article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
several papers have addressed , in the last few years , the problem of isotopy of symplectic surfaces representing the same homology class @xmath0 of a simply connected , symplectic @xmath1-manifold @xmath2 .
although a list of the results obtained in those papers is beyond the scope of this introduction , we just want to mention the existence of both uniqueness and non - uniqueness results , depending on the choice of the pair @xmath3 .
the prototypical example of a non - uniqueness result , further discussed in this paper , is the presence of infinitely many non - isotopic symplectic tori representing any multiple of the fiber class of ( some ) elliptic surfaces .
the idea behind this construction , developed by fintushel and stern in @xcite , is to obtain nonequivalent symplectic tori by braiding parallel copies of a symplectic torus .
all known examples of non - uniqueness , so far , have been genus @xmath4 representatives of a homology class of self - intersection @xmath5 . on the other hand ,
the few higher genus cases under control have always led to uniqueness results as in @xcite .
the interest in obtaining examples of non - isotopic symplectic surfaces of higher genus has been pointed out by many researchers , in particular , in @xcite , @xcite and @xcite .
there is no special reason to expect that non - isotopy phenomena should be restricted to genus @xmath4 , self - intersection @xmath5 surfaces .
in fact , if we also take into account @xmath1-manifolds with non - trivial fundamental groups , smith has shown in @xcite the existence of higher ( odd ) genus examples , and examples of symplectic curves with cusp singularities have been provided ( for the projective plane ) in @xcite .
however , all attempts to build higher genus examples in simply connected @xmath1-manifolds by suitably doubling " the braiding construction have failed .
the reason for the failure is the impossibility of detecting non - isotopy ( for potential examples ) either by using the seiberg - witten invariants or by using a more classical topological invariant like the fundamental group of the complement . in the latter approach ,
the reason is that most of the basic building blocks , genus @xmath4 surfaces built through the braiding construction , can not be distinguished by means of the @xmath6 either , as first observed in @xcite .
however , there is a second mechanism , different from the original braiding construction , that allows us to produce non - isotopic symplectic tori in a symplectic @xmath1-manifold ( see @xcite and @xcite ) .
this alternative construction originates from the presence of non - isotopic , nullhomologous lagrangian tori in a large class of ( simply connected ) symplectic @xmath1-manifolds ( see @xcite and @xcite ) .
these symplectic tori are obtained by summing " a preferred symplectic torus with non - isotopic nullhomologous tori .
the non - isotopy can be detected , as for the braiding construction , by using seiberg - witten theory , but the advantage of this second construction is that , in many cases , @xmath6 is sufficient to distinguish the tori , as shown in @xcite . and
if we double " the construction , the fundamental groups of the complements may retain enough information to distinguish the resulting higher genus surfaces .
in fact , following the ideas outlined above , we will be able to prove the following .
[ main ] for any fixed integer @xmath7 , there exist simply connected , symplectic @xmath1-manifolds @xmath8 containing infinitely many homologous , pairwise non - isotopic , symplectic surfaces @xmath9 of genus @xmath10 .
furthermore , there is no pair homeomorphism between @xmath11 and @xmath12 unless @xmath13 .
we briefly preview the idea of the construction .
first , consider the symplectic simply connected 4-manifold @xmath14 obtained by the knot surgery construction of @xcite from the elliptic @xmath15 surface @xmath16 and a non - trivial fibered knot @xmath17 in @xmath18 .
@xmath14 contains , for each @xmath7 , an infinite family of homologous , symplectic , non - isotopic tori @xmath19 , together with a symplectic surface @xmath20 of genus @xmath21 of self - intersection @xmath5 .
this surface @xmath20 intersects each of the tori @xmath22 at @xmath23 positive transverse points . by doubling @xmath14 along @xmath20
we obtain a symplectic , simply connected @xmath1-manifold @xmath24 . by summing together two copies of the torus @xmath22 , one copy from each side , we obtain a family of homologous genus @xmath10 , self - intersection @xmath5 symplectic surfaces @xmath25 . by direct computation
, we will show that if the fibered knot @xmath17 used in the construction of @xmath26 is non - trivial , then infinitely many of these surfaces @xmath25 have non - homeomorphic complements @xmath27 , distinguished by their fundamental groups .
this proves , in particular , that the surfaces are not isotopic .
in this section we will construct an infinite family of non - isotopic , symplectic surfaces of genus greater than or equal to @xmath28 for a class of simply connected , symplectic @xmath1-manifolds .
these 4-manifolds are the manifolds denoted by @xmath26 , where @xmath17 is any fibered knot , introduced in @xcite .
first we recall the construction of @xmath29 .
let @xmath30 be a fibered knot and let @xmath31 denote its genus .
let @xmath32 denote the fiber of the fibration of the knot exterior , the minimal genus seifert surface for the knot @xmath17 . for each fibered knot
@xmath17 we can construct a symplectic @xmath1-manifold @xmath14 , homeomorphic to the elliptic surface @xmath16 ( the @xmath15 surface ) , obtained by knot surgery on @xmath16 with the knot @xmath17 ( see @xcite ) .
denote by @xmath33 the fibered @xmath34-manifold obtained by the @xmath5-surgery on @xmath35 along @xmath17 .
@xmath36 is a symplectic @xmath1-manifold with a framed , self - intersection @xmath5 symplectic torus @xmath37 , where @xmath38 is the core of the surgery solid torus .
the manifold @xmath14 can be presented as @xmath39 where the gluing diffeomorphism identifies factorwise the boundary 3-tori @xmath40 and @xmath41 ( reversing the orientation on the last factor ) .
the presentation of @xmath14 as a fiber sum shows , by gompf s theory ( see @xcite ) , that it admits a symplectic structure restricting , outside the gluing locus , to the one of the summands . inside @xmath42 a disk section @xmath43 of @xmath44 can be glued to a seifert surface @xmath45 to form a closed , symplectic , genus @xmath31 surface @xmath46 of self - intersection @xmath47 . by taking @xmath46 and a regular torus fiber @xmath48 and resolving their normal intersection
, we obtain a symplectic surface @xmath49 of self - intersection @xmath5 and genus @xmath50 inside @xmath42 .
such a surface can be endowed with a natural framing , inherited from the canonical framings of the fibers of @xmath16 and @xmath36 .
we define @xmath51 where the gluing map @xmath52 is an orientation - reversing self - diffeomorphism on @xmath53 that is the identity on the @xmath49 factor and complex conjugation on the @xmath54 factor . the following result is proved in @xcite .
@xmath29 is a closed , simply connected , spin , irreducible , symplectic @xmath1-manifold .
the signature of @xmath29 is @xmath55 , and its intersection form is given by @xmath56 $ ] . in order to obtain higher genus symplectic surfaces in @xmath26
, we will start from some symplectic tori in @xmath14 .
the symplectic @xmath1-manifold @xmath14 contains , for every @xmath7 , an infinite family @xmath57 of pairwise non - isotopic symplectic tori representing the homology class @xmath58 = q[s^{1 } \times { \bf m } ] \in h_2(e(2)_k;{\mathbb{z}})$ ] indexed by the integers @xmath59 coprime to @xmath23 ( see @xcite and @xcite ) .
there are two equivalent ways to present these tori .
we start with the approach of @xcite .
first , by looking at the construction of @xmath14 in ( [ knotsur ] ) , we can recognize a rim torus @xmath60 given by the image in @xmath14 of @xmath61 .
as we can assume , up to isotopy , that a copy of the longitude @xmath62 lies on the fiber of @xmath33 ( given by the seifert surface @xmath32 capped off with a disk of the dehn filling ) , the torus @xmath60 can be assumed to be lagrangian . inside @xmath36
the torus @xmath63 is the boundary of a solid torus ( as @xmath62 bounds a disk of the dehn filling ) but after the knot surgery this is not generally true anymore for @xmath60 ( for non - trivial @xmath17 ) .
it remains , however , nullhomologous , as it bounds the @xmath34-manifold @xmath64 .
take now a copy of @xmath65 pushed slightly inside @xmath66 , in such a way that it is transverse to the fibration and intersects each fiber in a curve isotopic to the longitude .
the simple closed curves on this torus are the @xmath67-cables of @xmath17 , where @xmath68 ( remember that their linking number with @xmath17 is @xmath23 ) , with @xmath62 and @xmath69 themselves being the @xmath70- and the @xmath71-cables respectively . up to isotopy we can assume that , with the exception of @xmath70-cable , every @xmath67-cable is a curve transverse to the fibration , and its intersection with a fiber consists of positive points for @xmath7 .
denote by @xmath72 the @xmath67-cable of @xmath17 , for any @xmath7 .
we can say that this closed curve is obtained by _ circle sum _
( in the sense of @xcite ) of @xmath23 copies of the meridian and @xmath59 copies of the longitude of @xmath17
. it will be useful in what follows to observe that the cable knot @xmath72 is contained in an enlarged tubular neighborhood @xmath73 of @xmath17 ( which we can assume to have , say , twice the radius of @xmath74 for some metric on @xmath35 ) .
note that @xmath75 also fibers over @xmath76 .
its fiber @xmath77 is strictly contained in @xmath32 , with @xmath78 isotopic to @xmath79 inside @xmath32 .
the cable @xmath72 intersects the fiber @xmath32 at @xmath23 points @xmath80 contained in the knotted annulus @xmath81 , whose boundary is the union @xmath82 .
figure [ rela ] illustrates the relations between these fibers and @xmath72 . and fibers of the two knot exteriors ]
as @xmath72 is transverse to @xmath32 the torus @xmath83 is symplectic in @xmath36 . for a fixed value of @xmath23 ,
the tori @xmath84 are homologous , as the homology class of @xmath72 in @xmath85 is given by @xmath86 = q [ \mu(k)]$ ] and the dehn filling does not affect this relation . ( if @xmath87 more is true , namely the tori @xmath88 are isotopic , the isotopy being supported in the solid torus of the dehn filling . ) after knot surgery , the images @xmath22 in @xmath14 of these tori remain homologous . in fact , the image of the homology class of the @xmath67-cable of @xmath17 , through the injective map @xmath89 is @xmath58 $ ] .
however , infinitely many of the tori @xmath22 become non - isotopic .
roughly speaking , the torus @xmath22 is the sum of @xmath23 copies of @xmath48 with @xmath59 copies of the nullhomologous lagrangian torus @xmath60 , and can be interpreted in terms of the cabling of @xmath48 itself .
the family of tori @xmath22 will be the starting point of our construction , but note that we can extend the previous argument to include , instead of just @xmath67-cables of the knot @xmath17 , any curve in @xmath90 obtained as the circle sum of a number of copies of the meridian and a closed curve ( or a union of closed curves ) lying on the surface @xmath32 ( and thus homologous , but not necessarily isotopic to the longitude ) .
with arguments similar to the ones above , we can achieve transversality for the resulting curves ; the symplectic tori obtained this way differ by the addition of nullhomologous lagrangian tori .
details of this construction are presented in @xcite . a second way to construct the family @xmath22 , presented in @xcite ,
is the following .
denote by @xmath91 the hopf link . in the knot exterior @xmath92
we consider a standard pair of meridian and longitude , that we denote , to avoid confusion with @xmath69 and @xmath62 , by @xmath93 and @xmath94 respectively .
we have then the decomposition @xmath95 \,\cup\ , \big[s^1\!\times(s^3\setminus \nu h)\big]\,\cup\ , \big [ s^1\!\times ( s^3\setminus { { \sf n}}k)\big]\ , , \ ] ] with the first gluing identifying @xmath96 with @xmath97 and the second one identifying @xmath98 with @xmath99 factorwise , reversing the orientation on the last factor .
( the bar over a curve refers to the negative orientation . )
the equivalence between this construction and the construction of ( [ knotsur ] ) ( even in the symplectic category , with a suitable presentation as symplectic fiber sum ) simply follows by observing that the exterior of the hopf link is diffeomorphic to @xmath100 . the torus @xmath22 is given by the image of @xmath101 $ ] , where @xmath102 is the @xmath67-cable of @xmath103 , as illustrated in figure [ fig : doublehopf ] for @xmath87 .
the transversality of @xmath102 to the fibration of @xmath104 with fiber given by a disk spanning @xmath103 , pierced once by @xmath105 , is evident , and this entails that @xmath22 is symplectic . for fixed @xmath23 ,
the curves @xmath102 are not homologous in @xmath104 , but they become so after gluing in @xmath92 with the prescription above , as the meridian @xmath106 becomes nullhomologous .
therefore , the tori @xmath22 are homologous in @xmath14 for fixed @xmath23 . in @xmath18 ]
the equivalence between the two constructions of @xmath22 follows when we observe that the gluing of @xmath92 to @xmath104 in ( [ etgu - park construction ] ) is the _ splicing _ of @xmath17 and @xmath108 along @xmath109 , and the result of this operation is to produce a copy of @xmath90 , with the image of @xmath103 giving the knot @xmath17 and the image of the @xmath67-cable of @xmath103 ( i.e. @xmath110 ) giving the @xmath67-cable of @xmath17 ( i.e. @xmath111 ) .
( see @xcite , proposition 1.1 for a detailed discussion of this construction . )
stated otherwise , we can think of the curve @xmath102 in the second construction as spiraling along an inner torus of @xmath112 ; after the gluing this inner torus gets identified with a copy of @xmath65 pushed inside @xmath90 , and hence @xmath102 is identified with @xmath72 in the first construction .
note that , after the splicing , the disk spanning @xmath103 and pierced by @xmath105 ( that appears shaded in figure [ fig : doublehopf ] ) is identified with @xmath113 , which appears as the neck in figure [ rela ] . now that we have defined the family @xmath114 we can construct , out of two copies of @xmath22 , the desired genus @xmath10 surface in @xmath26 . since in @xmath42 , @xmath115\cdot[\sigma_g]=q[f]\cdot[\sigma_g]=q[f]\cdot[s ] = q\,$ ] for each @xmath59 , we can internally sum two copies of @xmath23-times punctured @xmath22 from each of @xmath116 halves in ( [ d_k construction ] ) , and thus obtain a family of homologous , genus @xmath10 , self - intersection 0 surfaces @xmath117 inside @xmath29 .
( we are going to follow the convention that @xmath118 . )
see figure [ fig : xi ] . as the double of @xmath119 since
each @xmath22 is symplectic in @xmath42 and our ambient @xmath1-manifold @xmath29 is a symplectic sum ( cf.@xmath120@xcite ) of two copies of @xmath42 along a symplectic surface @xmath49 , each @xmath25 is a symplectic submanifold of @xmath29 . in the next section
we discuss how to distinguish the isotopy classes of submanifolds @xmath25 by comparing the fundamental groups of their complements in @xmath29 .
the isotopy class of the torus @xmath22 inside @xmath14 can be detected via the seiberg - witten invariant of the fiber sum manifold @xmath121 ( that depends ultimately on the diffeomorphism type of the pair @xmath122 ) .
it would be interesting ( and probably quicker ) to be able to use a similar approach to distinguish the diffeomorphism type of the pair @xmath123 .
unfortunately , we currently lack a satisfactory machinery within the seiberg - witten theory to carry this out . however
, we can use a classical approach to this problem , namely the study of the fundamental group of the exterior of @xmath25 .
this allows us to prove an even stronger result , namely the existence of pairs @xmath123 that are not even homeomorphic .
we were first led to this approach by the observation that the tori @xmath22 themselves can be distinguished by means of the fundamental group alone , as illustrated in @xcite .
the reason underlying that phenomenon seems to be related to the role of the nullhomologous lagrangian tori in the construction of @xmath22 . it remains an interesting question to determine whether there are homeomorphic pairs ( or at least pairs with the same fundamental group of the exterior ) which are not diffeomorphic .
the main result of this section is the following theorem .
[ thm : fundamental ] let @xmath25 be the genus @xmath10 , symplectic surface inside @xmath29 constructed in section @xmath124
. then we have @xmath125 where @xmath126 is the rational homology sphere obtained by dehn @xmath127-surgery on @xmath35 along the knot @xmath17 , so that @xmath128 with @xmath69 and @xmath62 denoting the meridian and the longitude of @xmath17 respectively .
theorem [ main ] now follows from theorem [ thm : fundamental ] and the fact that when @xmath17 is a non - trivial knot , for a fixed value of @xmath23 , infinitely many choices of @xmath59 give mutually non - isomorphic groups @xmath129 .
a proof of this latter statement is contained in @xcite .
for some knot @xmath17 , we can explicitly determine infinite values of @xmath59 that lead to pairwise non - homeomorphic pairs .
as an example , consider the figure eight knot .
this knot is known to be fibered and hyperbolic .
moreover , it is known ( see @xcite ) that , with @xmath130 explicit exceptions all having @xmath131 , its @xmath127-dehn surgery is a hyperbolic manifold . at this point
we can invoke the result of @xcite which guarantees that , restricting ourselves to positive @xmath59 , no two surgeries give homeomorphic @xmath34-manifolds .
mostow rigidity then implies that all the resulting manifolds have distinct fundamental groups .
the proof of theorem [ thm : fundamental ] will occupy the remainder of this section .
we start by observing that we have the following decomposition into halves @xmath132\:\cup\:\big[e(2)_k
\setminus \nu(\sigma_g\cup t_{p , q } ) \big].\ ] ] we want to apply seifert - van kampen theorem to this decomposition .
our first step is the computation of the fundamental group of the halves .
we claim that @xmath133 recall from ( [ knotsur ] ) and the construction in section [ sec : construction ] that we have @xmath134 \,\cup\ , \big[s^1\!\times(s^3\setminus \nu ( k \cup k_{p , q}))\big].\ ] ] our strategy is to use decomposition ( [ decomposition : t_q complement ] ) to compute the fundamental group in ( [ pi_1(half ) ] ) .
note that the surface @xmath49 intersects both components of decomposition ( [ decomposition : t_q complement ] ) .
we proceed to compute the fundamental group of the complement of @xmath135 in each piece separately .
denote by @xmath136 and @xmath137 the two components in decomposition ( [ decomposition : t_q complement ] ) .
the first result we need is a straightforward modification of proposition 2.1 of @xcite , and we present its proof here for completeness . observe that @xmath138 ( i.e. @xmath139 ) , from the definition in ( [ decomposition : t_q complement ] ) , does not depend on @xmath17 .
[ lemma : x_1 ] let @xmath140 . then @xmath141 .
note that @xmath138 is @xmath142\setminus \nu \sigma'$ ] , where @xmath143 is a genus @xmath4 , self - intersection @xmath5 surface with boundary , gotten from the union @xmath144 of another copy of regular fiber and a disk section of @xmath145 by resolving their normal intersection point .
it is well known that @xmath142 $ ] is simply connected ( see e.g. @xmath146@xcite ) .
it now follows from seifert - van kampen theorem that the fundamental group in question is normally generated by the meridian of @xmath143 . by perturbing the elliptic fibration of @xmath16 if necessary
, we may assume that there is at least one singular cusp fiber which is topologically a sphere .
this implies that the meridian of @xmath143 ( which is isotopic to a meridian of @xmath43 ) bounds a topological disk , which is a cusp fiber with a disk removed .
we now consider the remaining piece in decomposition ( [ decomposition : t_q complement ] ) .
first , observe that @xmath147 admits a fibration over @xmath76 , with fiber @xmath148 given by the spanning surface @xmath32 of @xmath17 with @xmath23 disjoint disks @xmath149 removed , in correspondence with the intersection points @xmath80 of @xmath32 with @xmath72 .
these @xmath23 disks all lie in the knotted annulus @xmath113
. the existence of such fibration should be quite apparent from the constructions in section [ sec : construction ] , and comes from gluing , through the splicing , the fibration of @xmath75 with fiber @xmath77 and the fibration of @xmath150 induced by the spanning disk of @xmath103 ( that gives the punctured knotted annulus above ) . a section of this fibration is given by a copy of @xmath69 .
denote by @xmath151 , @xmath152 , the generators of @xmath153 and denote by @xmath154 , @xmath155 , the homotopy classes of the loops around the disks @xmath156 . locating the base point on @xmath157 we can assume that representatives of the generators @xmath158 and @xmath159 are all contained in @xmath160 .
the fundamental group of @xmath148 ( a free group on @xmath161 generators ) is generated by these elements .
[ lemma : x_2 ] let @xmath162 $ ]
. then @xmath163 is isomorphic to the group @xmath164=[x , b_i]=[x , c_{j}]= 1 \mid i=1,\dots , g(k ) \
{ \rm and}\ j=1,\dots , q\}}\ : , \ ] ] where @xmath165 is the free group on one generator , and @xmath166 are the generators of @xmath167 . note that @xmath137 is the total space of a fiber bundle : @xmath168 & \;\;s^1\!\times(s^3\setminus \nu ( k \cup k_{p , q } ) ) \ar[d]^{\;\pi\:=\:{\rm id}\times\pi } \\ &
s^1 \!\times \mu(k ) } \ ] ] suppose @xmath169 for some point @xmath170
. then we can write @xmath171 for some small disk @xmath172 .
it follows that @xmath173 is the total space of the restricted bundle : @xmath168 & \;\;\big[s^1\!\times(s^3\setminus \nu ( k \cup k_{p , q}))\big]\setminus \pi^{-1}(\mathbb{d}_{t_0 } ) \ar[d]^{\;\pi| } \\ & [ s^1 \!\times \mu(k)]\setminus \mathbb{d}_{t_0 } } \ ] ] the new base @xmath174\setminus \mathbb{d}_{t_0}$ ] is a torus with a disk removed and hence homotopy equivalent to a wedge of two circles .
the generator @xmath175 corresponding to the @xmath176 factor in @xmath177 now no longer commutes with @xmath69 , but @xmath175 still commutes with the generators of @xmath178 since the map @xmath179 had trivial monodromy in the @xmath176 direction .
presentation ( [ does ] ) follows immediately .
now we are ready to prove ( [ pi_1(half ) ] ) .
the intersection of @xmath180 and @xmath181 is given by the 3-torus @xmath182 minus its intersection with @xmath183 .
this intersection is given by the neighborhood of a copy of the meridional circle @xmath184 .
we have in fact @xmath185 , so that @xmath186 \oplus ( { \mathbb{z}}[\gamma_{1 } ] \ast { \mathbb{z}}[\gamma_{2}]),\ ] ] where @xmath187 and @xmath188 form a homotopy basis for @xmath189 .
the generators @xmath184 , @xmath187 and @xmath188 are identified , respectively , with the classes of @xmath190 , @xmath69 and @xmath76 through the gluing map in ( [ knotsur ] ) . by using the fact that @xmath191 is simply connected ( lemma [ lemma : x_1 ] )
, we deduce that @xmath192 where the last equality comes from observing that the elements of @xmath193 identified with the elements of @xmath194 become trivial .
this accounts for the nullhomotopy of generators @xmath195 $ ] , @xmath69 and @xmath62 appearing in group ( [ does ] ) .
we are left , therefore , with the exercise of computing the fundamental group of the exterior of a link given by a knot @xmath17 and its @xmath67-cable , and then quotienting by the relations that @xmath62 and @xmath69 are trivial . in principle
, we could use the existence of the fibration of @xmath196 , but this requires an explicit knowledge of the monodromy of the fibration .
another approach is much more viable .
first , it is useful to keep in mind the construction of @xmath196 obtained by splicing @xmath92 and @xmath150 .
now we can proceed ( mimicking one of the standard computations of the fundamental group of a torus knot exterior ) as follows . presenting the knot and its cable as companion and satellite ,
we can write @xmath197 the following three paragraphs explain the terms appearing on the right hand side of ( [ svk ] ) .
first , @xmath198 is a @xmath34-manifold with torus boundary that is given by the exterior @xmath92 of the knot @xmath17 union a collar of its boundary , extended in the outward direction .
@xmath72 lies on the boundary of @xmath199 . by carving out the intersection with a tubular neighborhood of @xmath72
, we obtain the manifold @xmath200 that is a deformation retract of @xmath198 .
recall that we denote by @xmath93 and @xmath94 the meridian and the longitude of @xmath92 .
we point out that these should not be confused with the meridian and the longitude @xmath69 and @xmath62 of @xmath66 ; in particular , they are not homotopic to @xmath69 and @xmath62 in @xmath196 .
the 3-manifold @xmath201 is a solid torus knotted in @xmath35 as a tubular neighborhood of @xmath17 , with a neighborhood of the core ( @xmath74 ) removed .
@xmath72 lies on the outer boundary of @xmath202 . by carving out the intersection with a tubular neighborhood of @xmath72 we obtain @xmath203 that is , again , a deformation retract of @xmath201 and
in particular is homotopy equivalent to a 2-torus .
the generators of @xmath204 are given exactly by @xmath69 and @xmath62 .
@xmath198 and @xmath201 intersect in a 2-torus , which contains the curve @xmath72 , and which has a natural homotopy basis given by @xmath205 and @xmath206 ( the identification being in @xmath207 ) , so that @xmath72 is homotopic to @xmath208 .
if we avoided removing the neighborhood of @xmath72 , the intersection of @xmath198 and @xmath201 being a 2-torus , seifert - van kampen theorem would give the obvious identifications @xmath205 , @xmath206 in the resulting 3-manifold ( and decomposition ( [ svk ] ) would simply be a redundant presentation of @xmath90 ) . instead , the intersection of @xmath209 and @xmath210 is an annulus @xmath211 , given by the exterior of @xmath72 in the torus @xmath212 .
such an annulus deformation - retracts to its core , which is homotopic in @xmath212 to a parallel copy of @xmath72 , whose image in @xmath213 is given , respectively , by @xmath214 and by @xmath215 .
therefore an application of seifert - van kampen theorem gives @xmath216 \oplus { \mathbb{z}}[\lambda(k)]\big)}{{{\sf m}(k)}^{q}{{\sf l}(k)}^{p } = \mu(k)^{q } \lambda(k)^{p}}.\ ] ] we can now complete our argument .
when we quotient by the relations @xmath217 in ( [ befq ] ) we obtain that @xmath218 and considering that @xmath198 deformation - retracts to @xmath75 with @xmath93 and @xmath94 corresponding to the meridian and the longitude of @xmath17 respectively , we obtain that @xmath219 this finishes the proof of ( [ pi_1(half ) ] ) .
the second step in the proof of ( [ pi_1(xi_q complement ) ] ) consists of applying seifert - van kampen theorem to decomposition ( [ decomposition : xi_q complement ] ) using ( [ pi_1(half ) ] ) . note that the fundamental group of @xmath220 is generated by @xmath221 ( a free group @xmath222 on @xmath223 generators @xmath151 , @xmath152 ) and the meridian @xmath93 , subject to the monodromy relations .
the relation @xmath224 adds an extra relation between the meridian and the generators of @xmath221 , since @xmath225 $ ] . in sum
, we can write ( [ pi_1(half ) ] ) as @xmath226 where @xmath227 denotes the monodromy map of the fiber bundle @xmath228 .
let @xmath229 denote the intersection of the two halves in ( [ decomposition : xi_q complement ] ) .
precisely , @xmath229 is diffeomorphic to the cartesian product @xmath230 , where @xmath231 is the disjoint union of @xmath23 disks centered at the intersection points @xmath232 of @xmath49 with @xmath22 .
note that @xmath233 , where @xmath143 is the genus @xmath4 surface with boundary described in the proof of lemma [ lemma : x_1 ] .
the loops of the form @xmath234 , which are copies of the meridian of @xmath20 in @xmath26 , are contractible in each half - component of ( [ decomposition : xi_q complement ] ) , as in the proof of lemma [ lemma : x_1 ] .
remember now that the surface @xmath77 is contained in @xmath235 , as the latter is built by gluing to @xmath77 an annulus ( see figure [ fig : doublehopf ] ) in @xmath236 with @xmath23 disks removed , and then capping off with @xmath143 . when applied to ( [ decomposition : xi_q complement ] ) , seifert - van kampen theorem gives , as @xmath237 , the free product of the fundamental groups of the two halves , modulo the identification induced by the image of the generators of @xmath238 .
[ lemma : inclusion ] the fundamental group @xmath239 of each half - component of decomposition @xmath240 is generated by the image of @xmath241 under the homomorphism induced by the inclusion map . in order to prove this lemma we need to keep track of the representatives of the generators of the fundamental group of each half - component . from presentation
( [ each ] ) we see that the generators for both copies of @xmath242 denoted by @xmath158 and @xmath159 are contained in the image of @xmath241 for obvious reasons , resulting in the identification of the generators denoted by the same symbols .
the lone remaining generator , for each half - component , is @xmath93 .
now we can observe that , as an element of @xmath243 , @xmath93 can be written as a word in the generators of this group , for which we choose the presentation used in the proof of lemma [ lemma : x_2 ] . by construction
this word will be the same for both halves of decomposition ( [ decomposition : xi_q complement ] ) .
it follows that @xmath93 , as an element of @xmath244 , can be written as a word in @xmath69 , the @xmath158 s , the @xmath159 s and in the @xmath154 s .
( a more careful study of @xmath245 would enable us to identify this word , but we will not need its explicit form . ) when we quotient the group @xmath244 by the relations @xmath246 to obtain @xmath247 , we see that ( the image of ) @xmath93 can be written as a word in the generators of ( the image of ) @xmath248 , and _ a fortiori _ of @xmath238 . as the word is the same on each half of the decomposition ,
the generators @xmath93 of two halves are identified . as a consequence of lemma [ lemma : inclusion ] the free product with amalgamation identifies all the corresponding generators of @xmath249 of two halves of ( [ decomposition : xi_q complement ] ) and we get the simple formula @xmath250 with the latter group presented ( using standard notation ) in ( [ stand ] ) . this concludes the proof of theorem [ thm : fundamental ] .
we thank steven boyer and olivier collin for their help in obtaining some results of @xcite that have been used here .
we also thank tolga etg and ronald fintushel for stimulating discussions .
we are also grateful to the anonymous referee for her or his useful observations . | we construct an infinite family of homologous , non - isotopic , symplectic surfaces of any genus greater than one in a certain class of closed , simply connected , symplectic four - manifolds .
our construction is the first example of this phenomenon for surfaces of genus greater than one . |
_ multiple hard parton interactions _ ( mpi ) is an important element of the picture of strong interactions at high energies .
the issue attracts a lot of attention .
a series of theoretical studies were carried out in the last decade @xcite .
attempts have been made to incorporate multi - parton collisions into event generators @xcite .
multi - parton interactions can serve as a probe for _ non - perturbative correlations _ between partons in the nucleon wave function and are crucial for determining the structure of the underlying event at lhc energies .
they constitute an important background for new physics searches at the lhc .
a number of experimental studies were performed at the tevatron @xcite .
new measurements are underway at the lhc @xcite .
double hard parton scattering in hadron hadron collisions ( dpi ) can contribute to production of _ four _ hadron jets with large transverse momenta @xmath0 , of _ two _ electroweak bosons , or `` mixed '' ensembles comprising three jets and @xmath1 , two jets and @xmath2 , etc .
in this paper we present the numerical results for a variety of final states .
however , for the sake of definiteness , in what follows we refer to production of four final state jets in the collision of partons 1 and 2 from one incident hadron with partons 3 and 4 from the second hadron : @xmath3 , @xmath4 .
double hard interaction is a process difficult to approach theoretically .
formally speaking , it calls for analysis of four - parton operators that emerge in the squared matrix element describing a two - parton state in a hadron .
relevant objects quasi - parton operators were introduced and classified , and evolution of their matrix elements studied by bukhvostov et .
al in the eighties in a series of papers @xcite .
an approach to mpi based on the operator product expansion and on the notion of transverse momentum dependent parton distributions ( tmd ) is being developed in @xcite .
mpi is a multi - scale problem . not only because the separate parton parton interactions may differ in hardness .
more importantly , _
each _ single hard interaction possesses _ two _ very different hardness scales .
the distinctive feature of dpi is that it produces two pairs of nearly back - to - back jets , so that [ eq : deltas ] _ 13 ^ 2 & & ( _ 1 + _ 3)^2 q_1 ^ 2 = j_1 ^ 2_3 ^ 2 , + _ 24 ^ 2 & & ( _ 2 + _ 4)^2
q_2 ^ 2 = j_2 ^ 2_4 ^ 2 .
hence , in the collision of partons 1 and 3 the first ( larger ) scale is given by the invariant mass of the jet pair , @xmath5 , while the second scale is the magnitude of the _ total transverse momentum _ of the pair : @xmath6 .
it is important to stress that in the mpi physics there is _ no factorization _ in the usual sense of the word .
the cross sections do not factorize into the product of the hard parton interaction cross sections and the multi - parton distributions depending on momentum fractions @xmath7 and the hard scale(s ) . in @xcite
we have introduced necessary theoretical tools for approaching the problem by introducing the _ generalized double parton distributions _ ( @xmath8s ) in the momentum space .
the double - parton gpd depends on one extra variable as compared to the double - parton distribution the transverse momentum mismatch @xmath9 between the partons in the wave function and the wave function conjugated . in the mixed space of longitudinal momenta and transverse coordinates ,
an object equivalent to @xmath8 has been introduced by treleani and paver in the early eighties @xcite , and was long present in the literature since , see , in particular , @xcite . in order to construct a viable model for the two - parton distribution stemming from the non - perturbative parton wave function of the proton and , in particular , for the @xmath10-dependence of the corresponding part of @xmath8
, we have exploited the existing experimental information obtained by the hera experiments .
this model disregards ( rather arbitrarily , for lack of anything better ) any _ long - range correlations _ between the partons in the proton wave function , which approximation seems reasonable in a limited range of parton momenta @xmath11 ( see section [ sec : model ] for details ) .
this allows one to expresses the non - perturbative @xmath8 via the standard generalized parton distributions ( gpds ) studied at hera . in @xcite
we have studied how do perturbative qcd ( pqcd ) phenomena affect two - parton correlations .
one obvious effect is `` scaling violation '' due to parton multiplication processes in higher orders .
another important effect is _ short - range correlations _ induced by perturbative parton splitting , giving rise to _ three - parton collisions_. such contributions we will label as 2(``@xmath12 '' of @xcite , `` @xmath13 '' of @xcite ) .
this contribution to double hard interactions emerges as a result of collision of two partons from one hadron with the two offspring of the perturbative splitting of a single parton taken from another hadron .
perturbative splitting of one parton into two as a contributor to double - parton distributions was being discussed in the literature for a long time , see , e.g. , @xcite . however , for a relevant object double - parton gpd it was embedded into the parton evolution picture only recently .
the pqcd expressions for fully differential distributions were derived in the leading collinear approximation in @xcite . in the present paper
we examine how various contributions to the differential distribution , being integrated over transverse momentum imbalances , give rise to the expression for the _ integrated cross section _ in terms of the product of @xmath8s of colliding hadrons .
the results of numerical studies of the integrated dpi cross sections based on the theory and the model for non - perturbative two - parton distributions in the proton developed earlier are reported .
we demonstrate that the 2 processes contribute significantly to the dpi cross section and , in particular , are capable of explaining the longstanding tevatron puzzle @xcite .
namely , that the dpi cross section extracted by the the cdf and d0 at @xmath14 turned out to be a factor of two larger than the expectation based on the approximation of independent partons in the proton with the transverse distance spread extracted from the hera data @xcite . this observation does not exclude the presence of genuinely non - perturbative correlations between partons in the proton .
however , we find it interesting that the pqcd induced correlations alone could explain the scale of the enhancement found by the tevatron experiments .
we also discuss the pattern of the @xmath15- and @xmath16-dependence of the effective interaction area ( aka `` effective cross section '' ) induced by pqcd correlations at tevatron and lhc energies .
the paper is organized as follows . in section [ sec : hidden ] we reflect upon intrinsic complexity of theoretical mpi analysis . in section [ sec
: gpd ] we remind the reader the basics of our approach to generalized two - parton distributions , @xmath8 , and discuss the approximations adopted for constructing the model for @xmath8 in terms of the standard gpds . in section [ sec : p2pcor ] we discuss importance of perturbative parton correlations and give a semi - quantitative estimate of their magnitude .
section [ sec:12 ] is devoted to three - parton interactions ( 2 subprocesses ) .
here we reexamine the role of the so - called `` short split '' term in 2 subprocesses which is concentrated in the kinematical region of nearly equal transverse momentum imbalances , @xmath17 see @xcite .
we show here that the `` short split '' contribution to the _ integrated cross section _ is actually contained in @xmath18 that one obtains integrating the simple ddt - like formula derived in the complimentary kinematical region @xmath19 ( @xmath20 ) .
thus we correct ( and simplify ) the expression for the integrated dpi cross section ( original eq .
( 28 ) of @xcite was plagued by double counting ) . in section [ sec : model ]
we discuss the independent parton model for the non - perturbative part of two - parton correlations .
the numerical results for tevatron and lhc energies are presented in section [ sec : numerics ] .
we conclude and discuss the results and perspectives of dpi studies in section [ sec : conc ] .
an important question is , whether mpi admit an intuitive probabilistic parton interpretation as do the classical single hard interaction processes . when one considers inclusive one - parton distribution in a hadron ( pdf ) , the quantum state of the parton in the light - cone hadron wave function ( w.f . )
coincides with that in the wave function conjugated ( w.f.c . ) .
this lays down the foundation for the probabilistic qcd improved parton picture . in the case of the two - parton correlation
the situation is different . here only the overall quantum characteristics of the parton pair as a whole ( its total energy - momentum , its spin and color states ) should be identical in the w.f . and w.f.c .
as a result , the momentum , spin and color state of a single parton in the pair do not necessarily match in the wave function and in the wave function conjugated , thus endangering the very probabilistic interpretation of the process under consideration .
a general approach to double hard interactions has been developed in @xcite .
it turned out that the _ transverse momentum _ of the parton in the w.f . and that of its counterpart in the w.f.c
. are indeed necessarily different , with their difference @xmath21 being conjugate to the relative transverse distance between the two partons in the hadron .
this has led to introduction of the notion of the _ generalized double parton distribution _ , @xmath8 , which depends on a new momentum parameter @xmath21 .
it is important to stress that the cross section of the double hard process does not factorize into the product of the hard cross section and parton distributions .
instead , it contains a convolution of the product of two @xmath8s over @xmath22 .
non - diagonality in the longitudinal momentum fractions is there too .
representing incident partons as plain waves with definite momenta , one ignores the fact that the two partons originate from one and the same finite size hadron . therefore , when one picks two partons with momenta @xmath23 , @xmath24 from the hadron wave function , one has to integrate over @xmath25 at the level of the amplitude , which integration ensures that the longitudinal separation between the partons does not exceed the size of the parent hadron . when one considers independent hard interactions of two pairs of partons , @xmath26 and @xmath27 , taken from colliding hadrons , integrals over @xmath28 and @xmath29 do not manifest themselves , since all four longitudinal momentum fractions @xmath7 are uniquely determined by the kinematics of the two produced hard systems .
not so for 2 processes . in this case
a flow of large momenta between the two hard vertices is possible , and the integration over @xmath30 is instrumental in getting rid of a fake singularity of the scattering matrix element ( a simple explanation of the origin of this unphysical singularity can be found in @xcite , see also @xcite ) . as a result , a small offset between the parton longitudinal momenta in the w.f . and w.f.c .
emerges , @xmath31 .
however , this mismatch turns out to be negligible in the dominant kinematical region where the squared total transverse momentum @xmath32 of the jet pair is much smaller than the overall hardness @xmath16 of the process : @xmath33 .
as noticed by gaunt @xcite , the non - diagonality in the longitudinal momentum space is likely to get induced in 2 processes by higher order qcd effects .
this happens when incoming partons , in the w.f . and w.f.c .
, exchange ( real or virtual ) gluons with transverse momenta in the interval @xmath34 . in @xcite arguments
were raised in favor of smallness of the crosstalk effects in the dpi cross section .
otherwise , this would have led to unwelcome complication of the problem : one would be forced to deal with an unknown function of _ four _ longitudinal momentum fractions ( three independent variables ) in place of the two of @xmath35 .
hence for the time being we disregard this complication .
we do not dwell into potential non - diagonality in color and in spin variables either .
one may argue that such non - diagonal configurations are likely to be suppressed as they can be related with form factors for proton transition between two states with different quantum numbers of the proton constituents .
for example , swapping inside the proton the colors of two partons that sit at distances of the order of the nucleon radius , corresponds to excitation that can be visualized as adding an extra piece of color string whose energy , @xmath36 , would excite and destroy the proton .
in @xcite we have developed a formalism to address the problem of multi - parton interactions .
qft description of double hard parton collisions calls for introduction of a new object _ the generalized two - parton distribution _ , @xmath37gpd .
defined in the momentum space , it characterizes two - parton correlations inside hadron @xcite : @xmath38 here the index @xmath39 refers to the hadron , @xmath23 and @xmath24 are the light - cone fractions of the parton momenta , and @xmath40 the corresponding hard scales .
as has been mention above , the two - dimensional vector @xmath9 is fourier conjugate to the relative distance between the partons @xmath41 and @xmath42 in the impact parameter plane .
the distribution obviously depends on the parton species ; we suppress the corresponding indices for brevity .
the double hard interaction cross section ( and , in particular , that of production of two dijets ) can be expressed through the generalized two - parton distributions @xmath37gpd .
@xmath8s enter the expressions for the differential distributions in the jet transverse momentum imbalances @xmath43 in the kinematical region , as well as for the `` total '' dpi cross section ( integrated over @xmath43 ) . in the latter case the hardness parameters of the @xmath8s are given by the jet transverse momenta @xmath44 , while in the differential distributions by the imbalances @xmath45 themselves .
the corresponding formulae derived in the leading collinear approximation of pqcd can be found in ref .
@xcite .
it is important to bear in mind that the dpi cross section _ does not factorize _ into the product of the hard parton interaction cross sections and the two two - parton distributions depending on momentum fractions @xmath7 and the hard scales , @xmath46 . instead , [ eq : xss ] & = & , + 1 _ & & .
the _ effective interaction area _
@xmath47 is given by the convolution of the @xmath8s of incident hadrons over the transverse momentum parameter @xmath21 normalized by the product of single - parton inclusive pdfs .
for expression for the dpi cross section to make sense , the integral over @xmath10 has to be convergent .
this is well the case when the two partons are taken from the non - perturbative ( np ) proton wave function .
indeed , a typical inter - parton distance in the proton is large , of the order of the hadron size @xmath48 .
accordingly , one expects the corresponding correlator in the momentum space to be concentrated at a finite np scale @xmath49 and to fall fast at large @xmath50 ( exponentially or as a sufficiently high power of @xmath10 ) .
however , there is another source of two - parton correlations .
this is purely perturbative ( pt ) mechanism when the two partons emerge from _ perturbative splitting _ of one parton taken from the hadron wave function . in this scenario
the production of the parton pair is concentrated at much smaller distances . as a result
, the corresponding contribution to @xmath37gpd turns out to be practically independent on @xmath50 in a broad range , up to the hard scale(s ) characterizing the hard process under consideration ( @xmath50 only affects the lower limit of transverse momentum integrals in the parton cascades , causing but a mild logarithmic dependence )
. given essentially different dependence on @xmath10 , one has to treat the two contributions separately by casting the @xmath37gpd as a sum of two terms : [ eq:2terms ] d_h ( x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; ) & = _ [ 2]d_h(x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; ) & + _ [ 1]d_h(x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; ) .
here subscripts @xmath51}d$ ] and @xmath52}d$ ] mark the first and the second mechanisms , correspondingly : two partons from the wave function versus one parton that perturbatively splits into two .
in this section we discuss the role of the pt parton correlations and show that , given a sufficiently large scale of hard interactions , they turn out to be as important as np ones .
let us chose a scale @xmath53 that separates np and pt physics to be sufficiently low , so that parton cascades due to evolution between @xmath53 and @xmath44 are well developed . to get a feeling of relative importance of the pt correlation , as well as to understand its dependence on @xmath15 and the ratio of scales , @xmath16 vs. @xmath53 , the following lowest order pt estimate can be used .
imagine that at the scale @xmath53 the nucleon consisted of @xmath54 quarks and @xmath55 gluons ( `` valence partons '' ) with relatively large longitudinal momenta , so that triggered partons with @xmath56 resulted necessarily from pt evolution .
in the first logarithmic order , @xmath57 , the inclusive spectrum can be represented as @xmath58 where we suppressed @xmath15-dependence as irrelevant . if both gluons originate from the same `` valence '' parton , then [ eq:1d ] _ [ 1]d 12n_cd + ( n_qc_f^2 + n_gn_c^2)^2 , while independent sources give @xmath51}\!d$ ] : [ eq:2d ] & ( n_q(n_q-1)c_f^2 + 2n_qn_gc_fn_c + n_g(n_g-1)n_c^2)^2 & = d^2 - ( n_qc_f^2 + n_gn_c^2)^2 .
recall that the @xmath10-dependence is different in and .
however , at @xmath59 the second terms cancel in the sum and we get for the correlator [ eq : corrmodel ] -1 .
the correlation is driven by the gluon cascade - the first term in and is not small ( being of the order of unity ) .
it gets diluted when the number of independent `` valence sources '' at the scale @xmath53 increases .
this happens , obviously , when @xmath7 are taken smaller . on the other hand , for large @xmath60 and increasing ,
the effective number of more energetic partons in the nucleon is about 2 and decreasing , so that the relative importance of the 2 processes grows .
we conclude that the relative size of pt correlations is of order one , provided @xmath61 .
moreover , the pt parton correlations can not be disregarded without running a risk of violating general principles .
this can be illustrated by looking at the momentum sum rule for double parton distributions .
an obvious momentum sum rule should be satisfied .
namely , that the integral over @xmath62 with the weight @xmath24 ( summing over all parton species ) should produce in the end @xmath63 times the inclusive one - parton distribution @xmath64 , that is , the total longitudinal momentum carried by all the partons but the triggered one : [ eq : momsum ] _ i_2 & dx_2 x_2 d^(i_1,i_2)_h(x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; = 0 ) & = ( 1-x_1 ) d_h^i_1(x_1 , q_1 ^ 2 ) ( here we have explicitly restored the parton species indices @xmath65 , @xmath66 ) .
this sum rule , together with other ones concerning valence quantum numbers , has been discussed by gaunt and stirling in @xcite as means for restricting the form of double parton distributions . setting the @xmath10 argument of @xmath8 to @xmath59
corresponds to taking integral over the relative transverse distance between partons in the proton .
derivation of is carried out in appendix [ sec : momentum ] .
it explicitly demonstrates that the pt parton splitting enters on equal grounds with the contribution due to two partons taken both from the initial np hadron wave function .
actually , the np and pt contributions _ do not _ enter the physical dpi cross section in arithmetic sum , driving one even farther from the familiar factorization picture based on universal ( process independent ) parton distributions . as explained in @xcite , a double hard interaction of two pairs of partons that _ both _ originate from pt splitting of a single parton from each of the colliding hadrons ,
does not produce back - to - back dijets .
in fact , such an eventuality corresponds to a one - loop correction to the usual @xmath67 jet production process and should not be looked upon as multi - parton interaction .
the term @xmath52}d_{h_1}\times { } _ { [ 1]}d_{h_2}$ ] has to be excluded from the product @xmath68 , the conclusion we share with gaunt and stirling @xcite .
so , we are left with two sources of genuine two - parton interactions : four - parton collisions described by the product of ( pt - evolved ) @xmath37gpds of np origin ( 2 ) , [ eq:2212 ] [ eq:22 ] _ [ 2]d_h_1(x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; ) _ [ 2]d_h_2(x_3,x_4 , q_1 ^ 2,q_2 ^ 2 ; - ) , and three - parton collisions due to an interplay between the np two - parton correlation in one hadron and the two partons emerging from a pt parton splitting in another hadron ( 2 ) , described by the combination [ eq:12 ] & _ [ 2]d_h_1 ( x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; ) _ [ 1]d_h_2 ( x_3,x_4 , q_1 ^ 2,q_2 ^ 2 ; - ) + & _ [ 1]d_h_1 ( x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; ) _ [ 2]d_h_2(x_3,x_4 , q_1 ^ 2,q_2 ^ 2;- ) .
given that @xmath69}\!d$ ] falls fast at large @xmath10 , a mild logarithmic @xmath10-dependence of @xmath52}d$ ] can be neglected in the product in .
separation of pt and np contributions is a delicate issue . by definition of the _ perturbative _ correlation function , @xmath52}d$ ]
vanishes when @xmath72 , @xmath73 are taken equal to the separation scale @xmath53 that one chooses to set the lower limit for applicability of pqcd calculations .
strictly speaking , @xmath53 can be chosen arbitrarily : both the np input function @xmath69}\!d$ ] and the pt - calculable correlation @xmath52}d$ ] contain @xmath53-dependence , but their sum does not depend on this formal parameter . at the same time , the character of the @xmath10-dependence of @xmath69}\!d$ ] depends , obviously , on the choice of the @xmath53 scale . indeed , by increasing the value of @xmath53 one will shuffle a part of the perturbative splitting contribution from @xmath74}\!d$ ] into @xmath69}\!d$ ] .
as a result , the `` np correlator '' @xmath69}\!d$ ] , contaminated by a short range pt correlation , would acquire a `` tail '' at large @xmath50 which would spoil convergence of the @xmath10 integration in .
thus , in order to preserve the logic of the np pt separation , one is led to introduce a specific resolution scale , @xmath75 , at which scale the np correlation @xmath69}\!d$ ] falls fast with increase of @xmath50 .
so defined , @xmath71 is no longer an arbitrary `` factorization scale '' but a _ phenomenological parameter _ whose value ( which one expects to be of order of @xmath76 ) should be established from the data . in order to derive the dpi cross section , one has to start with examination of the double differential transverse momentum distribution and then integrate it over jet imbalances @xmath77 .
why this step is necessary ?
the parton distribution @xmath78 the core object of the qcd - modified parton model arises upon logarithmic integration over the transverse momentum up to the hard scale , @xmath79 .
analogously , the double parton distribution @xmath80 embeds _
independent integrations _ over parton transverse momenta @xmath81 , @xmath82 up to @xmath72 and @xmath73 , respectively .
however , the 2 dpi cross section contains a specific contribution ( `` short split '' , see below ) in which the transverse momenta of the partons 1 and 2 are strongly correlated ( nearly opposite ) .
this pattern does not fit into the structure of the pqcd evolution equation for @xmath70 where @xmath83 and @xmath84 change independently .
given this subtlety , a legitimate question arises whether the expression for the integrated 2 cross section based on the notion of the two - parton distribution @xmath74}\!d$ ] takes the short split into account .
below we demonstrate that in fact it does .
the differential distribution over jet imbalances was derived in @xcite in the leading collinear approximation of pqcd .
it resembles the `` ddt formula '' for the drell - yan spectrum @xcite and contains two derivatives of the product of @xmath8s that depend on the corresponding @xmath77 as hardness scales , and the proper sudakov form factors depending on ( the ratio of ) the @xmath44 and @xmath85 .
in particular , in the region of _ strongly ordered _ imbalances , [ eq : dlog1 ] ; _ 13 ^ 2 _ 24 ^ 2 , _ 13 ^ 2 _ 24 ^ 2 , the differential 2 cross section reads [ eq : ddt31 ] & = & \ { _ [ 1]d_h_1(x_1,x_2,_13 ^ 2,_24 ^ 2 ; ) _ [ 2]d_h_2(x_3,x_4,_13 ^ 2 , _ 24 ^ 2 ; ) + & & s_1(q_1 ^ 2,_13 ^ 2 ) s_3(q_1 ^ 2,_13 ^ 2 ) s_2(q_2 ^ 2,_24 ^ 2 ) s_4(q_2 ^ 2,_24 ^ 2 ) } + \ { h_1 h_2 } .
the differential distribution for the 2 dpi mechanism has a similar structure , see eq .
( 25 ) of @xcite .
in addition to , there is another source of double collinear enhancement in the differential 2 cross section .
it is due to the kinematical region where the two imbalances nearly compensate one another , [ eq : deltaprime ] ^2 = ( _ 13 + _
24)^2 ^2 = _ 13 ^ 2_24 ^ 2 , and the dominant integration region is complementary to that of : [ eq : dlog2 ] ; ^2 ^2 .
this enhancement characterizes the set of 2 graphs in which there is no accompanying radiation with transverse momenta exceeding @xmath86 .
in this situation , the parton that compensates the overall imbalance , @xmath87 is radiated off the incoming , quasi - real , parton legs as shown in fig .
