to manifest itself in the corresponding time - dependent decay rates is through contributions to @xmath21@xmath22 mixing .
in contrast , the decay @xmath0 discussed in this paper exhibits also cp - violating effects that are due to the interference between `` tree '' and `` penguin '' , i.e. fcnc , processes .
therefore , new physics may well show up in the corresponding cp asymmetries , thereby affecting the extracted value of @xmath3 .
a similar comment applies to the @xmath8 strategy .
the outline of this paper is as follows : in section [ dao ] , the @xmath31 decay amplitudes are parametrized in a completely general way within the framework of the standard model .
moreover , expressions for the observables of the corresponding time - dependent decay rates are given .
the strategy to determine @xmath3 with the help of these observables is discussed in section [ gam - det ] , whereas we turn to the analogous strategy using @xmath8 decays in section [ bddgam - det ] .
the main results are summarized in section [ sum ] .
= 5.5truecm = 5.5truecm
the decays @xmath32 are transitions into a cp eigenstate with eigenvalue @xmath33 and originate from @xmath34 quark - level decays .
we have to deal both with current current and with penguin contributions , as can be seen in fig .
[ fig : bdpsiks ] .
let us turn to the mode @xmath35 first .
its transition amplitude can be written as @xmath36 where @xmath37 denotes the current current contributions , i.e. the `` tree '' processes in fig .
[ fig : bdpsiks ] , and the amplitudes @xmath38 describe the contributions from penguin topologies with internal @xmath39 quarks ( @xmath40 .
these penguin amplitudes take into account both qcd and electroweak penguin contributions .
the primes in ( [ bd - ampl1 ] ) remind us that we are dealing with a @xmath41 transition , and @xmath42 are the usual ckm factors . making use of the unitarity of the ckm matrix and applying the wolfenstein parametrization @xcite , generalized to include non - leading terms in @xmath43 @xcite , we obtain @xmath44,\ ] ] where @xmath45 with @xmath46 , and @xmath47 the quantity @xmath48 is defined in analogy to @xmath49 , and the relevant ckm factors
are given as follows : @xmath50 the decay @xmath51 is related to @xmath52 by interchanging all down and strange quarks , i.e. through the so - called @xmath2-spin subgroup of the @xmath12 flavour symmetry of strong interactions . using again the unitarity of the ckm matrix and a notation similar to that in ( [ bd - ampl2 ] ) , we obtain @xmath53,\ ] ] where @xmath54 and @xmath55 correspond to ( [ aap - def ] ) and ( [ ap - def ] ) , respectively .
it should be emphasized that ( [ bd - ampl2 ] ) and ( [ bs - ampl ] ) are completely general parametrizations of the @xmath32 decay amplitudes within the standard model , relying only on the unitarity of the ckm matrix . in particular
, these expressions also take into account final - state - interaction effects , which can be considered as long - distance penguin topologies with internal up- and charm - quark exchanges @xcite . if we compare ( [ bd - ampl2 ] ) and ( [ bs - ampl ] ) with each other
, we observe that the quantity @xmath56 is doubly cabibbo - suppressed in the @xmath35 decay amplitude ( [ bd - ampl2 ] ) , whereas @xmath57 enters in the @xmath51 amplitude ( [ bs - ampl ] ) in a cabibbo - allowed way .
this feature has important implications for the cp - violating effects arising in the corresponding time - dependent decay rates .
the time evolution for decays of initially , i.e. at time @xmath28 , present neutral @xmath11 or @xmath58 mesons into a final cp eigenstate @xmath59 , satisfying @xmath60 is given as follows @xcite : @xmath61\ ] ] @xmath62,\ ] ] where the @xmath63 denote the decay widths of the @xmath11 mass eigenstates , @xmath64 , and @xmath65 is their mass difference . for the @xmath11 decays considered in this paper , the `` unevolved '' decay amplitudes take the form @xmath66\equiv { \cal n}z \label{z - def}\\ \overline{a}&=&\eta\,{\cal n}\left[1-b\ , e^{i\rho } e^{-i\gamma}\right]\equiv \eta\,{\cal n}\,\overline{z}\,,\label{zbar - def}\end{aligned}\ ] ] and we have @xmath67}\nonumber\\ & & = ( 1+\eta\cos\phi)-2\,b\cos\rho\left[\cos\gamma+\eta\cos(\phi+\gamma ) \right]+b^2\left[1+\eta\cos(\phi+2\,\gamma)\right]\end{aligned}\ ] ] @xmath68}\nonumber\\ & & = ( 1-\eta\cos\phi)-2\,b\cos\rho\left[\cos\gamma-\eta\cos(\phi+\gamma ) \right]+b^2\left[1-\eta\cos(\phi+2\,\gamma)\right]\end{aligned}\ ] ] @xmath69 @xmath70.\ ] ] here the phase @xmath71 denotes the @xmath11@xmath58 mixing phase : @xmath72 where @xmath73 is tiny in the standard model because of a cabibbo suppression of @xmath74 .
note that the observables @xmath75 , @xmath76 , @xmath77 and @xmath78 satisfy the relation @xmath79 for the following considerations , it is useful to introduce the time - dependent cp asymmetry @xmath80\ ] ] with @xmath81 \label{amix - def}\\ { \cal a}_{\delta\gamma}&\equiv&\frac{r_{\rm h}-r_{\rm l}}{r_{\rm h}+r_{\rm l}}= -\,\eta\left[\frac{\cos\phi-2\,b\,\cos\rho\,\cos(\phi+\gamma)+ b^2\cos(\phi+2\,\gamma)}{1 - 2\,b\cos\rho\cos\gamma+b^2}\right],\label{ag - def}\end{aligned}\ ] ] and the observable @xmath82 in the cp asymmetry ( [ acp - def ] ) , we have separated the `` direct '' from the `` mixing - induced '' cp - violating contributions
. it is interesting to note that not only @xmath83 , but also @xmath84 does not depend on the @xmath11@xmath58 mixing phase @xmath71 . the observables @xmath83 , @xmath85 and @xmath86
are not independent quantities , and satisfy the relation @xmath87 the formulae given above describe the time evolution of all kinds of neutral @xmath11 decays into a final cp eigenstate , where the `` unevolved '' decay amplitudes take the form specified in ( [ z - def ] ) and ( [ zbar - def ] ) .
let us turn , in the following section , to the @xmath7 observables , which may provide an interesting strategy to determine @xmath3 .
the observables introduced in ( [ adir - def])([ag - def ] ) can be obtained directly from the time evolution of the decay rates corresponding to ( [ at ] ) and ( [ atbar ] ) and do not depend on the overall normalization @xmath88 . however , owing to ( [ obs - rel ] ) , we have only two independent observables , depending on the three `` unknowns '' @xmath89 , @xmath90 and @xmath3 , and on the @xmath11@xmath58 mixing phase @xmath71 .
consequently , in order to determine these `` unknowns '' , we need an additional observable , which is provided by @xmath84 . unfortunately , the time - dependent decay rates fix only the quantity @xmath91 through @xmath92,\ ] ] where `` phsp '' denotes an appropriate , straightforwardly calculable phase - space factor .
consequently , the overall normalization @xmath88 is required in order to determine @xmath84 . in the case of the decay
@xmath0 , this normalization can be fixed through the cp - averaged @xmath1 rate with the help of the @xmath2-spin symmetry . in the case of @xmath1 , we have @xmath93 whereas we have in the @xmath0 case @xmath94 consequently , we obtain @xmath95 ^ 3 \frac{\langle\gamma\rangle}{\langle\gamma'\rangle}= \frac{1 - 2\,a\,\cos\theta\cos\gamma+a^2}{1 + 2\,\epsilon\,a'\cos\theta'\cos\gamma+\epsilon^2\,a'^2}\,,\ ] ] where @xmath96\left[1-(x - y)^2\right]}\ ] ] is the usual two - body phase - space function , and @xmath97 and @xmath98 can be determined from the `` untagged '' @xmath7 rates with the help of ( [ rn - def ] ) and ( [ untag - rate ] ) . since the @xmath2-spin flavour symmetry of strong interactions implies @xmath99 we can determine @xmath100 , @xmath101 and @xmath3 as a function of the @xmath29@xmath102 mixing phase by combining @xmath103 with @xmath104 and @xmath105 or @xmath106 .
in contrast to certain isospin relations , electroweak penguins do not lead to any problems in these @xmath2-spin relations . as we have already noted , the @xmath21@xmath22 mixing phase @xmath107 is expected to be negligibly small in the standard model .
it can be probed with the help of the decay @xmath108 ( see , for example , @xcite ) .
large cp - violating effects in this decay would signal that @xmath109 is not tiny , and would indicate new - physics contributions to @xmath21@xmath22 mixing . strictly speaking , in the case of @xmath0
, we have @xmath110 , where @xmath111 is related to the @xmath112@xmath113 mixing phase and is negligibly small in the standard model . on the other hand
, we have @xmath114 in the case of @xmath1 .
since the value of the cp - violating parameter @xmath115 of the neutral kaon system is small , @xmath111 can only be affected by very contrived models of new physics @xcite .
an important by - product of the strategy described above is that the quantities @xmath116 and @xmath117 allow us to take into account the penguin contributions in the determination of @xmath4 from @xmath118 , which are presumably very small because of the cabibbo suppression of @xmath119 in ( [ bd - ampl2 ] ) .
moreover , using ( [ su3 - 2 ] ) , we obtain an interesting relation between the direct cp asymmetries arising in the modes @xmath1 and @xmath0 and their cp - averaged rates : @xmath120 ^ 3 \frac{\langle\gamma\rangle}{\langle\gamma'\rangle}.\ ] ] an analogous relation holds also between the @xmath121 and @xmath122 cp - violating asymmetries @xcite . at `` second - generation '' @xmath11-physics experiments at hadron machines , for instance at lhcb , the sensitivity may be good enough to resolve a direct cp asymmetry in @xmath1 . in view of the impressive accuracy that can be achieved in the era of such experiments , it is also an important issue to think about the theoretical accuracy of the determination of @xmath4 from @xmath1 .
the approach discussed above allows us to control these presumably very small hadronic uncertainties with the help of @xmath0 .
interestingly , the strategy to extract @xmath3 from @xmath123 decays does not require a non - trivial cp - conserving strong phase @xmath101 .
however , its experimental feasibility depends strongly on the value of the quantity @xmath100 introduced in ( [ a - def ] ) .
it is very difficult to estimate @xmath100 theoretically .
in contrast to the `` usual '' qcd penguin topologies , the qcd penguins contributing to @xmath7 require a colour - singlet exchange , as indicated in fig .
[ fig : bdpsiks ] through the dashed lines , and are `` zweig - suppressed '' .
such a comment does not apply to the electroweak penguins , which contribute in `` colour - allowed '' form .
the current current amplitude @xmath124 is due to `` colour - suppressed '' topologies , and the ratio @xmath125 , which governs @xmath100 , may be sizeable .
it is interesting to note that the measured branching ratio @xmath126 @xcite probes only the combination @xmath127 of current current and penguin amplitudes , and obviously does not allow us to separate these contributions . it would be very important to have a better theoretical understanding of the quantity @xmath128 . however , such analyses are far beyond the scope of this paper , and are left for further studies .
if we use @xmath129 ^ 3 \frac{\tau_{b_s}}{\tau_{b_d}}\ ] ] and ( [ su3 - 1 ] ) , we expect a @xmath0 branching ratio at the level of @xmath130 .
the general expressions for the observables ( [ adir - def])([ag - def ] ) and ( [ h - def ] ) are quite complicated .
however , they simplify considerably , if we keep only the terms linear in @xmath100 . within this approximation , we obtain the simple result @xmath131 allowing us to determine @xmath3 from the cp - averaged @xmath7 rates and the mixing - induced cp asymmetry arising in @xmath0 . in the general case , where no approximations are made
, there is also a `` transparent '' strategy to determine @xmath3 .
the point is that the cp - violating asymmetries @xmath83 and @xmath85 allow us to fix contours in the @xmath3@xmath100 plane , which are described by @xmath132},\ ] ] where @xmath133 with @xmath134 and @xmath135 it should be emphasized that these contours are _ theoretically clean_. it is also possible to combine the direct and mixing - induced cp asymmetries arising in @xmath136 in an analogous way @xcite , allowing us to fix certain contours as well @xcite .
so far , we have not yet used the observable @xmath103 . combining it with @xmath85
, we can fix another contour in the @xmath3@xmath100 plane : @xmath137 if we use @xmath86 instead of @xmath85 , we obtain the same expression for @xmath100 as given in ( [ contours2 ] ) , where @xmath138 and @xmath139 specified in ( [ u - def ] ) and ( [ v - def ] ) are replaced by @xmath140 the intersection of the contours described by ( [ contours1 ] ) and ( [ contours2 ] ) fixes both @xmath100 and @xmath3 .
let us illustrate this approach in a quantitative way by considering a simple example . assuming a negligible @xmath21@xmath22 mixing phase , i.e. @xmath141 , and @xmath142 , which lies within the range allowed at present for this angle , implied by the usual indirect fits of the unitarity triangle , as well as @xmath143 and @xmath144
, we obtain the @xmath0 observables @xmath145 , @xmath146 , @xmath147 and @xmath148 .
the corresponding contours in the @xmath3@xmath100 plane are shown in fig .
[ fig : bscont ] , where the solid lines are obtained with the help of ( [ contours1 ] ) , and the dot - dashed lines correspond to ( [ contours2 ] ) .
interestingly , in the case of the contours shown in fig .
[ fig : bscont ] , we would not have to deal with `` physical '' discrete ambiguities for @xmath3 , since values of @xmath100 larger than 1 would simply appear unrealistic .
if it should become possible to measure @xmath86 with the help of the widths difference @xmath149 , the dotted line could be fixed . in this example , the approximate expression ( [ gam - approx ] ) yields @xmath150 , which deviates from the `` true '' value of @xmath142 by only 8% .
it is also interesting to note that we have @xmath151 in our example .
= 5truecm = 5.5truecm before turning to the @xmath8 decays in the next section , let us say a few words on the @xmath12-breaking corrections . whereas the contours in the @xmath3@xmath100 plane related to ( [ contours1 ] ) , i.e. the solid curves in fig . [ fig : bscont ] , are _ theoretically clean _ , those described by ( [ contours2 ] ) , i.e. the dot - dashed and dotted lines in fig .
[ fig : bscont ] , are affected by @xmath2-spin - breaking corrections .
because of the small parameter @xmath152 in ( [ h - def ] ) , these contours are essentially unaffected by possible corrections to ( [ su3 - 2 ] ) , and rely predominantly on the @xmath2-spin relation @xmath153 . in the `` factorization '' approximation , we have @xmath154 where the form factors @xmath155 and @xmath156 parametrize the quark - current matrix elements @xmath157 and @xmath158 , respectively @xcite .
we are not aware of quantitative studies of ( [ su3-break ] ) , which could be performed , for instance , with the help of sum rule or lattice techniques . in the light - cone sum - rule approach
, sizeable @xmath12-breaking effects were found in the case of the @xmath159 form factors @xcite .
it should be emphasized that also non - factorizable corrections , which are not included in ( [ su3-break ] ) , may play an important role .
we are optimistic that we will have a better picture of @xmath12 breaking by the time the @xmath0 measurements can be performed in practice .
the decays @xmath161 are transitions into a cp eigenstate with eigenvalue @xmath162 and originate from @xmath163 quark - level decays .
we have to deal both with current current and with penguin contributions , as can be seen in fig .
[ fig : bdd ] . in analogy to ( [ bd - ampl2 ] ) and ( [ bs - ampl ] ) , the corresponding transition amplitudes can be written as @xmath164\label{bdds - ampl}\\ a(b_d^0\to d_d^+d_d^-)&=&-\lambda\,\tilde{\cal a}\left[1-\tilde a\ , e^{i\tilde\theta}e^{i\gamma}\right],\label{bddd - ampl}\end{aligned}\ ] ] where the quantities @xmath165 , @xmath166 and @xmath167 , @xmath168 take the same form as in the @xmath169 case .
in contrast to @xmath7 , there are `` colour - allowed '' current current contributions to @xmath160 , as well as contributions from `` exchange '' topologies , and the qcd penguins do not require a colour - singlet exchange , i.e.are not `` zweig - suppressed '' .
usually , @xmath170 decays appear in the literature as a tool to probe @xmath4 @xcite .
in fact , if penguins played a negligible role in these modes , @xmath4 could be determined from the corresponding mixing - induced cp - violating effects .
however , the penguin topologies , which contain also important contributions from final - state - interaction effects , may well be sizeable , although it is very difficult to calculate them in a reliable way .
the strategy proposed here makes use of these penguin topologies , allowing us to determine @xmath3 , if the overall @xmath170 normalization is fixed through the cp - averaged , i.e. the `` untagged '' @xmath171 rate , and if the @xmath15@xmath16 mixing phase @xmath17 is determined with the help of @xmath1 .
it should be emphasized that no @xmath172 oscillations have to be resolved to measure the untagged @xmath171 rate .
since the phase structures of the @xmath173 and @xmath174 decay amplitudes are completely analogous to those of @xmath175 and @xmath35 , respectively , the formalism developed in the previous section can be applied by performing straightforward replacements of variables .
taking into account phase - space effects , we have @xmath176 \frac{\langle\tilde\gamma\rangle}{\langle\tilde\gamma'\rangle}\,,\ ] ] where the cp - averaged rates @xmath177 and @xmath178 can be determined with the help of ( [ rn - def ] ) and ( [ untag - rate ] ) , and the function @xmath179 is as given in ( [ phsp ] ) .
let us illustrate the strategy to determine @xmath3 , again by considering a simple example .
assuming @xmath180 , @xmath181 , @xmath142 and a @xmath15@xmath16 mixing phase of @xmath182 , we obtain the @xmath170 observables @xmath183 , @xmath184 and @xmath185 . in this case
, studies of cp violation in @xmath1 would yield @xmath186 , which is the central value of the most recent cdf analysis @xcite , implying @xmath187 or @xmath188 . here
we have assumed that @xmath189 $ ] , as implied by the measured value of @xmath115 .
a similar comment applies to the range for @xmath3 .
the former solution for @xmath17 would lead to the contours in the @xmath3@xmath190 plane shown in fig .
[ fig : bdcontours1 ] .
the contours corresponding to @xmath191 are shown in fig .
[ fig : bdcontours2 ] .
since values of @xmath192 appear unrealistic , we would obtain the two `` physical '' solutions of @xmath193 and @xmath194 for @xmath3 , which are due to the twofold ambiguity of @xmath17 .
there are several strategies to resolve this discrete ambiguity in the extraction of @xmath4 @xcite , which should be feasible in the era of `` second - generation '' @xmath11-physics experiments . as in the @xmath169 case ,
only the contours involving the observable @xmath195 are affected by @xmath12-breaking corrections .
because of the small parameter @xmath196 , they are essentially due to the @xmath2-spin - breaking corrections to @xmath197 where the restrictions form the heavy - quark effective theory for the @xmath198 form factors have been taken into account by introducing appropriate isgur wise functions @xmath199 with @xmath200 @xcite .
studies of the light - quark dependence of the isgur wise function were performed within heavy - meson chiral perturbation theory , indicating an enhancement of @xmath201 at the level of @xmath202 @xcite . applying the same formalism to @xmath203
gives values at the 1.2 level @xcite , which is of the same order of magnitude as the results of recent lattice calculations @xcite .
further studies are needed to get a better picture of the @xmath12-breaking corrections to the ratio @xmath204 . since `` factorization
'' may work reasonably well for @xmath205 , the leading corrections are expected to be due to ( [ su3-breakbdd ] ) .
the experimental feasibility of the strategy to extract @xmath3 from @xmath8 decays depends strongly on the size of the penguin parameter @xmath190 , which is difficult to predict theoretically .
the branching ratio for @xmath173 is expected at the @xmath206 level @xcite ; the one for @xmath174 is enhanced by @xmath207 , and is correspondingly expected at the @xmath208 level . already at the asymmetric @xmath19 @xmath11-factories starting very soon , it should be possible to perform time - dependent measurements of the decay @xmath170 , whereas @xmath171 and its `` untagged '' rate may be accessible at cdf or hera b .
however , unless the penguin effects in @xmath170 are very large , the approach to determine @xmath3 discussed in this section appears to be particularly interesting for `` second - generation '' experiments , such as lhcb .
the @xmath19 @xmath11-factory experiments should nevertheless have a very careful look at the decay @xmath170 , and those at hadron machines should study its @xmath2-spin counterpart @xmath171 .
the observables of the time - dependent @xmath0 rate , in combination with the cp - averaged @xmath1 rate , provide an interesting strategy to determine the angle @xmath3 of the unitarity triangle .
this approach is not affected by any final - state - interaction effects , and its theoretical accuracy is only limited by the @xmath2-spin flavour symmetry of strong interactions . as a by - product , it allows us to take into account the penguin effects in the determination of @xmath4 from @xmath1 , which are presumably very small .
an analogous strategy is provided by the time evolution of @xmath170 decays and the untagged @xmath171 rate .
these new strategies may be promising for `` second - generation '' @xmath11-physics experiments , for example lhcb .
their experimental feasibility strongly depends on the size of the penguin effects in @xmath7 and @xmath160 , which are very difficult to calculate and require further theoretical studies .
recent experimental results of the cleo collaboration on certain non - leptonic @xmath11 decays , which are dominated by penguin contributions , have shown that these topologies may well lead to surprises .
opal collaboration ( k. ackerstaff _
_ ) , _ eur . phys .
j. _ * c5 * ( 1998 ) 379 ; cdf collaboration ( f. abe _
_ ) , _ phys . rev .
_ * 81 * ( 1998 ) 5513 ; for an updated analysis , see preprint cdf / pub / bottom / cdf/4855 . for reviews ,
see , for instance , _ the babar physics book _ ,
harrison and h.r .
quinn ( slac report 504 , october 1998 ) ; y. nir , published in the proceedings of the 18th international symposium on lepton photon interactions ( lp 97 ) , hamburg , germany , 28 july1 august 1997 , eds . a. de roeck and a. wagner ( world scientific , singapore , 1998 ) , p. 295
[ hep - ph/9709301 ] ; m. gronau , _ nucl .
phys.proc .
suppl . _ * 65 * ( 1998 ) 245 ; r. fleischer , _ int .
_ * a12 * ( 1997 ) 2459 .
see , for example , y. grossman and h.r .
quinn , _ phys .
_ * d56 * ( 1997 ) 7259 ; j. charles , a. le yaouanc , l. oliver , o. pne and j .- c .
raynal , _ phys .
_ * b425 * ( 1998 ) 375 ; a.s .
dighe , i. dunietz and r. fleischer , _ phys .
_ * b433 * ( 1998 ) 147 . | a completely general parametrization of the time - dependent decay rates of the modes @xmath0 and @xmath1 is given , which are related to each other through the @xmath2-spin flavour symmetry of strong interactions .
owing to the interference of current current and penguin processes , the @xmath0 observables probe the angle @xmath3 of the unitarity triangle . using the @xmath2-spin symmetry ,
the overall normalization of the @xmath0 rate can be fixed with the help of the cp - averaged @xmath1 rate , providing a new strategy to determine @xmath3 .
this extraction of @xmath3 is not affected by any final - state - interaction effects , and its theoretical accuracy is only limited by @xmath2-spin - breaking corrections . as a by - product , this strategy allows us to take into account also the penguin effects in the determination of @xmath4 from @xmath1 , which are presumably very small , and to predict the direct cp asymmetry arising in this mode . an analogous strategy is provided by the time - dependent @xmath5 rate , if its overall normalization is fixed through the cp - averaged @xmath6 rate .
cern - th/99 - 78 + hep - ph/9903455 * extracting @xmath3 from @xmath7 and @xmath8 * robert fleischer + _ theory division , cern , ch-1211 geneva 23 , switzerland
_ cern - th/99 - 78 + march 1999 |
it is well understood that the electromagnetic fields giving rise to all optical phenomena have ultimately to be represented by quantum operators .
these couple to the degrees of freedom of matter and their modification due to this interaction constitutes the quantum counterpart of the action of optical components .
ordinary optical devices operate in the linear regime of interaction , but the important area of nonlinear optics is based on higher order effects of the field - matter coupling . in most situations , the optical phenomena can be described entirely in terms of classical fields , but many recent investigations require that the quantum character of the field is accounted for .
such research constitutes the topics of quantum optics @xcite .
however , many quantum effects are of interest even in the linear regime of operation : quantum noise @xcite sets the limit to communication by optical channels and amplifiers , quantum interference shows up in precision measurements and tests of fundamental issues and also in reading and writing quantum information .
manipulation of quantum information such as quantum computations usually requires the inclusion of higher order effects , i.e. nonlinear interactions between the qubits @xcite . in this paper , we are going to discuss the dynamics of single photon wave packets in various two - dimensional atomic configurations .
these are taken to be models of optical networks , where we explicitly include the atomic nature of the optical components distributed over the volume under investigation .
this approach provide us with a completely microscopic quantum - mechanical picture of how photon wave packets interact with optical elements represented as collections of two level atoms . for practical reasons
, we have to restrict our work to one - photon states , but this is not such a serious limitation as it may seem .
all _ linear _ optical effects are based on the single - photon interacting with material structures , and consequently we have a general description of quantum optics phenomena in the linear regime . the need to consider multi - photon effects arises only in connection with the quantum treatment of nonlinear optics .
there are two basic ways to approach the quantization of optical systems . in the conventional one , we determine a complete set of eigenmodes of the total universe , and express the fields of interest in terms of these .
any matter present is described through its interaction with the fields , and the coupled field - matter problem is then solved to the best of our ability . this is the approach utilized in traditional qed , and its development is found in many standard texts .
the alternative approach , designed for quantum optics applications , is to determine the eigenmodes of the system at the classical level , the matter involved is then treated as boundary conditions on the field modes .
especially the new area of cavity qed research @xcite , utilizes this point of view , and it provides the basis both for quantum communication theories and many fundamental investigations . in the field of optics ,
the components are usually treated as boundary conditions only , and the complete optical device is considered to be an optical network .
this approach has been discussed thoroughly in the classical regime of operation @xcite . for linear devices
, the classical treatment can be taken over into the quantum regime by the use of suitable quantum optics tools @xcite . in principle ,
any device understood classically , can be treated quantum mechanically with such an approach . the specific quantum features manifest themselves in the initial conditions and the restrictions on observability imposed by quantum theory @xcite .
another specifically quantum mechanical effect is the occurrence of spontaneous decay . within a one - dimensional model of the modes of the universe , this is discussed in ref .
@xcite , where both free weisskopf - wigner decay and cavity modified decay are discussed .
such phenomena have been the object of much interest within qed research , for an extensive list of references see ref .
@xcite . within the model chosen there
, one can see the emergence of the exponential law and the inhibition of decay observed in a photonic band gap structure .
in general , the model provides insight into the role of atomic media in the irreversible transfer of excitation energy into the field modes of the universe . in this paper
we combine the two views discussed above : we retain a description in terms of a complete set of two - dimensional eigenmodes of the universe .
the optical components are described in terms of their atomic constituents .
all atomic structures are represented by spatially localized two - level atoms .
these are treated as point - like partciles in accordance with the dipole approximation assumed to be valid .
the state of the field is taken to be a single photon wave packet with a narrow energy distribution . in this case
, the state can be described by a truncated expansion in terms of modes of the universe .
the spatially distributed two - level atoms describing the structures are taken to be initially in their ground states .
the atoms can be chosen to resonate with the central frequency of the photon wave packet or be well off resonance ; various effects can be modeled in this way .
when the single photon is absorbed , only one of the atoms is excited , and the field is reduced to its ground state .
such a choice limits the hilbert space needed in the calculations to manageable size , but allows us to investigate many simple networks of significance in linear optics .
all such effects are , in principle , describable at the single photon level ; only nonlinear optics effects require more photons , which would make the hilbert space expand beyond the limits of available computer resources .
our approach based on a complete set of eigenmodes allows us to investigate the dynamic performance of many linear systems . in order to illustrate the method , we select the simplest optical components : mirrors , beam splitters , focusing devices and interferometers .
the over - all performance of the components follows directly from their classical theory , but our approach allows us to investigate the microscopic ( quantum ) behavior of the setup .
quantum coherence between various spatial regions in the device is directly visible in the states calculated , and the time and space scales of the various interferometric structures can be read off the results .
combined with various models of measurements , our calculations contain considerably more information than a simple classical computation .
here we only discuss the measurement of frequency and the possible occurrence of filtering action in the atomic structures , which does not in itself depend too much on the quantum nature of the fields . but modeling the frequency detection by atomic absorption , we utilize the full character of the model , which allows further extension to quantum correlation measurements if we so desire .
our work is based on the model put forward in @xcite which we extend to two dimensions .
the quantized modes of the universe are introduced in sec .
ii together with their interaction with the spatially distributed atoms . in sec .
iii we specify the details of the model and indicate how the calculations have been carried out .
section iv presents the various simple components analyzed in this paper .
we describe how they are modeled and show the results of the detailed solution of the time evolution . finally in sec .
v we present our conclusions and discuss possible extensions and applications of the work .
the field is enclosed inside a two dimensional cavity determined by the relations @xmath0 the periodic boundary conditions restrict the allowed values in @xmath1-space to a discrete set @xmath2 in computer simulations , the @xmath1-values must be restricted by giving some upper limit for the integer @xmath3 which corresponds to a specific frequency cut - off . the electric and magnetic field can be expanded @xcite using the mode functions @xmath4 where the summation @xmath5 is over all @xmath1-values ( [ kvalues ] ) and two polarization indices @xmath6 .
the frequency @xmath7 is the same for both polarizations @xmath8 the general * k*-vector in two dimensions can be written @xmath9 the polarization vectors which obey the usual right hand rule conventions are @xmath10 the @xmath1-vector and polarization indexes satisfy the relations @xcite
@xmath11 the energy - density operator is @xmath12 using ( [ efree ] ) and ( [ bfree ] ) gives @xmath13\end{aligned}\ ] ] 2 in our simulations , we have restricted the polarization of the field to @xmath14 .
the modes with @xmath15 are taken to have have zero amplitudes .
in addition to that we restrict the number of excitations of our basis vectors to one . for these kind of basis vectors the terms @xmath16 and @xmath17 do not give any contribution .
these terms can be omitted from the expressions .
for the states described above , the expectation values are obtained by replacing the operators with the coefficients of the corresponding statevectors @xmath18 and @xmath19 .
the normally - ordered terms in the energy density become ( normal ordering is indicated by colons ) @xmath20 where @xmath21 the two - fold summation over the @xmath1-space is seen to factorize and the formulas for @xmath22 and @xmath23 are fourier transforms of two different functions . for numerical simulations
these two properties are essential as will be seen later .
we note that if the polarization is such that the modes with @xmath24 are taken to have zero amplitudes , then the two terms @xmath25 and @xmath26 in the expression for the energy density are equal . integrating ( [ energydensity ] ) over
the spatial coordinates and using the integral @xmath27 gives the familiar form @xmath28 which in the normally ordered form reads @xmath29
in the previous chapter the formulas for the field in the vacuum were derived . in this chapter
we add an assembly of two level atoms to the cavity and give the corresponding hamiltonians .
the general form of the statevector with one excitation is also given .
the material presented here is based on the similar simulations in one dimension done by v.buek et.al .
the simulations in two dimensions are numerically more demanding , but we have been able to develop efficient numerical methods which make these simulations possible . the total hamiltonian @xmath30 can be divided into three parts @xmath31 where the field hamiltonian is given by equation ( [ hf ] ) . the atomic hamiltonian is the sum over all one - atom hamiltonians @xmath32 where @xmath33 is the transition frequency of the @xmath34-th atom and @xmath35 is pauli s spin matrix . in the interaction
hamiltonian the dipole approximation is used . for simplicity
the dipole operator is taken to be @xmath36 i.e. it has a component in the @xmath37 direction only .
the general dipole vector would have components in the @xmath38- and @xmath39-directions too .
the interaction hamiltonian has the form @xmath40 where @xmath41 is the electric field operator ( [ efree ] ) at the position of the atom .
the rotating wave approximation ( rwa ) is to be used , and we neglect the @xmath42 - and @xmath43 terms .
in addition to that we replace the mode frequency in the electric field operator by the atomic frequency and use the dot products @xmath44 and @xmath45 to get @xmath46 in what follows we omit the polarization index in subscripts of field operators .
the coupling constant is @xmath47 only those modes whose resonance frequency is close to the atomic frequency interact significantly with the atom , so we can replace the mode frequency @xmath48 by the atomic frequency @xmath33 in equation ( [ couplingconstant ] ) . in all simulations
we have restricted the total number of excitations to one .
consequently , the most general statevector of the atom - field system has the form @xmath49 the first sum contains all the basis vectors where the excitation is in one of the field modes and all the atoms are in the ground state . in the second sum
the field modes are in the vacuum state and one of the atoms is excited .
the complex numbers @xmath50 and @xmath51 are the probability amplitudes of the corresponding basis vectors .
we have dropped the polarization indices because in our simulations only the basis vectors with the polarization vector @xmath14 are excited as was discussed earlier .
the general gaussian one photon statevector is of the form @xmath52 where the mode coefficient @xmath50 is @xmath53 the parameters @xmath54 and @xmath55 are @xmath56 if the cross - variance @xmath55 vanishes the formula for @xmath50 reduces to two independent gaussian distributions @xmath57\end{aligned}\ ] ] all initial distributions used in our simulations are of the form ( [ gauss ] ) .
the distribution ( [ gauss ] ) in @xmath1-space is centered around @xmath58 with the corresponding central frequency @xmath59 . if @xmath60 the distribution is symmetric . if @xmath61 the distribution is wider in the @xmath39-direction ( and vice versa ) .
the variances in @xmath1-space and configuration space are inversely proportional .
if @xmath62 is small , the energy density distribution in configuration space is wide in the x - direction .
the normally ordered energy distribution associated with the state ( [ generalgauss ] ) or ( [ gauss ] ) is well localized near the point @xmath63 in the configuration space .
essential for this is the phase part @xmath64 of the coefficient @xmath50 .
if the form of the phase was different the intensity profile would not be gaussian .
the time evolution of the gaussian wave packet inside an empty cavity is determined by the hamiltonian @xmath65 ( [ hf ] ) with the corresponding evolution operator @xmath66 .
applying this to the state ( [ statevec ] ) gives for the time - evolution of the coefficients @xmath67 .
the absolute value of the coefficients remain the same , only the phase changes . for the phase part
we get @xmath68,\ ] ] where @xmath69 .
the time evolution inside the empty cavity reduces to the time evolution of the parameter @xmath70 .
we remember that the phase factor determine the shape of the normal ordered intensity profile .
because the time evolution of the phase is different for different modes , the normal ordered intensity does not preserve its original gaussian shape .
if the direction of the vector @xmath71 is more or less the same for all basis vectors which have nonzero coefficients , the shape of the energy density distribution remains approximately the same longer .
the situation is like this when the state vector in @xmath1-space is centered around some @xmath1-value far from the origin and the variances are small .
it turned out to be faster to carry out the numerical integration in the interaction picture .
the transformation hamiltonian is @xmath72 .
the interaction hamiltonian in the rotating frame is @xmath73 which is obtained by the following replacement @xmath74 in equation ( [ hi ] ) , and we get @xmath75 the statevectors in the interaction picture become @xmath76 and the schrdinger equation for the wavefunction is @xmath77 integration of the schrdinger equation in the interaction picture is faster than the original equation because only the interaction hamiltonian is present . our choice for the integration method of the time dependent schrdinger equation is a classical four stage fourth order runge - kutta method . if the wavefunction at time @xmath78 is @xmath79 the wavefunction at a later time @xmath80 ( @xmath81 small ) @xmath79 is given by the following algorithm @xcite @xmath82 the timestep @xmath81 is a fixed constant .
the essential part of the integration from the numerical point of view is how to evaluate the right hand part of the equation ( [ schrinint ] ) as efficiently as possible .
the first term in the equation ( [ hiinint ] ) gives @xmath83 and the second one @xmath84 hence the new coefficients for the atomic ( @xmath85 ) and field ( @xmath86 ) basis vectors become @xmath87 where @xmath88
both @xmath89 and @xmath90 are two dimensional fourier transforms , so in numerical calculations the fast fourier transform ( fft ) can be used . the speed increase obtained by using fft instead of the direct summation is enormous especially in simulations with a large number of atoms . in some simulations
it can be said that only this method makes these simulations possible .
there are several natural checks for the numerical simulations .
first of all the norm of the wavefunction has to remain unity for all times .
the system is closed so the total energy of the system must be constant all the time .
the field energy can be calculated using either the formula ( [ hf ] ) or integrating the energy density over the whole cavity .
the two methods should give the same results .
in the following simulations the spectrum is detected using so - called analyzer atoms @xcite .
many atoms with a very small dipole coupling constant are put into specific locations in the cavity .
all the atoms have different transition frequencies in between @xmath91 and @xmath92 @xmath93 also the dipole constants are all different and very small @xmath94 where @xmath95 is a very small constant , typically c=0.0001 or so . the form ( [ dipole ] ) of @xmath96
gives the same decay constant @xmath97 for all the atoms because in two dimensions @xmath97 is directly proportional to the product @xmath98 .
because the dipole coupling is small , the atoms have very small decay constants and linewidths and only the radiation which is exactly on resonance with the atom can excite it .
therefore the excitation of the atoms as function of @xmath99 can be interpreted as a spectrum of the field at the position of the atoms . because the interaction between the radiation and the atoms is small ,
the state of the field does not change appreciably .
the method can be used to detect the local time dependent spectrum .
two - time averages , usually used in spectrum calculations , are not needed .
a more detailed description of the method and comparisons with the time dependent spectra defined using two - time averages @xcite can be found in the paper by m.havukainen and s.stenholm @xcite , where it was used to detect the spectrum of a radiation emitted by a laser driven three level atom .
in this chapter the results of several simulations are presented .
first we show that the energy density profile of the free photon does not preserve its shape if @xmath59 is small as was explained earlier . in the second and third simulation
, atoms are used as mirrors and beam splitters . using these components it is possible to build many optical systems .
we present an interferometer as an example .
we also present a simulation of a two - slit experiment .
finaly , we also briefly study a spontaneous decay of a two level atom into the vacuum of electromagnetic modes in the two - dimensional cavity . in the first simulation the time evolution of the free photon wave packet
is studied .
the initial wave packet is gaussian ( [ gauss ] ) with parameters @xmath100 , @xmath101 , @xmath102 , @xmath103 and @xmath104 .
the probabilities @xmath105 of the field modes are shown fig .
[ freekspace ] .
the central frequency of the photon wave packet is so small that the @xmath1-vectors of the modes with nonzero amplitudes are not parallel .
we would expect this to be observed as was explained earlier .
the time evolution of the energy density at two time values is shown in fig .
[ figfree ] .
the initial gaussian photon wave packet has an energy density centered at @xmath106 , @xmath107 .
the wave packet is moving to the right . during the free evolution energy density
becomes delocalized . from the figure we see that at @xmath108 the width in the @xmath39-direction is much larger than the initial value .
this spread of the width of the original wave packet is a standard quantum - mechanical effect .
it is possible to `` build '' mirrors and beam splitters using two level atoms . in the next simulation many atoms with large dipole constants
were arranged into a slab configuration .
we take a @xmath109 angle between the slab and the @xmath38-axis .
we assume all atoms to have the same transition frequencies and dipole constants .
the initial photon wave packet has a gaussian distribution ( [ gauss ] ) with parameters @xmath110 , @xmath103 and @xmath111 .
the atoms in the slab are exactly on resonance with the incoming photon wave packet ( i.e. the central frequency of the wave packet @xmath112 is equal to the transition frequency of the atoms ) .
the dipole constant is large @xmath113 .
we assume that the mirror is composed of eight layers of atoms as close to each other as possible .
in our case we assume to distance between neighboring layers of the atoms to coincide with the grid in the configuration space ( the grid spacing is @xmath114 and for the given orientation of the mirror the distance between the different atomic layers is chosen to be @xmath115 ) .
the central wavelength of the incoming photon wave packet is @xmath116 so the difference between the neighboring atoms much shorter than the wavelength of the incoming wave packet .
we plot the energy density of the one - photon wave packet reflected by the mirror in fig .
[ figmirror ] .
firstly we plot the initial wave packet at @xmath117 ( a ) .
the photon is coming towards the atoms of the mirror .
these atoms become excited by the incoming wave packet .
the `` secondary '' radiation which is emitted by the atoms interfere with the incoming wave packet .
this secondary radiation can formally be expressed as a sum of the two terms - the first destructively interfere with the incoming wave packet . as a consequence of this interference
the incoming wave packet is `` destroyed '' ( i.e. becomes extinct ) . the other part of the radiation which is `` collectively '' radiated by the atoms of the mirror represents the reflected wave packet .
in fact , the process of reflection of the wave packet by atoms of the mirror represents purely quantum ( microscopic ) version of the ewald - oseen extinction theorem @xcite . in fig .
[ figmirror](b ) we have chosen conditions such that at @xmath108 all the radiation is reflected by the atoms .
the direction of propagation of the reflected wave packet is the same as expected in the classical theory .
the energy density compared to the incoming wave packet is changed but is still clearly localized . note that the energy density is not perfectly symmetrical .
the reason is the same as in the simulation with a free photon , i.e. the distribution in @xmath118-space is broad and near the origin so the spread of the wave packet is clearly seen .
additionally , the interference between components of radiation emitted by different atoms of the mirror plays a role . in the left part of fig .
[ figphotonatmandbs ] we see the energy density of the photon wave packet close to the surface of the mirror . we see that the incoming and reflected parts interfere .
we also see that no energy is transmitted by the atomic slab . in this sense
the atoms serve as a mirror .
nevertheless , one has to remember that the atoms during the process of reflection of the original wave packet become excited , that is the mirror under consideration has its own `` internal '' ( quantum ) degrees of freedom , so the part of the original energy can be ( transiently ) absorbed by the mirror .
this also result in the fact that this quantum mirror might become entangled with the reflected wave packet .
we note that the parameters of the atoms in this simulation were carefully chosen in such a way that the atoms really form a mirror .
if the parameters are changed then part of the radiation can be transmitted , that is the collection of the atoms can play a rle of a beam splitter . in the previous simulation
we have shown that it is possible to build an almost perfect mirror using two level atoms , assuming the parameters are chosen correctly . using slightly different parameters , we find that the atoms can behave as a beam splitter .
there are several ways how to modify the `` mirror '' configuration to obtain a beam splitter for instance , we can consider a smaller number of atoms , or we can decrease the dipole constants , or change the resonance frequencies of the atoms .
we tried all the possibilities and the most satisfactory results were obtained by detuning the atoms .
the frequencies of the atoms are now taken to be @xmath119 .
the center frequency of the incoming photon wave packet is @xmath120 , i.e. the detuning is really large .
the time evolution of the energy density of the electromagnetic field in this case is shown in fig .
[ figbs ] .
the line in the middle represents the positions of the detuned atoms .
there is only one layer of atoms instead of eight as in the mirror simulation . at @xmath117
the photon is propagating towards the atoms .
here again the incoming wave packet excites the atoms .
now the quantum interference between the incoming and emitted radiation is such that part of the original wave packet is transmitted by the layer of the atoms .
the other part is reflected . in fig .
[ figbs ] we clearly see that at @xmath108 the original wave packet is split into two parts propagating up and to the right .
the energy is divided equally , that is the atoms form a 50 - 50 beam splitter for the incoming photon . here
we stress that the beam splitter under consideration has its own internal degrees of freedom and transiently it becomes excited .
nevertheless , after a while the atoms completely emit the excitation energy and the beam splitter is in a ground state - at this point it is completely disentangled from the one - photon radiation field which is now in a pure superposition state with two macroscopically distinguishable components ( reflected and transmitted ) . in the right hand part of fig .
[ figphotonatmandbs ] energy density of the photon wave packet is shown close to the `` surface '' of the beam splitter . to the left of the atoms the incoming and reflected wave packets interfere .
we also see that a fraction of the original radiation is able to `` pass '' the atoms and to continue to propagate to the right .
the wavelength of the photon was chosen shorter than in the mirror simulation , which can be seen from the interference structure .
we have also studied spectral properties of reflected and transmitted parts of the original wave packet . in this situation
200 atoms were used to detect the time dependent spectra of the two outgoing parts of the photon by applying the method described earlier .
both spectra were identical to the spectrum of the incoming photon .
this means that our quantum beam splitters and mirrors are linear devices , which is important if we want to build optical networks out of the considered optical elements .
another illustration of the power of our microscopic model of optical elements is the parabolic mirror .
in fact , it is possible to `` build '' out of two - level atoms mirrors of arbitrary shapes . in the next simulation ,
the photon wave packet is propagating towards a parabolic mirror the shape of which is described by the equation @xmath121 .
the focus of the parabola is at the point @xmath122 , @xmath123 .
the time evolution of the energy density is shown in fig .
[ figparabola ] . at @xmath117
the gaussian photon is propagating towards the parabola . the little circle in between the photon and the parabola shows the position of the focus . at @xmath124
we see the photon wave packet being reflected from the parabola .
we see the interference between the incoming and the re - emitted radiation .
we note that at @xmath125 most of the radiation goes through the focus . in the last figure ( @xmath126 ) the photon wave packet propagates to the left . using a quantum beam splitter and two quantum mirrors we can `` construct '' a single - photon interferometer ( see fig . [ figifshift00 ] ) . here
the one - photon wave packet comes towards the beam splitter ( @xmath117 ) and is divided into two parts which propagate towards the mirrors ( @xmath126 ) .
the distances of the mirrors from the beam splitter are exactly the same .
the mirrors reflect the radiation back to the beam splitter . at @xmath127
the two reflected parts reach the beam splitter .
each wave packet considered individually would be splitted by the beam splitter into two parts going left and up ( i.e. transmitted and reflected ) .
on the other hand due to the quantum interference between the components of the radiation field coming from the two mirrors we can observe something completely different : if the optical paths of the two components are equal then their relative relative phase is such that quantum interference results in an emergence of a single - photon wave packet traveling up ( @xmath128 ) . on the contrary , if the distances of the two mirrors from the beam splitter are not equal then the relative phase of the two components which interfere on the beam splitter after being reflected by the mirrors can result in a wave packet traveling left ( see fig .
[ figifshift043 ] ) . here
the difference of the optical paths is approximately one central wavelength of the original wave packet .
we see that in this case most of the energy travels in a form of a wave packet to the left .
the microscopic quantum model we study in this paper can be also used to study the two - slit experiment .
let us assume the photon wave packet which has a very broad energy density in the @xmath39-direction , i.e. this wave packet models a plane wave which approaches the mirror with two slits , see fig .
[ figtwoslit ] . on the left
we have placed another mirror . without it , the part of the plane wave which is reflected from the double slit mirror would disappear at the left and reappear on the right because of the periodic boundary conditions we use in our simulations .
the original one - photon wave packet ( @xmath117 ) propagates towards the mirror with two slits . at @xmath129
the `` plane '' wave packet is reflected from mirror .
some of the energy propagates through the slits ( i.e. there is a nonzero probability that the original one - photon wave packet can be transmitted through the mirror via the slits ) .
we see that through each slit a part of the energy propagates to the right - the interference between these components of the electromagnetic field are clearly seen ( see @xmath108 and @xmath130 ) .
the `` plane '' wave packet which has been reflected by the mirror with two slits is then reflected by the left mirror and then again by the the double slit mirror . here part of the energy `` goes '' through the slits again forming a second , more complex , interference pattern ( see @xmath108 ) .
this process of bouncing of the original wave packet between two mirror continues and each time a fraction of the energy passes through the slits .
it is interesting to compare the interference structure with the theoretical prediction derived within classical optics .
the formula can be calculated using huygens principle @xcite . according to this , every point at the slit
can be considered to be a source of secondary wavelets .
the total intensity profile is a superposition of these wavelets .
let us first consider a text book treatment of a one slit mirror , fig .
[ slitfigure ] .
the plane wave packet of the frequency @xmath99 is coming from the left towards a slit of a width @xmath131 .
the field strength on the right coming from a specific point at the slit is proportional to the phase factor @xmath132 .
the phase difference of the radiation coming from two different spatial points in the slit is @xmath133 , if the distance from the mirror is long enough . according to the huygens principle
the total radiation is a superposition @xmath134 for two slits of width @xmath131 and a separation @xmath135 we have two integrals @xmath136 which gives for the intensity @xmath137 on the other hand , we can use results of our numerical simulations and evaluate the intensity of the radiation which has been created to the right of the two - slit mirror during the first reflection of the original wave packet : @xmath138 to neglect the contribution of the second reflection we take the lower bound of the integral over the polar coordinate @xmath139 to be @xmath140 . the theoretical prediction ( [ theoreticalintensity ] ) and
the intensity derived from our simulations ( [ intensitytheta ] ) are shown in fig .
[ figthetaintensity2curves ] .
both intensities are normalized in such a way that their maximum is equal to unity . near @xmath141
the agreement between the two results is very good . for larger values of @xmath142
there is a difference between the two lines which is understandable because we are comparing a classical result with a numerical simulation of a quantum model with realistic features such as the nonzero width of a mirror composed of two - level atoms or the wave packet which is not a plane wave , etc .
taking these differences into account it is surprising that the two pictures coincide so well .
till now we have considered in our simulations that the field has been initially excited and all the atoms were initially in the ground state . obviously , our model can be also applied to a situation when one of the atoms is excited and the field is initially in the vacuum state ( i.e. , we still restrict ourselves to the one - excitation subspace of the total hilbert space ) . in this section
we briefly discuss the problem of a spontaneous decay of a two - level atom in a two - dimensional cavity .
we consider the atomic transition frequency to be @xmath143 and the dipole constant is @xmath144 .
the atom is situated at the origin ( @xmath145 , @xmath107 ) of the two - dimensional cavity .
the number of the field modes is 256@xmath146256=65536 . in fig .
[ decayexcitation ] we present the natural logarithm of the excitation probability of the atom as a function of time . from here
we can conclude that the decay of the atom is approximately exponential with a decay constant @xmath147 . in fig .
[ decaykslides ] we present probabilities of the excitation of the modes @xmath148 ( @xmath149=0 ) . because the direction of the constant dipole vector of the atom is chosen to be in the @xmath150-direction ,
the amplitude profile is the same on any line which goes through the origin of the momentum space .
as expected for times large enough the modes with @xmath151 are dominantly excited .
in fact the peaks are not exactly at the resonance frequency , there is a small shift which is identified to be a lamb shift ( from the figures we can not see this but the shift can be determined from numerical values obtained in the simulation ) .
we plot the energy density of the one - photon wave packet emitted by the decaying atom in fig .
[ decayintensity ] . because of the rotational symmetry of the problem we plot just one `` cut '' ( @xmath123 ) in the energy density as a function of @xmath38 .
the energy density is presented for two times @xmath152 and @xmath153 . at both times there is a peak in the center where the atom is positioned .
this means that at these two times the atom still emits the radiation ( which is in agreement with the chosen decay rate @xmath154 ) .
we turn our attention to the fact that at @xmath152 the energy density is nonzero only for @xmath155 .
analogously for the time @xmath153 the energy density is nonzero only for @xmath156 .
this reflects the fact that the causality is preserved in our simple quantum - mechanical treatment of the decay of the two - level atom in the cavity .
here we have presented just few features of the decay , the complete description of the process deserves more detailed discussion .
for instance , one might be interested on how the decay depends on the mode spectra , the position of the atom , what are the values of the lamb shift , how the decay depends on the frequency cut - off , etc .
we will address these questions elsewhere .
we have shown the results of many quantum mechanical simulations with two level atoms and a one photon wave packet inside a two dimensional cavity .
the initial basis vectors are restricted to admit only one excitation .
because a rotating wave interaction between the radiation and the atoms is used , basis vectors with more excitations acquire no excitation . for these kinds of states the special numerical technique which utilizes fft ( fast fourier transform )
may be used .
using fft the simulations become orders of magnitude faster , allowing more modes and atoms to be included . the atoms are at fixed positions and it is possible to build complicated structures with different kinds of atoms .
several layers of atoms which are on resonance with the incoming radiation form a quantum mechanical mirror if the density of the atoms is high enough .
the mirror may have an arbitrary shape . in our simulations
usual flat and parabolic mirror were used .
one layer of detuned atoms forms a beam splitter .
we have shown that , using mirrors and beam splitters , it is possible to build complicated optical networks . as an example
the time evolution of a photon in an interferometer was studied .
usually the optical components are taken to be classical objects which give boundary conditions to the quantum mechanical time evolution or determine the modes used in a quantization . in our simulations
the whole system including beam splitters and mirrors is in a well - defined quantum mechanical state .
in addition to the simulations shown in this paper is is possible to build more complicated networks of beam splitters and mirrors .
one interesting possibility is to build cavities of arbitrary shape and study the time evolution of the photon intensity inside the cavity .
it is also possible to use moving atoms in the simulations allowing moving beam splitters and mirrors to be built .
one extension of the current model would be to take basis states with more than one photon excitation into account .
however , the number of basis states with a given excitation increases so rapidly that it is unlikely to be possible to use the methods of this paper for fields of higher intensity .
thus all the phenomena of nonlinear optics require novel computational approaches .
we want to thank the academy of finland and the slovak academy of sciences for the financial support .
this work was supported in part by the royal society .
computers of the center for scientific computing ( csc ) were used in the simulations . in many computer programs the c++ class library `` blitz '' developed by todd veldhuizen was used ( http://monet.uwaterloo.ca/blitz/ ) .
finally we want to thank k - a .
suominen for reading the preprint and correcting several misprints . | within a framework of a two - dimensional microscopic purely - quantum mechanical model we analyze dynamics of single - photon wave packets interacting with optical elements ( beam splitters , mirrors ) modeled as systems of two - level atoms .
that is , we utilize a two dimensional cavity to simulate the quantum behavior of simple optical components and networks thereof .
the field is quantized using the canonical procedure , and only the basis states with one unit of excitation are included .
this , however , covers the linear optical phenomena .
the field is taken to interact with localized atoms through a dipole interaction . using different configurations of atoms and choosing their frequencies to be resonant or off - resonance , we can model mirrors , beam splitters , focusing devices and multicomponent systems .
thus we can model arbitrary linear networks of optical components .
we show the time evolution of a photon wave packet in an interferometer as an example .
as the state of the field is known at each instant , spectral properties and spatial coherence can immediately be obtained from the simulations .
we also know the states of the two level atoms constituting the components , which allows us to consider their quantum behavior . here
the decay of an excited atom into the vacuum state of the electromagnetic field in the two - dimensional cavity is studied . 2 |
the relationship between unconventional superconductivity and magnetism in heavy - fermion systems and transition metal oxides is one of the most interesting research areas in condensed matter physics . in both cases
it is widely believed that the magnetic degrees of freedom play an essential role in the formation of superconductivity .
furthermore , unconventional superconductivity yields strong feedback on the magnetic spin excitations in these systems below the superconducting transition temperature @xmath4 .
one example is the famous so - called resonance peak observed in high-@xmath4 cuprates by means of inelastic neutron scattering ( ins ) @xcite whose nature is still actively debated.@xcite remarkably , it has been found that ins reveals the formation of a new magnetic mode in the superconducting state of the uranium based heavy - fermion compound upd@xmath0al@xmath1 with @xmath4 = 1.8 k.@xcite its sharply peaked intensity , its temperature dependence and the energy position well below 2@xmath2 ( with @xmath2 being the maximum of the superconducting gap ) strongly resembles the resonance peak seen in high-@xmath4 cuprates .
this is particularly remarkable , since the origin of superconductivity in cuprates and upd@xmath0al@xmath1 seems to be different .
while frequently discussed scenarios in cuprates are a spin - fluctuation mediated cooper pairing or the electron - phonon interaction , in upd@xmath0al@xmath1 a magnetic - exciton mediated pairing model has been proposed@xcite based on available experiments .
the latter model is built on the dual nature of the 5@xmath5 electrons .
it consists of localized 5@xmath6 crystalline electric field ( cef ) states which disperse into a magnetic exciton band due to intersite interactions and a conduction electron band @xcite formed by itinerant @xmath7 electrons with enhanced hybridization .
the model successfully explains the formation of unconventional superconductivity in this compound@xcite based on the virtual exchange of the magnetic excitons between itinerant quasiparticles .
it is important to note that the resonant spin excitations in superconducting cuprates can be seen as a direct consequence of the @xmath8-wave symmetry of the superconducting order parameter .
namely , the resonance peak occurs only if the order parameter changes sign in the first brillouin zone ( bz).@xcite thus , ins can be considered as a bulk probe for the unconventional nature of superconductivity in these compounds .
therefore it is important to search for such an effect in other unconventional superconductors as well . in this paper
we analyze the consequences of the unconventional pairing on the magnetic excitations in upd@xmath0al@xmath1 .
we show that in addition to the magnetic exciton dispersion present in the normal state , unconventional superconductivity induces the formation of a bound state below @xmath4 with an associated resonance peak in the magnetic spectrum at the antiferromagnetic ( af ) wave vector * q*=@xmath3 where @xmath9 is a lattice constant along the crystallographic @xmath10 axis .
its frequency is well below 2@xmath2 and in good agreement with experimental data .
we show that similar to cuprates the resonance peak in upd@xmath0al@xmath1 is a consequence of an unconventional superconducting order parameter which changes sign at regions of the fermi surface connected by the antiferromagnetic wave vector * q*. we analyze the influence of antiferromagnetism on the formation of the resonance peak and surprisingly find that its intensity is enhanced due to the reconstruction of the fermi surface .
we find that the dispersion of the resonance peak away from * q * is controlled by the momentum dependence of excitations of the localized magnetic moment ( magnetic exciton ) .
the resonance peak in upd@xmath0al@xmath1 has also been studied by bernhoeft _ _
e__t al.@xcite within a phenomenological two component spin susceptibility model
. however , in our microscopic calculations we show that antiferromagnetic order plays a crucial role in the formation of the resonance peak below @xmath4 .
following previous consideration by mchale _ _ e__t al .
@xcite we use the low - energy hamiltonian describing the interaction of the itinerant @xmath5 electrons and magnetic excitons originating from localized 5@xmath6 crystalline electric field ( cef ) states : @xmath11 where @xmath12 , and @xmath13 mev is the energy gap of the 5@xmath6 electrons between the ground and first excited states in the crystalline electric field .
the dispersion of the magnetic excitons is approximately described by @xmath14 $ ] with @xmath15 , where @xmath16 is the coupling constant between the itinerant electrons and the localized magnetic moments .
we adopt parameter values @xmath17 = 5.5 mev and @xmath18 .
note , here we follow ref . in assuming that only the @xmath19 component of the conduction electrons can excite magnetic excitons .
therefore the spin - space isotropy is broken in a maximal ( ising ) way . as a result
the usual classification of cooper pairs into spin - triplet and spin - singlet states is not valid and the notation _ _ e__qual and opposite spin pairing states should be better used instead .
however , we will still speak of singlet and triplet cooper - pairing states as commonly done .
( [ eq : hamilt ] ) gives rise to fermionic and bosonic self - energies and is particularly relevant for electron - hole states separated by the antiferromagnetic wave vector * q*=(0,0,@xmath20 ) .
previously it has been shown that this interaction explains superconductivity in upd@xmath0al@xmath1.@xcite we define the electron and magnetic exciton green s functions as follows : @xmath21 where @xmath22 and @xmath23 is essentially the pseudospin susceptibility .
the bare magnetic exciton green s function is given by @xmath24 .
due to the interaction of the magnetic excitons with conducting electrons , the feedback effect on the former results in @xmath25 where the magnetic exciton self - energy is given by @xmath26 here , the spin susceptibility of the conduction electrons in the superconducting state is the fermi surface of the conducting electrons is almost like a cylinder with weak dispersion along the @xmath10 direction . neglecting the anisotropy of dispersion in the plane yields @xmath32 ( @xmath33 , @xmath34 ) and @xmath35 . here
, we approximate the hexagonal unit cell by a circle with radius @xmath36 chosen so that the hexagon and the circle have the same area . furthermore , we assume a parabolic dispersion in the plane .
due to the ising - type anisotropy of the interaction between conduction electrons and magnetic excitons it has been previously found@xcite that both pure paramagnetic , _
, spin - singlet states @xmath37 with @xmath38 momentum dependence and spin - triplet states ( @xmath39 ) @xmath40 with @xmath41 have the highest ( degenerate ) superconducting transition temperature .
thus , in the following we will consider the two superconducting order parameters , @xmath42 and @xmath43 as the most relevant ones in this model .
let us now discuss the consequences of the behavior of im@xmath44 for the magnetic exciton dispersion which follows from eq .
( [ eq : exciton ] ) .
the dispersion of the magnetic exciton in the presence of coupling to the conduction electrons is given by @xmath45 in the normal paramagnetic state , re@xmath44 is a constant at low frequencies determined in our case by the curvature of the fermi surface along the @xmath46 direction . at the same time
im@xmath47 , where @xmath48 is a landau damping constant .
thus , the bare magnetic exciton acquires a linewidth and renormalizes slightly by a certain constant which changes the original position of @xmath49 downwards . in the superconducting state
the renormalization is strongly dependent on the gap symmetry .
for example , in the conventional @xmath50-wave state re@xmath44 is less than its normal state value .
first , the @xmath50-wave superconducting gap gives a negative contribution to @xmath51 as follows from the coherence factor in eq .
( [ sust0 ] ) , and second it yields the spin excitation gap structure in im@xmath44 as follows from the @xmath52 function .
therefore , the feedback of the conduction electrons on the magnetic exciton becomes weaker and yields a shift of the magnetic exciton dispersion towards higher frequencies in the superconducting state with respect to its normal state value . at the same time , the unconventional character of the superconducting gap entails immediate consequences for eqs .
( [ sust0 ] ) and ( [ eq : pole ] ) .
namely , at the fermi surface ( i.e. , for @xmath53 ) one finds @xmath54 as well as @xmath55 where * q * is the antiferromagnetic wave vector .
thus , in both cases the anomalous coherence factor equals @xmath56 for all @xmath46 momenta which is in strong contrast to the usual @xmath50-wave symmetry of the superconducting gap .
simultaneously , the @xmath52 function in eq .
( [ sust0 ] ) starts to contribute for @xmath57 ^ 2 + ( |\delta_{k_z}|+|\delta _ { k_z+q_z}|)^2}$ ] which is determined by the value of the gap , @xmath58 and the quasiparticle dispersion along @xmath46 axis , @xmath59 .
note that @xmath60 and @xmath61 for the singlet and triplet superconducting gap , respectively .
it is interesting to note that for the singlet case there is no gap in im@xmath44 and therefore , it is nearly the same as in the normal state@xcite [ see fig .
[ fig1](a ) ] .
then re@xmath44 is a constant at low frequencies . therefore the magnetic exciton dispersion will be simply shifted in proportion to the total value of re@xmath44 exactly as in the normal state .
correspondingly , no strong feedback on the magnetic exciton due to superconductivity takes place .
. then due to a combined effect of the anomalous coherence factor and the @xmath52 function , a discontinuous jump in im@xmath44 occurs at about @xmath64 for the triplet order parameter . via kramers - kronig transformation
the discontinuous jump in im@xmath44 yields a logarithmic singularity in re@xmath44 .
note , the logarithmic singularity in fig .
1.(b ) is suppressed by a weak damping .
furthermore , below @xmath65 the re@xmath44 is increased with respect to its normal state value for @xmath66 as shown in fig .
[ fig1](b ) .
a frequency dependence of the re@xmath67 can yield more than one solution of eq .
( 7 ) . in order to demonstrate
how those solutions can be found above and below t@xmath68 we illustrate in fig .
2 the possible characteristic behavior of @xmath69 . here
, we assume that in normal state the re@xmath44 is almost frequency independent and magnetic exciton s peak position shifts slightly in superconducting state with respect to its normal state value . due to the gap structure in superconducting state , depending on the value of @xmath16 , a new pole may occur at energies less than @xmath70 . if im@xmath44 is small or zero at these frequencies , the total im@xmath71 , i.e. , the spectral function of magnetic excitations , shows a resonance peak which occurs only in the superconducting state .
this agrees well with experimental ins data@xcite .
moreover , at higher energies one finds in addition two more poles in eq .
( [ eq : pole ] ) .
the latter yields an additional structure in im@xmath71 which is a renormalized magnetic exciton with finite damping .
this typical behavior of the susceptibility can be found in fig .
so far we have ignored the coexistence of antiferromagnetism and superconductivity in upd@xmath0al@xmath1 for the conduction electrons.@xcite antiferromagnetic order results in upd@xmath0al@xmath1 due to the interaction of neighboring uranium ions .
this leads to a dispersion for exciton state and eventually to an antiferromagnetic instability and a new ground state.@xcite the unit cell is doubled and the brillouine zone is correspondingly reduced . the new dispersion of the conducting electrons enters the expression for the spin susceptibility .
then the solutions of eq .
( 7 ) must be redetermined .
this is done in the following .
the total hamiltonian is @xmath72 where @xmath73 denotes the value of the effective antiferromagnetic staggered field .
this term leads to a splitting of the quasiparticle energy dispersion into two bands.@xcite in particular , the hamiltonian ( [ eq : totalh ] ) can be easily diagonalized by a unitary transformation@xcite and the resulting energy dispersions are @xmath74 with @xmath75 $ ] . here , we have introduced @xmath76 and @xmath77 .
the af fermi surface of the two bands @xmath78 consists of two disjoint cylinders in the reduced af bz @xmath79 ( see ref . ) . in a pure af state , the real part of the susceptibility @xmath51 at low energies
is determined by the intraband processes and can be approximated by @xmath80 where @xmath81 is the af coherence factor . at wave vector * q * for @xmath82 the susceptibility is proportional to the density of states and decreases rapidly to zero as one increases frequency .
this is a consequence of the equality @xmath83 .
therefore the renormalization of the magnetic excitons due to conduction electrons can be safely ignored .
most importantly , in the superconducting state coexisting with af , the imaginary part of the spin susceptibility of the conduction electrons including intraband and interband scattering is given for @xmath84=0 k , @xmath85 , and @xmath86 by here the first and the third terms describe the intraband quasiparticle pair creation while the second term refers to the corresponding interband process .
note that eq .
( [ eq : susaf ] ) contains two types of coherence factors , _ i_.e .
, due to superconductivity and antiferromagnetic order , respectively . as usual ,
the low - energy behavior of im@xmath44 is dominated by the intraband contributions .
we also assume that the presence of antiferromagnetism does not change the qualitative behavior of the superconducting gap , _ i_.e .
, the position of the line node and the corresponding change of sign of the superconducting order parameter remain the same although some higher harmonics may appear.@xcite as in eq .
( [ sust0 ] ) , the superconducting coherence factors equal 2 for @xmath46 momenta close to the fermi surface . at the same time , the reconstructed conduction bands in the af state have only one - half the original period , i.e. , @xmath88 .
therefore all parts of the fermi surface can be connected by the antiferromagnetic wave vector * q * and simultaneously have a sign change @xmath89 of the gap except at the nodal points . as a result
, im@xmath90 is non - zero for @xmath91 due to the contribution of the nodal states both in the singlet and the triplet cooper - pairing cases . with increasing frequency im@xmath44 increases up to energies of about @xmath70 and then decreases
[ see fig .
[ fig3](a ) ] .
the functional dependence of im@xmath44 at low frequencies resembles the behavior of the density of states except that the structure occurs at around @xmath70 .
correspondingly , the real part of @xmath92 is the same as in a pure af state .
however , away from @xmath82 it does not drop as in the pure af state but increases quadratically up to about 2@xmath2 due to the structure of im@xmath44 induced by the superconducting gap . only then does re@xmath44 drop to small values . altogether re@xmath93 increases in the superconducting state for @xmath94 due to the unconventional nature of the superconducting order parameter .
however , the pure resonance ( bound state ) in im@xmath71 is not realized due to finite damping .
an additional pole in im@xmath71 still exists at frequencies smaller than 2@xmath2 due to a strong increase of re@xmath44 at small frequencies as shown in fig .
[ fig3](b ) . at higher frequencies
it becomes small and thus another pole appears corresponding to the broadened original magnetic exciton .
thus , im@xmath71 has a two - pole structure as shown in fig .
[ fig3](b ) . note , in order to increase the intensity of the low - energy pole in im@xmath71 we include the contribution of the higher harmonics to the gap function .
as already mentioned , they are due to the presence of af order @xcite . finally we discuss the dispersion of the magnetic excitations away from * q * along the @xmath95 direction . in fig .
[ fig4 ] we show the calculated momentum and frequency dependencies of im@xmath96 .
it is clear that as soon as @xmath97 the original magnetic exciton has a strong upward dispersion in the normal state .
therefore , effects connected with renormalization induced by superconductivity will also be shifted towards higher energies . the pole induced by superconductivity shows dispersion similar to the magnetic exciton
[ see fig .
[ fig4 ] ] . for
both singlet and triplet order parameters our results are in fair agreement with recent ins data.@xcite namely , in the superconducting state one finds two distinct energy dispersions , one being the resonancelike feature with high intensity inside the superconducing gap and the second , that of localized magnetic excitons renormalized by the conducting electrons .
another interesting point worth noting is that due to the doubling of the unit cell and the equality @xmath98 , the effect of the @xmath99 and @xmath38 gaps leads to a very similar behavior for im@xmath44 .
the slight difference in the absolute magnitude arises from the different densities of states at those regions of the fermi surface where the maximum of the singlet and the triplet gaps occur .
altogether this does not change the functional form of im@xmath44 .
we mention that thermal conductivity results in a rotating field @xcite are compatible with both @xmath38 and @xmath99 order parameters while the observed knight shift @xcite seems to favor the former .
interestingly , we also found that a recently proposed superconducting gap with @xmath100 symmetry @xcite does not lead to the formation of low - energy spin excitations around wave vector * q * in the superconducting gap .
the reason is that its momentum dependence yields no change of the sign of the superconducting order parameter , @xmath101 , and thus no constructive contribution can result from the anomalous coherence factor . in conclusion
, we have investigated the effects of superconductivity on the magnetic excitations in the unconventional superconductor upd@xmath0al@xmath1 .
in particular , due to the change in sign of the superconducting order parameter the conduction electron susceptibility is enhanced in the superconducting state which yields an additional pole ( bound state ) in the total susceptibility .
we further analyzed the role played by antiferromagnetism and found that its presence increases the spectral weight of the resonance due to the doubling of the unit cell .
however , the resonance peak in the af phase becomes a virtual bound state due to finite damping .
finally we point out that upd@xmath0al@xmath1 is another known example where the unconventional nature of the superconducting order parameter yields a structure in the magnetic susceptibility as in layered high-@xmath4 cuprates
. therefore it can be regarded as a model system of unconventional superconductivity studied by inelastic neutron scattering .
99 j. rossat - mignod , l.p .
regnault , c. vettier , p. bourges , p. burlet , and j. bossy , physica c * * 1**85 - 189 , 86 ( 1991 ) .
see for review p.a .
lee , n. nagaosa , and x .-
wen , rev .
* 7**8 , 17 ( 2006 ) .
n.k . sato , n. aso , k. miyake , r. shiina , p. thalmeier , g. varelogiannis , c. geibel , f. steglich , p. fulde , and t. komatsubara , nature * * 4**10 , 340 ( 2001 ) . g. zwicknagl , a. yaresko and p. fulde , phys .
b * * 6**8 , 052508 ( 2003 ) ; g. zwicknagl , a.n . yaresko and p. fulde,__i__bid . * * 6**5 , 081103(r ) ( 2002 ) .
p. mchale , p. fulde , and p. thalmeier , phys .
b * * 7**0 , 014513 ( 2004 ) . j. brinckmann and p. a. lee , phys . rev .
lett . * * 8**2 , 2915 ( 1999 ) ; y .- j .
kao , q. si , and k. levin , phys .
b * * 6**1 , r11898 ( 2000 ) ; m. r. norman , _ _ i__bid . * * 6**1 , 14751 ( 2000 ) , * * 6**3 , 092509 ( 2001 ) ; f. onufrieva and p. pfeuty , _ _ i__bid . * * 6**5 , 054515 ( 2002 ) ; i. eremin , d. k. morr , a. v. chubukov , k. bennemann , and m. r. norman , phys . rev . lett . * * 9**4 , 147001 ( 2005 ) . n. bernhoeft , a.hiess , n. metoki , g.h .
lander , and b. roessli , j. phys . : condens .
matter * * 1**8 , 5961 ( 2006 ) ; n. bernhoeft , eur .
j. b * * 1**3 , 29 ( 2000 ) .
a. krimmel , p. fischer , b. roessli , h. maletta , c. geibel , c. schank , a , grauel , a. loidl , and f. steglich , z. phys .
b * * 8**6 , 161 ( 1992 ) .
p. thalmeier , eur .
j. b * * 2**7 , 29 ( 2002 ) .
y. inada , h. yamagami , y. haga , k. sakurai , y. tokigawa , t. honma , e. yamamoto , y. onuki , and t. yanagisawa , j. phys .
. jpn . * * 6**8 , 3643 ( 1999 ) .
a. amici , p. thalmeier , and p. fulde phys .
lett . * * 8**4 , 1800 ( 2000 ) . j .-
ismer , i. eremin , and d.k .
morr , phys .
b * * 7**3 , 104519 ( 2006 ) .
a. hiess , n. bernhoeft , n. metoki , g.h .
lander , b. roessli , n.k .
sato , n. aso , y. haga , y. koike , t. komatsubara , and y. onuki , j. phys .
: condens .
matter * * 1**8 , r437 ( 2006 ) . t. watanabe , k. izawa , y. kasahara , y. haga , y. onuki , p. thalmeier , k. maki , and y. matsuda , phys . rev .
b * * 7**0 , 184502 ( 2004 ) y. kitaoka , h. tou , k. ishida , n , kimura , y. onuki , e. yamamoto , y. haga , and k. maezawa , physica b * * 2**81 - 282 , 878 ( 2000 ) h. won , d. parker , k. maki , t. watanabe , k. izawa , and y. matsuda , phys .
b * * 7**0 , 140509(r ) ( 2004 ) . | we analyze the influence of unconventional superconductivity on the magnetic excitations in the heavy fermion compound upd@xmath0al@xmath1 .
we show that it leads to the formation of a bound state at energies well below 2@xmath2 at the antiferromagnetic wave vector * q*=@xmath3 .
its signature is a resonance peak in the spectrum of magnetic excitations in good agreement with results from inelastic neutron scattering .
furthermore we investigate the influence of antiferromagnetic order on the formation of the resonance peak .
we find that its intensity is enhanced due to intraband transitions induced by the reconstruction of fermi surface sheets .
we determine the dispersion of the resonance peak near * q * and show that it is dominated by the magnetic exciton dispersion associated with local moments .
we demonstrate by a microscopic calculation that upd@xmath0al@xmath1 is another example in which the unconventional nature of the superconducting order parameter can be probed by means of inelastic neutron scattering and determined unambiguously . |
the understanding of low - energy electronic structures and excitations in real materials is an important subject of condensed - matter physics and material science .
interesting phenomena such as a non - fermi - liquid behavior and unconventional superconductivity are caused by the instability of electronic structures near the fermi level .
a common feature is often found in the band structures showing such phenomena ; isolated bands appear near the fermi level .
the width of these isolated bands is typically the order of 1 ev , which is comparable to local electronic interactions .
thus , in these isolated bands , the kinetic and potential energies compete with each other , and the competition is often discussed within a local - interaction approximation , as in the hubbard model . in real materials , however , there exist various elementally excitations not described by the local electronic interaction .
the plasmon in metallic systems or the exciton in insulating systems are well known examples of such nonlocal excitations , which result from the long - range coulomb interaction . in the above - mentioned isolated - band systems
, the plasmon excitation can occur in this band , and its energy scale can be very small ( of the order of 1 ev ) , which is comparable to the bandwidth and the size of the local coulomb interaction . in this study , we investigate the effect of the low - energy - plasmon fluctuation on the electronic structure of real isolated - band systems from first principles . for this purpose
, we choose two materials , a quasi - one dimensional organic conductor ( tmtsf)@xmath1pf@xmath2 ( ref . ) , where tmtsf stands for tetramethyltetraselenafulvalene , and a three - dimensional perovskite transition - metal oxide srvo@xmath3 ( ref . ) .
these materials are typical isolated - band systems and are studied as benchmark materials of the correlated metal , where the local coulomb - interaction effect on the electronic properties is investigated with much interest .
@xcite in the present work , we focus on the low - energy plasmon effect on the electronic structure .
@xcite through the comparison between theoretical and experimental results on the plasmon - related properties and spectral functions , we verify low - energy plasmon effects on the electronic structure of the real system . the organic conductor ( tmtsf)@xmath1pf@xmath2 is a representative quasi - one - dimensional material , @xcite and basically behaves as a good metallic conductor .
@xcite at low temperature ( around 12 k ) , it undergoes a transition to a spin - density - wave phase .
@xcite in the high - temperature metallic region , photoemission spectroscopy has observed small spectral weight near the fermi level .
@xcite from this observation and the quasi - one - dimensional nature , the origin of the renormalization has been discussed in view of the tomonaga - luttinger liquid .
@xcite on the other hand , this material exhibits clear low - energy plasma edges around 0.1 - 1 ev in the reflectance spectra .
@xcite therefore , this plasmon excitation would also be a prominent renormalization factor of the electronic structure .
the transition - metal oxide srvo@xmath3 is another well known correlated metal .
@xcite many high - resolution photoemission measurements @xcite including the bulk - sensitive version @xcite were performed and clarified a strong renormalization of the low - energy isolated band and the satellite peak just below this band .
the origin has actively been discussed in terms of the local electronic correlation .
@xcite on the other hand , the reflectance and electron - energy - loss spectra have clarified low - energy plasmon excitations around 1.4 ev .
@xcite in density - functional band structure , srvo@xmath3 has isolated bands of the @xmath4 orbitals around the fermi level , whose bandwidth is about 2.7 ev .
@xcite also , the constrained random phase approximation gives an estimate of the local electronic interaction of @xmath52 - 3 ev .
@xcite thus , the energy scale of the experimentally observed plasmon excitation is comparable to the bandwidth and the local coulomb interaction .
hence , the low - energy plasmon fluctuation would certainly be relevant to the low - energy properties . to study how the plasmon excitation affects the electronic structure , we perform _ ab initio _ calculations based on the @xmath0 approximation .
@xcite the @xmath0 calculation considers the self - energy effect due to the plasmon fluctuation and describes properly quasiparticle energies in the valence region . on the other hand ,
the description for the plasmon satellite in the spectral function is known to be less accurate .
@xcite in order to improve this deficiency , _ ab initio _
@xmath0 plus cumulant ( @xmath0+@xmath6 ) calculations have recently been performed .
@xcite the accuracy of the @xmath0+@xmath6 method has been verified in bulk silicon @xcite and simple metals , @xcite where the satellite property is satisfactorily improved . in the present study
, we apply the @xmath0+@xmath6 method to the study of the above mentioned isolated - band systems , and show that the low - energy plasmon fluctuations modify substantially the low - energy electronic structure .
the present paper is organized as follows . in sec .
ii , we describe the @xmath0 and @xmath0+@xmath6 methods to calculate dielectric and spectral properties .
computational details and results for ( tmtsf)@xmath1pf@xmath2 and srvo@xmath3 are given in sec .
we also discuss the comparison between theory and experiment , focusing on the renormalization of the electronic structure due to the plasmon excitation .
the summary is given in sec .
in this section , we describe _ ab initio _ @xmath0 and @xmath7 methods .
the latter is a post-@xmath0 treatment and uses the self - energies calculated with the @xmath0 approximation .
below , we first describe details of the @xmath0 calculation . the non - interacting green s function is written as @xmath8 where @xmath9 and @xmath10 are wavevector and frequency , respectively . @xmath11 and @xmath12 are the kohn - sham ( ks ) wavefunction and its energy , respectively , and @xmath13 is the fermi level of the ks system .
the @xmath14 parameter is chosen to be a small positive value to stabilize numerical calculations .
a polarization function of a type @xmath15 is written in a matrix form in the plane wave basis as @xmath16 with @xmath17 and @xmath18 where @xmath19 is a reciprocal lattice vector . with this polarization function
, the symmetrized dielectric matrix in reciprocal space is defined by @xmath20 with @xmath21 being the crystal volume . in the @xmath22 limit , the dielectric matrix in eq .
( [ symeps ] ) is expressed by @xcite @xmath23 \displaystyle -\frac{4\pi}{v } 2\sum_{{\bf k}}\sum^{unocc}_{\alpha}\sum^{occ}_{\beta } \biggl(\frac{p^{\mu}_{\alpha\beta}({\bf k } ) } { \epsilon_{\alpha{\bf k}}-\epsilon_{\beta{\bf k}}}\biggr ) \frac{m_{\alpha\beta}^{{\bf g'}}({\bf k , k})^*}{|{\bf g'}| } x_{\alpha{\bf k},\beta{\bf k}}(\omega ) & \mbox{(${\bf g=0},{\bf g'\ne 0}$ ) , } \\[+20pt ] \displaystyle -\frac{4\pi}{v } 2\sum_{{\bf k}}\sum^{unocc}_{\alpha}\sum^{occ}_{\beta } \frac{m_{\alpha\beta}^{{\bf g}}({\bf k , k})}{|{\bf g}| } \biggl(\frac{p^{\mu}_{\alpha\beta}({\bf k } ) } { \epsilon_{\alpha{\bf k}}-\epsilon_{\beta{\bf k}}}\biggr)^ * x_{\alpha{\bf k},\beta{\bf k}}(\omega ) & \mbox{($ { \bf g \ne 0 } , { \bf g'=0}$ ) , } \\[+20pt ] \displaystyle \delta_{{\bf gg'}}-\frac{4\pi}{v } \frac{1}{|{\bf g}| } \chi_{{\bf gg'}}({\bf 0},\omega)\frac{1}{|{\bf g'}| } & \mbox{($otherwise$ ) } \end{array } \right .
\label{eq0 } \end{aligned}\ ] ] where @xmath24 approaches zero along the cartesian @xmath13 direction and @xmath25 is a matrix element of a momentum as @xmath26|\phi_{\beta{\bf k}}\rangle , \label{p_ij } \end{aligned}\ ] ] with @xmath27 being the nonlocal part of the pseudopotential . in the first line of eq .
( [ eq0 ] ) , the last term on the right hand side is the drude term of the intraband transitions around the fermi level , @xcite where @xmath28 is the bare plasma frequency .
the other terms in eq .
( [ eq0 ] ) result from the interband transitions .
we next describe the calculation of the self - energy .
the operator of the exchange self - energy is defined by @xmath29 where @xmath30 is the bare coulomb interaction . in practice
, we use an attenuation coulomb interaction @xmath31 with a cutoff @xmath32 instead of @xmath33 to treat the integrable singularities in the bare coulomb interaction .
@xcite the matrix element of the exchange self - energy is thus @xmath34 where @xmath35 and @xmath36 are the `` body '' and `` head '' components of the exchange self - energy , respectively .
the former body matrix element is expressed as @xmath37 with @xmath38 the prime in the sum of eq .
( [ sx ] ) represents the summation excluding the contribution of the head term of @xmath39 . corresponding to the replacement of @xmath33 with @xmath40 ,
the related fourier transform is modified from @xmath41 to @xmath42 in eq .
( [ sx ] ) .
accordingly , the body matrix element in eq .
( [ sx ] ) is supplemented by the head component in eq .
( [ sxb+sxh ] ) @xmath43 the operator of the correlation self - energy is defined by @xmath44 where @xmath45 is the correlation part of the symmetrized screened coulomb interaction @xmath46 . here , @xmath47 is the inverse dielectric function , which is calculated by inverting the symmetrized dielectric matrix in eqs .
( [ symeps ] ) and ( [ eq0 ] ) . for a practical calculation of the matrix element of @xmath48 in eq .
( [ gw_sigma ] ) , we introduce the following model screened interaction @xcite @xmath49 with @xmath50 in eq .
( [ modelw ] ) , the real frequency @xmath10 is discretized into @xmath51 , and @xmath52 and @xmath53 are the pole and amplitude of the model interactions , respectively .
the matrix element @xmath54 comprises a square matrix [ see sec .
iii(a ) ] .
since the frequency - dependent part @xmath55 is decoupled from the amplitude one , the frequency integral in @xmath56 can be analytically performed .
the matrix element of @xmath57 consists of the body and head components as follows : @xmath58 the body matrix element in the above is given by @xmath59 where the numerator is given by @xmath60 with @xmath61 note that , in the practical calculation , the frequency @xmath10 for the self - energy in eq .
( [ eq : sgm ] ) is distinguished from the frequency @xmath51 for the screened interaction in eq .
( [ modelw ] ) .
the body matrix element @xmath62 in eq .
( [ eq : sgm ] ) is supplemented with the head component in eq .
( [ scb+sch ] ) @xmath63 with @xmath64 with these ingredients , the spectral function is calculated via the wannier - interpolation method ( see below ) .
the spectral function at an arbitrary @xmath9 is @xmath65 where @xmath66 is obtained by diagonalizing non - symmetric complex matrix in the wannier basis @xmath67 where @xmath68 is the fourier transform of the ks hamiltonian matrix in the wannier basis as @xmath69 with @xmath70 here , @xmath71 is a @xmath72 point in the regular mesh and @xmath73 is the total number of the @xmath72 points in the regular mesh . also , @xmath74 is the @xmath75th wannier orbital at the lattice point @xmath76 , and the transform @xmath77 is obtained in the wannier - function - generation routine .
@xmath78 in eq .
( [ hij ] ) is the fourier transform of the self - energy in the wannier basis as @xmath79 with @xmath80 the matrix element @xmath81 is defined by @xmath82 with @xmath83 being the exchange - correlation potential in density - functional theory . in eq .
( [ akw ] ) , the energy shift @xmath84 is introduced to correct the mismatch of the fermi level between the ks and one - shot @xmath0 systems .
this parameter is determined from the equation on the spectral norm @xmath85 where @xmath86 is the total number of electrons in the system and @xmath87 is the total number of sampling @xmath72 points after the interpolation .
note that @xmath13 is set to the fermi level for the ks system .
the flow of the calculation is as follows : we first perform density - functional calculations to obtain the band structures and the wannier functions for bands associated with the self - energy calculations .
we then calculate the self - energies for the irreducible @xmath72-points @xmath88 in a regular mesh , including band off - diagonal terms as @xmath89 in the selected energy region .
the self - energies at a @xmath90 point symmetrically equivalent to @xmath91 are the same as @xmath92 , but , in the time - reversal symmetry case , @xmath81 is obtained by @xmath93 .
then , we transform @xmath81 in eq .
( [ sigma - od ] ) to the wannier representation @xmath94 with eq .
( [ sijrw ] ) . with these data ,
we evaluate the self - energy at an arbitrary @xmath95 via eq .
( [ sijkw ] ) .
note that the calculated spectral function includes band - off - diagonal effects , which is discussed in appendix a. finally , we calculate the spectral function of eq .
( [ akw ] ) considering the energy shift @xmath84 in eqs .
( [ akw ] ) and ( [ norm ] ) .
the @xmath7 approach is a theory beyond the @xmath0 approximation , @xcite which is based on systematic diagrammatic expansions .
this approach is suitable for dealing with long - range correlations , i.e. , various types of the plasmon - fluctuation diagrams not included in the usual @xmath0 diagram . in the initial stage of the study , it was applied to a system of core electrons interacting with a plasmon field .
@xcite currently ,
_ ab initio _
@xmath0+@xmath6 calculations have been known to give a better description for satellites due to the plasmon excitation .
@xcite the green s function with the cumulant expansion is defined in the time domain by @xcite @xmath96 where @xmath97 for the first term on the right hand side and @xmath98 for the second term .
the @xmath99 and @xmath100 are the cumulants for the hole and particle states , respectively .
the spectral function is calculated by the fourier transform as @xmath101 which consists of the hole @xmath102 and particle @xmath103 contributions .
the spectral function for the hole part is written as @xmath104 where the band sum is taken over the occupied states . to the lowest order in the screened interaction @xmath105
, the cumulant is obtained by @xcite @xmath106 where @xmath107 . in the present study
, the cumulant is expanded around the quasiparticle energy @xmath108 , @xcite which is a solution of @xmath109 where @xmath84 is an energy shift to correct a mismatch of the fermi level between the ks and one - shot @xmath0+@xmath6 systems . by considering the fourier transform of @xmath110 in eq .
( [ chdef ] ) and the spectral representation of @xmath111 , and after some manipulation , @xcite the expression of the cumulant is obtained , which consists of the quasiparticle and satellite parts as @xmath112 with @xmath113 and @xmath114 note that @xmath115 is negative for the hole part . to show the expansion point explicitly , we add @xmath108 as an index in the cumulant eqs .
( [ ctoth ] ) , ( [ cqph ] ) , and ( [ csh ] ) . within the one - shot calculation ,
the position of the cumulant expansion may be taken at the non - interacting energy @xmath12 .
@xcite by taking the expansion point at @xmath108 , the results may include some sort of the self - consistency effect . the hole self - energy in eq .
( [ cqph ] ) is defined by @xmath116 it should be noted that the derivative @xmath117 in eq .
( [ cqph ] ) is related to the @xmath115=0 component of the satellite cumulant in eq .
( [ csh ] ) as @xcite @xmath118 this expression is practically used for the evaluation of the derivative ; we avoid the numerical calculation of the derivative by a finite difference and use the integral expression on the right hand side of eq .
( [ del - eq - c ] ) for the derivative , since we find that the latter treatment is numerically more stable . the stable calculation of the derivative is important in keeping the sum rule on the @xmath7 spectrum . in the practical calculation
, we divide @xmath119 in eq .
( [ ahkwgwc ] ) into the two parts for stable calculations as @xcite @xmath120 the first quasiparticle term on the right hand side can be calculated analytically as @xmath121 where @xmath122 the latter convolution term in eq .
( [ ahkw ] ) is calculated via the numerical integration as @xmath123 where the integrand in the right - hand side decays to zero rapidly . the particle part of the spectral function in eq .
( [ akwgwc ] ) is given by @xmath124 with @xmath125 in this case @xmath115 is positive .
the band sum in eq .
( [ apkwgwc ] ) runs over the unoccupied states .
the @xmath126 and @xmath127 are given by @xmath128 and @xmath129 respectively .
the @xmath130 in eq .
( [ cqpp ] ) is the particle self - energy as @xmath131 note that the self - energies in the above equations are defined as the causal one , so that the imaginary part of the self - energy is positive for @xmath132 and negative for @xmath133 . in the particle part
, @xmath134 holds similarly to eq .
( [ del - eq - c ] ) . the contribution from the quasiparticle part to @xmath135 is @xmath136 with @xmath137 and the convolution - part contribution is @xmath138
density - functional calculations were performed with @xcite with plane - wave basis sets , where we employed norm - conserving pseudopotentials @xcite and generalized gradient approximation ( gga ) for the exchange - correlation potential .
@xcite maximally localized wannier functions @xcite were used for the interpolation of the self - energy . for the atomic coordinates of ( tmtsf)@xmath1pf@xmath2 ,
the experimental structure obtained by a neutron measurement @xcite at 20 k was adopted .
the cutoff energies in wavefunction and in charge densities are 36 ry and 144 ry , respectively , and a 15@xmath13915@xmath1393 @xmath72-point sampling was employed .
the cutoff for the polarization function in eq .
( [ eq : chi ] ) was set to be 3 ry , and 200 bands were considered , which covers an energy range from the bottom of the occupied states near @xmath14030 ev to the top of the unoccupied states near 15 ev , where 0 ev is the fermi level .
the frequency grid of the polarization was taken up to @xmath141 = 86 ev in a double - logarithmic form , for which we sampled 99 energy - points for [ 0.1 ev : 43 ev ] and 10 points for [ 43 ev : 86 ev ] with initial grid set to be 0.0 ev . the @xmath72 sum over bz in eqs .
( [ eq : chi ] ) , ( [ eq0 ] ) , and ( [ wpl ] ) was evaluated by the generalized tetrahedron method .
@xcite the drude term of @xmath142 is evaluated at the slightly shifted frequency @xmath143 ( au ) .
the self - energy in eq . ( [ eq : sgm ] ) was calculated for the frequency range [ @xmath144 ev : @xmath145 ev ] , where @xmath145 was set to be 100 ev . in the practical calculation , we sampled 450 points for [ @xmath140100 ev : @xmath14010 ev ] with the interval of 0.2 ev , 1000 points for [ @xmath14010 ev : 10 ev ] with the 0.02 ev interval , and 450 points for [ 10 ev : 100 ev ] with the 0.2 ev interval . with this energy range ,
the high - frequency tail of the self - energy is sufficiently small .
this convergence is important to preserve the norm of the spectral function . for srvo@xmath3
, band calculations were performed for the idealized simple cubic structure , where the lattice parameter was set to be @xmath146=3.84 .
the cutoff energies for wavefunction and charge densities are 49 ry and 196 ry , respectively , and an 11@xmath13911@xmath13911 @xmath72-point sampling was employed .
the cutoff energy for the polarization function was set to be 10 ry and 130 bands were considered , which cover from the bottom of the occupied states near @xmath14020 ev to the top of the unoccupied states near 90 ev .
the frequency range of the polarization function was taken to be @xmath147 ev , where finer logarithmic sampling was applied to [ 0.1 ev : 110 ev ] with 189 points and coarser sampling was done for [ 110 ev : 220 ev ] with 10 points , with initial grid set to be 0.0 ev . in the self - energy calculation , @xmath145 was set to be 200 ev , thus , the frequency dependence of the self - energy was calculated for [ @xmath140200 ev : 200 ev ] , where we sampled 200 points for [ @xmath140200 ev : @xmath14040 ev ] with the 0.8 ev interval , 1600 points for [ @xmath14040 ev : 40 ev ] with the 0.05 ev interval , and 200 points for [ 40 ev : 200 ev ] with the 0.8 ev interval . in the fitting of the model screened interaction [ eqs .
( [ modelw ] ) and ( [ pole - z ] ) ] , the positions of the poles in the model interaction are set as follows : @xcite @xmath148 with @xmath149 .
the total number of \{@xmath150 } is the same as that of the frequency grid \{@xmath51 } for the polarization function .
we checked that the _ ab initio _
screened interaction @xmath151 are satisfactorily reproduced by this model function @xmath152 .
the broadening @xmath14 in eqs .
( [ eq : chi ] ) , ( [ theta ] ) , ( [ eq : sgm ] ) , ( [ csh ] ) , ( [ imsgmh ] ) , ( [ csp ] ) , and ( [ imsgmp ] ) was set to be 0.02 ev for ( tmtsf)@xmath1pf@xmath2 and 0.05 ev for srvo@xmath3 .
the value of @xmath14 is desirable to be small enough , but the lower - bound in the practical calculation is determined by the resolution of the band dispersion , depending primary on the @xmath72-mesh density . also , the cutoff @xmath32 in the attenuation potential [ eqs .
( [ sx ] ) and ( [ wc ] ) ] is 28.27 for ( tmtsf)@xmath1pf@xmath2 and 11.53 for srvo@xmath3 , respectively . for ( tmtsf)@xmath1pf@xmath2 , the shift @xmath84 in the spectral function @xmath153 in eq .
( [ akw ] ) was found to be 1.06 ev for the @xmath0 calculation and 0.90 ev for the @xmath0+@xmath6 one , respectively . for srvo@xmath3 , @xmath84 = 2.1 ev for the @xmath0 and 2.35 ev for the @xmath0+@xmath6 .
in @xmath0+@xmath6 , two numerical integrals on time and frequency appear [ eqs .
( [ ash ] ) and ( [ asp ] ) for the time integral and eqs .
( [ csh ] ) and ( [ csp ] ) for frequency integral ] , which must be treated carefully .
we performed time integrals in eq .
( [ ash ] ) numerically for the range [ @xmath154 au : 0 au ] and those in eq .
( [ asp ] ) for [ 0 au : @xmath155 au ] , where @xmath155 is 50 ( au )
. the total number of the time grid @xmath156 is 50000 , with the interval @xmath157 au .
note that @xmath158 is necessary to reproduce the norm of the spectral weight correctly .
the frequency integral in eqs .