[ figshortkin ] . at the same time , the virtual partons after the core splitting `` 0''@xmath88 `` 1''+``2 '' enter their respective hard collisions without radiating any offspring on the way .
the @xmath89 splitting neighbors the hard vertices , therefore the name `` short split '' ( aka `` endpoint contribution '' , @xcite ) .
a closed expression for the differential distribution in jet imbalances due to short split , derived in the leading collinear approximation , is given by eq .
( 27 ) of @xcite : [ eq : ddt32 ] & = _ c p_c^(1,2 ) ( ) s_1(q_1 ^ 2,^2 ) s_2(q_2 ^ 2,^2 ) & \ { s_c(^2,^2 ) s_3(q_1 ^ 2 , ^2 ) s_4(q_2 ^ 2,^2 ) _ [ 2]d_h_2(x_3,x_4,^2 , ^2 ; ) } + \ { h_1 h_2 } .
the short split becomes less important when the scales of the two hard collisions separate .
indeed , the logarithmic integration over @xmath32 is kinematically restricted from above , @xmath90 . as a result ,
when transverse momenta of jets in one pair much exceed those of the second pair , e.g. , @xmath91 ( see ) , the contribution of the short split becomes suppressed as @xmath92 here @xmath93 and @xmath94 are the double logarithmic sudakov form factors of the partons `` 1 '' and `` 3 '' that enter the hard interaction with the larger hardness scale .
short split induces a strong correlation between jet imbalances which is worth trying to look for experimentally .
the relative weight of the short split depends on the process under consideration . for most dpi processes in the kinematical region
we have studied , it typically provides 1015% of the @xmath48 value .
however , it happens to be more important when the nature of the process favors parton splitting .
this is markedly the case of the double drell - yan pair production where the short split contribution reaches 3035% . on the contrary
, it turns out to be practically negligible for the same - sign double @xmath2-meson production ( see discussion in the conclusions section below ) . in the preceding publication
we have treated the contribution of the short split to the total dpi cross section as an addition , @xmath95 , to the 2 effective interaction area @xmath96 , see eqs .
( 28 ) , ( 32c ) of @xcite .
however , this was not a right thing to do . as it turns out , the short split contribution to the _ integrated cross section _
is already contained in . to see how this happens
, one has to examine the structure of the ddt formula for the 2 differential cross section more closely .
the distribution function @xmath52 } \!d$ ] in in the leading collinear approximation is given by eq .
( 18 ) of @xcite : [ eq : termpt ] _ [ 1 ] d^(a , b)_h(x_1,x_2 ; q_1 ^ 2,q_2 ^ 2 ; ) & = _
a,b,c ^ d^c_h(y , k^2 ) & p_c^(a,b)(z ) ^a_a(,q_1 ^ 2 ; k^2 ) ^b_b ( , q_2 ^ 2 ; k^2 ) . here
@xmath97 mark the registered partons , and @xmath98 are the indices of the partons involved in the splitting @xmath99 with the dglap probability @xmath100 .
the distribution @xmath101 describes the standard probability of finding a parton @xmath102 inside the incident hadron @xmath39 at the transverse momentum scale @xmath103 , and the functions @xmath104 in the second line stand for the distribution of parton @xmath105 , probed at scale @xmath106 , in the initial parton @xmath107 at lower virtuality scale @xmath103 .
let us examine the structure of the terms one gets applying two derivatives to substituted into : ^ ( k^2 ) ( k^2;q_1 ^ 2,q_2 ^ 2 ) . taking derivative over @xmath108 of the product @xmath109
corresponds to picking up from the parton chain an accompanying parton @xmath110 ( with the largest transverse momentum ) which compensates the imbalance , @xmath111 .
apart from logarithmic dependences of @xmath112 and @xmath113 , the derivative in may act upon the upper limit of the @xmath103integration in .
[ eq : dirs ] [ eq : dir0 ] ^ ( k^2 )
( k^2;q_1 ^ 2,q_2 ^ 2 ) .
differentiating the upper limit of the @xmath103-integral describes another legitimate situation when one of the partons `` 1 '' and `` 2 '' does not radiate before entering the hard interaction . in this case
the smaller of the jet imbalances is determined by the splitting momentum @xmath114 : [ eq : dir1 ] finally , applying _ both _ derivatives to the integration limit gives rise to [ eq : dir2 ] |_k^2=q_1
^ 2=q_2 ^ 2 . contrary to the first two contributions and , the last term obviously violates the condition of applicability of the ddt formula . instead , in the region @xmath115 it is the short split that contributes in the leading collinear approximation , so that eq .
has to be used to describe the differential spectrum in place of . however , as far as the _ total cross section _ is concerned , the integrals over imbalances of the short split and of the fake singular term turn out , as by miracle , to be the same . indeed , integrating the short split over @xmath116 up to @xmath32 , we get [ eq : xshort ]
_ c p_c^(1,2 ) ( ) ^ \{q_1 ^ 2,q_2 ^ 2 }
_ i=1 ^ 4 s_i _ [ 2 ] d _ h_2(x_3,x_4 , ^2,^2 ; ) , where @xmath117 .
on the other side , taking the integrand of @xmath52 } d$ ] in the point @xmath118 , and using @xmath119 we evaluate the momentum integrals , @xmath120 to arrive at the very same expression . it is worth noticing that this correspondence does not depend on the precise form of the upper integration limit in .
the result does not change , within the leading logarithmic accuracy , if one replaces a sharp @xmath121-function cut by a smooth damping factor that cuts the logarithmic @xmath103 integration at @xmath122 .
thus , for the integrated dpi cross section we obtain two contributions to the effective interaction area : [ eq : eff43 ] [ eq : eff4 ] & = & _ [ 2]d_h_1 ( x_1,x_2 , q_1
^ 2,q_2 ^ 2 ; ) _ [ 2]d_h_2 ( x_3,x_4 , q_1 ^ 2,q_2 ^ 2 ; - ) , + [ eq : eff3 ] & = & .
let us stress in conclusion that a compact and intuitively clear expression containing the product of the @xmath37gpds @xmath51}\mbox{d}$ ] and @xmath52}\mbox{d}$ ] in applies only to the _ integrated _ 2 cross section .
when addressing the _ differential _ distributions , one has to employ the `` ddt - like formula '' in the region of strongly ordered transverse momenta , and a quite different expression in the kinematical region of nearly opposite jet pair imbalances .
to proceed with quantitative estimates , one needs a model for the non - perturbative two - parton distributions in a proton . a priori , we know next to nothing about them .
the first natural step to take is an _ approximation of independent partons_. it allows one to relate @xmath37gpd with known objects , namely @xcite [ eq : model ] [ eq : dgg ] _ [ 2]d(x_1,x_2 , q_1 ^ 2,q_2 ^ 2 ; ) g(x_1,q_1 ^ 2;^2 ) g(x_2 , q_2 ^ 2;^2 ) . here @xmath123 is the non - forward parton correlator ( known as generalized parton distribution , gpd ) that determines , e.g. , hard vector meson production at hera and which enter in our case in the diagonal kinematics in @xmath15 ( @xmath124 , see fig .
[ figgpd ] ) .
the modeling by is not perfect .
first of all , it does not respect an obvious restriction @xmath125 .
so , @xmath7 have to be taken not too large ( say , @xmath126 ) .
actually , @xmath7 must be taken even smaller .
the gpd in fig .
[ figgpd ] is an elastic amplitude , while the corresponding block in the dpi represents the inclusive cross section ( the cut - through amplitude ) . for the analogy to hold
, the interaction amplitude has to be close to imaginary .
this condition calls for @xmath127 . on the other hand
, @xmath7 should not be _ too small _ to stay away from the region of the regge - gribov phenomena where there are serious reasons for parton correlations to be present at the non - perturbative level ( see discussion in @xcite ) .
thus , we fix the domain of applicability of the model for @xmath8 as @xmath128 .
the gpd , in its turn , can be modeled as [ eq : gdf ] g(x_1,q_1 ^
2;^2 ) d(x_1 , q_1 ^ 2 ) f_2g(^2 ) , with @xmath129 the usual one - parton distribution determining dis structure functions and @xmath130 the so - called two - gluon form factor of the hadron .
the latter is a non - perturbative object ; it falls fast with the `` momentum transfer '' @xmath50 .
this form factor can be parametrized differently .
for example ,
by a dipole formula [ eq : fdipole ] f_2g(^2 ) = ( 1 + ) ^-2 , where an effective parameter @xmath131 extracted from the fnal and hera @xmath132 exclusive photoproduction data lies in the ballpark of @xmath133 and decreases with further decrease of @xmath15 @xcite .
substituting into gives [ eq : seff1 ] _
^-1 = f_gg^4(^2 ) = , _ 32 mb .
it is about a factor of two larger than the value measured by the tevatron experiments @xcite .
pqcd induced parton correlations ( 2 dpi processes ) are capable of explaining this discrepancy .
turning to the 2 term , we neglect a mild logarithmic @xmath10-dependence of @xmath134}d$ ] in and use the model for @xmath135}d$ ] to obtain ^-1 73 ^ -1 , where we substituted the value of the integral ( cf . )
@xmath136 we will parametrize the result in terms of the ratio [ eq : rdef ] r = . for the effective interaction area , [ eq : sfull ] _
^-1 = _ 4 ^ -1 + _ 3 ^ -1 , we then have [ eq : smodel ] _
= 11+r 11+r ( the phenomenological value @xmath137 was used ) . within the framework of the np two - parton correlations model ,
there is but one free parameter @xmath71 .
the dpi theory applies to various processes and holds in a range of energies and different kinematical regions . therefore ,
having fixed the @xmath71 value , say , from the tevatron data , one can consider all other applications ( in particular , to lhc processes ) as parameter - free theoretical predictions .
in numerical calculations we used the grv92 parametrisation of gluon and quark parton distributions in the proton @xcite .
we have checked that using more advanced grv98 and cteq6l parametrisations does not change the numerical results .
the explicit grv92 parametrisation is speed efficient and allows one to start the pt evolution from rather small virtuality scales .
the combination @xmath138 was used as the lower cutoff for logarithmic transverse momentum integrals in the parton evolution , which induced a mild ( logarithmic ) @xmath10-dependence on top of the relevant power of the two - gluon form factor @xmath139 . to quantify the role of the 2 dpi subprocesses , we calculated the ratio @xmath48 in the kinematical region @xmath140 for tevatron ( @xmath141 ) and lhc energies ( @xmath142 ) .
we chose to consider three types of ensembles of colliding partons : 1 .
@xmath143 quark and three gluons which is relevant for `` photon plus 3 jets '' cdf and d0 experiments , 2 .
four gluons ( two pairs of hadron jets ) , 3 .
@xmath144 plus two gluons , illustrating @xmath145 production .
4 . @xmath144 plus @xmath146 , corresponding to the @xmath147 channel . in fig .
[ figcdf10 ] we show the profile of the 2 to 2 ratio @xmath48 for the @xmath148jets process in the kinematical domain of the cdf experiment @xcite .
the calculation was performed for the dominant `` compton scattering '' channel of the photon production : @xmath149 .
the longitudinal momentum fractions of two gluons producing second pair of jets are @xmath23 and @xmath150 .
the typical transverse momenta were taken to be @xmath151 for the jet pair , and @xmath152 for the photon
jet system . in fig .
[ figcdf10 ] @xmath48 is displayed as a function of rapidities of the photon
jet , @xmath153 , and the 2-jet system , @xmath154 .
we observe that the enhancement factor lies in the ballpark of @xmath155 .
processed through eq .
, it translates into @xmath156 mb .
this expectation has to be compared with the cdf finding @xmath157 .
a recent reanalysis of the cdf data points at an even small value : @xmath158 , @xcite . in our previous report
@xcite we have included a plot of the @xmath15-dependence of @xmath48 for central production at tevatron and lhc energies .
that plot turned out to be confusing : a rather sharp @xmath15-dependence it has demonstrated seemed to contradict the cdf findings of approximate constancy of @xmath47 .
in fact , this variance with @xmath15 was in the major part resulting from the kinematical link between @xmath15 and @xmath16 for a given collision energy ( @xmath159 ) . the results of numerical calculation for a fixed hardness @xmath16 shown in fig .
[ figcdf10 ] for the cdf kinematics exhibit a very mild @xmath15-dependence of the @xmath48 factor and thus of the @xmath47 .
the ratio @xmath48 is practically constant in the kinematical domain of the d0 experiment on photon+3 jets production @xcite and is very similar to that of the cdf experiment shown above in fig .
[ figcdf10 ] .
so , for the d0 kinematics we instead display in figs .
[ figd0pt1 ] , [ figd0pt2 ] the enhancement factor @xmath160 in dependence on @xmath161 of the secondary jet pair for photon transverse momenta 10 , 20 , 30 , 50 , 70 , and 90 ( from bottom to top ) .
central rapidity photon+3 jets production in @xmath162(@xmath163)gluon collisions in the d0 kinematics .
, scaledwidth=40.0% ] same as fig .
[ figd0pt1 ] for @xmath164 .
, scaledwidth=40.0% ] the corresponding prediction for @xmath47 is shown in fig .
[ figd0seff ] in comparison with the d0 findings .
@xmath47 as a function of the hardness of the second jet in the kinematics of the d0 experiment for @xmath165.,scaledwidth=95.0% ] both the absolute value and ( a hint at ) the @xmath161-trend look satisfactory . in fig .
[ figlhcgg05 ] we show the 2 to 2 ratio for production of two pairs of back - to - back jets with transverse momenta @xmath166 produced in collision of gluons at the lhc energy @xmath167 7 tev .
dependence on the hardness parameters of the dpi process of double gluon
gluon collisions is illustrated in fig .
[ figjet ] . for the sake of illustration ,
we have chosen the value of the pt cutoff parameter @xmath168 , and calculated the enhancement factor @xmath169 for five values of the transverse momenta of the jets in one pair , @xmath170 20 , 40 , 60 , 80 , 100 .
[ figjet ] demonstrates its dependence on the transverse momenta of jets in the second pair , @xmath171 .
@xmath169 for two dijets at lhc : @xmath170 20 , 40 , 60 , 80 , 100 ( from bottom to top).,scaledwidth=90.0% ] we observe that within the chosen range , @xmath48 increases by about 1525% with increase of the hardness of one of the jet pairs .
this corresponds to approximately 10% drop in @xmath47 . finally , in fig .
[ figlhcwjj05 ] we show the rapidity profile of the @xmath48 ratio for the process of production of the vector boson , @xmath172 , accompanied by an additional pair of ( nearly back - to - back ) jets with transverse momenta @xmath173 produces in a gluon gluon collision .
it is interesting to notice that the effect of perturbatively induced parton
parton correlations is maximal for equal rapidities of the @xmath2 and the jet pair , and diminishes when they separate .
this feature is more pronounced when the cutoff parameter @xmath71 is taken larger , so that the pt correlation becomes smaller and , at the same time , exhibits a stronger rapidity dependence .
the recent atlas study @xcite reported for this process the value @xmath174 mb which is consistent with the expected enhancement due to contribution of the 2 dpi channel , see .
[ figseffwjj ] shows a slight variation of @xmath47 with the jet scale in @xmath175 production .
the dependence of the enhancement factor @xmath169 on the @xmath176 parameter is shown for the typical kinematics of the cdf photon+3 jets experiment in fig .
[ figjets50q0md1 ] , and for central production of two pairs of @xmath177 jets at the lhc ( @xmath178 ) in fig .
[ figjets50q0md2 ] .
in the previous paper @xcite we have analyzed the perturbative correlation that arises due to 2 splitting of the parton in one of the colliding hadrons and derived the corresponding expressions in the leading collinear approximation of the pqcd .
here we presented results of numerical evaluation of this contribution to the dpi cross section measured at the tevatron and found the theoretical results to be consistent with the data for the value of the model parameter @xmath179 . with
@xmath71 fixed , theoretical expectations for certain exemplary dpi processes at lhc energies become parameter - free predictions
. theoretical derivation of the effective interaction area @xmath47 ( `` effective cross section '' ) relied on certain assumptions and approximations .
our approach to perturbative qcd effects in dpi developed in @xcite was essentially probabilistic .
in particular , we did not discuss the issue of possible interference between 2 and 2 two - parton amplitudes .
one can argue that such eventuality should be strongly suppressed .
indeed , spatial properties of accompanying radiation produced by so different configurations make them unlikely to interfere , since in the 2 mechanism a typical transverse distance between two partons from the hadron w.f .
is of order of the hadron size , while in the 2 case it is much smaller and is determined by a hard scale .
moreover , we disregarded potential contributions from non - diagonal interference diagrams that are due to crosstalk between partons in the amplitude and the amplitude conjugated .
such contributions are absent in 2 collisions but emerge specifically in the 2 dpi process in which the two partons from the hadron wave function are relatively close to one another in the impact parameter plane @xcite .
our prediction for @xmath47 was obtained as the ratio of the dpi and spi cross sections derived in the leading collinear approximation of pqcd ( lla ) .
this means that evolution of perturbative parton cascades was treated at the one - loop level , and the matrix elements of hard parton interactions treated in the born approximation . in the dpi problem ,
the subleading non - logarithmic corrections to the lla are bound to be sizable .
indeed , when deriving the total dpi cross section within the lla accuracy , one integrates the differential distribution over jet imbalances , @xmath180 , up to the scale given by the transverse momenta of the jets , @xmath44 . in reality ,
due to experimental cuts that are imposed in order to extract jets in the _ back - to - back _ kinematics the true hard scale of the dpi cross section is lower .
being formally a subleading @xmath181 correction , it will affect both the 2 and 2 cross sections . which way
the subleading pqcd effects will change the ratio is so far unknown . to establish the true hard scales of the parton distributions entering the dpi cross section formula
, one has to carry out the nlla analysis which would include taking into consideration concrete details of the jet finding algorithms employed in the experimental setup .
finally , our prediction for the dpi cross sections was based on a model assumption of the absence of np two - parton correlations in the proton .
this assumption is arbitrary .
one routinely makes it for the lack of any firsthand information about such correlations .
in @xcite we have pointed out a source of genuine non - perturbative two - parton correlations that should come onto the stage for very small @xmath15 values , @xmath182 , and estimated its magnitude via inelastic diffraction in the framework of the regge gribov picture of high energy hadron interactions . also it was argued in @xcite that strong quark antiquark correlations may arise from dynamical chiral symmetry breaking . in order to be able to reliably extract the dpi physics
, one has to learn how to theoretically predict parton collision processes with production of two hard systems ( four jets in particular ) .
this is the dominant channel , and it is only in the back - to - back kinematics that the 2 and 2 dpi processes become competitive with it . among first subleading pqcd corrections to the amplitude , there is a loop graph that looks like two - by - two parton collision .
this resemblance is deceiving though . unlike the 2 and 2 contributions , this specific correction does not depend on the spatial distribution of partons in the proton ( information encoded in @xmath47 ) , is not power enhanced in the region of small transverse momenta of hard systems , and therefore does not belong to the dpi mechanism @xcite . treating
the amplitude at the one - loop level corresponds to the two - loop accuracy for the cross section . until this accuracy is achieved ,
the values of @xmath47 extracted by experiments should be considered as tentative .
our first conclusion is that in the kinematical region explored by the tevatron experiments , the @xmath15-dependence of @xmath47 turns out to be rather mild .
this by no means implies , however , that @xmath47 can be looked upon as any sort of a universal number . on the contrary , we see that the presence of the perturbative correlation due to the 2 dpi mechanism results in dependence of @xmath47 not only on the parton momentum fractions @xmath7 and on the hardness parameters , but also on the type of the dpi process . for example , in the case of golden dpi channel of production of two same sign @xmath2 bosons @xcite the discussed mechanism leads to expectation of significantly larger @xmath47 than for , say , @xmath2 plus two jets process .
indeed , the comparison of the values of @xmath48 for central production of two gluon jet pairs , @xmath175 and @xmath183 ( with jet transverse momenta @xmath184 ) , gives ( @xmath185 ) .
rl r(jj+jj ) & = 1.18 ( 0.81 ) + r(w+jj ) & = 0.75 ( 0.45 ) + r(w^+w^+ ) & = 0.49 ( 0.26 ) } q_0 ^ 2=0.5(1.0 ) ^2 . as a result of the varying magnitude of the perturbative correlation
, the effective interaction areas @xmath47 come out to be significantly different for the three processes : rl jj+jj : & _ = 14.6 17.6 , + w+jj : & _ = 18.3 22.0 , + w^+w^+ : & _ = 21.5 25.4 .
the smaller value for each effective interaction area corresponds to more developed perturbative parton cascades than the larger one ( @xmath168 versus @xmath186 ) .
contrary to the @xmath183 channel , the _
double drell - yan process _ favors the 2 mechanism , @xmath187 . as a result
, the effective interaction area here turns out to be even smaller . for central production of two @xmath188 bosons at @xmath189
we get r(zz ) & = 1.03 ( 0.73 ) , zz : _ & = 15.9 18.5
. an important feature of the 2 mechanism is its dependence on the hardness of the process . with increase of @xmath44 ,
the 2 to 2 ratio @xmath48 should increase rather fast thus pushing @xmath47 to smaller values . at the same time , with decrease of the @xmath161 of the jets this contribution decreases .
as we have seen above , such a trend is consistent with the d0 data for @xmath190 . by pushing the hardness scales down to @xmath191 ,
one enters the domain of the physics of _
minijets_. here one should have @xmath192 25 mb for @xmath193 , the much larger value than the @xmath194-independent @xmath47 which was assumed in the monte carlo models like pythia for a long time . it would be interesting to implement in the mc models a more realistic account of mpi in which @xmath195 would decrease with increase of @xmath161 .
acknowledgements : this research was supported by the united states department of energy and the binational science foundation .
the proof of the sum rule involves two types of graphs , see fig .
[ momgraph ] .
consider first two partons 1 and 2 whose parents are taken from the hadron wave function at some scale @xmath53 that separates np and pt stages .
distribution of each parton then independently evolves up to the hard scale @xmath16 according to the standard pqcd rules .
single parton distribution is presented in the form of convolution d^f_h(x , q^2 ) = _ b _ x^1 ^f_i ( , q^2;2 ) w^i_h(y;2 ) of the np input function @xmath196 with the parton evolution function @xmath197 obeying the initial condition ^f_i(x,2;2 ) = ^f_i ( 1-x ) .
applying the momentum integral to the single parton distribution @xmath198 gives @xmath199 with @xmath200 the average energy fraction of the proton carried by the initial parton @xmath201 ( parent of @xmath66 ) at the scale @xmath53 , @xmath202 due to the momentum sum rule for the parton wave function of the proton , summing over parton species @xmath201 produces the total energy carried by _ all initial partons _ but the parent of the second registered parton @xmath65 : @xmath203 the parent parton energy @xmath204 is not observable .
what is fixed by the measurement is @xmath23 , while @xmath204 is being integrated over : [ eq : sumnp ] where we have introduced an average conditional parent parton energy @xmath205 which depends on @xmath23 and @xmath16 .
it is important to notice that this quantity depends also on the unphysical scale @xmath53 which separates the domains of pt and np description ( unlike the physical inclusive parton distribution in the proton @xmath129 on the r.h.s . of ) .
with @xmath212 . combining and , we get a full logarithmic scale derivative of the product of the d - functions : @xmath213 .\ ] ] now we integrate over the intermediate virtuality and make use of the boundary conditions , @xmath214 with @xmath215 the np input parton distribution .
we obtain [ eq : pfin ] we conclude that for consistency of the dpi picture , the perturbative parton correlation ( and thus the 2 subprocesses ) should be taken into full consideration at that very moment when one allows distributions of partons picked from the hadron wave function to evolve with the hard scale(s ) .
a. del fabbro and d. treleani , phys .
d * 61 * , 077502 ( 2000 ) [ arxiv : hep - ph/9911358 ] ; phys .
d * 63 * , 057901 ( 2001 ) [ arxiv : hep - ph/0005273 ] ; a. accardi and d. treleani , phys .
d * 63 * , 116002 ( 2001 ) [ arxiv : hep - ph/0009234 ] .
b. blok , yu . dokshitzer , l. frankfurt and m. strikman , phys .
d * 83 * , 071501 ( 2011 ) [ arxiv:1009.2714 [ hep - ph ] ] .
b. blok , yu . dokshitser , l. frankfurt and m. strikman , eur .
j. c * 72 * , 1963 ( 2012 ) [ arxiv:1106.5533 [ hep - ph ] ] . for
pythia summary see t. sjostrand and p.z .
skands , eur .
j. c * 39 * , 129 ( 2005 ) [ arxiv : hep - ph/0408302 ] .
abazov _ et al . _
[ d0 collaboration ] , phys .
d * 81 * , 052012 ( 2010 ) .
abazov _ et al . _
[ d0 collaboration ] , phys .
d * 83 * , 052008 ( 2011 ) .
proceedings of the 1st international workshop on multiple partonic interactions at the lhc verlug deutsches elektronen - synchrotron , 2010 .
p. bartalini and l. fano , arxiv:1103.6201 [ hep - ex ] .
bukhvostov , e.a .
kuraev and l.n .
lipatov , sov .
j. nucl .
phys . * 38 * , 263 ( 1983 ) [ yad .
38 * , 439 ( 1983 ) ] .
bukhvostov , e.a .
kuraev , l.n .
lipatov and g.v .
frolov , jetp lett . * 41 * , 92 ( 1985 ) [ pisma zh .
fiz . * 41 * , 77 ( 1985 ) ] .
bukhvostov and g.v .
frolov , yad .
* 45 * , 1136 ( 1987 ) .
gaunt , c .- h .
kom , a. kulesza and w.j .
stirling , eur .
j. c * 69 * ( 2010 ) 53 [ arxiv:1003.3953 [ hep - ph ] ] .
dokshitzer sov .
jetp * 46 * , 641 ( 1977 ) ; + yu.l .
dokshitzer , v.a .
khoze , a.h .
mueller and s.i .
troian , `` basics of perturbative qcd '' , ed .
j. tran thanh van , editions frontires , gif - sur - yvette , 1991 | we examine the role played in double parton interactions ( dpi ) by the parton parton correlations originating from perturbative qcd parton splittings . also presented
are the results of the numerical analysis of the integrated dpi cross sections at tevatron and lhc energies . to obtain the numerical results the knowledge of the single - parton gpds gained by the hera experiments was used to construct the non - perturbative input for _ generalized double parton distributions_. the perturbative two - parton correlations induced by three - parton interactions contribute significantly to resolution of the longstanding puzzle of an excess of multi - jet production events in the back - to - back kinematics observed at the tevatron . |
recently , there has been renewed interest in the long - standing theoretical possibility that supermassive stars ( smss ) , with masses of @xmath9@xmath10m@xmath1 inhabited the early universe ( see e.g. , volonteri 2012 ) and their possible fates ( e.g. , iben 1963 ; fowler & hoyle 1964 ; appenzeller & fricke 1972 ; shapiro & teukolsky 1979 ; bond et al . 1984 ; fuller et al .
1986 ) .
one of the main motivations for their study comes from observations of quasars at @xmath11 @xmath12@xmath13 which are inferred to be powered by black holes ( bhs ) with masses exceeding @xmath14m@xmath1 ( e.g. , willott et al . 2003 ; fan et al .
2006 ; mortlock et al .
2011 ) . given the short time ( @xmath15myr ) available for such massive bhs to grow via accretion from their initial seed masses , as derived from the most recent cosmological parameters inferred by e.g. , the _ wilkinson microwave anisotropy probe _
( komatsu et al . 2011 ) , and the suppression of bh growth due to the strong radiative feedback from both stars ( e.g. , whalen et al . 2004 ; wise & abel 2007 ; oshea & norman 2008 ) and the bhs themselves ( e.g. , pelupessy et al . 2007 ; alvarez et al . 2009 ; milosavljevi ' c et al .
2009 ; jeon et al .
2012 ; park & ricotti 2012 ) , it now appears more likely than ever that the seeds of the most massive early bhs must have been quite massive themselves ( e.g. , @xmath16m@xmath1 ; see johnson et al .
2012a ; also e.g. , shapiro 2005 ; volonteri & rees 2006 ; natarajan & volonteri 2012 ) .
whereas the majority of the first , population ( pop ) iii stars may have had masses of @xmath17@xmath18m@xmath1 ( e.g. , abel et al . 2002 ; bromm & larson 2004 ; yoshida et al . 2008 ; greif et al .
2011 ) , the best candidates for the seeds of smbhs are thus much more massive ( and rare ) supermassive primordial stars . an additional , and independent , reason to consider smss in the early universe is that the conditions required for their formation are now thought to be realized much more often than was previously assumed .
the most widely discussed avenue for the formation of smss is via the direct gravitational collapse of hot ( @xmath19k ) primordial gas in so - called atomic cooling dark matter ( dm ) halos at @xmath20 ( e.g. , bromm & loeb 2003 ; begelman et al .
2006 ; lodato & natarajan 2006 ; spaans & silk 2006 ; regan & haehnelt 2009 ; choi et al . 2013
; latif et al .
2013a , b ) . in this scenario
, the gas in the protogalaxy remains at the virial temperature of @xmath21 @xmath22k because h@xmath23 molecules have been photodissociated by the lyman - werner ( lw ) background , leading to the rapid formation of smss via the accretion of gas at rates @xmath24@xmath25 times higher than in the formation of most pop iii stars from h@xmath26-cooled gas .
the flux of radiation required to keep the gas h@xmath23-free depends on its spectrum , with lower fluxes required if it is produced by metal - enriched stars instead of pop iii stars ( e.g. , shang et al . 2010 ) .
recent work by independent groups has shown that pop ii star - forming galaxies in the early universe are able to produce sufficient h@xmath23-dissociating radiation to prevent the cooling of primordial gas in a substantial fraction of atomic cooling halos , thereby leading to the seeding of these halos with smss that can collapse into bhs ( see dijkstra et al .
2008 ; agarwal et al .
2012 ; petri et al . 2012 ; johnson et al .
indeed , agarwal et al .
( 2012 , 2013 ) find that a large fraction of the smbhs in the centers of galaxies today may have been seeded by smss .
strengthening these conclusions are other recent results which suggest that lower lw fluxes may be required for the formation of smss , due to a reduced role of h@xmath23 self - shielding ( wolcott - green et al . 2011 ) and the presence of significant turbulence or magnetic fields ( van borm & spaans 2013 ) .
complementary studies have been undertaken to understand the growth and evolution of smss , as well .
modeling the growth of accreting protostars with masses up to @xmath27m@xmath1 , hosokawa et al .
( 2012 ) have shown that they emit little high energy radiation that could halt their continued accretion , and inayoshi et al .
( 2013 ) have shown that pulsational instabilities are likewise unable to halt their growth .
johnson et al . ( 2012b ) modeled the growth of smss to much higher masses and showed that , even if they are able to emit the copious ionizing radiation characteristic of main sequence pop iii stars , which may occur once the accretion rate becomes sufficiently low ( schleicher et al .
2013 ) , radiative feedback is not able to stop their growth up to at least @xmath28m@xmath1 . at the highest accretion rates expected for these objects ( @xmath29m@xmath1 yr@xmath30 ; wise et al . 2008 ; shang et al . 2010 ; johnson et al .
2011 ) , the masses of primordial smss are only limited by the @xmath31myr that they have to accrete gas before they collapse to bhs ( begelman 2010 ) . whereas the majority of smss are expected to collapse to black holes with little or no associated explosion ( e.g. , fryer & heger 2011 ; see also fuller & shi 1998 ; linke et al .
2001 ) , it is possible that some fraction instead explode as extremely energetic supernovae ( sne ; e.g. , fuller et al .
1986 ; montero et al . 2012 ;
whalen et al . 2012a , 2013d ) .
in particular , heger et al .
( 2013 ) have found from stellar evolution calculations including post - newtonian corrections to gravity that smss with masses in a narrow range around @xmath32m@xmath1 end their lives as extraordinarily luminous sne . with energies of almost @xmath33erg ,
these thermonuclear explosions are among the most energetic in the history of the universe . here
we expand on the radiation hydrodynamics simulations presented by whalen et al .
( 2012a ) and simulate the long - term evolution of a sms sn in its cosmological environment , in order to show how these gargantuan explosions impact both the formation of the first galaxies and the chemical signature of the first stars . in the next section , we describe the multi - scale simulations that we have carried out to model the evolution of the explosion from its breakout from the surface of the star to the propagation of the blast wave into the intergalactic medium ( igm ) . in section 3
, we present our results on the energetics and dynamics of the explosion , as well as on metal enrichment and second - generation star formation . in section 4 ,
we conclude with a brief discussion of our results .
here we describe the two simulations that we have carried out .
the first is a 1-d radiation hydrodynamics calculation using the los alamos national laboratory rage code ( gittings et al .
2008 ) which allows to track the propagation of the sn blast wave out to several parsecs from the explosion site , deep within the host atomic cooling halo .
for the second , we map the results of the first into a 3-d cosmological simulation using the gadget hydrodynamics code ( springel et al . 2001 ;
springel & hernquist 2002 ) .
our use of these two simulation codes for the phases of the sn in which they are most well - suited to accurately model the explosion , from small ( au ) scales to large ( kpc ) scales , constitutes a significant improvement over previous cosmological simulations of sn feedback .
= 7.in = 8.5 cm for the sms progenitor of the sn we adopt the @xmath34m@xmath1 stellar model described in whalen et al .
( 2012a ) , which was evolved until the onset of explosion , using the _ kepler _ code ( weaver et al .
1978 ; woosley et al . 2002 ) .
the explosion was then followed using _ kepler _ and was confirmed ( chen et al . , 2013 ) using the castro code ( almgren et al .
the explosion completely disrupts the star and yields an explosion energy of 7.7 @xmath35 10@xmath36 erg ( heger et al .
the rage code is then used to simulate the sn from the breakout of the shock from the surface of the star until the blast wave has propagated through a circumstellar medium with a density @xmath37@xmath38 ( and a density profile @xmath39 @xmath40 ) out to several parsecs from the explosion site . up to this point , the stellar evolution calculation and the simulation of the early phases of the sn are the same as described in whalen et al .
( 2012a ) , to which
we refer the reader for more detailed discussion of the calclulations ( see also frey et al .
2013 ) .
the blue curves in figure 1 show the velocity ( left panel ) and density ( right panel ) profiles of the sn ejecta at the end of the rage simulation , at which point the shock has propagated out to @xmath41pc from the explosion site .
the @xmath34m@xmath1 of ejecta , @xmath42m@xmath1 of which is heavy elements produced during the evolution and explosion of the progenitor , are traveling at almost @xmath43km@xmath44s@xmath30 outward from the center of the host atomic cooling halo .
it is from this point that we map these velocity and density profiles into a self - consistently evolved atomic cooling halo in a much larger cosmological volume . to simulate the subsequent evolution of the sn blast wave in the appropriate cosmological environment , we map the velocity and density profiles obtained from the smaller - scale rage simulation into the center of a @xmath45m@xmath1 atomic cooling dm halo , the type of which is expected to host the formation of smss in the early universe .
the halo is identified in a @xmath46mpc@xmath47 ( comoving ) cosmological volume which has been evolved from @xmath48 down to @xmath49 under the influence of a uniform , elevated h@xmath26-dissociating ( lyman - werner ; lw ) radiation field , which prevents the gas from cooling and is assumed to lead to formation of a single sms .
further details of the cosmological simulation up to this point are described in johnson et al .
( 2011 ) , who considered the impact of the alternative end state of such a sms , a rapidly accreting bh . to map
the output of the ( eulerian ) rage simulation into the ( lagrangian ) smoothed particle hydrodynamics ( sph ) gadget simulation , we assigned the central @xmath50 sph particles , constituting the @xmath34m@xmath1 of gas within @xmath51pc of the densest particle in the halo , outward ( radial ) velocities and hydrogen number densities so as to match those from the rage simulation .
the fits that we obtain are shown in fig . 1 , with the inner @xmath34m@xmath1 in sph particles constituting the ejecta denoted by orange triangles and the unpertured particles residing in the outskirts of the halo denoted by black circles . whereas we fit the velocity profile very well , due to the mapping from an eulerian to lagrangian code the density profile is somewhat noisier .
nonetheless , the basic features of the density profile containing the vast majority of the mass are represented , and the overall energy and momentum are also well - matched . from the initial conditions shown in fig .
1 we restart the cosmological sph simulation , the results of which we present in the next section .
beyond mapping into it the blast profile of the sms sn , we have chosen to leave the gas in the host halo otherwise unchanged .
we have made this simplifying choice , in light of the large uncertainties in the radiative output of rapidly accreting smss , due to which it is unclear how the radiation emitted during the brief ( @xmath52myr ) lifetime of the star will impact the medium within the host halo . even if the star emits copious ionizing radiation , as main sequence pop iii stars are expected to do
, it may be that the h ii region created by the star is confined to the innermost regions of the halo ( johnson et al .
2012 ; see also hosokawa et al .
2012 on the possibility of even less energetic radiation being emitted during the protostellar phase ) .
it is possible , on the other hand , that ionizing radiation is able to escape out into the halo , if the accretion flow is highly anisotropic ( e.g. , due to the presence of an accretion disk ; e.g. , mckee & tan 2008 ) , or if accretion is intermittent ( e.g. , clark et al .
2011 ; smith et al . 2011 ; vorobyov et al .
2013 ) , in which case radiation could break out during periods of reduced accretion .
whereas the small - scale structure of the interstellar gas in the simulation we present here is subject to the limited resolution of the cosmological simulation into which we place the expanding blast wave , we note that at sub - resolution scales ( @xmath53pc ) it is possible that a substantial amount of energy in the explosion is radiated away ( e.g. , kitayama & yoshida 2005 ; whalen et al .
2008 ; de souza et al . 2011 ; vasiliev et al .
. we shall address how such additional radiative losses would affect the dynamics of the blast wave and the metal enrichment of the host halo and igm in future work .
here we present the results of our cosmological hydrodynamics simulation , with particular attention paid to the dynamics of the expanding blast wave and to the enrichment of the igm by the metal - rich sn ejecta .
the injection of almost @xmath54
erg at a single explosion site has a dramatic impact on the host halo .
figure 2 shows the properties of the gas in the vicinity of the halo , as a function of the distance from the explosion site at its center .
as the left - most panels show , the blast results in the complete evacuation of high - density ( @xmath55 @xmath56@xmath38 ) gas within @xmath57myr , with the material overtaken by the shock being carried out beyond the virial radius of the halo ( at @xmath58pc ) at up to @xmath27km@xmath44s@xmath30 .
this material is also shock heated to temperatures up to @xmath59k , resulting in its almost complete ionization , as shown in the right - panels .
the gas , however , rapidly cools due to inverse compton scattering of cmb photons , h and he atomic line emission , and bremsstrahlung , as also found ( using the same gadget code ) in the less - energetic ( @xmath60erg ) pair instability supernova ( psn ) explosion in a @xmath61m@xmath62 dm halo simulated by greif et al .
( 2007 ) , as well as in the 1-d calculation of a very energetic ( @xmath63erg ) pop iii sn in a slightly less massive ( @xmath64m@xmath1 ) dm halo presented in kitayama & yoshida ( 2005 ) . as shown in figure 3 ,
most ( @xmath65 ) of the @xmath66erg initially in the blast is radiated away via these processes within @xmath22yr .
nevertheless , the momentum of the blast is conserved and the shock continues to propagate into the igm , sweeping up the majority of the mass after @xmath67myr . by @xmath68myr , at which time @xmath69 of the energy has been radiated away , the shock has propagated out to @xmath70@xmath57kpc and has swept up almost @xmath71m@xmath1 . figure 4 shows the properties of the gas in the vicinity of the explosion site within a @xmath72pc ( comoving ) slice of the cosmological volume , at @xmath73myr , @xmath74myr and @xmath75myr after the explosion of the sms . comparing the radial velocity field ( second column from the left ) to the cosmological density field ( far left column ) ,
it is clear that the blast wave propagates most rapidly into the low - density voids while its progress is halted in the direction of the high - density filaments , at the intersection of which lies the host halo .
figs . 2 and 4
also show the same general trend that , at late times , the most strongly shock - heated and highest - velocity material is located behind the shock front several kpc from the explosion site .
the cooler material within the ( @xmath67kpc ) virial radius of the host halo is able to begin recollapsing after @xmath75myr . as we discuss next , this gas is likely to form second - generation stars that are enriched to fairly high metallicities .
= 8.7 cm = 8.6 cm = 7.1 in as is also expected for less energetic psne from pop iii stars with masses of @xmath76m@xmath1 ( e.g. , heger et al .
2003 ) , a large fraction of the ejecta from our sms sn consists of heavy elements .
in particular , @xmath77m@xmath1 of newly - synthesized metals are ejected in the explosion .
these heavy elements are mixed with the primordial gas , enriching it to relatively high metallicity .
figure 5 shows the average metallicity to which the gas in the vicinity of the explosion site is enriched by the ejecta , after @xmath78myr , @xmath74myr and @xmath79myr .
as the blast wave propagates outward into the low - density igm gas at greater and greater radii becomes enriched .
as metals are carried out of the host halo , metallicities at the smallest radii fall .
after @xmath79myr , however , the average metallicity of the gas out to almost @xmath74kpc ( physical ) is enriched to of the order of @xmath80z@xmath1 , and the densest gas which is recollapsing into the host halo is enriched to @xmath81z@xmath1 . as shown in figure 6 , by @xmath79myr the gas at the center of the sn remnant , shown in figure 7 , has recollapsed to densities of @xmath82@xmath38 , significantly more dense than the highest - density ( @xmath55 @xmath8310@xmath44@xmath38 ) gas that remained within the virial radius of the host halo @xmath84myr after the sn .
not only is the gas recollapsing after @xmath79myr , but , as shown in the right panel of fig .
6 , a portion of the sn ejecta is entrained in this gas .
if this ejecta is well - mixed with the dense primordial gas , then we expect its metallicity to be @xmath85z@xmath1 , as indicated in fig . 5 .
gas enriched to such a high metallicity is predicted to readily fragment into low - mass stars ( e.g. , bromm et al .
2001 ; santoro & shull 2006 ; schneider et al .
2006 ) , even in the presence of the elevated lw background radiation field expected in regions of the universe where smss form ( e.g. , omukai et al .
therefore , we expect that second - generation stars would form from this sms sn - enriched gas , and that a large fraction of these would have masses low enough ( @xmath86m@xmath1 ) that they could still be present in the galaxy today . while the nucleosynthetic signature of very massive pop iii pair - instability sne ( i.e. , @xmath87@xmath88m@xmath1 ) has yet to be uncovered in extremely metal - poor stars ( cayrel et al .
2004 ; beers & christlieb 2005 ; frebel et al 2005 ; lai et al 2008 ; joggerst et al 2010 ; joggerst & whalen 2011 ) , it may have been found in high - redshift damped lyman alpha absorbers ( cooke et al .
2011 ) , and a number of metal - poor stars in a recent extension to the segue survey have now been selected for spectroscopic followup on suspicion that they too may harbor this pattern ( ren et al .
furthermore , most stars forming in the ashes of very massive primordial sne were likely enriched to metallicities above those targeted by surveys of metal - poor stars to date ( karlsson et al .
2008 ) , and our simulations predict that second - generation stars formed from gas enriched by sms sne will have metallicities above @xmath80z@xmath1 , which are also above this threshold . we conclude that , while sms sne are almost certainly rare events , their chemical signature might well be found in some of the ancient but not very metal - poor stars inhabiting our galaxy today .
we have carried out a multi - scale simulation of the explosion of a @xmath89m@xmath1 sms , which , with an energy approaching @xmath33erg , is among the biggest explosions in the history of the universe . with our multi - code approach ,
we have captured self - consistently , and for the first time , the essential radiation hydrodynamic features of the explosion at early times and the interaction of the blast wave with its host protogalaxy and with the igm at later times . although the atomic cooling halos expected to host the formation of smss are at least two orders of magnitude more massive than the dm halos in which the first primordial stars are expected to form , we have found that these sms sne are energetic enough to completely evacuate them of dense gas .
the metal - enriched ejecta is dispersed well beyond the @xmath90kpc virial radius of the host halo , out to @xmath91@xmath74kpc into the low - density igm , after @xmath92myr . by this time
, @xmath93 of the kinetic energy in the explosion has been radiated away ; nonetheless , the expansion of the shock into the igm continues even at these late times , as the remaining kinetic energy is still comparable to that expected for a @xmath94m@xmath1 pop iii psn . because of the deep potential well of the halo and ongoing accretion from filaments , after @xmath79myr a fraction of the metal - enriched gas in the sn remnant , shown in fig .
7 , has recollapsed to high densities .
given the relatively high metallicity of this gas ( @xmath95z@xmath1),z@xmath1 ( e.g. , wise & abel 2008 ; greif et al .
2010 ; ritter et al . 2012 ; vasiliev et al . 2012
; see also wise et al .
2012 ; safranek - shrader et al .
incidentally , we note that greif et al .
( 2007 ) arrived at comparable results by tracking metal enrichment using a much smaller number of sph particles than we have used to track it here .
thus , we expect that our results with regard to metal enrichment should be broadly consistent with what would be found using other approaches .
] it is most likely that it will fragment vigorously and form a cluster of second - generation stars .
enriched to this metallicity , any of these stars which survive to the present - day would exhibit metallicities much higher than the most metal - poor stars that have been found in surveys of the galactic halo .
indeed , although sms sn are likely rare events , it is possible that their chemical signature could be found in very old , relatively high - metallicity pop ii stars which inhabit the milky way today ( or in present - day dwarf galaxies ; frebel & bromm 2012 ) . as we expect sms sne to produce very little @xmath96ni ( heger et al .
2013 ) , their chemical signature may be distinguishable from those of most very massive ( i.e. , @xmath87@xmath97m@xmath1 ) pop iii explosions ( psne ) , which can produce iron - group elements ( heger & woosley 2002 ) . similar to these psne , however , sms sne would not make any _ s_-process or _
r_-process contributions . finding the chemical signature of sms sne in low - mass , long - lived stars ( or in the igm
; see e.g. , cooke et al .
2011 ) would be one way of verifying that such exotic events occurred in the early universe .
another possibility is that these gargantuan explosions could be found in all - sky surveys such as those planned for the _ wide - field infrared survey telescope _ ( wfirst ) and the _ wide - field imaging surveyor for high - redshift _
( wish ) , as shown recently by whalen et al .
( 2012a ) .
given that smss are expected to form in regions subjected to a large flux of lw radiation from nearby ( within @xmath98kpc ; dijkstra et al .
2008 ; agarwal et al .
2012 ) star - forming galaxies , detection of sms sne would pinpoint locations on the sky where rapidly - forming first galaxies could be found in follow - up observations by the _ james webb space telescope_. indeed , these explosions are so large that the sn remnants they leave behind , as shown in fig . 7 , are likely to be many times larger than , and in fact are likely to envelop , any such neighboring star - forming galaxies . could later stages of sms sne be detected by other means
? they might appear in the radio at @xmath99 cm .
meiksin & whalen ( 2013 ) recently found that synchrotron emission from hypernovae in relatively dense media will be visible at @xmath99 cm to existing radio facilities such as evla and emerlin in addition to the _ square kilometer array _ ( ska ) . with their much higher energies and similar circumstellar densities
, sms sne may be much brighter in the radio and detectable in all - sky surveys in spite of their small numbers .
we are now calculating the radio signatures of sms sne in @xmath100 protogalaxies .
likewise , oh et al .
( 2003 ) and whalen et al .
( 2008 ) examined the potential imprint of pop iii sne on the cosmic microwave background ( cmb ) via the sunyayev - zeldovich ( sz ) effect .
they found that a population of @xmath87@xmath101m@xmath1 psne might impose excess power on the cmb on small scales , but with two caveats .
first , the explosions must occur in large h ii regions because the sn shock must still be hot by the time it encloses a large volume of cmb photons .
explosions in dense media dissipate too much heat to have an appreciable sz signature or upscatter many cmb photons .
second , although a population of pop iii sne might collectively impose features on the cmb individual remnants achieve radii just below the current resolution of _ atacama cosmology telescope _ or the _ south pole telescope_. our models explode in dense environments and at redshifts at which inverse compton cooling losses are lower than for the first sne , but they eventually reach radii that would allow them to be resolved by current instruments .
we are currently evaluating the sz signatures of sms sne .
this work was supported by the u.s .
department of energy through the lanl / ldrd program , and jlj acknowledges the support of a ldrd director s postdoctoral fellowship at los alamos national laboratory .
the rage and gadget simulations were carried out on the lanl institutional computing clusters pinto and mustang , respectively .
djw acknowledges support from the baden - wrttemberg - stiftung by contract research via the programme internationale spitzenforschung ii ( grant p- ls - spii/18 ) . ah and kc were supported by the us doe program for scientific discovery through advanced computing ( scidac ; de - fc02 - 09er41618 ) , by the us department of energy under grant de - fg02 - 87er40328 , by the joint institute for nuclear astrophysics ( jina ; nsf grant phy08 - 22648 and phy110 - 2511 ) . ah acknowledges support by an arc future fellowship ( ft120100363 ) and a monash university larkins fellowship .
kc was supported by a kitp / ucsb graduate fellowship and by a umn stanwood johnston fellowship .
the authors thank avery meiksin for helpful discussion .
work at lanl was done under the auspices of the national nuclear security administration of the u.s .
department of energy at los alamos national laboratory under contract no .
de - ac52 - 06na25396 .
abel , t. , bryan , g. l. , norman , m. l. 2002 , sci , 295 , 93 almgren , a. s. , et al .
2010 , apj , 715 , 1221 agarwal , b. , khochfar , s. , johnson , j. l. , neistein , e. , dalla vecchia , c. , livio , m. 2012 , mnras , 425 , 2854 agarwal , b. , davis , a. j. , khochfar , s. , natarajan , p. , dunlop , j. s. 2013 , mnras , submitted ( arxiv:1302.6996 ) alvarez , m. a. , wise , j. h. , abel , t. 2009 , apj , 701 , l133 appenzeller , i. , fricke , k. 1972 , a&a , 21 , 285 begelman m. c. 2010 , mnras , 402 , 673 begelman m. c. , volonteri m. , rees m. j. 2006 , mnras , 370 , 289 beers , t. c. , christlieb , n. 2005 , ara&a , 43 , 531 bond , j. r. , arnett , w. d. , carr , b. j. 1984 , apj , 280 , 825 bromm , v. , ferrara , a. , coppi , p. s. , larson , r. b. 2001 , mnras , 328 , 969 bromm , v. , loeb , a. 2003 , apj , 596 , 34 bromm v. , larson r. b. 2004 , ara&a , 42 , 79 cayrel , r. , et al .
2004 , a&a , 416 , 1117 choi , j .- h .
, shlosman , i. , begelman , m. c. 2013 , apj , submitted ( arxiv:1304.1369 ) clark , p. c. , glover , s. c. o. , smith , r. j. , greift , t. h. , klessen , r. s. , bromm , v. 2011 , sci , 331 , 1040 cooke , r. , pettini , m. , steidel , c. c. , rudie , g. c. , jogenson , r. a. 2011 , mnras , 412 , 1047 de souza , r. s. , rodrigues , l. f. s. , ishida , e. e. o. , opher , r. 2011 , mnras , 415 , 2969 de souza , r. s. , ishida , e. e. o. , johnson , j. l. , whalen , d. j. , mesinger , a. 2013 , mnras , submitted ( arxiv:1306.4984 ) dijkstra , m. , haiman , z. , mesinger , a. , wyithe , j. s. b. 2008 , mnras , 391 , 1961 fan , x. , et al . 2006 , aj , 131 , 1203 fowler , w. a. , hoyle , f. 1964 , apjs , 9 , 201 frebel , a. , et al .
2005 , nat , 434 , 871 frebel , a. , bromm , v. 2012 , apj , 759 , 115 frey , l. , even , w. , whalen , d. j. , et al .
2013 , apjs , 204 , 16 fryer , c. l. , heger , a. 2011 , an , 332 , 408 fuller , g. m. , woosley , s. e. , weaver , t. a. 1986 , apj , 307 , 675 fuller , g. m. , shi , x. 1998 , apj , 502 , l5 gittings , m. , et al .