( [ csh ] ) and ( [ csp ] ) was numerically evaluated with the simpson s formula for the interval @xmath159 divided into 21 subintervals , which is also important in obtaining the correct time dependence of the satellite cumulant . also , in the @xmath72 integration to obtain the @xmath7 density of states @xmath160 , the random @xmath72-point sampling was performed to improve the statistical average , where the wannier interpolation technique was efficiently applied . with this condition
, we obtained well converged spectra .
figure [ ggaband ] shows calculated gga band structures of ( tmtsf)@xmath1pf@xmath2 [ panel ( a ) ] and srvo@xmath3 [ ( b ) ] . in both systems , isolated bands are found around the fermi level . in ( tmtsf)@xmath1pf@xmath2 ,
the isolated bands consist of highest - occupied molecular orbital ( homo ) of two molecules in the unit cell .
[ for detailed atomic geometry of ( tmtsf)@xmath1pf@xmath2 , refer to refs . and .
] we call them the homo " bands which are shown with the green - dotted curves .
in addition , in this figure , homo@xmath1401 " and homo@xmath1402 " bands are shown by blue- and black - dotted curves , respectively .
@xcite in srvo@xmath3 , the isolated bands ( green - dotted curves ) around the fermi level are formed by the @xmath4 orbitals of the vanadium atom and bands around [ @xmath1407 ev : @xmath1402 ev ] come from the oxygen-@xmath161 orbitals ( blue - dotted curves ) .
@xcite pf@xmath2 and ( b ) srvo@xmath3 .
the fermi level is at zero energy . in the panel ( a ) ,
the green - dotted , blue - dotted , and black - dotted curves denote the homo , homo@xmath162 , and homo@xmath163 bands , respectively . in ( b ) , the green - dotted and blue - dotted curves correspond to the @xmath4 and o@xmath164 bands , respectively.,scaledwidth=35.0% ] to confirm the low - energy plasmon excitation in the above isolated - band systems , we calculated the reflectance spectra with the random - phase approximation @xmath165 where @xmath166 is obtained by inverse of the dielectric matrix in eq .
( [ eq0 ] ) .
figure [ reflectance ] ( a ) is the result for ( tmtsf)@xmath1pf@xmath2 , where dark - red and light - green colors represent the results in the light polarization of @xmath167 and @xmath168 , respectively , and the @xmath146 axis ( @xmath146@xmath169@xmath170 ) is the one - dimensional conducting axis
. the calculated results ( solid curves ) are compared with experimental results ( circles ) .
we see that the theoretical plasma edges satisfactorily reproduce the experimental ones around 0.8 ev for @xmath167 and 0.1 - 0.2 ev for @xmath168 , which indicates that the present scheme correctly captures the low - energy plasmon excitation .
the panel ( b ) shows the result for srvo@xmath3 .
we again see a reasonable agreement between the theory and experiment for the plasma edge ( 1.8 ev for theory and 1.4 ev for the experiment ) .
( solid curve ) based on the random - phase approximation and experimental one ( open circles ) of ( a ) ( tmtsf)@xmath1pf@xmath2 measured at 25 k and ( b ) srvo@xmath3 measured at room temperature .
the experimental data are taken from ref .
for ( tmtsf)@xmath1pf@xmath2 and ref . for srvo@xmath3 .
in ( tmtsf)@xmath1pf@xmath2 , the results for @xmath167 and @xmath168 are displayed by dark red and light green , respectively.,scaledwidth=40.0% ] table [ param - wwu ] summarizes parameters characterizing low - energy electronic structures of the two isolated - band systems , i.e. , the homo bands of ( tmtsf)@xmath1pf@xmath2 and the @xmath4 bands of srvo@xmath3 .
the table includes the calculated bare plasma frequency @xmath171 in eq .
( [ wpl ] ) , bandwidth @xmath105 , and effective local - interaction parameter @xmath172 with @xmath173 and @xmath21 being onsite and nearest - neighbor interactions , respectively , calculated with the constrained random phase approximation .
@xcite we note that @xmath171 is by definition different from the plasma edge in the reflectance spectra in fig .
[ reflectance ] ; the latter energies are lowered from the bare @xmath171 by the presence of the individual electronic excitations .
we see that @xmath171 has the same size as @xmath105 and @xmath172 , indicating that energy scale of the plasmon excitation would compete with those of kinetic and local electronic - interaction energies of electrons in the isolated band .
the long - range interaction and related plasmon fluctuations would clearly be important for electronic structure .
c@ c@ c@ c@ c@ c + [ -8pt ] & @xmath171 & @xmath105 & @xmath173 & @xmath21 & @xmath172 + [ 3pt ] + [ -5pt ] & 1.25 ( @xmath167 ) & & & & + & 0.20 ( @xmath168 ) & & & & + [ 4pt ] srvo@xmath3 & 3.54 & 2.55 & 3.48 & 0.79 & 2.69 + [ 2pt ] [ param - wwu ] to study effects of the low - energy plasmon excitation in the isolated bands on the electronic structure , we calculated a spectral function @xmath174 for the homo bands of ( tmtsf)@xmath1pf@xmath2
. figure [ akw - tmtsf ] displays the calculated spectra , where the panels ( a ) and ( b ) are the @xmath0 result via eq .
( [ akw ] ) and the @xmath0+@xmath6 one via eq .
( [ akwgwc ] ) , respectively . for comparison ,
the gga band structure is superposed with blue - solid curves . in the @xmath0 spectrum ,
clear incoherent peaks appear ; along the y-@xmath175 line , the spectral intensities of plasmaron states @xcite emerge about 1 ev above ( below ) the unoccupied ( occupied ) part of the homo bands .
@xcite also , along the x - m line , the spectra are more broadened and spread in the range from @xmath1401.5 to 0 ev .
interestingly , these sharp plasmaron peaks do not appear in the @xmath0+@xmath6 spectrum .
instead , the @xmath0+@xmath6 spectrum exhibits a broad incoherent structure throughout bz .
this is because the @xmath0+@xmath6 treatment further takes into account the long - range correlation effect @xcite or various types of the self - energy diagram involving the plasmon fluctuation , which is not included in the standard @xmath0 calculations . in the panel ( c ) ,
density of states @xmath176 is shown , where the @xmath0+@xmath6 , @xmath0 , and gga results are plotted by red - solid , blue - dotted , and thin - black - solid curves , respectively .
compared to the gga spectrum , the @xmath0 and @xmath7 spectra show an appreciable band renormalization around the fermi level by the plasmon excitation .
we again confirm that the distinct plasmon satellite ( plasmaron ) around @xmath1402 ev and + 1 ev in the @xmath0 spectrum disappears in the @xmath0+@xmath6 spectrum .
pf@xmath2 calculated with ( a ) @xmath0 approximation and ( b ) @xmath7 method .
blue - solid curves are the gga results .
the fermi level is at zero energy .
the colorbar is in linear scale .
( c ) comparison of the density of states among the @xmath7 ( red - solid curve ) , @xmath0 ( blue - dotted curve ) , and gga ( black - thin curve ) results . , scaledwidth=45.0% ] figure [ pes - tmtsf ] is the comparison between theoretical photo - emission spectra @xmath177 and experimental one ( green open circles ) obtained with heii radiation ( @xmath178=40.8 ev ) at 50 k ( ref . ) . thick - red - solid , blue - dotted , and thin - black - solid curves are the @xmath0+@xmath6 , @xmath0 , and gga results , respectively .
the spectra were calculated for the homo , homo@xmath162 , and homo@xmath163 bands to cover the energy range measured in the experiment .
we see that the gga spectrum around the fermi level is largely reduced in the @xmath0 and @xmath0+@xmath6 spectra by the self - energy effect due to the plasmon excitation .
the @xmath0+@xmath6 spectrum agrees with the overall profile of the experimental spectrum better than the @xmath0 result ; the broader spectrum is obtained in the @xmath7 result around @xmath1793 ev than in the @xmath0 one .
the discrepancy in the level position in this region between the theory and experiment arises probably from the level underestimation of the flat gga homo@xmath1401 band [ see fig .
[ ggaband ] ( a ) ] . , and homo@xmath163 bands . red - thick - solid , blue - dotted , and black - thin - solid curves denote the @xmath0+@xmath6 , @xmath0 , and gga results , respectively .
a lorentzian broadening of 0.02 ev is applied to the calculated spectra.,scaledwidth=40.0% ] next , we consider the low - energy plasmon - fluctuation effect on the electronic structure of the transition - metal oxide srvo@xmath3 .
figure [ akw - svo ] shows the spectral function calculated for the @xmath4 and o@xmath180 bands , where the panels ( a ) and ( b ) show the @xmath0 and @xmath0+@xmath6 results , respectively .
the gga bands are depicted with blue - solid curves .
the low - energy plasmon satellite emerges around 1 ev above ( below ) the unoccupied ( occupied ) part of the @xmath4 bands , but the intensity is weaker than that of ( tmtsf)@xmath1pf@xmath2 . similarly to the tmtsf case , the @xmath0+@xmath6 for srvo@xmath3 makes the plasmon satellite broader than the @xmath0 result . on the o@xmath180 bands ,
the self - energy effect is appreciable ; the imaginary part of the self - energy , which is related to the lifetime of the quasiparticle states , becomes large as the binding energy is apart from the fermi level .
thus , the spectrum of the o@xmath180 band becomes rather broad .
the panel ( c ) shows the comparison of calculated density of states .
we see that the gga spectrum in the @xmath4 and o@xmath180 bands is largely renormalized in the @xmath0 and @xmath7 spectra .
the shape of the @xmath0 spectrum resembles the @xmath0+@xmath6 one , though the latter is somewhat more broadened
. and o@xmath180 bands of srvo@xmath3 calculated with ( a ) @xmath0 approximation and ( b ) @xmath7 method .
blue - solid curves are the gga results .
the fermi level is at zero energy .
the colorbar is in linear scale .
( c ) comparison of the density of states among @xmath0+@xmath6 ( red - solid curve ) , @xmath0 ( blue - dotted curve ) , and gga ( black - thin curve ) results .
, scaledwidth=45.0% ] figure [ pes - svo ] compares theoretical photo - emission spectra with experimental ones taken from refs . and .
light - green circles and black dots are the experimental photoemission spectra with the photon energies @xmath181 60 ev ( ref . ) and @xmath181 900 ev ( ref . ) , respectively . on the overall profile , the _ ab initio _
@xmath0+@xmath6 spectrum reasonably reproduces the experimental results such as the relative intensity between the @xmath4 and o@xmath180 bands .
@xcite on the other hand , we see that the theoretical incoherent intensity around @xmath1822 ev is weaker than the experimental one .
+ @xmath6 , @xmath0 , and gga results , respectively . in the experiment ,
light - green circles and black dots denote the photoemission spectra with the photon energies @xmath181 60 ev ( ref . ) and @xmath181 900 ev ( ref . ) , respectively .
a lorentzian broadening of 0.05 ev is applied to the calculated spectra.,scaledwidth=40.0% ] in fig .
[ xas - svo ] , we compare theoretical spectra of an unoccupied region in the @xmath4 bands @xmath183 with experimental one taken from the soft x - ray absorption spectrum ( green circles ) ( ref . ) .
the effect of the low - energy plasmon excitation on the electronic structure appears to be stronger in the unoccupied region than in the occupied one .
@xcite the _ ab initio _
@xmath0+@xmath6 spectrum resembles the experimental spectrum more closely than the @xmath0 result ; the cumulant expansion reduces the quasiparticle intensity and shifts the plasmon satellite to a lower energy .
we note that the theoretical spectra do not include the contribution from the @xmath184 states .
also , in srvo@xmath3 , since the local - interaction effect competes with the plasmon excitation , the view of the competition of the several factors would be important for quantitative understanding , which remains to be explored
. bands .
the experimental data are taken from the soft x - ray absorption spectrum ( ref . ) . red - thick - solid , blue - dotted , and black - thin - solid curves are the @xmath0+@xmath6 , @xmath0 , and gga results , respectively .
a lorentzian broadening of 0.05 ev is applied to the calculated spectra.,scaledwidth=40.0% ]
we have performed _ ab initio _
@xmath0 plus cumulant - expansion calculations for an organic conductor ( tmtsf)@xmath1pf@xmath2 and a transition - metal oxide srvo@xmath3 to study the low - energy plasmon - fluctuation effect on the electronic structure .
the bands around the fermi level of these materials are isolated from the other bands , and the low - energy plasmon excitations derived from these isolated bands exist .
our calculated reflectance spectra well identify the experimental low - energy plasmon peaks . by calculating the cumulant - expanded green s function based on the @xmath0 approximation to the self - energy
, we have simulated spectral functions and compare them with photoemission data .
we have found agreements between them , indicating that the low - energy plasmon excitation certainly affects the low - energy electronic structure ; it reduces the quasiparticle spectral weight around the fermi level and leads to the weight transfer to the satellite parts .
this effect was found to be more or less appreciable in ( tmtsf)@xmath1pf@xmath2 than in srvo@xmath3 . in particular , in ( tmtsf)@xmath1pf@xmath2 , the spectrum at the standard @xmath0 level exhibits a clear plasmaron state , but considering the plasmon - fluctuation effects not treated in the standard @xmath0 calculation leads to the disappearance of the state in the @xmath7 . since the low - energy isolated - band structure
is commonly found in various materials , the low - energy plasmon effect pursued in the present work can provide a basis for understanding the electronic structure of real systems .
the recent progress in the photoemission experiment for the correlated materials ( refs . ) requires a concomitant progress on the theoretical side and the present _ ab initio _ many - body calculations would provide firm basis for them . in the present study ,
we have focused on the long - range correlation and treated it effectively with the cumulant - expansion method , while the short - ranged correlation , which is appropriately described by the @xmath185-matrix framework , is neglected .
this is a future challenge which remains to be explored .
in addition , recent photoemission spectroscopy reveals appreciable differences in the electronic structure between the bulk and thin film systems .
@xcite the electronic structure of the surface is very sensitive to the atomic configurations at the surface @xcite and therefore careful analyses for the atomic structure and its effect on electronic structure are required .
the _ ab initio _
calculations for the surface effect are clearly important for the deep understanding of the spectroscopy of the real materials .
this is also a future challenge .
we would like to thank norikazu tomita and teppei yoshida for useful discussions .
calculations were done at supercomputer center at institute for solid state physics , university of tokyo .
this work was supported by grants - in - aid for scientific research ( no . 22740215 , 22104010 , 23110708 , 23340095 , 23510120 , 25800200 ) from mext , japan , and consolidator grant correlmat of the european research council ( project number 617196 ) .
we briefly describe effects of band - off - diagonal terms of the self - energy on the spectral functions . with neglecting the band - off - diagonal terms , the spectral function is calculated as @xmath186 where the matrix element of the self - energy with respect to the ks state @xmath187 @xmath188 is the diagonal term of eq .
( [ sigma - od ] ) .
the @xmath84 in eq .
( [ akwdiag ] ) is the energy shift to correct the mismatch of the fermi level between the initial and final states [ see eq .
( [ norm ] ) ] .
figure [ akw - svo - diag ] displays the gw spectral function of srvo@xmath3 ( a ) without and ( b ) with band off - diagonal terms of self - energy .
also , the panel ( c ) shows the comparison between the calculated density of states .
we do not see discernible difference between the two ; the band off - diagonal terms of the self - energy is negligible .
this is because , in srvo@xmath3 , the @xmath189 orbitals of the v atom are well localized and the hybridization with o@xmath180 orbital is small . similarly to srvo@xmath3 , the band - off - diagonal effect on the spectral function is found to be very small in ( tmtsf)@xmath1pf@xmath2 . and
o@xmath180 bands of srvo@xmath3 ( a ) without and ( b ) with the band off - diagonal matrix elements of the self - energy .
blue - solid curves are the gga result .
the fermi level is at zero energy .
the colorbar is in linear scale .
( c ) comparison of the density of states without ( dark - red - solid curve ) and with ( light - green - dotted curve ) the off - diagonal terms.,scaledwidth=45.0% ] d. jerome , a. mazaud , m. ribault and k. bechgaard , j. phys .
( paris ) * 41 * , l95 ( 1980 ) . for a review ,
see , e.g , t. ishiguro , k. yamaji , and g. saito , _ organic superconductors _
( springer - verlag , heidelberg , 1998 ) . for a review , see , e.g , k. kuroki , j. phys .
. jpn . * 75 * , 051013 ( 2006 ) .
s. ishibashi , sci .
technol . adv
. mater . * 10 * , 024311 ( 2009 ) .
e. pavarini , s. biermann , a. poteryaev , a. i. lichtenstein , a. georges , and o. k. andersen , phys .
lett . , * 92 * , 176403 ( 2004 ) .
i. a. nekrasov , k. held , g. keller , d. e. kondakov , th .
pruschke , m. kollar , o. k. andersen , v. i. anisimov , and d. vollhardt , phys .
b * 73 * , 155112 ( 2006 ) .
f. lechermann , a. georges , a. poteryaev , s. biermann , m. posternak , a. yamasaki , and o. k. andersen , phys .
b * 74 * , 125120 ( 2006 ) .
b. gallois , j. gaultier , c. hauw , t .- d .
lamcharfi and a. filhol , acta cryst . *
b42 * , 564 ( 1986 ) .
martin dressel , isrn condensed matter physics , * 2012 * 732973 ( 2012 ) . c. s. jacobsen , d. b. tanner , and k. bechgaard , phys .
46 * 1142 ( 1981 ) .
d. malterre , b. dardel , m. grioni , p. weibel , y. baer , journal de physique iv , * 3 * , 97 ( 1993 ) .
b. dardel , d. malterre , m. grioni , p. weibel , y. baer , j. voit and d. jrme , europhys .
* 24 * , 687 ( 1993 ) .
f. zwick , s. brown , g. margaritondo , c. merlic , m. onellion , j. voit , and m. grioni , phys .
, * 79 * , 3982 ( 1997 ) .
f. zwick , m. grionia , g. margaritondoa , v. vescolib , l. degiorgib , b. alavic , g. grnerc , solid state commun . * 113 * , 179 ( 2000 ) .
v. vescoli , f. zwick , w. henderson , l. degiorgi , m. grioni , g. gruner , and l.k .
montgomery , eur .
j. b * 13 * , 503 ( 2000 ) .
i. h. inoue , o. goto , h. makino , n. e. hussey , m. ishikawa , phys .
b * 58 * , 4372 ( 1998 ) .
h. makino , i. h. inoue , m. j. rozenberg , i. hase , and y. aiura , and s. onari , phys .
b * 58 * , 4384 ( 1998 ) .
k. morikawa , t. mizokawa , k. kobayashi , a. fujimori , h. eisaki and s. uchida , f. iga , and y. nishihara , phys .
b * 52 * , 13711 ( 1995 ) .
a. sekiyama , h. fujiwara , s. imada , s. suga , h. eisaki , s. i. uchida , k. takegahara , h. harima , y. saitoh , i. a. nekrasov , g. keller , d. e. kondakov , a.v .
kozhevnikov , th .
pruschke , k. held , d. vollhardt , and v. i. anisimov , phys .
lett . , * 93 * , 156402 ( 2004 ) .
k. maiti , u. manju , s. ray , p. mahadevan , i. h. inoue , c. carbone , and d. d. sarma , phys .
b * 73 * , 052508 ( 2006 ) .
g. kotliar , s. y. savrasov , k. haule , v. s. oudovenko , o. parcollet , and c. a. marianetti , rev .
phys . , * 78 * 865 ( 2006 ) .
m karolak , t o wehling , f lechermann and a i lichtenstein , j. phys . : condens .
matter * 23 * , 085601 ( 2011 ) .
l. huang and y. wang , europhys .
lett , * 99 * 67003 ( 2012 ) .
m. casula , a. rubtsov , and s. biermann , phys .
b * 85 * , 035115 ( 2012 ) .
h. lee , k. foyevtsova , j. ferber , m. aichhorn , h. o. jeschke , and r. valent , phys .
b * 85 * , 165103 ( 2012 ) .
s. biermann , f. aryasetiawan , and a. georges , phys .
, * 90 * , 086402 ( 2003 ) .
j. m. tomczak , m. casula , t. miyake , f. aryasetiawan and s. biermann , europhys .
* 100 * , 67001 ( 2012 ) . c. taranto , m. kaltak , n. parragh , g. sangiovanni , g. kresse , a. toschi , and k. held , phys .
b * 88 * , 165119 ( 2013 ) .
r. sakuma , p. werner , and f. aryasetiawan , phys .
b * 88 * , 235110 ( 2013 ) .
r. sakuma , c. martins , t. miyake , and f. aryasetiawan , phys .
b * 89 * , 235119 ( 2014 ) . j. m. tomczak , m. casula , t. miyake , and s. biermann , phys . rev .
b * 90 * , 165138 ( 2014 ) .
l. hedin , phys . rev . * 139 * , a796 ( 1965 ) .
m. s. hybertsen and s. g. louie , phys . rev .
b * 34 * , 5390 ( 1986 ) .
g. onida , l. reining , a. rubio , rev .
phys , * 74 * , 601 ( 2002 ) .
f. aryasetiawan and o. gunnarsson , rep .
. phys . * 61 * , 237 ( 1998 ) .
m. s. hybertsen and s. g. louie , phys .
lett . * 55 * , 1418 ( 1985 ) .
r. w. godby , m. schlter , and l. j. sham , phys .
lett . * 56 * , 2415 ( 1986 ) .
r. w. godby , m. schlter , and l. j. sham , phys .
b * 35 * , 4170 ( 1987 ) .
j. e. northrup , m. s. hybertsen , and s. g. louie , phys .
lett . * 59 * , 819 ( 1987 ) .
n. hamada , m. hwang , and a. j. freeman , phys .
b * 41 * , 3620 ( 1990 ) .
f. aryasetiawan , phys .
b * 46 * , 13051 ( 1992 ) .
f. aryasetiawan and o. gunnarsson , phys .
lett . * 74 * , 3221 ( 1995 ) .
w. g. aulbur , l. joensson , and j. w. wilkins , solid state phys .
* 54 * , 1 ( 2000 ) .
b. arnaud and m. alouani , phys .
b * 62 * , 4464 ( 2000 ) .
w. ku and a. g. eguiluz , phys .
lett . * 89 * , 126401 ( 2002 ) .
m. usuda , n. hamada , t. kotani , and m. van schilfgaarde , phys .
b * 66 * , 125101 ( 2002 )
. t. kotani and m. van schilfgaarde , solid state commun . * 121 * , 461 ( 2002 ) .
a. yamasaki and t. fujiwara , phys .
b * 66 * , 245108 ( 2002 ) .
s. lebgue , b. arnaud , m. alouani , and p. e. bloechl , phys .
b * 67 * , 155208 ( 2003 ) .
s. v. faleev , m. van schilfgaarde , and t. kotani , phys .
lett . * 93 * , 126406 ( 2004 ) .
m. shishkin and g. kresse , phys .
b * 74 * , 035101 ( 2006 )
. t. kotani , m. van schilfgaarde , and s. v. faleev , phys .
b * 76 * , 165106 ( 2007 ) c. friedrich , s. blgel , and a. schindlmayr , phys .
b * 81 * , 125102 ( 2010 ) .
r. sakuma , t. miyake , and f. aryasetiawan , phys .
b * 86 * , 245126 ( 2012 )
. t. miyake , c. martins , r. sakuma , and f. aryasetiawan , phys .
b * 87 * , 115110 ( 2013 ) .
y. nohara , s. yamamoto , and takeo fujiwara , phys .
b * 79 * , 195110 ( 2009 ) .
l. hedin , physica scripta .
* 21 * , 477 ( 1980 ) .
f. aryasetiawan , l. hedin , and k. karlsson , phys .
lett . * 77 * , 2268 ( 1996 ) .
m. guzzo , g. lani , f. sottile , p. romaniello , m. gatti , j. j. kas , j. j. rehr , m. g. silly , f. sirotti , and l. reining , phys .
lett . * 107 * , 166401 ( 2011 ) .
j. lischner , d. vigil - fowler , and s. g. louie , phys .
lett . * 110 * , 146801 ( 2013 ) .
j. lischner , d. vigil - fowler , and s. g. louie , phys .
b * 89 * , 125430 ( 2014 ) .
j. j. kas , j. j. rehr , and l. reining , phys .
b * 90 * , 085112 ( 2014 ) .
j. lischner , g. k. plsson , d. vigil - fowler , s. nemsak , j. avila , m. c. asensio , c. s. fadley , and s. g. louie , phys . rev .
b bf 91 , 205113 ( 2015 ) .
m. s. hybertsen and s. g. louie , phys .
b * 35 * , 5585 ( 1987 ) .
p. puschnig and c. ambrosch - draxl , phys .
b * 66 * , 165105 ( 2002 ) . c. ambrosch - draxl and j. o. sofo , computer physics communications * 175 * , 1 ( 2006 )
j. spencer and a. alavi , phys .
b * 77 * , 193110 ( 2008 ) .
the hole green s function based the cumulant expansion is written in the time domain as @xmath190 where @xmath191 and @xmath192 are the non - interacting green s function and the cumulant , respectively , and we adopt the compressed notation @xmath193 , etc .
the green s function based on the dyson equation is @xmath194 in the lowest order , the following integral equation for the cumulant holds : @xmath195 from this equation , the time dependence of the cumulant is determined . in the practical calculation ,
the @xmath0 approximation is employed in the self - energy in the right hand side in eq .
( [ ch12 ] ) .
j. yamauchi , m. tsukada , s. watanabe , and o. sugino , phys .
b * 54 * , 5586 ( 1996 ) . n. troullier and j. l. martins , phys . rev .
b * 43 * 1993 ( 1991 ) . l. kleinman and d. m. bylander , phys .
. lett . * 48 * 1425 ( 1982 ) .
j. p. perdew , k. burke , and m. ernzerhof , phys .
lett . * 77 * , 3865 ( 1996 ) .
n. marzari and d. vanderbilt , phys .
b * 56 * , 12847 ( 1997 ) .
i. souza , n. marzari , and d. vanderbilt , phys .
b * 65 * , 035109 ( 2001 ) .
t. fujiwara , s. yamamoto , and y. ishii , j. phys .
* 72 * , 777 ( 2003 ) . for ( tmtsf)@xmath1pf@xmath2
, we constructed the maximally localized wannier orbitals for three separated energy windows , which contains homo , homo@xmath1401 , and homo@xmath1402 bands respectively .
as an initial guess for the construction of the homo wannier orbital in ( tmtsf)@xmath1pf@xmath2 , we used the @xmath161-type gaussians with the width 1.67 , which is placed on the center of the central c bond of the tmtsf molecule , in a direction perpendicular to the bond .
also , an initial guess for the homo@xmath162 wannier orbital is a superposition of the @xmath161-type gaussians with width 1.18 put on the c atoms forming the center bond , in a direction perpendicular to the bond . in the homo@xmath163 band , the wannier orbital
is found to be localized in only one of the two molecules in the unit cell .
therefore , we take as an initial guess for the homo@xmath163 wannier orbital a superposition of the @xmath161-type gaussians with width 1.18 put on the se atoms in the molecule , in a direction perpendicular to the molecular plane . for srvo@xmath3 , we constructed wannier functions from the single energy window containing v 3d and o 2p orbitals .
as the initial guess , @xmath189- and @xmath161- gaussians are set on the v and o sites , respectively .
k. nakamura , r. arita , and m. imada , j. phys .
. jpn . * 77 * , 093711 ( 2008 ) .
k. nakamura , y. yoshimoto , t. kosugi , r. arita , and m. imada , j. phys .
. jpn . * 78 * , 083710 ( 2009 ) .
k. nakamura , y. yoshimoto , and m. imada , phys .
b * 86 * , 205117 ( 2012 ) . t. kiss , a. chainani , h. m. yamamoto , t. miyazaki , t. akimoto , t. shimojima , k. ishizaka , s. watanabe , c .- t . chen , a. fukaya , r. kato and s. shin , nature communications , doi:10.1038/ncomms2079 ( 2012 ) .
f. zwick , s. brown , g. margaritondo , c. merlic , m. onellion , j. voit , and m. grioni , phys .
lett . * 79 * , 3982 ( 1997 ) .
r. claessen , m. sing , u. schwingenschlgl , p. blaha , m. dressel , and c. s. jacobsen , phys .
* 88 * , 096402 ( 2002 ) .
m. takizawa , m. minohara , h. kumigashira , d. toyota , m. oshima , h. wadati , t. yoshida , a. fujimori , m. lippmaa , m. kawasaki , h. koinuma , g. sordi , and m. rozenberg , phys .
b * 80 * , 235104 ( 2009 ) .
t. yoshida , m. hashimoto , t. takizawa , a. fujimori , m. kubota , k. ono , and h. eisaki , phys . rev .
b * 82 * , 085119 ( 2010 ) .
b. dardel , d. malterre , m. grioni , p. weibel , and y. baer , f. levy , phys .
lett . , * 67 * , 3144 ( 1991 ) .
a. sekiyama , a. fujimori , s. aonuma , h. sawa , and r. kato , phys .
b * 51 * 13899(r ) ( 1995 )
. h. ando , t. yokoya , k. ishizaka , s. tsuda , t. kiss , s. shin1 , t. eguchi , m. nohara , and h. takagi , j. phys . : condens .
matter * 17 * 4935 ( 2005 ) .
m. grioni , s. pons , and e. frantzeskakis , j. phys . : condens .
matter * 21 * , 023201 ( 2009 ) k. koizumi , k. ishizaka , t. kiss , m. okawa , r. kato , and s. shin , j. phys .
jpn , * 82 * 025004 ( 2013 ) . | we present _ ab initio _
@xmath0 plus cumulant - expansion calculations for an organic compound ( tmtsf)@xmath1pf@xmath2 and a transition - metal oxide srvo@xmath3 .
these materials exhibit characteristic low - energy band structures around the fermi level , which bring about interesting low - energy properties ; the low - energy bands near the fermi level are isolated from the other bands and , in the isolated bands , unusually low - energy plasmon excitations occur . to study the effect of this low - energy - plasmon fluctuation on the electronic structure , we calculate spectral functions and photoemission spectra using the _ ab initio _ cumulant expansion of the green s function based on the @xmath0 self - energy .
we found that the low - energy plasmon fluctuation leads to an appreciable renormalization of the low - energy bands and a transfer of the spectral weight into the incoherent part , thus resulting in an agreement with experimental photoemission data . |
current models of galaxy formation and evolution favor hierarchical growth of galaxies from smaller systems ( e.g. , @xcite ) . within the paradigm of hierarchical galaxy formation , galaxies evolve along the hubble sequence , transforming back and forth between e / s0 and spiral / irregular morphology through a series of quiescent and violent periods .
recognition that galaxies can transform from late to early type dates back at least to @xcite , who proposed that elliptical galaxies can form from mergers of similar mass late - type galaxies .
recent simulations find that while similar mass ( 1:13:1 ) mergers of disk galaxies result in classical ellipticals , unequal mass ( 4.5:110:1 ) or gas - rich mergers of galaxies result in s0s and hybrid galaxies with properties of both early and late types @xcite .
substantial observational evidence confirms that early - type galaxies can form through major mergers ( e.g. , @xcite ) . in agreement with the simulations ,
@xcite argue that larger , slow - rotating elliptical galaxies form from dry major mergers while fainter , fast - rotating ellipticals and s0s seem to form from minor mergers or very gas - rich major mergers .
can galaxies evolve in the other direction , from early to late type ?
simulations suggest that galaxy evolution in this direction involves the regrowth of stellar disks @xcite .
morphological transformation from early to late type requires that several conditions be met , which recent studies have found satisfied in some early - type galaxies .
first , the transformation from early- to late - type requires a substantial reservoir of cold gas the raw material for disk regrowth .
historically , early - type galaxies were thought to be gas - poor @xcite .
subsequent surveys , however , found that the ratio of gas mass to blue luminosity ( @xmath4 ) in early - type galaxies ranges from upper limits of @xmath5 to measured @xmath6 for large atomic gas reservoirs similar to those of spiral galaxies @xcite .
the 70% hi detection rate by @xcite suggests that atomic gas reservoirs are actually relatively common in field early - type galaxies .
recent surveys also find a significant amount of molecular gas ( @xmath7@xmath8 ) in 2878% of early - type galaxies , depending on the survey @xcite .
the large range in cold gas content hints that there may be distinct sub - populations of early - type galaxies the conventional red and dead early types , and a population that is still accreting gas and forming stars .
in addition to the _ presence _ of cold gas , the distribution of the gas is important in assessing the potential for disk regrowth .
the existence of giant hi disks and rings @xcite , as predicted by simulations of mergers between gas - rich galaxies @xcite , makes it possible for stellar disks to form from already - present gas disks .
the hi structures around some early - types galaxies are known to have regular velocity fields , with a continuity between ionized and neutral gas @xcite .
@xcite find a rotationally supported hi disk in the elliptical galaxy ngc 520 , which , if star formation is triggered in the disk , may generate a late - type galaxy .
kiloparsec - scale disks of regularly rotating molecular gas have also been found in the center of some early - type galaxies @xcite .
all these observations suggest that the gas disks around some , possibly most early - type galaxies have reached an equilibrium arrangement suitable for disk growth .
while the presence of cold gas disks in an equilibrium configuration is necessary for stellar disk growth , it does not imply that star formation is actually occurring .
a key question is whether e / s0 galaxies with such disks are currently evolving morphologically , and whether such galaxies constitute a significant fraction of the early - type galaxy population . to answer these questions , one must be able to distinguish between old , gas - poor early - type galaxies and those undergoing morphological transformation via disk regrowth .
* hereafter kgb ) recently identified a population of e / s0s that reside alongside spiral galaxies on the blue sequence in color vs. stellar mass parameter space .
they argue that among these `` blue - sequence e / s0s , '' a subset with low - to - intermediate masses may represent the missing link i.e. , galaxies that are actively ( re)growing stellar disks and plausibly transitioning from early to late type .
find that these blue - sequence e / s0s fall between spirals and red - sequence e / s0s in scaling relations ( stellar mass @xmath9 vs. radius and vs. velocity dispersion @xmath10 ) , implying that blue - sequence e / s0s might form a transitional population between the two groups .
compared to conventional red - sequence e / s0s , blue - sequence e / s0s consistently have bluer outer disks , and often bluer centers as well suggesting on - going star formation in both disks and disky bulges . further supporting this disk - building picture ,
find evidence for kinematically distinct disks ( i.e. , counterrotating or polar ) in a notable sub - population of blue - sequence e / s0s .
unlike red - sequence e / s0s , whose mass function peaks at high masses , blue - sequence e / s0s are rare for @xmath11 12@xmath12 , and are common only for @xmath13 .
their abundance increase sharply to 2030% below @xmath14 , coincident with the mass threshold below which galaxies become notably more gas rich ( @xcite , based on @xcite ) . at intermediate masses between @xmath15 and @xmath16 ,
find that blue - sequence e / s0s include examples of both major mergers that are likely to fade onto the red sequence after exhausting their gas , and also settled galaxies that may be evolving toward later - type morphologies .
the former dominate at higher masses , and the latter at lower masses .
one outstanding question that remains is the extent of disk growth possible in low - to - intermediate mass blue - sequence e / s0s . in this paper
, we approach this question from the point of view of atomic gas reservoirs . addressed this question with limited archival data for the nearby field galaxy survey ( nfgs , @xcite ) , finding preliminary evidence for gas reservoirs comparable to spiral galaxies . in this paper
we present more complete , higher - quality data for all nfgs e / s0s with @xmath17 and a sampling of more massive e / s0s .
we compare the atomic gas masses ( normalized to @xmath18 ) of blue - sequence e / s0s with those of other galaxies , and we examine the possible fractional growth in @xmath18 given current rates of star formation .
the question of how efficiently the atomic gas might flow inward and condense into molecular gas is beyond the scope of this paper ; however , we discuss preliminary evidence from our own and others work that suggests efficient conversion is plausible ( elaborated in a forthcoming paper : ( * ? ? ? * hereafter k09b ) ) .
section [ section.sample ] describes our statistically representative sample of red- and blue - sequence e / s0s from the nfgs , and presents new green bank telescope ( gbt ) data for the galaxies with @xmath17 , which are expected to show the most disk growth .
we also present a tabulation of hi data for the full nfgs .
section [ section.gasreservoirs ] compares atomic gas masses of blue - sequence e / s0s with those of red - sequence e / s0s and galaxies of later - type morphology within the nfgs .
we also compare to the sample of @xcite .
section [ section.starformation ] examines the fractional stellar mass growth possible for blue - sequence e / s0s given the current star formation rate , in two limiting scenarios constant ( i.e. , allowing gas infall ) and exponentially declining ( i.e. , closed box ) star formation . in section [ section.sfmechanism ] , we discuss gas exhaustion and gas inflow timescales , and we examine evidence for bursty star formation in blue - sequence e / s0s , which likely implies efficient conversion of hi to @xmath19 .
we conclude with a discussion of the evolutionary fates of blue - sequence e / s0s .
appendix [ section.appendix ] details features of our new hi data for interesting individual galaxies . in this paper , we assume @xmath20 = 70 @xmath21 .
to understand the role of blue - sequence e / s0s in the morphological evolution of galaxies , we need to examine these galaxies alongside different types of galaxies in various stages of evolution .
the nfgs provides an ideal parent sample for such study , spanning the natural diversity of galaxies in the local universe in terms of mass , luminosity , and morphological type .
@xcite selected the nfgs galaxies from the cfa redshift catalog , which they binned by absolute magnitude and sub - binned by morphological type . after applying a luminosity - dependent minimum redshift to avoid galaxies of a large angular size , @xcite chose every nth galaxy from each bin , scaling n to approximate the local galaxy luminosity function .
the resulting nfgs spans the full range of morphological types and eight magnitudes in luminosity , providing a distribution of galaxies that is statistically consistent with that of the local universe .
archival nfgs data include @xmath22 photometry @xcite , integrated spectrophotometry @xcite , and ionized gas and stellar kinematic data @xcite .
all nfgs galaxies also have @xmath23 photometry from 2mass @xcite .
stellar masses are estimated by fitting stellar population models to @xmath24 photometry and integrated spectrophotometry as described in ( updating @xcite ; see also @xcite ) .
this limits the `` full nfgs sample '' to 176 galaxies for which stellar masses are available . for consistency with the red / blue sequence dividing line of
, we use @xmath25 colors with those authors extinction corrections and @xmath26-corrections .
total magnitudes are also extinction corrected .
we also use star formation rates calculated by from extinction - corrected h@xmath27 spectral line data , integrated by scanning the slit across each galaxy and calibrated against iras - based star formation rates ( sfrs ) following @xcite .
we consider a subsample of nfgs e / s0 galaxies for detailed study of the @xmath18 regime where report abundant blue - sequence e / s0s .
our focus sample includes all 14 blue - sequence e / s0s with @xmath29 , with the limit chosen where the blue - sequence e / s0s tail off ( figure [ fig.nfgssample ] ) . to make a fair comparison
, we include all 11 nfgs red - sequence e / s0s with @xmath30 .
we also include two galaxies that lie on the dividing line between the red and blue sequences , which we will refer to as `` mid - sequence '' e / s0s following the naming convention of .
our cutoff mass is very close to the bimodality mass ( @xmath31 ) discussed in , so henceforth we refer to this sample as the `` sub-@xmath28 '' e / s0s .
the sample mass limit @xmath32 excludes a large population of high - mass red - sequence e / s0s ; this is appropriate because many properties scale with stellar mass , and including the high - mass galaxies would bias statistical comparison .
figure [ fig.nfgssample ] shows the sub-@xmath28 sample in @xmath25 color vs. stellar mass parameter space , illustrating similar mass distributions with good coverage of both the red and blue sequences .
divide the two sequences with the dashed line in figure [ fig.nfgssample ] , which is chosen with respect to the locus that hugs the upper boundary of the distribution of most late - type galaxies .
the line levels out at @xmath25 values of 1.14 and 1.64 , in agreement with @xcite .
the rest of the nfgs ( spirals and irregulars ) is plotted in the background in the same figure for comparison .
because the full nfgs was selected to be broadly representative of the local universe , we expect our sub-@xmath28 e / s0 sample to encompass a wide range of evolutionary stages as well .
color vs. stellar mass for galaxies in the nearby field galaxies survey ( nfgs , @xcite ) .
symbols denote morphological type classified by eye using monochrome @xmath33 or @xmath34 band images . the red sequence , the main locus of high - mass and/or cluster e / s0s , lies above the dashed line , while the blue - sequence , typically populated by spiral galaxies , lies below .
note the presence of a population of galaxies with early - type morphology in the region of color vs. stellar mass space populated by spirals .
for consistency with the red / blue sequence dividing line of , we use @xmath25 colors with those authors extinction corrections and k - corrections .
the sub-@xmath28 e / s0 sample of galaxies are boxed , with the two mid - sequence e / s0s on the borderline between the red and blue sequences also enclosed in diamonds .
note that in this paper `` sub-@xmath35 '' means @xmath36 , i.e. we use a cutoff slightly above @xmath37 , due to our original sample selection.[fig.nfgssample ] ] we present the compilation of hi data for the full nfgs in table [ table.nfgshi ] .
hi fluxes are available for most of the nfgs from the hyperleda database @xcite , which consists of data compiled from the literature . because the hyperleda database is compiled from observations made at different telescopes by different observers , @xcite homogenize the hi data to account for differences in observational parameters such as beamsize , spectral resolution , and flux scale . for nfgs galaxies lacking hi data in hyperleda
, we gather data from the literature when possible .
we obtain upper limits for three galaxies from @xcite , four hi fluxes and one upper limit from the cornell egg hi digital archive @xcite , and six upper limits and one flux measurement from the hi parkes all sky survey ( hipass , @xcite ) . because these measurements are often upper limits or taken at telescopes not included in the homogenization effort of @xcite , the fluxes for these galaxies are not homogenized to the hyperleda dataset .
good quality literature data are lacking for many early - type galaxies in the nfgs , so we have obtained gbt hi observations for many of these galaxies , with priority for our sample of sub-@xmath28 e / s0s .
the observations were obtained with the gbt spectrometer in l - band , with 50 mhz bandwidth , one spectral window , and nine sampling levels , in ten minute on - off source pairs ( five minutes per position ) during march and october of 2007 .
the total on - source time for each galaxy was determined during the observing runs based on the strength of the hi emission relative to the noise .
the spectra were reduced using gbtidl @xcite .
individual 30 second records with large harmonic radio frequency interference ( rfi ) were flagged , and persistent rfi spikes near the velocity range of the galaxy were interpolated across in a few cases .
the scans were accumulated and averaged for all data for an individual galaxy , and a polynomial of order @xmath385 was fitted over a range of @xmath3920 mhz to subtract the baseline .
hanning and fourth order boxcar smoothing , with decimation , were then applied to all the baseline - subtracted data , resulting in channel resolution of 0.0244 mhz ( @xmath395 @xmath40 ) .
flux calibration is derived from simultaneous observations of an internal noise diode whose intensity is stable .
figures [ fig.gbtspectra_red ] and [ fig.gbtspectra_blue ] present the spectra of red- , blue- , and mid - sequence e / s0s obtained with the gbt , and table [ table.newhi ] lists the observational parameters and measured quantities . col .
( 6 ) lists the total on - source time in seconds ; col .
( 7 ) is the heliocentric recession velocity measured at the mid - point of 20% flux after excluding companions ( @xmath41 , see [ section.companions ] )
. cols .
( 8) and ( 9 ) are the hi line widths measured at the 20% and 50% level ( @xmath42 ) .
( 10 ) gives the velocity - integrated flux and error ( @xmath43 ) , both in jy @xmath44 , and col .
( 11 ) gives the dispersion of the baseline channels measured in a line - free part of the spectrum ( @xmath45 ) in mjy .
we estimate the error of our flux measurements following @xcite , who derived the following analytical expression for uncertainty in total hi flux ( @xmath46 ) : @xmath47 where @xmath45 is the rms dispersion of the baseline in jy and @xmath48 is the velocity resolution of the spectrum in @xmath44 .
the errors in the measured heliocentric velocity ( @xmath49 ) and velocity widths ( @xmath50 ) are estimated following @xcite , using : @xmath51 @xmath52 @xmath53 where @xmath54 is the steepness of the profile , @xmath55 , and s / n is the ratio of the peak signal to @xmath45
. we can not homogenize our hi fluxes to the hyperleda system of @xcite , because these authors do not include the gbt in their homogenization calculations .
however , because of the large beam of the gbt at 21 cm ( @xmath399@xmath56 ) compared to the small optical sizes of our galaxies ( @xmath392@xmath56 ) , no beam - filling correction is needed . also
, flux calibration of gbt hi spectral line data is extremely stable and accurate , so large offsets in the flux scale between our gbt data and the hyperleda data are unlikely . in fact , comparison between our new gbt hi data and existing hi data from hyperleda for 11 galaxies show that over half have gbt fluxes within 20% of published hi fluxes , even though the hyperleda data are of poorer quality . in all 11 cases
the s / n of the gbt hi data is better than that of the published data .
we discuss the hi fluxes and line profiles of individual galaxies in further detail in appendix a. we check for companions to our gbt galaxies by setting the search radius in the nasa / ipac extragalactic database ( ned ) to match the beam of the gbt . for galaxies with known or obvious companions , we measure the hi flux twice : the first time we measure the flux within the velocity range from ionized - gas rotation curves ( or stellar rotation curves , if ionized gas is not available ) , and the second time we include all flux in the beam within a reasonable velocity range near the galaxy ( @xmath57 ) .
we mark the ionized - gas / stellar velocity range used to measure fluxes for the first method in figures [ fig.gbtspectra_red ] and [ fig.gbtspectra_blue ] with double vertical lines .
we present both measurements in table [ table.newhi ] : the first row contains values measured in the ionized - gas / stellar velocity range , while the second row includes companion flux ( if any ) within @xmath58300 @xmath59 .
the first method most likely underestimates hi content for the two galaxies ( ic 1639 , ic 195 ) for which we must determine the velocity range from stellar rotation curves , as those rotation curves are still rising at the last measured point .
the ionized - gas rotation curves used for the other galaxies , on the other hand , are flat , so the underestimation of hi flux for those galaxies should be small .
the second method likely overestimates a galaxy s hi reservoir , since cold gas from a companion may not be readily accessible for star formation in the target galaxy .
_ we use fluxes measured with only the first method in our analysis , as conservative estimates of hi gas mass .
_ we discuss how our results may change if we include companion gas in
[ section.simulations ] .
we estimate the observed maximum rotation speed , @xmath60 for each galaxy following @xcite : @xmath61 . where @xmath62 and @xmath63 are specified for a given velocity width @xmath64 measured at % level @xmath65 and velocity resolution @xmath66 , enabling galaxies from disparate hi datasets to be compared . for our gbt data , we adopt the recommended values of @xmath67 and @xmath68 to convert our @xmath69 widths into @xmath70 ( table [ table.nfgshi ] , col .
( 7 ) ) .
we find the difference between heliocentric measurements of the hi velocity and the optical velocity for nfgs galaxies to be small , centered around 0 @xmath44 , with a standard deviation of 33 @xmath44 . of the 200 galaxies in the nfgs
, we have hi information for 170 : new gbt observations for 27 galaxies , hyperleda hi data for 128 galaxies , and other literature data for 15 galaxies .
the 30 galaxies with no hi information are distributed reasonably evenly between morphological types ( 53% early , 40% late , and 7% pec / im ) as well as sequences ( 43% blue , 43% red , and 14% unknown because no stellar mass estimate is available ) .
we plot @xmath25 color as a function of @xmath71-band luminosity in figure [ fig.hisample ] to show the distribution of the nfgs sample with hi information .
there is a somewhat higher frequency of missing data among the brightest galaxies ( also the most distant , @xcite ) .
the hi gas masses given in col .
( 9 ) of table [ table.nfgshi ] are calculated from @xmath72 following @xcite : @xmath73 where @xmath74 is the local group and virgocentric flow corrected recessional velocity from @xcite , which we use to be consistent with the stellar mass estimates .
we multiply the hi gas mass by a factor of 1.4 to account for the presence of helium .
color as a function of @xmath71-band luminosity , both extinction corrected , for galaxies in the nfgs . hi detections are shown as dots , hi upper limits are represented by @xmath75 s , and galaxies with no hi data are shown as open circles .