2008 , cs&d , 1 , 5005 greif t. h. , johnson j. l. , bromm v. , klessen r. s. 2007 , apj , 670 , 1 greif t. h. , glover , s. c. o. , bromm , v. , klessen , r. s. 2010 , apj , 716 , 510 greif t. h. , springel , v. , white , s. d. m. , glover , s. c. o. , clark , p. c. , smith , r. j. , klessen , r. s. , bromm , v. 2011 , apj , 737 , 75 heger , a. , woosley , s. e. 2002 , apj , 567 , 532 heger , a. , et al .
2013 , in prep hosokawa , t. , omukai , k. , yorke , h. w. 2012 , apj , submitted ( arxiv:1203.2613 ) hummel , j. a. , pawlik , a. h. , milosavljevi ' c , m. , bromm , v. 2012 , apj , 755 , 72 iben , i. 1963 , apj , 138 , 1090 inayoshi , k. , omukai , k. 2012 , mnras , 422 , 2539 inayoshi , k. , hosokawa , t. , omukai , k. 2013 , mnras , accepted ( arxiv:1302.6065 ) jeon , m. , pawlik , a. h. , greif , t. h. , glover , s. c. o. , bromm , v. , milosavljevi ' c , m. , klessen , r. s. 2012 , apj , 754 , 34 joggerst , c. c. , whalen , d. j. 2011 , apj , 728 , 129 joggerst , c. c. , almgren , a. , bell , j. , heger , a. , whalen , d. j. , woosley , s. e. 2010 , apj , 709 , 11 johnson , j. l. , dalla vecchia , c. , khochfar , s. 2013 , mnras , 428 , 1857 johnson , j. l. , khochfar , s. , greif , t. h. , durier , f. 2011 , mnras , 410 , 919 johnson , j. l. , whalen , d. j. , fryer , c. l. , li , h. 2012b , apj , 750 , 66 johnson , j. l. , whalen , d. j. , li , h. , holz , d. e. 2012a , apj , submitted ( arxiv:1211.0548 ) karlsson , t. , johnson , j. l. , bromm , v. 2008 , apj , 679 , 6 kistler , m. d. , beacom , j. f. 2006 , phrvd , 74 , 063007 kitayama t. , yoshida n. 2005 , apj , 630 , 675 komatsu , e. , et al .
2011 , apjs , 192 , 18 lai , d. k. , bolte , m. , johnson , j. a. , lucatello , s. , heger , a. , woosley , s. e. 2008 , apj , 681 , 1524 latif , m. a. , schleicher , d. r. g. , schmidt , w. , niemeyer , j. 2013a , mnras , 430 , 588 latif , m. a. , schleicher , d. r. g. , schmidt , w. , niemeyer , j. 2013b , mnras , submitted ( arxiv:1304.0962 ) linke , f. , font , j. a. , janka , h .- t .
, mller , e. , papadopoulos , p. 2001 , a&a , 376 , 568 lodato , g. , natarajan , p. 2006 , mnras , 371 , 1813 mckee , c. f. , tan , j. c. 2008 , apj , 681 , 771 meiksin , a. , whalen , d. j. 2013 , mnras , 430 , 2854 milosavljevi ' c , m. , bromm , v. , couch , s. m. , oh , s. p. 2009
, apj , 698 , 766 montero , p. j. , janka , h .-
t . , m " uller , e. 2012,apj , 749 , 37 morlino , g. , blasi , p. , amato , e. 2009 , astropart .
phys . , 31 , 376 mortlock , d. j. , et al .
2011 , nat , 474 , 616 natarajan , p. , volonteri , m. 2012 , mnras , 422 , 2051 oh , s. p. , cooray , a. , kamionkowski , m. 2003 , mnras , 342 , l20 omukai , k. , schneider , r. , haiman , z. 2008 , apj , 686 , 801 oshea b. w. , norman m. l. 2008 , apj , 673 , 14 pan , t. , kasen , d. , loeb , a. 2012 , mnras , 422 , 2701 park , k. , ricotti , m. 2012 , apj , 747 , 9 pelupessy , f. i. , di matteo , t. , ciardi , b. 2007 , apj , 665 , 107119 petri , a. , ferrara , a. , salvaterra , r. 2012 , mnras , accepted ( arxiv:1202.3141 ) planck collaboration , a&a submitted ( arxiv:1303.5076 ) regan j. a. , haehnelt m. g. 2009 , mnras , 396 , 343 ren , j. , christlieb , n. , zhao , g. 2012 , raa , 12 , 1637 ritter , j. s. , safranek - shrader , c. , gnat , o. , milosavljevi ' c , m. , bromm , v. 2012 , apj , 761 , 56 safranek - shrader , c. , milosavljevi ' c , m. , bromm , v. 2013 , mnras , submitted ( arxiv:1307.1982 ) santoro , f. , shull , j. m. 2006 , apj , 643 , 26 scannapieco , e. , madau , p. , woosley , s. , heger , a. , ferrara , a. 2005 , apj , 633 , 1031 schleicher , d. r. g. , palla , f. , ferrara , a. , galli , d. , latif , m. 2013 , a&a , submitted ( arxiv:1305.5923 ) schneider , r. , omukai , k. , inoue , a. k. , ferrara , a. 2006 , mnras , 369 , 825 sethi , s. , haiman , z. , pandey , k. 2010 , apj , 721 , 615 shang , c. , bryan , g. l. , haiman , z. 2010 , mnras , 402 , 1249 shapiro , s. l. 2005 , apj , 620 , 59 shapiro , s. l. , teukolsky , s. a. 1979 , apj , 234 , l177 smith , r. j. , glover , s. c. o. , clark , p. c. , greif , t. h. , klessen , r. s. 2011 , mnras , 414 , 3633 spaans , m. , silk , j. 2006 , apj , 652 , 902 springel v. , yoshida n. , white s. d. m. , 2001 , newa , 6 , 79 springel v. , hernquist , l. 2002 , mnras , 333 , 649 tanaka , m. , moriya , t. j. , yoshida , n. , nomoto , k. 2012 , mnras , 422 , 2675 tanaka , m. , moriya , t. j. , yoshida , n. 2013 , mnras , submitted ( arxiv:1306.3743 ) van borm , c. , spaans , m. 2013 , a&a , submitted ( arxiv:1304.4057 ) vasiliev , e. o. , vorobyov , e. i. , matvienko , e. e. , razoumov , a. o. , shchekinov , y. a. 2012 , arep , 56 , 895 volonteri , m. , rees , m. 2006 , apj , 650 , 669 volonteri , m. 2012 , sci , 337 , 544 volonteri , m. , begelman , m. c. 2010 , mnras , 409 , 1022 vorobyov , e. i. , desouza , a. l. , basu , s. 2013 , apj , submitted ( arxiv:1303.3622 ) weaver , t. a. , zimmerman , g. b. , woosley , s. e. 1978 , apj , 225 , 1021 whalen d. , van veelen b. , oshea b. w. , norman m. l. 2008 , apj , 682 , 49 whalen , d. j. , heger , a. , chen , k .- j .
, even , w. , fryer , c. l. , stiavelli , m. , xu , h. , joggerst , c. c. 2012a , apj , submitted ( arxiv:1211.1815 ) whalen , d. j. , abel , t. , norman , m. l. 2004 , apj , 610 , 14 whalen , d. j. , et al .
2013a , apj , 768 , 195 whalen , d. j. , et al .
2012b , apj , submitted ( arxiv:.1211.4979 ) whalen , d. j. , et al . 2013b , apj , 768 , 95 whalen , d. j. , et al .
2013c , apjl , 762 , 6 whalen , d. j. , et al .
2013d , apj , accepted ( arxiv:1305.6966 ) willott , c. j. , mclure , r. j. , jarvis , m. j. , 2003 , apj , 587 , l15 wise , j. h. , abel , t. 2007 , apj , 671 , 1559 wise , j. h. , turk , m. j. , abel , t. 2008 , apj , 682 , 745 wise , j. h. , abel , t. 2008 , apj , 685 , 40 wise , j. h. , turk , m. j. , norman , m. l. , abel , t. 2012 , apj , 745 , 50 wolcott - green , j. , haiman , z. , bryan , g. l. 2011 , mnras , 418 , 838 woosley , s. e. , heger , a. , weaver , t. a. 2002 , revmp , 74 , 1015 yamazaki , r. , kohri , k. , katagiri , h. 2009 , a&a , 495 , 9 yoshida , n. , omukai , k. , hernquist , l. 2008 , sci , 321 , 669 yuan , q. , yin , p. , bi , x. 2010 , arxiv:1010.1901 | supermassive primordial stars are expected to form in a small fraction of massive protogalaxies in the early universe , and are generally conceived of as the progenitors of the seeds of supermassive black holes ( bhs ) .
supermassive stars with masses of @xmath0m@xmath1 , however , have been found to explode and completely disrupt in a supernova ( sn ) with an energy of up to @xmath2erg instead of collapsing to a bh .
such events , @xmath3 times more energetic than typical sne today , would be among the biggest explosions in the history of the universe . here
we present a simulation of such a sn in two stages . using the rage radiation hydrodynamics code
we first evolve the explosion from an early stage through the breakout of the shock from the surface of the star until the blast wave has propagated out to several parsecs from the explosion site , which lies deep within an atomic cooling dark matter ( dm ) halo at @xmath4 .
then , using the gadget cosmological hydrodynamics code we evolve the explosion out to several kiloparsecs from the explosion site , far into the low - density intergalactic medium . the host dm halo , with a total mass of @xmath5m@xmath1 , much more massive than typical primordial star - forming halos , is completely evacuated of high density gas after @xmath6myr , although dense metal - enriched gas recollapses into the halo , where it will likely form second - generation stars with metallicities of @xmath7z@xmath1 after @xmath8myr . the chemical signature of supermassive star explosions may be found in such long - lived second - generation stars today . |
[ [ a. ] ] a. + + in @xcite the author investigated the asymptotic behaviour of solutions of an @xmath18system @xmath19 where the unperturbed system @xmath0 has a phase portrait like a _ linear centre _ and the perturbation @xmath20 as @xmath21 . in this paper
we give natural generalizations of this problem .
[ [ b. ] ] b. + + first of all we consider an @xmath22system @xmath23 where @xmath24 and @xmath25 for all @xmath26 .
we also assume that the phase portrait of each subsystem @xmath27 is similar to a _ linear centre _ , i.e. origin is a unique equilibrium point and the other trajectories are closed curves surrounding the origin . the phase portrait of the whole system ( 1.1 ) resembles the cartesian product @xmath28 where @xmath29 and @xmath17 denotes the @xmath3dimensional torus .
the first example of such a system is a linear system @xmath30 where all eigenvalues of matrix @xmath31 are pure imaginary and have only prime jordan blocks .
this system can be reduced by a linear transformation to a system of @xmath3 independent mathematical pendulums @xmath32 where all @xmath33 .
the next more general example is a system of @xmath3 independent duffing equations @xmath34 where every function @xmath35 does not destroy the centre defined by the linear part in each equation .
we often meet the same systems in different applications of mathematics , mechanics , physics and engineering sciences .
there are other important examples of similar systems , but we do not mention all of them here . the only exception is a hamiltonian integrable system mentioned below .
we rewrite system ( 1.1 ) in brief form @xmath36 where @xmath37 , @xmath38 .
and we consider a perturbation of ( 1.1 ) @xmath19 where @xmath39 and function @xmath6 satisfies the main assumption * _ @xmath40 uniformly with respect to @xmath9 on compact subsets of @xmath22 as @xmath41 .
_ an equivalent formulation of ( a1 ) is : * _ there are positive continuous scalar functions @xmath42 and @xmath43 such that @xmath44 is increasing , @xmath45 is decreasing , @xmath46 and _
@xmath47 the proof of this equivalence is standard enough and we omit it here .
the following theorems are proved for system ( 1.3 ) . * theorem 4.2 . * if @xmath11 then system ( 1.3 ) has bounded solutions for @xmath48 . * theorem 4.4 . * if @xmath11 then any bounded solution of ( 1.3 ) approaches ( orbitally ) a certain @xmath3-dimensional invariant torus of the unperturbed system ( 1.1 ) . * theorem 5.2 . * if @xmath49 then any bounded solution of ( 1.3 ) approaches a certain solution of unperturbed system ( 1.1 ) .
it will be shown that for @xmath50 and @xmath51 these theorems will no longer be valid .
[ [ c. ] ] c. + + in section 3 we construct for the system ( 1.1 ) action - angle variables ( `` quasi - polar '' coordinates ) in which ( 1.1 ) has the form @xmath52 , @xmath53 where @xmath54 are `` quasi - polar '' radii and @xmath55 are `` quasi - polar '' angles .
it gives us possibility to consider instead of ( 1.1 ) the more general system in the form @xmath56 where @xmath57 , @xmath58 , @xmath59 ( i.e. each component @xmath60 ) .
the phase portrait of ( 1.4 ) resembles the cartesian product of a more complicated form than ( 1.2 ) , namely @xmath61 where @xmath62 and @xmath63 the perturbations of ( 1.4 ) need to be taken in the form @xmath64 where @xmath65 , @xmath66 ( let us recall that @xmath58 means indeed that @xmath67 and functions @xmath68 and @xmath69 are periodic in each component of the vector @xmath70 . ) we shall suppose that @xmath68 and @xmath69 satisfy the following main assumption * _ @xmath71 uniformly with respect to @xmath72 on compact subsets of @xmath73 and for all @xmath58 as @xmath41 . _
an equivalent formulation of ( a3 ) is : * _ there are positive continuous scalar functions @xmath42 and @xmath43 such that @xmath44 is increasing , @xmath45 is decreasing , @xmath46 and _
@xmath74 we prove the following theorems for ( 1.5 ) .
* theorem 4.1 .
* if @xmath11 then system ( 1.5 ) has solutions bounded with respect to @xmath72-coordinates for @xmath48 .
* theorem 4.3 . * if @xmath11 then any bounded solution of ( 1.5 ) approaches ( orbitally ) a certain invariant torus of the unperturbed system ( 1.4 ) .
* theorem 5.1 . * if @xmath49 then any bounded solution of ( 1.5 ) converges to a certain solution of the unperturbed system ( 1.4 ) .
we note at the end of this section that problems of asymptotic integration are classical in the theory of differential equations .
if we eliminate the @xmath75equation and the variable @xmath75 from ( 1.5 ) we obtain a well investigated in the literature problem of the perturbations of zero system ( see @xcite @xcite , monographs @xcite and @xcite give the general theory ) . from this point of view theorems 4.14.4 are quite natural and expected statements .
on the contrary , theorems 5.15.2 represent more delicate results .
we consider the construction of the `` quasi - polar '' variables and proof of theorems 5.15.2 as one of the most important parts of our paper .
the aim of this section is to prove the following statement .
let @xmath76 be a two - dimensional subsystem of ( 1.1 ) .
then it has a differentiable positively definite integral @xmath77 on @xmath78 .
this is a known result .
but because we shall need further some detailed properties of @xmath79 we give an explicit proof of this theorem . for convenience
we omit the index @xmath80 and will write simply @xmath9 and @xmath81 instead of @xmath82 and @xmath83
. we belief that no confusion with the system ( 1.1 ) can arise .
let @xmath84 and @xmath85 . the system @xmath0 can be rewritten in the form @xmath86 consider the system `` orthogonal '' to ( 2.1 ) @xmath87 it is evident that the right - hand vector fields of ( 2.1 ) and ( 2.2 ) are orthogonal at each point .
this implies that ( 2.2 ) has also a unique equilibrium point at the origin and other trajectories of ( 2.2 ) cross orthogonally all cycles of ( 2.1 ) at a unique point . for
if any trajectory of ( 2.2 ) intersects some cycle of ( 2.1 ) at more than one point then there is a point on this cycle where right - hand vectors of ( 2.1 ) and ( 2.2 ) will tangent each other .
let @xmath88 be a certain nonzero solution of ( 2.2 ) and @xmath89 denote its maximal interval of existence .
let @xmath90 denote the trajectory of @xmath88 .
without loss of generality we can assume that @xmath90 crosses the cycles of ( 2.1 ) from inside to outside with increase of @xmath91 ( otherwise one needs to consider the system @xmath92 instead of ( 2.2 ) ) .
it follows from poincar bendixon theory that @xmath93 as @xmath94 and @xmath95 as @xmath96 .
indeed , system ( 2.2 ) can not have closed curves ( cycles ) because then they must intersect the cycles of ( 2.1 ) at more than one point .
hence the @xmath44-limit set of @xmath88 is not empty and may consist of only origin . by the same reasons the @xmath97-limit set of @xmath88 must be empty , i.e. @xmath95 as @xmath96 .
it is clear that when time @xmath91 increases from @xmath98 to @xmath99 the solution @xmath88 crosses each cycle of ( 2.1 ) .
we have also that @xmath100 because @xmath88 remains bounded as @xmath94 .
let @xmath101 denote the solution of ( 2.1 ) with initial condition @xmath102 .
we define the mapping @xmath103 by the formula @xmath104 it is clear that @xmath105 because the cycle @xmath106 intersects @xmath90 transversally .
let @xmath107 be the mapping which to every @xmath108 associates the time @xmath91 such that @xmath109 . since @xmath88 is differentiable and @xmath110 ( it is not equilibrium point ) we have @xmath111 .
finally we define on @xmath112 the function @xmath113 by the equality @xmath114 we point out some properties of @xmath115 .
* @xmath116 , * @xmath117 , * @xmath118 , thus @xmath115 is an integral for system ( 2.1 ) on @xmath119 , * equation @xmath120 defines a unique cycle of ( 2.1 ) + ( indeed , curve @xmath88 intersects all cycles of ( 2.1 ) with increasing @xmath91 at only one point ; hence @xmath115 monotonically increases along @xmath90 ; therefore each level curve of @xmath121 consists of a unique cycle ) .
* @xmath122 .
( indeed , @xmath123 ) . by setting @xmath124 we construct a continuous positively - definite integral for ( 2.1 ) on whole @xmath78 ( we assume naturally @xmath125 ) .
it is clear that @xmath126 since differentiability of @xmath127 can be broken only at @xmath128 let us consider the behaviour of the gradient @xmath129 as @xmath130 .
there are two possibilities .
* @xmath131 .
* @xmath132 . in the first case
we define @xmath133 .
we thus have @xmath134 and @xmath135 is therefore the required integral for ( 2.1 ) .
in the second case we will construct the integral @xmath135 in the form @xmath136 where a scalar function @xmath137 is chosen to smooth out the discontinuity of @xmath138 at origin .
to do this we first define the function @xmath139 for @xmath140 .
we have @xmath141 , since the sets @xmath142 are compact and depend continuously on @xmath143 .
the following properties are evident : * @xmath144 .
* @xmath145 . * @xmath146 .
finally we define @xmath147 .
it is easy to see that @xmath148 and @xmath149 .
we set @xmath150 and verify that @xmath135 is the required integral for ( 2.1 ) .
all we need is to check that @xmath151 exists at @xmath128 .
but it is fulfilled , because @xmath152 we point out some properties of @xmath153 . *
@xmath154 , where @xmath155 or @xmath156 . * @xmath157 , i.e. @xmath153 is positively definite .
* @xmath153 monotonically increases along @xmath158 .
* each level curve @xmath159 defines a unique cycle of ( 2.1 ) .
the theorem is proved .
for simplicity we will further assume that @xmath160 because it is easy to construct a @xmath161function @xmath162 and consider the new integral @xmath163 instead of @xmath135 .
in this section we construct new coordinates for more convenient notation of systems ( 1.1 ) and ( 1.3 ) .
this procedure is a standard one for hamiltonian systems , see for example the textbook @xcite .
but we would notice that our coordinates are not identical to the hamiltonian action - angle coordinates .
we introduce them in a more complicated and sly way , so that the change of variables does not break the asymptotic properties of the perturbations and solutions . to each @xmath78component @xmath82 of vector @xmath9
we put in correspondence two scalar coordinates @xmath164 and @xmath165 .
we define the first `` quasi - polar '' coordinate ( the `` radius '' ) @xmath164 by @xmath166 . in accordance to section 2 ,
the equation @xmath167 for each @xmath168 defines a unique cycle of @xmath76 .
moreover the function @xmath166 is invertible on @xmath90 , since @xmath79 monotonically increases along @xmath90 .
let @xmath169 denote this inversion .
we have * @xmath170 , * @xmath171 , * @xmath172 .
we denote the period of any cycle @xmath167 by @xmath173 .
it is equal to the time for the next intersection with @xmath90 of the solution passing at @xmath174 through the point @xmath175 .
since @xmath90 is a tranversal , we have @xmath176 .
we now prove the following the value @xmath177 is well - defined on @xmath178 .
obviously , @xmath179 for @xmath180 .
it remains to prove that @xmath181 .
let us use the notations from the proof of theorem 2.1 .
changing in ( 2.1 ) coordinates @xmath182 and @xmath183 into the standard polar coordinates @xmath184 , @xmath185 , we get for @xmath44 @xmath186 obviously , @xmath187 and @xmath188
. then hadamard s lemma implies @xmath189 where @xmath190 .
hence , @xmath191 . let @xmath192 denote the solution passing through the initial point @xmath193 at @xmath174 .
let @xmath194 be the period of it .
it is clear that @xmath195 . on other hand , @xmath196 . for any @xmath197 $ ]
the solution @xmath192 belongs to a certain compact set , therefore @xmath198 is uniformly bounded with respect to @xmath199 . if @xmath200 as @xmath201 , then @xmath202 this contradiction proves the proposition .
change time in the system @xmath76 by formula @xmath203 since @xmath204 along the solutions we have @xmath205 and @xmath206 all solutions of ( 3.1 ) have the common period @xmath207 , thus the centre defined by ( 3.1 ) is an isochronous one .
let @xmath208 be the solution of ( 3.1 ) satisfying the initial condition @xmath209 .
we define the second `` quasi polar '' coordinate ( the `` angle '' ) @xmath165 in such a way that @xmath210 .
i.e. @xmath211 is the time of the first intersection of solution @xmath208 with @xmath90 .
it is clear that @xmath165 depends smoothly on @xmath82 .
thus the general formula for the transformation to the new coordinates is @xmath212 let us investigate some properties of the `` quasi polar '' variables
. for any function @xmath213 the notations @xmath214 and @xmath215 denote the partial derivatives of @xmath216 with respect to the first and second argument correspondingly .
let @xmath217 be the jacobian matrix of the change of variables ( 3.2 ) .
we have @xmath218.\ ] ] since @xmath219 , the matrix @xmath220 is 1periodic with respect to @xmath165 . on the other hand @xmath221
is the derivative of the solution @xmath222 with respect to initial values , i.e. it is a fundamental matrix of a so - called linear system in variation .
hence @xmath221 is nonsingular and @xmath223 .
it follows from periodicity that the values @xmath224 and @xmath225 are bounded away from zero and infinity for each fixed @xmath164 and @xmath226 . since @xmath227 is a solution of ( 3.1 )
, we have @xmath228 this yields @xmath229 $ ] .
but derivative @xmath230 is a tangent vector to @xmath90 at the point @xmath169 , because @xmath169 is some parametrization of curve @xmath90 .
hence @xmath231 is to be orthogonal to the vector @xmath232 .
thus matrix @xmath233 is non singular @xmath234 .
it is easy to check the following statement .
let @xmath235 be an arbitrary @xmath236smooth system and @xmath237 be the solution of it with initial condition @xmath238 , then @xmath239 .
this implies @xmath240 and therefore @xmath241= ( d_2\psi_k ) \big [ v'_k,\quad t_k\,f_k\big ] \\ & = d_2\psi_k(\theta_k , v_k(r_k ) ) j_k(r_k,0).\end{aligned}\ ] ] each matrix in the last expression is nonsingular and possesses `` good '' properties .
so this equality guarantees that the jacobian matrix @xmath242 is nonsingular for fixed @xmath243 and arbitrary @xmath244 .
moreover , @xmath245 , @xmath246 , @xmath247 and @xmath248 are bounded away from zero and infinity .
similar properties of boundedness will also be fulfilled if we take @xmath249 where @xmath250 is some compact set of @xmath178 and @xmath165 as before will be an arbitrary one . in the new variables system ( 3.1 ) will have the following form @xmath251 the original system @xmath252 is transformed into @xmath253 ( replacing @xmath254 ) . finally the full system ( 1.1 ) is transformed into @xmath255 where @xmath54 , @xmath55 , @xmath256 , @xmath257 . changing the variables in the perturbed system ( 1.3 ) , we obtain @xmath258^{-1}g(t,\psi(\theta , v(r)))= { 0\choose a(r)}+{p(t , r,\theta)\choose q(t , r,\theta)}\ ] ] ( we use natural notations @xmath259 , @xmath260 , + @xmath261 ) .
the functions @xmath68 and @xmath69 possess the following properties : * they are in @xmath262 , * they are 1periodic with respect to each components of @xmath75 , * if @xmath6 satisfies the assumptions ( a1 ) ( or ( a2 ) ) then @xmath68 and @xmath69 satisfy ( a3 ) ( or ( a4 ) ) . indeed , let @xmath263 be a compact set and @xmath264 .
there exist @xmath265 such that @xmath266 for all @xmath267 and @xmath67 .
hence @xmath268 ( we note that if @xmath264 then cycle @xmath269 does not leave some compact set in @xmath22 for all @xmath67 ) . thus , we have shown that systems ( 1.1 ) and ( 1.3 ) may be reduced to systems ( 1.4 ) and ( 1.5 ) while preserving assumption ( a3 ) ( or ( a4 ) ) .
we want to underline that it is not so difficult to reduce formally system ( 1.3 ) to ( 1.5 ) ( see theory of hamiltonian system in @xcite ) .
the problem is to keep the asymptotic properties ( a1-a2 ) as well as the boundedness and covergence of solutions under the inverse transformation . for this
we use the trick with isochronicity . without it
the jacobian matrix may be unbounded and this manner of reasoning fails .
let us consider the system ( 1.5 ) .
let @xmath68 satisfy assumption ( a3 ) or ( a4 ) with @xmath11 .
then ( 1.5 ) has solutions bounded with respect to @xmath72-coordinates .
we shall find a comparison equation for @xmath270 .
we have @xmath271 here @xmath272 denotes inner product .
this gives for @xmath273 @xmath274 ( we can assume that @xmath275 for sufficiently large @xmath91 , because @xmath276 ) .
let us consider the comparison equation @xmath277 it is well known ( see @xcite ) that every function @xmath278 satisfying inequality ( 4.1 ) does not exceed the solution @xmath279 of ( 4.2 ) with initial condition @xmath280 .
thus the existence of bounded solutions of ( 4.2 ) will imply the same for ( 1.5 ) .
let @xmath281 .
since @xmath282 , function @xmath283 is monotone increasing and invertible .
the solution of ( 4.2 ) with initial condition @xmath284 has the form @xmath285 there are two possibilities . *
a*. @xmath286 .
then all solutions of ( 4.2 ) must be bounded . for if some solution @xmath287 as @xmath21 , then the left
hand side of ( 4.3 ) tends to @xmath288 whereas the right hand side tends to @xmath289 .
we got a contradiction . *
b*. @xmath290 .
fix @xmath291 and take @xmath292 so large that @xmath293
. then the solution of ( 4.2 ) satisfying the initial condition @xmath284 is bounded .
otherwise , tending @xmath294 in ( 4.3 ) and assuming that @xmath287 we get the impossible equality @xmath295 . thus the equation ( 4.2 ) has bounded solutions in both cases . so the theorem is proved . the case * a * occurs when @xmath68 satisfies the well known inequality ( see @xcite ) @xmath296 where @xmath297 and @xmath298 . in this case all solutions of ( 1.5 ) are bounded with respect to @xmath72 .
conversely , there are unbounded solutions in the case * b * as the next example shows .
consider the system @xmath13 , @xmath299 and its perturbation @xmath300 , @xmath299 .
the perturbed system has unbounded extendable to infinity solution @xmath301 .
all solutions lying inside the sector @xmath302 are bounded and those lying outside it are unbounded ( moreover , they are unextendable to infinity ) .
the condition @xmath11 in theorem 4.1 can not be generalized to include the case @xmath50 .
see example 4.2 below .
now we give reformulations of theorem 4.1 and note 2.1 for system ( 1.3 ) .
suppose that perturbation @xmath6 in ( 1.3 ) satisfies ( a1 ) or ( a2 ) with @xmath11
. then ( 1.3 ) has bounded solutions .
as it follows from section 3 the system ( 1.3 ) written in `` quasi - polar '' coordinates satisfies all conditions of theorem 4.1 .
all solutions of ( 1.3 ) are bounded if @xmath303 , where @xmath297 and @xmath298 .
the present example shows that condition @xmath11 can not be generalized to include the case @xmath50 .
we take the standard two - dimensional centre @xmath304 and perturb it in the following way @xmath305 it is evident that @xmath306 and @xmath307 satisfy ( a2 ) with @xmath50 , @xmath308 , @xmath309 .
but all solutions of the perturbed system have the form @xmath310 and are unbounded .
now we return to system ( 1.5 ) and show that under the assumptions of theorem 4.1 every bounded solution approaches a certain invariant torus of ( 1.4 ) .
let @xmath311 be a bounded solution of ( 1.5 ) with respect to @xmath72-coordinates and @xmath312 with @xmath11 .
there exists @xmath313 such that @xmath314 evidently , if ( a3 ) or ( a4 ) holds then every bounded solution of ( 1.5 ) satisfies @xmath315 .
we have @xmath316 assumption @xmath315 implies that @xmath317 converges .
hence the limit @xmath318 exists .
the theorem is proved .
the condition @xmath11 can not be generalized to include the case @xmath50 , see example 4.4 .
the next example demonstrates that assumptions of theorem 4.3 do nt imply even orbital convergence of bounded solutions of ( 1.5 ) to certain solution of ( 1.4 ) .
however this convergence occurs if @xmath319 ( see @xcite ) , see also note 5.2 .
consider the following system @xmath320 and its perturbation @xmath321 solutions of ( 4.5 ) are @xmath322 where @xmath323 , @xmath324 , @xmath325 , @xmath326 are arbitrary constants ( initial values ) .
solutions of ( 4.6 ) are @xmath327 all solutions of ( 4.6 ) are bounded with respect to @xmath328 and @xmath329 .
it is clear that @xmath330 and @xmath331 as @xmath41 .
the orbital convergence of ( 4.8 ) to ( 4.7 ) means that @xmath332 and there is an infinitely monotonically increasing function @xmath333 such that @xmath334 or @xmath335 from ( 4.9 ) we have @xmath336 .
substituting this in ( 4.10 ) we obtain @xmath337 however , this is possible if and only if @xmath338 , @xmath339 , @xmath340 .
consequently there are solutions of ( 4.6 ) which do nt orbitally converge to any solution of ( 4.5 ) .
now we reformulate theorem 4.3 for system ( 1.3 ) .
let @xmath341 be a bounded solution of ( 1.3 ) and @xmath342 with @xmath11 .
then @xmath341 approaches a certain invariant torus of unperturbed system ( 1.1 ) .
it follows immediately from theorem 4.3 and section 3 . for theorem 4.4
notes 4.44.6 are also valid .
the next example demonstrates that assumption @xmath11 can not be generalized to include the case @xmath50 .
let s consider again the two - dimensional linear centre ( 4.4 ) with the following perturbation @xmath343 obviously every bounded solution of ( 4.11 ) satisfies the conditions of theorem 4.4 with @xmath50 .
but ( 4.11 ) has the bounded solutions @xmath344 which oscillate between the circles with radii @xmath345 and @xmath346 .
hence no one of them can approach any circle @xmath347 , which is just the solution of the unperturbed system .
we begin with the main statement for system ( 1.5 ) .
let @xmath311 denote any bounded solution of ( 1.5 ) .
theorem 4.1 implies the convergence @xmath348 as @xmath8 .
every solution of ( 1.4 ) lying on the invariant torus @xmath349 has the form @xmath350 , where @xmath351 is the initial value .
it is necessary and sufficient to prove the existence of a @xmath351 such that @xmath352 thus it is sufficient to prove that the limit value @xmath353 exists . since @xmath354 it is sufficient to prove that the obtained integral converges . by ( a4 ) we have @xmath355 there is a constant @xmath356 such that @xmath357 for @xmath358 .
hence @xmath359 next we need to estimate @xmath360 .
first of all we notice that @xmath361 since @xmath362 , it satisfies the lipschitz inequality with some constant @xmath363 ( on such a compact set @xmath250 from @xmath73 that @xmath364 for @xmath365 ) @xmath366 it follows from ( 5.2 ) and ( 5.3 ) that @xmath367 .
inequality @xmath49 implies convergence of @xmath368 .
hence limit @xmath369 exists .
this proves the theorem .
consider the system @xmath370 and its perturbation @xmath371 the cycles of ( 5.4 ) are @xmath372 @xmath373 .
let @xmath374 denote the solution of ( 5.4 ) passing through point ( 1,0 ) as @xmath174 .
changing the variables by the formulas @xmath375 ( action - angle coordinates ) we transform ( 5.4 ) into @xmath376 and ( 5.5 ) into @xmath377 the solutions of ( 5.6 ) are @xmath378 @xmath379 and those of ( 5.7 ) are @xmath380,$ ] @xmath381\,ds.$ ] obviously , @xmath382 but @xmath383 from proof of theorem 5.1 does not exist .
@xmath384\,ds - t\big ) \\
\qquad = \lim_{t\to+\infty}r_0 ^ 2\big(\int_2^t ( \exp[-2/(s\ln s)]-1)\,ds-2\big)=-\infty,\end{gathered}\ ] ] because obtained integral diverges .
this follows from the asymptotic equivalence @xmath385 - 1 \sim -2/(s\ln s)\ ] ] and divergence of @xmath386 we wouls like to make an interesting remark about the possibility to decrease the lower bound in the inequality @xmath49 in theorems 5.1 and 5.2 .
suppose @xmath387 is a constant . for system ( 1.1 )
it is equivalent to @xmath388 it means that each subsystem @xmath76 in ( 1.1 ) , @xmath389 defines an isochronous centre . for ( 1.5 ) we can also talk about a property of isochronicity .
then the statements of theorems 5.15.2 are valid for @xmath390 indeed , we have in this case @xmath391 where convergence of this integral follows from ( 5.1 ) . the property of isochronicity in hamiltonian systems ( and not only for such systems ) is discussed in numerous papers .
we refer only to @xcite , where further references can be found . only for asymptotic perturbations of such systems ,
our note 5.2 is valid .
we note that the main reason of appearance of the condition @xmath49 is just the necessity to estimate @xmath392 .
hence we may suppose that @xmath69 satisfies ( a3 ) with @xmath11 ( but @xmath68 with @xmath49 as before ) and theorem 5.1 will also be valid .
finally we point out a possibility of the situation for system ( 1.1 ) when some subsystems of it are isochronous centres and all other not .
it is not difficult to see that for `` isochronous '' angles @xmath165 the orbital convergence ( _ note : _ even usual convergence ) is realized in theorem 4.4 .
i am very grateful to the royal swedish academy of sciences for their financial support of this work .
i would also like to thank g. sderbacka for carefully reading this paper and his valuable remarks . | we consider the class of autonomous systems @xmath0 , where @xmath1 , @xmath2 whose phase portrait is a cartesian product of @xmath3 two - dimensional _
centres_. we also consider perturbations of this system , namely @xmath4 , where @xmath5 and @xmath6 is asymptotically small , that is @xmath7 as @xmath8 uniformly with respect to @xmath9 . the rate of decrease of @xmath6 is assumed to be @xmath10 where @xmath11 .
we prove under this conditions the existence of bounded solutions of the perturbed system and discuss their convergence to solutions of the unperturbed system .
this convergence depends on @xmath12 .
moreover , we show that the original unperturbed system may be reduced to the form @xmath13 , @xmath14 , and taking @xmath15 , @xmath16 , where @xmath17 denotes the @xmath3-dimensional torus , we investigate the more general case of systems whose phase portrait is foliated on invariant tori .
we notice that integrable hamiltonian systems are of the same nature .
we give also several examples , showing that the conditions of our theorems can not be improved .
journal of nonlinear mathematical physics 2000 , v.7 , n 2 , . [ firstpage ] |
varying coefficient models have been extensively studied in the literature , and they are useful for characterizing nonconstancy relationship between predictors and responses in regression models ; see , for example , @xcite . in this paper
we consider the time - varying coefficient model @xmath1 where @xmath2 is the response , @xmath3 is the predictor , @xmath4 is the transpose operator , @xmath5 for some smooth function @xmath6 \to\mathbb r^p$ ] and @xmath7 is the error .
we assume that @xmath8 .
estimation of the coefficient function @xmath9 in model ( [ eqnlreg ] ) has been considered by @xcite among others .
an important special example of ( [ eqnlreg ] ) is the time - varying autoregressive model @xcite by letting @xmath10 .
there are important differences between our model ( [ eqnlreg ] ) and those under longitudinal or functional setting . here
we assume that only one realization @xmath11 is available , while in the longitudinal and functional setting , many subjects are measured at multiple times , so that one has multiple realizations .
a natural problem for model ( [ eqnlreg ] ) is to test whether certain or all components in @xmath12 are time - invariant .
there is a huge literature on the problem of testing parameter stability ; see , for example , @xcite . for model ( [ eqnlreg ] ) , we are interested in testing @xmath13 for some vector @xmath14 , where @xmath15 is a real - valued matrix . with an appropriately chosen @xmath16 , the null hypothesis ( [ eqnh0 ] )
can be formulated to test whether a certain part of coefficients is zero or time - invariant . in the latter case
, @xmath17 needs to be replaced by an estimate @xmath18 .
zhou and wu @xcite built simultaneous confidence tubes for the regression coefficient function @xmath19 , which can be used as an @xmath20-test for ( [ eqnh0 ] ) .
the latter test often does not have a good power if the alternative hypothesis consists of nonzero smooth functions . in section [ subsechypotest ]
we propose a more powerful @xmath21-test which is based on the weighted integrated squared errors .
our setting is much more general than the one in @xcite where @xmath22 is assumed to be @xmath23-mixing and stationary . in comparison
, we allow nonstationary predictor and error processes which can be nonstrong mixing ; see section [ secprelim ] for our nonstationary framework and basic assumptions . if some of the coefficients are time - invariant , model ( [ eqnlreg ] ) becomes the ( semiparametric ) partially time - varying coefficient model @xmath24 where @xmath25 are groups of parametric and nonparametric components , respectively .
based on an estimate of ( [ eqnlreg ] ) , we simply take an integration / average over the parametric part to obtain an estimate of @xmath26 that achieves the @xmath0-convergence rate . an asymptotic theory is given in section [ subsecparameterestimation ] .
this method was previously used in @xcite for estimating state - domain semivarying coefficient models .
the latter paper assumed that @xmath27 where @xmath28 , @xmath29 , are independent and identically distributed ; see @xcite for the case with stationary mixing processes .
our time - domain model ( [ eqnregsemi ] ) is very general , and it includes both ( [ eqnlreg ] ) and usual linear regression models .
gao and hawthorne @xcite considered a special example of ( [ eqnregsemi ] ) with @xmath30 , @xmath31 and @xmath32 , so that only the intercept term in ( [ eqnregsemi ] ) is time - varying . section [ subsecvarsel ] deals with the problem of selecting significant predictors .
fan et al .
@xcite proposed an extended aic for choosing locally significant variables .
abramovich et al .
@xcite considered the problem of order selection for time - varying autoregressive models by requiring that multiple realizations are available . using the dependence framework in section [ secprelim ] , we are able to solve this problem under parameter instability and temporal dependence
. in particular , we propose an information criterion , consisting of measures of goodness - of - fit and model complexity , that can consistently select the true set of relevant predictors based only on one realization .
section [ secimplementation ] provides simulation studies and an application .
proofs are given in the .
since the coefficient function @xmath33 in ( [ eqnlreg ] ) is smooth , we can naturally estimate it ( along with its derivative ) by @xmath34 where @xmath35 is the kernel function , and @xmath36 is a bandwidth sequence satisfying @xmath37 and @xmath38 . throughout the paper
we assume that the kernel function @xmath35 is a symmetric and bounded function in @xmath39 $ ] with @xmath40 .
for example , it can be the epanechnikov kernel @xmath41 or the bartlett kernel @xmath42 . observe that ( [ eqnllmineqn ] ) has the closed form solution @xmath43 where for @xmath44 , @xmath45 with the convention that @xmath46 , and @xmath47 to establish an asymptotic theory for @xmath48 , we need to impose appropriate regularity conditions on the covariates @xmath49 and errors @xmath50 .
for testing the hypothesis ( [ eqnh0 ] ) , @xcite assumed that @xmath22 is @xmath23-mixing and stationary .
to allow nonstationary predictor and error processes that can be nonstrong mixing , we assume that @xmath51 where @xmath52 is a shift process of independent and identically distributed ( i.i.d . )
random variables @xmath53 , @xmath54 and @xmath55 and @xmath56 are measurable functions such that @xmath57 and @xmath58 are well defined for each @xmath59 $ ] . this setup is also used in @xcite . for a random vector @xmath60 , we write @xmath61 , @xmath62 , if @xmath63 , where @xmath64 is the euclidean vector norm , and we denote @xmath65 .
a process @xmath66 is said to be stochastically lipschitz continuous ( @xmath67 in short ) if there exists @xmath68 , such that @xmath69 holds uniformly for all @xmath70 $ ] .
then , under condition ( a2 ) below , ( [ eqnformulation ] ) defines locally stationary processes .
let @xmath71 be an i.i.d .
copy of @xmath72 and @xmath73 be the coupled shift process .
we define the functional dependence measure @xmath74 } \bigl\|\mathbf{j}(t ; \bolds { \mathcal f}_k ) - \mathbf{j } ( t ; \bolds{\mathcal f}_{k,\{0\ } } ) \bigr\|_q \quad\mathrm{and}\quad \theta_{m , q}(\mathbf{j } ) = \sum _ { j = m}^{\infty } \delta_{j , q}(\mathbf{j}).\ ] ] let @xmath75 be the long - run covariance matrix , and @xmath76 . under the short - range dependence condition
@xmath77 , both of them are uniformly bounded over @xmath78 $ ] .
let @xmath79 , and we shall make the following assumptions : smoothness : @xmath80 $ ] ; local stationarity : @xmath81 ; short - range dependence : @xmath82 for some @xmath83 ; the smallest eigenvalue of @xmath84 is bounded away from zero on @xmath85 $ ] .
a sufficient condition for ( a3 ) is that @xmath86 for some @xmath83 .
let @xmath16 be a pre - specified matrix and @xmath87 .
then @xmath88 is an estimate of @xmath17 . for the partially time - varying coefficient model ( [ eqnregsemi ] ) ,
let @xmath89 be a matrix with rows @xmath90 , where @xmath91 is the vector with unit @xmath92th component and zeros elsewhere , then @xmath93 , @xmath78 $ ] .
although @xmath26 can be consistently estimated by @xmath94 for any @xmath95 , the smoothed estimate @xmath96 can have a better rate of convergence .
[ thmcltahat ] assume and @xmath97 for some @xmath98 .
let @xmath99 and @xmath100 .
if @xmath101 for some @xmath102 , then @xmath103 in theorem [ thmcltahat ] , the term @xmath104 can be interpreted as the bias due to nonparametric estimation , and it vanishes under the null hypothesis ( [ eqnh0 ] ) .
hence the parametric component @xmath105 in the semi - parametric model ( [ eqnregsemi ] ) can have a @xmath0-consistent estimate @xmath106 . for testing the null hypothesis ( [ eqnh0 ] ) ,
let @xmath107 be a continuous mapping from @xmath85 $ ] to symmetric positive - definite matrices in @xmath108 .
consider the weighted integrated squared error @xmath109 if @xmath17 is unknown , an estimate can be used .
for example we can use @xmath110 , which has the parametric convergence rate ; see theorem [ thmcltahat ] .
chen and hong @xcite considered the special case that @xmath22 is a stationary @xmath23-mixing process .
their generalized hausman test @xcite relates to ( [ eqntn ] ) with @xmath16 being the identity matrix and @xmath111 .
such a choice of weight matrices should be used if we are interested in prediction .
alternatively , we can use @xmath112 , the identity matrix to form the integrated squared errors .
let @xmath113 , by theorem 1 in @xcite , @xmath114 has the asymptotic covariance @xmath115 .
hence , we can choose @xmath116 to serve as a normalizer . in this case , ( [ eqntn ] )
is ( proportionally ) an integral of the squared local @xmath117-statistics . for a matrix @xmath16 ,
define @xmath118 and @xmath119 .
let @xmath120 and @xmath121 . since @xmath122 , we have @xmath123 $ ] and is symmetric .
let @xmath124 \\[-8pt ] \nonumber \xi_{\mathbf{a},\mathbf{w},l } & = & \operatorname{tr } \biggl\{\int_0 ^ 1 \bolds\xi_{\mathbf{a},\mathbf{w}}(t)^l \,dt \biggr\}.\end{aligned}\ ] ] theorem [ thmclttn ] provides asymptotic normality for @xmath125 .
[ thmclttn ] assume , @xmath126 and @xmath127 for some @xmath128 .
if @xmath129 for some @xmath130 , then @xmath131 if in addition @xmath132 , then ( [ eqnclttn ] ) holds for @xmath133 .
let @xmath134 be the cumulative standard normal distribution function and @xmath135 be the corresponding @xmath136th quantile .
we reject the null hypothesis ( [ eqnh0 ] ) at level @xmath137 if @xmath138 let @xmath139 \to\mathbb{r}^{s}$ ] be of class @xmath140 , and @xmath141 be a sequence of nonnegative real numbers .
proposition [ thmpower ] provides the asymptotic power of the test ( [ eqntestusual ] ) under the local alternative @xmath142 [ thmpower ] assume conditions of theorem [ thmclttn ] .
if @xmath143 , then the power of the test ( [ eqntestusual ] ) satisfies @xmath144 in this section we shall propose an information criterion for time - varying coefficient models that can consistently identify the true set of relevant predictors .
recall that @xmath145 is the whole set of potential predictors , and @xmath146 is the coefficient estimate .
let @xmath147 be the true set of relevant predictors .
for a candidate subset @xmath148 , we can compute the variable selection information criterion @xmath149 \\[-8pt ] \eqntext{\mbox{where } \displaystyle\operatorname{rss}(d ) = \sum _ { i=1}^n \bigl\{y_i - \mathbf{x}_{d , i}^\top\tilde{\bolds{\beta}}_d(i / n ) \bigr\}^2.}\end{aligned}\ ] ] here @xmath150 is a tuning parameter .
we select a subset @xmath151 that minimizes @xmath152 , thus balancing goodess - of - fit and model complexity .
smaller @xmath150 leads to more predictors , and vice versa . theorem [ thmmodelselect ] provides theoretical properties of our procedure . [ thmmodelselect ] assume , @xmath153 , @xmath97 for some @xmath98 .
let @xmath154 and @xmath155 . if @xmath156 for some @xmath157 , @xmath158 then , for any @xmath159 , @xmath160 .
@xmath161theorems [ thmcltahat ] and [ thmclttn ] both involve unknown quantities depending on the covariance matrices @xmath162 and @xmath163 , @xmath59 $ ] . the problem of estimating
covariance matrices has been extensively studied ; see @xcite among others .
let @xmath164 , @xmath165 and @xmath166 be bandwidth sequences satisfying @xmath167 , @xmath168 , @xmath169 and @xmath170 .
let @xmath171 $ ] , @xmath172 , @xmath173 $ ] and @xmath174 for @xmath78 $ ] , we estimate @xmath162 and @xmath163 , respectively , by @xmath175 and @xmath176 where @xmath177 are local linear weights with bandwidth @xmath178 and @xmath179 .
proposition [ thmcovest ] provides consistency of our covariance matrix estimates .
[ thmcovest ] assume , @xmath180 and @xmath181 for some @xmath182 .
if both @xmath183 and @xmath184 are in class @xmath185 , then @xmath186 } \bigl\| \hat{\mathbf{m}}(\mathbf{g } , t ) - \mathbf{m}(\mathbf{g } , t ) \bigr\| = o\bigl\{(n\varpi_n)^{-1/2 } + \varpi_n^2\bigr\},\ ] ] and @xmath187 } \bigl\| \hat{\bolds{\lambda}}(\mathbf{l } , t ) - \bolds{\lambda}(\mathbf{l } , t ) \bigr\| = o\bigl\{\varrho_n^{1/2 } + ( n \tau_n \varrho_n)^{-\nu } + ( \tau_n \varrho_n)^{\nu/(1+\nu ) } + \tau_n^2 \bigr\}.\hspace*{-35pt}\ ] ] the bound in ( [ eqnbndm ] ) is optimized and becomes @xmath188 if @xmath189 .
the optimal bound in ( [ eqnbndlambda ] ) is complicated and it depends on @xmath190 , the decay rate of dependence .
in particular , if @xmath191 , then the optimal bound in ( [ eqnbndlambda ] ) is @xmath192 if @xmath193 and @xmath194 ; otherwise it is @xmath195 if @xmath196 and @xmath197 , or @xmath198 and @xmath199 . in computing @xmath200 ,
since @xmath50 is usually unknown , we shall replace it by @xmath201 , the estimated local linear residuals . by the sandwich formula ,
let @xmath202 and , as in ( [ thmo19303 ] ) , correspondingly define @xmath203 and @xmath204 . by ( [ eqntestusual ] )
, we reject the null hypothesis ( [ eqnh0 ] ) at level @xmath137 if @xmath205 if @xmath17 is known , then in ( [ eqo19307 ] ) we can use @xmath17 instead of @xmath18 . the criterion ( [ eqo19307 ] )
usually does not have a good performance because of the slow convergence in ( [ eqnclttn ] ) .
note that the statistic @xmath206 is asymptotically pivotal , so we propose a simulation - assisted testing procedure that can substantially improve the finite - sample performance .
in particular , we generate i.i.d .
standard normal random variables @xmath207 , @xmath208 , and i.i.d .
standard multivariate normal random vectors @xmath209 , @xmath210 , that are also independent of @xmath211 .
we compute the corresponding @xmath212 , and repeat this for many times to obtain its empirical quantile @xmath213 .
we reject the null hypothesis ( [ eqnh0 ] ) at level @xmath137 if @xmath214 .
our procedure has a similar flavor as the wilks type of phenomenon discussed in @xcite .
a major difference is that we allow dependent and nonstationary errors .
bandwidth selection for nonparametric hypothesis testing is a nontrivial problem , and it has been studied in @xcite among many others . as commented by wang @xcite , there exists no uniform guidance for an optimal choice . on the positive side ,
our simulation results in section [ subsecsimulation ] indicate that the empirical acceptance probabilities are not quite sensitive to the choice of bandwidth .
hence one can simply choose @xmath215 that has the asymptotic mean integrated squared error ( amise ) optimal rate . as an alternative , we consider the generalized cross - validation ( gcv ) selector by @xcite , and estimate the covariance matrix @xmath216 to correct for dependence @xcite . specifically , let @xmath217 ; then for any bandwidth @xmath218 , one can write the local linear smoothed fitted values as @xmath219 , where @xmath220 is the corresponding hat matrix .
we choose the bandwidth @xmath221 that minimizes @xmath222 ^ 2}.\ ] ] an estimate of the covariance matrix @xmath223 can be obtained by using the banding technique as in @xcite .
the gcv selector ( [ eqngcv ] ) works reasonably well in our simulation studies .
we shall now provide data - driven choices of @xmath164 , @xmath165 and @xmath166 in the estimation of covariance matrices . from the construction in section [ subseccovest ] , we truncate the long - run covariance matrix estimate at lag @xmath224 and , by the proof of theorem [ thmcltahat ] , @xmath225 \biggr ) \rightarrow n\bigl(0 , \sigma_k^2\bigr),\ ] ] where @xmath226 is the integrated long - run variance of the process @xmath227 .
we propose to choose @xmath228 .
note that the final estimate is a local linear smoother of @xmath229 , @xmath210 , and we can apply the gcv method to select @xmath230 .
the latter can also be applied to @xmath231 , @xmath232 , to select @xmath233 , and we take @xmath234 to avoid numerical singularities .
these data - driven choices of bandwidths are able to capture dependence and nonstationarity and have a good performance in our simulation studies . for
the information criterion ( [ eqnvic ] ) , if @xmath235 and @xmath236 , then ( [ eqnchin ] ) becomes @xmath237 note that condition ( [ eqnchinsimplified ] ) is more restrictive than the traditional bayesian information criterion ( bic ) because of parameter instability . under the latter setting , a heavier penalty on model complexity
is usually needed to suppress the over - fitting problem ; see @xcite for a similar finding on cross - validation methods . as a rule of thumb , we suggest using @xmath238 .
this simple choice performs reasonably well as can be seen from our simulation study .
the choice of bandwidth becomes further complicated due to model uncertainty .
we suggest a two - stage selection procedure : let @xmath221 be the selected bandwidth by gcv with all available predictors , and we use the information criterion ( [ eqnvic ] ) to select a pilot set of relevant predictors ; then we select the bandwidth @xmath239 by applying the gcv method to this pilot set . modeling a nonstationary process by autoregressive models with time - varying coefficients
has attracted considerable attention .
a traditional approach is to project the coefficient function onto a basis of temporal functions , and estimates the basis coefficients ; see , for example , @xcite .
other contributions on parameter estimation can be found in @xcite among others .
abramovich et al .
@xcite considered the problem of order selection by requiring multiple realizations .
let @xmath240 be continuous functions .
we shall prove that the time - varying autoregressive process @xmath241 has an approximate solution of form ( [ eqnformulation ] ) , and the difference is of a negligible order .
hence the results in section [ secmain ] can be directly applied to address the problem of parameter estimation , hypothesis testing and order selection for time - varying autoregressive models .
recall that @xmath242 .
let @xmath243 , @xmath244 and @xmath245 then ( [ eqntvar ] ) can be written as @xmath246 we shall make the following assumptions : the starting point @xmath247 .
the coefficient functions @xmath248 , @xmath249 , are lipschitz continuous on @xmath85 $ ] .
@xmath250 for all @xmath251 with @xmath252 uniformly in @xmath78 $ ] .
conditions ( t2 ) and ( t3 ) entail local stationarity and short - range dependence , respectively ; see also @xcite .
proposition [ proptvar ] states that the autoregressive process ( [ eqntvarvector ] ) can be approximated by ( [ eqnformulation ] ) with a uniform approximation error of order @xmath253 .
[ proptvar ] assume .
if @xmath254 , then there exists a measurable function @xmath255 and a constant @xmath68 such that @xmath256 in proving asymptotic results of section [ secmain ] , the key quantity is the partial sum process @xmath257 and @xmath258 , @xmath259 . by proposition [ proptvar ]
, there exists a measurable function @xmath255 such that @xmath260 and @xmath261 a careful check of the proofs of our results in section [ secmain ] indicates that , due to the above relation , they are still valid for the time - varying autoregressive process ( [ eqntvar ] ) . the generalized likelihood ratio test ( glrt , @xcite ) is a popular method for nonparametric hypothesis testing .
it was used by cai , fan and li @xcite for testing the coefficient constancy , and generalized by fan and huang @xcite to semiparametric models .
properties of glrt have been extensively studied for i.i.d .
samples . however , its validity for dependent data is not guaranteed .
we shall here briefly review the glrt and compare it with our method . for the null hypothesis @xmath262 for some vector @xmath263 ,
the glrt statistic is defined as @xmath264 where @xmath265 is the least squares estimate . to construct the null distribution of @xmath266
, we use the conditional bootstrap as suggested by @xcite .
let @xmath267 and @xmath268 be i.i.d .
we generate the bootstrap sample @xmath270 , @xmath232 , and compute the test statistic @xmath271 .
we approximate the distribution of @xmath272 by that of @xmath273 .
consider the ar - arch process with time - varying conditional variance @xmath274 e_i , \\
e_i & = & \bigl(1 + 0.25 e_{i-1}^2 \bigr)^{1/2 } \varepsilon_i,\end{aligned}\ ] ] where @xmath275 are i.i.d .
let @xmath277 and the bandwidth @xmath278 .
we consider testing whether the coefficient of @xmath279 is a constant . for @xmath280
, we use the identity weights @xmath281 and obtain its cut - off value by the simulation - assisted procedure in section [ subsechypotestsimubased ] with 5000 simulated @xmath282 . for @xmath266 , the cut - off value is obtained by 5000 bootstrapped @xmath273 .
we generate 5000 realizations of the ar - arch process and use q q plots to examine the performance .
the results are presented in figure [ figqqplots ] .
it shows that the glrt fails to provide valid @xmath283-values in the presence of dependence and nonstationarity .
for example , the empirical acceptance probabilities are 79% , 86.4% and 94.7% for the 90% , 95% and 99% nominal levels , respectively .
as shown in figure [ figqqplots](b ) , our dependence - adjusted procedure provides a satisfactory approximation of @xmath284 . at 90% , 95% and 99% nominal levels ,
our empirical acceptance probabilities are 89.5% , 95.0% and 98.7% , respectively . [ cols="^,^ " , ] @lcd3.2c@ & + & + & * under - fitted * & & + + 0.1 & 0.0 & 100.0 & 0.0 + @xmath285 & 0.0 & 100.0 & 0.0 + 0.2 & 0.0 & 100.0 & 0.0 + @xmath286 & 0.0 & 100.0 & 0.0 + 0.3 & 0.0 & 100.0 & 0.0 + @xmath287 & 0.0 & 100.0 & 0.0 + 0.4 & 0.0 & 100.0 & 0.0 + 0.5 & 0.0 & 100.0 & 0.0 + 0.6 & 0.0 & 100.0 & 0.0 + 0.7 & 0.0 & 100.0 & 0.0 + 0.8 & 0.0 & 100.0 & 0.0 + 0.9 & 0.0 & 100.0 & 0.0 + + 0.1 & 0.0 & 99.8 & 0.2 + @xmath288 & 0.0 & 100.0 & 0.0 + @xmath289 & 0.0 & 100.0 & 0.0 + 0.2 & 0.0 & 100.0 & 0.0 + @xmath290 & 0.0 & 100.0 & 0.0 + 0.3 & 0.0 & 100.0 & 0.0 + 0.4 & 0.0 & 100.0 & 0.0 + 0.5 & 0.0 & 100.0 & 0.0 + 0.6 & 0.0 & 100.0 & 0.0 + 0.7 & 0.2 & 99.8 & 0.0 + 0.8 & 1.1 & 98.9 & 0.0 + 0.9 & 3.3 & 96.7 & 0.0
+ we shall here carry out a simulation study to examine the finite - sample performance of our hypothesis testing procedure in section [ subsechypotest ] and the information criterion for variable selection in section [ subsecvarsel ] .
let @xmath291 be the @xmath292th order legendre polynomial and @xmath293 be the diagonal matrix with @xmath292th diagonal component @xmath294 .
let @xmath295 , @xmath296 , be i.i.d .
rademacher random variables and @xmath297 .
then @xmath298 , @xmath296 , forms a sequence of independent random vectors with correlated components .
let @xmath299 and @xmath300 .
consider : a linear model with heteroscedastic errors : for @xmath208 , @xmath301 a linear model with autoregressive effects : for @xmath208 , @xmath302 let @xmath303 and the bandwidth @xmath304 , @xmath305 .
we use the information criterion ( [ eqnvic ] ) to estimate the set of relevant predictors . for model ( ii ) , the whole set of potential predictors is taken to be @xmath49 along with three lags of the response variable .
a realization is categorized as under - fitting if we miss at least one relevant predictor , and over - fitting if the selected set contains at least one irrelevant predictor without under - fitting .
the results are summarized in table [ tabsimulationvic ] based on 5000 realizations .
given models ( i ) and ( ii ) , we use the hypothesis testing procedure to test whether @xmath306 has time - invariant contributions .
three types of weight matrices are used : the identity weights @xmath307 , the normalizer weights @xmath308 and the prediction weights @xmath309 . for each configuration , we use the simulation - assisted hypothesis testing procedure in section [ subsechypotestsimubased ] to obtain cut - off values @xmath310 and @xmath311 with 5000 simulated @xmath312 .
we then generate 5000 realizations of both models ( i ) and ( ii ) , and calculate the corresponding test statistic @xmath313 .
empirical acceptance probabilities are reported in table [ tabsimulationtesting ] .
@lcccccc@ & & & + & & & + @xmath314 & * 90% * & * 95% * & * 90% * & * 95% * & * 90% * & * 95% * + + 0.1 & 92.7 & 96.9 & 93.2 & 97.1 & 92.4 & 96.5 + @xmath285 & 92.2 & 96.5 & 92.9 & 96.7 & 92.0 & 96.4 + 0.2 & 91.7 & 96.0 & 92.4 & 96.2 & 91.7 & 95.9 + @xmath286 & 90.9 & 95.8 & 91.4 & 96.4 & 90.3 & 95.5 + 0.3 & 90.3 & 95.2 & 90.9 & 95.7 & 90.0 &
94.8 + @xmath287 & 90.5 & 95.1 & 90.9 & 95.8 & 90.3 & 95.1 + 0.4 & 90.9 & 95.5 & 91.7 & 96.1 & 91.1 & 95.5 + 0.5 & 90.4 & 94.8 & 91.3 & 95.2 & 90.4 & 95.0 + 0.6 & 90.7 & 95.3 & 91.2 & 95.5 & 90.7 & 95.2 + 0.7 & 91.6 & 95.6 & 91.7 & 95.7 & 91.5 & 95.7 + 0.8 & 89.1 & 94.9 & 89.0 & 94.8 & 89.0 & 94.8 + 0.9 & 89.9 & 95.0 & 90.4 & 95.4 & 89.8 & 94.9 + + 0.1 & 93.1 & 97.0 & 93.2 & 97.0 & 92.9 & 96.9 + @xmath288 & 92.6 & 96.6 & 92.2 & 96.1 & 92.1 & 96.2 + @xmath289 & 91.1 & 96.1 & 91.0 & 95.5 & 90.7 & 95.8 + 0.2 & 91.3 & 96.3 & 91.2 & 96.2 & 91.0 & 96.0 + @xmath290 & 90.9 & 95.4 & 90.8 & 95.7 & 90.7 & 95.2 + 0.3 & 90.8 & 95.1 & 89.9 & 94.8 & 90.4 & 95.1 + 0.4 & 90.0 & 95.1 & 89.8 & 94.9 & 89.9 & 95.1 + 0.5 & 89.5 & 94.6 & 88.2 & 93.9 & 89.2 & 94.5 + 0.6 & 88.6 & 93.6 & 87.7 & 93.1 & 88.5 & 93.6 + 0.7 & 88.6 & 93.8 & 87.4 & 93.6 & 88.1 & 93.8 + 0.8 & 87.4 & 92.6 & 87.0 & 92.2 & 87.4 & 92.7 + 0.9 & 88.3 & 94.4 & 88.1 & 94.0 & 88.2 & 94.3 + it can be seen that the empirical acceptance probabilities are fairly close to their nominal levels ( 90% and 95% ) , and the information criterion ( [ eqnvic ] ) performs quite well . in addition , the results are not sensitive to choices of weight matrices and bandwidths . for models ( i ) and
( ii ) , medians of the selected bandwidths based on the gcv criterion ( [ eqngcv ] ) are @xmath315 and @xmath316 , respectively .
observe that for model ( i ) , the performance is also quite satisfactory if we choose bandwidths @xmath317 , 0.25 and @xmath318 with @xmath319 .
a similar claim can be made for model ( ii ) as well .
we apply our model selection method to the hong kong circulatory and respiratory data which contains daily measurements of pollutants and hospital admissions in hong kong between january 1 , 1994 and december 31 , 1995 ( @xmath320 ) .
four pollutants , sulphur dioxide ( in @xmath321 ) , nitrogen dioxide ( in @xmath321 ) , dust ( in @xmath321 ) and ozone ( in @xmath322 ) , are considered here .
the purpose is to understand the association between daily hospital admission ( @xmath2 ) and levels of sulphur dioxide ( @xmath323 ) , nitrogen dioxide ( @xmath324 ) , dust ( @xmath325 ) and ozone ( @xmath326 ) .
figure [ figtsplots ] provides their time series plots . in the analysis
, we regularize the data so that each variable has zero mean and unit variance . letting @xmath327
be the intercept , we consider the time - varying coefficient model @xmath328 the dataset has been studied by @xcite by assuming that the observations are i.i.d . , while zhou and wu @xcite found substantial dependence among the fitted residuals .
we shall here model the process by ( [ eqnformulation ] ) and apply our model selection method in section [ secmain ] .
the selected bandwidth and tuning parameter are @xmath329 and @xmath330 , respectively .
the information criterion ( [ eqnvic ] ) selects the intercept ( @xmath331 ) , nitrogen dioxide ( @xmath324 ) and dust ( @xmath325 ) as relevant predictors .
fan and zhang @xcite did not consider the ozone effect ( @xmath326 ) and concluded that sulphur dioxide ( @xmath323 ) is not statistically significant .
we then apply the hypothesis testing procedure in section [ subsechypotestsimubased ] to examine whether the selected variables really have time - varying contributions . with 5000 simulated @xmath282 ,
the results are summarized in table [ tabapplication ] .
hence , at 10% significance level , we conclude that @xmath332 and @xmath333 are time - varying while @xmath334 can be treated as time - invariant , suggesting the model @xmath335 where @xmath336 , and @xmath337 and @xmath338 are plotted in figure [ figbetatilde ] .
@ld2.2cd3.2cd2.2c@ & & & + & & & + & & & & & & + @xmath332 & 69.77 & 0.00 & 120.77 & 0.02 & 69.77 & 0.00 + @xmath334 & 6.88 & 0.14 & 12.47 & 0.19 & 6.85 & 0.15 + @xmath333 & 16.27 & 0.02 & 30.13 & 0.09 & 23.06 & 0.01 + ( left ) and @xmath333 ( right ) . ]
for @xmath339 , write @xmath340 and @xmath341 . for a matrix @xmath16 , recall that @xmath342 and @xmath343 .
the proofs of the following two propositions are straightforward , and the details are omitted .
[ propmatrixnorm ] let @xmath344 be a real matrix
. then @xmath345 ; if @xmath346 has same dimension as @xmath16 , then @xmath347 ; if @xmath348 , then @xmath349 and @xmath350 ; and @xmath351 assume that @xmath16 is a nonsingular square matrix and that @xmath352 is a matrix with the same dimension . if @xmath353 , then @xmath354 is nonsingular and @xmath355 .
let @xmath356 , @xmath357 .
define the projection operator @xmath358 let @xmath359 be a zero mean process . write @xmath360 , @xmath208 .
lemmas [ lembndlin ] and [ lembndquad ] provide @xmath361-bounds for linear and quadratic forms of @xmath362 , respectively . to prove theorem [ thmcltahat ] , we need lemmas [ lembndlin ] and [ lemsupbnd ] .
[ lembndlin ] assume @xmath363 , @xmath364 .
write @xmath365 .
let@xmath366 , @xmath59 $ ] , be a sequence of real matrix functions , and define @xmath367 .
then : @xmath368^{1/q ' } \theta_{0,q}(\mathbf{j})$ ] ; @xmath369 } |\mathbf s_n(t)|\|_q \leq c_q n^{1/q ' } \mathcal{a}_n \theta_{0,q}(\mathbf{j})$ ] , where @xmath370 } [ \overline{\rho}\{\mathbf{a}_{1,n}(t)\ } + \sum_{k=1}^{n-1 } \overline{\rho}\{\mathbf{a}_{k+1,n}(t ) - \mathbf{a}_{k , n}(t)\}]$ ] .
let @xmath371 and @xmath372 , @xmath373 , form martingale differences . by the burkholder and the minkowski inequalities , we have @xmath374 since @xmath375 , ( i ) follows .
we now prove ( ii ) . by doob s inequality and the summation by parts formula
, we have @xmath376 } |\sum_{k=1}^n \mathbf{a}_{k , n}(t ) \mathbf d_{k , l , n}| \|_q \leq c_q \mathcal{a}_n n^{1/q ' } \delta_{l , q}(\mathbf{j})$ ] , entailing ( ii ) .
[ lembndquad ] assume @xmath377 , @xmath378 .
let @xmath379 be real matrices and @xmath380 .
then @xmath381 where @xmath382 .
let @xmath383 be the @xmath384-dependent approximated process and @xmath385 be the corresponding quadratic form .
if @xmath386 , @xmath387 .
hence @xmath388 where @xmath389 by lemma [ lembndlin ] and the arguments of proposition 1 in @xcite ,
we have @xmath390 . so lemma [ lembndquad ] follows .
[ lemsupbnd ] assume @xmath391 } \|\mathbf{j}(t;\bolds{\mathcal f}_0)\|_\iota < \infty$ ] , @xmath83 , and @xmath392 for some @xmath393 .
let @xmath394 .
then @xmath395 let @xmath396 . by theorem 2(ii ) in @xcite ,
there exist constants @xmath397 such that , for all @xmath398 and @xmath399 , @xmath400 \\[-8pt ] \nonumber & & \qquad\leq c_1 \frac{n b_n } { ( \lambda v_n)^\iota } + 2 \exp\bigl\{-(\lambda v_n)^2/(n b_n c_2)\bigr\}.\end{aligned}\ ] ] note that @xmath401 \subseteq\bigcup_{j \leq1/b_n } [ jb_n , ( j+1)b_n]$ ] . using the summation by parts formula , since @xmath402 has support @xmath403 $ ] , we have ( [ eqo19859a ] ) in view of ( [ eqnmaxinq ] ) and @xmath404\ ] ] by choosing a sufficiently large @xmath405
. for @xmath44 , let @xmath406 proof of theorem [ thmcltahat ] by lemma [ lembndlin ] , we have @xmath407 by @xmath384-dependence approximation , under conditions ( a2 ) , ( a3 ) and ( a4 ) , we obtain @xmath408 by lemmas [ lembndlin ] and [ lemsupbnd ] , and the argument in the proof of theorem 3 in @xcite , we have @xmath409 } |\tilde{\bolds{\beta}}(t ) - \bolds{\beta}(t)| = o_p(\varphi_n)$ ] and @xmath410 \\[-8pt ] \nonumber & & \qquad = o_p(\varphi_n \rho_n).\end{aligned}\ ] ] therefore , @xmath411 under our bandwidth conditions , @xmath412
. so theorem [ thmcltahat ] follows .
let @xmath413 .
lemma [ lemlrcov ] provides continuity properties of long - run covariance matrices for stochastically lipschitz continuous processes .
[ lemlrcov ] assume @xmath67 and @xmath414 .
then : for any nonnegative sequence @xmath415 , @xmath416 ; if , in addition , @xmath417 for some @xmath182 , then @xmath418 ; and if @xmath419 } \underline\rho\{\bolds{\lambda}(\mathbf{j } , t)\ } > 0 $ ] , then and hold for the inverse @xmath420 .
we first observe that @xmath421 & \leq & \sum_{s \in\mathbb{z } } \bigl\|\mathcal{p}_s \bolds \vartheta_i(t_1 ) \bigr\| \bigl\|\mathcal{p}_s \bolds \vartheta_j(t_2 ) \bigr\| \\ & \leq & \sum_{s \in\mathbb{z } } \delta_{i - s,2 } ( \mathbf{j } ) \delta_{j - s,2}(\mathbf{j } ) = \gamma_{|j - i|,2 } ( \mathbf{j}).\end{aligned}\ ] ] the lipschitz continuity implies @xmath422\bigr ) \leq c \bigl\{\gamma_{|i - j|,2}(\mathbf{j } ) \wedge|t_2 - t_1|\bigr\}\ ] ] uniformly .
hence @xmath423 which entails ( i ) by the dominated convergence theorem .
let @xmath424 which goes to infinity as @xmath425 .
since @xmath426 , we have @xmath427 , ( ii ) follows .
then ( iii ) follows by proposition [ propinversenorm ] .
let @xmath428 be a continuous mapping from @xmath85 $ ] to symmetric matrices in @xmath429 . for @xmath430
, define @xmath431 $ ] . before we prove theorem [ thmclttn ] , we shall first establish a parallel result for @xmath432 let @xmath433 and @xmath434 \gamma_{k,2 } ( \mathbf{l})$ ] .
[ lemclttndiamond ] assume @xmath435 and @xmath436 . if @xmath37 and @xmath437 , then @xmath438 \rightarrow n \bigl(0 , 4 k^*_2 \lambda_{\mathbf{w}_0,2}\bigr),\ ] ] and @xmath439 let @xmath440 and @xmath441 be its @xmath384-dependent counterpart and @xmath442 be the corresponding long - run covariance matrix .
then @xmath443 uniformly as @xmath444 .
let @xmath445 , @xmath446 , and @xmath447 . the central limit theorem ( [ eqnclttndiamond ] ) is a multivariate generalization of theorem a1 in @xcite by using propositions [ propmatrixnorm ] and [ propinversenorm ]
. we shall only detail steps that require special attention on the dimensionality .
essentially , we need to show that @xmath448 \\ & & \qquad = k_2^ * \lambda_{\mathbf{w}_0,2},\end{aligned}\ ] ] where @xmath449 . since @xmath450 , by lemma [ lemlrcov ] , we have @xmath451 \biggl(\int_0 ^ 1 w_{i , n}(t ) w_{j , n}(t ) \,dt \biggr)^2 \\ & & \qquad\quad { } + o\bigl\{n^2b_n \bigl(n^2b_n \bigr)^{-2}\bigr\ } + o\bigl\{\ell(m ) n^2b_n \bigl(n^2b_n\bigr)^{-2}\bigr\ } + o\bigl\{mn \bigl(n^2b_n\bigr)^{-2}\bigr\}\end{aligned}\ ] ] for some function @xmath452 as @xmath444 . then ( [ eqnclttndiamond ] ) follows . for ( [ eqntndiamondmean ] ) , by the proof of theorem 1 in @xcite , we have @xmath453 where @xmath454 since @xmath455 , ( [ eqntndiamondmean ] ) follows . proof of theorem [ thmclttn ] let @xmath456 \cup[1-b_n,1]$ ] .
lemma [ lembndlin ] implies @xmath457 and @xmath458 . by the proof of theorem 1 in @xcite
, we have @xmath459
. hence , @xmath460 by ( [ eqnbahadurrep ] ) and lemma [ lemsupbnd ] , we have @xmath461 and @xmath462 let @xmath463 . since @xmath464 under our bandwidth conditions , @xmath465 .
for ( [ eqnclttn ] ) , by lemma [ lemclttndiamond ] , it suffices to show that both @xmath466 and @xmath467 are of order @xmath468 . by the proof of lemma [ lemlrcov ] ,
@xmath469 since @xmath470 .
let @xmath471 .
then @xmath472 \gamma_{k,2}(\mathbf{l } ) \leq\frac{r_n(r_n+1 ) } { 2nb_n } \theta_{0,2}(\mathbf{l})^2 + \theta_{r_n,2}(\mathbf{l } ) \theta_{0,2}(\mathbf{l } ) = o \bigl(r_n^{-\nu}\bigr).\ ] ] hence , we have @xmath473 since @xmath128 , ( [ eqnclttn ] ) follows .
note that @xmath474 where @xmath475 , and by ( [ eqnllestapp ] ) and lemmas [ lembndlin ] and [ lemsupbnd ] , @xmath476 note that @xmath477 , and theorem [ thmclttn ] follows .
proof of proposition [ thmpower ] under the local alternative ( [ eqnlocalalternative ] ) , we have @xmath478 and @xmath479 where by ( [ eqnbahadurrep ] ) and lemmas [ lembndlin ] and [ lemsupbnd ] , @xmath480 the weight matrix @xmath481 and @xmath482 since @xmath483 , ( [ eqnpower ] ) follows from lemma [ lemclttndiamond ] .
recall that @xmath148 is a subset with complement @xmath484 , and @xmath147 is the true set of relevant predictors .
let @xmath485 , @xmath486 . then @xmath487 .
lemma [ lemrss ] provides bounds for @xmath488 for both cases @xmath489 and @xmath490 .
[ lemrss ] assume , @xmath126 , @xmath127 for some @xmath491 , @xmath37 and @xmath38 . then if @xmath492 , then @xmath493 and if @xmath490 , then @xmath494 for ( i ) , since @xmath492 , we have @xmath495 and @xmath496 where , by the proof of theorem [ thmcltahat ] , @xmath497 ^ 2 = o_p(n\varphi_n^2)$ ] and , by ( [ eqnbahadurrep ] ) and lemmas [ lembndlin ] and [ lemsupbnd ] , @xmath498 since , by lemma [ lembndquad ] , @xmath499 we have @xmath500 . since @xmath501 , ( i ) follows . for ( ii ) , since @xmath502 , we have @xmath503 where , by ( i ) , @xmath504 ^ 2 - \sum_{i=1}^n e_i^2 = o_p\bigl\{n\varphi_n ( \varphi_n + \rho_n)\bigr\}\ ] ] and , by lemma [ lembndlin ] and the argument on the quantity @xmath505 in ( i ) , @xmath506 \mathbf{x}_{\bar d , i}^\top\bolds{\beta}_{\bar d}(i /
n ) \cr & = & o_p\bigl(n^{1/2 } + b_n^{-1 } + n\varphi_n\rho_n + n^{1/2}b_n^2 \bigr).\end{aligned}\ ] ] in addition , by lemma [ lembndlin ] , @xmath507 lemma [ lemrss ] follows .
proof of theorem [ thmmodelselect ] by lemma [ lembndlin ] , @xmath508 . lemma [ lemrss ] implies @xmath509 for @xmath492 , and @xmath510 + o_p(1)\ ] ] for @xmath490 . since @xmath511 and @xmath512 , theorem [ thmmodelselect ] follows .
proof of proposition [ thmcovest ] by lemma [ lembndlin ] , @xmath513 } \bigl\|\hat{\mathbf{m}}(\mathbf{g } , t ) - e\bigl\{\hat { \mathbf{m}}(\mathbf{g } , t)\bigr\}\bigr\| = o\bigl\{(n\varpi_n)^{-1/2 } \bigr\},\ ] ] and , by lemma [ lembndquad ] , @xmath513 } \bigl\|\hat{\bolds{\lambda}}(\mathbf{l } , t ) - e\bigl\{\hat { \bolds{\lambda}}(\mathbf{l } , t)\bigr\}\bigr\| = o\bigl(\varrho_n^{1/2 } \bigr).\ ] ] by ( [ eqnbndt1t2 ] ) , we have @xmath514 proposition [ thmcovest ] follows by properties of local linear estimates . proof of proposition [ proptvar ] consider the process @xmath515 that satisfies the recursion @xmath516 then , for each @xmath78 $ ] , the process @xmath517 is stationary , and there exists a measurable function @xmath55 such that @xmath518 , @xmath519 . by condition ( t3 ) , @xmath520 } \overline\rho\{\mathbf{a}(t)\ } < 1 $ ] .
hence , by condition ( t2 ) and induction , we have @xmath521 \\[-8pt ] \nonumber & & { } + c \sum_{j=1}^{k-1 } \frac{j \rho_{\mathbf{a}}^j } { n } , \qquad k \geq 2.\end{aligned}\ ] ] since @xmath522 and @xmath523 , ( [ eqntvarmaxdiff ] ) follows by letting @xmath524 .
it suffices to show that @xmath525 . for this , by a similar argument of ( [ eqntvarmaxdiffk ] ) , we have for any @xmath526 , @xmath527 } \|\mathbf{z}_{t_1,i } - \mathbf{z}_{t_2,i}\| \leq\rho_{\mathbf{a}}^k \sup_{t_1,t_2 \in[0,1 ] } \| \mathbf{z}_{t_1,i - k } - \mathbf{z}_{t_2,i - k}\| + c |t_1-t_2| \sum_{j=0}^{k-1 } \rho_{\mathbf{a}}^j.\ ] ] since @xmath522 and @xmath528 , proposition [ proptvar ] follows by letting @xmath524 .
we are grateful to the editor , an associate editor , and two anonymous referees for their helpful comments and suggestions . | we consider parameter estimation , hypothesis testing and variable selection for partially time - varying coefficient models .
our asymptotic theory has the useful feature that it can allow dependent , nonstationary error and covariate processes . with a two - stage method , the parametric component can be estimated with a @xmath0-convergence rate .
a simulation - assisted hypothesis testing procedure is proposed for testing significance and parameter constancy .
we further propose an information criterion that can consistently the true set of significant predictors .
our method is applied to autoregressive models with time - varying coefficients .
simulation results and a real data application are provided . . . |
in an atomic force microscope , the motion of the tip as it is dragged across a substrate provides an efficient way to probe atomic friction of surfaces .
in fact , the moving tip can be regarded as a single asperity which determines the frictional force between macroscopic surfaces @xcite . under a constant load ,
the low - velocity motion of the tip on a surface exhibits stick - slip behavior as the moving tip hops over the corrugation potential defined by the substrate . for a quantitative description of friction at the microscopic level , theoretical understanding of the behavior of the tip on the substrate under different conditions
is required .
recent studies have argued that due to thermal fluctuations , the lateral frictional force increases logarithmically with increasing sliding velocity @xcite .
however , it is also found experimentally that the frictional force can have an opposite behavior , where it decreases with the sliding velocity depending on the ambient conditions @xcite and the nature of the substrate @xcite .
the origin of this decreasing velocity dependence is not fully understood and it is possible that distinct effects need to be taken into account to explain different experimental conditions .
current explanations of this behavior for point contact friction are based on extensions of the simplest effective model , the generalized tomlinson model @xcite , taking into account additional effects such as dissipation due to deformation of the afm tip @xcite or the inclusion of additional time scales for reorganization of the surface during sliding @xcite . in particular , the interpretation of recent experiments on surfaces covered with a monolayer in terms of an additional time scale due to surface restructuring @xcite still lacks detailed theoretical support . in this work , we study the velocity dependence of the frictional force within a simple theoretical model taking into account the coupling of the afm tip to excitation modes of the surface in addition to the coupling to the periodic corrugation potential .
we model the generic excitation modes by damped harmonic oscillators analogous to the model of an adsorbate coupled to phonon modes of the substrate @xcite .
we find that when the excitation mode frequency is comparable to the characteristic frequency of the particle moving across the surface potential , the velocity dependence of the frictional force is non monotonic displaying a velocity range where the frictional force can decrease with increasing velocity .
these results provide theoretical support within a simpler model for the interpretation of recent experiments @xcite which find a frictional force that decreases with velocity on surfaces covered with a monolayer .
the lateral motion of the tip can be described by a model of a single particle in an external one - dimensional potential representing the substrate and coupled elastically to the moving support .
the equation of motion for the particle is given by @xcite @xmath0 where the noise variables @xmath1 satisfy @xmath2 and @xmath3 is the total potential including the interaction of the tip to the substrate with periodic potential @xmath4\ ] ] and the elastic interaction between the tip and the moving support @xmath5 where @xmath6 is the position of the support .
the friction damping parameter @xmath7 in eq .
( [ tipeq ] ) results from the coupling of the tip to the surface excitations with a much shorter time scale than the motion of the tip ( e.g. electronic excitations ) which results in a @xmath8-function correlated random force @xmath1 with instantaneous markovian damping without memory as described by eq .
( [ noise ] ) .
the microscope support is moving at constant velocity , @xmath9 , where @xmath10 is the support velocity .
the average force on the support due to the tip motion is then given by @xmath11 this is regarded as the frictional force with the substrate since without damping we should have @xmath12 when averaging out the negative and positive force region of the periodic potential . for sufficiently large frictional damping stick and
slip motion is expected . as shown in ref .
, using this model the frictional force @xmath13 increases with increasing velocity @xmath10 . at small @xmath10
, it increases logarithmically but ultimately one would expect @xmath13 to be just @xmath14 , for sufficiently large @xmath10 .
we extend the model described by eq .
( [ tipeq ] ) by allowing the tip to couple to other excitations in the substrate which have a finite time scale .
these other excitations are modeled by damped simple harmonic oscillators as done in ref . .
these excitations can be labeled by the mode index @xmath15 where @xmath16 labels the symmetry group and @xmath17 runs from @xmath18 to @xmath19 where @xmath20 is the total number of modes . each mode
is characterized by a displacement variable @xmath21 , which couples to the tip motion .
the total hamiltonian can be written as @xmath22\end{aligned}\ ] ] the third term in the above hamiltonian represents a counterbalance term .
this is added so that when the tip moves very slowly relative to all the time scales given by @xmath23 , then at any instantaneous position of @xmath24 , @xmath25 and the actual effective potential felt by the tip is @xmath26 which is the adiabatic potential .
however , when the tip is moving very fast relative to such time scales , then the substrate has no time to respond , and one recovers the rigid substrate potential @xmath27 .
the explicit form of @xmath28 is chosen as @xmath29 the coupling @xmath30 in eq .
( [ coupl ] ) depends on the frequency @xmath31 of the mode .
this is chosen here to have the form @xmath32 so that the total effect of a finite number of damped oscillators corresponds to a continum distribution @xcite of excitation modes with a density of states @xmath33 , analogous to that for 3-d phonons in the debye model .
the equations of motion for the tip and the mode displacement variables can be written in a dimensionless form as @xmath34 where the additional noise variables @xmath35 for the damped harmonic oscillator also satisfy a fluctuation - dissipation relation @xmath36 without loss of generality , we can set @xmath37 for simplicity . in eq .
( [ dimeq ] ) , @xmath38 , @xmath39 , @xmath40 , @xmath41 and @xmath42 .
the lattice constant @xmath43 is taken as the unit of length , @xmath44 the unit of energy , and @xmath45 the unit of time .
we have performed numerical simulations of the coupled dynamical equations ( [ dimeq ] ) using brownian molecular dynamics @xcite .
the parameters for the spring and microscopic damping were fixed to @xmath46 and @xmath47 and for damped harmonic oscillators , @xmath48 and @xmath49 .
time steps ranged from @xmath50 to @xmath51 .
numerical results for the velocity dependence of the friction force obtained from eq .
( [ force ] ) assuming different angular frequencies @xmath31 for coupling to one odd or even mode are shown in figs 1 and 2 .
they are qualitative similar and show that when the frequency of the excitation mode is sufficiently small the velocity dependence of the friction force is non monotonic . while at sufficiently small or large velocities it increases with velocity
there is an intermediate range of velocities where the frictional force decreases as the velocity increases .
the approximate linear decrease in the semilog plot of figs 1 and 2 , suggest that this decrease is logarithmic . to understand this non monotonic behavior
it should be noted that there are several important time scales in the problem .
first , there is a velocity dependent time scale , @xmath52 , which is the average time for the tip to traverse one lattice period of pinning potential .
the pinning potential introduces a time scale , @xmath53 , the vibrational period of the tip in the well . in the stick and
slip regime where the tip is stuck in the well and perform many vibrational oscillations before hopping to the next minima , we have @xmath54 . the time for crossing the barrier is relatively fast , on the order of @xmath55 . with the coupling to a new excitation mode of frequency @xmath56
, another time scale @xmath57 is introduced .
we now consider what is the effect of varying @xmath52 on the frictional force @xmath13 . in the absence of the coupling to the excitation mode ,
the logarithmically increase of the friction force with velocity @xcite was obtained for velocities up to @xmath58 .
now consider what happen if we choose an excitation mode such that @xmath59 is several times larger than @xmath55 ( @xmath60 ) and we vary @xmath61 .
as far as the intrinsic frictional damping is concerned , when @xmath61 approaches @xmath59 , the damping should increase and also becomes non - markovian .
eventually , when @xmath61 becomes very large , the damping should saturate to a markovian value .
so based on this consideration , the effective @xmath7 should increase and saturate as a function of decreasing @xmath10 .
the frictional force @xmath13 should exhibit similar behavior .
in addition , the even mode can lead to a velocity dependent substrate relaxation that changes the effective potential as discussed previously .
however , for the present choice of parameters , the velocity dependence of the effective non - adiabatic frictional force due to the coupling to the new excitation mode dominates as evidenced by the results in fig .
1 and 2 , showing similar results for coupling to an even or odd modes .
the effect of varying the temperature at fixed mode frequency @xmath31 is also of interest . as can been from fig .
3 , the qualitative non monotonic behavior remains but there is an increase in the magnitude of the negative slope of the velocity - force curve as the temperature decreases .
we have also performed additional calculations to check the effect of including more excitation modes to model a more realistic continuum model of substrate excitations .
@xmath62 independent damped harmonic oscillators with frequencies in a range @xmath63 $ ] with @xmath64 were included in these calculations .
as can been seen from fig .
4 , the behavior for the velocity dependence for different maximum frequencies @xmath65 is qualitative similar to the behavior in fig . 1 at the same temperature .
the main effect of the additional modes below @xmath65 is a tendency to saturation of the frictional force at lower frequencies .
we study the velocity dependence of the frictional force of the tip of an atomic force microscope as it is dragged across a surface using a single particle model , taking into account memory effects and thermal fluctuations . in our model
, there is an additional coupling of the particle to excitation modes of the surface modeled by damped harmonic oscillators @xcite .
we find that when the excitation mode frequency is comparable to the characteristic frequency of the motion of the tip across the surface potential , the velocity dependence of the frictional force is non monotonic displaying an intermediate velocity range where the frictional force can decrease with increasing velocity .
therefore , the sliding behavior on the surface depends on the nature the surface through the characteristic frequency of such excitation modes .
recently , non - monotonic velocity dependence of friction was observed experimentally @xcite on surfaces covered with a monolayer .
the experimental results were interpreted as resulting from the chemical nature of the surface , with friction increasing or decreasing with velocity depending on the presence of cross - linked h - bonds .
it was argued that for higher sliding velocity there is not enough time for reordering to occur and the friction force should decrease logarithmically with velocity , while at lower velocities the breaking and reordering of the h - bonds contribute a new source of friction that increases with decreasing velocity .
our results provide theoretical support for this idea through an explicit calculation within a simple model .
the breaking of glassy domains of h - bonds at a critical stress and reordering introduces a new time scale , which corresponds in our model to the inverse frequency of the additional excitation mode of the surface .
the non - monotonic velocity dependence of the friction then follows naturally when the time scale of the motion of the tip across the surface becomes comparable to this new time scale for the excitation mode .
however , since we find a non monotonic behavior without invoking any glassy behavior , our results suggest that similar behavior should be found even on surfaces without disorder when additional excitation modes are present .
we thank t. ala - nissila for helpful discussions and suggestions .
e.g. was supported by fundao de amparo pesquisa do estado de so paulo - fapesp ( grant no . 07/08492 - 9 ) .
s.c.y . also acknowledges fapesp ( grant no .
09/01942 - 4 ) for supporting a visit to instituto nacional de pesquisas espaciais . | we study the velocity dependence of the frictional force of the tip of an atomic force microscope as it is dragged across a surface , taking into account memory effects and thermal fluctuations .
memory effects are described by a coupling of the tip to low frequency excitation modes of the surface in addition to the coupling to the periodic corrugation potential .
we find that when the excitation mode frequency is comparable to the characteristic frequency corresponding to the motion of the tip across the surface , the velocity dependence of the frictional force is non monotonic , displaying a velocity range where the frictional force can decrease with increasing velocity .
these results provide theoretical support for the interpretation of recent experiments which find a frictional force that decreases with velocity on surfaces covered with a monolayer . |
a microwave reentrant cavity consists of a cylindrical shaped structure with a narrow gap between a reentrant post and the cavity lid .
such a device presents an interesting electromagnetic field configuration with an azimuthal magnetic component around the post and an intense axial electric field in the gap spacing , which makes its dominant resonance frequency very sensitive to the gap dimensions . electrical quality factors ( q - factors )
up to 10@xmath0 have already been reported when operated in the superconducting regime and undergone an intense cleaning process @xcite .
without such a rigorous cleaning q - factors above @xmath1 are regularly achieved .
in fact , microwave reentrant cavities have being investigated over the past 50 years @xcite and repeatedly employed as transducers in resonant - mass gravitational wave detectors @xcite , finding also applicability in different research areas , such as solid state microwave oscillators @xcite , particle accelerators @xcite , dielectric characterization @xcite , electron spin resonance spectroscopy @xcite and test of fundamental physics @xcite .
most of the aforementioned applications would benefit from a microwave cavity with a tunable superconducting mode , fine tunability and large dynamic range .
for this reason , tunable cavities coupled to piezoelectric actuators have been developed for decades @xcite .
more recently , a system was presented that uses the central post of a reentrant cavity as a movable structure and was able to offer fine electronic tuning over a range of up to 139 mhz at cryogenic temperatures , but with low q - factors @xcite .
the new tuning mechanism developed in this work offers an improved dynamic range due to the mechanically amplified piezoelectric actuator ( apa ) manufactured by cedrat technologies sa .
we also gain a substantial increase in resonator q - factor by combining a superconducting cavity with a different design , where a movable cavity lid with a choke mechanism was used to control the resonance frequency .
this allowed the post to be fixed , reducing extra resistive losses associated with the post not being in proper electrical contact with the body of the resonant cavity .
the system is able to tune the cavity resonance frequency over ranges of several hundreds of mhz at 10 mk with q - factors higher than 10@xmath2 .
moreover , this system does not degrade with only a few photons on average within the cavity , a condition required for the operation of hybrid quantum systems and quantum electrodynamics experiments , which necessarily couple the resonator to superconducting quantum devices such a qubit or a squid . in this context , a 3d microwave cavity would offer advantages over on chip resonators , being able to supply a strong and localized electromagnetic field with a low loss microwave mode .
moreover , to control the interaction between the cavity and the quantum device usually a magnetic field will be applied . however , this ability is limited by the critical field of the superconducting electronics , thus an electronic way to tune a resonator over a large dynamic range and under cryogenic conditions is a valuable alternative , alleviating the necessity of large dc magnetic fields . aside from the tunability properties of the system ,
microwave nonlinear effects due superconducting - to - normal - state transition were observed to affect the cavity resonance .
such behavior can be associated to hot surface regions formation when the resonator is subject to high microwave power driving .
the detailed tunable system design is shown in fig .
[ fig1 ] ( a ) .
the cylindrical reentrant cavity is 3.7 mm high and its radius is equal to 6.4 mm with a central post whose diameter is 1 mm . the cavity is made of niobium and is located at the bottom of the system , where two magnetic probes can be coupled to the cavity field . around the cavity is the choke : a cut ring with inner radius equal to one quarter of the resonant mode wavelength and whose function is to reduce the energy leakage between the cavity top and bottom . ( a ) schematic of the microwave tunable cavity .
inset : three - dimensional illustration of the system .
( b ) apa frame deformation when a dc voltage is applied to the device .
( c ) schematic diagram of the experiment setup for cryogenic measurements ( hv : high voltage power amplifier ; vna : vector network analyzer ; pzt : piezoelectric actuator).,scaledwidth=48.0% ] the cavity lid is movable and not connected to the bottom structure , hence in order to tune the gap spacing between the post and the top wall the apa is attached to the lid .
then , by exciting the piezoelectric material through a difference of potential it has its initial length increased in the horizontal direction deforming the apa frame as in fig
. [ fig1 ] ( b ) . the frame structure supplies mechanical amplification to the piezoelectric displacement and changes the initial frame height from @xmath3 to @xmath4 , where @xmath3 is the initial vertical dimension . therefore , because the top of the frame is held by a fixed support , the cavity lid is pulled up by @xmath5 and the displacement is translated into frequency tuning . the electronic tuning mechanism is shown in fig .
[ fig1 ] ( c ) , where the diagram presents the tunable system located in the mixing chamber of the dilution refrigerator .
very thin power cables connect the apa to a high voltage power supply outside the refrigerator , providing up to 150 v , which is the maximum dc voltage supported by the actuator .
the tunable cavity was also designed to allow initial tuning of the gap , which can be changed by adding or removing spacers between the apa frame and its holder .
this enables the setting of an initial frequency between 6 ghz and 9 ghz , but requires the tunable system to be taken out of the refrigerator , as it can only be done manually . in summary ,
the initial frequency is set initially at room temperature and additional finer tuning with a dynamic range of up to a ghz can be provided by the electronic mechanism when the cavity is already cold .
fig . [ fig2 ] shows the measured frequency tuning range ( @xmath6 ) for room temperature ( rt ) and cryogenic ( mk ) experiments as a function of the initially set resonance frequency . in this plot @xmath6
is the frequency change only due the electronic tuning when the maximum voltage is applied to the apa .
the results can be compared to simulation calculated from eq .
( [ eq1 ] ) , where @xmath7 is the simulated tuning range , @xmath8 is the apa frame vertical displacement after 150 v was applied to the actuator and @xmath9 is the cavity displacement sensitivity : @xmath10 a finite element method ( fem ) model was used to simulate the relationship between the cavity frequency ( _ f _ ) and the gap spacing ( _ d _ ) . the curve used to calculate @xmath9 is presented in fig .
[ fig2 ] inset .
the @xmath8 was estimated by neglecting the cavity lid weight and adopting the minimum possible frame vertical displacement at 150 v , which according to the apa s manufacturer is 189 @xmath11 m .
experimental data were taken at room temperature and 10 mk .
the former demonstrates reasonable consistency with the simulations , with many of the points agreeing very well with the simulated data . a few measurements close to 6 ghz exhibit a small decrease in the tuning range in comparison to the simulations
this is attributed to unwanted deformation of apa frame during the assembling process .
deformations on cooling to cryogenic temperatures are harder to control , and hence the tuning range is harder to predict and exhibits some scatter .
previous results have shown the dynamic range of a piezoelectric actuator when cooled to cryogenic temperatures is reduced by about @xmath12 @xcite , in this case the reduction in tuning range compared to the simulated room temperature performance is 66% on average , showing a less significant reduction in sensitivity compared to the previous results . measured dynamic range of the frequency tuning as a function of the mechanically set initial frequency for room temperature and cryogenic measurements .
inset : cavity displacement sensitivity simulated by the fem model.,scaledwidth=42.0% ] ( a ) measured nonlinear resonance response due superconducting - to - normal - state transition .
( b ) measured resonator q - factor for different cavity input power .
( c ) calculated cavity photon number for different input power.,scaledwidth=48.0% ] the best tuning performance of the cryogenic trials shows that the cavity resonance frequency can be tuned over a range as large as 1 ghz .
the aim , however , is to achieve a large tuning range combined to a high electrical q - factor . from this perspective the best measured q - factor , equal to 10@xmath2 ,
occurs when the initial frequency is 6.1 ghz , where the tuning range is about 500 mhz .
such result can be easily explained by leakage through the gap between cavity top and bottom .
the choke was designed to minimize this effect ; however , it is less efficient as the gap is increased .
nonlinear effects were also observed when the resonator was set to 6.1 ghz and was operated at relatively high microwave powers .
[ fig3 ] ( a ) has the transmission measurement of the cavity resonance for different input power showing a clear deformation of the resonance shape with increasing power .
this effect besides distorting the lorentzian curve , also shifted the resonance peak to lower frequencies .
such behavior is often associated to the breakdown of the superconductivity regime on the resonator surface .
the heating caused by elevated power intensity can increase the material surface resistance destroying the superconductivity in some hot regions and thus leading to the normal - state transition . as a consequence ,
this effect will degrade the resonator q - factor and may change its resonance frequency @xcite .
the narrow gap spacing of the reentrant cavity is the most likely region to become non - superconducting due to the intense local electric field , which in turn will cause the temperature to rise , explaining the flattening and broadening of the resonance peak when the cavity is overdriven by high power .
[ fig3 ] ( b ) shows how the effect of hot regimes influences the resonator q - factor . from -5 to -15 dbm the resonator is predominantly in the normal regime , but superconducting regions begin to appear leading to a nonlinear response of the cavity resonance and , therefore , broadening the transmission spectrum as these regions spread over the resonator surface .
the calculated effective q - factor reduces with power during this part of the transition to the superconducting regime .
below -15 dbm the superconductivity starts to dominate and the normal - state regions diminish .
the q - factor now starts to grow until the resonator surface becomes completely superconducting at approximately -40 dbm .
after this , the reduction in power does not affect the cavity resonance anymore and a q - factor order of 10@xmath2 is achieved nearly independent on power .
an interesting feature such a 3d cavity can offer is the possibility to operate in a very low loss regime with a small amount of photons . in this sense
, the number of photons @xmath13 inside a cavity can be calculated by dividing the total electromagnetic energy into the cavity @xmath14 by the energy of a single photon , as @xcite : @xmath15 where @xmath16 and the total energy stored in the tunable cavity is given by @xmath17 in the eq .
( [ eq3 ] ) @xmath18 and @xmath19 are the coupling coefficients of the cavity input and output ports , respectively .
they must be considered because the calculation of the energy into the cavity has to account for the strength of its coupling to the external circuit , as well as the loading of the q - factor by it . at 6.1 ghz
the couplings were @xmath20 and fig .
[ fig3 ] ( c ) presents how the photon number would vary for different input power .
measurements were undertaken at 10 mk with an input power of -127 dbm ; such conditions combined with a q - factor of @xmath1 implicates that we achieved approximately 35 microwave photons on average stored within the cavity , just 15 db from the single photon level .
this level of photons has been shown to be sufficient for cavity characterization for quantum applications @xcite .
in this work we presented a finely tunable microwave three - dimensional cavity based on a reentrant design , which was electronically controlled over a range of several hundred of megahertz .
this system demonstrated effective performance at dilution refrigerator temperatures and capability of operation in the superconducting regime , leading to the achievement of high - q microwave modes .
the resonator also presented thermal nonlinear effects associated to the breakdown of the superconductivity due high power microwave driving .
in contrast , its low power operation illustrated the high - q operation with few photons in the cavity .
a superconducting fine - tunable 3d microwave cavity with such characteristics is a desirable device for quantum - enabled experiments and can offer practical contribution to the research area .
18ifxundefined [ 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 * * , ( ) link : www.scopus.com [ * * , ( ) ] , @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) | in the context of engineered quantum systems , there is a demand for superconducting tunable devices able to operate with high q - factors at power levels equivalent to only a few photons . in this work
, we developed a 3d microwave reentrant cavity with such characteristics ready to provide a very fine - tuning of a high - q resonant mode over a large dynamic range .
this system has an electronic tuning mechanism based on a mechanically amplified piezoelectric actuator , which controls the resonator dominant mode frequency by changing the cavity narrow gap by very small displacements .
experiments were conducted at room and dilution refrigerator temperatures showing a large dynamic range up to 4 ghz and 1 ghz , respectively , and were compared to a fem model simulated data . at elevated microwave power input , nonlinear thermal effects were observed to destroy the superconductivity of the cavity due to the large electric fields generated in the small gap of the reentrant cavity . |
the critical current at which a voltage appears along the length of a superconductor is one of the most important factors considered in applications of superconductivity @xcite .
the critical current is known to depend upon both local pinning centers in the material and the shape of the conductor s cross section . even in the absence of bulk
pinning , isolated type - ii superconducting thin - film strips subjected to perpendicular magnetic fields show magnetic hysteresis due to geometrical edge barriers @xcite .
such strips have finite critical currents arising from the edge barriers @xcite .
it also has been shown that the critical currents become larger when slits are fabricated near the edges of the strips @xcite ; the edge - barrier effects are enhanced because the slits increase the number of edges that can prevent flux penetration into the inner strips . although at subcritical applied fields and currents the magnetic response of two parallel strips in the meissner state is reversible @xcite , when the strips are connected at their ends and an applied magnetic field exceeds a certain value , magnetic flux penetrates irreversibly and the magnetic response becomes hysteretic as a consequence of the edge barriers @xcite . a detailed analysis of the effects of edge barriers upon the magnetization hysteresis in samples consisting of one , two , and three parallel strips connected at their ends has been presented in @xcite .
to calculate the combined effects of geometrical barriers and bulk pinning is more difficult , but this has been accomplished for single strips in @xcite , and a theoretical analysis of the magnetic - field dependence of the critical current of an isolated superconducting strip due to both an edge barrier and uniform bulk pinning has been presented in @xcite . here
we extend the above calculations to account for both geometrical edge barriers and bulk pinning in an infinite number of strips using the @xmath0-array method @xcite , by which the magnetic - field and current - density distributions for an array of parallel superconducting strips arranged periodically along the @xmath1 axis in the plane @xmath2 can be obtained analytically from the solutions for an isolated superconducting strip @xcite . in particular , we consider the case for which each strip in the array carries an equal amount of current in the presence of a perpendicular applied magnetic field .
we then calculate the critical current @xmath3 of each strip accounting for both a geometrical edge barrier , which impedes the entry of vortices into the strip , and uniform bulk pinning , which impedes the motion of vortices across the strip .
our calculation is relevant to a number of recent studies of the ac properties of striated coated - conductor tapes , i.e. , tapes that have been subdivided into parallel thin strips ( filaments ) separated by narrow gaps .
ideally , to minimize ac losses in power engineering applications using multifilamentary conductors , the individual filaments should not only have small cross sections but also be twisted and transposed , such that the filaments are inductively equivalent , are decoupled from each other , and thus carry equal current @xcite . in practice , it is not possible to satisfy all these conditions , and compromises are generally accepted for practical reasons .
carr and oberly @xcite have suggested that that the ac losses in tapes several millimeters wide could be reduced by first subdividing the tapes by striations and then twisting the tapes .
analytical and experimental results pursuing this idea have been reported in @xcite .
ashworth and grilli @xcite have recently proposed a variation of this approach , in which the tape is also subdivided into narrow parallel filaments , but instead of twisting the tape , the filaments are interrupted periodically by transverse cross - cuts bridged with normal metal .
magnetic flux can enter at the cross - cuts , thereby decoupling the filaments and allowing equal currents to flow in each filament .
the additional ohmic losses in the normal bridges are more than compensated by a huge reduction in the far more important ac coupling losses . in the present paper
we examine the dc properties of an infinite array of parallel superconducting strips , which can be regarded as an approximation to a striated coated - conductor tape of finite width .
our paper focuses on the possibility that a geometrical edge barrier can increase the critical current of each strip above that due to bulk pinning alone , but it also should be possible to use our results in calculating the hysteretic ac losses of an array of parallel strips as in @xcite . our paper is organized as follows . in sec.[x_arrays ] we review our complex - field approach and the @xmath0-array method . in sec .