[ fig.hisample ] ] to calculate upper limits , we measure the rms dispersion of the baseline , @xmath45 , in a signal - free part of the spectrum within the velocity range from ionized - gas ( or stellar ) rotation curves .
we then estimate @xmath76 following : @xmath77 where n is the number of channels in the velocity range we measured from , and @xmath48 is the width of each channel in units of @xmath44 .
we estimate the hi mass upper limit using 3@xmath78 as the hi flux and following equation [ equation.himass ] .
all of our sub-@xmath35 blue - sequence e / s0s , as well as both mid - sequence e / s0s , are detected in hi and have atomic gas masses ( @xmath79 ) ranging from @xmath80 to almost @xmath81 .
in contrast , four of our eleven sub-@xmath35 red - sequence e / s0s were not detected , although we integrated down to upper limits of @xmath82@xmath83 .
the remaining seven galaxies have atomic gas masses ranging from @xmath80@xmath84 .
we plot the distribution of atomic gas mass as a function of stellar mass for the sub-@xmath35 blue- and red - sequence e / s0s , mid - sequence e / s0s , and other nfgs galaxies in figure [ fig.mgasvmstar ] .
we find that , at a given @xmath18 , blue - sequence e / s0s tend to have larger gas masses than red - sequence e / s0s .
this result confirms the preliminary results of , for hi data that are much more complete in terms of sampling sub-@xmath35 e / s0s .
since almost half of the red - sequence e / s0s are actually upper limits , the separation between the gas masses of blue- and red - sequence e / s0s is actually larger than it appears in figure [ fig.mgasvmstar ] . ] even more intriguing is the location of blue - sequence e / s0s _ in between _ spiral / irregular galaxies and red - sequence e / s0s in the @xmath85 vs. @xmath18 relation shown in figure [ fig.mgasvmstar ] . in fact , there is considerable overlap between blue - sequence e / s0s and spiral / irregular galaxies .
this supports the suggestion that blue - sequence e / s0s form a transitional class between spirals / irregular galaxies and traditional red - sequence e / s0s , as originally inferred from the fact that blue - sequence e / s0s also lie between spiral / irregular galaxies and red - sequence e / s0s in the @xmath9-radius and @xmath9-@xmath10 relations . since we are interested in the potential for morphological transformation in blue - sequence e / s0s , it is informative to consider @xmath86 the mass of the atomic gas relative to the current stellar mass of a galaxy .
we plot the distribution of atomic gas - to - stellar mass ratios for sub-@xmath35 e / s0s on both sequences in figure [ fig.gsvmstar]a and list their values in table [ table.masses ] .
a kolmogorov - smirnov test on the distributions of @xmath87 for sub-@xmath35 red- and blue - sequence e / s0s rejects at the 99% level the possibility that these galaxies derive from the same parent population . while most of the sub-@xmath35 blue - sequence e / s0s and both the mid - sequence e / s0s have @xmath86 in the range of 0.1 to 1.0 , all but two of the sub-@xmath35 red - sequence e / s0s have @xmath88 .
following @xcite figure 4.51 , the formation of new stars in an extended disk constituting 25% of original total stellar mass will change the typical s0 galaxy to an sa galaxy .
this suggests that at least half of the sub-@xmath35 blue - sequence e / s0s _ do _ have large enough gas reservoirs for major morphological transformation _ if _ all the gas is converted into stars in the disk .
we discuss in later sections whether the gas is actually forming stars .
but for now , we can say that at least half of our sub-@xmath35 blue - sequence e / s0s do have the _ potential _ to transform their morphologies simply based on their atomic gas masses . in contrast , sub-@xmath35 red - sequence e / s0s lack sufficient atomic gas to do the same with rare exceptions .
figure [ fig.gsvmstar]a shows that while sub-@xmath35 red - sequence e / s0s tend to have lower values of @xmath86 , sub-@xmath35 blue - sequence e / s0s tend to have higher values .
there are , however , a couple of outliers that do not seem to follow this trend .
the most prominent of the outliers are the red - sequence e / s0 with a large gas reservoir , ugc 5923 ( log @xmath89 ) , and the blue - sequence e / s0 with a very small gas reservoir , ic 1639 ( log @xmath90 ) .
taking the stellar masses of these galaxies into consideration , however , provides plausible explanations for their gas - to - stellar mass ratios .
ugc 5923 , the gas - rich red - sequence e / s0 , has the lowest stellar mass of all red - sequence e / s0s in the nfgs with @xmath91 , so it is not surprising that this galaxy has a fractionally large gas reservoir , despite its red color .
@xmath92 emission indicates that there is some low - level star formation in ugc 5923 , but not enough to push the galaxy towards the blue sequence .
this galaxy does , however , appear dusty and has the highest internal extinction of all sub-@xmath35 red - sequence e / s0s , which suggests that it could also be forming stars _ now _ behind an obscuring dust screen . but for dust , this object might well follow the trend of the blue - sequence e / s0s in figure [ fig.gsvmstar ] .
ic 1639 , the gas - poor blue - sequence e / s0 , is on the other end of the stellar mass scale as the galaxy with the largest stellar mass in the sub-@xmath35 e / s0 sample ( @xmath93 ) . argue that blue - sequence e / s0s at these higher stellar masses are more often associated with violent encounters than disk building .
it is possible that this galaxy underwent an interaction with its larger companion in the not - so - distant past , which triggered a burst of star formation ( hence its blue color ) and quickly exhausted its gas reservoir .
the outliers described in the previous section indicate the importance of taking stellar mass into account when considering the gas - to - stellar mass ratio .
hence we plot @xmath87 as a function of @xmath18 ( figure [ fig.gsvmstar]b ) for all galaxies in the nfgs .
we fit the trend of decreasing atomic gas - to - stellar mass ratio with increasing stellar mass in figure [ fig.gsvmstar]b with a line in the form of @xmath94 for all nfgs galaxies with hi data , using survival analysis ( the asurv package : @xcite ) to include galaxies with @xmath85 upper limits .
table [ table.gsfits ] lists the coefficients for the different fits ( forward , bisector ) for different populations ( blue sequence , red sequence , spiral / irregulars , and e / s0s ) .
as mentioned earlier , most of the sub-@xmath35 blue - sequence e / s0s have @xmath87 in the range of 0.1 to 1.0 , while all but two of the sub-@xmath35 red - sequence e / s0s have @xmath95 .
the separation between the two populations is even more striking in figure [ fig.gsvmstar]b , and re - emphasizes the difference in potential for morphological transformation between sub-@xmath35 blue- and red - sequence e / s0s .
we compare the masses of gas reservoirs in our sample of sub-@xmath35 e / s0s with an ongoing survey of cold gas ( hi and co ) in a volume - limited sample of nearby e / s0s by @xcite and @xcite .
@xcite report that e / s0s contain less than @xmath3910% of the cold gas mass ( @xmath96 ) predicted for gas return by analytical stellar evolution models @xcite .
in particular , @xcite predict that @xmath97 in solar units , while @xcite conclude that @xmath98 for s0s and suggest that this also applies to ellipticals based on preliminary data @xcite .
we plot @xmath99 vs. @xmath100 in figure [ fig.blumvgasfrac ] , showing @xmath101 as a dashed line .
we divide @xmath102 by @xmath103 rather than @xmath9 to facilitate comparison with @xcite ; although the larger scatter compared to figure [ fig.gsvmstar]b suggests that @xmath100 does not work as well as @xmath18 , @xmath18 is not available for the e / s0s from @xcite and @xcite .
-band luminosity ( in solar units ) vs. @xmath33-band luminosity .
the @xmath33-band luminosities are corrected for foreground and internal extinction .
@xmath102 values for e / s0s from sage & welch ( 2006 , 2007 ; @xmath75 s ) account for hi , he , and @xmath19 , while @xmath102 values for blue- , red- , and mid - sequence e / s0s from this paper ( filled and empty circles , same symbols as figures [ fig.mgasvmstar ] and [ fig.gsvmstar]b ) account for only hi and he .
the cold gas cutoff at @xmath104 found by @xcite is shown as a dashed line .
downward arrows indicate upper limits.[fig.blumvgasfrac ] ] while there are a few e / s0 galaxies from @xcite and @xcite with values of @xmath105 0.1 , the majority of their e / s0s have much smaller values of @xmath99 .
in contrast , all but two of our sub-@xmath35 blue - sequence e / s0s fall above the 0.1 @xmath99 cutoff , as do both of the mid - sequence and a few red - sequence e / s0s . note that our data points do not include molecular gas as opposed to the @xcite data points , so the values of @xmath99 for our galaxies reflect only the neutral atomic ism , and the actual values could be even higher .
this result suggests that the sub-@xmath35 blue - sequence e / s0s in our sample have some of the most massive fractional gas reservoirs among early - type galaxies , in comparison to red - sequence e / s0s in our sample as well as e / s0s from the literature .
the differences in how galaxies in the two samples are selected may explain the disparity in @xmath99 between them .
the @xcite sample , selected from the nearby galaxies catalog , inherits the parent sample s biases against optically small galaxies ( diameter @xmath106@xmath107 ) and hi - poor systems @xcite .
hence , it is likely that the @xcite e / s0s are more massive and therefore more likely to be on the red sequence than our sample of e / s0s .
the distribution of @xmath100 for the two different samples in figure [ fig.blumvgasfrac ] hints at this , with our sample of sub-@xmath35 e / s0s predominantly at lower @xmath100 and the @xcite sample dominating at higher values . note that the @xmath100 of blue - sequence e / s0s in our sub-@xmath35 sample may be systematically enhanced by star formation , decreasing @xmath108 , so the difference in @xmath109 between our sub-@xmath35 e / s0s and the @xcite galaxies may be even larger .
although our sub-@xmath35 blue - sequence e / s0s have larger gas mass reservoirs than seen by @xcite , we still find much smaller reservoirs than that predicted for gas return by the analytical stellar evolution models discussed in @xcite , ranging from 10% to 30% of the predicted value .
our findings support sage & welch s conclusion that stellar mass loss is not the primary source of hi in e / s0s .
now that we have established that sub-@xmath35 blue - sequence e / s0s typically have substantial fractional atomic gas reservoirs and therefore potential for morphological transformation , we consider the question of whether this gas is being converted into stars at a rate that can lead to morphological transformation in a reasonable amount of time .
although we lack information about the spatial distribution of the atomic gas , we can make some simplifying assumptions and create limiting scenarios for the evolutionary trajectory of blue - sequence e / s0s given the current rate of star formation . to construct truly realistic scenarios for the evolutionary path of our galaxies
, we would have to account for all sinks and sources of gas .
the sink terms are the rates at which gas is converted into stars and ionized and/or expelled due to stellar winds and supernovae .
the source terms include the rate at which fresh gas is brought in from external sources ( minor mergers , interactions , etc . ) and at which gas is returned by stellar evolution .
a detailed accounting of all these processes , however , is beyond the scope of this paper . while there are many theoretical studies of the hierarchical assembly of galaxies ( many of which include estimates of the frequency of mergers ) , simulations still lack the resolution to predict the frequency of very minor mergers and interactions .
we discuss current simulations in more detail in
[ section.simulations ] .
we consider here the range of plausible evolutionary trajectories for our blue - sequence e / s0s by presenting two simplified , but limiting , scenarios .
we should note that we assume that the atomic gas is distributed in such a way that it can be made available for star formation ( e.g. , via conversion to molecular gas ) at a rate comparable to the current star formation rate .
this may not always be the case , but we argue that this is a plausible assumption in [ section.simulations ] and [ section.inflow ] . in the first scenario , we assume that the current star formation rate remains constant over time , which is an assumption many population synthesis models make .
preserving a constant star formation rate requires an increasing star formation efficiency for a closed box , or open - box inflows .
this scenario provides a reasonable upper limit on the possible growth in stellar mass per unit time and a lower limit on the amount of time it takes for that mass to form .
the second scenario represents the other limit : an exponentially declining star formation rate ( i.e. , closed box with no gas return ) . for each galaxy , we start out with the current cold gas reservoir ( @xmath110 ) and star formation rate ( @xmath111 ) .
as time progresses , the star formation rate declines exponentially following @xmath112 as gas is converted into stars , where @xmath113 ( e.g. , @xcite ) .
this scenario defines the lower limit on the amount of gas converted to stars within a fixed time , as the star formation rate is declining exponentially . because we lack information regarding the frequency of internal / external gas replenishment , these two scenarios ( exponentially declining sfr with no new gas , and constant sfr allowing gas infall ) represent simplified limits which likely bracket the actual amount of growth in the stellar component in these galaxies .
given the two limiting scenarios described in the previous section , how much growth in stellar mass is possible for the different galaxies ?
we plot the ratio of new stellar mass formed relative to the current stellar mass as a function of current stellar mass 1 , 2 , 3 , and 4 gyr in the future in figures [ fig.newstarsformed]ad , respectively . here
we note that 11 e / s0s from the sub-@xmath35 sample ( two on the blue sequence and nine on the red sequence ) do not appear in figure [ fig.newstarsformed ] .
these galaxies have spectra that are integrated over the galaxy , but no h@xmath27 emission was detected , so these galaxies do not appear in any of the figures using star formation rates .
future work with _
galex _ and _ spitzer _ data will provide better estimates of the star formation rates in these galaxies . for each galaxy ,
a vertical line in figure [ fig.newstarsformed ] represents the range of possible fractional increase in stellar mass , with the lower end representing the exponentially declining sfr scenario and the upper end representing the constant sfr scenario .
we mark the vertical line with a horizontal dash to note the stellar mass fraction at which the original gas reservoir runs out .
any growth in the stellar component indicated by the line above the horizontal dash requires inflow of gas , which we allow in the constant star formation scenario .
galaxies that have not converted all of the original gas mass in the constant star formation rate scenario after the time specified for each figure do not have a horizontal dash .
since these two scenarios are the limiting cases , it is likely that the actual amount of stellar mass growth is somewhere in between the two ends of the line .
figure [ fig.newstarsformed]b shows that many of the sub-@xmath28 blue - sequence e / s0s ( dark solid lines ) have the potential to increase their stellar masses by a large fraction in just 2 gyr , with seven of the fourteen sub-@xmath35 blue - sequence e / s0s crossing or above the 10% line . in 3 gyr ,
the time by which most sub-@xmath35 blue - sequence e / s0s have exhausted their gas reservoirs ( [ section.timescale ] ) , nine of the fourteen sub-@xmath35 blue - sequence e / s0s will cross or be above the 10% line .
the remaining five blue - sequence e / s0s have lower sfrs , so they are below the 10% line after 3 gyr has passed .
in fact , the sfrs for ngc 7360 , ic 1639 , and ic 195 are so low they do not form @xmath20.1% of their current stellar masses within 3 or even 4 gyr , so they do not appear on any of the panels in figure [ fig.newstarsformed ] . this large spread in the fractional stellar mass growth in blue - sequence e / s0s
reflects the spread in star formation rates , which in turn may be indicative of differences in burst stages of the galaxies within this population .
we discuss this in more depth in
[ section.inflow ] .
the two mid - sequence e / s0s in our sample also have potential for substantial morphological transformation .
ugc 7020a ( @xmath114 ) can form new stellar mass in the range of 2040% of its current stellar mass within 3 gyr , and ugc 6570 ( @xmath115 ) can form new stellar mass @xmath3910% over the same period .
the low - mass red - sequence e / s0 with a surprisingly large gas mass reservoir discussed earlier , ugc 5923 ( @xmath116 ) , has a relatively low star formation rate , but can still form new stellar mass @xmath210% of its current stellar mass within 3 gyr .
ngc 5338 ( @xmath117 ) has a very small gas reservoir but might form significant stellar mass if there is replenishment of gas .
the rest of the sub-@xmath35 red - sequence e / s0s have such low sfrs that they are below the 1% line ( i.e. , off the plot ) in all four panels of figure [ fig.newstarsformed ] .
we plot the same lines of fractional stellar mass growth for spiral and irregular galaxies with @xmath118 in grey in figures [ fig.newstarsformed]ad for comparison . for sub-@xmath35 blue - sequence e / s0s ,
the ratio of new stellar mass formed in 13 gyr to original stellar mass is comparable to that of the spiral / irregular distribution ( figure [ fig.newstarsformed]ac ) .
this suggests that , as a population , sub-@xmath35 blue - sequence e / s0s have potential for growth in the stellar component similar to that of spiral / irregular galaxies in the near term future ( @xmath3913 gyr ) .
the horizontal dashes marking where the original gas reservoirs runs out , however , suggest that the inflow of fresh gas is important to the long - term evolution of our galaxies ( figure [ fig.newstarsformed]c , d ) .
without maps of the distribution of hi , we can not say for certain that the atomic gas is distributed in such a fashion that it is readily available for star formation . in [ section.timescale ] and [ section.simulations ] below , we argue that regardless of the distribution of the atomic gas ( extra - planar , in the disk , or in companions ) , the timescale for inward travel of gas is most likely shorter than the duration of star formation in blue - sequence e / s0s .
this suggests that assuming that the atomic gas reservoir is readily available for star formation is reasonable given internal or external mechanisms to drive gas inflow . in [ section.inflow ] we will discuss evidence for frequent , externally driven inflow in the sub-@xmath35 blue - sequence e / s0 population .
the timescale that can be directly estimated from the atomic gas mass and star formation rate of a galaxy is the gas exhaustion time ( @xmath119 ) the amount of time it would take to convert all the gas into stars , assuming the current star formation rate remains constant .
this is the same @xmath120 as the one used in the exponentially declining star formation rate calculation for the second scenario in [ section.scenarios ] , although by definition the gas reservoir will never be exhausted in this scenario since the sfr decreases exponentially in parallel with decreasing gas mass .
_ most _ of the gas mass , however , will be converted into stellar mass within the timescale @xmath120 , so this is an interesting timescale to consider for both scenarios .
table [ table.gastimes ] lists the gas exhaustion times for galaxies in the sub-@xmath35 e / s0 sample that have star formation rates .
the large spread in gas exhaustion times may be reflective of the diversity of evolutionary states within the blue - sequence e / s0 population , which we explore further in [ section.inflow ] .
note that the range of gas exhaustion timescales we find includes shorter timescales than found by kgb because of preferential incompleteness at low hi masses in the archival data used by kgb . for a more informative picture
, we plot the fraction of sub-@xmath35 galaxies that will _ not _ have exhausted their original atomic gas reservoirs at a given future time in figure [ fig.timescales ] .
we include the mid - sequence e / s0s with the blue - sequence e / s0s in this figure as they seem to behave similarly .
while the fraction of star - forming spiral / irregular galaxies seems to have a gradual , smooth decline in this figure , the blue - sequence e / s0s have a sharp drop - off at @xmath393 gyr . in agreement with figure [ fig.newstarsformed]ad ,
figure [ fig.timescales ] suggests that sub-@xmath35 spiral / irregular galaxies typically continue forming new stars long after star formation in sub-@xmath35 blue - sequence e / s0s is extinguished by the exhaustion of the original gas reservoir .
there is no drop - off for the sub-@xmath35 red - sequence e / s0s , as we have sfrs for only two of the eleven galaxies .
the fact that most of the sub-@xmath35 red - sequence e / s0s have extremely low star formation rates , however , supports the conclusion that the near - term evolutionary trajectories of blue- and red - sequence e / s0s will be quite different . ] in [ section.gasreservoirs ] and [ section.starformation ] , we have demonstrated that blue - sequence e / s0s have substantial fractional atomic gas reservoirs that , if readily available for star formation , can translate into significant growth in stellar mass and consequent morphological transformation .
the key uncertainty is whether this gas is or can be made available for star formation on a reasonable timescale compared to the gas exhaustion timescale just discussed .
the gbt spectra , at 21 cm , have a resolution of only @xmath399@xmath56 , which ( for our sample ) translates to an uncertainty radius of 10110 kpc in the location of atomic gas , depending on the distance to the galaxy . without maps with sufficient angular resolution ( e.g. ,
vla hi maps ) , we do not know whether the cold atomic gas is accessible for conversion into @xmath19 for star formation or not .
broadly , there are three possible distributions for the hi gas : on a trajectory falling into the galaxy , in an hi disk that is a part of the galaxy ( the most plausible configuration ) , or in companions .
_ infalling gas : _ if the gas is somewhere outside the galaxy on its way inward , we expect it to travel inwards to the galaxy on a dynamical timescale .
the 9@xmath56 gbt beam at 1.4 ghz corresponds to radii of 10110 kpc from the centers of galaxies in the sub-@xmath28 e / s0 sample , with a median of 34 kpc .
we list the dynamical timescale for inward travel of gas from the edge of the beam for each galaxy in the sub-@xmath35 sample in col .
( 5 ) of table [ table.gastimes ] .
the dynamical timescale estimates for inward travel of gas range from 0.4 to 2.5 gyr , with a median of 1.1 gyr .
these timescale estimates are smaller than the gas exhaustion times for all but two of the galaxies .
these two galaxies happen to be the most distant of all sub-@xmath35 e / s0s in table [ table.gastimes ] , so the edge of the gbt beam corresponds to @xmath2100 kpc .
therefore it is not surprising that these galaxies have long timescales for infall from the edge of the beam that are greater than the gas exhaustion time . assuming that the gas is at the edge of the beam is the most extreme case ; it is much more likely that the gas is much closer to the galaxy , which would reduce the infall time as @xmath121 .
for example , the dynamical time for inflow of gas from the predicted hi radius ( 217 kpc , scaling from the blue optical radius using an assumed ratio of 2.11 ; @xcite ) is much shorter than the gas exhaustion time in all cases , ranging from 50400 myr , with a median of 150 myr ( table [ table.gastimes ] , col .
@xcite find evidence for the infall of extra - planar gas onto star - forming spiral galaxies on short timescales at rates comparable to their star formation rates ( @xmath39few @xmath122 ) , which supports our estimates above .
the dynamical timescales we estimate here are for inward travel of gas all the way to the center of the galaxy .
however , the gas is capable of forming stars far from the centers of galaxies , depending on parameters such as local surface density and mid - plane pressure ( e.g. , @xcite ) .
thus the distance infalling gas has to travel to reach star forming regions and the corresponding infall timescale may be even shorter than our estimates above .
_ disk gas : _ comparison between ionized - gas rotation curves and the new gbt hi profiles suggests that most of the atomic gas is likely distributed in rotating disks for all the gbt galaxies .
we identify four particularly interesting cases .
the ionized gas data show marginal or no rotation for ugc 6805 , ugc 7020a , and ugc 6003 , but their hi profiles have the appropriate widths for rotation given their stellar masses ( e.g. , based on the @xmath9-rotation velocity relation in ) .
there is no ionized gas detection at all for ngc 3522 , so the hi profile for this galaxy indicates the presence of a previously unknown gas disk . if the gas is in a stable orbit in the disk of a galaxy , it will not necessarily travel inwards towards star forming regions on a short timescale compared to the gas exhaustion time .
depending on the density of the gas , the gas disk may or may not collapse to form stars .
a dense disk of gas could dovetail nicely with either of our star formation scenarios above it will collapse and form a stellar disk at the rate of the global sfr . if , however , the hi disk is too diffuse and spread out in an extended disk , the gas will continue in circular orbit in a dormant fashion unless it is perturbed by internal instabilities or events such as minor mergers or interactions .
recent simulations find that the frequency of mergers increases as the ratio of masses between the progenitors ( @xmath123 ) decreases ( e.g. , @xcite and references therein ) . at the finest resolution currently available with the millennium simulation ,
@xcite find minor merger rates of 0.2 to 0.7 mergers per halo per gyr at z = 0 for @xmath124 1:30 to 1:100 , respectively .
these minor merger rates correspond to one merger every 5 gyr on the high mass end ( 1:30 ) and one per 1.4 gyr on the low mass end ( 1:100 ) .
the minor merger rate at the high mass end ( 1:30 ) is perhaps a bit long relative to the gas exhaustion times for our galaxies , but minor mergers with progenitor mass ratios down to 1:100 and even smaller may still be capable of inducing gas inflow and star formation . due to the lack of resolution ,
however , @xcite do not consider progenitor mass ratios smaller than 1:100 , where mergers are extrapolated to occur on timescales shorter than 1 gyr .
there is also observational evidence suggesting that tidal interactions with small companions occur relatively frequently in field galaxies , bringing fresh infall of gas .
at least 25% of field galaxies observed in hi in several surveys show asymmetric features , indicating that they have recently undergone or are undergoing tidal interactions @xcite .
moreover , if one takes lopsided structure ( azimuthal distortions in the stellar disk ) and kinematics as evidence of interaction , the fraction would increase to above 50% of field galaxies @xcite .
the frequency of minor mergers , taking into account recent studies of how gaseous disks could survive mergers and interactions @xcite , suggests that if some of the galaxies in our sub-@xmath35 e / s0 sample have large , diffuse , and extended hi disks , they may not lie dormant for too long before a minor merger or interaction induces the gas to flow inwards and triggers star formation . once a minor merger or interaction occurs , we expect the gas to travel inwards on the order of a dynamical timescale or shorter @xcite .
as we discussed above , these are short timescales relative to the frequency of minor mergers in simulations , and so we take the merger rate as the limiting factor in this case , not the subsequent inward travel time . here
we have considered only external mechanisms that could drive the disk gas inwards , but secular mechanisms that are internally driven ( e.g. , cloud - cloud collisions that provide an effective viscosity , bars that may be able to form independently of external perturbations , resonances , instabilities , etc . ) likely also play a role in the inflow of disk gas . because these mechanisms do not require external triggers , including the effects of internal mechanisms
will shorten the gas inflow timescales we estimate from minor mergers / interactions alone .
_ companion gas : _ as described in [ section.companions ]
, we adopt a conservative approach and limit the hi flux measurement to the velocity range indicated by the primary ionized - gas ( or stellar ) rotation curve for galaxies with known companions .
since we restrict the flux measurements , our fluxes are most likely underestimates . of the four sub-@xmath28 e / s0s with known companions in ned ( ugc 12265n , ic 195 , ic 1639 , and ngc 4117 ) , only ugc 12265n has a non - negligible sfr and appears in the plots in
[ section.starformation ] .
the other three will increase their atomic gas - to - stellar mass ratios by factors of 1.63.7 if we include the companion gas ( ic 195 : 0.12 @xmath125 0.20 , ic 1639 : 0.01 @xmath125 0.03 , ngc 4117 : 0.05 @xmath125 0.12 ) , but do not change the results of [ section.gasreservoirs ] significantly . because these three galaxies have no detected star formation
, the fractional stellar mass growth over time remains negligible .
including the companion gas , the atomic gas - to - stellar mass ratio for ugc 12265n would quadruple from 0.3 to 1.2 , giving it the second highest value of @xmath86 for sub-@xmath28 blue - sequence e / s0s .
the amount of growth over time ( figure [ fig.newstarsformed ] ) would not change by much ; the upper limit would remain the same , but the lower limit would increase since @xmath120 is larger with the extra gas . if the atomic gas is readily accessible for star formation , the gas exhaustion timescale would then increase from 1.4 gyr to 5.6 gyr .
following equation 4 of @xcite , we estimate the timescale for a merger via dynamical friction between ugc 12265n and its companion ( ugc 12265s ) 12 kpc away ( neglecting the unknown line - of - sight distance ) to be @xmath3970 myr.^2 \ , [ \frac{10^{10}\,m_{\odot}}{m_s}]\ , [ \frac{v_c}{\sqrt(2)\cdot 100 { \rm km\,s^{-1 } } } ] $ ] .
we assume that ln @xmath126 = 3.3 , @xmath127 , and the companion is half as massive as ugc 12265n .
if the companion is more massive than our assumption , the merger time would decrease by the same factor . if the companion is less massive and/or further away , the timescale for merger would increase as @xmath128 and @xmath129 .
] this is shorter than its gas exhaustion timescale , which suggests that much of the companion gas is available for star formation in the near future and the amount of growth in the stellar component for this galaxy is greater than our conservative estimates in the previous sections . given the 1:2 progenitor mass ratio for this system , the resultant burst of star formation is likely to be more extreme than the scenarios discussed in [ section.scenarios ] , enhancing the star formation rate by a factor of two or more ( e.g. , @xcite ) .
whether the merger between ugc 12265n and its companion and the subsequent star formation will result in late - type morphology is unclear .
a burst of central star formation may leave ugc 12265n with early - type morphology ( e.g. , @xcite ) , but the gas - richness of this pair suggests that a disk - dominated remnant may be more likely (;
@xcite ) . for sub-@xmath35 nfgs galaxies , where @xmath130 is the outer disk color ( between 5075% light radii ) minus the central color ( within 50% light radii ) , corrected for the typical color gradient of a galaxy at that galaxy s blue luminosity ( see @xcite ) .
[ fig.dbrcorrhist ] ] in [ section.simulations ] above , we presented the possibility of minor mergers and/or interactions acting as a trigger which induces gas inflow and star formation . here
we examine observational evidence that inflow events , most likely triggered externally , are in fact occurring in blue - sequence e / s0s .
in addition to having bluer inner and outer disk colors than red - sequence e / s0s , half of the blue - sequence e / s0s in the nfgs have centers ( within 50% light radii ) that are _ bluer _ than their outer disks ( 5075% light radii ) .
figure [ fig.dbrcorrhist ] plots the distribution of a related quantity , @xmath130 , for sub-@xmath35 nfgs galaxies .
@xmath130 is the color gradient of a galaxy `` corrected '' for the mean color difference of galaxies of the same @xmath131 @xcite , since larger galaxies may have central colors diluted by preexisting red bars or bulges .
figure [ fig.dbrcorrhist ] shows that all but one of the sub-@xmath35 blue - sequence e / s0s and both mid - sequence e / s0s are on the more blue - centered end of the distribution for spiral / irregular galaxies . in the broader nfgs ,
larger values of @xmath130 reflect central star formation enhancements and correlate strongly with morphological peculiarities and the presence of nearby companions @xcite , suggesting a triggered gas inflow scenario .
@xcite also find evidence for gas inflows in interacting blue - centered galaxies based on radial trends in gas metallicity .
these results are consistent with preliminary evidence for a relationship between @xmath132 and molecular - to - atomic gas mass ratios in s0sb galaxies , to be reported in , which implies that hi can in fact flow inwards and become @xmath19 , fueling central star formation . if the range of blue - centeredness we see in the sub-@xmath35 blue - sequence e / s0s indicates different burst stages resulting from episodic gas inflow , then perhaps the specific star formation rate ( ssfr ) in these galaxies should scatter about some average ssfr expected based on the median 3 gyr gas exhaustion timescale and the atomic gas mass of each galaxy .
we consider the difference between this expected ssfr ( @xmath133 ) and the observed ssfr as a function of @xmath130 ( figure [ fig.ssfrdbrcorr ] ) .
the error bars for @xmath134 are formal errors , and do not include systematic uncertainties from @xmath135 not being corrected for dust and/or from delays between the bluest stellar population colors and the peak of star formation .
such uncertainties may cause some of the significant scatter seen in this figure .
nonetheless , there is a suggestive trend with the more blue - centered galaxies having enhanced ssfrs , and the less blue - centered galaxies having reduced ssfrs .
the spearman rank correlation coefficient for the data in figure [ fig.ssfrdbrcorr ] is @xmath136 , with the probability of a null correlation at 19% .
however , if we exclude ic 692 ( @xmath137 = 0.13 ) from the calculation , the correlation coefficient becomes @xmath138 , with a probability of a null correlation decreasing to 3% .
given the error bars , we infer that the relationship is likely real .
this supports the inflow - driven burst picture , although follow - up on the apparent outliers would be informative .
the correlation between blue - centeredness and several properties morphological peculiarities , the presence of companions , increasing molecular - to - atomic gas mass ratios , and enhanced ssfrs suggests that the inflow of gas in sub-@xmath35 blue - sequence e / s0s may be episodic and triggered externally .
episodic inflows of gas in sub-@xmath35 blue - sequence e / s0s would imply that these galaxies are in various stages of bursty star formation , which is in agreement with the large ranges of prospective stellar mass growth ( [ section.massgrowth ] ) and gas exhaustion times ( [ section.timescale ] ) we see for this population .
this result also suggests that our focus on hi rather than molecular gas ( to be discussed in wei et al .
, in prep . ) does not render our estimates of stellar mass growth and gas exhaustion time in [ section.massgrowth ] and [ section.timescale ] highly inaccurate .
although stars typically form from molecular gas , the estimates of inflow and minor merger / interaction timescales suggest that the transition from hi to @xmath19 may occur quickly enough that the potential for morphological transformation can be realized even after material already in the molecular phase has been consumed or otherwise dispersed .
these inflow events may also play an important role in bridging between the sub-@xmath35 red- and blue - sequence e / s0 populations , triggering star formation in sub-@xmath35 red - sequence e / s0s such as the gas - rich ugc 5923 ( discussed in [ section.outliers ] ) and moving e / s0s onto the blue sequence . ) and observed specific star formation rate for sub-@xmath35 blue- and mid - sequence e / s0s as a function of blue - centeredness .
one of our blue - sequence e / s0s , ugc 9562 , is excluded from this figure because it is a polar ring galaxy and not expected to behave normally with respect to inflow .
[ fig.ssfrdbrcorr ] ] observations of the sauron e / s0 sample @xcite have found that many intermediate - mass early - type galaxies seem to contain a disk - like stellar component ( e.g. , @xcite ) possibly the remnant of a burst of star formation induced by the inflow of gas as discussed in this section .
it is possible that some of our sub-@xmath35 blue - sequence e / s0s are in an active , post gas - infall star - forming phase but will eventually become passive early - type galaxies like those observed by @xcite .
some of the differences between the sauron e / s0s and our blue - sequence e / s0s , however , likely arise from the differences between the two samples . while there are some e / s0s below the bimodality mass @xmath139 in the sauron sample , these are in the minority
; in contrast , the disk - building sub - population of blue - sequence e / s0s becomes numerically important only below this mass , and more notably below the threshold mass at @xmath140 ( ) .
it is also possible that the accretion of fresh gas from external sources , as discussed in [ section.simulations ] , will reignite star formation in some red - sequence e / s0s ( in both samples ) and continue building disks (; @xcite ) .
the high frequency of blue - sequence e / s0s below the threshold mass ( @xmath395% of the general galaxy population and @xmath3925% of e / s0s , ) would be difficult to explain by quenching mergers alone , but could be explained if these systems star - forming phase is either long or episodic .
@xcite recently identified a population of e / s0s that reside on the blue sequence in color vs. stellar mass space , where spiral galaxies typically reside .
blue - sequence e / s0s increase in numbers below the bimodality mass ( @xmath141 ) and especially below the gas richness threshold ( @xmath395@xmath142 ) .
these sub-@xmath35 blue - sequence e / s0s are characterized by fairly regular morphology , and many appear to be rebuilding disks .
blue - sequence e / s0s also fall between red - sequence e / s0s and spirals in the @xmath9-radius and @xmath9-@xmath10 relations , suggesting that they may be a transitional population . in this paper
, we examined the atomic gas content of blue - sequence e / s0s below @xmath39@xmath35 to determine whether they have large enough gas reservoirs to transform to later - type morphology . in a representative sample drawn from the nearby field galaxy survey
, we find that blue - sequence e / s0s have substantial atomic gas masses in the range of @xmath7@xmath143 , comparable to the gas masses of spiral and irregular galaxies in the same stellar mass range . blue- and red - sequence e / s0s have distinctly different atomic gas - to - stellar mass ratios , with most blue - sequence e / s0s in the range of 0.11.0 and most red - sequence e / s0s @xmath1440.1 .
this suggests significantly greater potential for morphological transformation in blue - sequence e / s0s than red - sequence e / s0s . combining these atomic gas masses with current rates of star formation
, we find that many of the sub-@xmath35 blue - sequence e / s0s can form new stars in the range of 1060@xmath145 of their current stellar masses within 3 gyr , in both constant ( i.e. , allowing gas infall ) and exponentially declining ( i.e. , closed box ) star formation scenarios , provided that the atomic gas reservoir is available for star formation . in the constant star formation scenario
, we find that about half of the sub-@xmath35 blue - sequence e / s0 systems will exhaust their gas reservoirs in @xmath33 gyr if no fresh gas infall is permitted . because of the lack of spatial resolution in our hi data
, we can not say for certain that the gas is readily available for star formation .
we find evidence , however , which indicate that fresh gas may be brought inwards and made available for star formation on timescales shorter than the gas exhaustion timescale .
we estimate the dynamical timescale for infall of extra - planar gas to be on average @xmath1441 gyr , shorter than the gas exhaustion time for most of our galaxies . the frequency of inflow for gas trapped in diffuse disks may be dominated by the rate of minor mergers / interactions , which simulations find to occur every 1.4 gyr or less for progenitor mass ratios of 1:100 or smaller @xcite . as evidence of such events
, we find that sub-@xmath35 blue - sequence e / s0s are more often blue - centered than the general galaxy population , where blue - centeredness is measured relative to the typical color gradient of galaxies at a given luminosity .
@xcite find blue - centeredness to correlate with morphological peculiarities and companions , which supports the externally triggered gas inflow scenario . for blue - sequence e / s0s
, we find a relationship between blue - centeredness and variations in specific star formation rates relative to typical reference values , suggesting that such inflows may be episodic and trigger bursts of star formation . in summary
, this work clearly confirms that blue - sequence e / s0s have both the gas reservoirs and the potential for sustained star formation necessary for significant disk growth , consistent with evolution toward later - type morphology if the spatial distribution of the gas is extended . while our sample of 27 blue- and red - sequence e / s0s is sufficient for the analysis presented in this paper , it is important to extend our work to a larger sample of sub-@xmath35 e / s0s for more robust statistics . the multi - wavelength resolve survey underway
@xcite would be an ideal data set for such a study .
active follow - up includes obtaining iram co(10 ) and co(21 ) spectra to quantify the molecular gas content in these galaxies , which may be a large fraction of the gas mass content and further extend the potential for morphological transformation in blue - sequence e / s0s .
we are also obtaining vla hi and carma co(10 ) maps to resolve the distribution of the atomic and molecular gas in blue - sequence e / s0s .
hi maps will allow us to look for extended gas disks , small companions , and/or signs of interactions .
co(10 ) maps may reveal inner disky `` pseudobulges '' growing in tandem with extended disks , as would be expected during the formation of late type galaxies .
we thank the anonymous referee for his / her helpful comments .
we are grateful to a. bolatto , e. gawiser , s. khochfar , d. mar , s. mcgaugh , j. rose , e. shaya , and p. teuben for insightful conversations .
we would like to thank the gbt operators and the green bank staff for support during this program ( 07a-072 , 07c-148 ) .
the national radio astronomy observatory is a facility of the national science foundation operated under cooperative agreement by associated universities , inc .
this material is based upon work supported by the national science foundation under grant no .
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 .
we acknowledge the usage of the hyperleda database ( http://leda.univ-lyon1.fr ) and the cornell digital hi archive ( http://arecibo.tc.cornell.edu/hiarchive ) .
the parkes telescope is part of the australia telescope which is funded by the commonwealth of australia for operation as a national facility managed by csiro .
for many galaxies ( ugc 6655 , ugc 7020a , ngc 3011 , ic 1141 , ugc 6570 ) the gbt hi profiles have much stronger s / n than those from the literature , hence any differences in flux measurements between gbt and literature are likely to be attributed to the noisy profiles of the literature data .
this can be seen in the error bars in figure [ fig.fluxdiff ] .
ngc 4117 : the literature data are also of poorer quality and suffer from confusion with a nearby companion ngc 4118 ( 1.5@xmath56 and 643 @xmath44 ) .
we separate out the companion by using the width of the ionized gas from 8501050 @xmath44 , which also agrees with our preliminary co data for this galaxy .
ugc 12265n : one member of a galaxy pair ; the hi is completely blended in with its smaller companion , ugc 12265s , only @xmath147 and @xmath3970 @xmath44 away ( see figure [ fig.gbtspectra_blue ] ) .
preliminary vla hi data suggest that ugc 12265n contains about 1/4 of total hi within the gbt beam , so we use this fraction of the total hi flux for our analysis .
ngc 3522 : hi spectrum from @xcite seem to have comparable s / n to our gbt spectrum , due to our rather short integration time .
the @xmath393@xmath56 beam of the arecibo telescope used by @xcite , however , could be missing some extended hi flux , which would explain the higher flux measurement on our part .
ic 1639 , ic 195 : both have known companions within the beam of the gbt . because these galaxies do not have ionized - gas rotation information , we measure the hi flux within the range of stellar rotation ( 53445444 @xmath44 and 34983798 @xmath59 , respectively ) .
ngc 1029 : does not have ionized - gas or stellar rotation data , so we exclude the companion by measuring the hi flux within a width given by arecibo hi observations of this galaxy .
the smaller 3@xmath56 beam of the arecibo telescope does not detect the companion and finds a @xmath148 of 353 @xmath44 @xcite .
the following galaxies are classified as undetected , though hi emission was detected in velocity ranges corresponding to nearby galaxies .
for these three galaxies , we give both the hi upper limit for the target galaxy as well as the hi flux from the nearby galaxy . ngc 4308
: the measured hi velocity suggests that the flux belongs to a nearby companion , ugc 7438 ( @xmath395@xmath56 , 699 @xmath44 ) , though this is not certain . assigning the gas to ngc 4308 does not change any results since the measured atomic gas mass is still extremely low , @xmath149 of the stellar mass of the galaxy . _
galaxies with flat velocity profiles : _ ngc 3011(b ) , ngc 1552(r ) , ngc 3522(r ) , and ngc 3499(r ) all have hi profiles that are reminiscent of a gas disk extending either slightly into or well beyond the flat part of the rotation curve .
these profiles are not as sharply peaked as the typical double - horned profile , suggesting that the gas disk does not extend too far into the flat part of the rotation curve .
none of these galaxies have large companions within the beam of the gbt , though all of these profiles seem to have slight asymmetries , suggesting there may be distortions in the gas disk . _ galaxies with rising velocity profiles : _ ugc 6003(b ) , ugc 6655(b ) , ugc 6570(mid ) , ugc 7020a(mid ) , and ngc 5338(r ) have sharply peaked hi profiles , which are indicative of gas disks that are still in the rising part of the rotation curve .
most of these are relatively symmetric and do not have any known large companions within the beam of the gbt .
_ galaxies in between the first two cases : _ ugc 6805(b ) , ic 1141(b ) , and ugc 5923(r ) have hi profiles that seem to be a combination of the previous two cases their hi profiles all have a single peak like the rising rotation curve case , but there is a slight flattening suggesting that the gas is reaching the flat part of the rotation curve .
ugc 6805 seems to have a gas disk with rotation in the rising part of the curve ; although the center seems slightly flattened , it is hard to tell given the noise in the hi profile .
ic 1141 and ugc 5923 , on the other hand , seem to have centers that are much more flattened , suggesting that the gas disk may extend a little further into the flat part of the rotation curve .
these galaxies also do not have any known large companions . _ galaxies with known companions : _ ngc 4117(r ) , ugc 12265n(b ) , ic 1639(b ) , and ic 195(b ) all have optically confirmed companions .
ngc 4117 has an asymmetric double horn profile between 800 and 1100 @xmath44 , and a companion ( ngc 4118 ) @xmath150 and 643 @xmath44 away .
ugc 12265n is actually a galaxy pair with ugc 12265s , and its hi profile is completely blended in since the two are less than 100 @xmath44 apart in velocity space .
the larger peak in the hi profile for ic 1639 is actually its larger companion , ic 1640 ; ic 1639 is the smaller peak at 2581 @xmath59 . ic 195 has two companions : ic 196 , which is @xmath151 away , but only about 15 @xmath44 apart in velocity space ( so it is completely blended in ) , and arp 290 , which probably contributes to the hi profile at 3509 @xmath44 , located @xmath152 away . , j. , & merrifield , m. 1998 , galactic astronomy ( galactic astronomy : by james binney and michael merrifield .
princeton , nj : princeton university press , 1998 .
( princeton series in astrophysics ) qb857 .b522 1998 ) , j. , & tremaine , s. 2008 , galactic dynamics : second edition ( galactic dynamics : second edition , by james binney and scott tremaine .
isbn 978 - 0 - 691 - 13026 - 2 ( hb )
. published by princeton university press , princeton , nj usa , 2008 . ) , w. k. , & richter , o .- g .
1989 , a general catalog of hi observations of galaxies . the reference catalog .
( a general catalog of hi observations of galaxies . the reference catalog .
huchtmeier , w.k . , richter , o .- g . , pp .
isbn 0 - 387 - 96997 - 7 .
springer - verlag berlin heidelberg 1989 ) , p. , garwood , r. w. , braatz , j. a. , radziwill , n. m. , & maddalena , r. j. 2006 , in astronomical society of the pacific conference series , vol .
351 , astronomical data analysis software and systems xv , ed .
c. gabriel , c. arviset , d. ponz , & s. enrique , 512+ 1 & a00113p3037 & 130 & ce & r & & & @xmath1440.12 & @xmath1441.36@xmath153 & & gbt + 2 & a00289p0556 & 313 & sd & b & 2085(7 ) & 48(2 ) & 0.78(0.13 ) & 1.65@xmath154 & & le + 3 & ngc193 & 408 & l- & r & 4220 ? & & @xmath1442.0 & @xmath1441.80@xmath155 & & hr + 4 & a00389m0159 & 439 & sa & b & 5299(7 ) & 55(3 ) & 4.94(1.28 ) & 6.54@xmath156 & & le + 5 & a00442p3224 & 484 & sb & b & 4857(8 ) & 197(6 ) & 13.12(4.83 ) & 1.56@xmath157 & & le + 6 & a00510p1225 & 545 & sc & & 18339(8 ) & 175(9 ) & 1.03(0.33 ) & 1.66@xmath158 & & le + 7 & ngc315 & 597 & e & r & 4921 & & @xmath1441.33 & @xmath1441.63@xmath155 & & hr + 8 & a00570p1504 & 615 & sab & b & 5507(3 ) & 164(12 ) & 1.72(0.21 ) & 2.53@xmath159 & & cegg + 9 & a01047p1625 & 685 & i m & & 156(4 ) & 32(1 ) & 11.32(2.04 ) & 5.07@xmath160 & & le + 10 & ngc382 & 688 & e & r & & & & & & le + 11 & ic1639 & 750 & ce & b & 5381(12 ) & 22(10 ) & 0.70(0.08 ) & 9.55@xmath161 & @xmath75 & gbt + 12 & a01123m0046 & 793 & sc & b & & & @xmath1441.98 & @xmath1441.01@xmath162 & & hi + 13 & a01187m0048 & 892 & sa & b & 5228(5 ) & 32(2 ) & 2.37(0.33 ) & 3.04@xmath163 & & cegg + 14 & ngc516 & 946 & l & r & & & @xmath1440.09 & @xmath1442.59@xmath164 & @xmath75 & gbt + 15 & a01300p1804 & 1104 & i m & b & 684(4 ) & 44(2 ) & 7.62(0.97 ) & 2.35@xmath165 & & le + 16 & a01344p2838 & 1154 & sbc & b & 7758(7 ) & 124(5 ) & 5.03(2.38 ) & 1.47@xmath166 & & le + 17 & a01346p0438 & 1155 & sbc & b & 3181(8 ) & 85(4 ) & 2.50(0.42 ) & 1.18@xmath167 & & le + 18 & a01374p1539b & 1176 & i m & & 632(3 ) & 16(0.4 ) & 23.23(2.24 ) & 5.98@xmath168 & & le + 19 & ngc695 & 1315 & sc & b & 9729(6 ) & 148(5 ) & 3.93(0.84 ) & 1.84@xmath169 & & le + 20 & ngc784 & 1501 & sm & & 198(4 ) & 41(1 ) & 45.50(4.39 ) & 2.87@xmath170 & & le + 21 & a02008p2350 & 1551 & sdm & b & 2671(5 ) & 47(2 ) & 11.53(1.11 ) & 4.15@xmath171 & & le + 22 & ic195 & 1555 & s0/a & b & 3648(2 ) & 104(7 ) & 6.85(0.16 ) & 4.39@xmath172 & @xmath75 & gbt + 23 & ic197 & 1564 & sbc & b & 6316(6 ) & 122(9 ) & 8.26(0.86 ) & 1.55@xmath173 & & cegg + 24 & ic1776 & 1579 & sc & b & 3410(6 ) & 48(2 ) & 12.19(3.16 ) & 6.64@xmath174 & & le + 25 & a02056p1444 & 1630 & sb & b & 4437(3 ) & 133(10 ) & 6.39(0.67 ) & 5.94@xmath175 & & cegg + 26 & ngc825 & 1636 & sa & r & 3397(5 ) & 217(10 ) & 5.89(1.52 ) & 3.20@xmath176 & & le + 27 & ngc927 & 1908 & sc & b & 8266(10 ) & 94(4 ) & 4.43(1.15 ) & 1.43@xmath177 & & le + 28 & a02257m0134 & 1945 & sdm & b & 1768(8 ) & 49(3 ) & 2.30(0.2 ) & 3.25@xmath170 & & le + 29 & ngc984 & 2059 & sa & r & 4355(7 ) & 278(18 ) & 10.62(4.59 ) & 1.00@xmath178 & & le + 30 & ngc1029 & 2149 & s0/a & r & 3620(28 ) & 127(9 ) & 8.32(0.15 ) & 5.19@xmath179 & & gbt + 31 & a02464p1807 & 2296 & ce & & & & @xmath1446.9 & 3.39@xmath180 & & hr + 32 & a02493m0122 & 2345 & sm & b & 1506(5 ) & 40(2 ) & 19.68(2.5 ) & 2.03@xmath181 & & le + 33 & ngc1298 & 2683 & ? e & r & 6452(3 ) & 36(2 ) & 0.48(0.09 ) & 9.43@xmath182 & & gbt + 34 & a03202m0205 & 2704 & sa & b & & & @xmath1441.92 & @xmath1446.03@xmath155 & & hi + 35 & ngc1552 & 3015 & l & r & 4784(8 ) & 129(9 ) & 0.7(0.09 ) & 7.76@xmath183 & & gbt + 36 & ngc2692 & 4675 & sa & r & 4026(12 ) & 75(4 ) & 3.03(0.42 ) & 2.37@xmath184 & & le + 37 & a08567p5242 & 4713 & sb & b & 9036(7 ) & 249(9 ) & 3.09(0.27 ) & 1.27@xmath185 & & le + 38 & a09045p3328 & 4787 & sdm & b & 552(8 ) & 53(1 ) & 6.00(1.02 ) & 1.07@xmath186 & & le + 39 & ngc2780 & 4843 & sab & b & 1962(7 ) & 130(10 ) & 0.25(0.06 ) & 5.83@xmath187 & & le + 40 & a09125p5303 & 4879 & i m & r & & & & & & le + 41 & ngc2799 & 4909 & sm & b & 1755(25 ) & 146(13 ) & 7.55(2.5 ) & 1.70@xmath188 & & le + 42 & ngc2824 & 4933 & l & r & & & & & & le + 43 & ngc2844 & 4971 & sa & b & 1489(6 ) & 145(5 ) & 6.46(0.82 ) & 9.75@xmath189 & & le + 44 & ngc3011 & 5259 & s0/a & b & 1543(2 ) & 76(5 ) & 1.24(0.08 ) & 1.93@xmath190 & @xmath75 & gbt + 45 & ngc3009 & 5264 & sc & b & 4564(12 ) & 194(13 ) & 2.95(0.34 ) & 3.42@xmath191 & & le + 46 & ic2520 & 5335 & pec & b & 1238(8 ) & 85(8 ) & 6.00(1.69 ) & 6.31@xmath192 & & le + 47 & a09557p4758 & 5354 & sm & b & 1171(4 ) & 76(3 ) & 17.14(1.48 ) & 1.85@xmath193 & & le + 48 & ngc3075 & 5360 & sc & b & 3582(8 ) & 124(4 ) & 15.07(7.12 ) & 1.00@xmath194 & & le + 49 & a09579p0439 & 5378 & sb & b & 4161(10 ) & 117(4 ) & 6.05(2.86 ) & 5.31@xmath195 & & le + 50 & ngc3104 & 5414 & i m & b & 603(3 ) & 41(2 ) & 17.78(1.36 ) & 4.33@xmath196 & & le + 51 & a10042p4716 & 5451 & i m & r & 629(9 ) & 52(2 ) & 2.84(0.64 ) & 7.17@xmath197 & & le + 52 & ngc3165 & 5512 & i m & r & 1333(4 ) & 61(2 ) & 3.45(0.78 ) & 3.63@xmath198 & & le + 53 & a10114p0716 & 5522 & sc & b & 1220(4 ) & 90(2 ) & 27.92(2.98 ) & 2.62@xmath199 & & le + 54 & ngc3179 & 5555 & l & r & & & @xmath1440.19 & @xmath1445.10@xmath153 & & gbt + 55 & a10171p3853 & 5577 & sm & b & 2028(9 ) & 45(2 ) & 3.30(0.93 ) & 8.61@xmath200 & & le + 56 & ngc3213 & 5590 & sbc & b & 1346(5 ) & 67(4 ) & 1.47(0.35 ) & 2.01@xmath201 & & le + 57 & ngc3264 & 5719 & sdm & b & 940(7 ) & 60(2 ) & 14.93(1.44 ) & 1.16@xmath202 & & le + 58 & ngc3279 & 5741 & sc & b & 1393(6 ) & 156(4 ) & 3.55(0.96 ) & 4.81@xmath203 & & le + 59 & a10321p4649 & 5744 & sc & b & & & & & & le + 60 & a10337p1358 & 5760 & scd & b & 3010(6 ) & 139(5 ) & 3.78(0.33 ) & 1.86@xmath193 & & le + 61 & ic2591 & 5763 & sbc & b & 6797(10 ) & 160(5 ) & 8.75(4.13 ) & 2.04@xmath204 & & le + 62 & a10365p4812 & 5791 & sc & b & 857(5 ) & 67(3 ) & 3.72(1.05 ) & 2.34@xmath205 & & le + 63 & a10368p4811 & 5798 & sc & b & 1519(6 ) & 72(5 ) & 5.37(1.52 ) & 9.36@xmath206 & & le + 64 & ngc3326 & 5799 & sb & b & 8153(8 ) & 100(6 ) & 2.77(0.88 ) & 8.97@xmath207 & & le + 65 & a10389p3859 & 5819 & sbc & b & & & & & & le + 66 & a10431p3514 & 5870 & sa & b & 1992(6 ) & 100(7 ) & 4.59(1.3 ) & 1.24@xmath208 & & le + 67 & a10448p0731 & 5892 & sb & b & & & @xmath1441.53 & @xmath1444.97@xmath155 & & hi + 68 & a10465p0711 & 5923 & s0/a & r & 713(4 ) & 65(5 ) & 3.12(0.16 ) & 4.71@xmath209 & @xmath75 & gbt + 69 & a10504p0454 & 6003 & s0/a & b & 5819(5 ) & 67(5 ) & 1.47(0.09 ) & 2.46@xmath210 & @xmath75 & gbt + 70 & ngc3454 & 6026 & sc & b & 1109(5 ) & 91(3 ) & 5.62(1.26 ) & 5.62@xmath211 & & le + 71 & a10592p1652 & 6104 & sbc & b & 2947(5 ) & 110(2 ) & 7.41(1.26 ) & 3.59@xmath212 & & le + 72 & ngc3499 & 6115 & s0/a & r & 1495(12 ) & 119(9 ) & 0.39(0.06 ) & 7.15@xmath213 & @xmath75 & gbt + 73 & ngc3510 & 6126 & sd & b & 704(3 ) & 79(2 ) & 33.57(2.9 ) & 7.26@xmath214 & & le + 74 & mrk421 & 6132w & pec & & & & & & & le + 75 & ngc3522 & 6159 & l- & r & 1221(8 ) & 112(8 ) & 2.00(0.16 ) & 2.47@xmath215 & @xmath75 & gbt + 76 & a11040p5130 & 6162 & sc & b & 2208(5 ) & 95(4 ) & 12.65(1.74 ) & 4.04@xmath188 & & le + 77 & ic673 & 6200 & sa & b & 3859(8 ) & 151(5 ) & 17.14(8.09 ) & 1.31@xmath216 & & le + 78 & a11068p4705 & 6201 & l- & r & & & & & & le + 79 & a11072p1302 & 6206 & sc & b & & & & & & le + 80 & ngc3605 & 6295 & l- & r & & & @xmath1440.11 & @xmath1441.61@xmath217 & @xmath75 & gbt + 81 & a11142p1804 & 6296 & sc & r & 976(6 ) & 82(3 ) & 2.10(0.72 ) & 8.28@xmath218 & & le + 82 & ngc3633 & 6351 & sa & b & 2599(6 ) & 150(5 ) & 2.90(0.46 ) & 1.07@xmath219 & & le + 83 & ic692 & 6438 & e & b & 1163(8 ) & 35(3 ) & 2.50(0.71 ) & 2.71@xmath220 & @xmath75 & le + 84 & a11238p5401 & 6446 & sd & & 645(4 ) & 57(2 ) & 28.71(3.65 ) & 1.13@xmath221 & & le + 85 & a11310p3254 & 6545 & sb & b & 2630(6 ) & 66(7 ) & 1.00(0.28 ) & 4.20@xmath222 & & le + 86 & ic708 & 6549 & ?
e & r & & & & & & le + 87 & a11332p3536 & 6570 & s0/a & mid & 1628(2 ) & 49(3 ) & 1.35(0.05 ) & 2.60@xmath223 & @xmath75 & gbt + 88 & a11336p5829 & 6575 & sc & b & 1216(5 ) & 95(3 ) & 13.12(1.27 ) & 1.68@xmath193 & & le + 89 & ngc3795a & 6616 & scd & b & 1148(5 ) & 45(2 ) & 11.86(1.63 ) & 1.39@xmath224 & & le + 90 & a11372p2012 & 6625 & sc & b & & & & & & le + 91 & ngc3795 & 6629 & sc & b & 1212(6 ) & 103(4 ) & 7.21(0.62 ) & 7.72@xmath205 & & le + 92 & a11378p2840 & 6637 & l- & b & 1836(6 ) & 61(5 ) & 1.84(0.87 ) & 4.31@xmath225 & @xmath75 & le + 93 & a11392p1615 & 6655 & l & b & 744(2 ) & 26(2 ) & 0.92(0.04 ) & 1.70@xmath226 & @xmath75 & gbt + 94 & ngc3846 & 6706 & sm & b & 1451(8 ) & 66(3 ) & 6.34(0.87 ) & 1.00@xmath221 & & le + 95 & ngc3850 & 6733 & sc & & 1149(8 ) & 69(3 ) & 7.91(1.34 ) & 9.43@xmath227 & & le + 96 & a11476p4220 & 6805 & l & b & 1132(6 ) & 49(3 ) & 0.39(0.05 ) & 3.78@xmath228 & @xmath75 & gbt + 97 & ngc3913 & 6813 & sd & & 954(3 ) & 19(1 ) & 10.52(1.12 ) & 9.06@xmath229 & & le + 98 & ic746 & 6898 & sb & b & 5028(3 ) & 123(4 ) & 6.95(1.1 ) & 9.31@xmath230 & & le + 99 & a11531p0132 & 6903 & sc & & 1892(6 ) & 78(3 ) & 15.35(2.44 ) & 3.54@xmath188 & & le + 100 & ngc3978 & 6910 & sbc & b & 9962(8 ) & 77(3 ) & 3.33(0.71 ) & 1.64@xmath231 & & le + 101 & a11547p4933 & 6930 & sc & & 778(5 ) & 50(2 ) & 27.41(4.06 & 1.53@xmath232 & & le + 102 & a11547p5813 & 6931 & sm & b & 1195(8 ) & 49(2 ) & 4.63(0.89 ) & 5.62@xmath233 & & le + 103 & ngc4034 & 7006 & sc & b & 2367(8 ) & 91(6 ) & 3.21(0.28 ) & 1.19@xmath172 & & le + 104 & a11592p6237 & 7009 & i m & b & 1120(12 ) & 74(6 ) & 6.00(1.02 ) & 6.79@xmath234 & & le + 105 & a12001p6439 & 7020a & l & mid & 1515(1 ) & 39(3 ) & 2.30(0.05 ) & 3.86@xmath235 & @xmath75 & gbt + 106 & ngc4117 & 7112 & l & r & 934(2 ) & 122(9 ) & 5.10(0.20 ) & 4.36@xmath236 & @xmath75 & gbt + 107 & ngc4120 & 7121 & sc & b & 2246(7 ) & 92(4 ) & 10.42(0.9 ) & 3.53@xmath237 & & le + 108 & a12064p4201 & 7129 & sab & r & 947(9 ) & 60(6 ) & 1.22(0.25 ) & 9.69@xmath238 & & le + 109 & ngc4141 & 7147 & sc & r & 1902(7 ) & 70(3 ) & 6.89(0.60 ) & 1.90@xmath219 & & le + 110 & ngc4159 & 7174 & sdm & b & 1753(12 ) & 77(5 ) & 5.68(2.09 ) & 1.29@xmath239 & & le + 111 & ngc4204 & 7261 & sdm & & 858(9 ) & 36(1 ) & 24.10(4.61 ) & 5.80@xmath240 & & le + 112 & ngc4238 & 7308 & sc & b & 2765(6 ) & 113(6 ) & 5.78(0.50 ) & 2.78@xmath241 & & le + 113 & ngc4248 & 7335 & sdm & r & 484(14 ) & 29(3 ) & 4.63(1.31 ) & 8.05@xmath242 & & le + 114 & a12167p4938 & 7358 & sc & b & 3658(6 ) & 132(5 ) & 5.27(1.01 ) & 4.04@xmath243 & & le + 115 & ngc4272 & 7378 & l- & & & & & & & le + 116 & ngc4288 & 7399 & sm & b & 535(4 ) & 77(3 ) & 26.18(3.6 ) & 5.38@xmath244 & & le + 117 & ngc4308 & 7426 & l & r & & & @xmath1440.06 & @xmath1441.01@xmath217 & @xmath75 & gbt + 118 & a12195p3222 & 7428 & sdm & & 1138(6 ) & 28(1 ) & 7.21(2.04 ) & 9.43@xmath245 & & le + 119 & a12195p7535 & & ce & & & & & & & le + 120 & a12263p4331 & 7608 & i m & & 536(5 ) & 24(1 ) & 20.42(2.39 ) & 3.80@xmath246 & & le + 121 & a12295p4007 & 7678 & sd & b & 682(9 ) & 24(1 ) & 5.03(1.42 ) & 1.49@xmath247 & & le + 122 & a12300p4259 & 7690 & sdm & b & 538(11 ) & 35(2 ) & 14.52(2.46 ) & 2.73@xmath248 & & le + 123 & a12304p3754 & 7699 & sd & b & 500(7 ) & 81(3 ) & 20.04(2.34 ) & 2.67@xmath249 & & le + 124 & ngc4509 & 7704 & sm & b & 937(8 ) & 38(4 ) & 3.89(1.84 ) & 2.72@xmath250 & & le + 125 & a12331p7230 & 7761 & sb & b & & & & & & le + 126 & a12446p5155 & 7950 & i m & b & 498(4 ) & 35(1 ) & 5.89(0.63 ) & 1.37@xmath251 & & le + 127 & ngc4758 & 8014 & sbc & b & 1243(6 ) & 79(3 ) & 8.59(1.46 ) & 1.24@xmath252 & & le + 128 & ngc4795 & 8037 & sa & r & 2812(5 ) & 155(7 ) & 1.70(0.23 ) & 7.88@xmath253 & & le + 129 & ngc4807 & 8049 & l+ & r & & & & & & le + 130 & ngc4841b & 8073 & e & & & & & & & le + 131 & ngc4926 & 8142 & l- & r & & & & & & le + 132 & ngc4961 & 8185 & sbc & & 2534(4 ) & 93(3 ) & 12.19(1.94 ) & 4.98@xmath254 & & le + 133 & a13065p5420 & 8231 & sb & b & & & & & & le + 134 & ic4213 & 8280 & scd & b & 815(4 ) & 81(4 ) & 8.51(0.99 ) & 4.44@xmath255 & & le + 135 & a13194p4232 & 8400 & scd & b & & & & & & le + 136 & ngc5117 & 8411 & sc & & 2392(4 ) & 96(3 ) & 9.25(1.47 ) & 3.51@xmath188 & & le + 137 & ngc5173 & 8468 & e & b & 2428(6 ) & 81(7 ) & 4.99(1.06 ) & 2.00@xmath163 & @xmath75 & le + 138 & a13281p3153 & 8498 & sab & b & 7320(8 ) & 290(19 ) & 11.43(5.40 ) & 3.19@xmath256 & & le + 139 & ngc5208 & 8519 & l & r & & & @xmath1442.72 & @xmath1446.36@xmath155 & & hi + 140 & ngc5230 & 8573 & sc & b & 6856(6 ) & 65(2 ) & 8.75(1.39 ) & 2.12@xmath257 & & le + 141 & a13361p3323 & 8630 & sm & b & 2438(7 ) & 83(3 ) & 6.70(0.58 ) & 2.48@xmath258 & & le + 142 & ngc5267 & 8655 & sb & b & & 70(4 ) & 0.78(0.16 ) & 1.47@xmath237 & & le + 143 & a13422p3526 & 8693 & sbc & b & 2438(8 ) & 106(4 ) & 8.36(3.95 ) & 3.41@xmath259 & & le + 144 & ngc5338 & 8800 & l & r & 801(9 ) & 20(1 ) & 0.77(0.12 ) & 1.93@xmath260 & @xmath75 & gbt + 145 & ngc5356 & 8831 & sb & b & 1372(5 ) & 122(4 ) & 4.39(0.84 ) & 6.75@xmath261 & & le + 146 & a13550p4613 & 8876 & s0/a & r & & & @xmath1440.15 & @xmath1444.78@xmath164 & @xmath75 & gbt + 147 & ngc5407 & 8930 & l & r & & & & & & le + 148 & ngc5425 & 8933 & sc & b & 2072(8 ) & 103(5 ) & 6.64(0.78 ) & 2.01@xmath241 & & le + 149 & a14016p3559 & 8984 & l & r & & & & & & le + 150 & ngc5470 & 9020 & sb & & 1026(8 ) & 109(4 ) & 7.69(1.47 ) & 6.75@xmath261 & & le + 151 & ngc5491 & 9072a & sc & b & 5888(7 ) & 214(3 ) & 6.17(1.05 ) & 1.10@xmath262 & & le + 152 & ngc5532 & 9137 & l & r & & & & & & le + 153 & ngc5541 & 9139 & sc & b & & & & & & le + 154 & ngc5596 & 9208 & l & r & 3220(24 ) & 94(7 ) & 0.52(0.14 ) & 3.17@xmath263 & @xmath75 & gbt + 155 & ngc5608 & 9219 & sm & b & 664(6 ) & 46(2 ) & 9.33(1.09 ) & 4.19@xmath264 & & le + 156 & a14305p1149 & 9356 & sc & b & 2226(6 ) & 103(3 ) & 15.92(2.53 ) & 5.18@xmath265 & & le + 157 & ngc5684 & 9402 & l & r & & & & & & le + 158 & ngc5762 & 9535 & sa & b & 1792(8 ) & 80(3 ) & 7.48(1.77 ) & 1.71@xmath171 & & le + 159 & a14489p3547 & 9560 & pec & b & 1218(4 ) & 80(4 ) & 4.47(0.95 ) & 5.99@xmath266 & & le + 160 & a14492p3545 & 9562 & l+ & b & 1257(3 ) & 76(3 ) & 11.97(2.29 ) & 1.79@xmath267 & @xmath75 & le + 161 & ic1066 & 9573 & sab & b & 1576(6 ) & 90(3 ) & 7.69(2.93 ) & 1.43@xmath268 & & le + 162 & a14594p4454 & 9660 & sc & b & 608(9 ) & 29(2 ) & 6.00(1.69 ) & 2.75@xmath269 & & le + 163 & a15016p1037 & & ce & & & & & & & le + 164 & ic1100 & 9729 & scd & b & & & & & & le + 165 & ngc5874 & 9736 & sbc & b & 3128(5 ) & 139(5 ) & 6.82(1.08 ) & 4.12@xmath270 & & le + 166 & ngc5875a & 9741 & sc & b & & & & & & le + 167 & ngc5888 & 9771 & sb & b & & & & & & le + 168 & ic1124 & 9869 & sab & b & 5313(8 ) & 160(11 ) & 3.48(0.90 ) & 5.19@xmath271 & & le + 169 & ngc5940 & 9876 & sab & & 10214(6 ) & 78(3 ) & 1.58(0.25 ) & 7.90@xmath272 & & le + 170 & a15314p6744 & 9896 & sc & & 6466(6 ) & 115(4 ) & 4.34(0.42 ) & 9.72@xmath273 & & le + 171 & ngc5993 & 10007 & sb & b & 9562(9 ) & 76(6 ) & 5.83(1.65 ) & 2.67@xmath274 & & le + 172 & ic1141 & 10051 & s0/a & b & 4389(3 ) & 93(7 ) & 1.82(0.07 ) & 1.98@xmath275 & @xmath75 & gbt + 173 & ic1144 & 10069 & s0/a & b & & & @xmath1440.07 & @xmath1445.37@xmath276 & & gbt + 174 & ngc6007 & 10079 & sbc & b & 10544(5 ) & 184(3 ) & 8.99(2.65 ) & 4.89@xmath277 & & le + 175 & a15523p1645 & 10086 & sc & b & & & @xmath1442.06 & @xmath1446.53@xmath153 & & le + 176 & a15542p4800 & 10097 &
l & r & & & & & & hi + 177 & ngc6020 & 10100 & ?
e & r & & & @xmath1442.10 & @xmath1442.26@xmath155 & & hi + 178 & ngc6123 & 10333 & s0/a & r & 3964(12 ) & 34(3 ) & 3.01(0.85 ) & 2.78@xmath254 & @xmath75 & le + 179 & ngc6131 & 10356 & sc & b & 5061(12 ) & 93(7 ) &
4.85(1.37 ) & 6.86@xmath278 & & le + 180 & ngc6185 & 10444 & sa & b & & & & & & le + 181 & ngc7077 & 11755 & s0/a & b & 1148(6 ) & 47(2 ) & 2.00(0.41 ) & 1.69@xmath279 & @xmath75 & le + 182 & ngc7194 & 11888 & ?
e & r & 8105 ( ? ) & 106 ( ? ) & 6.17(0.58 ) & 2.02@xmath280 & & hi + 183 & a22306p0750 & 12074 & sc & b & 1989(7 ) & 70(3 ) & 1.90(0.32 ) & 4.15@xmath281 & & le + 184 & ngc7328 & 12118 & sab & b & 2824(6 ) & 143(4 ) & 10.33(1.75 ) & 4.33@xmath282 & & le + 185 & ngc7360 & 12167 & e & b & 4685(7 ) & 157(8 ) & 3.96(1.87 ) & 4.31@xmath283 & @xmath75 & le + 186 & a22426p0610 & 12178 & sdm & b & 1931(4 ) & 103(7 ) & 22.18(1.92 ) & 4.46@xmath191 & & le + 187 & a22551p1931n & 12265n & l & b & 5717(2 ) & 103(7 ) & 6.64(0.13 ) & 1.07@xmath284 & @xmath75 & gbt + 188 & ngc7436 & 12269 & e & r & & & & & & le + 189 & ngc7460 & 12312 & sb & b & 3190(6 ) & 90(3 ) & 1.16(0.46 ) & 6.28@xmath285 & & le + 190 & ngc7537 & 12442 & sbc & b & 2678(3 ) & 142(3 ) & 24.55(2.62 ) & 8.74@xmath286 & & le + 191 & ngc7548 & 12455 & l & r & 7991(6 ) & 83(4 ) & 0.29(0.07 ) & 9.02@xmath287 & & le + 192 & a23176p1541 & 12519 & sd & b & 4378(5 ) & 145(5 ) & 4.23(0.32 ) & 4.07@xmath159 & & le + 193 & ngc7620 & 12520 & scd & b & 9583(6 ) & 108(3 ) & 6.70(1.06 ) & 3.11@xmath288 & & le + 194 & a23264p1703 & 12620 & l & r & 6849 & & @xmath1440.16 & @xmath1443.69@xmath153 & & cegg + 195 & ic1504 & 12734 & sb & b & 6274(6 ) & 204(7 ) & 7.28(0.63 ) & 1.39@xmath289 & & le + 196 & ngc7752 & 12779 & sd & b & 5044(8 ) & 170(10 ) & 8.67(1.95 ) & 1.06@xmath290 & & le + 197 & a23514p2813 & 12835 & e & r & & & @xmath1440.20 & @xmath1444.79@xmath153 & & gbt + 198 & a23542p1633 & 12856 & i m & b & 1776(5 ) & 66(2 ) & 11.53(1.11 ) & 2.02@xmath224 & & le + 199 & a04345m0225 & 3104 & l & & 9739(8 ) & 87(6 ) & 1.37(0.39 ) & 4.99@xmath291 & & le + 200 & ngc1517 & 2970 & sc & & 3483(3 ) & 79(3 ) & 10.71(0.93 ) & 6.04@xmath292 & & le + 1 & a00113 + 3037 & 130 & 00:13:56.9 & + 30:52:59 & 2970 & & & & @xmath1440.12 & 0.93 + 11 & ic1639 & 750 & 01:11:46.5 & -00:39:52 & 2700 & 5381(12 ) & 83(35 ) & 54(23 ) & 0.19(0.05 ) & 1.00 + & & & & & & 5523(12 ) & 253(36 ) & 80(24 ) & 0.70 0.08 ) & 1.00 + 14 & ngc516 & 946 & 01:24:08.1 & + 09:33:06 & 4440 & & & & @xmath1440.09 & 0.80 + 22 & ic195 & 1555 & 02:03:44.6 & + 14:42:33 & 1140 & 3648(2 ) & 264(7 ) & 229(4 ) & 4.28(0.12 ) & 1.46 + & & & & & & 3631(3 ) &
474(10 ) & 393(7 ) & 6.85(0.16 ) & 1.46 + 30 & ngc1029 & 2149 & 02:39:36.5 & + 10:47:36 & 870 & 3620(28 ) & 340(83 ) & 267(55 ) & 7.41(0.14 ) & 1.56 + & & & & & & 3609(30 ) & 365(91 ) & 277(60 ) & 8.32(0.15 ) & 1.56 + 33 & ngc1298 & 2683 & 03:20:13.1 & -02:06:51 & 840 & 6452(3 ) & 91(10 ) & 86(6 ) & 0.48(0.09 ) & 1.87 + 35 & ngc1552 & 2683 & 04:20:17.7 & -00:41:34 & 3180 & 4784(8 ) & 318(24 ) & 281(16 ) & 0.70(0.09 ) & 0.98 + 44 & ngc3011 & 5259 & 09:49:41.2 & + 32:13:15 & 2100 & 1543(2 ) & 179(7 ) & 170(5 ) & 1.24(0.08 ) & 1.25 + 54 & ngc3179 & 5555 & 10:17:57.2 & + 41:06:51 & 1740 & & & & @xmath1440.19 & 1.31 + 68 & a10465 + 0711 & 5923 & 10:49:07.6 & + 06:55:02 & 870 & 713(4 ) & 188(11 ) & 149(7 ) & 3.12(0.16 ) & 2.31 + 69 & a10504 + 0454 & 6003 & 10:53:03.8 & + 04:37:54 & 1740 & 5819(5 ) & 234(16 ) & 152(11 ) & 1.47(0.09 ) & 1.17 + 72 & ngc3499 & 6115 & 11:03:11.0 & + 56:13:18 & 6990 & 1495(12 ) & 370(35 ) & 261(23 ) & 0.39(0.06 ) & 0.58 + 75 & ngc3522 & 6159 & 11:06:40.4 & + 20:05:08 & 900 & 1221(8 ) & 308(24 ) & 246(16 ) & 2.00(0.16 ) & 1.82 + 80 & ngc3605 & 6295 & 11:16:46.6 & + 18:01:02 & 900 & & & & @xmath1440.11 & 1.33 + 87 & a11332 + 3536 & 6570 & 11:35:49.1 & + 35:20:06 & 3600 & 1628(2 ) & 155(7 ) & 113(4 ) & 1.35(0.05 ) & 0.82 + 93 & a11392 + 1615 & 6655 & 11:41:50.6 & + 15:58:25 & 3600 & 744(2 ) & 86(5 ) & 63(3 ) & 0.92(0.04 ) & 0.97 + 96 & a11476 + 4220 & 6805 & 11:50:12.3 & + 42:04:28 & 3600 & 1132(6 ) & 137(17 ) & 114(11 ) & 0.39(0.05 ) & 0.90 + 105 & a12001 + 6439 & 7020a & 12:02:37.6 & + 64:22:35 & 2100 & 1515(1 ) & 136(4 ) & 91(3 ) & 2.30(0.05 ) & 0.91 + 106 & ngc4117 & 7112 & 12:07:46.1 & + 43:07:35 & 900 & 950(2 ) & 197(6 ) & 197(4 ) & 2.20(0.13 ) & 1.80 + & & & & & & 839(2 ) & 480(7 ) & 450(5 ) & 5.10(0.2 ) & 1.80 + 117 & ngc4308 & 7426 & 12:21:56.9 & + 30:04:27 & 8070 & & & & @xmath1440.06 & 0.61 + & & & & & & 702(5 ) & 69(15 ) & 56(10 ) & 0.13(0.03 ) & 0.61 + 144 & ngc5338 & 8800 & 13:53:26.5 & + 05:12:28 & 900 & 801(9 ) & 110(27 ) & 48(18 ) & 0.77(0.12 ) & 2.23 + 146 & a13550 + 4613 & 8876 & 13:56:58.0 & + 45:58:24 & 2100 & & & & @xmath1440.15 & 1.24 + 154 & ngc5596 & 9208 & 14:22:28.7 & + 37:07:20 & 1920 & 3220(24 ) & 385(73 ) & 209(49 ) & 0.52(0.14 ) & 1.40 + 172 & ic1141 & 10051 & 15:49:46.9 & + 12:23:57 & 4140 & 4389(3 ) & 237(8 ) & 207(5 ) & 1.82(0.07 ) & 0.92 + 173 & ic1144 & 10069 & 15:51:21.7 & + 43:25:03 & 5700 & & & & @xmath1440.07 & 0.62 + & & & & & & 12206(6 ) & 215(19 ) & 175(13 ) & 0.47(0.05 ) & 0.62 + 187 & a22551 + 1931n & 12265n & 22:57:36.0 & + 19:47:26 & 1140 & & & & 1.66(0.13 ) & 1.44 + & & & & & & 5717(2 ) & 315(7 ) & 227(5 ) & 6.64(0.13 ) & 1.44 + 197 & a23514 + 2813 & 12835 & 23:53:56.7 & + 28:29:34 & 4080 & & & & @xmath1440.20 & 0.79 + & & & & & & 6728(5 ) & 204(16 ) & 182(10 ) & 0.43(0.06 ) & 0.79 + ic 1639 & @xmath293 & 0.01 + ic 195 & @xmath294 & 0.12 + ngc 3011 & @xmath295 & 0.12 + ugc 6003 & @xmath296 & 0.29 + ic 692 & @xmath297 & 0.53 + ugc 6637 & @xmath298 & 0.38 + ugc 6655 & @xmath299 & 0.26 + ugc 6805 & @xmath300 & 0.07 + ngc 5173 & @xmath301 & 0.15 + ugc 9562 & @xmath302 & 3.35 + ic 1141 & @xmath303 & 0.12 + ngc 7077 & @xmath304 & 0.34 + ngc 7360 & @xmath305 & 0.18 + ugc 12265n & @xmath296 & 0.30 + ugc 6570 & @xmath306 & 0.10 + ugc 7020a & @xmath307 & 0.25 + ngc 516 & @xmath308 & 0.003 + ugc 5923 & @xmath309 & 0.51 + ngc 3499 & @xmath310 & 0.01 + ngc 3522 & @xmath311 & 0.07 + ngc 3605 & @xmath312 & 0.002 + ngc 4117 & @xmath313 & 0.05 + ngc 4308 & @xmath314 & 0.003 + ngc 5338 & @xmath297 & 0.04 + ugc 8876 & @xmath315 & 0.004 + ngc 5596 & @xmath316 & 0.03 + ngc 6123 & @xmath317 & 0.10 & & & & + & forward & -0.45 @xmath58 0.05 & 4.09 & 0.45 + & bisector & -0.72 & 6.70 & + * spiral / irregular : * & & & & + & forward & -0.45 @xmath58 0.05 & 4.04 & 0.44 + & bisector & -0.70 & 6.49 & + * red sequence : * & & & & + & forward & -0.67 @xmath58 0.12 & 5.23 & 0.58 + & bisector & -1.46 & 13.0 & + * e / s0 : * & & & & + & forward & -0.67 @xmath58 0.12 & 5.17 & 0.59 + & bisector & -1.57 & 13.9 & + ngc 3011 & b & 6.7 & 0.06 & 1.0 + ugc 6003 & b & 1.8 & 0.12 & 2.5 + ic 692 & b & 12.1 & 0.15 & 1.3 + ugc 6637 & b & 6.4 & 0.17 & 1.6 + ugc 6655 & b & 3.1 & 0.09 & 1.0 + ugc 6805 & b & 1.1 & 0.06 & 1.1 + ngc 5173 & b & 21.1 & 0.17 & 0.7 + ugc 9562 & b & 38.3 & 0.10 & 1.6 + ic 1141 & b & 2.5 & 0.07 & 1.3 + ngc 7077 & b & 2.7 & 0.17 & 1.1 + ngc 7360 & b & 168.5 & 0.15 & 1.1 + ugc 12265n & b & 1.4 & 0.17 & 2.4 + ugc 6570 & mid & 2.7 & 0.22 & 0.9 + ugc 7020a & mid & 1.9 & 0.28 & 1.0 + ugc 5923 & r & 18.7 & 0.04 & 0.7 + ngc 5338 & r & 0.8 & 0.40 & 0.4 | recent work has identified a population of low - redshift e / s0 galaxies that lie on the blue sequence in color vs. stellar mass parameter space , where spiral galaxies typically reside . while high - mass blue - sequence e / s0s often resemble young merger or interaction remnants likely to fade to the red sequence , we focus on blue - sequence e / s0s with lower stellar masses ( @xmath0 a few @xmath1 ) , which are characterized by fairly regular morphologies and low - density field environments where fresh gas infall is possible .
this population may provide an evolutionary link between early - type galaxies and spirals through disk regrowth . focusing on atomic gas reservoirs , we present new gbt hi data for 27 e / s0s on both sequences as well as a complete tabulation of archival hi data for other galaxies in the nearby field galaxy survey .
normalized to stellar mass , the atomic gas masses for 12 of the 14 blue - sequence e / s0s range from 0.1 to @xmath21.0 , demonstrating that morphological transformation is possible if the detected gas can be converted into stars .
these gas - to - stellar mass ratios are comparable to those of spiral and irregular galaxies and have a similar dependence on stellar mass . assuming that the hi is accessible for star formation
, we find that many of our blue - sequence e / s0s can increase in stellar mass by 1060% in 3 gyr in both of two limiting scenarios , exponentially declining star formation ( i.e. , closed box ) and constant star formation ( i.e. , allowing gas infall ) . in a constant star formation scenario ,
about half of the blue - sequence e / s0s require fresh gas infall on a timescale of @xmath33 gyr to avoid exhausting their atomic gas reservoirs and evolving to the red sequence .
we present evidence that star formation in these galaxies is bursty and likely involves externally triggered gas inflows .
our analysis suggests that most blue - sequence e / s0s are indeed capable of substantial stellar disk growth on relatively short timescales . |
in a recent workshop convened in aspen , rob kennicutt gave a striking summary of the convergence of estimates of the hubble constant and the corresponding shrinking of error bars .
one of the most important contributions of the hubble space telescope has been , for the first time , accurate estimates of the errors . as kennicutt pointed to a plot of the value of the hubble constant
versus time since 1990 and the convergence of estimates to a value in the mid to high 60 s ( in units of km s@xmath1 mpc@xmath1 ) one of the authors of this paper ( jcw ) piped up to say , i knew that , " and kennicutt , without missing a beat , his back to the audience , said
i knew some butthead would claim he knew the right value all along . "
one purpose of this review is to summarize the arguments that have led theorists studying physical models of type ia supernovae ( sn ia ) to predict that when the dust settled , the hst key project would find the hubble constant to be @xmath2 km s@xmath1 mpc@xmath1 , as has , in fact , been the case .
there is some cause for taking this expected but welcome confirmation as proof that we are on the right track in our interpretation of the physical nature of sn ia and encouragement to employ our growing physical understanding to determine the systematic effects that are expected to dominate the use of sn ia to determine the deceleration parameter and perhaps even the cosmological constant .
the status of work on sn ia is given in 2 .
the use of type ii supernovae ( sn ii ) as a complementary tool to measure the hubble constant is discussed in 3 .
a brief summary is given in 4 .
the standard model of a sn ia is a thermonuclear explosion of a carbon / oxygen white dwarf . to first order , models for such an explosion put constraints on the luminosity that in turn constrain the hubble constant .
the reason is that the nuclear energy liberated has two correlated , but independently measurable effects .
the energy of the explosion determines the expansion velocity which can be determined from the doppler shift of spectral features .
the energy is supplied by burning carbon and oxygen to intermediate mass and iron - peak elements .
the principle iron peak element produced is that having equal numbers of protons and neutrons , as does the fuel , and that is .
the decay of and its daughter product , however , determine the bolometric luminosity of the explosion with dominating near maximum light .
thus , to first order , the amount of and hence the luminosity , is closely related to the expansion velocity .
this correlation was pointed out by sutherland & wheeler ( 1984 ) and its direct implications for the hubble constant were summarized by arnett , branch , & wheeler ( 1985 ) .
it was not possible at that time to put precise limits on the hubble constant because of simplifications and some uncertainties in the radiative transfer , but some limits were clear .
if the amount of nuclear burning were too low , then the white dwarf could not be unbound , never mind give the requisite expansion velocity of @xmath0 10,000 km s@xmath1 .
this meant that this class of model could not be consistent with a hubble constant of 100 km s@xmath1 mpc@xmath1 and even 70 km s@xmath1 mpc@xmath1 caused some discomfort . at the other extreme
, such a model could not produce more than its entire mass , a chandrasekhar mass , of .
this constraint and the observation that rise times of sne ia are at least 2 weeks , put an absolute upper limit on the brightness and a lower limit on h@xmath3 of 40 km s@xmath1 mpc@xmath1 .
this is clearly an unphysical lower limit because we know from the spectrum that a substantial portion of the outer layers of the ejecta of an sn ia are not iron peak , but intermediate mass elements .
exactly how outrageous this lower limit was could not be assessed precisely at the time , but a model with about 0.6 @xmath4 of nickel ( model w7 of nomoto , thielemann , & yokoi 1984 ) gave a reasonable light curve and , even more importantly , spectrum ( harkness 1986 , 1991 ) .
the luminosity of this model applied to normalize the hubble diagram of sn ia implied that a preferred value of the hubble constant was @xmath0 60 km s@xmath1 mpc@xmath1 .
subsequent work has made crucial refinements in both the physics of the explosion and the sophistication of the radiative transfer that allows one to go from the bolometric luminosity given by radioactive decay to the observed multi - color light curves .
this work brought a recent critical advance in understanding that kept pace with important observational developments . on the observational side , the long suspected inhomogeneity of sne ia ( pskovskii 1977 ) has been confirmed ( phillips 1993 ; hamuy , et al .
there is no longer any basis for assuming they are intrinsically ideal standard candles . "
rather , this inhomogeneity gives an opportunity for a deeper understanding of the underlying evolutionary and physical processes .
this diversity must also be taken into account in using sne ia to determine cosmological distance scales .
this can be done to some extent by empirical methods ( riess , press , & kirshner 1996 ) , but a physical understanding is both intrinsically preferable and necessary to fully integrate the changing properties of sne ia with the structure and evolution of their host galaxies .
of particular interest is the light curve brightness / decline relation .
observations show that the dimmer sne ia decline more rapidly than brighter events ( phillips 1993 ; hamuy , et al . 1996 ) .
a one - parameter version of this relation has been used to empirically calibrate sne ia in the context of estimates of the hubble constant ( riess , press , & kirshner , 1996 ) and the deceleration parameter ( kim , et al .
1996 ) , but one expects and observes scatter around the mean brightness / decline relation that must be understood and incorporated in further work . on the physical side ,
there have been important breakthroughs in understanding the possible nature of the thermonuclear combustion of sn ia .
there is now strong reason to believe that the explosion proceeds from a phase driven by a subsonic turbulent combustion , a deflagration , followed by a transition to a shock - driven , supersonic detonation phase ( khokhlov 1991 ) .
spherically symmetric models that treated the density at which this deflagration to detonation transition occurred as a free parameter revealed a new effect . the nuclear energy that gave rise to the expansion velocity
could be provided by burning predominantly to intermediate mass elements , especially silicon . the critical nuclear energy to unbind the white dwarf and provide
the expansion velocity is thus proportional to the sum of the si peak material and ni ( hflich , khokhlov & wheeler 1995 ) .
depending on the value of the transition density , however , more or less would be produced which would in turn produce different values of the luminosity .
this class of models , while generically closely related to earlier ones , broke the tight relation between the mass of and the expansion velocity .
models with less ( but with substantial total thermonuclear energy release , @xmath0 10@xmath5 ergs ) were not only dimmer , but also cooler .
the lower temperatures resulted in lower opacity and faster rates of decline .
these models account qualitatively and even quantitatively for the observed brightness - decline rate relation ( hflich , khokhlov & wheeler 1995 ; hflich , et al .
1996b ) .
these new developments were applied to the determination of the hubble constant by hflich and khokhlov ( 1996 ) .
this paper presented a large variety of models computed with different assumptions concerning the combustion physics , for each of which multi - color light curves were computed .
the result was an ensemble of models from which one could chose the ones that best matched the observations , or , more precisely , one could reject models that clearly failed to match the observations .
models that provided reasonable fits to the light curves around maximum in a wide variety of photometric bands constrain the luminosity rather tightly .
thus one can estimate the brightness ( as well as reddening ) of each individual supernova and determine a corresponding point on the hubble diagram .
the result is the currently best guess for the hubble constant using only physical models and observations of individual supernovae with no reference to local calibrators .
the result obtained by hflich & khokhlov is 67@xmath69 km s@xmath1 mpc@xmath1 ( fig .
1 ) . the error is a 2 @xmath7 " value ( i.e. the 95 % probability limit for a non - gaussian error distribution ) obtained by assessing the probability that a given range of models fit a given supernova .
the value and error estimate are in excellent accord with recent estimates of h@xmath3 based on distances to cepheid variables . from hst observations of @xmath8-cephei stars in ic4182 , ngc5253 , and ngc4536 distance moduli have been found to be @xmath9 , @xmath10 , and @xmath11 mag , respectively ( freedman , et al . 1994 ; sandage , et al .
1994 ; tammann 1996 ) which also compare well with the estimates of hflich & khokhlov of @xmath12 , @xmath13 , and @xmath11 mag , respectively .
the spread for nearby sne ia in fig .
1 is consistent with the cobe dipole field .
hflich & khokhlov were also able to show ( based on one very distant supernova , sn 1988u ) that @xmath14 does not vary significantly from redshift of 0.1 to 0.5 and to put rough limits on the deceleration parameter , q@xmath3 = 0.7@xmath61 .
6.5 cm -4.0 cm the bolometric luminosity in the models basically depends on the decay energies which are well known .
the translation between total luminosity and brightness in a given band may be a point of concern .
the bolometric correction , defined to be the conversion factor between the total luminosity and that in the v band , can be tested with observed spectra in a model - independent way because it depends only on the flux distribution .
figure 2 gives the spectrum of sn 1992a and that of the sun normalized to give the same total flux .
models give a bolometric correction of @xmath0 0.1@xmath15 , consistent with the normalized observed sn spectrum in the v band .
a bc of 0.3 would clearly be too large to be inconsistent with the flux in the observed supernova spectrum in the v band .
the models reproduce the empirical constraints on bc to within @xmath6 0.1@xmath15 , implying a systematic error of less than 10 % ( hflich & khokhlov 1996 ) .
6.5 cm -6.cm this work strongly suggests that the physical and radiative transfer models of sn ia are on the right track and can be used with some confidence in the next exciting phase as one attempts to use high redshift supernovae ( perlmutter , et al 1995 ; kim , et al .
1996 ; schmidt 1997 ) to determine q@xmath3 and @xmath16 . with the great productivity of the distant supernova searches , internal errors due to
small number statistics will soon become of vanishing significance .
the critical issue will be to determine the systematic effects as one looks back to earlier eras when metallicities , progenitor evolution , even galaxy evolution may be significantly different .
riess , press , & kirshner ( 1996 ) have shown the power of correcting for the brightness / decline relation . by so doing ,
the scatter around the hubble flow line is reduced from @xmath17 if sn ia are assumed to be standard candles to @xmath18 if their multi - color techniques are used to determine peak luminosity .
some reservation about this empirical technique has been expressed by noting that some of the nearest sn ia are the brightest , a distinctly counter - intuitive result .
this needs to be better understood , but it is probably the result of low number statistics for nearby supernovae and the fact that spiral galaxies have , in the mean , somewhat brighter supernovae . the local sample is dominated by spiral galaxies .
another aspect of this issue was recently uncovered by wang , hflich , & wheeler ( 1997 ) .
they examined the radial distribution of supernovae in galaxies . using the well - calibrated calan - tololo sample of sn ia
, they showed that at galactocentric radii less than @xmath19 kpc , sn ia show the full dispersion of peak luminosity reflected in the brightness / decline relation . beyond @xmath19 kpc ,
however , the dispersion in peak brightness drops dramatically . the intrinsic dispersion with no correction whatever for the brightness / decline relation is @xmath20 , comparable to that obtained after correction by riess , et al .
the reason for this remarkable change in the properties of sn ia demands explanation in terms of progenitor evolution , metallicity , etc .
it suggests that their may yet be a way of picking a sample of sn ia which do represent nearly standard candles , " but it remains abundantly clear that this is not the case for the full sample of sn ia .
future work , both on understanding the progenitor evolution and physics of sn ia and on purely empirical calibration methods must focus on departures from a one - parameter light curve brightness / decline relation . the observational data , for instance from the calan - tololo survey already show a dispersion that can not be fit by a one parameter curve ( hamuy , et al . 1996 , hflich , et al . 1996b ) .
sn 1994d is a particular case in point .
it is too bright and blue for its light curve shape ( hflich 1995a ) .
there are also abundant theoretical reasons to think that there should be a scatter in properties in terms of initial metallicity , rotation , and mass accretion rate . even if the progenitor evolution is identical , the randomness associated with the site of ignition of carbon many points and with subsequent turbulent burning is sure to impose some dispersion of final properties .
current work on the physics of sn ia promises deeper understanding of these issues
. the carbon runaway will grow out of a preliminary phase of quasi - static convective carbon burning .
the resulting dynamical runaway may start in the center or off center in one or many points ( so called spotty ignition ) as determined by the temperature distribution on the convective phase .
the first full 3-d calculations of the deflagration of a carbon / oxygen white dwarf have been done by khokhlov ( 1995 ) .
these calculations show that the expansion and freezing of the flow is important and seems to be consistent with models in which expansion quenches the subsonic deflagration , thus allowing a turbulent mixing of warm ash and cold fuel .
subsequent recompression is likely to trigger a detonation .
this problem of deflagration to detonation transition is currently under intense study ( khokhlov , oran , & wheeler 1997a , b ; niemeyer and woosley 1997 ) .
future 3-d calculations will use adaptive mesh techniques to extend the resolution of the turbulent flame to the scale where turbulence disrupts the laminar flame ( the gibson " scale ) to provide a more accurate model of the turbulent burning phase and a natural transition to detonation .
these models should go a long way to removing the current free parameter of the transition density from deflagration to detonation and leave the progenitor evolution as the major unknown .