[ bulk - pinning ] we apply the complex - field approach and the @xmath0-array method to obtain the complex field and the critical currents in a periodic array of parallel superconducting strips subject only to bulk pinning . in sec.[edge_barrier_only ]
we calculate the magnetic - field dependence of the critical current of one of the strips when the infinite array is subject only to geometrical edge barriers . in sec .
[ both_effects ] we calculate the critical current of one of the strips when both geometrical edge barriers and bulk pinning are present . we discuss the results and present our conclusions in sec.[conclusions ] .
we begin by reviewing the properties of a long superconducting strip of thickness @xmath4 centered on the @xmath5 axis in the region @xmath6 [ see figure 1 ] . and width @xmath7.,width=302 ] ( we attach tildes to all quantities related to the single strip . ) since we are not concerned with details of how the current density varies across the film thickness , we consider only the sheet - current density @xmath8 .
if the film thickness @xmath4 is less than the london penetration depth @xmath9 , we assume that the two - dimensional screening length ( or pearl length ) @xcite @xmath10 is much smaller than @xmath11 .
we consider the general case when the strip carries a current in the @xmath5 direction and is subjected to a perpendicular applied field @xmath12 in the @xmath13 direction .
outside the strip , the net magnetic field @xmath14 obeys @xmath15 and @xmath16 . for a two - dimensional problem such as this ,
it is convenient to introduce a complex magnetic field @xmath17 , which is an analytic function of the complex variable @xmath18 outside the strip .
the real and imaginary parts of @xmath19 obey the cauchy relations , @xmath20 and @xmath21 , which guarantee that @xmath22 obeys @xmath23 and @xmath24 , respectively .
the biot - savart law can be expressed as @xmath25 using the property that @xmath26 , where @xmath27 is a positive infinitessimal and @xmath28 denotes the principal value , we obtain @xmath29 and @xmath30 the net current carried in the @xmath5 direction is @xmath31 note from ( [ htilde ] ) that an expansion of @xmath19 in powers of @xmath32 yields @xcite @xmath33 .
solutions for the magnetic - field and current - density distributions for a long thin strip are known for many different physical situations .
it has been shown by mawatari in @xcite how known solutions @xmath19 and @xmath34 for a single isolated strip can be used to generate the corresponding solutions @xmath35 and @xmath36 for an @xmath0 array , i.e. , an infinite periodic array of identical coplanar strips of width @xmath37 and periodicity @xmath38 in the plane @xmath39 , as sketched in figure 2 . , width @xmath40 , periodicity @xmath38 , and separation @xmath41.,width=302 ] the complex field @xmath42 , which is an analytic function of the complex variable @xmath43 outside the strips , describes the components of the net magnetic field @xmath44 produced by a sheet - current density @xmath45 and a magnetic field @xmath46 applied in the @xmath47 direction .
the periodicity requires that the magnetic field and sheet - current density obey @xmath48 and @xmath49 .
corresponding solutions for the periodic complex field @xmath35 can be obtained from @xmath19 using the conformal mapping @xcite @xmath50 or its inverse @xmath51 these equations also describe the relations between @xmath52 and @xmath1 , @xmath11 and @xmath53 , or @xmath54 and @xmath55 , which will appear in later expressions .
then @xmath56 , @xmath57 , and the biot - savart law yields @xcite @xmath58du , \label{cot}\ ] ] where the magnetic field applied in the @xmath47 direction in the @xmath59 plane is @xcite @xmath60 and the current carried in the @xmath61 direction by one of the strips shown in figure [ xarray ] is @xmath62 note by comparing ( [ iz0tilde ] ) and ( [ iz0 ] ) that @xmath63 is in general not the same as @xmath64 .
the relations ( [ hahatilde ] ) and ( [ iz0 ] ) can be obtained from ( [ cot ] ) by applying the requirement that @xmath65 along the @xmath1 axis , we have the properties @xmath66 , @xmath67 , and @xmath68 in the following sections , we shall obtain @xmath35 using @xmath69 , @xmath57 , ( [ tzeta ] ) and ( [ lims ] ) , and various trigonometric identities .
when the critical current of a single superconducting strip is due solely to bulk pinning , characterized by the bulk - pinning critical current density @xmath70 ( here assumed to be field - independent , an assumption to be justified in sec .
[ conclusions ] ) , the corresponding critical sheet - current density is @xmath71 , and @xmath72 for @xmath73 , such that the critical current is @xmath74 and the complex field is @xmath75 the corresponding complex field surrounding the @xmath0 array shown in figure [ xarray ] is @xmath76}}{\sin{[\pi(\zeta - w)/l]}}\big ] , \label{hkp}\ ] ] where @xmath77 . since @xmath78 , we immediately find for one of the strips in the @xmath0 array that its critical current is @xmath79 , its average critical sheet - current density is @xmath80 , and its average critical current density is @xmath70 . note that when @xmath81 , the self - field contribution to @xmath82 is positive and diverges logarithmically when @xmath83 ; it is negative and has a similar divergence when @xmath84 . in the limit that @xmath85
, however , we recover the complex potential generated by an infinite film carrying a sheet current @xmath86 in a perpendicular applied field @xmath46 , @xmath87 where the upper ( lower ) sign holds when @xmath88 ( @xmath89 ) .
consider a single bulk - pinning - free superconducting strip in which the critical current is determined by a geometrical edge barrier ( or surface barrier at the edge ) .
as discussed in @xcite , the current and field distributions at the critical current have two possible forms , depending upon the strength of the applied field @xmath12 . for small values of @xmath12 ,
the distributions are simply the meissner response to the applied field and current , but for larger values of @xmath12 , there is a domelike magnetic field distribution inside the strip .
the complex magnetic field describing the meissner response to a magnetic field @xmath12 applied in the @xmath90 direction and to a net current @xmath91 applied in the @xmath61 direction is @xcite @xmath92 and the sheet - current density in the strip @xmath93 is @xmath94 if the strip contains a domelike magnetic - flux distribution in the region @xmath95 , where the sheet - current density is zero , the complex magnetic field is @xcite @xmath96^{1/2},\ ] ] the sheet - current density in the strip @xmath93 is @xmath97 and the net current in the @xmath5 direction is @xcite @xmath98 the corresponding quantities for the @xmath0 array are , for the meissner response to an applied field @xmath46 in the @xmath47 direction and an applied current @xmath99 in the @xmath61 direction , the complex field @xmath100^{1/2 } } \label{hi}\ ] ] and the sheet - current density in the strip @xmath101 @xmath102 where @xmath103 and @xmath104 .
if all the strips in the @xmath0 array contain domelike magnetic - flux distributions identical to the one in the region @xmath105 , the complex field is @xmath106^{1/2 } , \label{hdome}\ ] ] the sheet - current density in the central strip @xmath101 is @xmath107 where @xmath108 and the net current in the @xmath61 direction in the strip @xmath109 is @xmath110 .
\label{idome}\ ] ] the divergences in the above expressions for @xmath111 and @xmath112 at @xmath113 are artifacts of ignoring the finite thickness @xmath4 of the film .
it is well known that these divergences are cut off at a length scale @xmath114 , the larger of @xmath115 and @xmath116 [ i.e. , @xmath117 . to determine the critical current of one of the strips of the @xmath0 array , we use the approximations applied in @xcite and @xcite .
accordingly , we estimate the the sheet - current density at an edge by evaluating the diverging inverse square root in the expression for @xmath112 at a distance @xmath114 from that edge ; for example , for the edge at @xmath118 , in ( [ kzhi ] ) and ( [ kzdome ] ) we replace @xmath1 by @xmath53 in the numerators and by @xmath119 in the denominators and use the fact that the cut - off length scale obeys @xmath120 . to account for the edge barrier , we assume that vortices nucleate and enter the superconductor when the magnitude of @xmath112 at either edge of the strip reaches the critical value @xmath121 at which the barrier is overcome .
for an ideal edge , @xmath122 is equal to the ginzburg - landau depairing current density @xmath123 @xcite , but for an extremely defected edge , @xmath122 may become very small . applying this procedure to ( [ kzhi ] ) for @xmath124 , we obtain the following approximation to the zero - field surface - barrier critical current for any one of the strips in the @xmath0 array , @xmath125 where @xmath126 . alternatively , by estimating the local magnetic field at the edge of the strip @xcite , one obtains ( [ is00 ] ) but with @xmath127 , where @xmath128 is the critical field at which the barrier is overcome .
we expect that @xmath128 is at least as large as the lower critical field @xmath129 , but under favorable circumstances it may approach the bulk thermodynamic critical field @xmath130 . as a function of the applied field
@xmath46 , the surface - barrier critical current is found from ( [ kzhi ] ) to obey @xmath131 for small @xmath132 , where @xmath133 when @xmath134 , ( [ is00 ] ) and ( [ ish0 ] ) reduce to corresponding results found in @xcite and @xcite in low fields for isolated bulk - pinning - free strips . in the linear region of @xmath135
given in ( [ ish0 ] ) , the flux flow producing the voltage at a current just above @xmath136 can be described as the nucleation of vortices at @xmath137 as they overcome the geometrical edge barrier , followed by rapid motion across the strip and annihilation at @xmath138 .
this can occur only when @xmath139 for all @xmath109 . for increasing @xmath46 ,
however , as can be seen from ( [ kzhi ] ) , @xmath140 becomes zero at the critical current @xmath135 when @xmath141 . combining this condition with ( [ ish0 ] ) , we find that this occurs when @xmath142 , where @xmath143 and the subscripts ( @xmath144 ) denote the onset of a dome for zero bulk pinning . note that we also may write @xmath145 .
the linear portion of @xmath135 vs @xmath46 given in ( [ ish0 ] ) ends when @xmath142 or @xmath146 . for applied fields @xmath147 ( or @xmath148 ) ,
each strip contains a domelike magnetic field distribution at the critical current , and @xmath135 is no longer a linear function of @xmath46 .
the flux flow producing the voltage at a current just above @xmath136 can be described as the nucleation of vortices at @xmath137 , as they overcome the geometrical edge barrier , followed by rapid motion from @xmath137 to @xmath149 , the right boundary of the magnetic field dome .
the vortices inside the dome move collectively very slowly , and vortices at @xmath150 get pushed out of the dome and annihilate at @xmath138 .
we therefore calculate the critical current by setting @xmath151 in ( [ kzdome ] ) , evaluating @xmath112 at @xmath137 using the above approximation method , and setting it equal to @xmath152 . using ( [ is00 ] ) and ( [ h ] )
we then find that the value of @xmath153 at the critical current is @xmath154 , where @xmath155 = \tan(\pi w / l)/4h^2 . \label{tandome}\ ] ] ( note that @xmath156 when @xmath132 = 1/2 . ) substituting this into ( [ idome ] ) , we obtain @xmath135 for @xmath157 ( or @xmath158 ) .
the critical current of one of the strips of the @xmath0 array is thus given by @xmath159 when there is no bulk pinning and the critical current is determined only by a geometrical edge barrier ( surface barrier ) .
we expect our theory to be accurate only at sufficiently low applied fields that @xmath160 or @xmath161 , where @xmath162 is the reduced field at which @xmath163 . from ( [ tandome ] )
we obtain @xmath164^{1/2 } \geq ( w/2\lambda_c)^{1/2 } \gg 1 $ ] and @xmath165 .
in this section we calculate the complex fields and critical current @xmath3 of one of the strips in the @xmath0 array when the critical current is due not only to a geometrical edge barrier but also bulk pinning , characterized by a bulk - pinning field - independent critical sheet - current density @xmath80 .
an important parameter is the ratio , @xmath166 as in @xcite . ]
@xmath167 where @xmath168 [ see sec.[bulk - pinning ] ] is the critical current due to bulk pinning in the absence of an edge barrier , @xmath169 [ see sec . [ edge_barrier_only ] ]
is the critical current due to the edge barrier in the absence of bulk pinning when @xmath170 , and @xmath171 .
we first consider the case of relatively weak bulk pinning when @xmath172 .
in low fields ( @xmath173 or @xmath174 , region i of figure 3 ) , and bulk pinning parameter @xmath175 . in region
i , the strip is vortex - free [ @xmath176 . in region ii , there are two zones : a vortex dome ( @xmath177 ) and a vortex - free zone ( @xmath178 ) . in region
iii , there are three zones : a vortex dome ( @xmath105 ) between two vortex - free zones ( @xmath179 and @xmath178 ) . in region
iv , there are four zones : an antivortex dome ( @xmath180 ) where @xmath181 , a vortex dome ( @xmath182 ) where @xmath183 , and two vortex - free zones ( @xmath179 and @xmath178 ) . in each vortex - free zone @xmath184 , and under each vortex dome @xmath185 .
the curve @xmath186 separates regions i and iii , the curves @xmath187 separate regions iii and iv , and the curves @xmath188 separate regions ii and iii .
these curves are shown for @xmath189 ( solid ) , @xmath190 ( dot - dashed ) , and @xmath191 ( dashed).,width=302 ] vortices nucleate at the right edge of the strip ( @xmath118 ) when @xmath63 slightly exceeds @xmath135 . when @xmath36 in ( [ kzhi ] ) is greater than @xmath80 for all @xmath1 in the strip ( @xmath109 ) , nucleating vortices are driven entirely across the strip to the opposite side at @xmath192 , where they annihilate .
the critical current is then @xmath193 , and the normalized critical current is @xmath194 the same result as in ( [ ish0 ] ) .
thus if @xmath175 is small , the critical current @xmath195 for small @xmath46 is still completely dominated by the edge barrier and is independent of the strength of the bulk pinning . from ( [ kzhi ] )
we see that @xmath36 vs @xmath1 has a minimum , @xmath196 , at @xmath197 , where @xmath198 the minimum deepens [ i.e. , @xmath196 decreases ] with increasing @xmath46 , and @xmath199 when @xmath200 and @xmath201 or @xmath202 , such that @xmath203\tan \theta,\ ] ] where @xmath204.\ ] ] as a result , for small @xmath205 , the linear behavior of @xmath206 vs @xmath46 , given by @xmath207 vs @xmath132 in ( [ ici ] ) , holds only for @xmath208 . when @xmath205 and @xmath209 , or when @xmath210 and @xmath132 has any value , i.e. , for @xmath132 and @xmath175 outside region i of figure 3 , domelike magnetic - field distributions occur at the critical current @xmath206 . using the x - array method @xcite ,
we can obtain the complex field @xmath35 and associated sheet - current density @xmath36 in one of the strips of the @xmath0 array from the corresponding solution for an isolated strip @xcite . for the @xmath0 array , we find @xmath211 , where @xmath212 , \label{hzetai}\ ] ] where @xmath213^{1/2 } , \label{tildep}\ ] ] @xmath214 \label{tildeqa } \\ & = & -\frac{2(\tilde w-\tilde b)}{\sqrt{(\tilde w-\tilde a)(\tilde w+\tilde b)}}\big[{\mathbf{\pi } } \big(\frac{\tilde b-\tilde a}{\tilde w-\tilde a},\tilde q\big ) \nonumber \\ & & -\frac{(\tilde \zeta+\tilde w)}{(\tilde \zeta-\tilde b ) } { \mathbf{\pi}}\big(\frac{(\tilde \zeta-\tilde w)(\tilde b-\tilde a)}{(\tilde \zeta-\tilde b)(\tilde w-\tilde a)},\tilde q\big)\big ] , \label{tildeqb}\end{aligned}\ ] ] and @xmath215 the integral in ( [ tildeq ] ) is expressed in terms of complete elliptic integrals of the third kind @xcite @xmath216 , where @xmath217 is called either the characteristic or the parameter , and @xmath218 is called the modulus .
the mappings of ( [ tzeta ] ) and ( [ zeta ] ) define the relations between @xmath219 and @xmath59 , @xmath220 and @xmath221 , @xmath222 and @xmath153 , or @xmath11 and @xmath53 , and also guarantee that @xmath35 is periodic in the @xmath1 direction with period @xmath38 .
the applied magnetic field @xmath46 in the @xmath47 direction and the current @xmath63 carried in the @xmath61 direction by one of the strips shown in figure [ xarray ] , obtained from ( [ lims ] ) , obey @xmath223\ } \label{hadomes}\ ] ] and @xmath224\ } , \label{izdomes}\ ] ] where @xmath225 and @xmath226 the quantities of primary interest to us are @xmath227 and @xmath228 , @xmath229 , \label{hy } \\ & & \;\;\;\;\;\;\;\tilde a < \tilde x < \tilde b \;{\rm or}\ ; |\tilde x| > \tilde w , \nonumber \\ & = & 0 , -\tilde w < \tilde x < \tilde a \;{\rm or}\ ; \tilde b < \tilde x
< \tilde w.\end{aligned}\ ] ] @xmath230 , \label{kzleft } \\ & & \;\;\;\ ; -\tilde w < \tilde x < \tilde a , \nonumber \\ & = & k_p , \;\tilde a < \tilde x < \tilde b , \label{kzmiddle } \\ & = & + 2 \tilde p_0(\tilde a , \tilde b , \tilde x)[\tilde h_a + \frac{k_p}{2 \pi}\tilde q(\tilde a , \tilde b , \tilde x ) ] , \label{kzright } \\ & & \;\;\;\;\;\;\;\ ; \tilde b < \tilde x < \tilde w , \nonumber\end{aligned}\ ] ] where @xmath231 for the calculation of @xmath82 for @xmath105 ( i.e. , @xmath232 ) , taking the real part of @xmath233 in ( [ hy ] ) corresponds to taking the principal value of the integral of ( [ tildeq ] ) . for values of @xmath210 [ see ( [ p ] ) ] , and small values of @xmath132 , i.e. , for @xmath132 and @xmath175 in region iv of figure [ pvsh ] , the vortex distribution in each of the strips at the critical current can be described as a double dome , consisting of a vortex dome adjacent to an antivortex dome .
just above the critical current , vortices nucleate at @xmath137 , where @xmath234 , move rapidly to the left through an otherwise vortex - free region ( @xmath178 ) , and then move slowly to the left through a vortex - filled region ( the vortex dome ) , @xmath182 .
antivortices nucleate at @xmath192 , where @xmath235 , move rapidly to the right through an otherwise vortex - free region ( @xmath179 ) , and then move slowly to the right through an antivortex - filled region ( the antivortex dome ) , @xmath180 .
vortices and antivortices annihilate at @xmath236 , where the two domes meet and @xmath237 .
four equations must be solved simultaneously for @xmath221 , @xmath153 , @xmath12 , and the critical current @xmath206 for known values of @xmath132 , @xmath175 , and @xmath171 in region iv of figure [ pvsh ] .
one condition is that @xmath238 , which we evaluate by replacing @xmath52 in the denominator of ( [ tildep0 ] ) by @xmath239 , where @xmath240 and @xmath241 , and by replacing @xmath52 by @xmath11 in @xmath242 ( [ tildeqa ] ) and in the numerator of ( [ tildep0 ] ) . this yields from ( [ p ] ) and ( [ kzright ] ) @xmath243 where we use the normalized quantities @xmath244 and @xmath245 , as well as the definition @xmath246 a second condition is that @xmath247 , which we evaluate in a similar manner with the help of ( [ p ] ) , ( [ tildeqb ] ) , and ( [ kzleft ] ) .
the result is @xmath248 equations ( [ k+w ] ) and ( [ k - w ] ) have nearly the same form as ( 9 ) and ( 10 ) in @xcite and reduce exactly to those equations in the limit as @xmath249 or @xmath250 .
the third and fourth equations needed are obtained from ( [ hadomes ] ) and ( [ izdomes ] ) , which we write in dimensionless form at @xmath251 as @xmath252 , \label{icregioniv}\end{aligned}\ ] ] where the real quantities @xmath253 and @xmath254 are defined via @xmath255 and @xmath256 , @xmath132 , @xmath257 , and @xmath175 are defined via ( [ h ] ) , ( [ tildeh ] ) , and ( [ p ] ) , and @xmath258 . for given values of the strip width @xmath37 , periodicity length @xmath38 , dimensionless applied field @xmath132 , and dimensionless bulk pinning strength @xmath175 , numerical solutions of ( [ k+w ] ) , ( [ k - w ] ) , ( [ hregioniv ] ) , and ( [ icregioniv ] ) yield the strip s dimensionless critical current @xmath207 as well as the three other unknowns , @xmath220 ( or @xmath221 ) , @xmath222 ( or @xmath153 ) , and @xmath257
. shown in figure [ hyiv ] ( a ) and ( b ) are sample plots of @xmath82 and @xmath36 for @xmath132 and @xmath175 in region iv .
( in units of @xmath259 ) and ( b ) @xmath36 ( in units of @xmath80 ) vs @xmath1 ( in units of @xmath53 ) for @xmath260 , @xmath261 , and @xmath262 ( region iv ) , for which @xmath263 and @xmath264 , showing three strips ( @xmath265 @xmath266 and @xmath267 ) and the gaps around them ( @xmath268 , -@xmath269 , @xmath270 and @xmath271.,width=302 ] simplifications occur when calculating the self - field critical current when @xmath272 and @xmath46 = 0 ( @xmath273 ) , for which @xmath274 ( @xmath275 ) and @xmath82 is antisymmetric about the point @xmath276 .
equations ( [ tildep ] ) and ( [ tildeq ] ) then can be evaluated as @xmath277 @xmath278 , \label{tildeqequal}\ ] ] where @xmath279 is the complete elliptic integral of the first kind of modulus @xmath218 . in ( [ hadomes ] ) and ( [ izdomes ] ) , @xmath280 is pure real and @xmath281 is pure imaginary , such that from ( [ hregioniv ] ) we obtain @xmath282 when @xmath283 . equations ( [ k+w ] ) and ( [ k - w ] ) , resulting from the requirement that @xmath284 at the critical current , reduce to a single equation , @xmath285 and the expression corresponding to ( [ icregioniv ] ) for the reduced critical current @xmath286 becomes , with the help of ( [ tbph0 ] ) , @xmath287 , \label{iciv0}\ ] ] where @xmath171 . when @xmath288 , we find that @xmath289 , and @xmath290 for any value of @xmath291 . when @xmath292 , ( [ iciv0 ] ) reduces to @xmath293 plots of @xmath82 and @xmath36 vs @xmath1 for @xmath294 are similar to figure [ hyiv ] ( a ) and ( b ) , except that @xmath82 is an antisymmetric function of @xmath1 , centered at the origin , and @xmath36 is a symmetric function of @xmath1 . for increasing values of the applied field @xmath46 ( or the reduced field @xmath132 ) , the vortex dome expands and the antivortex dome shrinks . for reduced fields
@xmath132 in the range @xmath296 , we have @xmath297 for @xmath1 slightly larger than @xmath221 .
in other words , the coefficient of @xmath298 in ( [ hy ] ) is negative .
however , when @xmath299 , this coefficient becomes zero . for @xmath300 ,
this coefficient is positive , indicating that at the critical current , only a vortex dome is present ( region iii in figure [ pvsh ] ) .
the value of @xmath295 is determined chiefly by the condition that the coefficient of @xmath298 in ( [ hy ] ) is zero , which yields with the help of ( [ tildeqb ] ) , ( 17.7.7 ) of @xcite , and ( 414.01 ) of @xcite , @xmath301=0 , \label{h1}\end{aligned}\ ] ] where @xmath302 the value of @xmath295 for given values of @xmath210 and @xmath171 is obtained as the value of @xmath132 that satisfies ( [ h1 ] ) , ( [ k+w ] ) , and ( [ k - w ] ) .
the solutions of these equations also yield the values of @xmath303 and @xmath304 at the critical current when @xmath299 .
in region iii of the diagram of @xmath175 vs @xmath132 shown in figure [ pvsh ] , at the critical current , the field and current distributions within each strip divide into three zones . for @xmath178 , we have @xmath305 and @xmath306 ; for @xmath307 , we have @xmath308 and @xmath309 ; and for @xmath179 , we have we have @xmath305 and @xmath306 . just above the critical current @xmath3
, the voltage along the length of the strip is produced by ( a ) small numbers of vortices nucleated at @xmath310 as they overcome the geometrical edge barrier , followed by their rapid motion from @xmath137 to @xmath149 , ( b ) large numbers of vortices moving slowly through the vortex dome from @xmath149 to @xmath311 , and ( c ) small numbers of vortices leaking out of the vortex dome and moving rapidly from @xmath311 to @xmath192 .
the four equations determining the values of @xmath221 , @xmath153 , @xmath257 , and the reduced critical current @xmath286 for given values of @xmath132 , @xmath175 and @xmath291 in region iii are ( [ k+w ] ) , ( [ hregioniv ] ) , ( [ icregioniv ] ) , and ( [ h1 ] ) , since the latter equation also can be shown to give the condition that @xmath312 at @xmath311 .
shown in figure [ hyiii ] ( a ) and ( b ) are sample plots of @xmath82 and @xmath36 for @xmath132 and @xmath175 in region iii .
( in units of @xmath259 ) and ( b ) @xmath36 ( in units of @xmath80 ) vs @xmath1 ( in units of @xmath53 ) for @xmath260 , @xmath313 , and @xmath262 ( region iii ) , for which @xmath314 and @xmath315 , showing three strips ( @xmath265 @xmath266 and @xmath267 ) and the gaps around them ( @xmath268 , -@xmath269 , @xmath270 and @xmath271.,width=302 ] for increasing values of @xmath132 , the left boundary of the vortex dome moves ever closer to the left edge of the strip , i.e. , @xmath317 or @xmath318 , and numerical difficulties arise .
although other criteria could be used , in this paper we state that for practical purposes the vortex - free zone on the left side shrinks to negligible width when @xmath319 . for given values of @xmath175 and @xmath291 , this occurs at a reduced field @xmath320 , which can be obtained by solving ( [ k+w ] ) , ( [ h1 ] ) , and ( [ hregioniv ] ) for @xmath132 , @xmath257 , and @xmath153 after replacing @xmath220 by @xmath321 . for values of @xmath322 in figure
[ pvsh ] , all the quantities at the critical current can in principal be calculated using the same equations as for region iii .
however , these quantities can be calculated with fewer numerical difficulties and with high accuracy by setting @xmath151 in the above equations , which results in a number of simplifications .
the complex field still obeys @xmath56 , but now , to good approximation , @xmath323 , \label{hzetaii}\ ] ] where @xmath324 and @xmath325^{1/2}. \label{tildeqiia}\end{aligned}\ ] ] this equation is equivalent to ( 14b ) of @xcite , which was misprinted without the prefactor before the second @xmath326 term .
three equations must be solved simultaneously for @xmath222 ( or @xmath153 ) , @xmath257 , and @xmath286 for given values of @xmath132 , @xmath175 , and @xmath291 in region ii of figure [ pvsh ] .
the condition that @xmath238 , evaluated as in deriving ( [ k+w ] ) , becomes @xmath330 and the other two equations needed are ( [ hregioniv ] ) and ( [ icregioniv ] ) , in which the real quantities @xmath253 and @xmath254 are defined via @xmath331 and @xmath332 . shown in figure [ hyii ] ( a ) and ( b ) are sample plots of @xmath82 and @xmath36 for @xmath132 and @xmath175 in region ii .
( in units of @xmath259 ) and ( b ) @xmath36 ( in units of @xmath80 ) vs @xmath1 ( in units of @xmath53 ) for @xmath333 , @xmath334 , and @xmath262 ( region ii ) , for which @xmath335 and @xmath336 , showing three strips ( @xmath265 @xmath266 and @xmath267 ) and the gaps around them ( @xmath268 , -@xmath269 , @xmath270 and @xmath271.,width=302 ] using the above approach , one can calculate the normalized critical current @xmath337 ( [ is00 ] ) for arbitrary values of the reduced field @xmath132 ( [ h ] ) , pinning parameter @xmath175 ( [ p ] ) , and @xmath338 , where the strip width is @xmath37 and the periodicity length is @xmath38 . shown in figure [ icvsh ]
are sample results for @xmath339 , when the widths of the strips and the gaps are equal .
( critical current normalized to @xmath169 ) vs reduced field @xmath132 for @xmath339 and various values of the bulk pinning parameter @xmath175 .
the solid curves exhibit @xmath286 vs @xmath132 , the dashed curve shows @xmath286 at @xmath340 , and the dotted curve shows @xmath286 at @xmath341 . for @xmath205 ,
@xmath286 decreases linearly with @xmath132 ( [ ici ] ) up to @xmath342 and then decreases more slowly in regions iii and ii .
one curve shows @xmath286 for the special case of @xmath289 . for @xmath343
, @xmath286 increases by a few percent in the double - dome region iv and then decreases more gradually in regions iii and ii . in all cases ,
@xmath286 asymptotically approaches @xmath344 for large @xmath132 ( short dotted lines along the right side of the figure).,width=302 ] because of the effects of the geometrical barrier , at sufficiently low applied fields and small values of @xmath175 , the overall critical current @xmath3 can be considerably enhanced above the critical current @xmath345 due to bulk pinning alone .
this occurs because the sheet - current density @xmath36 in the vortex - free zones near the sample edges can carry a supercurrent with a density well above @xmath80 .
however , in high applied fields the cross section of the strip becomes nearly completely filled with vortices , and the sheet - current density in the vortex - filled zones can not exceed @xmath80 .
in this paper we have considered an infinite array ( x array ) of coplanar equally spaced superconducting strips and have calculated the critical current of each strip in the presence of a perpendicular applied field , taking into account not only geometrical barriers at the edges of the strips but also bulk pinning characterized by a field - independent critical sheet - current density @xmath80 .
we have carried out these calculations using the x - array method of @xcite , which enabled us to use solutions from @xcite for an isolated strip .
geometrical barriers at the edges of the strips enhance the critical current above what it would be if it were due to bulk pinning alone . at low fields ,
these barriers delay the entrance of vortices and permit the strip edges to remain vortex - free and carry a sheet - current density well in excess of @xmath80 . any vortices and antivortices entering the strips arrange themselves into domelike distributions , and under these domes the sheet - current density remains at @xmath80 .
as the perpendicular applied field increases , more and more vortices are forced into the strip , causing the vortex domes to expand and the high - current - density vortex - free zones to shrink , such that the overall critical current approaches that due to bulk pinning alone , @xmath346 .
the effects predicted in this paper for parallel arrays of narrow superconducting strips would be most easily observed experimentally using magneto - optical or scanning techniques in materials with low bulk pinning , such as bi-2212 ( bi@xmath347sr@xmath347cacu@xmath347o@xmath348 ) , where geometrical - barrier effects were first observed @xcite , or in low - pinning type - i superconductors , such as pb , where magnetic flux domes consisting of the intermediate state have been observed @xcite . however , it is important also to consider the possibility of observing these effects in materials with strong bulk pinning . as a practical application of the above theory to coated conductors made of superconducting ybco ( yba@xmath347cu@xmath349o@xmath350 ) ,
let us calculate the critical - current enhancement due to geometrical barriers for an isolated strip and then estimate how this enhancement is affected by striations .
we first consider a long flat strip of thickness @xmath4 = 1 @xmath351 m and width @xmath37 = 5 mm .
we assume that its self - field bulk - pinning critical current density is @xmath70 = 1 ma/@xmath352 ( 10@xmath353 a / m@xmath354 ) , such that the bulk - pinnng critical sheet - current density is @xmath355 a / m and the bulk - pinning critical current is @xmath356 a. we use ( [ is00 ] ) with @xmath357 , @xmath358 , and @xmath359 to estimate the zero - field surface - barrier critical current @xmath169 .
we use the experimental results of @xcite as discussed in @xcite , from which we estimate that @xmath129 = 180 oe = 1.4 @xmath360 a / m and @xmath130 = 3.2 koe = 2.6 @xmath361 a / m at 77 k. we use the conservative estimate that @xmath362 , which yields @xmath169 = @xmath363 = 4.5 a. the corresponding scaling field @xmath364 which appears in the denominator of ( [ h ] ) and ( [ tildeh ] ) , becomes @xmath365 3.6 oe = 286 a / m . the corresponding value of the pinning parameter is @xmath366 ,
for which we obtain from ( [ tbph0 ] ) and ( [ iciv00 ] ) that the self - field critical current @xmath3 at @xmath367 is very nearly equal to @xmath368 .
moreover , @xmath3 remains very nearly equal to to @xmath368 as the applied field increases .
our solutions show that the vortex - free regions are very close to the edges at the critical current ( i.e. , @xmath153 and @xmath369 are nearly equal to @xmath53 ) , such that the geometrical - barrier enhancement of @xmath3 is negligible .
next we suppose that the 5 mm strip is subdivided into 50 parallel strips , each of width @xmath37 = 98 @xmath351 m , separated by gaps of width 2 @xmath351 m , with period @xmath38 = 100 @xmath351 m . to approximate the behavior of the resulting striated conductor , we apply the above x - array results . using ( [ is00 ] ) with @xmath357 , @xmath358 , @xmath370 , and @xmath371 , we find that the zero - field surface - barrier critical current of one of the narrow strips is @xmath372 = 2.9 a. the corresponding bulk - pinning critical current is @xmath345 = 0.98 a , and the bulk - pinning parameter ( [ p ] ) is @xmath175 = 0.22 . since @xmath172 , the reduced critical
current is @xmath290 , and the self - field critical current of one of the strips is @xmath373 = 2.9 a , determined by the geometrical barrier alone .
since there are 50 such strips , the estimated total self - field critical current is 143 a. the corresponding engineering critical current density , taking the cross section to be 5 mm @xmath374 1 @xmath351 m ( ignoring the cross section of the substrate ) is @xmath375 a / m@xmath354 = 2.9 ma/@xmath352 , as opposed to the 1 ma / am@xmath354 critical current density for the unstriated strip .
the reduced critical current @xmath286 vs @xmath132 for this case is similar to that shown in figure [ icvsh ] for @xmath175 = 0.2 .
however , the scaling field for @xmath286 vs @xmath132 from ( [ hscale ] ) is small : @xmath376 = 5.7 oe = 450 a / m . to summarize , the x - array results using the above assumptions predict that the self - field critical current for the striated conductor can be significantly enhanced above that due to bulk - pinning alone ( by approximately a factor of three using the above numbers ) .
however , the application of a relatively small perpendicular magnetic field ( of the order of tens of oe or hundreds of a / m using the above numbers ) can be expected to produce a strong reduction of the enhancement and to return the critical current nearly to that due to bulk - pinning alone . throughout this paper
we have assumed that the bulk - pinning critical current density @xmath70 is field - independent .
we now justify this assumption as follows .
we see from figure [ icvsh ] that the critical current has a significant dependence upon @xmath377 only when @xmath378 . in the above two paragraphs we found for state - of - the - art ybco coated conductors for which @xmath379 ma/@xmath352 ( 10@xmath353 a / m@xmath354 ) that @xmath380 oe = 796 a / m or @xmath381 10 g = 1 mt . experimentally it has been found , for example in @xcite , that @xmath70 at 77 k in strong - pinning ybco films is very nearly independent of field for @xmath382 100 oe @xmath383 8000 a / m or @xmath384 g = 10 mt .
therefore it is an excellent approximation to assume that @xmath70 is field - independent over the range of applied fields for which the effects of geometrical barriers or edge pinning are relevant . in @xcite
, we argued that in state - of - the art unstriated ybco coated conductors the pinning is due almost entirely to bulk pinning and that geometrical barriers ( or edge pinning ) have a negligible effect upon the critical current .
the above calculations indicate that for striated coated conductors the additional edges produced by the striations could produce significant enhancements of the critical current in self - field , but these enhancements are strongly suppressed by relatively small applied magnetic fields . in applications where the magnetic flux density is of the order of 0.1 t or higher , where @xmath70 has a strong field dependence , we expect that geometrical barriers will have no significant effect upon ybco coated conductors , even if they are striated .
99 campbell a m and evetts j e 1972 _ critical currents in superconductors _ (
taylor & francis , london ) indenbom m v , kronmller h , li t w , kes p h and menovsky a a 1994 _ physica c _ * 222 * 203 schuster th , indenbom m v , kuhn h , brandt e h and konczykowski m 1994 _ phys .
lett . _ * 73 * 1424 zeldov e , larkin a i , geshkenbein v b , konczykowski m , majer d , khaykovich b , vinokur v m and shtrikman h 1994 _ phys . rev
. lett . _ * 73 * 1428 benkraouda m and clem
j r 1996 _ phys .
b _ * 53 * 5716 doyle t b , labusch r and doyle r a 1997 _ physica c _ * 290 * 148 maksimov
i l and elistratov a a 1998 _ appl .
lett . _ * 72 * 1650 benkraouda
m and clem j r 1998 _ phys .
* 15103 mawatari y and clem j r 2001 _ phys .
lett . _ * 86 * 2870 babaei brojeny a a , mawatari y , benkraouda m and clem j r 2002 1454 zhelezina n v and maksimova g m 2002 _ tech .
_ * 28 * 618 ainbinder r m and maksimova g m 2003 871 maksimova g m , zhelezina n v and maksimov
i l 2004 _ physica c _ * 402 * 53 mawatari y and clem j r 2003 _ phys .
b _ * 68 * 0204505 maksimov
i l 1995 _ europhys .
lett . _ * 32 * 753 maksimov
i l and elistratov a a 1995 _ jetp lett . _
* 61 * 208 elistratov a a and maksimov
i l 2000 _ phys
. solid state _ * 42 * 201 maksimova g m , zhelezina n v and maksimov
i l 2001 _ europhys .
lett . _ * 53 * 639 kupriyanov m yu and likharev k k 1974 _ fiz .
tela _ * 16 * 2829 ( 1975 _ sov .
phys . solid state _ * 16 * 1835 ) elistratov a a , vodolazov d yu , maksimov
i l and clem j r 2002 _ phys .
b _ * 66 * 220506(r ) ; 2003 _ phys . rev .
b _ * 67 * 099901(e ) mawatari y 1996 _ phys . rev . b _ * 54 * 13215 wilson m n 1983 _ superconducting magnets _ ( clarendon press , oxford ) section 7.3 carr w j jr and oberly c e 1999 _ ieee trans . appl . supercond . _ * 9 * 1475 oberly c e , long l , rhoads g l and carr w j jr 2001 _ cryogenics _ * 41 * 117 cobb c b , barnes p n , haugan t j , tolliver j , lee e , sumption m , collings e and oberly c e 2002 _ physica c _ * 382 * 52 polak m , krempasky l , chromik s , wehler d and moenter b 2002 _ physica c _ * 372 - 376 * 1830 amemiya n , kasai s , yoda k , jiang z , levin g a , barnes p
n and oberly c e 2004 _ supercond .
technol . _ * 17 * 1464 barnes p n and sumption m d 2004 _ j. appl .
phys . _ * 96 * 6550 levin g a , barnes p n , amemiya n , kasai s , yoda k , jiang z and polyanksii a 2005 _ j. appl .
_ * 98 * 113909 wang l b , price m b , young j l , kwon c , levin g a , haugan t j and barnes p n 2005 _ physica c _ * 419 * 79 sumption m d , barnes p n and collings e w 2005 _ ieee trans
supercond . _
* 15 * 2815 majoros m , glowacki b a , campbell a m , levin g a , barnes p n and polak m 2005 _ ieee trans .
supercond . _ * 15 * 2819 tsukamoto o , sekine n , ciszek m and ogawa j 2005 _ ieee trans .
supercond . _
* 15 * 2823 levin g a , barnes p n , kell j w , amemiya n , jiang z , yoda k and kimura f 2006 _ appl .
lett . _ * 89 * 012506 demencik e , usak p , takacs s , vavra i , polak m , levin g a and barnes p n 2007 _ supercond .
technol . _
* 20 * 87 polak m , usak e , jansak l , demencik e , levin g a , barnes p n , wehler d and moenter b 2006 _ preprint _ cond - mat/0602422 levin g a , barnes p n and amemiya n 2006 _ preprint _ cond - mat/0609480 ashworth s p and grilli f 2006 _ supercond .
_ * 19 * 237 mawatari y 1997 in _ advances in superconductivity ix _
, ed nakajima s and murakami m ( springer , tokyo ) p 575 mller k - h 1997 _ physica c _ * 289 * 123 mller k - h 1999 _ physica c _ * 312 * 149 this paper also corrects typographical errors in expressions for the ac losses given in @xcite .
pearl j 1964 _ appl .
_ * 5 * 65 zeldov e , clem j r , mcelfresh m and darwin m 1994 _ phys . rev
b _ * 49 * 9802 aslamazov l g and lempicki s v 1983 _ zh .
phys . _ * 84 * 2216 vodolazov d yu , maksimov
i l and brandt e h 1999 _ europhys .
* 48 * 313 gradshteyn i s and ryzhik i m 2000 _ table of integrals , series , and products , 6th ed _ , ed a jeffrey and d zwillinger ( academic press , san diego ) 1967 _ handbook of mathematical functions _ , ed m abramowitz and i a stegun ( national bureau of standards , washington ) 2005 _ mathematica , version 5.2 _ , wolfram research , inc . , champaign , il selfridge r b and maxfield j e 1958 _ a table of the incomplete elliptic integrals of the third kind _ ( dover , new york ) byrd p f and friedman m d 1954 _ handbook of elliptic integrals for engineers and physicists _ ( springer , berlin ) castro h , dutoit b , jacquier a , baharami
m and rinderer l 1999 _ phys .
b _ * 59 * 596 hao z , clem j r , mcelfresh m w , civale l , malozemoff a p and holtzberg f 1991 _ phys . rev .
b _ * 43 * 2844 babaei brojeny a a and clem j r 2005 _ supercond .
technol . _ * 18 * 888 dam b , huijbregtse j m , klaassen f c , van der geest r c f , doornbos g , rector j h , testa a m , freisem s , martinez j c , stuble - pmpin b and griessen r 1999 _ nature _ * 399 * 439 gutirrez j , llords a , gzquez j , gibert m , rom n , ricart s , pomar a , sandiumenge f , mestres n , puig t and obradors x 2007 _ nature materials _ | we calculate the magnetic - field dependence of the critical current due to both geometrical edge barriers and bulk pinning in a periodic coplanar array of narrow superconducting strips .
we find that in zero or low applied magnetic fields the critical current can be considerably enhanced by the edge barriers , but in modest applied magnetic fields the critical current reduces to that due to bulk pinning alone . |
the ` particle in a box ' has served as a model for understanding various phenomena in solid state and nuclear physics the theory of dynamical systems terms these as ` billiards ' .
studies of their energy spectra and eigenfunctions , and their connections with quantum chaos have been very fruitful and exciting .
one of the properties of the eigenfunctions of these billiards is the organisation of regions with positive and negative signs .
these domains appear in rather complex forms @xcite ; their number displays a near - incomprehensible order if organised in increasing energy .
we are familiar with domains that appear in a system which has two states or phases .
for instance , in magnetic materials , there are regions of positively and negatively aligned spins .
their shapes and areas promise interesting statistical questions .
there has been a lot of interest in studying the nodal domain statistics in recent times of billiards in two dimensions @xcite .
here , we present a general result for the number of nodal domains of integrable plane polygonal billiards the geometries are rectangle , circle , ellipse , and triangles with angles ( 45 , 45 , 90 ) , ( 30 , 60 , 90 ) and ( 60 , 60 , 60 ) .
the great interest in these systems emerges from the simplicity they seem to present , and their ubiquity in a large number of contexts . the schrdinger equation for a particle inside a rectangular box satisfying dirichlet or neumann boundary conditions can be immediately solved .
the eigenfunctions are the well - known product of two sine or cosine functions and the nodal domains make a checkerboard . for the right isosceles and the equilateral triangle billiards , which are non - separable , the solutions are respectively two and three terms , each a product of two sinusoidal functions . for the equilateral triangle
, we discovered a difference equation satisfied by the number of nodal domains , @xmath5 @xcite , a discrete variable of the system . in this article , we show that @xmath5 for all the integrable billiards obey similar difference equations .
this general result is amazing , in view of widely varied findings observed for these billiards about various statistical measures @xcite .
interestingly , recent investigations of integrable lattice systems @xcite have also sought to probe the intricate connection between the theory of exactly integrable discrete systems and the formalism of difference equations .
let @xmath6 be a compact , connected domain on a surface with a smooth riemannian metric in two dimensions .
assuming dirichlet conditions along the boundary @xmath7 and denoting the laplace
beltrami operator by @xmath8 , the eigenvalue problem is formulated as @xmath9 a nodal domain of the wavefunction @xmath10 is a connected domain in @xmath11 where @xmath12 , which therefore defines a maximally connected region wherein the function does not change sign .
the subscript @xmath13 simply denotes the ordering of the spectrum such that @xmath14 .
the importance of the nodal set arises from the fact that the sequence of the number of nodal domains of the eigenfunctions of the schrdinger equation not only bears significant geometric information about the system @xcite but also provides a new criterion for chaos in quantum mechanics @xcite .
hence , it may be reasonable to conjecture that the difference equations also encode the geometry of the system itself .
any wavefunction @xmath10 on the domain @xmath11 is characterised by two quantum numbers for manifolds on @xmath15 which are represented hereafter as @xmath16 and @xmath17 , unless specified otherwise , where @xmath18 .
let @xmath19 denote the total number of nodal domains of the wavefunction @xmath20 .
furthermore , let @xmath21 be an equivalence relation defined on the set of wavefunctions as @xmath22 the relation @xmath21 defines a partition @xmath23 of the set of wavefunctions into equivalence classes @xmath24 $ ] where @xmath25 . here
, the parameter @xmath3 represents the number of linearly independent terms of which the wavefunction @xmath26 is a sum .
consideration of the sequence of @xmath19 for wavefunctions that belong to the same class illustrates the rich structure of the difference equations that arise in two - dimensional integrable billiards . to this end
, we introduce the forward difference operator , @xmath27 , operating on the first index @xmath28 of a generic function @xmath29 , with a finite difference @xmath30 , i.e. @xmath31 . _ if the metric space @xmath6 is integrable , then , in the absence of tiling , one of @xmath32 and @xmath33 holds for some @xmath34 , which is determined only by the geometry of the billiard . _
[ one ] the statement can be easily demonstrated by verifying it individually for all possible integrable billiards on @xmath15 .
the corresponding functions @xmath35 have been calculated in the respective sections .
the major utility of this theorem lies in the fact that it presents a method to conveniently compute expressions for the total number of nodal domains , even for non - separable billiards , which may seem intractable otherwise .
the reader is requested to note that the authors have taken the liberty of a slight abuse of notation in the sense that @xmath36 and @xmath37 have both been used interchangeably to denote the number of nodal domains , depending on whether the quantity is regarded as a numeric variable or as a function , respectively .
birkhoff conjectured that among all billiards inside smooth convex curves , ellipses are characterised by integrability of the billiard map @xcite .
examples of systems that are both integrable and separable are presented by the rectangular , circular and elliptic billiards .
separability arises from the observation that the corresponding schrdinger equations on @xmath11 are separable in certain coordinate systems @xcite .
owing to the structure of the nodal pattern , which is formed by a grid of intersecting nodal lines for both the billiards , the expressions for @xmath19 are relatively uncomplicated being equal to the number of nodal crossings .
the wavefunctions for both the billiards comprise of only a single term and hence @xmath38 , which implies that the equivalence class is defined as @xmath39 .
although such a construction of the class may seem artificial for the rectangular and circular billiards given the simple expressions for @xmath19 , the formulation helps to naturally extend a seamless transition to the non - separable cases where no such exact formulae are known .
in fact , the simplification obtained on approaching the nodal domain problem through the method of difference equations becomes perceptible for the elliptic billiard , for which the expression for @xmath19 is not trivially obvious .
[ h ] in the following discussion , we only consider billiards with the imposition of dirichlet boundary conditions , in which the wavefunctions vanish at the boundary , because such systems directly correspond to the model of a ` particle in a box ' through the physical assumption of an infinite potential outside the relevant region of the billiard .
for neumann boundary conditions , the generalisation of the results presented herein is trivial for separable integrable systems , since the same difference equations and expressions for @xmath36 hold .
however , the extension of these relations to the non - separable integrable billiards when neumann boundary conditions are imposed would certainly be an interesting exercise .
the nodal domain distribution for rectangular billiards has been considered in detail in @xcite .
let @xmath40 \times [ 0,l_y]$ ] be a rectangular drum such that @xmath41 .
the eigenfunctions for the system are given by @xmath42 choosing @xmath43 , the dirichlet spectrum simplifies to @xmath44 .
examination of the ` checkerboard ' pattern of nodal domains easily yields the number of nodal domains of a particular eigenfunction to be @xmath45 .
the difference equation satisfied by this system is @xmath46 the recurrence relation of eq .
can be solved analytically to obtain @xmath47 , where @xmath48 is a constant .
comparison with numerical results shows that @xmath48 = 0 .
the circle corresponds to a special case of an ellipse , which is an integrable billiard @xcite .
a circular domain of radius @xmath49 in two dimensions , physically corresponding to a circular infinite well , may be specified as @xmath50 . in polar coordinates ,
the solutions of the schrdinger equation ( for a particle of mass @xmath51 ) are separable into angular and radial components and are given by @xcite as @xmath52 ^ 2 r\ , \mathrm dr } } .\frac{\cos
m \theta}{\sqrt{\pi}};\:\ : k = \sqrt{\frac{2\mu e}{\hbar^2 } } , \label{eq : circ}\ ] ] where @xmath53 denotes the cylindrical bessel function of the first kind , which has non - divergent solutions as @xmath54 . the energy spectrum for the system is therefore @xmath55 ^ 2 $ ] , with @xmath56 representing the @xmath57 zero of the regular bessel function @xmath58 @xcite .
since the angular component of the eigenfunction , @xmath59 , is complex , the nodal domains are visualised for eq . by considering solely the real part of the wavefunction .
the results derived in this section remain invariant on consideration of the imaginary part .
study of the nodal pattern , which bears structural similarity to that of the rectangle , indicates the existence of a comparable difference equation for the nodal counts as @xmath60 the analytical solution to this recurrence relation is nearly identical to that for eq . except for a prefactor and
is given by @xmath61 , where @xmath62 is a constant .
on examining the numerically obtained number of nodal domains , it is observed that @xmath63 and @xmath36 simplifies to @xmath64 . when @xmath65 , the number of nodal domains is simply @xmath17 and the appropriate modification to eq . would therefore read as @xmath66 .
c c c | c c | c c @xmath16 & @xmath17 & @xmath67 & @xmath68 & @xmath69 & @xmath70 & @xmath69 14 & 11 & 3 & 154 & & 308 &
+ 25 & 11 & 3 & 275 & 121 & 550 & 242 + 36 & 11 & 3 & 396 & 121 & 792 & 242 + 57 & 11 & 3 & 517 & 121 & 1034 & 242 + 68 & 11 & 3 & 638 & 121 & 1276 & 242 + it is interesting to note that the bessel function @xmath53 tends to the standard trigonometric functions as @xmath71 , i.e. the function of the first kind resembles a sine or cosine curve , with a period that slowly shortens , eventually becoming @xmath72 . in the limit of @xmath71 ,
the boundary of a circle with a sufficiently large radius of curvature @xmath73 can be locally approximated by a straight line and thus the eigenfunctions , and consequently the nodal patterns , of the rectangular and circular billiards are intrinsically linked .
the fact that the difference equations and retain this connection corroborates the non - triviality of these recurrence relations and seems to suggest the presence of a more fundamental underlying reason for their existence .
the notions introduced in the treatment of the circle are generalised for the elliptic billiard , which is separable in elliptic coordinates , @xmath74 and @xmath75 , and the solutions thereof are described in terms of the radial and angular mathieu functions @xcite .
the radial and angular quantum numbers , @xmath76 and @xmath77 respectively , describing the system are defined in accordance with the notational convention for parity adopted in @xcite as this is in agreement with the einstein - brillouin - keller ( ebk ) quantisation for the symmetry reduced system . for the purpose of demonstration of the difference equations , we consider the states with even parity @xmath78 for which @xmath79 ( refer to figure 6 of @xcite ) . for @xmath80
, we have @xmath81 the solution to which is @xmath82 .
additionally , for the states where @xmath83 , it is to be noted that @xmath84 , which trivially implies that @xmath85 .
it is indeed straightforward to verify that this construction can be extended to the other parity combinations , @xmath86 ( figure [ a1e].c ) , @xmath87 and @xmath88 .
an interesting discussion on the hardening and softening of the pseudoradial and pseudoangular quanta respectively during the transition from the circle to the rectangle is presented in @xcite which underlines the inherent connection amongst these separable billiards and hence , between their ` checkerboard ' nodal patterns .
the eigenfunctions of the separable systems discussed in the previous section present a behaviour which is different from ` generic ' wavefunctions , the nodal lines of which , according to uhlenbeck s hypothesis , do not intersect @xcite .
the eigenfunctions of classically integrable but non - separable systems show avoided intersections and a few crossings of nodal lines @xcite , which greatly complicates the expression for @xmath19 .
the primary focus of our discussions in the following sections with respect to the non - separable integrable billiards has exclusively been the non - tiling wavefunctions thereof .
this apparent bias may be justified on the grounds that the number of nodal domains of a tiling state can be obtained by simply decomposing it into its constituent tiles and summing over the corresponding number in each individual non - tiling state .
for instance , the phenomenon of tiling is most easily observable when @xmath89 under such circumstances , @xmath90 , since the nodal pattern of the wavefunction @xmath26 is composed of exactly @xmath91 tiles .
nevertheless , the validity of theorem [ one ] is retained under tiling circumstances . to illustrate this point , we consider , as previously , the situation when @xmath89 , which implies that the state @xmath92 is irreducible into further tiles .
then , by the premise of theorem [ one ] , @xmath93 such that either @xmath32 or @xmath94 is satisfied @xmath95 .
it is easy to see that theorem [ one ] is satisfied , albeit with a different function @xmath96 for the tiling states , since at least one of the following holds : @xmath97 where the equality follows trivially from the fact that @xmath98 is divisible by @xmath99 .
[ h ] the right - angled isosceles triangle having the length of each equal sides as @xmath100 may be chosen as @xmath101 ^ 2 : y \le x \big\}$ ] .
the corresponding eigenfunctions are @xmath102 with the spectrum being determined by @xmath103 . since the wavefunction is a sum of two terms , the corresponding equivalence classes are defined by @xmath104 $ ] , where @xmath105 .
the total number of nodal domains @xmath36 can be expressed in terms of the number of nodal loops ( nodal curves which neither touch the boundary nor intersect themselves or any other nodal line ) , @xmath106 , and the total number of intersections of the nodal set with the boundary @xmath107 , as @xmath108 , where @xmath109 .
the nodal domains form a tiling structure when @xmath110 an empirical recursive formula for @xmath36 has been proposed by @xcite .
however , observations of evaluated counts of nodal domains indicate the existence of the simpler recurrence relations @xmath111 let @xmath112@xmath113 and @xmath114 . the functional solutions to eq
are given by @xmath115 , where @xmath116 is a parameter that depends on @xmath117 and @xmath17 .
comparison of this solution with the data predicted by the recursive formula of @xcite renders the exact expression for the number of domains , when @xmath117 is even , as @xmath118 , \label{isosceles_nu}\ ] ] with the @xmath119 sign being applicable when @xmath120 and the @xmath121 sign otherwise .
when @xmath117 is odd , this equation is to be modified to @xmath122 , \label{isosceles_odd}\ ] ] where @xmath123 is a term responsible for small fluctuations .
when @xmath124 and for @xmath125 exactly reduces to @xmath126 .
although @xmath123 increases with @xmath17 , it remains constant within @xmath104 $ ] for a particular value of @xmath17 and hence @xmath127 .
the explicit functional form of @xmath123 , however , remains undefined .
c c c | c c c c @xmath16 & @xmath17 & @xmath128 & @xmath19 & @xmath129 & @xmath130 & @xmath13138 & 13 & 12 & 103 & & 78 & + 64 & 13 & 12 & 194 & 91 & 156 & 78 + 90 & 13 & 12 & 285 & 91 & 234 & 78 + 116 & 13 & 12 & 376 & 91 & 312 & 78 + 142 & 13 & 12 & 467 & 91 & 390 & 78
+ -10 cm let @xmath6 be the equilateral triangle of area @xmath132 represented as @xmath133 \times \bigg[0,\frac{\sqrt{3}\pi}{2}\bigg ] : y\le\sqrt{3}x\bigg\ } \cup \bigg\{(x , y ) \in \bigg[\frac{\pi}{2},\pi \bigg ] \times \bigg[0,\frac{\sqrt{3}\pi}{2}\bigg ] : y\le\sqrt{3}(\pi - x)\bigg\}. \end{aligned}\ ] ] the eigenfunctions of the laplace
beltrami operator for the general equilateral triangle billiard , with @xmath134 being the length of each side , form a complete orthogonal basis and are given by @xcite as @xmath135 \sin{\bigg(n\frac{2\pi}{\sqrt{3}l}y\bigg ) } - ( \cos,\sin)\bigg[(2n - m)\frac{2\pi}{3l}x\bigg]\sin{\bigg(m\frac{2\pi}{\sqrt{3}l}y\bigg ) } \nonumber \\&+ & ( \cos,\sin)\bigg[-(m+n)\frac{2\pi}{3l}x\bigg]\sin{\bigg[(m - n)\frac{2\pi}{\sqrt{3}l}y\bigg ] } , \end{aligned}\ ] ] where @xmath16 and @xmath17 are restricted to values such that @xmath136 ( @xmath137 ) and the spectrum of eigenvalues is defined as @xmath138 .
tiling of the domains is observed when @xmath139 .
[ h ] each non - tiling wavefunction @xmath20 of the symmetric mode , for a fixed value of @xmath17 , may be categorised into one of at most @xmath140 equivalence classes depending on the value of the residue @xmath141 .
the sequence of nodal domain counts for each class , when analysed with regard to the second difference of @xmath19 , shows the existence of the recurrence relations @xmath142 this equation can be analytically solved to yield @xmath143 where @xmath116 and @xmath144 are two parameters dependent on @xmath145 and @xmath17 .
comparison of with the tables of evaluated domain counts clearly indicates @xmath146 .
this , when coupled with further observations about the nature of the parameter @xmath144 , helps to effectively reduce into two cases : @xmath147 although the general forms of @xmath148 and @xmath149 , which contribute to small variations in the nodal domain count , are unknown , it has been observed that @xmath150the functional relations satisfied by @xmath148 and @xmath149 are further expounded in detail in @xcite .
c c c | c c c c c @xmath16 & @xmath17 & @xmath151 & @xmath19 & @xmath152 & @xmath130 & @xmath153 & @xmath15424 & 7 & 3 & 44 & & 21 & & + 45 & 7 & 3 & 198 & 154 & 154 & 133 & + 66 & 7 & 3 & 499 & 301 & 434 & 280 & 147 + 87 & 7 & 3 & 947 & 448 & 861 & 427 & 147 + 108 & 7 & 3 & 1542 & 595 & 1435 & 574 & 147
+ the ( 30 , 60 , 90 ) scalene triangle may be regarded as a special extension of the equilateral billiard as its modes are a subset of those of the latter and precisely correspond to the states of the equilateral triangle which are antisymmetric about the altitude .
these modes can be obtained by choosing the sine functions for the first components of each term in the wavefunction .
however , the appropriate condition for tiling in this case is noted to be @xmath156 or @xmath157 .
[ h ] observations stemming from the numerical counting of the nodal domains suggest that the appropriate difference equation obeyed by the ( 30 , 60 , 90 ) triangular billiard is @xmath158 c c c | c c || c c c | c c @xmath16 & @xmath17 & @xmath145 & @xmath19 & @xmath159 & @xmath16 & @xmath17 & @xmath145 & @xmath19 & @xmath159 5 & 2 & 5 & 2 & & 7 & 3 & 7 & 3 & + 11 & 2 & 5 & 7 & 5 & 16 & 3 & 7 & 13 & 10 + 17 & 2 & 5 & 12 & 5 & 25 & 3 & 7 & 23 & 10 + 23 & 2 & 5 & 17 & 5 & 34 & 3 & 7 & 33 & 10 + 29 & 2 & 5 & 22 & 5 & 43 & 3 & 7 & 43 & 10 + the constancy of the first difference implies that the number of domains scales linearly as the quantum number @xmath16 , which presents a marked departure from the equilateral triangle for which the corresponding dependence is quadratic .
the fine interconnections of nodal lines between adjacent segments of domains not only complicate the manual counting of excited states but also render the billiard unsuitable for the application of the hoshen - kopelman algorithm @xcite and therefore , it is difficult to surmise an exact formula for the first difference .
however , our extensive analysis of a considerable number of lower states enables us to estimate that @xmath160 for the non - tiling situations .
plane polygonal billiards with internal angles @xmath161 are integrable as the resulting shapes tessellate the plane and all the polygonal domains possessing a complete set of trigonometric eigenfunctions of the laplacian , under either dirichlet or neumann boundary conditions , are documented by @xcite .
the only triangular solutions are the right isosceles , equilateral , and the ( 30 - 60 - 90 ) triangles .
in addition , the rectangular billiard is integrable
. non - polygonal , convex shapes are integrable if the helmholtz equation can be separated in an appropriate coordinate system .
for instance , the circular and elliptical billiards are integrable . in this article , for _ all _ the above - mentioned systems , we have shown that the number of domains , @xmath5 of an eigenfunction satisfies a difference equation .
since the arguments given for the circle and the ellipse are quite general , it is expected that @xmath162 for other separable planar billiards satisfy similar equations .
as classifying patterns in non - separable shapes and counting domains has been a very difficult problem , the theorem presented here marks a considerable advance . in short ,
_ the geometrical patterns have been algebraically represented_. perhaps the next study will be on the simplest non - integrable system the @xmath163-rhombus billiard @xcite .
the eigenfunctions of this system can not be represented as random waves @xcite .
the results presented here and recently found for the equilateral triangle billiard @xcite invite a similar study for the rhombus billiards . however , this non - integrable step " seems quite challenging .
we hope that the present result which is , in fact , a two - dimensional generalization of courant s work for a one - dimensional system , would suggest the nontrivial step without referring to the random - wave hypothesis .
we believe that the connection with a certain class of difference equations might play a significant role in realising the next step .
consider the eigenfunction of the right - angled isosceles triangular billiard given by .
we examine the origin of the similarity in the nodal domain patterns of @xmath164 and @xmath165 that was illustrated in figure [ examples ] .
we believe that a geometric argument for the similarity between these two states would naturally explain the similarity between all states belonging to the same equivalence class @xmath104 $ ] thereafter .
@xmath166 \sin ( ny ) - \sin(nx ) \sin [ ( m+2n)y]\\ \nonumber&=&\kappa \sin(mx)\sin(ny ) - \chi\sin(nx)\sin(my ) + \cos(mx)\sin(2nx)\sin(ny ) - \cos(my)\sin(nx)\sin(2ny),\end{aligned}\ ] ] where , @xmath167 , which implies that @xmath168 , @xmath169 using the notation @xmath170 which groups together the first two terms in the expansion of , we see that @xmath171 represents a linear combination of the two terms of @xmath26
. however , the weights assigned to each term of the sum is @xmath172 as in the eigenfunction , but rather @xmath173 and @xmath174 respectively .
the distortion of the nodal lines of the wavefunction @xmath175 from those of the original wavefunction ( @xmath176 ) with the variation of the parameters @xmath177 and @xmath178 has been discussed and diagrammatically presented in @xcite .
this perturbed equation is analogous to @xmath171 , except for the fact that the weights in the latter , albeit bounded , are functions of @xmath179 and @xmath180 and are therefore , position - dependent .
nevertheless , it is observed that the resultant nodal domains are not greatly distorted from the original domains of the triangular billiard and the initial and perturbed nodal lines are also comparable . this may be accounted for by the reasoning that when @xmath181 , @xmath182 and hence @xmath171 differs from the original wavefunction of the system by only a multiplicative constant , therefore leaving the structure and number of nodal domains unaffected .
the alteration of the shapes of the nodal domains of the wavefunction on the billiard with the values of @xmath173 and @xmath174 are documented in the mathematica notebook accompanying this article . from this heuristic framework
, it is evident that @xmath171 preserves the basic structure of the domains of @xmath20 and is therefore the term that is responsible for preserving the similarity of the nodal pattern as @xmath183 .
it is insightful to examine at this point the geometric nature of the third and fourth terms in the trigonometric expansion of .
we employ the notation @xmath184\big[\sin ( 2nx ) \sin(ny)\big ] - \big[\cos ( my)\big]\big[\sin ( nx ) \sin(2ny)\big]\\ \label{app4}&= & 2 \sin ( nx ) \sin ( ny ) \big[\cos ( nx ) \cos ( mx ) - \cos(ny)\cos(my ) \big].\end{aligned}\ ] ] it is apparent that resembles in the sense that it is also a linear combination of the individual terms of @xmath185 , with the respective weights being @xmath186 and @xmath187 .
however , since one quantum number is an integer multiple of the other , @xmath185 , and consequently @xmath188 corresponds to a tiling state .
the zeros of , corresponding to the nodal sets of each of the three terms in the product , occur along the lines @xmath189 , where @xmath190 , thereby forming a grid - like nodal structure .
neither @xmath171 nor @xmath188 is an eigenfunction of the triangle and hence do not necessarily satisfy the dirichlet boundary conditions .
[ h ] + as can be observed from figure [ convince ] , additional nodal domains arise in @xmath191 within regions that constituted a single , continuous domain for @xmath26 .
this is because the set @xmath192 ^ 2 : y \le x\ ; \lvert\ ; { \mathop{\mathrm{sgn}}}\big(\tau_{m , n}(x , y)\big ) = -{\mathop{\mathrm{sgn}}}\big(\sigma_{m , n } ( x , y)\big ) \big\ } \ne \emptyset,\ ] ] and has non - zero measure .
therefore , the functions @xmath193 and @xmath194 tend to cancel each other locally within the connected regions of the triangle wherein they differ in sign , leading to loss of convexity of the domain , if any .
the continuity of the nodal lines for the state thus formed , @xmath195 , is ensured by bers theorem @xcite , which states that in two dimensions , if @xmath196 , then in any neighbourhood of @xmath197 , the nodal line is either a smooth curve or a union of @xmath17 smooth curves intersecting at equal angles .
in fact , an implicit assumption of this theorem underlies our ansatz of the structure - preserving property of @xmath193 as it enables one to conjecture that a collection of smooth curves would remain reasonably smooth under small weighted perturbations , thereby resembling the initial set of nodal lines .
the region wherein the condition for the formation of a new connected domain is satisfied is : @xmath198 \big(\lvert \tau_{m , n } ( x , y ) \rvert - \lvert \sigma_{m , n } ( x , y ) \rvert \big ) > 0\big\}. \label{condition}\ ] ] let @xmath199 , such that @xmath200 . a local statement analogous to courant
s nodal domain theorem @xcite shows that @xmath201 where @xmath202 depends only on the riemannian metric and @xmath203 is a constant for a two - dimensional surface .
this decay shows the scaling of the regions of the triangular billiard within which new nodal domains may form as one moves to higher excited states corresponding to higher eigenvalues @xmath204 .
the size of a nodal domain can be characterised by its in - radius @xmath205 and it has been proved @xcite that @xmath206 where @xmath207 and @xmath208 are constants ( not to be confused with the equivalence classes ) . for two dimensions ,
it is known that @xcite @xmath209 the number of nodal domains may be crudely approximated as @xmath210 from , we have @xmath211 thus predicts that @xmath212 and scales as @xmath213 , which is in exact agreement with .
we believe that exactly the same argument involving trigonometric expansion of the wavefunction may be applied to the equilateral triangle billiard to explain the geometric similarity of the nodal patterns in @xmath214 $ ] .
the authors would like to thank h. r. krishnamurthy for an interesting discussion regarding the geometric similarity between wavefunctions belonging to the same equivalence class .
furthermore , the authors are grateful to howard wilson for kindly providing the matlab code and for his assistance when they were in the process of visualising the wavefunctions on the ellipse . | eigenfunctions of integrable planar billiards are studied in particular , the number of nodal domains , @xmath0 of the eigenfunctions with dirichlet boundary conditions are considered .
the billiards for which the time - independent schrdinger equation ( helmholtz equation ) is separable admit trivial expressions for the number of domains . here
, we discover that for all separable and non - separable integrable billiards , @xmath1 satisfies certain difference equations .
this has been possible because the eigenfunctions can be classified in families labelled by the same value of @xmath2 , given a particular @xmath3 , for a set of quantum numbers , @xmath4 .
further , we observe that the patterns in a family are similar and the algebraic representation of the geometrical nodal patterns is found . instances of this representation are explained in detail to understand the beauty of the patterns .
this paper therefore presents a mathematical connection between integrable systems and difference equations .
integrable billiards , nodal domains , quantum chaos 02.30.lk , 05.45.mt 81q50 , 37d50 |
ngc 1399 is the central dominant galaxy of the fornax cluster . due to its proximity
( 19 mpc for @xmath3 km s@xmath1 mpc@xmath1 ) , this very regular , almost spherical ( e0 ; @xcite ) cd galaxy has been extensively studied in a wide range of wavelengths , ranging form radio to x - rays .
the optical radial profile , first studied by @xcite and later in more detail by @xcite , reveals a large halo extending up to 250 kpc .
the galaxy is surrounded by a large number of globular clusters , 10 times in excess with respect to those of its nearer companion ngc 1404 and of the fornax galaxy ngc 1380 @xcite .
dynamical studies of the stellar population of ngc 1399 indicate that the central regions of the galaxy are characterized by a high velocity dispersion , decreasing with radius , and slow rotation , as expected in giant ellipticals @xcite . at larger radii
the velocity dispersion starts increasing again , as shown by studies of the planetary nebulae @xcite and globular clusters dynamical properties @xcite .
this suggests a different dynamical structure , than that of the inner stellar body , for the galaxy envelope and the globular cluster population .
ngc 1399 hosts a weak nuclear radio source @xcite with radio luminosity of @xmath4 ergs s@xmath1 between @xmath5 and @xmath6 hz .
the radio source has two jets ending in diffuse lobes , confined in projection within the optical galaxy .
@xcite suggest that the radio source is intrinsically small and confined by the thermal pressure of the hot ism .
x - ray data have shown the presence of an extended hot gaseous halo surrounding ngc 1399 .
observations made with the _ einstein _ ipc by @xcite constrained the gas temperature to be [email protected] kev .
the _ einstein _
ipc data were used also by @xcite to derive the mass distribution of the galaxy .
they found that both models with and without dark matter were compatible with those data , depending on the assumed temperature profile .
@xcite suggested the presence of a cooling flow in the center of the galaxy , depositing 0.8@xmath8 m@xmath9 y@xmath1 .
ginga observations @xcite led to the detection of extended emission out to a galactocentric radius of @xmath10250 kpc .
@xcite constrained the x - ray temperature to be 1.0@[email protected] kev with the broad band x - ray telescope .
* hereafter rffj ) , using the rosat pspc , have studied in detail the temperature profile of the hot inter - stellar medium ( ism ) out to 220 kpc finding an isothermal profile ( 0.9@[email protected] ) from 7 to 220 kpc , and a central cooling flow ( 0.6@[email protected] ) of at least 2 m@xmath9 y@xmath1 .
@xcite were able to identify with asca the presence of different components in the x - ray halo , associated respectively with the galaxy and the fornax cluster potential .
@xcite used the better resolution of the rosat pspc for studying in detail the cluster x - ray structure and finding a total binding mass of ( 4.3 - 8.1)@xmath12 m@xmath9 within 100 kpc , with a mass - to - light ratio increasing from 33 m@xmath9/l@xmath9 at 14 kpc to 70 m@xmath9/l@xmath9 at 85 kpc .
they also found a metal abundance of 0.6 solar .
@xcite analized asca data finding that either a two temperature spectral model or a cooling - flow model with fe solar abundances , are required to fit the thermal x - ray emission , and additional absorption is needed in the galaxy center .
@xcite found similar near solar metallicities but argued that the metallicity inferred by asca spectral fits is dependent on the assumed atomic physics model and is not solved by muti - components spectral models .
more recently @xcite used rosat data to pose an upper limit on the nuclear source brightness .
chandra _ data were used by @xcite to further constrain the nuclear source and by @xcite to study the point source population hosted by globular clusters . in this paper
we present the results of a deep image of the ngc 1399ngc 1404 field obtained from data collected between 1993 and 1996 with the rosat high resolution imager ( hri , for a description see @xcite ) .
we take advantage of the @xmath13 resolution of the hri to study in detail the structure of the galactic halo and relate the results to those obtained at larger scales with poorer resolution instruments .
we study the interactions between the nuclear radio source and the galactic halo and discuss the properties of the discrete sources population . a preliminary analysis of _ chandra _ data supports the rosat results .
+ throughout this paper we adopt @xmath3 km s@xmath1 mpc@xmath1 and a distance of 19 mpc ( 1=5.5 kpc ) .
the ngc 1399 field , including ngc 1404 , was observed at three separate times with the rosat hri : in february 1993 , between january and february 1996 and between july and august of the same year .
the total exposure time is 167.6 ks ( table [ observations ] ) .
the data were processed with the sass7_8 and sass7_9 versions of the rosat standard analysis software ( sass ) . for our data analysis
, we used the iraf / xray and ciao packages developed at the smithsonian astrophysical observatory and at the _ chandra _ x - ray center ( cxc ) , and other specific software as mentioned in the text .
ccccrcc + name & & sequence i d . &
exp.time & obs . date & p.i . + & r.a . &
dec & & ( sec ) & & + ngc 1399 & 03@xmath1438@xmath1531@xmath16 & 35@xmath172700 & rh600256n00 & 7265 & 1993 feb 17 & d .- w .
kim + & & & rh600831n00 & 72720 & 1996 jan 04 - 1996 feb 23 & g. fabbiano + & & & rh600831a01 & 87582 & 1996 jul 07 - 1996 aug 26 & g. fabbiano + the hri data processed with the sass versions prior to sass7_b of march 1999 ( as it happens for our data ) suffer from an error in the aspect time behavior @xcite .
this translates into an error on the position of the incoming photons and thus it results in a degradation of the point response function ( prf ) . because high resolution analysis is the primary objective of this work we run the correction routine asptime ( f. primini 2000 , private communication ) on the data , to improve the aspect solution .
visual inspection of the brightest pointlike sources in our field ( figure [ asp_corr ] ) demonstrates the improvement in the image quality .
additional problems that may preclude the attainment of the hri potential resolution are the imperfect correction of the spacecraft wobble and the wrong tracking of reference stars due to the variable pixel sensitivity across the detector @xcite . we followed the procedure suggested by harris and collaborators , of dividing the observations in time bins ( obi ) and realigning these segments by using the centroids of a reference pointlike source in each obi .
our data can be divided in 4 obis in rh600256n00 , 43 in rh600831n00 and 33 in rh600831a01 .
we checked each obi individually using the brightest point source in the field ( no.8 in table [ source_tab ] ) .
we found that in the first observation ( rh600256n00 ) the obi are well aligned and no correction is needed .
the second observation ( rh600831n00 ) has 7 obis out of 43 that are 7``-10 '' displaced ; however , the overall prf is 6 `` fwhm so that just a slight improvement can be obtained dewobbling the image .
the third observation ( rh600831a01 ) has a 8 '' prf but the signal to noise ratio of the individual obis is worse than in the previous observations so that the dewobbling procedure can not be applied .
given these results , we decided not to apply the dewobbling correction , and to take as good a resulting prf , in the composite image , of @xmath187 " fwhm .
the aspect - corrected observations were co - added to obtain a `` composite observation '' . to make sure that pointing uncertainties did not degrade the image we used the centroids of three bright pointlike sources in the field ( no.8 , 19 an 24 in table [ source_tab ] ) to align the images .
the applied corrections were all within 2 arcsec .
the resulting composite image is shown in figure [ ngc1399comp ] .
this image was then corrected for vignetting and variations of exposure time and quantum efficiency across the detector by producing an `` exposure map '' for each observation with the software developed by s.l .
snowden ( * ? ? ?
* hereafter smb ) . to study the large - scale brightness distribution of the ngc 1399 field we rebinned the exposure corrected data in 5@xmath195 pixels .
we then adaptively smoothed the image with the cxc ciao _ csmooth _ algorithm which convolves the data with a gaussian of variable width ( depending on the local signal to noise range of the image ) so to enhance both small and large scale structures .
the resulting image ( figure [ csmooth]a ) shows a complex x - ray morphology .
the center of the image is occupied by the extended halo of ngc 1399 .
the galaxy possess a central emission peak and an external extended and asymmetric halo .
this halo is not azymuthally symmetric with respect to the x - ray peak , but it extends more on the sw side .
the x - ray surface brightness distribution of the halo appears filamentary , with elongated structures and voids .
as it can be seen from a comparison with the optical digitized sky survey ( dss ) image ( figure [ csmooth]b ) , while the x - ray emission peak is centered on the optical galaxy , the x - ray halo extends radially much further than the optical distribution .
moreover most of the features seen in the x - ray image have no direct optical counterpart .
the x - ray emission of ngc 1404 instead is almost symmetric and largely consistent with the optical distribution , with the exception of the se tail reported by @xcite in the pspc data , that may be the signature of ram pressure stripping of the ngc 1404 corona infalling toward the denser ngc 1399 cluster halo .
a number of pointlike source in the field is also evident in figure [ csmooth]a .
the analysis of these sources is discussed in @xmath20 [ sources ] .
we examined the inner halo of ngc 1399 in greater detail using a 1" / pixel resolution .
the adaptively smoothed image ( figure [ centbox ] ) reveals an elongation of the inner halo structure in the n - s direction plus a large arc protruding to the west side of the halo .
several voids are present in the x - ray distribution , the largest being 1.5 arcmin nw of the emission peak .
the central peak appears circular with at most a slight n - s elongation .
these features are all above @xmath21 significance , since @xmath21 was the minimum significance level requested for an intensity fluctuations to be smoothed on a given scale . some features , as the sw clump centered at ra , dec=@xmath22,35@xmath172900 or the one at @xmath23,35@xmath172630
, are also detected with a wavelets detection algorithm ( see @xmath20 [ sources ] ) .
we can rule out that features on scales of 1 arcmin are of statistical nature , because their significance can be as high as @xmath24 ( see @xmath20 [ model ] ) .
these brightness fluctuations must be due to the local physical conditions of the hot gaseous halo . as a first step in the quantitative study of the x - ray emission
, we created a radial profile from the hri data , assuming circular symmetry .
count rates from the exposure - corrected composite image , were extracted in circular annuli centered on the x - ray centroid ra , dec=@xmath25 , [email protected] .
we took care to remove the contribution of all the detected point - like sources by excluding circles within the @xmath21 radius measured by the wavelets algorithm ( @xmath20 [ sources ] ) , from the source centroid .
we also excluded a 150 " circle centered on ngc 1404 ( see @xmath20 [ ngc 1404 ] ) .
the x - ray brightness profile is shown as a continuous line in figure [ profile ] .
the emission extends out to @xmath10 500 ( 46 kpc at the assumed distance of 19 mpc ) .
the radial profile flattens out past 500 " ( dash - dotted line in figure [ profile ] ) suggesting that we have reached the field background level and that any residual halo emission is below our sensitivity limit .
nevertheless we know from previous investigations of the fornax cluster @xcite that extended x - ray emission , bright enough to be detected in our data , is present at galactocentric radii @xmath26 .
we thus expected to see a smooth gradient in the x - ray profile continuing past 500 .
we checked to see if a very high background was present in some obis or either in some of the hri spectral channels but we found none that could explain our lack of sensitivity .
a possible explanation is that we did nt reach the background level around 500 and the actual background is much lower but , due to uncertainties in the exposure correction near the edge of hri field of view , we are not able to see the expected decline in the x - ray emission .
we used the smb software to calculate a background map for this field . the radial profile derived from this background map
is shown as a small dashed line in figure [ profile ] .
the smb model however may underestimate the `` true '' hri background somewhat because it models only the charged particle contribution .
we therefore corrected this value using the 52 ks rosat pspc observation rp600043n00 centered on ngc 1399 , taken on the 1991 august 15 ( figure [ pspc ] ) .
even if the pspc observation is much shorter than our hri image , the higher sensitivity and lower background of this instrument allows a better study of the large scale x - ray emission .
the exposure corrected pspc profile is shown in figure [ pspc_prof ] .
the background level ( dashed line ) was estimated from the 2400-3000 annulus .
we compared our radial profile to the one derived by @xcite and rffj , who made use of the same data , finding consistent results .
we then rescaled the smb background to find the best agreement between hri and pspc counts in the 50 - 400 arcsec region , after rebinning the hri data to match the wider pspc prf .
the `` true '' hri background level found in this way ( long - dashed line in figure [ profile ] ) lies in between the flattening level of the hri profile and the smb background and was used to derive the hri background - subtracted profile shown as empty circles .
this way of finding the hri background assumes that the pspc brightness profile is not affected by systematic errors on the background determination and can thus be used to rescale the hri counts .
however , even if the pspc background suffered from residual errors ( e.g. due to uncertainties in the exposure correction at large radii ) , they would not significantly affect the central region profile used to rescale the smb hri background because there the x - ray brightness is one order of magnitude larger than the pspc background level .
the hri radial surface brightness profile has a complex structure . in the ` central ' region ,
i.e. within 50 from the emission peak , the x - ray emission is well represented by a simple beta model : @xmath27^{-3\beta+0.5}\ ] ] with best fit parameters @xmath28 arcsec , @xmath29 and @xmath30 for 22 degrees of freedom ( figure [ fit ] ) . to determine the best fit parameters the model was convolved with the hri on - axis prf @xcite . at radii larger than 1 arcmin
the surface brightness profile shows a significant excess over this model ( see figure [ profile ] ) , indicating the presence of an additional ` galactic ' component .
restricting the fit to r @xmath11 40 to check if this more extended component contributes around 50 , gives consistent results with @xmath31 arcsec and @xmath32 ( @xmath33 for 17 degrees of freedom ) .
thus , for @xmath34 the x - ray emission falls as @xmath35 . as can be seen by the comparison with the hri prf ( shaded region in figure [ profile ] ) , whose uncertainty is due to residual errors in the aspect solution @xcite , the central component is extended and can not be due to the presence of a nuclear point source , that in fact is not detected in our data ( see @xmath20 [ xradio ] ) .
the pspc profile in figure [ pspc_prof ] shows that x - ray emission extends to much larger radii than visible in the hri data and suggests the presence of a third component dominant at @xmath36 .
large azimuthal differences are seen in the x - ray profile at radii larger @xmath37 arcmin .
we derived composite hri - pspc x - ray profiles in two sectors dividing the images along the line connecting the x - ray centroids of ngc 1399 and ngc 1404 : a ne sector ( p.a .
[email protected]@xmath38 ) and a sw sector ( p.a .
[email protected]@xmath38 ) . while the emission within 1 arcmin is azimuthally symmetric , the galactic component ( @xmath39 )
is much more extended in the sw direction and presents a steep decline in the ne sector , as suggested by figures [ csmooth]a and [ centbox ] .
the even more extended component visible in the pspc data follows the opposite behavior , being 3 times brighter around 1000 on the ne side than on the sw one .
while these asymmetries have been reported by @xcite , their pspc data alone do not have enough resolution to clearly separate the central component from the external halo .
jones and collaborators used a single power - law model to fit the whole distribution , thus obtaining a value of @xmath100.35 for the power law slope .
we derived the brightness profile of ngc 1404 extracting count rates in 2 annuli centered on the galaxy centroid .
the background determination for this galaxy is even more difficult than for ngc 1399 , due to the fact that it is embedded in the halo of the dominant galaxy .
thus we had no other choice than to estimate the background level from the region between 200 and 300 from the galaxy ( see figure [ 1404profile]a ) , obtaining @xmath40 counts arcmin@xmath41 sec@xmath1 .
the measured count rates are shown in figure [ 1404profile]a as a continuous line while the background subtracted profile is represented by open circles .
the ngc 1404 emission extends out to 110
( @xmath1010 kpc ) from the galaxy center and shows no sign of a multi component structure as the one detected in ngc 1399 .
the surface brightness is circularly symmetric ( cf .
figure [ csmooth]a ) and declines with a power law profile out to 90 . past this radius
the effect of ram pressure stripping becomes evident as a sharp cutoff in the radial profile . in figure
[ 1404profile]b we can see that in the north - west sector ( thin dashed line ) the galactic halo suddenly disappears in the x - ray background for @xmath42 ( @xmath43 kpc ) while in the south - east one ( continuous line ) the emission from the tail , seen in the adaptively smoothed images ( figure [ csmooth ] ) , extends up to @xmath44 ( 14 kpc ) .
we fitted the observed brightness distribution , within a 90 radius , with a beta model . in this case
the on - axis prf we used for ngc 1399 is no longer valid because ngc 1404 is located 10 arcmin south - west of ngc 1399 .
although an analytical correction of the on - axis prf is possible @xcite , it does nt take into account the prf azimuthal asymmetry .
we thus convolved the beta model with the on - axis prf and considered the resulting core radius as an upper limit to the correct value , obtaining @xmath45 arcsec and @xmath46 with @xmath47 for 42 d.o.f .
the best fit model is shown in figure [ 1404fit ] . because of the complex halo structure it is not possible to obtain a satisfactory fit of the entire surface brightness distribution of ngc 1399 with a single circularly symmetric model .
the _ sherpa _ modeling and fitting application of the ciao package allows us to adopt a more realistic approach by performing a two dimensional fit of the x - ray images , using a complex model obtained by adding together three bidimensional beta profiles of the form : @xmath48 where @xmath49 and @xmath50 @xmath51 expression ( [ 2dmodel ] ) represents the bidimensional extension of the beta model in equation ( [ beta ] ) where we also allow the model ellipticity ( @xmath52 ) and position angle ( @xmath53 ) to vary .
lcccccccc + component & & @xmath54 & @xmath55 & @xmath52 & @xmath53 & @xmath56 & @xmath57 + & r.a . &
dec & ( arcsec ) & & @xmath58 & ( rad ) & & ( d.o.f . )
+ central & 03@xmath1438@[email protected] & 35@xmath172701 & 3.93 & [email protected] & 0.0 & 0.0 & 1.2 & 1612 + galactic & 03@xmath1438@[email protected] & 35@xmath172740 & 2125@xmath60 825 & 41@xmath60 25 & 0.018@xmath60 0.018 & 1.09@xmath61 & 1.2 & 1612 + cluster & 03@xmath1438@[email protected] & 35@xmath172305 & 9047@xmath60 495 & 24@xmath59 3 & 0.14@xmath60 0.02 & 0.0 @xmath60 0.1 & 1.5 & 4790 + to model simultaneously the small and large scale components we used the following procedure : * ( a ) * we rebinned the hri image in 15@xmath1915 pixels , in order to have enough counts per pixel to fit the galaxy halo and to smooth out small scale asymmetries ; ngc 1404 and all the detected point - like sources ( @xmath20 [ sources ] ) were masked out before performing the fit . *
( b ) * the pspc data were rebinned in 45@xmath1945 pixels ; the regions containing the support structure where masked , together with all the sources detected in the hri field . because the pspc field of view ( fov ) is larger than the hri one , to mask all the remaining pspc sources we used the source parameters measured by the wavelets algorithm in the context of the galpipe project ( * ? ? ? * and references therein ) , which has produced a list of all sources present in the rosat pspc archive . * ( c ) * we built a model composed of three bidimensional @xmath55 components represented by expression ( [ 2dmodel ] ) plus a constant background . * ( d ) * the first two components , representing the central and galactic halo emission , were fitted to the 15@xmath1915 hri data , fixing the third component parameters to some initial values .
because at this resolution it is not possible to properly determine all the parameters of the central component , we fixed the core radius to the best - fit value obtained from the radial fit shown in figure [ profile ] . while the central component shows a slight n - s elongation ( @xmath62 , @xmath20[brightness ] ) at this resolution our attempts to vary also ellipticity @xmath52 and position angle @xmath53 resulted in no improvements of the fit , so that they were fixed to zero . * ( e ) * the third component , corresponding to the cluster emission ,
was fitted to the 45@xmath1945 pspc data , fixing the central and galactic halo parameters to the values obtained in step ( d ) .
the background value was fixed to the flattening level measured in @xmath20[ngc 1399 ] . *
( f ) * steps ( d ) and ( e ) were repeated iteratively until the best fit parameters converged within the errors .
the best - fit model parameters are shown in table [ fit_tab ] . in figure [ model_map ]
we show the model contours superimposed on the 45/pixel pspc image ( main panel ) ; the nuclear region is shown in greater detail in the right - bottom panel , with contours overlaid on the 15/pixel hri image .
the three components have different features : while the central one is centered on the galaxy and is highly peaked , the more extended galactic halo is displaced to the south - west thus accounting of the observed asymmetries shown in figure [ pies ] . in contrast with the circular symmetry of the smaller components ( @xmath63 ) the third component is elongated in the east - west direction and its center is displaced to the north - east of the optical galaxy .
these results suggest a different origin of each component , that will be discussed in detail in @xmath20 [ discussion ] .
the @xmath64 values obtained from the fit are quite large .
this is not surprising considering the large spatial anisotropy of the x - ray surface brightness ( see @xmath20 [ brightness ] ) .
the situation is worse for the pspc data ( see figure [ pspc ] ) , also because of the poor knowledge of the exposure corrections past @xmath65 from the center of the fov ( the position of the support ring ) and the large and irregular prf at these radii that makes difficult a correct source masking . in table
[ fit_tab ] the ` central ' and ` galactic ' components have the same @xmath64 value because they were fitted together to the hri data .
if we restrict the fit to the region within 1 arcmin and use higher resolution ( 2/pixel ) we obtain a @xmath66 , in agreement with the results obtained in @xmath20 [ ngc 1399 ] from the profile fitting . the relative contribution of each component to the global profile is shown in figure [ model_prof ] .
we must notice that here both global and individual components profiles were extracted in annuli centered on the x - ray centroid ( rather than centering each one on the respective component centroid ) so to allow a comparison of the observed data with the model .
the bottom panel shows the radial residuals respect to the best fit model .
the high @xmath64 values are in agreement with the values obtained from the bidimensional fit ; in particular the residuals are systematically higher at radii larger than 1100 due to the problems explained above .
the residuals of the bidimensional fit to the hri image are shown in figure [ residuals ] .
the presence of structures in the gaseous halo is evident and confirms the results obtained with the adaptive smoothing technique .
the highest residuals are coincident with the positions of point sources detected in the field , as can be seen from the overlaid contours . considering the fluctuations on scales of 60 ( @xmath67 kpc ) the statistical errors are reduced of a factor 4 with respect to the grayscale levels shown in figure [ residuals ] , and the significance of the residual emission is increased of the same factor .
this means , for instance , that the deep ` hole ' on the north - west side of the halo is approximately significant at the @xmath24 level above the fluctuations expected from the multicomponent model , while the excess emission regions on the north and south sides of the galactic center are significant at the @xmath21 level .
even though the multi - component bidimensional model may not be a proper physical representation of the hot x - ray halo this result shows that the filamentary structures seen in the halo are not due to statistical fluctuation over a smooth surface brightness distribution . this result
is further confirmed by the _ chandra _
data presented in @xmath20 [ chandra ] .
we have derived the hot gas density radial profile of ngc 1399 from the x - ray surface brightness profile , making use of the procedure described by @xcite , that allows to deproject the observed emission in concentric shells , assuming spherical symmetry and homogeneity of the hot gas .
since the global gas distribution has a complex and asymmetric structure , we can not apply this method to the total emission ; instead we applied it to the ` central ' and ` galactic ' components separately , whose ellipticities , derived from the bidimensional models ( see previous section ) , are small .
the spherical symmetry approximation is not valid for the cluster halo so that , in deriving the density profile , we assumed rotational symmetry around the minor axis .
we found that assuming symmetry around the major axis the difference is of the order of a few percents ( except at very large radii where exposure correction uncertainties are dominant ) and resulted in minor changes to our conclusions .
the central , galactic and cluster components were assumed to be isothermal with temperatures of respectively kt=0.86 , 1.1 and 1.1 kev , in agreement with the pspc temperature profiles derived by rffj .
@xcite showed that asca data suggest the presence of multiphase gas within the central 5 of ngc 1399 , better fitted by cooling flows models .
however , as he notes , the narrower rosat energy range and the lower spectral resolution is unable to clearly discriminate between single and multi - temperature models ( even though it reveals the presence of radial temperature gradients ) so that our assumption of isothermal components is adequate for the kind of analysis performed here .
the electron density profiles of the isothermal components are shown in figure [ density_prof ] .
the density profiles are centered on the centers of their respective components and thus the total density , shown as a thick line , does not take into account the centers offset . including the displacements of each component with respect to the galaxy center results in a slightly smoother profile in the outer regions , due to the azimuthal averaging , shown as a thin continuous line .
rosat and asca spectral fits ( rffj ; @xcite ) revealed that additional absorption , over the galactic value , may be required in the galaxy center .
we thus estimated an upper limit for the density of the central component assuming an absorption of @xmath68 cm@xmath41 ( see rffj ) , shown as downward triangles in figure [ density_prof ] . we must notice that adding the single density components , derived from the surface brightness decomposition , to obtain the total density profile is not strictly the correct approach since the surface brightness does not depend linearly on the gas density .
however regions of the image in which more than one component are significant are small and the good agreement of our results with the estimate of rffj ( who deprojected the total surface brightness ) indicates that we are not introducing significant errors in our total density estimate .
the same considerations hold for the mass profiles derived below . using the density profiles and the assumed temperatures we estimated the cooling time @xmath70 as a function of radius , where @xmath71 is the electron density shown in figure [ density_prof ] , @xmath72 @xcite and @xmath73 are
the hydrogen and total density and @xmath74 is the cooling function given by @xcite .
figure [ cooling_prof ] shows that the central and galactic components have cooling times much shorter than the age of the universe , assumed to be @xmath75 years , out to 250 and 350 ( 23 and 32 kpc ) respectively .
the cumulative gas mass profile is shown in figure [ mass_prof]a as a function of radius .
the flat profile of the galactic component over 600 ( 55 kpc ) reflects the steep slope of the beta profile .
the filled triangles show that the adoption of a high absorbing column in the galaxy center does not affect significantly the gas mass estimates , giving a difference of @xmath7633% .
we estimated the total gravitating mass using the expression @xcite : @xmath77 we tried four different temperature profiles based on published data : we used both the linear and power - law approximations derived by @xcite based on pspc spectral fits , a constant temperature of 1.1 kev that approximates the isothermal profile found by rffj and a power - law profile @xmath78 , with @xmath79 kev and @xmath80 , representing the temperature drop in the inner 150 `` ( 14 kpc ) .
the gravitating mass , extracted in concentric annuli centered on the x - ray centroid ( i.e. taking into account the offset between the different components ) , is shown in figure [ mass_prof]b .
the similarity of the continuous lines indicates that the gravitating mass is weekly dependent on the assumed temperature profile and is mostly determined by the density profile .
the largest deviation is seen in the model taking into account the central temperature drop , which is systematically lower in the inner 100 '' .
our profiles are in good agreement with rffj and fall within the range predicted by optical observations ( shaded region , @xcite ) .
we clearly distinguish the transition between the galaxy and cluster halo near 700 " ( 64 kpc ) , in agreement with the asca data ( filled circles , @xcite ) .
both components exceed the luminous mass , based on an extrapolation of stellar measurements ( thin dashed line ) , indicating the presence of a consistent amount of dark matter .
thanks to the rosat hri resolution in the inner 100 " ( 9 kpc ) we are able to resolve the dynamical behavior of the central component .
the gravitating mass follows closely the stellar mass profile , confirming that the central galaxy dynamics are dominated by the luminous matter @xcite .
following the method outlined in the previous section we derived the electron density , cooling time and mass profiles for ngc 1404 . since the deprojection algorithm tends to enhance the fluctuations present in the brightness profile , we used the best fit model to derive the deprojected electron density .
this is possible because the ngc 1404 radial brightness profile was well fitted by a single symmetric beta model out to 90 `` ( @xmath20 [ ngc 1404 ] ) .
figure [ density_prof_1404]a shows the good agreement between the deprojected density derived from the beta model ( continuous line ) and the one from the surface brightness profile ( filled circles ) .
the discrepancy in the inner 5 '' is due to the fact that the beta model is deconvolved for the instrumental prf .
we stopped the deprojection at 80 " to avoid uncertainties related to the background subtraction and to the asymmetries due to ram pressure stripping ( @xmath20 [ ngc 1404 ] ) , both of which tend to increase the fluctuations at large radii . the cooling time profile derived from the beta model
is shown in figure [ density_prof_1404]b and suggests that significative cooling must be present in the ngc 1404 halo because the cooling time is smaller than @xmath81 yr within the inner 80 " ( @xmath107 kpc ) .
we adopted a cooling function @xmath82 ergs @xmath83 s@xmath1 for kt=0.6 kev @xcite .
the integrated gas mass and the total gravitating mass derived from equation [ mass_eq ] are shown in figure [ mass_prof_1404 ] .
we assumed ngc 1404 to have an isothermal profile with kt=0.6 kev @xcite . to determine the total hri x - ray flux of ngc 1399 we measured the total net counts within a 500 radius from the x - ray centroid .
the background counts were extracted from the rescaled particle map ( see @xmath20 [ ngc 1399 ] ) in the 500850 annulus .
we then corrected the background subtracted counts for the contamination of pointlike sources within the 500 circle .
this was done by subtracting , for each source , the net counts measured by the wavelets algorithm ( see @xmath20 [ sources ] ) .
the final count rate , for the hot halo of ngc 1399 , is [email protected] counts sec@xmath1 . to convert this count rate into a flux we used the best fit raymond - smith ( rs ) model found by rffj from the rosat pspc data ( a main thermal component with @xmath841 kev plus a softer emission with @xmath85 ev contributing for 16% of the emission in the 0.20.3 kev range ) . using the heasarc - xspec software
, we simulated a rs spectrum fixing the absorbing column to the galactic value ( @xmath86 cm@xmath41 ) and the metal abundance to the solar value . with this spectrum , we calculated the conversion factor between counts and fluxes using the pimms software , that takes into account the hri spectral response function .
we obtain 1 count sec@xmath1=3.158@xmath87 erg s@xmath1 cm@xmath41 so that @xmath88 erg s@xmath1 cm@xmath41 in the 0.1 - 2.4 kev energy range . assuming a distance of 19 mpc , @xmath89 erg s@xmath1 . taking into account the uncertainties in the exposure correction and background extraction
, this result agrees within 10% with the estimate of rffj , converted to our adopted distance . in the case of ngc 1404
we used a single temperature raymond - smith spectrum with @xmath90 kev and metal abundance 0.2 solar , following @xcite .
we measured a count rate of @xmath91 counts sec@xmath1 within a 150 radius , extracting the background from the 200-300 annulus .
the corresponding ( 0.1 - 2.4 kev ) flux and luminosity are @xmath92 erg s@xmath1 cm@xmath41 and @xmath93 erg s@xmath1 .
we also searched for evidence of nuclear variability of the two galaxies in our data , finding none .
ngc 1399 is known to be a radio galaxy with a faint radio core ( @xmath94 mjy at 5 ghz , @xcite ) .