5.5 cm -2.6 cm = 5.5 cm 9.0 cm -2.5 cm = 8.5 cm the explosion will inevitably be a function of the redshift and hence age of the host galaxy , the mass accretion rate , metallicity , rotation and other factors .
an example of how the metallicity could alter the resulting spectrum is given in figure 3 from hflich , et al ( 1997 ) . in this calculation , the metallicity of the progenitor white dwarf is varied from solar to 1/3 solar .
it is important to note that this variation must be imposed on the progenitor , not added after the explosion , because it changes the nucleosynthesis during the explosion .
the principle effect is to alter the iron peak abundances and these in turn primarily affect the spectrum in the blue .
one result of this is substantially different colors , say b - v , as the small differences in the rest frame u band spectra are redshifted to b and then to v. figure 3 shows that the differences in b - v can be up to @xmath21 .
some groups doing supernova searches at large redshifts ( perlmutter , et al , 1995 ) search only in the ccd band and clearly will have to account for such variations .
other groups ( schmidt 1997 ) , employ specially constructed filters so that they are effectively always looking at the rest frame v band .
even with this technique , care must be taken that the red - shifted filter matches the standard v - band filter ( whatever standard one chooses ) and that the match remains precise over the observed redshift range .
sn ii represent an important complement to sn ia as a technique to measure cosmological distances . in principle
, the distance to each individual sn ii can be measured by the baade - wesslink " or
expanding photosphere " method ( note the photosphere actually retreats in mass with time ) in which the ratio of the observed apparent flux to the intrinsic absolute flux gives the anglular size and the velocity and elapsed time give the radius .
the critical aspect of this method is to determine the intrinsic flux of the atmosphere where scattering effects are important . in simple terms , scattering results in a
dilute black body " and this dilution factor which is , in general , a function of frequency and time , must be determined .
schmidt , et al ( 1994 ) have computed the behavior of this dilution factor over a range of temperatures and for a variety of models and have compared their results to sn 1987a and other sn ii .
their dilution factor is unity at low temperatures , drops rather rapidly by a factor of about 4 by 9000 k , and then remains constant at higher temperatures .
this dilution factor has been used to determine the distances to a number of sn ii and to derive an estimate of the hubble constant , h@xmath3 = 73 km s@xmath1 mpc@xmath1 .
there are a number of reasons to think that this issue can benefit from further examination .
the power of this method was illustrated by sn1987a where many groups agreed within 10 % in distance ( @xmath22 5 kpc , chilukuri & wagoner 1988 ; @xmath23 4 kpc , hflich 1987 , 1988 ; @xmath24 4 kpc , schmutz , et al .
1990 ; @xmath25 5 kpc , eastman & kirshner 1989 ) .
this was , however , by far the best observed sn with good time coverage and knowledge of the time of the explosion . for the baade - wesselink method , exact knowledge of the time of the explosion can be important .
the error in the distance scales as @xmath26/t .
typically , the time of the explosions of other supernovae are only known to within a few days at best , implying an error of @xmath27 in the first few weeks . later on ,
when this source of error becomes small , the photospheres of sn ii tend to be about 5000 k in the recombination phase of the plateau which puts the dilution factor on the steeply declining part of the function .
this clearly calls for especially careful treatment , by fitting individual supernovae . although all groups ( hflich 1991 , schmidt , et al .
1994 , baron , et al .
1995 ) agree on the general relation between dilution factor and photospheric temperature including the change of the dilution factor during the recombination phase , hflich ( 1991 ) and baron , et al .
( 1994 ) found a value of the dilution factor for high temperatures ( i.e. @xmath0 9000k ) that was higher by about 30 and 50 % , respectively , than that of schmidt , et al .
the estimate of baron , et al . was in the context of sn 1993j , not a standard sn ii , but their models were rather generic .
moreover , depending on the model parameters , both hflich and baron , et al . found larger scatter around the mean , even at large temperatures , than schmidt , et al .
clocchiatti , et al .
( 1995 ) found that in the early stages of the expansion of the fireball of sn 1993j , the atmosphere was hot , @xmath28 k , but the dilution factor was essentially unity , in sharp contrast with the asymptotic value of schmidt , et al .
the difference can be attributed to the apparent steep density profile of the atmosphere of sn 1993j ( brick walls radiate like black bodies ) , but this result suggests another cause for caution .
the range of estimates of the dilution factor by schmidt , et al .
, hflich , and baron , et al . might be regarded as the intrinsic range in uncertainty . at the very least the source of these discrepancies
should be understood .
the ability of polarimetry of supernovae to give unique information about intrinsic asymmetries has long been discussed ( shapiro & sutherland 1982 ; mccall , et al . 1984 ) .
sn 1987a also represented a breakthrough in this area by providing the first detailed record of the spectropolarimetric evolution ( mendez , et al .
1988 ; cropper , et al .
sn 1993j also provided a wealth of data which is still being analyzed ( trammell , hines , & wheeler 1993 ; hflich 1995b ; doroshenko , et al .
1995 ; hflich , et al .
1996a ; tran , et al .
more recently we have begun a program at mcdonald observatory to attempt to get routine spectropolarimetry on all accessible supernovae .
this program has nearly doubled the number of supernovae for which polarimetry is available .
the early qualitative conclusion was that all type ii are polarized at about the 1 percent level and that type ia are much less polarized , less than 0.1 - 0.2 percent .
( wang , et al . 1996 ) .
that trend continues without exception .
there are a number of ways in which the emission from supernovae could be polarized .
there could be asymmetries associated with the circumstellar medium , especially aided by the large scattering cross section of dust ( wang & wheeler 1996 ) .
there could be asymmetries in the ejecta that reflect initial distortions in the outer envelope due to rotation , filling a roche lobe , or other influences ( hflich 1991 ; jeffery 1991 ) .
a symmetric explosion in an envelope distorted by rotation or by filling a roche lobe can give a prolate envelope and an oblate core ( steinmetz & hflich 1992 ) .
the reason is that shocks first propagate up the axis where the density gradient is steeper , but then sideways to the equatorial axis where they meet and eject matter in the equatorial plane giving a prolate , pancake - like geometry .
pressure waves are then sent into the core by momentum conservation , compressing the core along the equatorial plane .
this tends to yield an oblate , cigar - shaped core .
both a distortion of the envelope and dust scattering may have played a role in sn 1987a . if the source of illumination is off - center even in an otherwise spherically symmetric density distribution , the emergent light will be polarized ( hflich 1995b ) .
this type of asymmetry is also established in other ways in sn 1987a and , to a certain extent , in sn 1993j .
finally , and perhaps most importantly , the intrinsic explosion process , core collapse or thermonuclear explosion , may impose asymmetry .
the polarization of sn ic 1997x , a bare , non - degenerate carbon / oxygen core , may give strong indication of this effect .
polarimetry may thus provide a unique ability to determine the asymmetry of the explosion mechanism and hence special constraints on the physics of the explosion independent of any subsequent instabilities or other symmetry - breaking phenomena .
the first several sn ii events with confirmed polarization , sn 1987a , sn 1993j , sn 1994y ( a narrow emission line event ) , were odd in some way , so there was some question of whether the trend identified by wang , et al ( 1996 ) that all sn ii were substantially polarized was due to a sample that was unrepresentative of normal sn ii .
observations of sn 1996w help to remove that concern .
sn 1996w was a perfectly normal type ii plateau event .
it was observed shortly after the first hump " in the visual at the very early stage of the plateau at the beginning of the recombination wave phase .
sn 1996w , like the other sn ii , was polarized at about the 1 percent level .
because sn iip are bright red supergiants with very large envelopes , rotational distortion of the envelope seems unlikely and any dust - scattered light should be rather dilute .
the envelope should be very optically thick at the observed phase , so it is unlikely that any effect of off - center nickel blobs would be manifest .
this may mean that the polarization observed in sn 1996w was due to some asymmetry associated with the explosion process .
the observations of sn 1996w imply that polarization of sn ii is ubiquitous .
this conclusion may have implications for use of sn ii to measure the distance scale . to generate the observed polarizations , envelope distortions of 10 - 50 percent are required ( hflich 1995b ) , although this may not be necessary in dust scattering models . for such distortions , the luminosity can be strongly asymmetric with differences of up to a factor two along the polar and equatorial directions . if the asymmetric luminosity were distributed in the same way as orientation on the sky , this would have no affect on a large sample of supernovae , because the effect would average out .
the luminosity is not , however , a linear function of the inclination angle .
therefore , the distribution of luminosity could have a net systematic bias which will affect the use of sn ii to make distance estimates .
further study of the polarization of supernovae is thus warranted in this regard .
the most recent estimate for the value of the hubble constant by the hubble key project ( freedman , et al .
1997 ) is 67@xmath29(internal)@xmath30(systematic ) km s@xmath1 mpc@xmath1 . in this work , local distances based on cepheid variables
are extrapolated out to the hubble flow using the tully - fisher relation , the surface brightness fluctuation method or sn ia , all of which provide the same relative distances .
the issue has been the absolute calibration , which the cepheids are now providing .
note that it is somewhat circular to use empirical determinations of sn ia as one means of extrapolation when this technique is itself dependent on the cepheids to determine the absolute scale . in any case
, the two methods agree .
riess , press , & kirshner get h@xmath3 = 67@xmath30 km s@xmath1 mpc@xmath1 . in an important complementary study , sandage , et al .
( 1996 ) have determined the distances to cepheid variables that have specifically been host to sn ia .
they have obtained h@xmath3 = 56@xmath31(internal ) km s@xmath1 mpc@xmath1 based on b magnitudes , and h@xmath3 = 58@xmath31(internal ) km s@xmath1 mpc@xmath1 based on v magnitudes for the sn ia .
this analysis assumes that sn ia are standard candles , which is not valid .
a correction for the brightness / decline relation would give a value that is quite consistent with those quoted above .
the current empirical values for the hubble constant based on sn ia and cepheids are consistent with the value obtained by direct comparison of models with observations of sn ia with no reference to local calibrators , h@xmath3 = 67@xmath32 km s@xmath1 mpc@xmath1 ( hflich & khokhlov 1996 ) .
further work on the physics of the explosion and the use of spectral , rather than broad band , evolution should be able to reduce the error in this method .
using sn ii , schmidt , et al ( 1994 ) obtain h@xmath3 = 73@xmath33(internal)@xmath34(systematic ) km s@xmath1 mpc@xmath1 .
this is consistent , within the errors , with estimates based on sn ia .
if , however , the dilution factor of baron , et al ( 1994 ) were used at face value , this estimate would drop to h@xmath3 @xmath0 50 km s@xmath1 mpc@xmath1 and with that of hflich ( 1991 ) , the value would be h@xmath3 @xmath0 65 km s@xmath1 mpc@xmath1 . some improvement could probably be made in the estimates based on sn ii , including possible systematic effects of asymmetries .
improvements are still necessary , but discussion of the hubble constant concerns uncertainties of 10 - 20 percent , no longer a factor of two .
recent developments show that things are still somewhat in a state of flux on the observational side .
there is a rumor that hipparchos will recalibrate the distances to cepheids and increase the distance scale by @xmath010 percent ; however , kochanek ( 1997 ) has examined the color effects of the cepheid period / luminosity relation and deduced that the distance scale must be decreased by @xmath010 percent .
these effects may prove to cancel , or one or the other may be otherwise qualified .
a colleague recently remarked that the theoretically - based methods using either sn ia or sn ii are doing quite well and if the observations threaten to depart from those results , one should question the observations .
this is no time for the theorists to lose heart !
we thank our colleagues at rice for the opportunity to summarize these issues and numerous colleagues who work on the distance scale for stimulating discussions over many years .
this work was supported in part by nsf grant ast9528110 , nasa grant nag52888 and a grant from the texas advanced research program .
arnett , w. d. , branch , d. , & wheeler , j. c. 1985 nature , 314 , 337 avrett , e. h. 1992 , in _ solar electromagnetic radiation _
donnelly ( noaa erl ) , p.20 baron , e. , et al .
1995 , apj 441 , 170 bessell , m. s. 1990 , pasp , 102 , 1181 chilukuri , m. , wagoner , r. v. 1988 , lect .
notes in physics , 305 , 295 clocchiatti , a. , et al .
1995 , apj , 446 , 167 cropper , m. , et al .
1988 , mnras , 231 , 695 doroshenko , v. t. , efimov , y. s. , & shakhovskoi , n. m. 1995 , astro .
lett . , 21 , 513 eastman , r. g. , kirshner , r. p. 1989
, apj 347 , 771 freedman , w. l. , et al .
1994 , apj , 457 , 628 freedman , w. l. , et al 1997 , in preparation and private communication harkness , r. p. 1986 , in _ radiation hydrodynamics in stars and compact objects _ , eds .
d. mihalas and k .- h .
a. winkler ( berlin : springer press ) , p. 166
harkness , r. p. 1991 , in _
sn 1987a and other supernovae _ , eds .
j. danziger & k. kjr ( garching : eso ) , p.447 hamuy , m. , et al .
1993 , aj , 106 , 2392 hamuy , m. , et al .
1996 , aj , 112 , 2408 hflich p. 1987 , in _
sn 1987a _ , ed .
i.j . danziger ( garching : eso ) , p. 449
hflich p. 1988
7 , 434 hflich p. 1991 , in _ supernovae _ , ed .
woosley ( new york : springer - verlag ) , p.415 hflich , p. 1995a
, apj , 443 , 89 hflich , p. 1995b
, apj , 440 , 821 hflich , p. , khokhlov , a. , & wheeler , j. c. 1995 , apj , 444 , 831 hflich , p. & khokhlov , a. 1996 , apj , 457 , 500 hflich , p. , wheeler , j. c. , hines , d. c. , & trammell , s. t. 1996a , apj , 459 , 307 hflich , p. , khokhlov , a. , wheeler , j. c. , phillips , m. m , suntzeff , n. b. , & hamuy , m. 1996b , apj , 472 , 81 hflich , p. , et al 1997 , in _ thermonuclear supernovae _ , eds .
p. ruiz - lapuente , r. canal , and j. isern , ( dordrecht : kluwer ) , p. 659
jeffery , d. j. 1991 , apjs , 77 , 405 khokhlov , a. 1991 , a&a , 245 , 114 khokhlov , a. 1995 , apj , 449 , 695 khokhlov , a. , oran , e. s. , & wheeler , j. c. 1997a , combustion & flame , 108 , 503 khokhlov , a. , oran , e. s. , & wheeler , j. c. 1997b , apj , 478 , 678 kim , a. , et al .
1996 , apj , 477 , 183 kirshner , r. p. , et al .
1993 , apj , 415 , 589 kochanek , c. s. 1997 , apj , in press kohl , j. l. , parkinson , w. h. , & kurucz , r. l. 1992 , in _ center to limb solar spectrum _ , nbs mccall , m. l. , reid , n. , bessell , m. s. , & wickramasinghe , d. 1984 , mnras , 210 , 839 mendez , m. , clocchiatti , a. , benvenuto , o. g. , feinstein , c. , & marraco , h. g. 1988 , apj , 334 , 295 niemeyer , j. c. & woosley , s. e. 1997 , apj , 475 , 740 nomoto , k. , thielemann , f .- k .
, & yokoi , k. 1984 , 268 , 644 perlmutter s. , et al . 1995 , apj , 440 , l41 phillips , m. m. 1993 , apj , 413 , l108 pskovskii , yu .
astr . , 21 , 675 riess , a. g. , press , w. h. , & kirshner , r. p. 1996 , apj , 473 , 588 sandage , a. , et al .
1994 , apj , 423 , l13 sandage , a. , et al .
1996 , apj , 460 , l15 schmidt , b. p. , et al .
1994 , apj , 432 , 42 schmidt , b. p. 1997
, in _ thermonuclear supernovae _ , eds .
ruiz - lapuente , r. canal , and j. isern , ( dordrecht : kluwer ) , p. 765 schmutz , w. , et al . 1990 , apj , 355 , 255 shapiro p. r. & sutherland , p. g. 1982 , apj , 263 902 steinmetz , m. & hflich p. 1992 , a&a , 257 , 641 sutherland , p. & wheeler , j. c. 1984 , apj , 280 , 282 tammann , g. 1996 , pasp 108 , 1083 trammell , s. r. , hines , d. c. , & wheeler , j. c. 1993 , apj , 414 , l21 tran , h. , et al .
1997 , pasp , in press wang , l. , wheeler , j. c. , li , z. w. , & clocchiatti , a. 1996 , apj 467 , 435 wang l. , & wheeler j. c. 1996 , apj , 462 , l27 wang , l. , hflich , p. , & wheeler , j. c. 1997 , apj in press | physical models of type ia supernovae ( sn ia ) that do not depend on secondary calibrators have indicated that the hubble constant must be @xmath0 65 km s@xmath1 mpc@xmath1 for well over a decade with the range of uncertainty shrinking with the sophistication of the models .
this estimate is in good agreement with those based on hst observations of cepheid variables and with purely empirical methods based on sn ia supernovae calibrated with cepheids .
the prospects of progress in understanding the physics of the explosion of sn ia and of their application to measure other cosmological parameters is reviewed . despite the rather complex nature of their atmospheres
, sn ia may give a more reliable estimate of distances than type ii .
the latter depend on remaining uncertainties in the scattering atmospheres , the helium abundance and possible systematic effects due to distortions of the envelope that will not average out even in a large sample .
psfig.tex 1.1 cm physical models of type ia supernovae ( sn ia ) that do not depend on secondary calibrators have indicated that the hubble constant must be @xmath0 65 km s@xmath1 mpc@xmath1 for well over a decade with the range of uncertainty shrinking with the sophistication of the models .
this estimate is in good agreement with those based on hst observations of cepheid variables and with purely empirical methods based on sn ia supernovae calibrated with cepheids .
the prospects of progress in understanding the physics of the explosion of sn ia and of their application to measure other cosmological parameters is reviewed . despite the rather complex nature of their atmospheres ,
sn ia may give a more reliable estimate of distances than type ii .
the latter depend on remaining uncertainties in the scattering atmospheres , the helium abundance and possible systematic effects due to distortions of the envelope that will not average out even in a large sample . |
the process of tidal capture , whereby a body in initially unbound orbit has a close approach to another gravitating mass , and through the excitation of internal modes of oscillation , becomes subsequently bound is of general importance in astrophysics .
it is believed to play a role in binary formation in globular clusters ( eg . press & teukolsky 1977 , hereafter pt , and references therein ) and to play a role in the tidal interactions of stars in galactic centres .
a related problem is the circularisation of an orbit from large eccentricity which may occur after a tidal capture . in this situation ,
each periastron passage is very similar to a close encounter in a weakly unbound orbit that causes a change in the energy contained within internal modes of oscillation at the expense of that in the orbit . in order to estimate capture probabilities and circularisation time
scales the excitation of normal modes of oscillation by a perturbing potential must be considered . in general modes associated with
spherical stars , or small perturbations of them , have been considered ( eg .
pt , ip and references therein ) . however , if the tidally perturbed body has multiple encounters , angular momentum transfer should lead it to rotate at an angular velocity related to that at periastron . in this case modes of oscillation with periods comparable to the rotation period
need to be considered ( eg .
papaloizou & pringle 1978 ) .
an approach for dealing with the problem of calculating tidal phenomena while taking into account the part of the oscillation spectrum with eigen frequencies comparable to the rotation frequency was recently proposed by pi .
they adopted a self - adjoint formalism for calculating the tidal response that used the density perturbation to describe the motions within the star and they made a low frequency approximation that the rotation and mode frequencies were very much less than the inverse sound crossing time .
they also evaluated the inertial mode spectrum for a polytrope with index @xmath14 and showed that inertial modes dominated the response at low frequencies or large periastron distances .
given the importance of rotation , the purpose of this paper is to develop further the work begun in pi and to use the results to obtain time scales for tidal phenomena such as circularisation of the orbit starting from high eccentricity .
we begin by noting that the self - adjoint formalism can be extended to general linear displacements without making a low frequency approximation . in this form , which uses the lagrangian displacement to describe the perturbations it has wider applicability than the formalism presented in pi which made a low frequency approximation .
we also wish to evaluate the expressions obtained for the energy and angular momentum exchanged during close periastron passages for realistic planet models .
accordingly we calculate sequences of models of one jupiter mass @xmath15 and @xmath16 each sequence has a range of radii in order to take account of the fact that the planet radius varies with age . for simplicity
we do not include a solid core noting that according to saumon & guillot ( 2004 ) observations of jupiter are not inconsistent with this .
we calculate the oscillation spectra together with required overlap integrals for the models using a set of basis or trial functions .
we then use the determined eigenmode properties for these models to estimate tidal phenomena such as the attainment of pseudo synchronism for which the angular momentum transfer at periastron passage is zero . such a state is expected to be attained , on account of the small inertia of the planet compared to that of the orbit , with subsequent energy exchanges occurring with the planet rotating at an equilibrium angular velocity .
the inclusion of inertial modes is important for determining this angular velocity which turns out to be close to that of the orbit at periastron .
calculations in ip incorporating only the @xmath3 modes gave a much higher value .
having calculated the equilibrium angular velocity we are able to consider multiple encounters and determine conditions for the onset of stochastic instability , that results in the stochastic gain of inertial mode energy over many periastron passages when there is no dissipation ( kochanek 1992 , mardling 1994a , b ) .
then we make estimates of circularisation time scales for the two different planet masses taking into account the evolutionary variation of the planet radius with age .
the plan of the paper is as follows . in section 2
we formulate the new self - adjoint formalism for the problem of linear perturbations proposed in our recent paper(pi ) using the lagrangian displacement to describe the motions and then use it to derive general expressions for the energy and angular momentum transfered when a rotating planet or a star passes through periastron in a parabolic or highly eccentric orbit around a central mass without making a low frequency approximation as in pi .
it can thus be applied to general situations in which @xmath17 @xmath18 @xmath19 and inertial modes are significant . in section 3
we show that the general formalism can be reduced to that of pi in an appropriate low frequency limit . in this case
only the inertial modes contribute to the tidal response for the baratropic stellar models we consider . in section 4
we discuss the tidal energy and angular momentum transfered through inertial modes giving very simple expressions that apply when the planet or star rotates in a state of pseudo synchronisation for which the angular momentum transfered in an encounter is essentially zero .
we apply these to the multi - passage problem when there is no dissipation finding conditions for the occurrence of the stochastic instability .
we find that this requires the circularisation process to start with a rather large semi - major axis @xmath4 for final periods of @xmath5 days .
however , this could be reduced to @xmath6 for the shortest observed final periods @xmath7 days . in section 5
we calculate the eigenspectrum of the inertial modes for the planet models which have been obtained using a realistic equation of state .
we find that the counterparts of the two global modes found in a polytrope with @xmath20 exist in these cases but that there are additional global modes associated with the transition from molecular to metallic hydrogen .
we go on to apply the results to the problem of tidal circularisation of the orbits of the extrasolar planets in sections 6 and 7 finding that inertial modes dominate the tidal interaction for @xmath15 planets ending up on circular orbits with final period in the @xmath10 day range and are potentially able to account for the initial circularisation at a @xmath11 day period within a few @xmath12 in that case
. however , in the case of @xmath13 oscillation modes excited in the star become more important . finally in section 8
we discuss our results .
in this section we further develop the self adjoint formalism describing the pulsations of a uniformly rotating fully convective object ( a planet or a star ) , referred to hereafter as a planet , introduced in our previous paper ( papaloizou @xmath21 ivanov 2005 , pi ) .
we give general expressions for the energy and angular momentum transfer from the orbit to the planet as a result of a close encounter or fly by around a gravitating centre . here , contrary to pi , we do not make the assumption that only low frequency modes respond effectively to the tidal forcing and the results obtained are valid for the excitation of any pulsational mode of a rotating planet .
the fundamental quantity used in our analysis is the lagrangian displacement vector @xmath22 . by definition ,
this is real .
we adopt a cylindrical coordinate system @xmath23 with origin at the centre of mass of the planet .
we make use of the fourier transforms of the displacement vector and any other perturbation quantity , say q , over the azimuthal angle @xmath24 and time @xmath25 in the form @xmath26 where the sum is over @xmath27 and @xmath28 and @xmath29 denotes the complex conjugate .
the reality of @xmath30 implies that the fourier transform , indicated by tilde satisfies @xmath31 the inner products of two complex scalars @xmath32 , @xmath33 , and vectors @xmath34 and @xmath35 are defined as @xmath36 where @xmath37 is the scalar product , @xmath38 is the density , and @xmath39 stands for the complex conjugate .
we assume that , viewed from an inertial frame , the planet is rotating with uniform angular velocity @xmath40 directed perpendicular to the orbital plane .
in this situation the hydrodynamic equations for the perturbed quantities take the simplest form in the rotating frame which we accordingly adopt .
since the planet is fully convective , the entropy per unit of mass of the planetary gas remains approximately the same over the volume of the planet , and the pressure @xmath41 can be considered as a function of density @xmath38 only , @xmath42 . a perturbation of the planet may be considered as adiabatic , and therefore the same functional dependence @xmath42 holds during perturbation as well .
this condition is often referred to as the baratropic condition . in the baratropic approximation
the linearised navier - stokes equations take the form ( see pi ) @xmath43 where @xmath44 @xmath45 is the density perturbation , @xmath46 is the adiabatic sound speed , @xmath47 is the stellar gravitational potential arising from the perturbations and @xmath48 is the external forcing tidal potential .
the viscous force per unit of mass is @xmath49 in cartesian coordinates it can be written in component form as @xmath50 where the comma denotes differentiation and @xmath51 is the kinematic viscosity .
the greek indices enumerate spatial coordinate directions and the usual summation convention over these is adopted .
note that the centrifugal term is absent in equation @xmath52 being formally incorporated into the potential governing the static equilibrium of the unperturbed star .
the convective derivative @xmath53 as there is no unperturbed motion in the rotating frame .
the linearised continuity equation is @xmath54 and the linearised poisson equation is @xmath55 provided that expressions for the density , sound speed and kinematic viscosity are specified in some unperturbed model of the planet , the set of equations @xmath56 is complete . without the viscous term on the right hand side of equation @xmath52
these equations coincide with the analogous equations used by pi .
the perturbed potential @xmath47 can be formally expressed in terms of the density perturbation with help of the inverse of the laplace operator , @xmath57 , and the density perturbation is expressed in terms of the displacement vector with help of the continuity equation .
we have @xmath58 now we express @xmath59 in terms of @xmath60 with help of equations @xmath61 , @xmath62 and @xmath63 .
we substitute the result into equation @xmath52 thus obtaining an equation containing only the displacement vector @xmath60 as a dynamical variable . after making fourier transform of this equation
we get @xmath64 where the operators @xmath65 and @xmath66 are defined through their action on the vector @xmath60 as @xmath67 and @xmath68 it can be easily checked that the so defined operators @xmath65 and @xmath66 are self adjoint in the sense of the normal product @xmath69 .
also for the models we consider with weak self - gravity , the operator @xmath66 may be taken to be positive definite for displacements with a non zero density perturbation somewhere and non negative for arbitrary displacements including those with zero density perturbation everywhere .
the source term @xmath70 is determined by the external forcing potential @xmath71 the presence of the viscous force @xmath72 leads to dissipation of energy of the perturbed motion on a dissipation time scale . for global disturbances in astrophysical systems
this time scale is typically very large compared to a time scale characterising the tidal forcing . in this situation
we can treat the action of the viscous force as a perturbation . neglecting the contribution of the viscous term
, we see equation @xmath73 has an analogous form to one discussed by pi for low frequency perturbations .
there it was shown that this form led to a standard type of eigen value problem associated with a new self adjoint operator @xmath74 . in order to bring equation @xmath73 to this standard type ,
let us formally consider the square root of the operator @xmath66 , @xmath75 , defined by the condition : @xmath76 this can be done on account of the non negativity of @xmath77 .
further , let us introduce a six dimensional vector @xmath78 with components @xmath79 it follows from equation @xmath73 that the eigenvalue problem can be written in the standard form @xmath80 where the operator @xmath74 has a matrix structure @xmath81 and the source six dimensional vector @xmath82 has the components @xmath83 it follows from equation @xmath84 that the operator @xmath74 has the required self adjoint form .
now we can look for solution to equation @xmath85 in the form : @xmath86 where @xmath87 are the eigenfunctions of @xmath74 which satisfy @xmath88 the eigen frequencies @xmath89 are necessarily real .
equation @xmath90 is equivalent to equation @xmath73 with the source term @xmath91 .
note that the eigen spectrum of @xmath74 or a part of it can be continuous .
it this case the summation in @xmath90 must be replaced by integration with an appropriate measure .
the eigen functions @xmath92 are orthogonal in the sense of the inner product @xmath93 where @xmath94 substituting @xmath95 into equation @xmath85 and using @xmath96 we obtain @xmath97 where we add a small imaginary part to the frequency @xmath98 with @xmath99 in potentially resonant denominators to account for dissipation . as we show below , for modes for which the effect of viscosity can be treated as a small perturbation , an explicit expression for @xmath100 can be obtained by taking into account the dissipation determined by the viscous force @xmath72 .
@xmath101 is the norm .
the coefficients @xmath102 determine decomposition of the source vector over the eigen functions , @xmath103 where summation is performed over all modes with a particular value of @xmath104 .
they can be written in the form @xmath105 where @xmath106 is the density perturbation associated with the eigenvector @xmath107 : @xmath108 , and to obtain the last equality , we integrate by parts .
it is important to note that the coefficients @xmath102 , the eigen frequencies @xmath109 and the associated displacement vector @xmath110 obey a certain identity relation @xmath111 which follows from equations @xmath112 , @xmath113 and the orthogonality relation @xmath96 for the eigen vectors .
this relation is also required to ensure that @xmath114 remains bounded as @xmath115 ( see below ) .
the solution for the displacement vector @xmath116 follows from equations @xmath112 , @xmath95 and @xmath117 @xmath118 where we use the relation @xmath119 to obtain the last equality .
the displacement vector @xmath60 is obtained from its fourier s transform @xmath116 by integration over @xmath120 and summation over @xmath104 @xmath121 this expression can be simplified in the limit @xmath122 at which only the poles in the expression in the braces contribute significantly to the integral over @xmath120 . in this limit
we may , therefore , write @xmath123 it can be easily shown from equation @xmath52 or equation @xmath73 that the expression for the canonical energy of the perturbations is given by @xmath124 where we integrate over the volume of the planet .
we evaluate the expression @xmath125 after a close encounter , @xmath126 .
assuming that the time after closest approach is much larger than the characteristic time scale of the encounter , but nevertheless much smaller than the characteristic viscous time scale @xmath127 , we can use equation @xmath128 setting @xmath129 substituting the expression for @xmath60 and its time derivative into @xmath125 and taking into account equation @xmath130 we get @xmath131 where we use the orthogonality relation @xmath96 and the expression for the norm @xmath96 .
the rate of viscous dissipation of energy is given by the standard expression @xmath132 an explicit expression for the amount of energy dissipated by viscosity after the fly by , @xmath133 , can be obtained from @xmath134 substituting @xmath135 into @xmath134 and integrating over time the resulting expression .
we have @xmath136 where we use the well known relation @xmath137 , and @xmath138 as we have mentioned above , we assume that viscosity is very small , and the imaginary part of the frequency is much smaller than the corresponding real part . in this case , one can see from equation @xmath139 that the leading contribution to @xmath140 is determined by diagonal terms in the summation series with @xmath141 . also , only a part of the integral over @xmath120 very close to the resonance , @xmath142 , gives a significant contribution .
we obtain @xmath143 the amount of energy transfered by tides into the modes must be equal to the amount of energy dissipated by viscosity : @xmath144 . from this condition and equation @xmath145 and @xmath146 we easily obtain an explicit expression for the inverse dissipation time @xmath147 similar arguments can be used for calculations of the amount of the angular momentum transferred to the modes after the fly by .
however , the expression for the angular momentum transfer directly follows from the general relation between canonical energy and angular momentum corresponding to a particular mode with given values of @xmath148 and @xmath104 ( eg .
friedman and schutz 1977 ) @xmath149 accordingly , we have from equation @xmath145 @xmath150 the expression for the energy of the modes in the inertial frame , @xmath151 , follows directly from equations @xmath145 and @xmath152 @xmath153 note that the modal energy in the inertial frame can be negative .
the formalism developed in the previous section can be applied for any pulsational mode of a rotating planet .
it follows from equations @xmath154 , @xmath145 and @xmath152 that in order to find the transfer of energy and angular momentum from the orbit to the modes we should be able to find solutions ( eigen functions and eigen frequencies ) of equation @xmath84 describing free oscillations of the planet .
this can be done numerically . however , in general , it is rather difficult to use numerical methods directly in solving equation @xmath84 without any further simplifications .
it turns out that these simplifications can be naturally introduced into the problem on hand . assuming that the periastron distance @xmath155 is sufficiently large
, there is a small parameter in the problem , namely the ratio of the typical frequency associated with the periastron passage , @xmath156 to the typical frequency associated with the planet , @xmath157 : @xmath158 . and @xmath159 are the radius and the mass of the planet , respectively , and @xmath160 is the mass of the object exerting tides on the planet . ] we would expect that the angular momentum transfer between the orbit and the modes leads to a so called state of pseudo synchronisation where @xmath161 , and therefore the angular velocity of the planet can be , typically , much smaller than the characteristic frequency @xmath162 . in such a situation
all modes of pulsation present in a convective planet can be divided in two distinct groups : high frequency modes with eigen frequencies of the order of or larger than @xmath162 , and low frequency modes with eigen frequencies of the order of @xmath163 . for the high frequency modes
, the effects of rotation on the structure of eigen functions and eigen frequencies can be taken into account as a perturbation ( see eg .
ip ) and references therein ) .
calculation of the eigen spectrum of the low frequency modes is much more difficult ( papaloizou @xmath21 pringle 1981 , hereafter pp ) .
however , a convenient approximation can also be made in this case which allows one to reduce the general equation @xmath84 to a much simpler equation for the low frequency modes only ( pp , pi ) . in order to reduce equation @xmath84 to an expression appropriate for low frequency modes it is convenient to use the potential @xmath164 ( see equation @xmath61 for definition ) as a dynamical variable instead of the displacement vector @xmath116 .
we express the components of @xmath116 in terms of @xmath164 using the fourier transformed variant of equation @xmath52 having set @xmath165 there . for a normal mode labelled by @xmath166 the components of the displacement vector
are related to the potential through : @xmath167 @xmath168 @xmath169 where @xmath170 . to simplify notation
we shall not use tildes and subscripts denoting @xmath104 for quantities associated with eigen modes .
now we substitute equations @xmath171 , @xmath172 and @xmath173 into right hand side of the continuity equation @xmath62 .
the density perturbation on the right hand side is expressed in terms of @xmath174 and perturbed gravitational potential @xmath175 with help of equation @xmath61 where we set @xmath176 .
we then obtain @xmath177 and @xmath178 @xmath179 note that the operator @xmath180 can be represented in an invariant form @xmath181 as follows from equation @xmath182 , @xmath180 is just a modified laplace operator .
the operators @xmath183 , @xmath184 and @xmath185 are analogous to the operators @xmath65 and @xmath66 introduced above .
they are self - adjoint with respect to the inner product @xmath69 .
also when @xmath59 is not a constant @xmath186 and @xmath184 are positive definite and @xmath185 is non negative definite .
note that an analogous equation used by pi has a factor @xmath187 instead of @xmath188 on the right hand side .
accordingly , the potential @xmath189 used by pi is different from the potential @xmath174 used in this paper : @xmath190 .
equation for the perturbed gravitational potential @xmath175 is obtained from equations @xmath61 and @xmath191 @xmath192 equations @xmath193 and @xmath194 form a complete set .
they are equivalent to equation @xmath90 and can be used instead of it to find the solutions for the eigen modes . assuming that @xmath195 it is easy to see that the terms on the right hand side of equation @xmath193 are much smaller than the terms on the left hand side . in the leading approximation in the small parameter @xmath196
they can be neglected and we obtain @xmath197 equation @xmath198 can be used for calculation of the low frequency eigen functions and eigen frequencies . then a formalism exactly equivalent to that given by pi is obtained .
equation @xmath198 can be brought into the standard self adjoint form analogous to equation @xmath90 ( pi ) .
for that let us consider two dimensional vector @xmath199 with components @xmath200 now it follows from @xmath198 that it can be written in the form @xmath201 where @xmath202 since the off diagonal terms in the matrix @xmath203 are adjoint of each other and the diagonal terms are self adjoint it is clear that the operator @xmath204 has the required self adjoint form .
the inner product induced by the operator @xmath204 is @xmath205 it is obvious that different solutions of equation @xmath206 are orthogonal with respect to this product . from equations
@xmath130 , @xmath207 and @xmath208 it follows that the norm @xmath209 corresponding to the displacement vector associated with @xmath174 is related to the norm @xmath210 induced by the inner product @xmath208 as @xmath211 the density perturbation determined by a solution of equation @xmath206 follows from equation @xmath61 @xmath212 where the gravitational potential @xmath213 is found from equation @xmath194 with @xmath174 as a source . the coefficients @xmath102 determining the energy and angular momentum transfer ( see equations @xmath145 and @xmath152 ) follow from equations @xmath154 and @xmath214 .
we have @xmath215 where we take into account that we can use real expressions for the potentials @xmath174 and @xmath48 .
the expression @xmath216 can be brought in a simpler form .
for that , let us formally solve equation @xmath194 @xmath217 where @xmath218 is the inverse of the operator @xmath219 .
it is easy to see that @xmath218 is self adjoint .
we may , therefore , write the last term on the right hand side of @xmath216 as @xmath220 where the potential @xmath221 is determined from equation analogous to equation @xmath194 but with negative of the potential @xmath48 as a source @xmath222 the polarisation potential @xmath221 has a simple physical meaning . if the external potential @xmath48 is assumed to be static , the potential @xmath221 is induced in the star by the corresponding perturbation of density
. finally we obtain @xmath223
in this section we obtain explicit expressions for the transfer of energy and angular momentum in terms of quantities characterising the perturbing tidal field and quantities characterising the planet .
these will be applied directly to the problem of evaluating the tidal response during a fly by or as a consequence of multiple encounters during circularisation from a highly eccentric orbit .
since the expressions for the energy transfer for the fundamental and @xmath224 modes are well known ( see pt , pi , ip and references therein ) we consider only the low frequency modes in this section . as we have mentioned before we consider highly eccentric orbits in this paper , with eccentricity @xmath225 . in this case
the energy and angular momentum transfer takes place mainly near the periastron @xmath155 and we make the usual approximation of the orbit as a parabolic one with the same value of @xmath155 for calculation of the perturbing potential @xmath226 .
the fourier s transform of the perturbing potential @xmath227 for the parabolic orbit in the quadrupole approximation has been calculated by pt .
we have @xmath228 where @xmath229 , @xmath230 .
@xmath231 is the ratio of @xmath159 to the mass of the star exerting the tides , @xmath160 : @xmath232 . in the application of our formalism to the problem of circularisation of the extra solar planets
@xmath233 is very small and will be neglected later .
@xmath234 is a typical frequency of periastron passage . the functions
@xmath235 \label{eq p58}\ ] ] specify the dependence of @xmath226 on @xmath98 , they are described by pt .
y=@xmath236 , where we use dimensionless frequencies @xmath237 @xmath238 is the associated legendre function , and @xmath239 are the spherical coordinates . when @xmath240 the quantities @xmath241 decay exponentially .
this is valid for the case of the high frequency fundamental and @xmath224 modes where @xmath242 for the eigen frequencies . on the other hand ,
the low frequency modes considered above have @xmath243 and do not experience this suppression .
the expressions for the energy and angular momentum transfer @xmath244 and @xmath152 are strictly valid only under assumption that the planet is unperturbed before the fly by , and we call the problem formulated in such a way as a first passage problem
. this assumption may not be correct when the orbit of the planet is elliptical and the viscous time @xmath245 is larger than the orbital period @xmath246 . in latter case
the energy and angular momentum exchange between the orbit and the modes depends on the state of the pulsational modes before the periastron passage .
however , the expressions for the first passage problem provides a basis for a treatment of that more complicated case , and firstly we would like to discuss this problem .
the generalisation to the case of a multi - passage problem is discussed in the next section . in our discussion
below we assume that effects determined by perturbations of the structure of the planet due to rotation are not significant for our purposes .
we use , accordingly , spherical models of the planet in order to find the form of the density and the sound speed entering in equations @xmath134 , @xmath152 . in the same approximation , taking into account that the perturbing potential is proportional to the product of a function of the radius @xmath247 and a spherical harmonic , we see that the variables in equation @xmath248 can be separated in spherical polar coordinates
. therefore we can express the sum @xmath249 entering in equation @xmath250 as @xmath251 where the function @xmath252 can be obtained from equation @xmath248 . in order to present the expressions @xmath134 , @xmath152 and @xmath253 in a simpler form
we express the quantities @xmath254 , @xmath151 and @xmath255 in terms of natural units of energy @xmath256 and angular momentum @xmath257 . as above
we also adopt dimensionless units for the spatial coordinates , density and the sound speed expressing them in units of @xmath258 , the averaged density @xmath259 , and @xmath260 , respectively . instead of the separation distance @xmath155
it is convenient to introduce the parameter ( pt ) @xmath261 we then substitute equations @xmath250 and @xmath262 into @xmath145 and @xmath152 .
we obtain @xmath263 @xmath264 where we separate contributions to @xmath254 and @xmath255 corresponding to different values of @xmath104 such as @xmath265 and take into account that @xmath266 and therefore @xmath267 .
all quantities in equations @xmath268 , @xmath269 are assumed to be in the dimensionless form , @xmath270 @xmath271 and @xmath272 the overlap integrals @xmath273 are analogous to the same quantities introduced by pt in their analysis of dynamic tides in a non rotating star .
they describe spatial coupling of the perturbing potential and of the eigen modes , and have the form : @xmath274 where @xmath275 .
the expression @xmath268 and @xmath269 have already been obtained by pi using a different method .
they can also be presented in another equivalent form as @xmath276 where the dependence of the functions @xmath277 and @xmath278 on @xmath279 follows from @xmath268 and @xmath280 accordingly , the energy transfer in the inertial frame follows from @xmath281 @xmath282 and the use of @xmath283 where all quantities in the brackets belong to a particular value of @xmath284 @xmath285 where all quantities in the brackets are evaluated for @xmath286 and @xmath287 as @xmath288 usually , the moment of inertia of the planet is much smaller than the moment of inertia of the orbit . in this case
the planet tends to rotate with equilibrium value of the angular velocity , @xmath289 determined by condition that there is no exchange of angular momentum between the planet and the orbit at this rotation rate , @xmath290 this state of rotation is often referred to as a state of pseudo - synchronisation. for a given model of planet , equation @xmath291 can be solved numerically to find @xmath289 .
the expression for the energy transfer at the state of pseudo - synchronisation , @xmath292 , follows from equation @xmath293 @xmath294 where @xmath295 depends only on the structure of a planet . for simplicity we set @xmath296 hereafter .
the expressions for a comparable mass ratio @xmath231 can be easily obtained from our equations by an appropriate re - scaling of variables we use .
an estimate of @xmath297 can be obtained from the following semi - analytical arguments . as we will see below only two global modes with @xmath298
mainly determine the energy transfer when @xmath299 and we can take into account only these two modes in the summation @xmath300 and @xmath301 .
one of these modes has a positive eigen frequency with @xmath302 and the overlap integral @xmath303 , and another mode has a negative eigen frequency with @xmath304 .
the equilibrium value of the angular velocity is always close to the frequency of periastron passage , with @xmath305 .
this information is sufficient to estimate @xmath297 .
indeed , we can express the second overlap integral from equation @xmath300 in terms of @xmath306 , the eigen frequencies and the quantities @xmath307 , where @xmath308 .
we get @xmath309 . taking into account that @xmath310 and @xmath311
, we get @xmath312 .
this value is in a good agreement with numerical results discussed below , see discussion and equation @xmath313 in section 6 .
note that for @xmath20 polytrope @xmath314 ( pi ) . in the circularisation problem expressions
@xmath315 , @xmath316 , @xmath317 , @xmath318 , @xmath319 , @xmath320 , @xmath321 and @xmath301 are strictly valid only when the energy stored in the modes can be dissipated between two successive periastron passages . in the opposite limit , when the dissipation time @xmath322 , where @xmath246 is the orbital period , the planet approaches periastron in a perturbed state .
the additional linear perturbation induced during a particular periastron passage must be added to the perturbation already present in the star .
since the mode energy is quadratic in the perturbed quantities , it can either decrease or increase after the periastron passage , depending on phase of the perturbations ( kochanek , 1992 , mardling 1995a , 1995b ) .
this issue has been studied analytically by ip .
it was found that when a change of the orbital period between two successive periastron passages is sufficiently large , the energy transfer proceeds in a stochastic way . in this case
the phase of perturbation in the planet approaching periastron is not correlated with the phase of the perturbation induced by tides near periastron , and the energy transfer during the periastron passage is essentially a random quantity , with dispersion of the order of the expressions for the energy transfer calculated above . in this situation , a build - up of mode energy due to
stochastic instability is possible .
a semi - analytical condition for this instability to happen has been obtained in ip .
here we generalise their formalism to take account of the specific features associated with inertial waves .
let us assume that dissipation is negligible and treat the tidal influence as a sequence of impulses separated by periods of time equal to the orbital period . in this case , any time between two impulses the planet undergoes free oscillation , and the lagrangian displacement vector can be represented by an expression analogous to @xmath323 @xmath324 where the dimensionless real coefficients @xmath325 characterise the amplitudes of perturbations corresponding to different modes and @xmath326 are the corresponding phases .
note that the origin of time is chosen to coincide with the moment of periastron passage .
the expressions for the canonical energy and angular momentum follow from @xmath327 , @xmath328 and @xmath329 @xmath330 when @xmath331 these expressions obviously coincide with the expressions @xmath315 and @xmath316 for the single passage problem . for inertial waves equations
@xmath332 can be brought to the form analogous to equations @xmath333 @xmath334 where we now show separately the contributions of terms with different @xmath104 to the canonical energy . @xmath335 and @xmath336 are the terms in the summation series for @xmath277 and @xmath278 , corresponding to a particular value of @xmath148 , see equations @xmath282 , @xmath337 and @xmath338 .
as we discus below in the tidal problem we can neglect the contribution of @xmath27 terms . in this case , the mode energy in the inertial frame can be represented as @xmath339 where @xmath292 is given by @xmath301 after the periastron passage , the expression @xmath323 must be added to @xmath329 , and the resulting expression can brought again to the form @xmath329 but with new values of the amplitudes and phases , @xmath340 and @xmath341 .
thus , the action of tidal forcing defines an iterative map @xmath342 .
it is convenient to introduce new variables @xmath343 and @xmath344 . in terms of these variables
this map has especially simple form ( ip ) @xmath345 where @xmath346 , and @xmath347 is a two dimensional rotational matrix .
note that in equation @xmath348 we take into account the change of origin of time after the periastron passage : @xmath349 .
when @xmath246 is a constant , it is easy to see from equation @xmath348 that the amplitudes @xmath340 do not grow with the number of iterations .
however , the orbital period depends on the orbital energy , @xmath350 , as @xmath351 and the orbital energy is proportional to the mode energy @xmath151 due to the law of energy conservation .
since @xmath352 is proportional to the squares of the modes amplitudes , @xmath353 , the map is non - linear . in order to simplify the problem ,
let us consider a situation when the mode energy is much smaller than the orbital energy such that the orbital energy and angular momentum remain in the neighbourhood of some particular fixed values .
in this case we can set @xmath354 , where @xmath355 is a constant , and @xmath356 where @xmath350 and @xmath357 are the orbital energy and semi - major axis corresponding to the period @xmath355 .
now the angles @xmath358 can be expressed in terms of @xmath151 as @xmath359 where @xmath360 , @xmath361 and @xmath362 where we use equations @xmath363 , @xmath364 , @xmath301 and @xmath365 .
equation @xmath366 is analogous to equation @xmath367 of ip .
equations @xmath368 , @xmath348 and @xmath366 fully determine the map for given values of @xmath369 @xmath370 @xmath371 and @xmath372 . when @xmath372 are sufficiently large , stochastic instability sets in .