@xcite studied in detail the 6 and 20 cm radio emission .
the 6 cm radio contours ( their figure 1 ) are superimposed on the ` csmoothed ' x - ray image in figure [ x - radio ] .
the radio core lies in the center of the galaxy , but the nuclear source is not distinguishable from the radio jets at this resolution .
we searched our x - ray data for a central point source that could be the counterpart of the radio core .
trying to model the nuclear emission with a beta model plus a delta function , convolved with the hri prf , resulted in minor variations in the beta model parameters and no improvement in the fit statistics , thus indicating no need for additional nuclear emission . to estimate an upper limit we varied the central point source brightness and repeated the fit until we reached a @xmath21 confidence level .
we obtain an upper count rate limit of @xmath95 counts s@xmath1 . assuming a power law spectrum with @xmath96 and galactic absorption , we get @xmath97 erg s@xmath1 cm@xmath41 in the 0.1 - 2.4 kev band , corresponding to @xmath98 erg s@xmath1 .
this value is approximately half of the one derived by @xcite making use of one of our hri observations ( rh600831a01 ) and 30% of the limit estimated by rffj by means of spectral analysis .
recently @xcite were able to further reduce the upper limit on the nuclear source by one order of magnitude making use of _ chandra _ data . by comparing the 6 cm radio contours with the x - ray halo ( figure [ x - radio ] )
we identify several interesting features : i ) in the x - rays the nuclear region is elongated in the n - s direction , following the radio jets .
this elongation is more pronounced in the southern direction , where there is an excess of x - ray emission following the western side of the jet ( region a ) ; ii ) at the point where the southern jet ends in the lobe there is a steep gradient in the x - ray emission .
the lobe is bent towards the west and aligned with a region of low x - ray emission ( region b ) ; iii ) the southern lobe ends in an irregular clump coincident with a region of higher x - ray emission ( region c ) ; iv ) no features as the ones seen in the southern lobe are evident in the northern one . here
the radio jet ends smoothly in a regular lobe with no sign of sharp transition and there is no evidence of cavities ( minima ) in the x - ray emission . instead
the jet and the lobe are aligned with an x - ray structure that seems to be coincident with the right side of the jet / lobe . at the north - west end of the northern lobe
there are two x - ray clumps ( region d ) : they are detected as a single source by the wavelet algorithm ( no.32 in table [ source_tab ] ) . to a more accurate visual inspection we found that the smaller and compact source is coincident with an optical ( probably background ) counterpart , while there is no counterpart for the diffuse clump ( @xmath20 [ sources ] ) suggesting that it is a local feature of the hot gaseous halo . from the residual image of the two - dimensional fit ( figure [ residuals ] )
we can calculate the significance of these features .
we find that the depletion corresponding to the southern lobe ( region b ) and the northern diffuse clump ( region d ) are significant respectively at the @xmath99 and @xmath100 level .
region c corresponds to a @xmath21 brightness enhancement ( @xmath20 [ model ] ) . to examine in greater detail the nuclear region we subtracted the halo model developed in @xmath20 [ model ] from the x - ray image to produce a residual map of the nuclear region with a 2" / pixel resolution ( figure [ x - radio_nuclear ] ) .
as discussed at the beginning of this section , there is no evidence of a nuclear point source aligned with the radio centroid .
instead residual emission is visible on both sides of the radio jets a few arcseconds ( @xmath101 kpc ) from the nucleus .
there is no evidence that any of these excesses may be due to the nuclear source because they are displaced with respect to the radio emission .
moreover the radio emission seems to avoid the x - ray excess on the eastern side of the nucleus , suggesting that these features may be correlated . in coincidence with the direction of the radio jets
there are two regions of negative residuals ( @xmath102 ) , extending up to @xmath103 , where additional excess emission , on both the northern and southern sides of the x - ray core and aligned with the radio jets , is clearly visible .
the southern excess corresponds to region a of figure [ x - radio ] .
residual emission is also present on the western side of the nuclear region .
visual inspection of the _ chandra _ image ( see @xmath20 [ chandra ] ) shows that this excess is probably due to the presence of point sources unresolved in the hri image . from the density profiles obtained in @xmath20 [ dens_par ] we were able to obtain the pressure profiles of the three halo components shown in figure [ press_prof ] .
the pressure profiles are compared to the minimum radio pressure measured by @xcite in the northern ( circles ) and southern ( triangles ) radio jets / lobes .
the figure confirms their result that the radio source can be confined by the ism whose pressure is higher than the radio one .
we must notice that they found a difference between radio and thermal pressure of more than one order of magnitude , while we find @xmath104 .
this is mostly due to the poor ism temperature constraint obtained from the _ einstein _ ipc data and to our better spatial resolution .
the vertical error bar in the figure indicates the pressure estimate for the diffuse clump observed in the d region . to calculate this value we assumed that the brightness enhancement is due to a homogeneous sphere of @xmath105 ( 800 pc ) radius at the same temperature of the galactic halo ( 1.1 kev ) . depending on
whether the background is extracted locally ( an annulus 9 " wide ) or from the beta model , the ` bubble ' pressure can vary up to 25% , so that the allowed range is shown as an error bar .
this estimate is dependent on the assumed geometry of the emitting region and on projection effects , so that it must be considered as an upper limit on the actual pressure .
we used the algorithm developed at the palermo observatory by f. damiani and collaborators @xcite to detect the discrete sources present in the hri field .
the algorithm measures the relevant parameters in a wavelet transformed space , using different wavelet scales to optimize the detection of both compact and extended sources .
the wavelet algorithm found 43 sources in our field , using a signal to noise threshold of 4.65 .
the threshold was chosen to assure a contamination of one / two spurious detections per field ( f. damiani private communication , see also @xcite ) .
three more sources were found with the algorithm , but being clearly associated to the ribbon - like features due to the presence of hot spots in the detector ( figure [ ngc1399comp ] ) , were excluded from the subsequent analysis .
after visual inspection of the adaptively smoothed image ( figure [ csmooth ] ) three more detections ( no.44 , 45 , 46 ) with a lower s / n ratio but showing clear features of pointlike sources , were added to the list .
the source positions in the field are shown in figure [ det ] .
sources no.42 and 43 are clearly extended and seem to be associated to structures in the ngc 1399 halo revealed by the adaptive smoothing algorithm ( @xmath20 [ brightness ] ) .
table [ source_tab ] lists the parameters measured with the wavelets algorithm for each source : position , 3@xmath106 radius , number of counts with statistical error , maximum signal to noise ratio ( for sources detected at different spatial scales ) , count rate and flux in the 0.5 - 2.0 kev band for both hri and pspc data .
the 0.5 - 2.0 kev energy band was chosen to allow an easily comparison with the @xcite background counts ( see below ) .
the the last three columns of table [ source_tab ] report the source short - term variability obtained with a kolmogorov - smirnov test of the hri light curves , long - term variability from comparison with pspc fluxes ( as explained below ) , and possible optical counterparts ( within the 3@xmath106 radius measured by the wavelets algorithm ) found by both visual inspection of the dss photographic plates and cross - checking with known sources present in the nasa extragalactic database ( ned ) .
the names of the counterparts correspond to those reported by @xcite or @xcite in their optical and nuv studies of the fornax field .
sources no.27 and 28 , representing respectively ngc 1399 and ngc 1404 , are not included in the table .
= 1.in cccrrrccccccc + source & r.a . & dec .
& 3@xmath106 radius & counts & max s / n & count rate & @xmath107 & @xmath108 & @xmath109 & time var . & opt.id . & ned obj . + no . & & ( arcsec ) & & & ( @xmath110 cnts sec@xmath1 ) & ( @xmath111 erg s@xmath1 cm@xmath41 ) & ( @xmath111 erg s@xmath1 cm@xmath41 ) & ( @xmath111 erg s@xmath1 cm@xmath41 ) & short - term / long - term & & + 1 & 3:38:20.0 & -35:24:42 & 8.7 & 35@xmath6012 & 4.7 & [email protected] & 4.3@xmath112 & 5.6 & 28@xmath113 & n / y & - & - + 2 & 3:38:32.0 & -35:26:45 & 8.7 & 54@xmath6017 & 6.1 & [email protected] & 6.7@xmath114 & 8.6 & - & n/@xmath115 & - & - + 3 & 3:38:25.3 & -35:22:42 & 8.7 & 49@xmath6014 & 6.7 & [email protected] & 6.1@xmath114 & 7.8 & - & n/@xmath115 & - & - + 4 & 3:38:45.9 & -35:22:50 & 8.7 & 42@xmath6012 & 6.1 & [email protected] & 5.3@xmath114 & 6.9 & 9@xmath116 & n / n & y & cgf 0202 + 5 & 3:38:31.8 & -35:26:02 & 11.9 & 165@xmath6023 & 13.0 & 10@xmath602 & 20@xmath117 & 26.0 & - & n/@xmath115 & - & cgf 0233 + 6 & 3:39:12.3 & -35:28:12 & 13.2 & 172@xmath6021 & 14.9 & 10@xmath601 & 21@xmath117 & 27.5 & 25@xmath118 & n / n & y & - + 7 & 3:38:31.8 & -35:33:42 & 12.2 & 56@xmath6016 & 6.0 & [email protected] & 6.9@xmath114 & 8.9 & 8@xmath116 & n / n & - & - + 8 & 3:38:41.2 & -35:31:35 & 10.8 & 1000@xmath6035 & 60.6 & 59@xmath602 & 123@xmath119 & 159.2 & 170@xmath120 & 99%/y & y & - + 9 & 3:38:49.8 & -35:24:57 & 12.2 & 47@xmath6014 & 5.0 & [email protected] & 5.8@xmath114 & 7.5 & - & n/@xmath115 & - & - + 10 & 3:39:08.1 & -35:31:22 & 12.2 & 56@xmath6016 & 6.2 & [email protected] & 6.9@xmath121 & 8.9 & 22@xmath119 & n / n & - & - + 11 & 3:38:32.4 & -35:27:05 & 13.7 & 171@xmath6030 & 11.5 & 10@xmath602 & 21@xmath117 & 27.2 & - & n/@xmath115 & - & - + 12 & 3:38:51.7 & -35:26:42 & 11.6 & 295@xmath6022 & 24.7 & 17@xmath601 & 36@xmath122 & 47.0 & 58@xmath119 & 95%/y & y & cgf 0102 + 13 & 3:38:48.6 & -35:28:35 & 12.2 & 77@xmath6020 & 7.7 & [email protected] & 9.3@xmath123 & 12.0 & - & n/@xmath115 & y & - + 14 & 3:39:09.7 & -35:27:07 & 12.2 & 46@xmath6014 & 5.0 & [email protected] & 5.7@xmath114 & 7.4 & - & n/@xmath115 & y & - + 15 & 3:38:01.9 & -35:26:27 & 17.3 & 62@xmath6019 & 4.7 & [email protected] & 7.5@xmath121 & 9.7 & - & n/@xmath115 & - & - + 16 & 3:38:29.4 & -35:25:05 & 17.3 & 75@xmath6023 & 4.8 & [email protected] & 9.1@xmath123 & 11.7 & - & n/@xmath115 & - & - + 17 & 3:38:31.6 & -35:30:58 & 12.2 & 70@xmath6018 & 6.8 & [email protected] & 8.6@xmath123 & 11.1 & - & n/@xmath115 & - & - + 18 & 3:38:03.7 & -35:33:27 & 17.3 & 65@xmath6019 & 5.4 & [email protected] & 8.5@xmath121 & 10.9 & - & n/@xmath115 & y & cgf 0354 + 19 & 3:38:08.5 & -35:34:20 & 15.0 & 908@xmath6034 & 52.3 & 57@xmath602 & 119@xmath119 & 152.7 & 127@xmath113 & 99%/n & y & - + 20 & 3:38:43.5 & -35:33:50 & 15.2 & 105@xmath6018 & 8.7 & [email protected] & 13.1@xmath124 & 16.8 & - & n/@xmath115 & y & - + 21 & 3:39:20.1 & -35:29:22 & 16.5 & 527@xmath6030 & 33.0 & 32@xmath602 & 68@xmath118 & 86.9 & 83@xmath113 & 99%/n & y & - + 22 & 3:38:24.1 & -35:14:52 & 17.3 & 60@xmath6018 & 4.9 & [email protected] & 8.0@xmath121 & 10.3 & 12@xmath118 & n / n & y & - + 23 & 3:38:12.2 & -35:16:17 & 17.3 & 59@xmath6018 & 4.8 & [email protected] & 7.7@xmath121 & 9.9 & - & n/@xmath115 & y & - + 24 & 03:37:58.5 & -35:19:42 & 16.7 & 379@xmath6027 & 24.9 & 24@xmath602 & 48@xmath122 & 61.6 & 75@xmath113 & n / y & y & - + 25 & 03:37:34.8 & -35:24:04 & 17.3 & 64@xmath6019 & 5.3 & [email protected] & 8.3@xmath121 & 10.6 & - & n/@xmath115 & - & - + 26 & 03:38:43.1 & -35:33:10 & 17.3 & 73@xmath6021 & 5.6 & [email protected] & 9.0@xmath123 & 11.6 & - & n/@xmath115 & y & - + 29 & 03:39:43.0 & -35:24:11 & 24.3 & 143@xmath6027 & 8.6 & 10@xmath602 & 21@xmath117 & 26.7 & 62@xmath120 & n / y & y & - + 30 & 03:38:31.0 & -35:14:17 & 26.4 & 586@xmath6038 & 26.5 & 38@xmath602 & 79@xmath118 & 102.1 & - & 99%/@xmath115 & y & fccb 1263 + 31 & 03:38:17.1 & -35:15:57 & 24.5 & 124@xmath6031 & 7.2 & 8@xmath602 & 16.4@xmath125 & 21.1 & - & n/@xmath115 & y & - + 32 & 03:38:25.7 & -35:25:27 & 24.5 & 145@xmath6037 & 6.2 & [email protected] & 17.7@xmath126 & 22.8 & 27@xmath113 & n / n & y & - + 33 & 03:37:36.5 & -35:33:37 & 24.2 & 921@xmath6040 & 41.4 & 60@xmath603 & 125@xmath119 & 161.5 & - & n/@xmath115 & y & - + 34 & 03:37:39.4 & -35:34:17 & 21.9 & 426@xmath6031 & 24.2 & 28@xmath602 & 58@xmath116 & 74.5 & - & n/@xmath115 & y & - + 35 & 03:37:50.1 & -35:21:27 & 34.6 & 139@xmath6037 & 5.8 & [email protected] & 17.6@xmath126 & 22.6 & - & n/@xmath115 & y & - + 36 & 03:39:27.8 & -35:18:52 & 34.6 & 129@xmath6035 & 5.6 & [email protected] & 18.0@xmath126 & 23.2 & 46@xmath120 & n / y & y & - + 37 & 03:37:36.3 & -35:14:47 & 49.0 & 153@xmath6044 & 5.1 & 11@xmath603 & 24@xmath117 & 30.7 & 25@xmath113 & n / n & - & - + 38 & 03:37:33.4 & -35:17:17 & 49.0 & 190@xmath6049 & 6.0 & 13@xmath603 & 27@xmath122 & 35.2 & - & n/@xmath115 & - & - + 39 & 03:37:29.0 & -35:41:11 & 69.3 & 580@xmath60119 & 14.1 & 67@xmath6014 & 140@xmath113 & 180.1 & 232@xmath127 & n/@xmath115 & y & - + 40 & 03:38:14.7 & -35:17:57 & 69.3 & 244@xmath6067 & 5.2 & 15@xmath604 & 31@xmath122 & 40.3 & - & n/@xmath115 & - & - + 41 & 03:38:36.7 & -35:41:32 & 62.9 & 366@xmath6061 & 8.9 & 24@xmath604 & 51@xmath116 & 65.8 & 57@xmath113 & n / n & - & - + 42 & 03:39:12.3 & -35:26:12 & 98.0 & 312@xmath6091 & 4.7 & 19@xmath605 & 39@xmath122 & 50.7 & - & n/@xmath115 & - & - + 43 & 03:38:12.2 & -35:30:02 & 98.0 & 365@xmath60101 & 5.0 & 22@xmath606 & 46@xmath122 & 59.1 & 60@xmath128 & n / n & - & - + 44 & 03:38:05.1 & -35:19:05 & 12.2 & 42@xmath6013 & 4.6 & [email protected] & 5.3@xmath114 & 6.8 & - & n/@xmath115 & - & - + 45 & 03:37:59.9 & -35:23:30 & 17.3 & 56@xmath6018 & 4.4 & [email protected] & 6.8@xmath121 & 8.8 & - & n/@xmath115 & - & - + 46 & 03:37:59.2 & -35:11:37 & 49.0 & 134@xmath6042 & 4.4 & [email protected] & 20@xmath117 & 26.6 & 28@xmath120 & n / n & - & - + we tried to determine which x - ray sources belong to the fornax cluster both by visual inspection of the photographic dss plates and by comparing their positions with those of published catalogs of the fornax region . sources no.2 and 11 are likely to belong to ngc 1399 because of their position within the optical effective radius ( @xmath129 , @xcite ) while sources no.16 and 32 appear to be coincident with the position of two globular clusters @xcite .
two sources are clearly associated with background objects : source no.4 corresponds to a background galaxy located at @xmath130 and represents the x - ray counterpart of the strongest component of the radio source pks 0336 - 35 @xcite ; source no.30 is coincident with two optical background galaxies classified as interacting by @xcite and clearly visible on the dss plates .
for the remaining sources there is no clear indication of whether they belong to the fornax cluster or not .
sources no.8 , 12 , 19 , 21 , 30 are variable at the 95% confidence level , according to our ks tests .
source no.30 is associated with a short x - ray burst occurred in the background galaxy between january and august 1996 .
in fact the source is not detected in both the august 1991 pspc ( rp600043n00 ) and the february 1993 hri observations ( rh600256n00 ) .
it is present , instead , in the last two hri observations ( rh600831n00 and rh600831a01 ) with a decrease in the observed count rate of a factor 8 between the two .
we must notice that our background level was also variable so that the presence of temporal variability was clearly detect only for the strongest sources ( @xmath131 erg s@xmath1 cm@xmath41 ) .
however , since the background subtraction was performed locally ( i.e. extracting the background in a region surrounding the source , see @xcite ) , this variability does not affect the source counts reported in table [ source_tab ] . to further check the reliability of the wavelets algorithm , we extracted the counts manually for each source using a local annulus to measure the background , finding consistent results ( within 1@xmath106 ) with those derived by the wavelets technique .
we cross - checked our source list with the list of sources detected on the pspc image by the galpipe project ( see @xmath20[model ] ) .
several sources detected in the hri field were bright enough and sufficiently isolated to allow a comparison of the pspc flux with the hri results and to check for long - term variability .
we converted the pspc count rates to fluxes in the 0.5 - 2.0 kev band ( 1 cps=6.397@xmath132 ergs cm@xmath41 s@xmath1 ) , assuming the same power law spectral model used for converting the hri counts ( see below ) .
the pspc fluxes and long - term variability ( i.e. when pspc and hri flux estimates differ more than 3@xmath106 ) are included in table [ source_tab ] .
this comparison also confirmed that sources no.39 and 46 , both near the very edge of the hri fov , are not an artifact of the poor s / n ratio or the steep emission gradient because both are present in the pspc data as well and detected with high significance .
we compared our source flux distribution to the expected background counts obtained from a deep survey of the lockmann hole @xcite .
we used the sensitivity map generated by the wavelet algorithm to weight each source for the effective detector area above the corresponding limiting flux . in this way we are able to normalize the measured counts at each flux to the total surveyed area .
we computed the flux of each source assuming a power law spectrum with photon index @xmath133 ( @xmath107 in table [ source_tab ] ) , as found by hasinger and collaborators for faint sources .
this gives a conversion factor of 1 cnt s@xmath1=@xmath134 ergs s@xmath1 cm@xmath41 in the 0.5 - 2.0 kev band ( the same used by hasinger ) .
we also tried a raymond - smith spectrum with @xmath135 and abundance @xmath136 solar , following the results of @xcite in their study of pointlike sources near ngc 507 and obtaining an almost identical result .
the flux distribution is shown in figure [ src_flux ] where our differential counts ( empty circles ) are compared to the hasinger ones ( continuous line ) .
the horizontal error bars take into account uncertainties in the the power law slope .
arrows represent the upper limit obtained fixing the absorption to the upper limit of @xmath137 cm@xmath41 ( @xmath138 in table [ source_tab ] ) , allowed from the spectral analysis performed by rffj .
our counts are higher than the hasinger estimate , giving a total of @xmath139 sources vs @xmath140 expected counts , but still compatible within 2@xmath106 . to take into account absorption uncertainties we estimated a lower limit on the number of expected sources adopting the higher absorption value found by rffj in the ngc 1399 field .
in this case the number of expected counts decreases to @xmath141 leading to an upper limit of the observed excess of @xmath142 .
the spatial distribution of detected sources is not uniform . in figure [ src_flux ]
we show the source excess in circular annuli centered on ngc 1399 .
we find that the excess is peaked on the dominant galaxy , suggesting that the inner sources are likely to be associated with ngc 1399 rather than being background objects .
in fact , excluding the central 300 ( 28 kpc ) , the measured counts drop to @xmath143 , in agreement within less than @xmath144 from the expected value .
the x - ray luminosity of the sources within the central 300 " range from @xmath145 erg s@xmath1 to @xmath146 erg s@xmath1 , if they are in ngc 1399 .
thus all of them have luminosities in excess of @xmath147 erg s@xmath1 , the eddinghton luminosity for a solar mass accreting object , suggesting that they may have massive black hole companions if they are accretion binaries .
while this work was in progress the _ chandra _ data on ngc 1399 became public .
a full data reduction of the _ chandra _ data , which is however already in progress @xcite , is beyond the aim of the present work .
however we decided to compare the rosat data with the _ chandra _ image to see if our conclusions are supported by the improved _ chandra _ resolution and sensitivity .
we used the longest _ chandra _ observation of 60 ks ( obs .
i d 319 , 55.942 ks live time ) , performed on 2000 january 18 with the acis - s detector .
we produced a 0.5 arcsec / pixel image in the energy band 0.3 - 10 kev , from the acis - s3 chip , which covers the central @xmath148 arcmin of ngc 1399 .
the image was then adaptively smoothed with the _ csmooth _ algorithm included in the ciao package .
the final image is shown in figure [ chandra_csmooth ] .
no exposure correction was applied because the _ csmooth _ algorithm works better on the uncorrected data .
however the exposure map is quite flat and we further checked that any feature present in the exposure map does not affect our conclusions .
figure [ chandra_csmooth ] shows a large number of pointlike sources present in the ngc 1399 field , most of which were unresolved in the hri image .
we marked with circles the sources detected in the hri field ( see @xmath20 [ sources ] ) .
chandra _ data show that sources 2 , 11 and 5 , identified as single sources in the hri image , are instead multiple sources .
the filamentary structures found in the hri image ( figure [ centbox ] ) are present in the _ chandra _ image as well . the ` arm ' protruding on the western side and the ` voids ' in the x - ray emission are clearly seen .
the hypothesis that the filamentary structures may be due to the presence of a large number of pointlike sources unresolved in the hri image , is ruled out by _ chandra _ data .
in fact the _ csmooth _ algorithm clearly separates point sources from the extended structures ( figure [ chandra_csmooth ] ) and shows that the two components are not spatially correlated . the correspondence between the halo features and the radio jets / lobes , discussed in @xmath20 [ xradio ] , are confirmed .
moreover the _ chandra _ data revealed the presence of a long arc extending south that follows closely the western edge of the radio lobe .
the arc is better visible on the raw _ chandra _ image shown in figure [ chandra_arc ] .
this feature has a low statistical significance ( @xmath149 ) but the correlation with the radio lobe argues in favor of a real structure .
the study of galactic and cluster gaseous halos has often been based on the assumption of homogeneity and spherical symmetry .
the deep hri image of the ngc 1399 halo has shown instead significant departures from these simplified assumptions . here
we briefly summarize the result of our analysis , and then discuss their physical implications .
we model the x - ray brightness distribution of ngc 1399 with three distinct components : ( 1 ) a central component ( dominating for @xmath150 ) , coincident with the optical galaxy ; ( 2 ) a ` galactic ' component ( @xmath151 ) displaced south - west of ngc 1399 and ( 3 ) an ellipsoidal ` cluster ' component ( @xmath36 ) centered north - east of the optical galaxy .
the cooling times of the central and galactic components are smaller than the hubble time , suggesting the presence of a central cooling flow .
the binding mass estimates , obtained assuming hydrostatic equilibrium , revealed that the dynamics of the central component are dominated by stellar matter while , at radii larger than 2 arcmin there is a large dark matter halo with two distinct components dominating respectively on galactic and cluster scales .
we also found significant filamentary structures and cavities in the x - ray emission .
some of these structures are correlated with the morphology and the position of the radio jets and lobes of ngc 1399 , suggesting interactions between the radio emitting plasma and the hot gas . for what concerns ngc 1404 the surface brightness distribution
is well represented by a single beta profile out to 90 . at larger radii
the effect of ram pressure stripping becomes evident as a sharp brightness cutoff in the north - west sector and an elongated tail on the opposite side of the galaxy .
the cooling times derived by the deprojected density profile are shorter than the hubble time inside the ngc 1404 halo .
the mass profile shows no evidence of multiple components .
finally we used a wavelets based algorithm to detect 46 sources in the hri field , 8 of which show significant time variability .
their distribution is consistent with the background source distribution at the @xmath21 level .
however there is excess over background counts centered on ngc 1399 .
the presence of a large number of pointlike sources near the galaxy center is further confirmed by _
chandra _ data .
the center of the gaseous halo is occupied by a strong emission peak ( see figure [ csmooth]a ) centered on the optical galaxy . such emission is often found in cooling flow galaxies ( @xcite ; see also @xcite ) because of the higher density of cooler hot gas in the galactic center .
rffj revealed the presence of a cooler region within the central 1.5 arcmin ( 14 kpc ) ; this is compatible with the extent of the region dominated by the x - ray peak , as shown by the radial profile in figure [ model_prof ] .
moreover both the central and galactic components have cooling times below the hubble time ( assumed to be @xmath75 years ) within respectively 250 ( 23 kpc ) and 350 ( 32 kpc ) , so that significant cooling must have taken place .
the cooling time of the cluster component , instead , is compatible with the hubble time , suggesting that the cooling flow is mainly of galactic origin .
the alternative hypothesis that the central peak can be due to the presence of a nuclear source @xcite is clearly ruled out by our data , because the observed profile ( figure [ profile ] ) is wider than the hri prf even taking into account the prf uncertainties .
the difference between the optical ( with slope @xmath152 , @xcite ) and x - ray profiles ( @xmath153 for the central component , @xmath20 [ ngc 1399 ] ) seen in figure [ pies ] also suggests that is unlikely that the central peak could be due to stellar sources .
@xcite analyzing asca data , found that the addition of a power - law component to their ngc 1399 spectral model significantly improves the quality of the fit .
they notice that the luminosity of the power - law component is compatible with the contribution expected from a population of unresolved hard x - ray binaries .
however , the hri count rate that would be produced by this power - law component falls more than an order of magnitude below the inner component count rate . we can thus exclude that this hard discrete population dominates the central component emission .
the gravitational mass estimate derived from the hydrostatic equilibrium ( equation [ mass_eq ] ) shows that the hot gas within 100 " follows closely the stellar mass profile .
this result is independent on the assumed temperature profile ( figure [ mass_prof ] ) being mostly determined by the hot gas density profile .
these x - ray data are thus in excellent agreement with the optical mass determinations of @xcite who find no evidence of dark matter in the galaxy core .
the origin of the central density enhancement may be due to hot gas inflow within the inner 1 arcmin .
the deprojection technique described in @xmath20 [ dens_par ] allowed to derive the cumulative gas mass profile of the central component ( figure [ mass_prof ] ) .
according to the mass deposition rate measured by rffj , which is in good agreement with our data , we expect to find @xmath154 of deposited gas within the central 60 ( @xmath67 kpc ) over @xmath6 yr , while we only observe @xmath155 in the hot phase .
taking into account a possible larger than line - of - sight central absorption , the observed mass would increase of @xmath156 , but the deposition rate would also increase ( see rffj ) , so that the result remains almost unchanged .
this difference suggests that @xmath157 of the hot gaseous mass that reached the inner regions of ngc 1399 has cooled enough to become invisible in x - rays .
however rffj argue that the missing mass is not seen in any other phase .
the situation gets even worse considering that we are ignoring stellar mass losses which must be considerable since the metallicity of the ism in the galactic center is consistent with stellar values @xcite .
if the age of the cooling flow is lower than @xmath6 yr , this disagreement would be reduced . equating the observed hot gas mass to the amount expected from the cooling flow
, we estimate a lower limit of @xmath81 years to the age of the cooling flow .
however , we must notice that the latest studies of cooling flows in clusters and galaxies with newton - xmm are showing that the classical cooling flow model is not able to reproduce the observed temperature distributions @xcite .
in this case no conclusion can be drawn from our data until the cooling flow physics will be better understood .
the x - ray brightness profiles ( figure [ pies ] ) shows that past 1 arcmin the hot gas distribution is more extended than the stellar distribution .
if the gas is in hydrostatic equilibrium , additional matter is required to explain this broader distribution .
the change in slope of the x - ray brightness profile for @xmath36 further requires the presence of two components that we identified with the ` galactic ' and ` cluster ' halo .
this complex distribution was previously found by @xcite using the asca data , and explained in terms of a different concentration of dark matter on galactic and cluster scales @xcite .
our data further show that this galactic component is not correlated with stellar matter , which is even more concentrated ( figure [ mass_prof]b ) and dominates the dynamics of the galactic core .
the kinematics inferred from the gcs and pne ( see figure 6 of @xcite ) indicate that for @xmath158 ( 5 kpc@xmath15990 kpc ) from the galaxy center , the velocity dispersion increases with radius .
this result is consistent with our measurements and suggests either the presence of a significant amount of dark matter associated with the outer radii of ngc 1399 or equivalently that the outer galactic dynamics is dominated by the cluster potential .
we must notice however that while on galactic scales the mass profiles may be explained by the presence of dark matter , the deviations from symmetry , i.e. the offset between the stellar component ( consistent with the central component ) and the galactic dark halo , are hardly explained in the galaxy formation framework .
for this reason in next sections we take into account possible departures from equilibrium .
an alternative scenario has recently been advanced by napolitano , arnaboldi & capaccioli ( 2001 , in preparation ) , that could explain our results by means of galaxy interactions .
their calculations show that the energy injection due to galaxy - galaxy encounters can account for the outer galaxy dynamics as seen in pne and gcs , and is in agreement with the observed x - ray temperature profile . in this case
there would be no need for additional dark matter in the ` galactic ' halo .
support for the tidal interaction picture comes from @xcite who showed that the gc system of ngc 1399 is 10 times more abundant than those of neighbors galaxies , while their properties are very similar . in the hypothesis that all galaxies initially had the same number of gcs per unit luminosity , the excess around ngc 1399 may be compatible with tidal stripping of gcs from nearby fornax galaxies .
the effects of tidal interactions on the gaseous halos of elliptical galaxies have been explored by ( * ? ? ?
* hereafter drc ) .
they found that the interaction strongly affects the gas flow , altering the ism density and thus increasing the cooling speed and luminosity .
this mechanism is enhanced when energy injection from sne ia is included and the gas remains turbulent independently from the epoch of the encounter .
in this process the surface brightness is scarcely affected and the global isophotes remain circular , although locally distorted .
the drc model predicts that the gas inflow produces steeper profiles in the central regions of the galaxy in the moment of maximum luminosity , indicating that the tidal perturbations may contribute to the formation of central brightness peaks such as the one seen in the ngc 1399 core .
the steep slope of the galactic component ( table [ fit_tab ] ) too may be due to tidal gas stripping because the galaxy undergoes significative mass losses ( up to 50% ) in the outer regions .
however , we must notice that in the case of ngc 1399 , the effect of tidal stripping may be smaller than what expected from the dcr models because pspc data @xcite suggest a lower sn rate than assumed by dercole and collaborators .
for these models to be consistent with the x - ray observations , tidal interaction must strongly affect the halo density distribution : the increase in the m / l ratio past 2 arcmin ( figure [ mass_prof ] ) is mainly determined by the shallower slope of the gas density profile in the outer galactic regions , since the isothermal model still exhibits the multi - component behavior ( @xmath20 [ dens_par ] ) .
thus a simple temperature increase would not produce the observed agreement with the optical data .
such strong profile alterations are not seen in the drc simulations .
conversely @xcite found that encounters between galaxies of different masses can produce both radial profiles with a central peak and a ` shoulder ' similar to our multi - component structure and azimuthal asymmetries as those seen in the ngc 1399 galactic halo .
however his simulations also show that the stellar profile follows the gas profile , which is not seen in our data .
the x - ray surface brightness analysis ( @xmath20 [ brightness ] ) revealed the presence of complex structures in the galactic halo .
we have shown in @xmath20 [ model ] that these brightness fluctuations are not due to artifacts of the smoothing algorithm , because they are present in the residuals of the bidimensional model , and have high statistical significance .
due to the weak dependence of the hri spectral response on the temperature around 1 kev ( see @xcite ) , they must be due to local density fluctuations of the hot gas .
while some of these features can be due to the interaction between the radio jets and the surrounding ism ( @xmath20 [ xradio ] ) , most of them are not related to any evident radio or optical feature and are spread all around the galaxy , suggesting a ` disturbed ' nature of the whole halo .
it is possible that these structures are the signature of interaction between the ngc 1399 halo and other fornax galaxies .
in fact drc predict significative density fluctuations in the hot gas and the formation of filamentary structures near the galaxy center as a consequence of galaxy - galaxy encounters .
it s not easy to determine if these perturbations are due to the nearby ngc 1404 or to more distant members because , according to drc , the perturbing effects lasts several gyrs .
we searched for possible signatures of interactions between ngc 1399 and ngc 1404 looking for excess emission in the region between the two galaxies .
the slight excess we found is only significant at the 2@xmath106 level , so that no significant conclusion can be drawn from our data . a different scenario that could explain the cluster and galactic halo displacement is one in which ngc 1399 is moving in the fornax cluster potential , thus experiencing ram pressure from the cluster hot gas .
the radial profiles in figure [ pies ] show a high degree of asymmetry with respect to the optical galaxy .
the 2d models indicate that the galactic component is displaced @xmath160 ( 5 kpc ) south - west with respect to the x - ray peak ( and thus optical galaxy ) , while the cluster component seems located @xmath161 arcmin ( 31 kpc ) north - east of ngc 1399 .
@xcite argue that this displacement is compatible with the hypothesis that the galaxy is slowly moving ( 40 km s@xmath1 ) south - east in a larger potential .
this would explain the displacement of the galaxy with respect to the cluster component but not the one of the galactic halo .
in the ram pressure scenario , the displacement and large scale asymmetry of the ` galactic ' halo may suggest that ngc 1399 is not sitting at the very center of the cluster potential and is slowly moving north - east toward the ` cluster ' component .
the effect of ram stripping has already been observed in nearby galaxies .
for instance @xcite observed the presence of a very irregular halo in the virgo galaxy ngc 4472 .
they found a steeper gradient in the x - ray emission in the direction of the cluster center associated with a temperature enhancement .
they calculate that ram stripping is able to explain the observed halo features if the galaxy is moving at @xmath162 km s@xmath1 in the cluster potential .
their rosat hri images also revealed filamentary structures and holes in the galactic halo , very similar to the features that we observed in the adaptively smoothed image of the inner halo ( figure [ centbox ] ) .
additional evidence that galaxies moving in the cluster environment can produce filaments of cooler gas behind them were found by @xcite or @xcite .
the accurate simulation of dynamic stripping in cluster and group elliptical galaxies performed by @xcite , show that the effect of dynamic stripping depends on the galaxy orbit and the duration of the subsonic and supersonic phases which determine the balance within gas losses and accretion onto the galaxy . in general their simulations reveal the formation of complex and elongated tails behind the galaxy ( similar to those seen in ngc 4472 ) and strong decentering of x - ray isophotes due to supersonic stripping .
such irregular structures are not seen in the ngc 1399 galactic halo which seems instead rather circular ( @xmath163 ) .
a very rough estimate of the velocity of ngc 1399 can be obtained supposing that the displacement of the optical galaxy , coincident with the x - ray centroid , with respect to the galactic halo is due to the deceleration of the halo caused by ram pressure from the cluster component . in the hypothesis that the motion is highly subsonic , we can treat the different components as homogeneous spheres and calculate the initial velocity required to produce the observed displacement after a time @xmath164 yr due to the ram pressure of the cluster halo @xmath165 . using the parameters derived form the bidimensional model ( @xmath20 [ model ] ) : @xmath166 , @xmath167 and a displacement @xmath168 , we obtain an initial velocity of @xmath169 km s@xmath1 , consistent with the initial assumption of subsonic motion . even considering a more recent interaction where
@xmath170 the galaxy motion remains highly subsonic ( @xmath171 km s@xmath1 ) .
this result is in agreement with the observations , showing that the halo morphology is only weakly distorted . in conclusion , even though extreme projection effects may hide stronger stripping features , ram pressure may be causing the galaxy halo displacement but seems unlikely to be responsible for the observed inner halo structure of ngc 1399 . for what concerns
the elongated cluster halo component , in many cases irregular features in the cluster halos seems associated with merger events ( see for instance @xcite ) .
simulations of cluster merging @xcite show that elongation and twisting of the x - ray isophotes are a general result of such interactions . in the case of
the fornax cluster @xcite found evidence of a subcluster 3 degrees south - west of ngc 1399 centered on ngc 1316 ( fornax a ) .
they claim that the two components are at the early stage of merging ; in this case merging is unlikely to be the cause of the elongation of the cluster x - ray isophotes . to further study the dynamical status of the cluster halo we would need temperature maps of the whole cluster , as can be obtained with _
chandra _ or xmm .
it is worth noticing that the x - ray centroid of the cluster component is displaced with respect to the optical distribution of fornax cluster galaxies , which tend to concentrate south - west of ngc 1399 ( see figure 1 from @xcite and figure 3 from @xcite ) .
if the merging picture is confirmed , this may imply that the galaxies in the two subclusters are infalling in the global potential well faster than the associated gaseous halos .
@xcite compared three x - ray dominant galaxies in poor groups : ngc 1399 , ngc 507 and ngc 5044 .
they show that even if these galaxies have similar properties ( temperatures , cooling core ) their x - ray brightness profiles are very different .
this suggests that the shape of the potential and the mass deposition mechanism may vary from galaxy to galaxy . if environmental effects ( tidal interactions , ram pressure stripping ) are important factors in determining the ngc 1399 halo structure , they must be considered as well in order to explain the observed differences .
in contrast with ngc 1399 , ngc 1404 possesses a very regular halo , well represented by a beta profile .
the x - ray brightness profile falls as @xmath172 , in good agreement with the optical brightness profile @xmath173 @xcite .
this result suggests that the the gas distribution of the hot halo of ngc 1404 is produced by the same gravitational potential that is binding the stellar population , out to @xmath174 ( @xmath175 kpc ) from the galaxy center ( figure [ 1404profile]b ) .
this result is in agreement with @xcite who found that luminosities and temperatures of x - ray compact galaxies are well explained by kinematical heating of gas supplied by stellar mass loss .
however she notice that ngc 1404 is significantly brighter than other compact ellipticals and explains this result in term of interaction with the ism .
indeed at radii larger than 80 the influence of dynamic stripping is evident from the presence of an elongated tail on the south - east side of the galaxy and the steep surface brightness gradient on the opposite side ( figure [ csmooth ] and [ 1404profile]b ) .
however the regularity of the brightness profile and its similarities to the optical profile within 90 " ( @xmath176 kpc ) implies that either the stripping efficiency is low or that it affects significantly only the outer galactic regions .
it is worth noticing that our brightness profile is shallower than expected from simulations of dynamically stripped galaxies suggesting that the models do not properly describe the influence of ram stripping on the inner galactic halo ( see @xcite ) .
the gas and total mass profiles derived in @xmath20 [ dens_par_1404 ] are in the same range of those of the central component of the ngc 1399 halo , within the inner 80 .
this is not surprising since we have seen that in ngc 1399 the central component seems to be in equilibrium with the stellar population , as it is the case for the ngc 1404 halo .
this result further supports the scenario in which the difference between x - ray compact and extended galaxies lies in the presence of additional components ( i.e. galactic and cluster components ) related to the central position of the dominant galaxies inside the cluster potential well ( e.g. @xcite ) .
signs of interaction between radio jet / lobes and the surrounding gaseous medium have been reported in powerful radio galaxies ( see for instance @xcite for cygnus a ; @xcite for evidence of such interaction in clusters hosting powerful radio sources ) .
ngc 1399 is a faint radio galaxy whose radio lobes are entirely contained within the optical galaxy .
nevertheless the structures seen in figure [ x - radio ] suggest that the same mechanisms at work in more powerful sources may be responsible for some of the galactic halo structures .
the residual image in figure [ x - radio_nuclear ] shows excess emission on both sides of the nucleus and a lack of x - ray emission along the radio jets in the inner 15 " ( @xmath177 kpc ) , as if the radio jets are displacing the ambient gas . for @xmath178 ( @xmath179 kpc ) more x - ray emission is seen in the north and south part of the galaxy , aligned with the direction of the radio jets .
@xcite have studied , by means of hydrodynamical simulations , how these features are produced during the propagation of the radio jet trough the icm .
they show that , for a @xmath37 kev gaseous halo and a detector response similar to the rosat hri ( see their figure 8) , we expect to see excess emission located on the side of the jets , due to the shocked gas near the tail of the radio lobes , as the one seen in figure [ x - radio ] ( region * a * ) and in figure [ x - radio_nuclear ] . in coincidence with the radio lobes , x - ray enhancements or deficits
are expected depending on the thickness of the shocked gas @xmath180 with respect to the lobe radius @xmath181 .
near the edge of the radio lobe , where @xmath182 , the shocked gas density is high enough to produce a brightness enhancement . on the contrary , in the tail of the lobes the compressed gas had time to expand , so that @xmath183 and the x - ray brightness is diminished , producing a ` cavity ' .
this picture is very similar to what we see in our data .
features * b * and * c * in figure [ x - radio ] may suggest respectively the presence of a cavity produced by the southern lobe in the icm and the excess emission associated with the edge of the lobe . the alternative hypothesis , that the excess is due to the presence of a bow shock at the leading end of the lobe , seems unlikely both because carilli and collaborators showed that the bow shock is hardly seen at energies @xmath184 kev and because the signatures of bow shocks are usually sharper than the one seen in our data , showing up as compact bright spots or arcs .
chandra _ image also revealed the presence of a long arc south of ngc 1399 .
the steeper gradient of the radio emission on the western side of the southern lobe , with respect of the eastern side , suggests that this feature may represent a layer of hot gas compressed by the radio emitting plasma .
signatures of interaction with the icm do not show up so clearly in the northern lobe , but this could be due to projection effects ( see for instance @xcite ) .
the brightness enhancement near the west side of the lobe could be due to the thin layer of compressed gas surrounding the lobe while the steeper radio gradient near region * d * suggests that the lobe is avoiding a region of higher gas density .
indeed the crude estimate indicates that the pressure of the bubble ( @xmath20 [ xradio ] ) can be up to 4 times higher than that of the surrounding ism . an upper limit of @xmath185 on the ratio of the electron to heavy particle energy ( or alternatively a lower limit of 0.005 on the filling factor ) can be derived by matching the radio pressure to the thermal pressure of the bubble ( figure [ press_prof ] ) , in agreement with @xcite .
inverse compton ( ic ) scattering of cosmic background photons by the radio emitting particles is often considered as an alternative process to explain the x - ray brightness enhancements near radio lobes @xcite .
we have evaluated this possibility for the x - ray excess near the edge of the southern radio lobe of ngc 1399 , following @xcite , and comparing the minimum energy field strength @xmath186 with the magnetic field @xmath187 required to produce a measured ratio of radio to x - ray flux .
we used the radio flux for the southern lobe measured by @xcite , corrected for the fact that just half of the lobe contributes to the x - ray excess , and we measured the x - ray flux density from the residuals shown in figure [ residuals ] , assuming that all the excess over the 2d model is due to ic scattering .
we find that it is unlikely that more than a few percent of the excess near the edge of the southern radio lobe may be due to ic scattering , even taking into account the poor knowledge of the parameters involved in the calculations .
pressure confinement of the radio lobes @xcite is supported by our data ( figure [ press_prof ] ) .
if the inclination of the radio source is @xmath188 , the radio jets remain collimated up to the radius within which the high density central component dominates ( @xmath189 , figure [ press_prof ] ) and diffuse in lobes as soon as they enter in the lower density environment .
a larger inclination is unlikely because the thermal pressure of the cluster component ( long - dashed line in figure [ press_prof ] ) is of the same order of magnitude as the radio lobes pressure , and would probably not be able to confine the radio emitting plasma , if this was extending at greater distances . finally , we analized the possibility that the power - law component found in the spectral model of @xcite for ngc 1399 ( see @xmath20 [ inner ] ) may be due to an active nucleus .
their estimated flux ( converted to our adopted distance ) is twice the value expected from our upper limit on a possible nuclear source , adopting their best fit parameters . considering that @xcite
further reduce this upper limit by an order of magnitude it appears that the bulk of the power - law emission comes from the discrete source population revealed by the _ chandra _ data ( see figure [ chandra_csmooth ] ) .
the presence of a population of individual x - ray sources in e and s0 galaxies was first suggested by @xcite .
additional evidence was then provided by rosat and asca observations ( e.g. @xcite ) . these x - ray source populations are now been detected down to much lower luminosities with _
chandra _ ( e.g. @xcite ) .
the hri data demonstrate the presence of such population , possibly of accretion binaries , also in ngc 1399 .
this result is confirmed by the _ chandra _ data which reveal a large number of pointlike sources surrounding the galaxy , many of which associated with globular clusters @xcite .
this population of accretion binaries is probably responsible for the hard spectral component @xcite required to fit the high energy spectrum of ngc 1399 .
as demonstrated by comparing the 5 " resolution rosat hri image with the 10 times sharper _ chandra _ acis image , subarcsecond resolution is essential for the study of x - ray source populations in galaxies .
we have analized a deep rosat hri observation centered on the dominant fornax cluster galaxy ngc 1399 and including ngc 1404 , in conjunction with the archival rosat pspc data of the same field .
we found an extended and asymmetric halo extending on cluster scales ( @xmath190 kpc ) .
the halo profile is not consistent with a simple beta model but suggests the presence of three different components .
the combined hri and pspc data were fitted with a multi - component bidimensional model , consisting of : ( 1 ) a ` central ' component ( dominating for @xmath150 ) centered on the optical galaxy whose distribution follows the luminous matter profile ; ( 2 ) a ` galactic ' component ( @xmath151 ) displaced @xmath191 ( 5 kpc ) south - west of ngc 1399 and ( 3 ) a ` cluster ' ellipsoidal component ( @xmath36 ) centered @xmath192 ( 31 kpc ) north - east of the galaxy
. the hri image also revealed the existence of filamentary structures and cavities in the galactic halo , due to density fluctuations in the hot gas .
the presence of these features is confirmed by a preliminary analysis of _ chandra _ data .
the large scale surface brightness distribution can be explained if the galaxy hosts a large dark halo with different dark matter distributions on galactic and cluster scales @xcite .
alternative models that explain the gcs abundance and optical velocity dispersion profiles through tidal interactions with fornax cluster galaxies ( @xcite ; napolitano , arnaboldi & capaccioli 2001 , in preparation ) are compatible with our data only if such encounters produce a significative flattening of the outer gas distribution .
tidal interactions may also explain the presence of the density fluctuations in the galactic halo @xcite .
ram pressure from the cluster halo ( e.g. @xcite ) is able to account for the decentering of the optical galaxy with respect to the galactic halo if ngc 1399 is moving subsonically ( 10 - 100 km s@xmath1 ) in the cluster potential .
the displacement of the center of the cluster component with respect to the nearest fornax galaxies may suggest that the cluster is not relaxed and may be undergoing a merger event .
we do not detect the x - ray counterpart of the nuclear radio source @xcite and pose an upper limit of @xmath98 erg s@xmath1 on its luminosity in the 0.1 - 2.4 kev band . in agreement with killeen and collaborators ,
our data support thermal confinement of the radio emitting plasma .
we found x - ray emission enhancements aligned with the radio jets that may be due to shocked gas and x - ray ` holes ' and enhancements coincident with the position of the radio lobes .
the latter are consistent with a scenario in which the hot gas is displaced by the radio plasma pressure ( e.g. @xcite ) , while alternative models of inverse compton scattering of cmb photons fail to account for the observed excesses .
ngc 1404 possess a very symmetric halo within 8 kpc well fitted by a simple beta model .
the agreement with the optical profile and the similarity with the central component of the ngc 1399 halo suggest that the ism distribution of ngc 1404 is produced by the same gravitational potential that is binding the stellar population . at larger radii
the influence of dynamic stripping is evident from the presence of an elongated tail on the south - east side of the galaxy and the steep surface brightness gradient on the opposite side .
we detected 43 discrete sources in our field .
their flux distribution is consistent at the @xmath149 level with the number expected from x - ray background counts .
however , there is a significant excess ` peak ' on the central galaxy in the spatial distribution of these sources , suggesting a population of galaxian x - ray sources in ngc 1399 .
this is confirmed by the _
chandra _ data , that reveal a large number of point - like sources associated with the optical galaxy .
this sources are likely to explain the hard component found by rosat and asca spectral analysis @xcite .
we would like to thank dan harris for useful discussions on the radio data interpretation , f. damiani and s. sciortino for advice on the use of the wavelets algorithm and the galpipe database .
helpful comments and suggestions were provided by the referee dr .
loewenstein and the editor dr .
+ we aknowledge partial support from the cxc contract nas8 - 39073 and nasa grant nag5 - 3584 ( adp ) , from the murst prin - cofin 1998 - 1999 ( resp .
g. peres ) and from the european social fund ( f.s.e . ) .
+ 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 .
allen , s. w. , di matteo , t. , & fabian , a. c. 2000 , mnras , 311 , 493 angelini , l. , loewenstein , m. , & mushotzky , r. f. in press in apjl , astro - ph/0107362 arnaboldi , m. , freeman , k. c. , hui , x. , capaccioli , m. , & ford , h. 1994 , messenger , 76 , 40 barnes , j. e. 2000 , to appear in `` stellar collisions , mergers , and their consequences '' , m. shara , ed .
, astro - ph/0010145 bhringer , h. , voges , w. , fabian , a.c . ,
edge , a.c . , & neumann , d.m . , 1993 ,
mnras 264 , l25 brunetti , g. , cappi , m. , setti , g. , feretti , l. , & harris , d.e .
2001 , a&a , 372 , 755 buote , d. a. 1999 , mnras , 309 , 685 canizares , c. r. , stewart , g. c. , & fabian , a. c. 1983 , apj , 272 , 449 carilli , c.l . ,
perley , r.a . , & harris , d.e .
1994 , mnras , 270 , 173 carter , d. , & malin , d. f. 1983 , mnras , 203 , 49 clarke , d.a . ,
harris , d.e . , & carilli , c.l .
1997 , mnras , 284 , 981 damiani , f. , maggio , a. , micela , g. , sciortino , s. 1997a , apj , 483 , 350 damiani , f. , maggio , a. , micela , g. , sciortino , s. 1997b , apj , 483 , 370 david , l. p. , jones , c. , forman , w. , & daines , s. 1994 , apj , 428 , 544 david , l. p. , harden , jr .
, f. r. , kearns , k. e. , & zombeck , m. v. 1996 , the rosat high resolution imager ( hri ) , calibration report 1996 february , revised , u.s . rosat science data center , sao , cambridge , ma dercole , a. , recchi , s. , & ciotti , l. 2000 , apj , 533 , 799 ( drc ) drinkwater , m. j. , gregg , m. d. , & colless , m. 2001 , apj , 548 , l139 ekers , r. d. , et al .
1989 , mnras , 136 , 737 fabbiano , g. , kim , d .- w . , &
trinchieri , g. 1994 , apj , 429 , 94 fabian , a. c. , nulsen , p. e. j. , & canizares , c. r. 1991 , a&a rev , 2 , 191 fabian , a.c . , sanders , j.s . ,
ettori , s. , taylor , g.b . , allen , s.w . ,
crawford , c.s .
, iwasawa , k. , johnstone , r.m . , & ogle , p.m. 2000 , mnras , 318 , l65 fabian , a. c. , sanders , j. s. , ettori , s. , taylor , g. b. , allen , s. w. , crawford , c. s. , iwasawa , k. , & johnstone , r. m. 2001 , mnras , 321 , l33 fabricant , d. , lecar , m. , & gorenstein , p. 1980 , apj , 241 , 552 feigelson , e. d. , laurent - muehleisen , s. a. , kollgaard , r. i. , & fomalont , e. b. 1995 , apj , 449 , l149 ferguson , h. c. 1989 , aj , 98 , 367 finoguenov , a. , & jones , c. 2001 , apj , 547 , l107 forbes , d. a. , grillmair , c. j. , williger , g. m. , elson , r. a. w. , & brodie , j , p. 1998 , mnras , 293 , 325 graham , a. w. , colless , m. m. , busarello , g. , zaggia , s. , & longo , g. 1998 , a&as , 133 , 325 goudfrooij , p. , hansen , l. , jrgensen , h. e. , nrgaard - nielsen , h. u. , de jong , t. , & van den hoek , l. b. , 1994 , a&as , 104 , 179 grillmair , c. j. , freeman , k. c. , bicknell , g. v. , carter , d. , couch , w. j. , sommer - larsen , j. , & taylor , k. 1994 , apj , 422 , l9 harris , d. e. , & grindlay , j. e. 1979 , mnras , 188 , 25 harris , d. e. , silverman , j. d. , hasinger , g. , & lehmann , i. 1998 , a&as , 133 , 431 harris , d. e. 1999 , american astronomical society , head meeting 31 , 17.12 hasinger , g. , burg , r. , giacconi , r. , hartner , g. , schmidt , m. , trumper , j. , zamorani , g. , 1993 , a&a , 275 , 1 hilker , m. , kissler - patig , m. , richtler , t. , infante , l. , & quintana , h. 1999 , a&as , 134 , 59 hilker , m. , infante , l. , & richtler , t. 1999 , a&as 138 , 55 irwin , j. a. , & sarazin , c. l. 1996 , apj , 471 , 683 ikebe , i. , et al .
1992 , apj , 384 , l5 ikebe , y. , et al .
1996 , nature , 379 , 427 jones , c. , stern , c. , forman , w. , breen , j. , david , l. , & tucker , w. 1997 , apj , 482 , 143 kaastra , j. s. , ferrigno , c. , tamura , t. , paerels , f. b. s. , peterson , j. r. , & mittaz , j. p. d. 2001 , a&a , 365 , l99 kaneda , h. , tashiro , m. , ikebe , y. , ishisaki , y. , kubo , h. , makshima , k. , ohashi , t. , saito , y. , tabara , h. , & takahashi , t. 1995 , apj , 453 , l13 killeen , n. e. b. , & bicknell , g. v. 1988 , apj , 325 , 165 killeen , n. e. b. , bicknell , g. v. , & ekers , r. d. 1988 , apj , 325 , 180 kim , d .- w . , fabbiano , g. , & trinchieri , g. 1992 , apjs , 80 , 645 kim , d .- w . , fabbiano , g. , matsumoto , h. , koyama , k. , & trinchieri , g. 1996 , apj , 468 , 175 kim , d .- w . , & fabbiano , g. 1995 , apj , 441 , 182 kissler - patig , m. , grillmair , c. j. , meylan , g. , brodie , j. p. , minniti , d. , & goudfrooij , p. 1999 , aj , 117 , 1206 kriss , g. a. , cioffi , d. f. , & canizares , c. r. 1983 , apj , 272 , 439 loewenstein , m. , mushotzky , r. f. , angelini , l. , arnaud , k. a. , & quataert , e. apj , 555 , l21 mackie , g. , et al .