since the angles @xmath369 are determined by initial conditions , the results should be averaged over these angles . this map has been investigated numerically by ip for the case of only one high frequency mode .
it has been found that the instability sets in when the average value of @xmath372 is larger than @xmath373 . in our case
the situation is slightly more complicated since there are at least two modes which play a role in the energy transfer .
also , we should specify the evolution of the angular velocity .
in order to investigate this question we consider numerically the simplest possible case of constant equilibrium rotation of a planet , when @xmath374 .
for definiteness , we use parameters described in the end of the previous section , namely @xmath375 and consider only @xmath298 modes with @xmath376 and @xmath377 . in this case
we have @xmath378 in the first case and @xmath379 in the second .
we iterate numerically the map for a large number @xmath380 of iterations for a given value of @xmath381 with @xmath382 , and for a given value of @xmath372 .
the stochastic instability is assumed to set in when the energy @xmath383 exceeds the maximal number of iterations , @xmath384 during the iterations .
this happens when @xmath372 exceeds some critical value @xmath385 .
the results of our computations are shown in figure 1 .
one can see from this figure that there is a strong dependence of @xmath386 on @xmath381 .
nevertheless we use the averaged value of @xmath386 , @xmath387 since our results are rather insensitive to the exact values of @xmath386 .
it is interesting to note that the transition from a quasi - periodic behaviour of @xmath383 to stochastic behaviour is very sharp . in figure 2
we show the evolution of @xmath383 for @xmath388 and two values of @xmath372 : @xmath389 and @xmath390 .
one can see from this figure that a small change of @xmath372 is sufficient for stochastic instability to set in . from the condition
@xmath391 we find that the stochastic instability sets in only when the semi - major axis is sufficiently large , @xmath392 using the values of all quantities in @xmath393 typical for the problem of circularisation of extra solar planets , we have @xmath394 where @xmath395 , @xmath396 and @xmath397 are the masses of jupiter and of the sun , respectively , and @xmath398 is the radius of jupiter .
remarkably , the typical value of @xmath399 is close to what has been obtained in the analysis of the stochastic instability associated with the fundamental modes ( ip ) .
this value is of the order of a typical initial semi - major axis from which we would expect the process of circularisation to begin for @xmath400 this corresponds to a final circularised period of @xmath401 d for a jupiter mass planet around a solar type star .
however if the final period is reduced to @xmath402 d , corresponding to the shortest period exoplanets , the initial semi - major axis is reduced to between @xmath403 and @xmath404 if the initial semi - major axis is smaller , the stochastic growth of the mode energy is not possible . however , in this case , we would expect a stronger dissipation of the mode energy due non - linear mode - mode interactions ( e.g. kumar @xmath21 goodman 1996 ) .
this possibility needs a further investigation .
in this section we construct realistic models of planets at different evolutionary stages for which we obtain the oscillation spectra numerically .
we discuss our numerical method , use it to solve equation @xmath405 , and discuss the results which are used in subsequent sections to evaluate the tidal responses of different planet models undergoing a fly by or circularisation from a highly eccentric orbit .
we find it convenient to adopt a system of dimensionless units for the spatial coordinates , density and the sound speed expressing them in units of @xmath258 , the averaged density @xmath259 , and @xmath260 , respectively .
we also use dimensionless frequencies @xmath406 we assume that the planets are spherically symmetric and adopt the equation of state of saumon , chabrier @xmath21 van horn ( 1995 ) . although it is likely that giant planets form through a nucleated core instability ( eg .
bodenheimer @xmath21 pollack 1986 ) , for this first set of calculations , for reasons of simplicity we shall assume that solid cores are absent .
in this context we note that the mass of any solid core in jupiter is highly uncertain , being sensitive among other things to the equation of state adopted , with recent estimates showing it to be consistent with zero ( see saumon & guillot 2004 ) . to make this consistent with a nucleated core instability scenario for formation
, these authors suggest that a core could be eroded and remixed during subsequent gas accretion of light elements .
but it should be noted that the presence of a large solid core could significantly affect the mode spectrum of the planet through the existence of wave attractors ( eg .
ogilvie & lin 2004 ) that could produce effects on a long time scale .
however , the presence of a small core is unlikely to affect phenomena of short duration such as the energy transfer that occurs during a close passage or fly by .
we calculate numerically the form of @xmath38 and @xmath46 as functions of the radius @xmath407 with these functions known , equation @xmath206 can be solved and the overlap integrals @xmath408 calculated . as in ip
we consider only two planet masses @xmath409 and @xmath410 . for a given mass we calculate planet models for six values of the radius .
these are @xmath411 ( @xmath412 ) when @xmath413 ( @xmath414 ) , and @xmath415 , @xmath416 , @xmath417 , @xmath418 , @xmath419 for both masses .
the different values of radius for a given mass correspond to different points on an evolutionary track .
on such a track , the radius decreases with time due to gravitational contraction . to determine the age corresponding to a given radius
, we use the relationship between the radius and luminosity , and hence the age of the planet calculated by burrows et al ( 1997 ) ( see fig . 8 of ip ) .
since a young hot planet of age @xmath420 has a radius of around @xmath421 the range of radii we consider corresponds to planet ages between @xmath420 and @xmath422 .
we illustrate the dependences of the density , the magnitude of the density gradient @xmath423 and square of the sound speed for several planet models in figures [ fig.3]-[fig.5 ] . as one can see from figure [ fig.3 ] for a particular planet , the ratio of central density to mean density increases with radius .
this effect is less prominent for the models with @xmath424 note a sharp drop of density at @xmath425 when @xmath426 and @xmath427 this is due to the phase transition of hydrogen from the metallic to molecular phase which is assumed to be of the first kind .
this phase transition also takes place when @xmath414 and @xmath428 but it is shifted to a larger @xmath429 and is not well resolved in this figure . in general , according to saumon , chabrier @xmath21 van horn ( 1995 ) , the phase transition is of the first kind when @xmath430 ( @xmath431 ) for @xmath409 ( @xmath414 ) . for larger radii the adiabats describing the interiors of the planets go above the critical point of saumon , chabrier @xmath21 van horn ( 1995 ) equation of state on @xmath432 phase plane . as we see below the effect of the phase transition on the structure of the eigen spectrum
is more prominent for @xmath409 . to illustrate this
it is of interest to compare this model with a similar model without the phase transition and we make such a model with smoothed out transition between the metallic and molecular phases ( dotted curves in figures 3 - 5 ) , for @xmath409 and @xmath433 this model has been calculated using a prescription provided by saumon , chabrier @xmath21 van horn 1995 .
the structure of the density profiles is better seen in figure [ fig.4 ] where the magnitudes of the density gradients are plotted .
for the models with the phase transition there is a discontinuity at the radius of the phase transition . for the models without the phase transition the modulus of the gradients increase with radius at small @xmath434 , and they sharply decrease at large @xmath434 .
the transition point between these two limiting cases corresponds to a melting point where the number density of the hydrogen molecules sharply decreases toward smaller values of @xmath434 .
this transition point shifts toward smaller @xmath434 with increase of the planet radius .
as seen from figure [ fig.5 ] the sound speed is changes non monotonically near the transition point . in general , the models corresponding to the larger mass are more similar to each other . in this case the radius of transition is shifted toward larger @xmath434 where the planets have rather low density .
the low density outer regions of the planet should not influence eigenfrequencies and eigenfunctions corresponding to global modes should they exist .
therefore , it is reasonable to suppose the presence of the phase transitions and the melting zones in the planet models with @xmath414 is less significant for our problem than for the case of @xmath409 . as we see below this assumption
is supported by our numerical results . for our numerical work
we use a method similar to that proposed by pp .
this involves reducing the operator equations to discrete form by use of a set of basis functions . for our numerical work
it is convenient to work with a matrix representation of equation @xmath436 this is accomplished by use of a basis of trial functions @xmath437 , @xmath438 which is assumed to be complete .
we introduce a fourier decomposition of an eigenfunction @xmath439 in terms of the trial functions as @xmath440 where the indices enumerating the eigen functions are enclosed in brackets from now on .
then we substitute equation @xmath441 into equation @xmath435 and taking the scalar product @xmath442 of the resulting expression with @xmath443 we get @xmath444 where @xmath445 and the standard summation rule over the repeating indices is implied
. we can introduce square roots and inverses of the above matrices in a similar manner to that adopted for the corresponding operators .
let us consider infinite dimensional vectors analogous to the vectors @xmath446 ( see @xmath447 ) , with components @xmath448 where the index @xmath160 stands for the pair ( @xmath449 and @xmath450 ) .
we have @xmath451 the eigen value problem @xmath452 can be formulated for the vectors @xmath453 in a way equivalent to @xmath405 , @xmath454 where the matrix @xmath455 has a structure equivalent to the operator @xmath204 ( equation @xmath456 ) with the operators being substituted by the corresponding matrices .
it is natural to use the solutions of the eigenvalue problem @xmath457 as the functions @xmath437 .
these functions are orthogonal with respect to the inner product @xmath458 . as follows from equation @xmath182 , for the spherical planet models the variables in equation @xmath459 are separable in spherical polar coordinates @xmath460 , @xmath461 , where @xmath462 .
@xmath463 is the order of the radial eigenfunction and the index @xmath464 stands for the pair of positive integers @xmath465 .
for the radial part we have @xmath466 where @xmath467 .
it follows from @xmath468 that the eigen values of @xmath186 are bound from below : @xmath469 when @xmath470 .
in our numerical work we truncate the series @xmath441 at some sufficiently large value @xmath471 .
the trial functions must be arranged in such a way that when @xmath464 is sufficiently large the corresponding trial functions oscillate with periods of spatial oscillation that decrease with increasing @xmath464 . in this case
, one can argue that errors will be small because of a negligible contribution of the oscillatory terms with high @xmath464 to the various integrals occurring in the formulation of our problem as discussed above .
of course , this idea must be tested , and we discuss such tests later . since for our problem @xmath472 is even and
the angular parts of the trial functions are such that @xmath473 the modes odd with respect to reflection in @xmath474 do not couple with the tidal field and we can choose @xmath475 to be even . thus @xmath476 with @xmath477 being a positive integer such that @xmath478 it is convenient to consider an equal number of the radial functions @xmath479 and angular functions @xmath480 thus we write @xmath481 in our numerical work , we use different values of @xmath384 in order to study the dependence on the number of trial functions , with the largest @xmath482 .
accordingly , we have @xmath483 after truncation of the basis set used in @xmath441 we have ordinary finite dimensional matrices @xmath484 , @xmath485 and @xmath486 .
we calculate the eigenvalues eigenvectors of these matrices .
the eigenvalues are used to calculate the inverses and square roots of the matrices entering in the matrix @xmath455 .
we find the components of @xmath455 and solve the eigenvalue problem @xmath487 .
the eigenvectors @xmath488 are used to find the coefficients @xmath489 , and we use these coefficients to find the eigenfunctions @xmath439 with help of equation @xmath441 . one can view the eigenvalues and eigenfunctions as belonging to the oscillation spectrum of the planet . for the case of a model without a solid core
this is ultimately likely to be discrete but everywhere dense ( see eg .
one should also consider the possibility of the existence of a continuous spectrum in some cases .
then the numerically obtained eigenfunctions may not show convergence to regular functions .
but note that , independently of this issue , we expect our numerical method to be convergent when the eigenfunctions we obtain are used to provide a basis for representing smooth functions varying on a global scale .
the issue of the possibility of ultimately irregular eigenfunctions belonging to a continuous spectrum does not affect this or preclude the convergence of the process of calculating the angular momentum and energy exchanges associated with a fly by . on account of the forcing by a global potential and short time duration of a fly by
, a basis representation capable of representing global scales would be expected to be adequate on physical grounds .
indeed for our models it turns out that only a few global eigenfunctions are important for determining the energy exchanged and these can be represented with a basis set that needs only to be capable of representing global functions .
this makes the issue qualitatively different to the one that occurs in the context of tidal forcing occurring in a circular orbit with fixed frequency for infinite time periods ( eg.ogilvie & lin 2004 ) .
it is obvious that the dimension of the matrix @xmath455 is @xmath490 .
thus , we have @xmath490 eigen values @xmath491 and eigen vectors . are degenerate . ]
the number of terms in the series occurring in @xmath318 , @xmath319 and @xmath492 determining the tidal response , is proportional to @xmath493 .
it is important to check whether the process adopted and the corresponding series converges and we discuss this issue below . as we have stressed above , the convergence of the series means that the tidal response is mainly determined by a few global modes with a large scale distribution of @xmath439 over the planet .
the results of our numerical calculations reported below allow us to suggest that the numerical method described above is stable and robust for the modes with sufficiently large values of @xmath495 .
however , there is a problem with this approach related to the case of small values of @xmath496 .
this problem is associated with the existence of a non - trivial null space of the operator @xmath497 .
it is easy to see that all modes satisfying @xmath498 are degenerate solutions of @xmath435 ( and @xmath405 ) with @xmath494 .
in fact any @xmath439 that is independent of @xmath499 will do .
the corresponding norm @xmath500 .
physically , for @xmath501 these modes correspond to introducing an infinitesimal amount of differential rotation and thus slightly shifting the static equilibrium .
as follows from equation @xmath502 the overlap integrals are formally infinite for these modes .
however , in the exact theory this does not lead to any difficulty since the contribution of a particular mode to the energy exchange is proportional to @xmath503 ( see e.g. equation @xmath318 .
this means that the degenerate set of modes is in fact not excited .
however , due to numerical inaccuracy of our method the modes formally belonging to the null space of @xmath497 acquire small non zero @xmath491 .
the corresponding overlap integrals are very large and they can lead , in principal , to a very large unphysical energy transfer .
this problem is not significant for the case of @xmath298 . in this case , the quantities @xmath504 entering in @xmath318 are very small for an interesting range of @xmath505 .
they suppress effectively the contribution of the spurious modes to the energy and angular momentum exchange . on the other hand , in the case of @xmath27 the corresponding quantities
@xmath506 do not suppress this contribution , and the spurious modes do lead to a very large unphysical energy transfer . therefore , in principle , the numerical method must be able to overcome this difficulty , by e.g. incorporating the solutions to @xmath507 in the set of the trial functions . this can be done by considering polynomials consisting of products of different power of @xmath508 and @xmath499 as a new set of trial functions ( pp ) instead of solutions to @xmath459 . however , this set is difficult to use when considering a large values of @xmath493 .
therefore , in our numerical work we simply identified the spurious modes corresponding to @xmath27 by comparing the spectra calculated using these two different sets of trial functions and removed them from the summation in the expression for the transfer of energy ( there is no angular momentum exchange in this case ) .
the resulting contribution to the energy exchange appears to be very small compared with what is obtained for the case of @xmath298 , and we will not discuss further the case @xmath27 below . in this section
we discuss in detail the case of a planet with @xmath409 and @xmath411 .
this case has the most complicated spectrum of all models we consider . in our opinion , this is mainly because of the presence of the phase transition at a sufficiently small radius @xmath425 .
the analysis of the eigenfrequencies and the eigenmodes developed for this case can be easily applied to planets with larger radii and masses which have a simpler structure . in figure [ fig.6 ]
we show the overlap integrals
@xmath511 ( see equation @xmath512 for the definition ) calculated for this case as a function of the eigenfrequency @xmath513 .
it can be shown analytically ( pp ) that @xmath514 .
the curves of different types correspond to different numbers of the trial functions , @xmath515 , @xmath516 and @xmath517 .
we show positions of particular @xmath511 representing results of calculations with @xmath515 and @xmath518 as circles and squares , respectively . as we have mentioned above the sharp rise near @xmath519
is due to numerical inaccuracy .
there is a similar unphysical rise near @xmath520 .
it is clearly seen from figure [ fig.6 ] that the width of these numerical features depends on the number of trial functions @xmath493 decreasing with increase of @xmath493 .
similar to the case of a polytropic star with index @xmath521 ( pi ) , there are two global modes with large values of the overlap integrals , with @xmath522 and @xmath523 .
the frequencies and amplitudes of these modes depends only weakly on the number of trial functions , and these modes are especially important for the problem on hand .
however , contrary to the case of a polytropic star with index @xmath521 , there are some other markedly large values of @xmath511 associated with modes at other eigenfrequencies ( eg . near @xmath524 and @xmath525 ) .
these can also be associated with global modes , which are , however , less stable with respect to changes of @xmath493 .
they can not be reliably identified with a small number of the trial functions ( @xmath515 in our case ) .
we discuss properties of these modes later in this section . in figure [ fig.7 ]
we show the quantities @xmath526 @xmath278 and @xmath527 defined by equations ( [ eq p891]-[eq p89 ] ) which determine the energy and angular momentum transfer through equations @xmath320 and @xmath321 . as for figure [ fig.6 ]
the curves of different types correspond to different values of @xmath493 .
it is seen from this figure that when @xmath528 is sufficiently large the curves representing large values of @xmath518 and @xmath517 are in a very good agreement with each other .
the curves representing @xmath527 and @xmath278 for @xmath515 are also in qualitative agreement with those corresponding to the larger value of @xmath493 .
however , the curve representing @xmath529 deviates significantly when @xmath530 .
when @xmath528 is small , the curves representing @xmath515 and @xmath516 have larger values compared to the case with @xmath531 .
this is due the contribution of spurious modes with eigenfrequencies very close to @xmath532 . in principle
, we can improve the agreement between different curves significantly simply by discarding these modes .
however , as we will see below it is possible to improve the stability of our numerical scheme with respect to the choice of different numbers of trial functions and suppress the influence of these spurious modes automatically . in order to improve the stability of our numerical scheme with respect to numerical noise we would like to separate the eigenspectrum into bins with a finite frequency width and adopt the sum of the overlap integrals within a particular bin instead of @xmath511 . to begin , we redefine the overlap integral according to the rule @xmath533 it follows from equation @xmath534 , that the quantity @xmath529 has especially simple form when expressed in terms @xmath535 , namely @xmath536 to define the frequency bins let us separate the eigenspectrum into regions between successive minima of @xmath537 and treat all @xmath538 and @xmath491 within such a region as belonging to the same bin .
note that modes defining minima are not included in the bins .
accordingly , we have @xmath539 as new binned overlap integrals and binned eigenfrequencies . here
, the summation is performed over all eigenfrequencies belonging to the same bin , and @xmath540 is a number of a bin . the result of comparison between the binned quantities @xmath541 corresponding to different @xmath493 and the quantities @xmath542 calculated with the largest @xmath531 is shown in figure [ fig.8 ] .
one can see from this figure that the binned quantities calculated for @xmath518 and @xmath531 are in a good agreement .
the results corresponding to @xmath515 agree with what is shown for larger values of @xmath493 for the two main global modes .
they are also in qualitative agreement for @xmath543 , but deviate in a part of the spectrum with positive values of @xmath491 .
it turns out that this deviation is not important for our purposes .
indeed , the positive part of the spectrum plays a role in the tidal response problem when @xmath279 is small . in this case
width of the function @xmath504 is rather large , and the tidal response is mainly determined by the main global mode with @xmath544 and the largest overlap integral .
when @xmath279 is large , the tidal response is mainly determined by the modes with negative @xmath491 which can then comove with the tidal perturbation near periastron .
the width of @xmath504 decreases with increase of @xmath279 , and , when the rotation rate is large , the tidal response is determined by several modes with sufficiently large overlap integral and sufficiently large absolute values of @xmath491 .
this effect explains in part why the energy transfer in the rotating frame determined by the quantity @xmath529 tends to some non - zero value with increase of the rotation rate ( see e.g. figure 7 and figure 9 ) . in figure [ fig.9 ]
we show the quantities @xmath527 , @xmath529 and @xmath545 calculated using the binned overlap integrals for the different values of @xmath493 and the corresponding quantities calculated using the non - binned overlap integrals with @xmath531 .
comparing figure [ fig.7 ] with figure [ fig.8 ] we see that the use of the binned quantities does allows us to improve significantly the stability of our numerical scheme with respect to different numbers of basis functions .
now , all curves are in a good agreement with each other for the whole considered range of @xmath279 . in figure [ fig.10 ]
we show the result of calculation of the overlap integrals for the model with smoothed out phase transition ( dashed curve ) together with the overlap integrals for the model with the phase transition .
as it is seen from this figure the two main global modes have practically the same overlap integrals for the both models . however , other prominent spikes present in the model with a phase transition are absent for the model without the phase transition .
therefore , there only two global modes in that case .
one can assume that the presence of the other global non standard modes in the realistic planet models is related to the presence of the phase transition .
in fact , as we will see below these modes are also present in the planet models with larger radii where the phase transition is absent .
it is reasonable to suppose that the presence of some of these modes is related to a sharp decrease of the density after the radius of the phase transition ( or the melting point for the models with larger radii , see figures [ fig.3 ] and [ fig.4 ] ) in the realistic planet models .
as we have pointed out above there is a set of modes characterised by relatively large values of the overlap integrals .
these modes are referred to as global modes .
physically , these modes have eigenfunctions for which @xmath439 varies on a large scale . in this situation ,
the overlap integrals are not averaged to small values after integration over the volume of the planet . as can be seen from figures [ fig.6 ] , [ fig.8 ] and [ fig.10 ]
there are more than ten modes with markedly large overlap integrals .
however , they are not equally important for the problem on hand .
the two most important modes are the main global modes with the largest values of the overlap integrals .
these are present in all our models as well as in polytropic stars .
other global modes with negative values of @xmath491 can also play an important role in the tidal response .
additionally , it is interesting to consider the positive and negative frequency global modes with the largest values of @xmath538 after the main modes .
these modes contain information about the difference of the realistic planet models from over - simplified polytropes , and may be of interest for another problems as well .
there are eight modes satisfying these criteria .
their eigen frequencies and overlap integrals are shown in table [ table1 ] . [
cols="<,<,>,>,>,>,<,>,<,>",options="header " , ] as follows from these table apart from the case of a planet with jupiter mass and radius which gives a rather large value @xmath546 , the values of @xmath297 and @xmath547 are rather close to each other . therefore ,
in our estimates of the time scale of tidal circularisation below we will use the typical values @xmath548 and @xmath375 .
accordingly , the energy transfer will be given by @xmath549 regardless of the planet radius and mass .
in figure [ fig.16 ] we show the dependence of @xmath292 on @xmath550 as a solid curve .
the dotted , dashed and dot - dashed curves correspond to the energy transfer associated with the fundamental modes , @xmath551 .
they are calculated with help of equations derived in ip and presented in appendix .. this has no influence on the conclusions of ip .
the correct expressions are shown in appendix . ] in a similar manner to the energy transfer associated with the inertial modes , @xmath551 depends on the angular velocity of the planet , and we present several different curves calculated for different values of @xmath552 .
the dashed curves are calculated for a non - rotating planet with @xmath553 , they correspond to a maximal energy transfer @xmath554 as a function of @xmath555 .
the dotted curves are calculated for the case @xmath556 .
the dot - dashed curves are calculated for the equilibrium angular velocity of the planet , @xmath557 , associated with the fundamental modes , see equation @xmath558 of appendix . these curves give a minimal energy transfer @xmath559 as a function of @xmath552 , for a given @xmath550 . as follows from appendix , in order to calculate @xmath551 the eigenfrequencies associated with the fundamental modes , @xmath560 as well as the corresponding dimensionless overlap integrals , @xmath561 must be known .
these quantities depend on the planet structure , and therefore they are different , in principle , for planets with different masses and radii , see ip for their numerical values . in figure [ fig.16 ]
we show the case of a planet with @xmath413 and @xmath562 ( the upper dashed , dotted and dot - dashed curves ) , where we have @xmath563 and @xmath564 .
the lower dashed , dotted and dot - dashed curves represent the case of a planet with the same value of mass and @xmath565 with @xmath566 and @xmath567 . as follows from the results provided in ip , when @xmath414 the dependence of the eigenfrequencies and overlap integrals on the planet radius is rather weak , and the corresponding energy transfer is always approximately represented by the upper dashed and dotted curves in figure [ fig.16 ] . as follows from figure [ fig.16 ] the inertial modes
dominate the energy transfer for sufficiently large values of @xmath550 .
when @xmath374 the contribution determined by the inertial waves dominates over that corresponding to the fundamental modes for @xmath568 in the case of @xmath562 and for @xmath569 in the case of @xmath565 .
thus , we expect that the circularisation of the planet orbits with sufficiently large periastron distances is mainly determined by the contribution associated with the inertial modes .
in this section we apply the results developed in the previous sections to estimate the effect of tides exerted by the central star on the planet on the planet orbital evolution .
we assume that the state of pseudo - synchronisation with @xmath570 is maintained in a situation when the inertial modes dominate over the fundamental modes in tidal response .
apart from the fact that we mainly consider the effects of inertial waves , the formulation of the problem is analogous to what was discussed in ip .
let us consider a planetary system consisting of a few massive planet and a central star and assume that a certain planet has lost a significant amount of its orbital angular momentum , most probably due to gravitational interaction with other planets in the planetary system .
such situations have been frequently observed in numerical simulations of gravitationally interacting planets ( papaloizou @xmath21 terquem 2001 , adams @xmath21 laughlin 2003 )
. the orbital semi - major axis of this planet , in this case , could be of the order of its typical value @xmath571 while the periastron distance @xmath155 could be very small , of the order of a few stellar radii .
for a such eccentric orbit the orbital angular momentum per unit of mass is related to the periastron distance as @xmath572 .
if we assume that the orbital evolution is determined by tides , the angular momentum is approximately conserved while the semi - major axis may decrease due to transfer of the orbital energy into oscillations of the planet by tides and subsequent dissipation . then , the energy may be radiated away from the planet .
the conservation of the angular momentum leads to the well known fact that the radius a final circular orbit , @xmath573 is twice as large as the periastron distance @xmath155 .
therefore , @xmath155 can be related to the observed orbital period of the planet after circularisation , @xmath574 , as @xmath575 where @xmath576 , @xmath577 is the stellar radius and @xmath578 is the radius of the sun .
we have , accordingly , from equation @xmath364 @xmath579 for the planet and the star with jupiter and solar values of mass and radius , respectively , @xmath580 .
let us consider the evolution of the semi - major axis of the eccentric planet orbit due to tidal interactions near the orbit periastron .
we can assume that the contributions from each of successive impulsive energy transfer can be added to each other only when one of two conditions is fulfilled : either the energy is dissipated during two successive periastron passages or the condition for the stochastic instability @xmath581 is satisfied .
in the latter case the evolution of the semi - major axis is stochastic , and we assume below that , in this case , our evolution equations are written for averaged over some distributions values of the corresponding quantities . under these assumption
we can write the evolution equation for semi - major axis , @xmath582 , in the form ( ip ) @xmath583 where @xmath584 .
the characteristic evolution time @xmath585 depends on the energy transfer due to dynamic tides associated with the inertial and fundamental modes in the planet , @xmath292 and @xmath551 , respectively , and also on the transfer of the orbital energy into stellar oscillations determined by dynamic tides exerted on the star @xmath586 where @xmath292 is given by equation @xmath313 and @xmath587 stands for the energy transfer associated with dynamic tides in the star .
note that the time @xmath588 implicitly depends on time @xmath25 through dependence of the planet radius @xmath258 on time due to the evolutionary cooling and a possible tidal heating of the planet .
equation @xmath589 can be integrated to give @xmath590 where @xmath591 is the initial value of the semi - major axis expressed in units of @xmath592 .
a more accurate compared to @xmath589 definition of the evolution time scale , @xmath593 , can be obtained from equation @xmath594 requiring that the semi - major axis @xmath595 is formally equal to zero when @xmath596 .
thus @xmath597 for a constant value of @xmath588 , @xmath598 . in this case
we can obtain a simple estimate of @xmath593 assuming that the transfer of energy is determined only by the inertial modes .
setting @xmath599 in equation @xmath600 and substituting equations @xmath313 , @xmath601 and @xmath600 into @xmath589 , we have @xmath602 note a very strong dependence of @xmath593 on @xmath258 and @xmath246 .
results of numerical calculations of @xmath593 are shown in figure [ fig.25 ] for the case of a planet with @xmath409 , and in figure [ fig.26 ] for a planet with @xmath414 . in order to calculate
@xmath593 we use expressions for the energy transfer associated with the fundamental modes given in appendix , the corresponding eigenfrequencies and overlap integrals are given in ip .
the dependence of @xmath258 on time is taken from burrows et al ( 1997 ) where a cooling history of isolated planets is considered .
thus we neglect effects of tidal heating on the planet size and structure reserving their discussion for a future work .
burrows et al ( 1997 ) consider evolution of the planet radius starting from the initial time @xmath603 and until a maximal time @xmath604 is reached .
we set the initial time @xmath605 .
we assume that the planet radius is not changed significantly for @xmath606 and set @xmath607 with @xmath608 for @xmath413 and @xmath609 for @xmath414 .
the lower curves of the same type in figures [ fig.25 ] and [ fig.26 ] correspond to @xmath610 and the upper curves are calculated for @xmath611 . in figure [ fig.25 ]
the solid curves describe the dependence of @xmath593 on @xmath246 calculated under assumption that only inertial waves contribute to the tidal evolution , and accordingly , in this case @xmath612 in equation @xmath600 . for the dashed curves we take into account all the contributions to the sum in @xmath600 .
is , strictly speaking , not correct .
however , we assume hereafter that this procedure gives correct order of magnitude estimates reported in this paper . ] in order to calculate @xmath587 we model the star as a non rotating @xmath613 polytrope and take into account the contributions of @xmath19 , @xmath3 and low order @xmath224 modes , see ip for details .
the contribution of the fundamental modes operating in the planet , @xmath551 is calculated for the case of rotation of the planet at the equilibrium angular velocity associated with the fundamental modes , see equation @xmath558 of appendix . in this case
@xmath551 is given by equation @xmath614 .
this may be justified according to the following arguments .
the inertial modes dominates over the fundamental modes at sufficiently large values of @xmath550 , and accordingly at sufficiently large values of @xmath246 , and the fundamental modes dominate at smaller values of @xmath246 .
taking into account that the contribution of fundamental modes decreases exponentially with increase of @xmath246 , the transition from the case of domination of the fundamental modes to the case of domination of inertial modes occurs near a certain value of @xmath246 . taking into account that a planet with a low value of the moment of inertia has the angular velocity always close to the equilibrium value determined by the modes dominating the energy transfer
, the dashed curve could give an order of magnitude estimate of @xmath593 for the whole range of @xmath246 .
the dotted curves show the results of calculation of @xmath593 for the case of a non rotating planet . in this case
the energy transfer determined by the inertial modes is absent and we set @xmath615 in equation @xmath600 and we use equation @xmath616 to calculate @xmath551 . as seen from figure [ fig.25 ] when @xmath409 the inertial modes dominate for @xmath617 .
the time scale of circularisation is smaller than a typical life time of a planetary system @xmath618 when @xmath619 for @xmath610 and when @xmath620 for @xmath611 .
it is interesting to note that the contribution of the fundamental modes sharply decrease when @xmath621 .
this is simply because the energy transfer associated with the fundamental modes is highly sensitive to the radius of the planet ( see appendix ) .
it can lead to a substantial change of the semi - major axis only when the planet radius is close to its initial value @xmath622 .
the cooling of the planet results in a decrease of the planet radius on a time scale @xmath420 and the contribution of the fundamental modes quickly become unimportant after this time is reached .
the tides exerted on the star are not significant at all for a planet with @xmath409 . in figure [ fig.26 ]
we show the results corresponding to a planet with @xmath414 .
contrary to the previous case now the inertial modes are not important for the whole range of @xmath246 and the circularisation time scale at large values of @xmath246 is determined by tides exerted by the planet on the star .
this may be explained as follows . at large values of @xmath246
the circularisation time scale is of the order of several @xmath12 . a planet with its mass in the range @xmath623 and its age of the order of a few @xmath12 has the radius close to @xmath624 .
the circularisation time scale is mainly determined by the expressions for the energy transfer calculated for this radius .
assuming that the inertial modes dominate we can use equation @xmath625 to estimate @xmath593 . as follows from this equation @xmath626 . when eg . @xmath627 and
@xmath610 we have @xmath628 for @xmath409 and @xmath629 for @xmath414 .
obviously , the inertial modes can not lead to a significant decrease of the semi - major distance in the latter case over the life time of the planetary systems .
unlike the situation with the inertial modes , the energy transfer determined by the stellar modes grows with the planet mass , and therefore the tides exerted on the star dominate when the planet mass is sufficiently large .
in this paper we have developed and extended a new self - adjoint formalism for the problem of small stellar oscillations proposed in our recent paper(pi ) .
we went on to apply this to the calculation of the tidal response of uniformly rotating fully convective planets and stars which undergo a single encounter in a parabolic orbit or a sequence of multiple encounters in a weakly bound orbit .
this was then applied to the problem of the tidal circularisation of the orbits of the extrasolar planets . in section 2
we began by showing that the general formalism , presented in pi when @xmath630 is used as variable characterising linear eigenmodes , can also be applied when the lagrangian displacement is used .
the advantage of this approach is that , as no low frequency approximation is needed , it allows us to consider pulsations of arbitrary frequency .
we derived general expressions for the energy and angular momentum transfer that occurs when a rotating planet or a star passes through periastron in a parabolic or highly eccentric orbit around a central mass the amplitudes and phases of the eigenmodes excited in the planet or a star as a result of the tidal encounter were also determined .
we showed that these expressions can be represented in terms of a spectral decomposition over the normal modes of the self - adjoint operator @xmath631 .
we also obtained a self - consistent expression for the rate of dissipation of an excited mode to leading order in the small parameter defined by the ratio of pulsation period to a characteristic dissipation time scale . in section 3
we focused on the case of low frequency inertial modes , showing how the more general formalism in terms of the lagrangian displacement can be reduced to that in pi in the low frequency limit .
this procedure allows one to separate out the contribution associated with the inertial modes and simplify the equations governing free pulsation of a rotating planet or a star . in that case
the corresponding self - adjoint problem is formulated through equations @xmath405 and @xmath456 .
note that the tidal response of the inertial modes to impulsive tidal forcing has not been described using other methods .
we showed in section 4 that when the rotating planet or a star approaches periastron in an unperturbed state the tidal energy and angular momentum transfered through the excitation of inertial modes are given by very simple expressions @xmath320 and @xmath301 .
we considered the multi - passage problem assuming that there was no significant dissipation of mode energy between successive periastron passages , and found a simple condition @xmath581 for the occurrence of stochastic instability that results in the stochastic gain of inertial mode energy over many periastron passages .
this was quantitatively similar to that given by ip for @xmath3 modes .
we found that for stochastic instability the circularisation process has to start with @xmath632 for final periods of @xmath5 days . or
@xmath6 for final orbital periods @xmath7 days .
in section 5 we applied our formalism to fully convective rotating giant planets . in order to evaluate the expressions for the energy and angular momentum transfer the eigenspectrum of @xmath456
must be found numerically , for a given model .
we calculated the eigenspectrum of the inertial modes for several planet models with a realistic equation of state .
the details and stability of the numerical method we used are discussed in section 5.2 .
we considered planet models with two different masses equal to @xmath633 and @xmath410 , with radii in the range @xmath634 .
we found that , as in the case of @xmath20 polytrope considered by pi , the tidal response was determined by a few global eigenmodes with a large scale distribution of perturbed quantities .
two main or standard global modes have eigenfrequencies close to those corresponding to the @xmath635 polytrope .
these modes have the largest overlap integrals characterising coupling between the tidal field and the eigenmodes .
however , we found that for realistic planet models , there could be non - standard global modes possibly related to a sharp change of structure occurring near to the point of ionisation of hydrogen .
the structure of the non - standard modes as well as their stability with respect to a change of the planet mass and radius was discussed .
the results obtained for the eigen spectra of realistic planet models were applied to the problem of tidal circularisation of the orbits of the extrasolar planets in sections 6 and 7 . in section 6
we calculated the rotational angular velocity for which the net angular momentum transferred was zero . because of the relative low inertia of the planet compared to the orbit it is expected to rapidly attain this angular velocity and achieve a state of pseudo synchronisation . the angular velocity associated with pseudo synchronisation
was found to be always close to @xmath2 times the angular velocity of a circular orbit at periastron .
orbital evolution arises through the transfer of orbital energy to the modes of oscillation .
transfers at successive periastron passages lead to orbital circularisation .
this is the case , either when the mode energy is dissipated directly between encounters , or when the system is in the regime of stochastic instability .
we compared the contribution associated with the inertial modes with the contribution associated with the fundamental modes , in a state of pseudo synchronisation , to the circularisation time scale in section 7 .
we found that the inertial modes led to effective circularisation of a giant planet with mass @xmath636 on a time scale smaller than or of the order of a few @xmath12 , when the initial semi major axis was less than @xmath9 and the final orbital period after circularisation @xmath637 in this case the inertial modes play the most important role at the longer periods .
however , the inertial waves are less important for the planets of larger mass @xmath410 that we considered . in that case the dynamic tides exerted on the star central star , which we assumed to be solar like , play a major role . in our opinion ,
the most important unresolved question related to the theory discussed above is concerned with internal dissipation .
of particular concern is the internal location of modal dissipation and the related issue of whether the energy deposited can be radiated away between successive periastron passages .
there are several potentially interesting channels of dissipation which may operate in our case .
the non - linear parametric instability discussed by kumar and goodman ( 1996 ) for the case of @xmath19 modes in convectively stable stars may be effective for the inertial modes as well .
this instability operates when there are eigenmodes in the stars with eigenfrequencies approximately half that of the eigenfrequencies of the tidally excited modes and sufficiently large coupling coefficients .
these modes may be unstable on account of parametric resonance .
taking into account that the eigenspectrum of the inertial modes is dense , unstable modes are expected to exist for sufficiently weak dissipative processes .
this question will be addressed in our future work .
our formalism is general enough that it can be extended to the case of convectively stable uniformly rotating stars and it can also be applied to other astrophysical situations where excitation of stellar pulsations with frequencies comparable to the angular velocity of rotation is important . in this particular work
we assumed that the orbital and planetary angular momentum vectors were aligned .
this is a very reasonable assumption when the system has many periastron passages and so readily attains a state of pseudo synchronisation during the circularisation process .
however , misaligned angular momenta should be considered if the main issue is tidal capture , as in globular clusters , because in that case there is no reason for the initial angular momentum vectors to be correlated .
these issues and extensions of the formalism will be considered in future work .
we are grateful to g. ogilvie for discussions .
pbi has been supported in part by rfbr grant 04 - 02 - 17444 .
99 adams , f. c. , laughlin , g. , 2003 , icarus , 163 , 290 bodenheimer , p. , pollack , j. b. , 1986 , icarus , 67 , 391 burrows , a. , marley , m. , hubbard , w. b. , lunine , j. l. , guillot , t. , saumon , d. , freedman , r. , sudarsky , d. , sharp , c. , 1997 , apj , 491 , 856 friedman , j. l. , schutz , b. f. , 1978 , apj , 221 , 937 ivanov , p. , papaloizou , j. , 2004 , mnras , 347 , 437 ( ip ) kochanek , c. s. , 1992 , apj , 385 , 604 kumar , p. , goodman , j. , 1992 , apj , 385 , 604 mardling , r. a. , 1995 , apj , 450 , 722 mardling , r. a. , 1995 , apj , 450 , 732 ogilvie , g. i. , lin , d. n. c. , 2004 , apj , 610 , 477 papaloizou , j. c. b. , ivanov , p. b. , 2005 , mnras , 364l , 66 ( pi ) papaloizou , j. c. b. , pringle , j. e. , 1981 , mnras , 195 , 743 ( pp ) papaloizou , j. c. b. , terquem , c. , 2001 , mnras , 325 , 221 press , w. h. , teukolsky , s. a. , 1977 , apj , 213 , 183 ( pt ) saumon , d. , charbrier , g. , van horn , h. m. , 1995 , apj suppl , 99 , 713 saumon , d. , guillot , t. , 2004 , apj , 609 , 1170
here we show the expressions for energy and angular momentum transfer associated with the the fundamental modes in the limit @xmath638 ( ip ) . for the energy transfer we have @xmath639 and for the angular momentum transfer we have @xmath640 where @xmath641 and @xmath642 @xmath643 and @xmath644 are the dimensionless eigenfrequency and dimensionless angular velocity of the planet , respectively and @xmath645 .
the dimensionless quantity @xmath646 determines the frequency splitting of the mode due to rotation , it is close to @xmath647 for our models ( see ip and references therein ) .
@xmath561 are the dimensionless overlap integrals , see pt and ip .
when @xmath648 the energy transfer as a function of @xmath649 is maximal , @xmath650 assuming that the fundamental modes dominate the tidal response , the state of pseudo - synchronisation is determined by condition @xmath651 . in this case
we obtain from equation @xmath652 @xmath653 when @xmath654 the energy transfer is minimal @xmath655 | we consider the general problem of the tidal capture or circularisation from large eccentricity of a uniformly rotating object .
we extend the self - adjoint formalism introduced in our recent paper(papaloizou & ivanov 2005 hereafter pi ) ) to derive general expressions for the energy and angular momentum transfered when the planet or a star passes through periastron in a parabolic or highly eccentric orbit around a central mass .
these can be used without making a low frequency approximation as was done in pi .
we show how these can be adapted to the low frequency limit in which only inertial modes contribute for baratropic planet models . in order to make quantitative estimates ,
we calculate the inertial mode eigenspectrum for planet models of one and five jupiter masses @xmath0 without a solid core , with different radii corresponding to different ages .
the spectra are found in general to be more complex than of a polytrope with index @xmath1 considered in pi , because of the existence global modes associated with the transition from molecular to metallic hydrogen .
nonetheless the main tidal response is still found to be determined by two global modes which have polytropic counterparts .
these also determine the uniform angular velocity in a state of pseudo synchronisation , for which the angular momentum transfered during an encounter is zero .
this is found to be close to @xmath2 times the circular orbit angular velocity at periastron for all models considered .
this is in contrast to the situation when only the @xmath3 mode is considered ( ivanov & papaloizou 2004 , hereafter ip ) and the equilibrium angular velocity is found to be much larger .
we consider the multi - passage problem when there is no dissipation finding that stochastic instability resulting in the stochastic gain of inertial mode energy over many periastron passages occurs under similar conditions to those already found by ip for the @xmath3 modes .
we find that this requires circularisation to start with a semi - major axis exceeding @xmath4 for final periods of @xmath5 days .
reducing to @xmath6 for final periods @xmath7 days .
finally we apply our calculations of the energy transfer during a periastron passage to the problem of the tidal circularisation of the orbits of the extrasolar planets in a state of pseudo synchronisation , expected because of the relatively small inertia of the planet , and find that inertial mode excitation dominates the tidal interaction for @xmath8 planets that start with semi- major axes less than @xmath9 and end up on circular orbits with final period in the @xmath10 day range .
it is potentially able to account for initial circularisation up to a final @xmath11 day period within a few @xmath12 but in the case of @xmath13 oscillation modes excited in the star are more important .
[ firstpage ] hydrodynamics ; stars : oscillations , binaries , rotation ; planetary systems : formation |
accretion onto black holes commonly results in relativistic , collimated outflows , or jets . in black hole x - ray binaries ( bhxbs ) , these jets can brighten , fade or even switch off all together on timescales of days months ( see * ? ? ? * for a review ) .
when jets are observed in bhxbs they usually manifest themselves as powerful , compact outflows commonly seen as a flat or slightly inverted radio spectrum ( @xmath1 0.5 , where @xmath2 ) and are spatially resolved in some cases ( e.g. * ? ? ?
* ; * ? ? ?
* miller - jones et al . in preparation ) .
both properties are in line with what is expected for partially self - absorbed overlapping synchrotron spectra @xcite .
these compact jets of bhxbs are typically launched continuously for several months at least during a single outburst , when the x - ray spectrum is a hard power law ( the hard state ; e.g. * ? ? ?
* ; * ? ? ?
* ) and are considered to be analogous to persistent compact radio sources associated with active galactic nuclei ( agn ) . in bhxbs ,
most of the energy radiated by these jets is emitted at higher energies , for example in the infrared / optical regime ( e.g. * ? ? ?
* ; * ? ? ?
* ) or even higher x - ray or possibly @xmath3-ray energies in some cases ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the direct detection of a break in the jet spectrum in one bhxb , gx 3394 at mid - infrared wavelengths has allowed a measurement of the total compact jet radiative power to be made , @xmath4 erg s@xmath5 ( assuming a distance of 8 kpc ) which is @xmath6 per cent of the total bolometric x - ray luminosity @xcite , ( and breaks have recently been constrained in several more bhxb jet spectra ; * ? ? ?
* ; * ? ? ?
* ) . these jets are radiatively inefficient , and the total ( kinetic plus radiative ) energy channelled into the jets can be approximately half of the available accretion energy or perhaps more @xcite . during accretion state transitions at high x - ray luminosities ( @xmath7 per cent of the eddington luminosity ;
@xmath8 ) , a different type of jet is commonly seen in bhxbs .
the compact jet fades and discrete , ballistic ejections appear as bright , optically thin radio flares , sometimes later resolved into two expanding structures either side of the core bhxb ( e.g. * ? ? ?
* ; * ? ? ?
these discrete ejections are typically seen once , or at most a few times during a bhxb outburst cycle .
each ejection has a total radiative luminosity of @xmath9
1 @xmath8 ( see e.g. table 1 in * ? ? ?
* ) , and may be associated with a fast , relativistic ejection ( see also * ? ? ?
* ) . for comparison , a compact jet such as that of gx 3394 mentioned above produces this much radiative energy in @xmath10 minutes during a bright hard state @xcite .
since hard states normally last several months , the transient jets probably provide a negligible fraction of the total radiative power of jets from bhxbs .
however , for both types of jet , the total kinetic power they contain is largely unknown , so comparisons can only be made between their radiative luminosities .
it is worth assessing whether the same physical process is responsible for launching both types of jet . the physical mechanism responsible for the launching , acceleration and collimation of jets has been debated for decades , and remains one of the major paradigms in the field of accretion in strong gravitational fields .
it is generally accepted that magnetohydrodynamic ( mhd ) processes are at work ( e.g. * ? ? ?
* ) , but two groups of models are usually considered that can both produce the poloidal magnetic fields and differential rotation necessary to launch powerful jets @xcite .
the blandford
znajek mechanism describes the extraction of the rotational energy of spinning bhs , in a region close to the bh where frame - dragging occurs ( @xmath11 gravitational radii ; @xmath12 ; e.g. * ? ? ?
* ; * ? ? ?
the alternative to spin - powered jets is purely accretion - powered jets ; the blandford
payne mechanism , where all the energy and rotation originates in the disc and its magnetic fields .
the blandford
znajek model predicts that the total jet power scales with the second power of the dimensionless spin parameter , @xmath13 , where @xmath14 and @xmath15 are the black hole angular momentum and mass ( see * ? ? ?
* for a similar scaling based on general - relativistic magnetohydrodynamics simulations ) .
the two leading methods currently adopted to infer bh spin in bhxbs are x - ray thermal continuum fitting , and x - ray reflection / fe k@xmath16 line fitting .
the former applies to spectra of bhxbs in the thermal dominant ( or high / soft ) state @xcite , while the latter fits ( low/)hard state spectra at eddington rates higher than about @xmath17 . in essence , the thermal continuum fitting method provides a measure of the area subtended by the innermost stable circular orbit ( isco ) , while the reflection method returns a measure of the gravitational redshift that shapes the red wing of the broadened fe k@xmath16 line .
both methods rely on the assumption that the accretion disc extends in to the isco , which is in turn a function of bh spin .
the reader is referred to e.g. @xcite and @xcite for further details on the two methodologies .
bhxbs , as opposed to agn , are particularly suited for this kind of study . in outburst ,
galactic bhxbs tend to be brighter than all agn in the sky , so with more photons , spectral fitting can be achieved with a higher level of precision .
in addition , the thermal continuum fitting method can not be applied to agn , since the peak of the disc emission falls in the uv band .
measures of bh jet powers also rely on different methods , arguably all more indirect than those adopted for estimating the spin parameter , such as the relative normalization of the radio / infrared ( ir ) versus x - ray luminosity in the hard state for compact jets , and the total energy in the bright radio flares that are thought to be at the origin of the large scale , transient jets . in a previous work ( * ? ? ?
* hereafter fgr10 ) we performed the first study to observationally test for empirical relations between bh spin ( @xmath18 ) and jet power .
x - ray and ir x - ray correlations exist in the hard state but some bhxbs are more radio - bright ( or infrared - bright ) than others , at a given x - ray luminosity .
the luminosity of the jet at radio and ir frequencies relative to the x - ray luminosity ( the normalization of the correlation ) was used to rank the bhxbs in order of jet power .
no single relation was found between current reported bh spin measurements and relative luminosity of compact jets .
this suggested that if the relative jet power estimates and bh spin estimates are accurate , compact jets may not be spin - powered . since the acceleration and collimation regions of compact jets are usually considered to be at least tens to hundreds of @xmath19 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , these results may not be surprising because this is essentially outside the region where bh spin energy can be tapped .
in addition , the jet is usually required to be anchored to a thick disc , which is thought to be present in the hard x - ray state ( e.g. @xcite ; see also @xcite ; however a thin disc that extends close to the isco may exist above @xmath20 in the hard state ; @xcite ) .
very recently , the actual base of the jet of the agn in m87 has been spatially resolved using radio interferometry , and its size was measured to be @xmath21 @xcite .
however even here , the wide opening angle of the sub - relativistic jet and the existence of a counter jet indicates that this jet is anchored within ( and powered by ) the accretion disk and not launched due to magnetic field lines crossing the event horizon @xcite .
the radio / ir normalization for a single source can change in time @xcite which , if x - ray luminosity is not the quantity that is changing , invalidates a single relationship between radio / ir luminosity normalization and jet power ( or jet radiative efficiency ) .
some mhd simulations of disc - emitting jets have also successfully described properties of the outflows commonly seen in bhxbs ( e.g. * ? ? ?
* ; * ? ? ?
in addition , @xcite found no significant correlation between the spin of neutron stars in x - ray binaries ( which can be measured accurately for some but not all sources considered ) and their relative jet powers inferred from radio or infrared jet luminosities ( although less data were available for neutron star jets , and those data span a smaller range in luminosity than the sample of bhxbs ) . finally , in a related study , @xcite considers the johannsen - psaltis metric around a bh , a speculative , alternative space - time metric to the kerr metric . in this scenario ,
current data of bhxbs suggest there may be a correlation between bh spin as defined by this metric , and compact jet power .
the only tentative relation found by fgr10 was between _ transient _ jet ejection energy and bh spin measured via reflection fits ( but not continuum fits ) , from four data points ( fig . 6 in fgr10 ) .
more recently , @xcite ( 2012 ; hereafter nm12 ) claim a relation between transient jet peak radio luminosity and bh spin measured from thermal continuum fitting , using a sub - sample of the data in the literature .
nm12 take the peak luminosity of the radio flare associated with the hard - to - soft state transition as their estimate of the jet power ( as opposed to an estimate of the total energy of the ejection ) . here , we test for a relation between transient jet power and reported bh spin using all the available data ( described in section 2 ) , adopting two differing methods to infer the jet power and two methods to estimate bh spin .
we make use of a bayesian fitting routine in order to quantify the significance of any relation ( the methodology is explained in section 3 ) . in all cases ,
no significant correlation is found .
if the methods used to infer jet power and bh spin are correct , we can rule out a single positive correlation between transient jet power and bh spin with high confidence .
a discussion and brief summary is provided in section 4 .
we also comment on a very recent paper , @xcite in section 4 .
this new work follows on from that of nm12 , adding one more source to their sample .
this new source was already included in our samples here , and we demonstrate that this paper does not affect our results or conclusions . [ cols= " < , < , < , < , < , < , < , < , < , < , < , < " , ] @xmath22for grs 112468 and gs 2000 + 25 , bh spin estimates were reported with no errors quoted ; here we assume the same errors as a062000 ( which has a similarly low spin ) .
@xmath23the distance and bh mass of some bhxbs are uncertain ( e.g. * ? ? ?
* ; * ? ? ?
* ) ; here we assume 8 kpc ( since these sources are towards the galactic centre ) and 10 m@xmath24 , respectively .
the peak radio flux density and rise times are tabulated for each radio flare over the hard - to - soft state transition reported for each source . in the sixth column , the year is that of the radio flare .
references : ( 1 ) = @xcite ; ( 2 ) = @xcite ; ( 3 ) = @xcite ; ( 4 ) = nm12 ; ( 5 ) = @xcite ; ( 6 ) = @xcite ; ( 7 ) = @xcite ; ( 8) = @xcite ; ( 9 ) = @xcite ; ( 10 ) = @xcite ; ( 11 ) = @xcite ; ( 12 ) = @xcite ; ( 13 ) = @xcite ; ( 14 ) = @xcite ; ( 15 ) = @xcite ; ( 16 ) = @xcite ; ( 17 ) = @xcite ; ( 18 ) = @xcite ; ( 19 ) = @xcite ; ( 20 ) = @xcite ; ( 21 ) = @xcite ; ( 22 ) = @xcite ; ( 23 ) = @xcite ; ( 24 ) = @xcite ; ( 25 ) = @xcite ; ( 26 ) = @xcite ; ( 27 ) = @xcite ; ( 28 ) = @xcite ; ( 29 ) = @xcite ; ( 30 ) = @xcite ; ( 31 ) = @xcite ; ( 32 ) = @xcite ; ( 33 ) = @xcite ; ( 34 ) = @xcite ; ( 35 ) = @xcite ; ( 36 ) = @xcite ; ( 37 ) = @xcite ; ( 38 ) = @xcite ; ( 39 ) = @xcite ; ( 40 ) = @xcite ; ( 41 ) = @xcite ; ( 42 ) = @xcite ; ( 43 ) = @xcite ; ( 44 ) = @xcite .
in table 1 we list all bhxbs that we are aware of with both bh spin ( measured either by the continuum fitting method or the reflection fitting method ) and observed transient jet flares .
bh spin and radio data are taken from table 1 of fgr10 , table 1 of @xcite and recent estimates noted in nm12 , plus some new bh spin estimates published since fgr10 but not used in nm12 and radio flare data from the same bhxbs ( see table 1 for all original references for the bh spin and radio data , as well as distances and bh mass estimates ) .
where more than one spin estimate exists using the same method for the same bh , we use the most recent spin measurement , inferred using data from modern x - ray telescopes , since these are likely to invoke the most sophisticated fitting techniques and use the highest quality data . throughout this paper
we consider both leading methods of measuring bh spin , and do not favour one method over the other .
it is not in the scope of this paper to argue that one method provides more accurate spin estimates than the other ; a scientifically fair study should consider all possible published measures . in the long term
, the hope is that the two methods will converge and give consistent results .
this appears to be currently the case for three bhxbs ( grs 1915 + 105 , cyg x1 and xte j1550564 ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . for some bhxbs ,
multiple spin measurements using the same method have been broadly consistent with each other , whereas for other bhxbs they have not .
two examples are the various spin estimates based on the disc continuum fitting method for gro j165540 ; @xmath25 @xcite , @xmath26 @xcite , @xmath27 @xcite , @xmath28 @xcite , and the reflection fitting method for cyg x1 ; @xmath29 @xcite , @xmath30 @xcite , @xmath31 @xcite . in almost all cases
the most recent measurement is likely to be the most accurate one , but we caution that this may not always be the case , and it depends on the many subtleties of the techniques and assumptions adopted in each individual case , which we can not go into here . generally speaking , for most bhxbs the error bars of most , if not all measurements overlap when one spin measurement method is considered , but in a few cases there are large differences ( like that of cyg x1 mentioned above ;
see also the discussion in fgr10 ) . since the peak flux of the bright radio flare of a bhxb can differ dramatically for different outbursts of the same source , we tabulate the brightest radio flare observed during state transition for each outburst of each source .
in addition to the sources considered here , there are several other notable cases of xbs with a large range in peak radio flare fluxes ( e.g. cyg x3 , cir x1 , ss 433 ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . for one source ,
4u 154347 , a radio flare seen during the state transition during its 2002 outburst was one of the few in the literature where the rise of the flare was reported and well sampled in time @xcite , rising at 0.81.0 ghz from @xmath32 mjy to @xmath33 mjy in 6.0 h , then peaking at @xmath34 mjy just 2.0 h later .
less than a day ( 19.0 h ) after this brightest detection , the source had faded to @xmath35 mjy at 0.8 ghz .
the light curve morphology is typical of a flare event often seen over the state transition ( see e.g. * ? ? ?
radio flares typically rise faster than they fade , so it is unlikely that the peak of this radio flare was much brighter than the observed peak of 22 mjy ( nm12 take this flux density as a lower limit ) . in addition , the rise times of all radio flares are consistent with being around 5 12 h ( table 1 ; see also table 1 of * ? ? ?
as an extremely conservative estimate , the peak could have been twice as bright as that observed , so we estimate a peak radio flux density of 2244 mjy at 0.81.0 ghz .
we use the method of nm12 of assuming @xmath36 for the spectral index of the optically thin radio flare in order to estimate the 5 ghz flux , which was 11.623.4 mjy .
gro j165540 had a very bright radio flare of 2 jy in 1994 which was associated with relativistic ejecta @xcite , whereas in 2005 , an optically thin radio flare was detected at @xmath37 mjy , @xmath38 mjy and @xmath39 mjy at 1.43 ghz , 4.86 ghz and 8.46 ghz , respectively , which then decayed over the next few days @xcite .
the radio light curve is well sampled , with daily coverage ( see also * ? ? ?
* ; * ? ? ?
* and references therein ) and the rise and decay of the flare were observed .
this was the brightest flare seen from gro j165540 during the hard to soft state transition of this outburst , and occurred within days of the source entering the soft state @xcite , which is fairly typical @xcite .
from these two outbursts with well sampled radio monitoring , it is clear that the peak luminosity of the flare associated with the state transition can vary by three orders of magnitude .
the fainter flare in 2005 had a longer rise time , and the resulting total energy required to produce the 1994 and 2005 radio flares may only differ by only a factor @xmath40 ( see section 3 ) .
+ four bright 5 ghz radio flares of grs 1915 + 105 are known to have peaked at flux densities between 150 and 912 mjy ( table 1 ) .
nm12 already noted that the bright radio flare of grs 1915 + 105 seen in 1997 peaked at a luminosity @xmath41 times fainter than the bright radio flare in 1994 , which invalidates a single relation between bh spin and peak radio flare luminosity because the bh spin can not change in three years ( nm12 only included the brightest of these two flares of grs 1915 + 105 in their following analysis of the bh spin jet power relation ) .
@xcite found from high resolution radio imaging of grs 1915 + 105 that even fainter radio flares are resolved into relativistic discrete ejecta .
we therefore include all four bright radio flares in table 1 and the following analysis .
these bright radio flares are not to be confused with the fainter , quasi - regular ` oscillation ' flares which peak at @xmath42 mjy ( e.g. * ? ? ?
* ) , and which are not included here ( but which could be due to the same physical process ) .
h1743322 also had peak radio flare flux densities in 2003 and 2009 that differed by a factor of @xmath40 @xcite .
we also include cygnus x1 in our analysis .
nm12 discard this source because , as they argue , the mass transfer is via a stellar wind from its high - mass companion star .
although cyg x1 is a wind - fed system , an accretion disc has been detected around the bh , its size has been estimated @xcite , and the disc has even been used to infer the bh spin ( e.g. * ? ? ?
* ; * ? ? ?
* ) . on large scales ,
the mode of accretion for wind - fed systems differs to that of roche lobe overflow systems the stellar wind ensures a relatively steady , high mass accretion rate onto the outer accretion disc .
the influence of bh spin on jet production is only important on distances of @xmath43 from the bh in which general relativistic frame dragging becomes important ( e.g. * ? ? ?
* ; * ? ? ?
obviously the mode of accretion towards the bh on these much smaller size scales is the same for wind - fed systems as ( transient ) roche lobe accreting systems ( i.e. , matter is accreted from the inner disc into the inner regions that are affected by frame dragging ) , and is not affected by the mode of accretion on large scales ( accretion disc or wind - fed accretion ) .
in addition , it is well known that cyg x1 performs state transitions that are associated with radio flares ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ?
* ) like other bhxbs . the brightest radio detection of cyg x1 ever seen over years of detailed , long - term monitoring with the ryle and green bank telescopes ( e.g. * ? ?
* ; * ? ? ?
* ) was during the hard to soft state transition in 2004 , where the radio luminosity peaked at 120 mjy ( * ? ? ? * in addition , a likely extended , discrete ejection has been seen ) .
in fact cyg x1 spends 10 per cent of its time in the soft state ( e.g. * ? ? ?
* ) in which the radio emission is suppressed ( e.g. * ? ? ?
* ; * ? ? ?
* ) , as expected after crossing the ` jet line ' when the ballistic ejection is usually launched @xcite .
cyg x1 should therefore certainly be included ; here we use the aforementioned brightest radio flare of 120 mjy at 15 ghz . again adopting @xmath36 ( nm12 ) we estimate a 5 ghz flux density of 186 mjy .
the two methods we adopt to estimate the relative jet power of the discrete ejections are explained in fgr10 and nm12 , respectively . in fgr10
we adopted the classical calculation of the minimum total energy required to energise a synchrotron flare , @xmath44 where @xmath45 is the peak radio luminosity @xcite .
this calculation is frequently used in the study of synchrotron radiation from numerous sources at wavelengths from radio to x - ray , and is a reliable measure the minimum total luminosity of an individual flare .
we calculate the jet power of each radio flare listed in table 1 as a fraction of the eddington luminosity ( using the known values of distance and bh mass ) .
the second method of measuring jet power is simply using the luminosity at the peak of each radio flare ( nm12 ) , which assumes the peak luminosity is a linear indicator of total jet kinetic power ( contrary to the generally accepted theoretical works ; e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
this method contains fewer uncertainties , but does not take into account the different temporal profiles and hence different total energies of each radio flare .
note that for both methods , the jet power as a fraction of the x - ray luminosity is not considered , because quasi - simultaneous x - ray luminosities were not available for all radio jet flares . in fig .
1 we plot bh spin versus relative jet power for all data listed in table 1 .
we assume an error of a factor of two for the jet power estimates .
for those sources for which different measurements / limits of the jet power have been reported over different epochs , the highest values correspond to filled black markers , while grey markers correspond to lower values / limits .
it is clear that there is a large scatter in the values of jet power for a single source , because transient jets had different peak fluxes and total energies during different outbursts .
bh spins of @xmath46 measured from disc continuum fitting have a range of jet flare powers spanning @xmath47 orders of magnitude ( left panels of fig .
1 ) . with the inclusion of all the available data ,
the claimed @xmath48 relation of nm12 is no longer as obvious ( fig .
particular sources to note are 4u 154347 , cyg x1 and gro j165540 ( 2005 outburst ) , which all had radio flares orders of magnitude fainter than expected from the relation , whereas the radio flare of grs 112468 was two orders of magnitude too bright for its low bh spin ( although we caution the reader that this is a bh spin measurement inferred from old data , which as yet has not been superseded ) .
spins of @xmath49 measured via reflection have jet powers ranging @xmath50 and @xmath41 orders of magnitude for jet powers estimated from their total energies and peak luminosities , respectively .
there is no clear single relation between jet power and bh spin in any of the four panels in fig . 1 . .
the bottom panels repeat the same exercise as in the top panels , but for an arbitrary sub - sample of data points .
the labels in the right column denote the inferred range of values for the slope of the correlation , while the percentages refer to positive slopes .
percentages lower than @xmath51 correspond to no statistically significant positive correlation ( since more than @xmath52 per cent of the draws result in a _ negative _ slope ) .
[ fig : bayes1],width=325 ] . in order to quantify our visual assessment
, we investigate the presence of a positive correlation between jet power and bh spin using a bayesian code @xcite .
we carry out a linear regression analysis on a relation of the form @xmath53 , with intrinsic random scatter included , and for the four possible permutations between the two different methods for estimating the transient jet power and and spin parameter , respectively .
three gaussians ( for the slope , intercept and intrinsic scatter ) are used in the independent variable mixture modelling , and a minimum of 10,000 iterations are performed .
the most likely parameter values are estimated as the median of 10,000 draws from the posterior distribution . for those sources for which
different values of the radio flare luminosity have been reported over different epochs , we adopt _ the highest _ value as the tracer for jet power ( corresponding to the filled black marks in fig .
1 ) .
the top panels of fig .
[ fig : bayes1 ] illustrate the results of the linear regression analysis for the case where a correlation has been claimed , i.e. for the jet power as inferred from the peak luminosity of the radio flare versus the spin parameter as estimated from the thermal continuum fitting method ( nm12 ) , here denoted as @xmath54 .
the histogram on the top right shows the distribution of the fitted slopes , with the draws returning a negative slope highlighted in black . for this specific case ,
only 60 per cent of the draws have slopes consistent with a positive correlation between the spin parameter and and jet power being present , _ ruling out a positive correlation with high confidence _ ( @xmath55 ) .
this is perhaps most apparent in the top left panel , where a random sub - sample of 150 fits ( out of 10,000 ) is over - plotted to the data . for comparison ,
the middle panels correspond to a similar analysis as the one discussed above , only for a _ simulated _ data set where the spin parameter values ( along with errors ) are identical to those in the top panels , while the jet power values are randomly drawn from a @xmath48 distribution with an intrinsic scatter of 0.5 dex ( additionally , each jet power value is assigned an error of 0.3 dex ) . in this case , only 0.4 per cent of the draws return negative slopes , which indicates a statistically significant ( @xmath56 ) positive correlation . by comparison with the top panels , this exercise serves to illustrate that the lack of a statistically significant correlation in the full data set is not merely a result of the small number statistics , and the code could indeed recover a statistically significant correlation with eight data points only .
lastly , the bottom panels illustrate how the results of the regression analysis change for five out of the eight data points considered in the top panels ( specifically with grs1124@xmath5768 , cyg x@xmath571 and 4u 1543@xmath5747 excluded from the analysis ; as in nm12 ) .
the sample is actually the same as that in @xcite ; this is discussed further in section 4 .
while the slope distribution is formally consistent with a value of two , the significant tail of negative draws ( 11 per cent ) corresponds to a marginally significant correlation only ( @xmath58 ) , although it is worth stressing that the statistical significance of the correlation for this data set is very sensitive to the error bars ( the bayesian code we are adopting can only treat with symmetric error bars . the same analysis as described above was carried out to assess the presence of a relation between the spin parameter @xmath59 and the jet power @xmath60 as measured from the different methods , namely spin from reflection ( @xmath61 ) and jet power from total energy ( denoted as @xmath62 , vs. @xmath63 ) .
plots analogous to those discussed in fig .
[ fig : bayes1 ] are shown in fig .
[ fig : bayes2 ] for the following combinations of parameters : jet power from peak radio flare luminosity vs. spin parameter from thermal continuum ( top left ) ; jet power from peak radio flare luminosity vs. spin parameter from reflection ( bottom right ) ; jet power from flare total energy vs. spin parameter from thermal continuum ( top left ) ; jet power from flare total energy vs. spin parameter from reflection ( top right ) . as apparent from the prominent negative tails of the fitted slope distributions ,
no positive , statistically significant correlation is found in all cases .
a similar conclusion is reached by fitting either @xmath60 as a function of @xmath59 ( solid lines ) , or @xmath59 as a function of @xmath60 ( dashed lines ) .
we have shown that no single statistically significant relation exists between inferred transient jet power and bh spin parameter , with the current estimates of these parameters .
furthermore , the very fact that one bhxb can produce radio ejections with very different total energies and peak luminosities invalidates a single relation between transient jet radiative energy and bh spin .
this latter discovery implies that many jet flares may be less luminous than the maximum value that would be produced by a purely spin - powered jet . for a typical source ,
there have been a limited number of outbursts observed to date , so it could be that truly spin - powered jets do reach a constant , maximum jet power , but few flares reach this critical maximum .
this could lead to underestimated jet luminosity estimates .
we note that one source , grs 112468 appears to have a high jet power but a low bh spin , which is not consistent with the above picture .
regular radio monitoring of populations of bhxbs over state transitions is required to measure the distribution of jet powers of discrete ejections .
it is worth noting that in our study we do not include the effects of relativistic beaming of the jet emission .
this is because both the orbital inclination angles ( which are usually assumed to be perpendicular to the jet axis ) and the jet lorentz factors are uncertain for many sources ( nm12 assumed a lorentz factor of @xmath64 for all sources ) .
very recently , a new paper has appeared from the same group as nm12 , which again claims a correlation between bh spin ( measured from continuum fitting ) and jet power ( measured from peak luminosities ) in transient ejections @xcite . in this work ,
the same subsample of bhxb spin measurements that appeared in nm12 was used , with an addition of one more spin measurement , that of h1743322 from @xcite .
this spin measurement is also included in our analysis .
however , it is worth noting a few discrepancies between our analysis and theirs . for this source we took jet flares from two outbursts , 2003 and 2009 , whereas in @xcite just the former jet flare is used
furthermore , we adopted the radio flare peak luminosity associated with the hard to soft transition in fig .
1c ( the same method as was adopted in nm12 ) but @xcite use a fainter flare , which peaked at 35 mjy at 5 ghz instead of the brighter , 96 mjy flare we used .
their arguement is that this fainter flare must have been the one associated with the launch of the jet ejection because the date on which this ejection occurred is known from resolved radio observations .
however , if the bh spin powers the 35 mjy flare but not the 96 mjy flare , then what powers the brighter flare ?
the 96 mjy flare had an optically thin spectrum and did not occur during the hard state @xcite , so we can rule out a compact jet origin to this brighter flare .
if one of those two flares was associated with an energetic boost from the spin of the bh , one would expect it to be the brighter one .
furthermore , since we showed that radio ejections observationally possess very different peak luminosities , the peak flux can not be used as a proxy for the total power of a discrete ejection .
@xcite do not take into account the time over which a flare is radiated ( which empirically can vary by more than one order of magnitude ; see table 1 ) , but instead adopt @xmath65 ( see equation b10 in the appendix of * ? ? ?
* ) , contrary to standard classical synchrotron equation , @xmath66 .
finally , a further discrepancy lies in the assumed relation between bh spin and jet power . in fitting the correlation ,
@xcite assume @xmath67 , as opposed to that adopted in nm12 , @xmath48 , assume the same jet lorentz factor for all sources , and fix the power law index to 2.0 ( as indicated in footnote 6 in their paper ) .
for comparison , the bayesian , linear regression analysis we present here leaves the slope as a free parameter , testing the probability that it is positive at all . as we discussed in detail in section 3 ,
a full bayesian treatment , with error bars on both jet power and spin parameters included , shows that a positive correlation of the form @xmath48 is at best marginally significant ( @xmath58 ) , even when the new h1743322 data point is included .
if any of the above methods of measuring bh spin or jet power are correct , then we can conclude that there is no single relation between bh spin and transient jet power .
are there any reasonable explanations that could explain the lack of a correlation ?
many works predict ( analytically or via simulations ) a relation between the spin of a black hole and the power of the resulting accretion - fed jet ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
positive relations are generally predicted but not necessarily as simple as the @xmath48 relation originally derived by @xcite .
it has recently been shown that rapidly _ retrograde _ spinning bhs ( where the bh spin is in the opposite direction to the accreting matter spiralling inwards ) can produce jets just as powerful as rapidly prograde spinning bhs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
interestingly , the only source with a ( slightly ) negative spin in table 1 , grs 112468 , had a relatively powerful jet ( fig . 1 ; lower left panel )
however , since several bhs with very similar spin estimates have produced flares of energies and peak luminosities spanning several orders of magnitude ( including different jet powers for different flares from the same source ) we can rule out a single relation between jet power of discrete ejections and current estimates of bh spin .
this is quantified by our bayesian analysis ( figs . 2 and 3 ) .
the spin parameters as estimated from the two methods do not agree with each other for a number of sources , and neither is to be preferred .
in addition , for compact , steady jets that exist in the hard state , radio or infrared luminosity normalization may turn out to be an unreliable tracer for jet power .
dual tracks exist with sources jumping between them and having different correlation slopes @xcite , which implies that the jet luminosity is affected by additional , as yet undefined parameters as well as the mass accretion rate and ( possibly ) the bh spin .
the radio / x - ray ( and ir / x - ray ) correlation normalization therefore may not be an accurate method of estimating the relative ranking of compact jet power , however the range in jet luminosities in the hard state ( for a given x - ray luminosity ) could still be improved . nevertheless , the results of fgr10 suggest there may not be a strong dependence of relative compact jet power on bh spin , even given the above caveats .
it was also predicted by @xcite that jet power also scales with the square of the magnetic field strength ; @xmath68 , because the mechanism relies on differential rotation between a spinning bh and its magnetosphere .
one would therefore not expect a @xmath48 correlation to exist unless the magnetic field is nearly constant across all sources ( this caveat is also mentioned , but then later ignored , in * ? ? ?
in the hard state , it may be that @xmath69 , the mass accretion rate ( e.g. * ? ? ?
* ) , and in @xcite it is argued that jet power is mostly driven by the magnetic flux , not the spin or accretion rate .
jet luminosity ( measured in radio or ir ) does scale positively with x - ray luminosity and hence with mass accretion rate , in the hard state .
however this is no indication of the blandford & znajek mechanism , as a positive relation between x - ray luminosity and jet power is of course also expected for the blandford
payne mechanism ( in which accretion powers the jet ) .
in addition , @xcite have recently argued that jets that precess on short timescales can not be spin powered . even if bhxb jets were powered by bh spin , the combination of several effects including but not limited to uncertainties in measuring the relevant parameters could still prevent us from finding a positive correlation at present .
dmr would like to thank jon miller and sera markoff for insightful discussions .
eg is grateful to brendan p. miller for his assistance with the bayesian analysis , and to monica colpi and laura maraschi for their insights .
dmr acknowledges support from a marie curie intra european fellowship within the 7th european community framework programme under contract no .
ief 274805 . 99 ball l. , kesteven m. j. , campbell - wilson d. , turtle a. j. , hjellming r. m. , 1995 , mnras , 273 , 722 bambi c. , 2012 , physical review d , 86 , 123013 barret d. , mcclintock j. e. , grindlay j. e. , 1996 , apj , 473 , 963 belloni t. m. , 2010 , in ` the jet paradigm - from microquasars to quasars ' , ed
. t. belloni , lect .
notes phys .
794 blandford r. d. , konigl a. , 1979 , apj , 232 , 34 blandford r. d. , payne d. g. , 1982 , mnras , 199 , 883 blandford r. d. , znajek r. l. , 1977 , mnras , 179 , 433 blum j. l. , miller j. m. , fabian a. c. , miller m. c. , homan j. , van der klis m. , cackett e. m. , reis r. c. , 2009 , apj , 706 , 60 blundell k. m. , schmidtobreick l. , trushkin s. , 2011 , mnras , 417 , 2401 bonsignori - facondi s. r. , padrielli l. , montebugnoli s. , barbieri r. , 1986 , a&a , 166 , 157 brocksopp c. , fender r. p. , larionov v. , lyuty v. m. , tarasov a. e. , pooley g. g. , paciesas w. s. , roche p. , 1999 ,
mnras , 309 , 1063 burbidge g. r. , 1959 , apj , 129 , 849 calvelo d. e. , tzioumis t. , corbel s. , brocksopp c. , fender r. p. , 2009
telegr . , 2135 cantrell a. g. , et al . , 2010 , apj , 710 , 1127 chaty s. , dubus g. , raichoor a. , 2011 , a&a , 529 , 3 corbel s. , fender r. p. , 2002
, apj , 573 , l35 corbel s. , et al . , 2012 , mnras , 421 , 2947 coriat m. , et al . , 2011 , mnras , 414 , 677 coriat m. , fender r. p. , dubus g. , 2012 , mnras , 424 , 1991 davis s. w. , hubeny i. , 2006 , apjs , 164 , 530 de villiers j .- p . , hawley j. f. , krolik j. h. , hirose s. , 2005 , apj , 620 , 878 dhawan v. , mirabel i. f. , rodrguez l. f. , 2000 , apj , 543 , 373 doeleman s. s. , et al .
2012 , science , 338 , 355 duro r. , et al . , 2011 ,
a&a , 533 , l3 fabian a. c. , et al . , 2012 , mnras , 424 , 217 falcke h. , biermann p. l. , 1995 , a&a , 293 , 665 fender r. p. , 2006
, in _ compact stellar x - ray sources _ ; cambridge university press : cambridge , u.k .
, pp 381 ( eds .
lewin , w. h. g. ; van der klis , m. ) fender r. p. , garrington s. t. , mckay d. j. , muxlow t. w. b. , pooley g. g. , spencer r. e. , stirling a. m. , waltman e. b. , 1999 , mnras , 304 , 865 fender r. p. , stirling a. m. , spencer r. e. , brown i. , pooley g. g. , muxlow t. w. b. , miller - jones j. c. a. , 2006 , mnras , 369 , 603 fender , r. p. , belloni , t. m. , gallo , e. 2004 , mnras , 355 , 1105 fender r. p. , homan j. , belloni t. m. , 2009 , mnras , 396 , 1370 fender r. p. , gallo e , russell d. m. , 2010 , mnras , 406 , 1425 ( fgr10 ) ferreira j. , petrucci p .- o .
, henri g. , saug l. , pelletier g. , 2006 , a&a , 447 , 813 gallo e. , corbel s. , fender r. p. , maccarone t. j. , tzioumis a. k. , 2004 , mnras , 347 , l52 gallo e. , fender r. p. , kaiser c. , russell , d. m. , morganti r. , oosterloo t. , heinz s. , 2005 , nat , 436 , 819 gallo e. , miller b. p. , fender r. , 2012 , mnras , 423 , 590 gandhi p. , et al . , 2011 , apj , 740 ,
l13 garofalo d. , 2009 , apj , 699 , 400 gelino d. m. , harrison t. e. , mcnamara b. j. , 2001 , aj , 122 , 971 gierliski m. , macioek - niedawiecki a. , ebisawa k. , 2001 , mnras , 325 , 1253 gou l. , mcclintock j. e. , steiner j. f. , narayan r. , cantrell a g. , bailyn c. d. , orosz j. a. , 2010 , apj , 718 , l122 gou l. , et al . , 2011 , apj , 742 , 85 greene , j. , bailyn , c. d. , orosz , j. a. 2001 , apj , 554 , 1290 greiner j. , cuby j. g. , mccaughrean m. j. , 2001 , nat , 414 , 522 hannikainen d. c. , 2009 , mnras , 397 , 569 hawley j. f. , krolik j. h. , 2006 , apj , 641 , 103 heinz s. , sunyaev r. a. , 2003 , mnras , 343 , l59 hiemstra b. , mndez m. , done c. , daz trigo m. , altamirano d. , casella p. , 2011
, mnras , 411 , 137 hjellming r. m. , rupen m. p. , 1995 ,
nat , 375 , 464 hjellming r. m. , calovini t. a. , han x. h. , cordova f. a. , 1988 , apj , 335 , l75 hynes r. i. , steeghs d. , casares j. , charles p. a. , obrien k. , 2004 , apj , 609 , 317 junor w. , biretta j. a. , livio m. , 1999 , nat , 401 , 891 kelly b. c. , 2007 , apj , 665 , 1489 koljonen k. i. i. , hannikainen d. c. , mccollough m. l. , pooley g. g. , trushkin s. a. , 2010 , mnras , 406 , 307 kolehmainen m. , done c. , 2010 , mnras , 406 , 2206 kuulkers e. , fender r. p. , spencer r. e. , davis r. j. , morison i. , 1999 , mnras , 306 , 919 lachowicz p. , zdziarski a. a. , schwarzenberg - czerny a. , pooley g. g. , kitamoto s. , 2006 , mnras , 368 , 1025 laurent p. , rodriguez j. , wilms j. , cadolle bel m. , pottschmidt k. , grinberg v. , 2011 , science , 332 , 438 livio m. , ogilvie g. i. , pringle j. e. , 1999 , apj , 512 , 100 longair m. s. , 1994 , high energy astrophysics .
cambridge univ . press , cambridge macdonald d. , thorne k. s. , 1982 , mnras , 198 , 345 markoff s. , falcke h. , fender r. , 2001 , a&a , 372 , l25 markoff , s. , nowak , m. , corbel , s. , fender , r. , falcke , h. 2003 , a&a , 397 , 645 markoff s. , nowak m. a. , wilms j. , 2005 , apj , 635 , 1203 mcclintock j. e. , remillard r. a. , 2006 , in compact stellar x - ray sources , eds .
lewin w. h. g. , van der klis m. , cambridge university press , p. 157
mcclintock j. e. , shafee r. , narayan r. , remillard r. a. , davis s. w. , li l .- x . , 2006 , apj , 652 , 518 mcclintock j. e. , remillard r. a. , rupen m. p. , torres m. a. p. , steeghs d. , levine a. m. , orosz j. a. , 2009 , apj , 698 , 1398 mcclintock j. e. , et al . , 2011 ,
class . & quantum grav .
, 28 , 114009 mckinney j. c. , 2005 , apj , 630 , l5 mckinney j. c. , tchekhovskoy a. , blandford r. d. , 2012 , mnras , 423 , 3083 meier d. l. , 1999 , apj , 518 , 788 meier d. l. , 2011 , in romero g. e. , sunyaev r. a. , belloni t. , eds .
275 , jets at all scales .
cambridge univ . press ,
cambridge , p. 13 meier d. l. , koide s. , uchida y. , 2001 , science , 291 , 84 migliari s. , miller - jones j. c. a. , russell d. m. , 2011 , mnras , 415 , 2407 miller j. m. , 2007 , ara&a , 45 , 441 miller j. m. , reynolds c. s. , fabian a. c. , miniutti g. , gallo l. c. , 2009 , apj , 697 , 900 miller - jones j. c. a. , fender r. p. , nakar e. , 2006 , mnras , 367 , 1432 miller - jones j. c. a. , et al . , 2012 , mnras , 421 , 468 mioduszewski a. j. , rupen m. p. , hjellming r. m. , pooley g. g. , waltman e. b. , 2001 , apj , 553 , 766 mirabel i. f. , rodrguez l. f. , 1994 , nat , 371 , 46 narayan r. , mcclintock j. e. , 2012 , mnras , 419 , l69 ( nm12 ) nixon c. , king a. , 2013 , apjl , in press ( arxiv:1301.5321 ) orosz j. a. , polisensky e. j. , bailyn c. d. , tourtellotte s. w. , mcclintock j. e. , remillard r. a. , 2002 , baas , 34 , 1124 orosz j. a. , mcclintock j. e. , aufdenberg j. p. , remillard r. a. , reid m. j. , narayan r. , gou l. , 2011a , apj , 742 , 84 orosz j. a. , steiner j. f. , mcclintock j. e. , torres m. a. p. , remillard r. a. , bailyn c. d. , miller j. m. , 2011b , apj , 730 , 75 pakull m. w. , soria r. , motch c. , 2010 , nat , 466 , 209 park s. q. , et al . , 2004 , apj , 610 , 378 peer a. , markoff s. , 2012 , apj , 753 , 177 polko p. , meier d. l. , markoff s. , 2010 , apj , 723 , 1343 preston r. a. , morabito d. d. , wehrle a. e. , jauncey d. l. , batty m. j. , haynes r. f. , wright a. e. , nicolson g. d. , 1983 , apj , 268 , l23 rahoui f. , et al . , 2012 , mnras , 422 , 2202 reid m. j. , mcclintock j. e. , narayan r. , gou l. , remillard r. a. , orosz j. a. , 2011 , apj , 742 , 83 reis r. c. , fabian a. c. , miller j. m. , 2010 , mnras , 402 , 836 reis r. c. , et al .
, 2011 , mnras , 410 , 2497 remillard r. a. , mcclintock j. e. , 2006 , ara&a , 44 , 49 rodriguez l. f. , gerard e. , mirabel i. f. , gomez y. , velazquez a. , 1995 , apjs , 101 , 173 rupen m. p. , dhawan v. , mioduszewski a. j. , 2005 , astron .
telegr . , 443 rushton a. , spencer r. e. , pooley g. , trushkin s. , 2010 , mnras , 401 , 2611 rushton a. , et al . , 2012 , mnras , 419 , 3194 russell d. m. , fender r. p. , gallo e. , kaiser c. r. , 2007a , mnras , 376 , 1341 russell d. m. , maccarone t. j. , krding e. g. , homan j. , 2007b , mnras , 379 , 1401 russell d. m. , maitra d. , dunn r. j. h. , markoff s. , 2010 , mnras , 405 , 1759 russell d. m. , et al .
, 2013 , mnras , 429 , 815 shafee r. , mcclintock j. e. , narayan r. , davis s. w. , li l .- x . , remillard r. a. , 2006 , apj , 636 , l113 shaposhnikov n , swank j. , shrader c. r. , rupen m. , beckmann v. , markwardt c. b. , smith d. a. , 2007 , apj , 655 , 434 shaposhnikov n. , titarchuk l. , 2009 , apj , 699 , 453 shaposhnikov n. , markwardt c. , swank j. , krimm h. , 2010 , apj , 723 , 1817 sikora , m. , begelman m. c. , 2013 , apjl , in press ( arxiv:1301.5638 ) spencer r. e. , swinney r. w. , johnston k. j. , hjellming r. m. , 1986 , apj , 309 , 694 steiner j. f. , et al .
, 2011 , mnras , 416 , 941 steiner j. f. , mcclintock j. e. , reid m. j. , 2012 , apj , 745 , l7 steiner j. f. , mcclintock j. e. , narayan r. , 2013 , apj , 762 , 104 stirling a. m. , spencer r. e. , de la force c. j. , garrett m. a. , fender r. p. , ogley r. n. , 2001 , mnras , 327 , 1273 tchekhovskoy a. , narayan r. , mckinney j. c. , 2010 , apj , 711 , 50 tchekhovskoy a. , mckinney j. c. , 2012 , mnras , 423 , l55 tigelaar s. p. , fender r. p. , tilanus r. p. j. , gallo e. , pooley g. g. , 2004 , mnras , 352 , 1015 tingay s. j. , et al .
, 1995 , nature , 374 , 141 trushkin s. a. , nizhelskij n. a. , 2010 , astron .
, 2899 tudose v. , fender r. p. , tzioumis a. k. , spencer r. e. , van der klis m. , 2008 , mnras , 390 , 447 williams p. k. g. , et al . , 2011 , apj , 733 , l20 wilms j. , pottschmidt k. , pooley g. g. , markoff s. , nowak m. a. , kreykenbohm i. , rothschild r. e. , 2007 , apj , 663 , l97 yang j. , brocksopp c. , corbel s. , paragi z. , tzioumis t. , fender r. p. , 2010 ,
mnras , 409 , l64 zhang s. n. , cui w. , chen w. , 1997 , apj , 482 , l155 | we revisit the paradigm of the dependency of jet power on black hole spin in accreting black hole systems . in a previous paper we showed that the luminosity of compact jets continuously launched due to accretion onto black holes in x - ray binaries ( analogous to those that dominate the kinetic feedback from agn ) do not appear to correlate with reported black hole spin measurements .
it is therefore unclear whether extraction of the black hole spin energy is the main driver powering compact jets from accreting black holes .
occasionally , black hole x - ray binaries produce discrete , transient ( ballistic ) jets for a brief time over accretion state changes . here
, we quantify the dependence of the power of these transient jets ( adopting two methods to infer the jet power ) on black hole spin , making use of all the available data in the current literature , which includes 12 bhs with both measured spin parameters and radio flares over the state transition . in several sources ,
regular , well - sampled radio monitoring has shown that the peak radio flux differs dramatically depending on the outburst ( up to a factor of 1000 ) whereas the total power required to energise the flare may only differ by a factor @xmath0 between outbursts .
the peak flux is determined by the total energy in the flare and the time over which it is radiated ( which can vary considerably between outbursts ) . using a bayesian fitting routine we rule out a statistically significant positive correlation between transient jet power measured using these methods , and current estimates of black hole spin . even when selecting subsamples of the data that disregard some methods of black hole spin measurement or jet power measurement , no correlation is found in all cases .
[ firstpage ] accretion , accretion discs , black hole physics , x - rays : binaries , ism : jets and outflows |
the existence of a large momentum or mass scale is the basis of any perturbative treatment of hadronic processes . in photoproduction of heavy quarkonia
there are two obvious scales : on the one hand the heavy quark mass and , on the other hand , the ( transverse ) momentum transfer .
if one considers @xmath0 production the charm - quark mass is still too small to really serve as a hard scale .
it is therefore likely that significant contributions come from non - perturbative regions in which hadronic fluctuations of the photon have a large transverse size .
this statement is supported by the success of pomeron phenomenology in describing diffractive photoproduction of @xmath0s @xcite .
taking , on the other hand , the momentum transfer as a large scale one can think of several perturbative mechanisms . a potential mechanism at intermediate momentum transfers
is depicted in fig .
[ overlap ] . only one protonic constituent , a gluon , participates in the hard - scattering subprocess .
a generalized gluon distribution describes the emission and reabsorption of the gluon .
such a kind of mechanism has been considered by huang and kroll for the photoproduction of light mesons @xcite .
they , however , found a strong suppression of this production mechanism and concluded that vector - meson dominance is still at work at moderately large momentum transfers where data exist . within the diquark model.,width=294 ] within the diquark model.,width=294 ] for asymptotically large momentum transfers it is generally assumed that the hard - scattering mechanism dominates @xcite .
the hard - scattering mechanism involves only valence fock states of the hadrons with the ( leading - order ) perturbative part being a coherent sum of tree graphs specific to the reaction under consideration .
the non - perturbative ingredients of the hard - scattering mechanism are hadron distribution amplitudes , i.e. probability amplitudes for finding the hadronic constituents carrying certain fractions of their parent hadron s momentum .
this is the reaction mechanism we concentrate on in our investigation of @xmath0 photoproduction . as the graph in fig .
[ hsm ] shows , we assume in addition that the valence fock state of the proton is a quark - diquark rather than a three - quark state .
this does not only simplify calculations , but diquarks also serve to model non - perturbative effects so that the hard - scattering mechanism may become applicable at reasonably large momentum transfers where data are available or can be expected .
we take into account scalar and vector diquarks and treat them as extended particles with form factors at the gauge - boson diquark vertices . for the present analysis
the form - factor parameters and also the quark - diquark distribution amplitude of the proton are adopted from previous work on baryon form factors , compton scattering and other reactions ( see , e.g. , ref .
the distribution amplitude for the @xmath0 , i.e. @xmath2 $ ] , is taken from ref .
its normalization is fixed by the experimentally known @xmath0 decay constant .
these are almost all the ingredients necessary for calculating the @xmath0-photoproduction amplitude .
at this point it is necessary to mention that constituent masses are usually neglected in the calculation of the perturbative part of the amplitude as long as they do not fix the energy or momentum - transfer scale ( like in quarkonium decays ) .
if we want to make predictions for momentum transfers which may become experimentally accessible in the near future , i.e. values not much larger than the charm - quark mass @xmath3 , we have to think about taking into account @xmath3 in the perturbative amplitude . following a common practice
we assume that the 4-momentum of every hadronic constituent is proportional to the 4-momentum of its parent hadron .
this implies that every constituent of hadron @xmath4 has an effective mass @xmath5 , with @xmath6 being the fraction of the hadron s 4-momentum carried by the constituent . keeping in mind that the perturbative scattering amplitude for the constituents is convoluted with the hadron distribution amplitudes this implies , e.g. , that @xmath7 for our @xmath0 distribution amplitude , which is strongly peaked at @xmath8 .
this prescription for constituent masses applies to ( on - shell ) particles which show up as external legs in the feynman diagrams . to obtain the constituent mass for an internal line
we express the corresponding momentum as a sum of hadron momenta multiplied with the appropriate momentum fractions and replace the hadron momenta by their respective mass .
if the resulting constituent mass contains a term @xmath9 this is neglected since it turns out that such terms violate @xmath10 gauge invariance . for our @xmath0 distribution amplitude the masses in
the the charm - quark propagators become in this way again @xmath7 . the feynman diagrams
are then calculated with these constituent masses and expressed in terms of the mandelstam variables @xmath11 , @xmath12 , and @xmath13 ( @xmath14 ) .
keeping the scattering angle fixed , we finally expand the perturbative amplitude in terms of @xmath15 and @xmath16 and neglect proton - mass corrections of order @xmath17 , @xmath18 , or higher .
this kind of mass treatment has the nice feature that it preserves @xmath10 gauge invariance with respect to the photon and @xmath19 gauge invariance with respect to the gluon .
a similar kind of mass treatment has already been suggested in ref .
@xcite for photoproduction of @xmath20 mesons . in that paper , however , massless mandelstam variables @xmath21 , @xmath22 , and @xmath23 have been used instead of @xmath11 , @xmath12 , and @xmath13 and the final mass expansion has not only been applied to the proton but also to the meson .
at @xmath24 gev . data are taken from ref .
@xcite.,width=287 ] figures [ jpsi60 ] and [ jpsi260 ] show the diquark - model predictions for @xmath25 $ ] for two hera energies along with recent zeus data .
these data are only available at low momentum transfers ( @xmath26 gev@xmath27 ) , i.e. outside the validity range of the perturbative diquark model , so that a direct comparison with our predictions is not possible . compared with a naive extrapolation of the data our predictions , however ,
seem to overshoot the @xmath0 photoproduction cross section in the few - gev momentum transfer region .
the finite constituent masses provide two effects which are both sizable , but which partly compensate each other .
for vanishing constituent masses only two of the twelve independent helicity amplitudes are different from zero , namely those for @xmath28 and @xmath29 ( dash - dotted line in fig .
[ jpsi60 ] ) . introducing constituent masses as described above and still taking only these two helicity amplitudes results in a considerable reduction of the cross section ( dashed line versus dash - dotted line in fig .
[ jpsi60 ] ) . with finite constituent masses , however ,
also the other helicity amplitudes become different from zero which increases the cross section again ( solid line in fig .
[ jpsi60 ] ) .
the overall effect of taking into account constituent masses is thus mainly a flattening of the cross section towards smaller momentum transfers .
a further flattening and reduction of the cross section can be expected if one takes into account that the minimum momentum transfer @xmath30 in exclusive @xmath0 photoproduction is @xmath31 gev@xmath27 . in order to come to definite conclusions about the validity of the diquark model for @xmath0 photoproduction the corresponding modifications for the gluon kinematics certainly have to be investigated . nevertheless , it is interesting to notice that , in contrast to other exclusive reactions in which the leading - order perturbative calculation often underestimates the data , it rather seems to come out too large for @xmath0 photoproduction .
this , at least , hints at the possibility that the perturbative production mechanism could provide a sizable contribution in exclusive @xmath0 photoproduction already at a few - gev of momentum transfer . it would therefore be interesting to have not only predictions from an effective model like ours , but also perturbative predictions on the more fundamental pure quark level . and
finally it would , of course , also be highly desirable to get experimental cross - section data in the few - gev momentum - transfer region to compare with .
9 a. donnachie and p.v .
landshoff , phys .
b 478 ( 2000 ) 146 . h. w. huang and p. kroll , eur .
j. c 17 ( 2000 ) 423 .
g. p. lepage and s. j. brodsky , phys .
d 22 ( 1980 ) 2157 .
r. jakob et al . , z. phys .
a 347 ( 1993 ) 109 .
t. feldmann and p. kroll , phys .
( 1997 ) 410 . c. f. berger and w. schweiger , phys
d 61 ( 2000 ) 114026 .
zeus collab . , s. chekanov et al . , preprint hep - ex/0201043 . | we report on an attempt to describe hard exclusive photoproduction of @xmath0 mesons , i.e. the reaction @xmath1 , by means of a modified version of the hard - scattering approach , in which the proton is treated as a quark - diquark rather than a three - quark system . in order to improve the applicability of the model at momentum transfers of only a few gev we take into account constituent - mass effects in the calculation of the perturbative scattering amplitude . with a standard @xmath0-meson distribution amplitude and diquark - model parameters adopted from preceding investigations of other photon - induced reactions our predictions for differential cross sections overestimate the naive extrapolation of the low - momentum transfer zeus data .
our results , however , reveal the importance of taking into account the charm - quark mass . |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.