1996 , in asp conference series , vol . 101 , `` astronomical data analysis software and systems v '' , g. h. jacoby and j. barnes eds .
makishima , k. , et al .
2001 , pasj , 53 , 401 matsumoto , h. , koyama , k. , awaki , h. , tsuru , t. , loewenstein , m. , & matsushita , k. 1997 , apj , 482 , 133 matsushita , k. , ohashi , t. , & makishima , k. 2000 , pasj , 52 , 685 matsushita , k. 2001 , apj , 547 , 693 mazzotta , p. , markevitch , m. , vikhlinin , a. , forman , w. r. , david , l. p. , & vanspeybroeck l. 2001 , apj , in press mcnamara , b.r . ,
2000 , apj , 530 , l81 peterson , j. r. , et al .
2001 , a&a , 365 , l104 rangarajan , f. v. n. , fabian , a. c. , forman , w. r. , & jones , c. 1995 , mnras , 272 , 665 ( rffj ) roettiger , k. , burns , j. o. , & locken , c. 1996 , apj , 473 , 651 saglia , r. p. , kronawitter , a. , gerhard , o. , & bender , r. 2000 , apj , 119 , 153 sarazin , c. l. , & white , r. e. 1987 , apj , 320 , 32 sarazin , c. l. 1988 , `` x - ray emission from clusters of galaxies '' , ed .
cambridge university press sarazin , c. l. , irwin , j. a. , & bregman , j. n. 2000 , apj , 544 , l101 sarazin , c. l. , irwin , j. a. , & bregman , j. n. 2001 , apj , 556 , 533 serlemitsos , p. j. , loewenstein , m. , mushotzky , r. f. , marshall , f. e. , & petre , r. 1993 , apj , 413 , 518 smith , e. p. , et al .
1996 , apjs , 104 , 287 snowden , s. l. , mccammon , d. , burrows , d. n. , & mendenhall , j. a. 1994 , apj , 424 , 714 ( smb ) schombert , j. m. 1986 , apjs , 60 , 603 sulkanen , m. e. , & bregman , j. n. 2001 , aj , 548 , l131 sun , m. , murray , s. s. , markevitch , m. , & vikhlinin , a. 2000 , to appear in the proceedings of the american astronomical society , head meeting # 32 , # 13.05 tamura , t. , et al .
2001 , a&a 365 , l87 toniazzo , t. , schindler , s. 2001 , mnras , 325 , 509 trinchieri , g. , & fabbiano , g. 1985 , apj , 296 , 447 white , d. a. 1992 , phd thesis , univ . | we present the analysis of a deep ( 167 ks ) rosat hri observation of the cd galaxy ngc 1399 in the fornax cluster , comparing it with previous work on this galaxy and with recent _ chandra _ data .
we find , in agreement with previous observations , an extended and asymmetric gaseous halo with a luminosity ( in the 0.1 - 2.4 kev energy band ) of @xmath0 erg s@xmath1 within 46 kpc ( assuming a distance of d=19 mpc ) .
using both hri and , at larger radii , archival pspc data , we find that the radial behavior of the x - ray surface brightness profile is not consistent with a simple beta model and suggests instead three distinct components .
we use a multi - component bidimensional model to study in detail these three components that we identify respectively with the cooling flow region , the galactic and the cluster halo . from these data
we derive a binding mass distribution in agreement with that suggested by optical dynamical indicators , with an inner core dominated by luminous matter and an extended dark halo differently distributed on galactic and cluster scales . the hri data and a preliminary analysis of _ chandra _ public data ,
allow us to detect significant density fluctuations in the halo .
we discuss possible non - equilibrium scenarios to explain the hot halo structure , including tidal interactions with neighboring galaxies , ram stripping from the intra - cluster medium and merging events . in the innermost region of ngc 1399
, the comparison between the x - ray and radio emission suggests that the radio emitting plasma is displacing and producing shocks in the hot x - ray emitting gas .
we do not detect the nuclear source in x - rays and we pose an upper limit of @xmath2 erg s@xmath1 ( 0.1 - 2.4 kev ) to its x - ray luminosity .
we found that the ngc 1404 halo is well represented by a single symmetric beta model and follows the stellar light profile within the inner 8 kpc .
the mass distribution is similar to the ` central ' component of the ngc 1399 halo . at larger radii ram pressure stripping from the intra - cluster medium
produces strong asymmetries in the gas distribution .
finally we discuss the properties of the point source population finding evidence of correlation between the source excess and ngc 1399 . |
radial velocity ( rv ) surveys of nearby stars have been employed in the search for extra - solar planets for nearly two decades ( see marcy , cochran & mayor 2000 ) . as these efforts continue into the next decade , they will be supplemented by precision astrometric searches , e.g. , by fame , keck interferometer , and sim @xcite . in previous papers , we examined the rv and astrometric techniques in detail , paying particular attention to the regime where the time - baseline of the observations is shorter than the orbital period of the extra - solar companion ( eisner & kulkarni 2001a , b ; hereafter ek2001a , b ) .
this regime is interesting because one expects giant planets to form in the colder regions of the proto - planetary nebula , and thus one expects such objects to possess periods of many years to centuries @xcite . in ek2001a , b we demonstrated that one can achieve a significant improvement in sensitivity ( over current techniques ) if the orbital amplitude _ and phase _ are included in the analysis . here
, we examine the benefits of combining simultaneous astrometric and rv observations .
specifically , we examine the sensitivity of a combined astrometric and rv detection technique applied to an edge - on orbit , where the full rv signature and one dimension of the astrometric signature can be observed .
the plan for the paper is straightforward .
first , we simulate large numbers of hypothetical data sets containing ( 1 ) noise only , and ( 2 ) signal and noise , and determine the frequentist type i and ii errors . as in ek2001a
, b we acknowledge that a frequentist approach is not as rigorous as a full bayesian analysis . however , this approach is simple enough that it is amenable to deriving ( semi-)analytical estimates of the sensitivity a principal goal of the paper .
we conclude by discussing the parameter space opened up by combining fame , keck interferometer , or sim astrometric surveys with ongoing precision rv studies .
we will assume edge - on circular orbits throughout this discussion .
the astrometric signature of an edge - on circular orbit is given by @xmath0 where @xmath1 and @xmath2 are the proper motion and parallax of the planetary system , respectively , and @xmath3 here , @xmath4 is the distance to the system , @xmath5 is the mass of the star , and @xmath6 is the mass of the planet .
we ignore the annual parallax .
however , annual parallax should be included in modeling of planets with periods around one year .
the rv signature of this orbit is given by the derivative of the orbital position along the line of sight : @xmath7 here @xmath8 is the radial velocity of the planetary system , and @xmath9 thus , we can express the sensitivity ( defined as the minimum - mass planet that can be detected ) of the rv and astrometric techniques in terms of @xmath10 : @xmath11 however , it is more difficult to identify planets with long periods than equation [ eq : sensitivity ] might suggest . in the so - called `` long - period regime '' , defined as @xmath12 where @xmath13 is the duration of the survey , we observe a fraction of the orbit . as a result , in this regime
, the sensitivity is expected to depend critically on the orbital phase .
the reflex velocity is covariant with @xmath8 and thus the rv technique is most sensitive when @xmath14 ( ek2001a ) .
in contrast , the astrometric signal of an edge - on orbit is covariant with @xmath15 and @xmath2 , and thus the astrometric technique is sensitive when @xmath16 ( ek2001b ) . thus the rv and astrometric techniques achieve their maximal sensitivities for different orbital phases , and we expect , on general grounds , that combining the two techniques should yield a substantial benefit in the long - period regime .
the signal analysis for the astrometric and rv techniques consists of fitting the observations to the models specified in equations [ eq : astrom - orbit ] and [ eq : rv - orbit ] . as noted by several authors ( e.g. * ? ? ?
* ; * ? ? ?
* ek2001a ) the most optimal fitting is obtained by using the technique of least squares .
first , we convert the physical model specified by equations [ eq : astrom - orbit ] and [ eq : rv - orbit ] to equations linear in the unknowns : @xmath17 @xmath18 here , @xmath19 , @xmath20 , @xmath21 , @xmath22 , and @xmath23 . in ek2001a , b we discuss the importance of the @xmath8 , @xmath1 and @xmath2 terms .
these three variables are not directly relevant in detecting or characterizing a companion planet but they are unknown and in the long - period regime are covariant with some of the orbital parameters @xcite .
thus the three variables must be solved for in order to correctly model the observations .
using equations [ eq : lsq - model1 ] and [ eq : lsq - model2 ] as our physical model , we perform the following analysis .
first , we simulate a large number of data - sets containing only gaussian nose ( i.e. no signal ) . for each of these data sets
, we perform a least squares fit to three models : a model using only astrometric measurements ( equation [ eq : lsq - model1 ] ) , a model using only rv measurements ( equation [ eq : lsq - model2 ] ) , and a model that utilizes both astrometric and rv measurements . in each case , for each simulated data set we fit for amplitude and phase .
we note here that for the rv+astrometry model , since the two measurements have different variances we minimize the @xmath24 ( where @xmath25 is the difference between the model and the rms - weighted measurements ) . specifically , we simulate @xmath26 data - sets , sampled at one month intervals for @xmath27 years ( with no loss of generality , we take the time interval to go from @xmath28 to @xmath29 ) , and we explore periods from 5 to 100 years .
we assume that the measurement noise in both the rv and astrometric surveys is characterized by gaussian noise with rms of @xmath30 and @xmath31 respectively .
the best achieved @xmath32 m s@xmath33 @xcite
. the anticipated astrometric precision of fame is between 50 and 100 @xmath2as @xcite , that of the keck interferometer ( narrow angle ) between @xmath34 and 50 @xmath2as @xcite , and that of sim between 1 and 10 @xmath2as @xcite .
we note that @xmath35 @xmath2as yields approximately equivalent sensitivity to rv technique with 3 m s@xmath33 rms for a planet orbiting a star located at distance @xmath36 pc with @xmath37 years ( equation [ eq : rv - amplitude ] ) .
next , for each of three models , we determine the ellipse ( in @xmath10@xmath38 space ) within which 99% of the fitted amplitudes and phases lie .
this ellipse , denoted by @xmath39 , describes the `` type i '' errors of the detection technique .
thus the inferred @xmath10 and @xmath38 have a 1% chance of being outside the @xmath39 ellipse ( in the absence of a signal ) . as discussed earlier ( [ sec : equations ] ) we expect rv and astrometric models to show orthogonal sensitivity .
indeed , as can be seen from figure [ fig : ellipses ] , the @xmath40 and @xmath41 are @xmath42 out of phase in the long period regime . on a basic level , we can understand the benefit of combining rv and astrometric observations by noting that the intersection of @xmath43 and @xmath44 is much smaller than either of the individual ellipses , and thus it is easier to detect signals over the level of the noise .
in fact , this is verified by the combined analysis : the ellipse for the combined analysis , @xmath45 , lies entirely within _ both _ @xmath40 and @xmath44 ( figure [ fig : ellipses ] ) .
analytic expressions for the type i errors for rv or astrometric techniques are given in ek2001a and ek2001b , respectively .
given these expressions , it is not difficult to infer an analytic expression for the type i errors in the case of combined astrometric+rv technique .
as noted earlier for @xmath46 m s@xmath33 and @xmath47 @xmath2as , the semi - minor axes for @xmath48 and @xmath49 are approximately equal ( @xmath36 pc ) , and thus @xmath39 for the combined analysis will be a circle whose radius is given by @xmath50 here , @xmath51 , @xmath52 , and the factor of @xmath53 reflects the fact that in the short - period regime , there are essentially twice as many measurements . as illustrated in figure [ fig : ac ] , this analytic function provides an excellent fit to the data .
next , we evaluate `` type ii '' errors for the three models .
type ii errors describe the probability of failing to detect a genuine signal due to contamination by noise . to understand the type ii statistics , we simulate a large number of data sets consisting of a simulated signal and noise ( see ek2001a , b for further details ) .
the signal is a sinusoidal wave with an amplitude @xmath54 ( astrometry ) , and the corresponding velocity amplitude is @xmath55 ( we set @xmath36 pc ) ; the phase , @xmath38 is randomly chosen from the interval @xmath56 $ ] ( uniform distribution ) . for each model , we increment the amplitude[s ] until 99% of the fitted orbital parameters lie outside of the appropriate @xmath57 ellipse ; this amplitude is denoted by @xmath58 ( for each method ) .
the benefit of combining rv and astrometric analysis accrues mainly from the fact that the error ellipses for the two techniques in @xmath59 parameter space are perpendicular to each other ( figure [ fig : ellipses ] ) .
rv+astrometry analysis will be most useful in cases where the error ellipses for the two techniques are roughly the same size ( otherwise , one error ellipse might lie entirely within the other , and no additional benefit would arise from combining the two techniques ) . as mentioned above ,
the current precision of rv techniques is @xmath60 m s@xmath33 @xcite , which means that for a 10 year survey , we must use astrometric measurements with @xmath61 @xmath2as precision ( for a system at @xmath36 pc ) to reap the maximal benefit from rv+astrometry technique .
this is approximately the sensitivity that will be obtained by future instruments like keck interferometer and fame @xcite .
as illustrated by figure [ fig : m99]a , rv+astrometry analysis ( with comparable rv and astrometric measurement accuracies ) applied to edge - on orbits attains approximately the same sensitivity as an astrometric analysis applied to face - on orbits .
this similarity stems from the fact that in both cases , the highly elliptical 1-d @xmath39 is circularized through the addition of a second dimension .
another way of thinking about this is that no matter what part of the orbit , when we observe both dimensions we can always see the full orbital curvature .
thus , combining astrometric and rv techniques in a large survey ensures good sensitivity for all orbital inclination angles .
it is also worth noting that rv+astrometry yields valuable gains in the short - period regime ( @xmath62 ) .
when @xmath30 is comparable to @xmath31 , the sensitivity of rv+astrometry is better by @xmath53 over rv or astrometry alone .
furthermore , noting that the sensitivity of astrometry to face - on orbits is @xmath63 ( ek2001b ) , we see that the short - period sensitivity of rv+astrometry is approximately independent of orbital inclination .
we have also examined the sensitivity when astrometric data is combined with rv data of a longer time - baseline , specifically for several upcoming missions ( recall that rv surveys will have been underway for 1520 years by the time astrometric surveys commence ) .
we investigate fame ( @xmath64 years , @xmath65 years ) , keck interferometer ( @xmath64 years , @xmath66 years ) , and sim ( @xmath67 years , @xmath66 years ) .
we find that in the cases of fame and keck interferometer , the addition of longer time - baseline rv measurements has a significant impact ( figures [ fig : m99]b[fig : m99]c ) .
in fact , rv+astrometry analysis can easily detect saturns when astrometric analysis alone does nt come close ( figure [ fig : m99]b ) . in the case of sim ,
the main benefit of rv+astrometry is that one can achieve optimal sensitivity over a wider range of inclination angles ( figure [ fig : m99]d ) .
we have also examined the sensitivities of the rv , astrometric , and combined techniques applied to shorter duration surveys , in order to compare the various sensitivities for short - period companions .
specifically , we investigate the prospect of finding companions around m dwarfs with a 2-year survey combining rv measurements with astrometric measurements from ao systems on large telescopes like keck or the palomar @xmath68 .
previous authors have successfully searched for companions around m - dwarfs using rv measurements and astrometric measurements on ao systems ( e.g. , * ? ? ?
* ) although they have not used the combined rv+astrometry analysis described here ( i.e. , they analysed the rv data and the astrometric data separately ) .
as illustrated by figure [ fig : bd ] , the rv technique gains significantly over astrometry for short periods ( because @xmath69 ; equation [ eq : rv - amplitude ] ) . if astrometry is to contribute meaningfully then @xmath70 mas .
there is some expectation that such a precision can be obtained for binary stars @xcite .
if so , combined rv+astrometry surveys of nearby m dwarfs can measure masses of jupiter and saturn - mass companions .
99 danner , r. , unwin , s. , & allen , r.j .
1999 , space interferometry mission : taking measure of the universe ( washington , dc : nasa ) dekany , r. , angel , r. , hege , k. , & wittman , d. 1994 , ap&ss , 212 , 299 delfosse et al . 1999delfosse+99delfosse , x. , forveille , t. , beuzit , j .- l . ,
udry , s. , mayor , m. , & perrier , c. 1999344897 marcy , g.d . ,
cochran , w.d . , & mayor , m. 2000 , protostars and planets iv , eds v. mannings , a.p . boss , & s.s .
russell , university of arizona press , p. 1285 | the astrometric and radial velocity techniques of extra - solar planet detection attempt to detect the periodic reflex motion of the parent star by extracting this periodic signal from a time - sampled set of observations .
the extraction is generally accomplished using periodogram analysis or the functionally equivalent technique of least squares fitting of sinusoids . in this paper
, we use a frequentist approach to examine the sensitivity of least squares technique when applied to a combination of radial velocity and astrometric observations . we derive a semi - analytical expression for the sensitivity and show that the combined approach yields significantly better sensitivity than either technique on its own .
we discuss the ramifications of this result to upcoming astrometric surveys with fame , the keck interferometer , and sim . |
one of the most dramatic discoveries of the modern cosmology in the past decade is that our universe is currently accelerating @xcite . a component that causes the accelerated expansion of the universe
is referred to as dark energy " . although the nature of such dark energy is still speculative ,
an overwhelming flood of papers has appeared which attempt to describe it by devising a great variety of models . among them are cosmological constant , exotic fields such as phantom or quintessence , modified gravity , etc , see @xcite@xcite for a recent review .
available models of dark energy differ in the value and variation of the equation of state parameter @xmath0 during the evolution of the universe .
the cosmological constant with @xmath1 , is located at a central position among dark energy models both in theoretical investigation and in data analysis @xcite . in quintessence
@xcite , chaplygin gas @xcite and holographic dark energy models @xcite , @xmath0 always stays bigger than @xmath2 .
the phantom models of dark energy have @xmath3 @xcite
. however , following the more accurate data analysis , a more dramatic result appears showing that the time varying dark energy gives a better fit than a cosmological constant and in particular , @xmath0 can cross @xmath2 around @xmath4 from above to below @xcite .
although the galaxy cluster gas mass fraction data do not support the time - varying @xmath0 @xcite , theoretical attempts towards the understanding of the @xmath0 crossing @xmath2 phenomenon have been started .
some dark energy models , such as the one containing a negative kinetic scalar field and a normal scalar field @xcite , or a single scalar field model @xcite and interacting holographic dark energy models @xcite have been constructed to gain insight into the occurrence of the transition of the dark energy equation of state and the mechanism behind this transition .
other studies on the @xmath1 crossing have been carried out in @xcite .
independent of the challenge we deal with the dark energy puzzle , in recent years , theories of large extra dimensions , in which the observed universe is realized as a brane embedded in a higher dimensional spacetime , have received a lot of interest . according to the braneworld scenario the standard model of particle fields
are confined on the brane while , in contrast , the gravity is free to propagate in the whole spacetime . in these theories the cosmological evolution on the brane
is described by an effective friedmann equation that incorporates non - trivially with the effects of the bulk onto the brane .
an interesting consequence of the braneworld scenario is that it allows the presence of five - dimensional matter which can propagate in the bulk space and may interact with the matter content in the braneworld .
it has been shown that such an interaction can alter the profile of the cosmic expansion and lead to a behavior that would resemble the dark energy .
the cosmic evolution of the braneworld with energy exchange between brane and bulk has been studied in different setups @xcite@xcite@xcite@xcite . in these models , due to the energy exchange between the bulk and the brane , the usual energy conservation law on the brane
is broken and consequently it was found that the equation of state of the effective dark energy may experience the transition behavior ( see e.g @xcite ) .
on the other hand , in string theory , in addition to the einstein action , some higher derivative curvature terms have been included to derive the gravity . in order to obtain a ghost - free theory ,
the combination of quadratic terms called gauss - bonnet term is usually employed as curvature corrections to the einstein - hilbert action @xcite . from a geometric point of view
, the combination of the einstein - hilbert and gauss - bonnet term constitutes , for 5d spacetimes , the most general lagrangian to produce second - order field equations @xcite .
the gauss - bonnet correction significantly changes the bulk field equations and leads to modifications in the braneworld friedmann equations .
therefore , the study of the effects of the gauss - bonnet correction term on the evolution of the universe in the braneworld scenario is well motivated .
influences of the gauss - bonnet correction on the dgp braneworld have been studied in @xcite .
the purpose of the present work is to investigate the effects of the bulk content in the general gauss - bonnet braneworld on the evolution of the universe .
although the effects of the gauss - bonnet correction term on the late time universe is small , we will see that it still plays an important role in the cosmic evolution . besides we will show that the combination of the dark radiation term and the matter content of the bulk plays the role of the dark energy on the brane and influences the evolution of the universe . in our model , in contrast to the previous models ( @xcite ) ,
we do not need to break down the standard energy momentum conservation law on the brane , although our model can allow such assumption if one is interested .
we will show that by suitably choosing model parameters , our model can exhibit accelerated expansion of the universe .
in addition , we will present a profile of the @xmath0 crossing @xmath2 phenomenon which is in good agreement with observations .
the paper is organized as follows . in section [ model ] , we present a braneworld model to describe the accelerated expansion and the effective equation of state of dark energy in the presence of the gauss - bonnet correction term in the bulk . in section
[ late ] , we study the cosmological consequences of the model and in particular , its effect on the evolution of the universe .
the last section is devoted to conclusions and discussions .
the theory we are considering is five - dimensional and has an action of the form @xmath5 where @xmath6 is the bulk cosmological constant and @xmath7 is the gauss - bonnet correction term @xmath8 here @xmath9 and @xmath10 are the bulk and brane metrics , respectively .
@xmath11 , @xmath12 , and @xmath13 are the scalar curvature and ricci and riemann tensors , respectively . throughout this paper
we choose the unit so that @xmath14 as the gravitational constant in five dimension .
we have also included arbitrary matter content both in the bulk and on the brane through @xmath15 and @xmath16 respectively , and @xmath17 is the positive brane tension .
the field equations can be obtained by varying the action ( [ act ] ) with respect to the bulk metric @xmath18 .
the result is @xmath19 where @xmath20 is the second - order lovelock tensor @xmath21 for convenience and without loss of generality , we can choose the extra - dimensional coordinate @xmath22 such that the brane is located at @xmath23 and bulk has @xmath24 symmetry .
we are interested in the cosmological solution with a metric @xmath25 where @xmath26 is a maximally symmetric @xmath27-dimensional metric for the surface ( @xmath28=const . , @xmath22=const .
) , whose spatial curvature is parameterized by k = -1 , 0 , 1 .
the metric coefficients @xmath29 and @xmath30 are chosen so that , @xmath31 and @xmath32 , where @xmath28 is cosmic time on the brane .
the total energy - momentum tensor has bulk and brane components and can be written as @xmath33 the first and the second terms are the contribution from the energy - momentum tensor of the matter field confined to the brane and the brane tension @xmath34 where @xmath35 , and @xmath36 , being the energy density and pressure on the brane , respectively .
in addition we assume an energy - momentum tensor for the bulk content of the form @xmath37 the quantities which are of interest here are @xmath38 and @xmath39 , as these two enter the cosmological equations of motion .
in fact , @xmath39 is the term responsible for energy exchange between the brane and the bulk . integrating the @xmath40 component of the field equations ( [ feq ] ) across the brane and imposing @xmath24 symmetry
, we have the jump across the brane @xcite @xmath41\frac{a'_{+ } } { a_{0 } } = -\frac{1}{6}(\rho+\sigma),\end{aligned}\ ] ] where @xmath42 is the discontinuity of the first derivative .
@xmath43 is the hubble parameter on the brane .
( [ jc1 ] ) is a cubic equation for the discontinuity @xmath44 , which has only one real solution , the other two being complex .
therefore , if we require our cosmological equations to have the right @xmath45 limit we are left with only one solution . however , this real root is too lengthy and complicated to present here . since we are interested to study the effect of the gauss - bonnet correction term on the evolution of the universe in the late time so it is reasonable to choose the gauss - bonnet coupling constant @xmath46 to be small , namely @xmath47 . using this fact we can expand the real solution for @xmath44 versus @xmath46 powers .
the result for @xmath48 up to order @xmath46 is @xmath49 in a similar way , integrating the @xmath50 component of the field equations ( [ feq ] ) across the brane and imposing @xmath24 symmetry , we can obtain the discontinuity in the metric function @xmath51 , which for @xmath48 can be written up to o(@xmath46 ) in the following form @xmath52 where dots denote time derivatives and primes denote derivatives with respect to @xmath22 . at this point
we find it convenient to absorb the brane tension @xmath17 in @xmath35 and @xmath36 with the replacement @xmath53 and @xmath54 .
therefore the junction conditions @xmath55 and @xmath56 can be simplified @xmath57 substituting the junction conditions @xmath58 and @xmath59 into the @xmath60 and @xmath61 components of the field equations @xmath62 , we obtain the modified friedmann equation and the semi - conservation law on the brane ( up to order @xmath46 ) @xmath63 and @xmath64.\end{aligned}\ ] ] we shall assume an equation of state @xmath65 to hold between the energy density and pressure of matter on the brane .
therefore we have @xmath66 one can easily check that in the limit @xmath67 , eqs .
( [ ju1])-([t2 ] ) reduce to the corresponding equations of the braneworld model without gauss - bonnet correction term @xcite .
remarkably , we can show that the friedmann equation ( [ fri2 ] ) is equivalent to the following equations @xmath68 with @xmath69 satisfying @xmath70+\frac{\dot{t}^5_{\ 5}}{3}.\end{aligned}\ ] ] using the definition for @xmath71 in eq .
( [ t1 ] ) , the latter equation up to order @xmath46 can be written as @xmath72+\frac{\dot{t}^5_{\ 5}}{3}.\end{aligned}\ ] ] eq .
( [ fri3 ] ) is the modified friedmann equation describing cosmological evolution on the brane .
the auxiliary field @xmath69 incorporates non - trivial contributions of dark energy which differ from the standard matter fields confined to the brane .
the bulk matter contributes to the energy content of the brane through the bulk pressure terms @xmath38 that appear in the right hand side of the friedmann equation .
in addition , the bulk matter contributes to the energy conservation equation ( [ t1 ] ) through @xmath73 which is responsible for the energy exchange between the brane and bulk .
the functions @xmath38 and @xmath74 are functions of time corresponding to their values on the brane .
the energy - momentum conservation @xmath75 can not fully determine @xmath38 and @xmath74 and a particular model of the bulk matter is required @xcite . in the limit @xmath45 , eqs .
( [ fri3 ] ) and ( [ chi ] ) reduce to ( after replacement @xmath76 ) @xmath77 if we invoke the usual definitation @xmath78 , @xmath79 and @xmath80 , we get ( @xmath14 ) @xmath81 which is noting , but the general set of the equations in rs ii braneworld model with bulk matter content plus brane - bulk energy exchange ( see for example @xcite ) . returning to the general friedman equation ( [ fri3 ] ) with gauss - bonnet correction term , we can show that this equation has the solution for @xmath82 of the form @xmath83^{1/2}. \end{aligned}\ ] ] the upper solution ( + ) has correct @xmath45 limit .
indeed , if we expand this solution verses @xmath46 we get ( up to o(@xmath46 ) ) @xmath84,\end{aligned}\ ] ] and eqs .
( [ t1 ] ) and ( [ chi ] ) become @xmath85,\label{t3}\\ & & \dot { \chi}+ 4\,h\left(\chi-\frac{t^5_{\ 5}}{6}\right)=\frac{4}{36}\,{t^0_{\ 5}}\,\rho\ , \left [ 1 - 8\,\alpha\ , \left ( { \frac { { \rho}^{2}}{108}}+\frac{\lambda}{6}-\frac{t^{5}_{\ 5}}{3}+\chi \right ) \right]+\frac{\dot{t}^{5}_{\ 5}}{3}.\label{chi2}\end{aligned}\ ] ] therefore , until now we have obtained the set of equations describing the dynamics of our universe ( eqs . ( [ fri4])-([chi2 ] ) ) in the general gauss - bonnet braneworld with both bulk matter content and bulk - brane energy exchange provided that the gauss - bonnet coupling constant @xmath46 is chosen sufficiently small .
it is worth noting that although @xmath46 is small , it has a dramatic effect on the dynamic behavior of the cosmic evolution . besides the appearance of the @xmath86 term on the right hand side of eq .
( [ fri4 ] ) shows that in high energy scale the gauss - bonnet correction term plays an important role .
( bold line ) and @xmath87 ( dashed line ) versus @xmath88 for @xmath89 and @xmath90.,title="fig:",width=491 ] + versus @xmath88 for @xmath91 .
@xmath92 ( bold line ) , @xmath93 ( continuous line ) , and @xmath94 ( dashed line).,title="fig:",width=491 ] + versus @xmath88 for @xmath95 .
@xmath92 ( bold line ) , @xmath93 ( continuous line ) , and @xmath94 ( dashed line).,title="fig:",width=491 ] + versus @xmath88 for @xmath96 .
@xmath92 ( bold line ) , @xmath93 ( continuous line ) , and @xmath94 ( dashed lined).,title="fig:",width=491 ] + versus @xmath88 for @xmath96 .
@xmath92 ( bold line ) , @xmath93 ( continuous line ) , and @xmath94 ( dashed line).,title="fig:",width=491 ] + for @xmath95 .
@xmath97 ( bold line ) , @xmath98 ( continuous line ) , and @xmath99 ( dashed line).,title="fig:",width=491 ] + for @xmath100 .
@xmath97 ( bold line ) , @xmath98 ( continuous line ) , and @xmath99 ( dashed line).,title="fig:",width=491 ] + for @xmath101 .
@xmath102 ( bold line ) , @xmath103 ( continuous line ) , and @xmath104 ( dashed line ) . , title="fig:",width=491 ] + for @xmath105 .
@xmath102 ( bold line ) , @xmath103 ( continuous line ) , and @xmath104 ( dashed line).,title="fig:",width=491 ] +
in this section we are going to explore some cosmological consequences of our model . to do this , first we separate back the matter energy density and the brane tension as usual form with the replacement @xmath106 . therefore eqs .
( [ fri4 ] ) and ( [ chi2 ] ) become @xmath107,\end{aligned}\ ] ] @xmath108+\frac{\dot{t}^{5}_{\ 5}}{3}.\end{aligned}\ ] ] we are interested in the scenarios where the energy density of the brane is much lower than the brane tension , namely @xmath109 . assuming the randall - sundrum fine - tuning @xmath110 holds on the brane and defining the parameter @xmath111 , eqs .
( [ fri5 ] ) and ( [ chi3 ] ) can be simplified in the following form @xmath112,\end{aligned}\ ] ] @xmath113+\frac{\dot{t}^{5}_{\ 5}}{3}.\end{aligned}\ ] ] now , one may adopt several strategies to find solutions of eqs .
( [ t3 ] ) , ( [ fri6 ] ) and ( [ chi4 ] ) .
for example , one may take a suitable ansatz for the time dependent functions @xmath114 and @xmath115 and using eq .
( [ chi4 ] ) to find the function @xmath69 .
then substitute @xmath69 , @xmath114 and @xmath115 into eq .
( [ t3 ] ) one can try to obtain @xmath35 , and finally one may find hubble parameter @xmath82 through eq .
( [ fri6 ] ) . in the following we are interested in the case in which the energy momentum conservation law on the brane holds , which is usually assumed in the braneworld scenarios .
indeed , we want to consider the effect of the bulk content on the evolution of the universe without brane - bulk energy exchange , therefore we set @xmath116 .
the case with brane - bulk energy exchange in the general gauss - bonnet braneworld will be addressed elsewhere .
it was argued that the energy exchange between the bulk and brane @xmath114 will lead to the effective dark energy equation of state crossing @xmath2 [ 16,17 ] .
here we will show that without the energy exchange , the effect of @xmath117 and the combined @xmath117 and the gauss - bonnet correction have the same role . inserting the condition @xmath116 in eq .
( [ t3 ] ) , it reduces to @xmath118 .
this equation has well known solution @xmath119 , where @xmath120 is the present matter density of the universe and we have omitted the o " subscript from the scale factor on the brane for simplicity . then , consider a general ansatz @xmath121 for the bulk pressure @xcite , where @xmath122 and @xmath123 are two arbitrary constants , one can easily check that eq .
( [ chi4 ] ) has a solution of the form @xmath124 where @xmath125 is a constant usually referred to as dark radiation term and @xmath126 . finally , inserting @xmath35 and @xmath69 into eq .
( [ fri6 ] ) , we can rewrite it in the standard form @xmath127 where @xmath128 is the @xmath129-dimensional newtonian constant and @xmath130 represents the effective dark energy density on the brane @xmath131,\ ] ] where @xmath132 .
the equation of state parameter of the effective dark energy on the brane can be defined by @xcite @xmath133 where @xmath134 accounts for terms in the friedmann equation except the brane matter with equation of state @xmath135 .
now , if we use the redshift parameter @xmath136 as our variable , we can easily show that @xmath137 { \nonumber}\\ & & \times \bigg{\ { } c \left ( 1+z \right ) ^{4}+a \left ( 1+z \right ) ^{-\nu } -\frac{\alpha}{18}\,\left[2\ , \left ( a \left ( 1+z \right ) ^{-\nu}+c \left ( 1+z \right ) ^{4 } \right ) \right .
{ \nonumber}\\ & & \left .
\times \left ( { \sigma}^{2}+18\,a \left ( 1+z \right ) ^{-\nu}+18\,c \left ( 1+z \right ) ^{4 } \right ) -{\frac { { \sigma}^{4}}{108 } } \ , \right]\bigg{\}}^{-1}.\end{aligned}\ ] ] the corresponding late time deceleration parameter can be written @xmath138,\ ] ] where @xmath139 is all part of the matter on the braneworld and we take its present value as @xmath140 . in the rest of the paper , we will obtain constraints on the parameters such as @xmath125 , @xmath141 , @xmath123 , @xmath46 and @xmath17 in our model . indeed , we want to show that under what parameter space constraints our model can describe the accelerated expansion of the universe with the equation of state of the effective dark energy @xmath142 crossing @xmath2 , as suggested by observations .
let us begin with the special case , in which the gauss - bonnet coupling constant @xmath46 is equal to zero .
in this case we have the usual randall - sundrum ii braneworld model and eq .
( [ w ] ) reduces to @xmath144 therefore , we are left with three parameters @xmath125 , @xmath141 , @xmath123 , and two of them are independent . requiring that at the present moment @xmath145 and @xmath0 crossed @xmath2 around @xmath4 as indicated by extensive analysis of observational data @xcite , we can obtain @xmath146 for these value of parameters and @xmath147 , from eq .
( [ q ] ) we have @xmath148 and in addition @xmath87 crosses @xmath149 around @xmath150 which is in good agreement with recent observational data @xcite@xcite .
in figure [ figurers ] we plot @xmath151 and @xmath87 for the above value of the parameters versus redshift parameter @xmath88 .
next , we consider the general gauss - bonnet braneworld with bulk matter content . in this case
we have five parameters only four of which are independent .
considering that the value of @xmath17 does not affect the general profile of our model and further according to the randall - sundrum fine - tuning relation it should be small , we first fix @xmath153 .
thus , we have now four parameters and among them three are independent .
numerical calculations show that the functions @xmath154 and @xmath155 are well behaved for @xmath156 , provided that @xmath157 . employing the present value of the equation of state parameter of dark energy @xmath158 and the moment it crossed @xmath2 , namely @xmath159
, we get @xmath160 if we impose the condition @xmath47 which was used in deriving our equations , we can get constraint on the free parameter @xmath141 .
in numerical calculations we find that for @xmath91 we should have @xmath161 , while for @xmath96 we should have @xmath162 to satisfy the condition on @xmath46 . in figures [ figure1 ] and [ figure2 ] we plot @xmath151 and @xmath87 for @xmath91 versus redshift parameter @xmath88 for different value of the gauss - bonnet coupling constant @xmath46 . from these figures
we observe that at large @xmath88 , the @xmath151 increases with the increase of @xmath46 , while @xmath87 decreases with the increase of @xmath46 .
this qualitative behavior is quite opposite when @xmath96 as one can see from figures [ figure3 ] and [ figure4 ] .
finally we plot in figures [ figure5]-[figure8 ] the parameter space for the functions @xmath163 and @xmath164 .
we find that in the case @xmath91 , @xmath46 and @xmath125 increase with the increase of @xmath123 while , in contrast , for @xmath96 , @xmath46 and @xmath125 decrease with the increase of @xmath123 .
in this work we have generalized the randall - sundrum ii braneworld with both bulk matter content and bulk - brane energy exchange by adding the gauss - bonnet curvature correction term in the bulk action .
we have investigated the effects of the bulk content in the general gauss - bonnet braneworld on the evolution of the universe and found that although the effect of the gauss - bonnet correction term in the late time universe is small , it still plays an important role in the universe evolution .
in contrast to the previous models ( @xcite ) , in our study we kept the energy momentum conservation law on the brane as usual and found that the combination of the dark radiation term and the matter content of the bulk can play the role of the dark energy on the brane and influence the evolution of the universe . by suitably choosing parameter space in our model , we can describe the super - acceleration of our universe with the behavior of the effective dark energy equation of state in agreement with observations . in @xcite , it was argued in a gauss - bonnet brane world with induced gravity that the gauss - bonnet term and the mass parameter in the bulk play a crucial role in the evolution of the universe . here in our general model
, we confirmed their argument .
it is easy to see from eqs.(25)-(27 ) that the gauss - bonnet correction influences the dynamics of our universe , especially in the early universe with high energy scale .
phenomenon on the gauss - bonnet role has been disclosed in figs.2 - 5 .
we observed that although the gauss - bonnet effect is not clear at the present moment , it influences the universe evolution in the past and was more important in the earlier period . in this work
we just restricted our numerical fitting to limited observational data . giving the wide range of cosmological data available , in the future we expect to further constrain our model parameter space and test the viability of our model .
riess , et al .
j. * 116 * ( 1998 ) 1009 ; s. perlmutter , et al .
, astrophys .
j. 517 ( 1999 ) 565 ; s. perlmutter , et al .
, astrophys .
j. 598 ( 2003 ) 102 ; p.
de bernardis , et al . ,
nature 404 ( 2000 ) 955 .
t. padmanabhan , phys . rep . * 380 * , ( 2003 ) 235 ; p. j. e.peebles , b. ratra , rev .
phys . * 75 * , ( 2003 ) 559 ; v. sahni , astro - ph/0403324 and references therein .
edmund j. copeland , m. sami and shinji tsujikawa , hep - th/060305 .
s. weinberg , rev .
61 * , ( 1989 ) 1 ; n. straumann , astro - ph/0203330 ; t. padmanabhan , hep - th/0406060 .
r. r. caldwell , r. dave and p. j. steinhardt , phys .
lett . * 80 * , 1582 ( 1998 ) ; p. j. e. peebles and a. vilenkin , phys .
d * 59 * , 063505 ( 1999 ) ; p. j. steinhardt , l. m. wang and i. zlatev , phys .
d * 59 * , 123504 ( 1999 ) ; m. doran and j. jaeckel , phys .
d * 66 * , 043519 ( 2002 ) ; a. r. liddle , p. parson and j. d. barrow , phys .
d * 50 * , 7222 ( 1994 ) .
r. r. calwell , phys .
b * 545 * , 23 ( 2002 ) ; r. r. caldwell , m. kamionkowski and n. n. weinberg , phys .
91 * , 071301 ( 2003 ) ; j. m. cline , s. y. jeon and g. d. moore , phys .
d * 70 * , 043543 ( 2004 ) .
u. alam , v. sahni and a. starobinsky , jcap * 0406 * ( 2004 ) 008 [ astro - ph/0403687 ] ; y.g .
gong , class .
* 22 * ( 2005 ) 2121 [ astro - ph/0405446 ] ; y. wang and m. tegmark , phys . rev .
* d71 * ( 2005 ) 103513 [ astro - ph/0501351 ] ; y. wang and p. mukherjee , astrophys .
j. * 606 * ( 2004 ) 654 [ astro - ph/0312192 ] ; r. daly and s. djorgovski , astrophys .
j. * 612 * ( 2004 ) 652 [ astro - ph/0403664 ] ; u. alam , v. sahni , t. saini and a. starobinsky , mon . not .
soc . * 354 * , ( 2004 ) 275 [ astro - ph/0311364 ] ; t. choudhury and t. padmanabhan , astron . astrophys .
* 429 * , ( 2005 ) 807 [ astro - ph/0311622 ] .
g. chen and b. ratra , astrophys .
j. * 612 * , ( 2004 ) l1
. b. feng , x. l.wang and x. m. zhang , phys .
b 607 ( 2005 ) 35 ; w. hu , phys . rev .
d * 71 * , ( 2005 ) 047301 ; z.k .
guo , y.s .
piao , x.m .
zhang and y.z .
zhang , phys .
b * 608 * , ( 2005 ) 177 ; x.f .
zhang , h.li , y.s .
piao and x. zhang , astro - ph/0501652 .
li , b. feng and x.m .
zhang , hep - ph/0503268 .
b. wang , y. gong , e. abdalla , phys .
b * 624 * , ( 2006 ) 141 ; b. wang , c. y. lin , e. abdalla , phys .
b * 637 * , ( 2006 ) 357 ; b. wang , j.d .
zang , c.y .
lin , e. abdalla and s. micheletti , astro - ph/0607126 .
s. nojiri and s.d .
odintsov , phys .
b * 562 * , ( 2003 ) 147 ; s. nojiri and s. d. odintsov , phys .
d * 70 * ( 2004 ) 103522 ; a. vikman , phys
d * 71 * , ( 2005 ) 023515 ; a. anisimov , e. babichev and a. vikman , jcap * 0506 * , ( 2005 ) 006 ; s. nojiri , s.d .
odintsov and s. tsujikawa , phys .
d * 71 * ( 2005 ) 063004 ; p. singh , gr - qc/0502086 ; r.g .
cai , h.s .
zhang and a.z .
wang , hep - th/0505186 ; h. stefancic , astro - ph/0504518 ; e. o. kahya and v. k. onemli , gr - qc/0612026 ; tomi koivisto , david f. mota , phys . lett .
b * 644 * , 104 ( 2007 ) , astro - ph/0606078 ; tomi koivisto , david f. mota , phys .
d * 75 * 023518 ( 2007 ) , hep - th/0609155 .
e. kiritsis , g. kofinas , n. tetradis , t. n. tomaras and v. zarikas , jhep * 0302 * ( 2003 ) 035 ; e. kiritsis , n. tetradis and t. n. tomaras , jhep * 0203 * ( 2002 ) 019 ; p. s. apostolopoulos and n. tetradis , phys .
d * 71 * 043506 ( 2005 ) ; p. s. apostolopoulos and n. tetradis , phys .
b * 633 * 409 ( 2006 ) ; e. kiritsis , jcap * 0510 * 014 ( 2005 ) ; k. i. umezu , k. ichiki , t. kajino , g. j. mathews , r. nakamura and m. yahiro , phys .
d * 73 * 063527 ( 2006 ) .
g. kofinas , g. panotopoulos and t.n .
tomaras , hep - th/0510207 .
cai , y. gong and b. wang , jcap * 0603 * , ( 2006 ) 006 ; p. s. apostolopoulos and n. tetradis , hep - th/0604014 . c. bogdanos and k. tamvakis , hep - th/0609100 ; c. bogdanos , a. dimitridis , k. tamvakis , hep - th/0611094 .
b. zwiebach , phys .
b * 156 * , 315 ( 1985 ) ; d. g. boulware and s. deser , phys . rev
. lett . * 55 * , 2656 ( 1985 ) .
d. lovelock , j. math . phys .
* 12 * , 498 ( 1971 ) .
x r. brown , r. maartens , e. papantonopoulos , v. zamarias , gr - qc/0508116 .
cai , h.s .
zhang and a. wang , hep - th/0505186 .
g. kofinas , r. maartens , e. papantonopoulos , jhep * 0310 * , 066 ( 2003 ) hep - th/0307138 .
cardone , a. troisi , s. capozziello , phys .
d , * 69 * , 083517 , 2004 ; | we investigate the effect of the bulk content in the general gauss - bonnet braneworld on the evolution of the universe .
we find that the gauss - bonnet term and the combination of the dark radiation and the matter content of the bulk play a crucial role in the universe evolution .
we show that our model can describe the super - acceleration of our universe with the equation of state of the effective dark energy in agreement with observations . |
in recent years , three dimensional ( 3d ) displays using computer - generated holograms ( cghs ) have been well studied , since cghs can faithfully reconstruct the light waves of 3d images @xcite .
such 3d display is referred as to `` electroholography '' .
electroholography can reconstruct 3d images by displaying cghs generated by diffraction calculations on a spatial light modulator ( slm ) such as amplitude or phase - modulated lcd panels . in electroholography ,
several important issues must be considered : calculation cost for cgh , narrow viewing angle , small reconstructed images and color reconstruction . as color reconstruction is an important issues ,
many color reconstruction methods have been studied .
for instance , color reconstruction methods using three slms to display three cghs corresponding to the red , green and blue components of a 3d image have been proposed @xcite .
color reconstruction methods using a single slm have also been proposed : space - division @xcite and depth - division @xcite methods .
another single slm method is the time - division method , which temporally switches rgb lights by synchronizing signals @xcite .
regardless of the color reconstruction methods as mentioned above , we need to compute three diffraction calculations corresponding to the rgb components of a color image .
we developed a calculation reduction method for color cghs using color space conversion .
color cghs have been calculated on rgb space . in this paper
, we calculate color cghs in other color spaces : for example , ycbcr color space , which is widely used in digital image processing like jpeg and mpeg formats . in ycbcr color space , a rgb image is converted to the luminance component ( y ) , blue - difference chroma ( cb ) and red - difference chroma ( cr ) components . in terms of the human eye , although the negligible difference of the luminance component is well - recognized , the difference of the other components is not . in this method ,
the luminance component is normal sampled and the chroma components are down - sampled .
the down - sampling allows us to accelerate the calculation of the color cghs .
we calculate diffraction from the components , and then convert the diffracted result in ycbcr color space to rgb color space . in section 2 ,
we explain the calculation reduction method for color cgh using color space conversion . in section 3
, we verify the proposed method on computer simulation .
section 4 concludes this work .
for simple notation we introduce diffraction operator @xmath0}}$ ] , where @xmath1 is the propagation distance , @xmath2 is the wavelength , @xmath3 is a source plane and @xmath4 is the position vector .
for example , the angular spectrum method @xcite , which is a major diffraction calculation method , is expressed by the diffraction operator as follows : @xmath5 } } = \int { \mathcal{f } \biggl[u_1(\bm x_1 ) \biggl ] } \exp \biggl ( 2 \pi i z \sqrt{\frac{1}{\lambda^2 } - |\bf f|^2 } ~ \biggr ) \exp(2\pi i { \bf f x_2 } ) d { \bm f } \nonumber \\ & = & { \mathcal{f}^{-1 } \biggl [ { \mathcal{f } \biggl[u_1(\bm x_1 ) \biggl ] } \exp \biggl ( 2 \pi i z \sqrt{\frac{1}{\lambda^2 } - |\bf f|^2 } ~ \biggr ) \biggl ] } , \label{eqn : fre}\end{aligned}\ ] ] where @xmath6 is the position vector in fourier domain , @xmath7 is a source plane , @xmath8 is a destination plane , @xmath9}$ ] is fourier transform , and @xmath1 is the propagation distance between the source and destination planes .
if we calculate color cghs in rgb color space , we need to calculate three diffracted fields @xmath10 and @xmath11 from the corresponding components of the rgb images , @xmath12 and @xmath13 as follows : @xmath14 } } \label{eqn : rgb_cgh1},\\ u_g(\bm x_2)={{{\rm d}^{z}_{\lambda_g } [ g(\bm x_1 ) ] } } \label{eqn : rgb_cgh2},\\ u_b(\bm x_2)={{{\rm d}^{z}_{\lambda_b } [ b(\bm x_1 ) ] } } \label{eqn : rgb_cgh3},\end{aligned}\ ] ] where @xmath15 and @xmath16 are the wavelengths of rgb lights , respectively . in this paper
, we call this the `` direct color cgh calculation '' .
if there are spatial light modulators that are capable of displaying complex amplitudes @xcite , we can reconstruct a clear rgb image without direct and conjugated lights from eqs.([eqn : rgb_cgh1])-([eqn : rgb_cgh3 ] ) by the three slm method @xcite , time - division method @xcite and space - division method @xcite and so forth . when using the amplitude or phase - modulated slms , we need to convert the complex amplitudes of eqs.([eqn : rgb_cgh1])-([eqn : rgb_cgh3 ] ) to the amplitude cghs by taking the real part or phase - only cgh by taking the argument . from here ,
let us consider the calculation of color cghs in the ycbcr color space .
converting a rgb image to a ycbcr image is as follows , @xmath17 where @xmath18 , @xmath19 and @xmath20 are the luminance , blue - difference chroma and red - difference chroma , respectively .
we convert the ycbcr image to the rgb image by , @xmath21 from eq.([eqn : y_r ] ) , the diffracted fields of rgb images passing through the ycbcr color space are expressed as , @xmath22}}=b_{11 } { { { \rm d}^{z}_{\lambda_r } [ y(\bm x_1 ) ] } } + b_{12 } { { { \rm d}^{z}_{\lambda_r } [ c_b(\bm x_1 ) ] } } + b_{13 } { { { \rm d}^{z}_{\lambda_r } [ c_r(\bm x_1 ) ] } } \label{eqn : diff_y_r1 } , \\
{ { { \rm d}^{z}_{\lambda_g } [ g(\bm x_1)]}}=b_{21 } { { { \rm d}^{z}_{\lambda_g } [ y(\bm x_1 ) ] } } + b_{22 } { { { \rm d}^{z}_{\lambda_g } [ c_b(\bm x_1 ) ] } } + b_{23 } { { { \rm d}^{z}_{\lambda_g } [ c_r(\bm x_1 ) ] } } \label{eqn : diff_y_r2 } , \\ { { { \rm d}^{z}_{\lambda_b } [ r(\bm x_1)]}}=b_{31 } { { { \rm d}^{z}_{\lambda_b } [ y(\bm x_1 ) ] } } + b_{32 } { { { \rm d}^{z}_{\lambda_b } [ c_b(\bm x_1 ) ] } } + b_{33 } { { { \rm d}^{z}_{\lambda_b } [ c_r(\bm x_1 ) ] } } \label{eqn : diff_y_r3}.\end{aligned}\ ] ] even though the direct color cgh calculation of eqs.([eqn : rgb_cgh1])-([eqn : rgb_cgh3 ] ) requires only three diffraction calculations , we need to calculate nine diffraction calculations in eqs .
( [ eqn : diff_y_r1])-([eqn : diff_y_r3 ] ) .
to improve this problem we use the following important relation of the diffraction operator : @xmath23 } } = { { { \rm d}^{z\lambda_2/\lambda_1}_{\lambda_2 } [ u(\bm x ) ] } } .
\label{eqn : ope}\ ] ] applying the relation of @xmath24 } } = { { { \rm d}^{z}_{\lambda_r } [ u(\bm x)]}}$ ] to eq.([eqn : diff_y_r2 ] ) , the following equation is derived , @xmath25}}&= & b_{21 } { { { \rm d}^{z \lambda_g/\lambda_r}_{\lambda_g } [ y(\bm x_1 ) ] } } + b_{22 } { { { \rm d}^{z \lambda_g/\lambda_r}_{\lambda_g } [ c_b(\bm x_1 ) ] } } + b_{23 } { { { \rm d}^{z \lambda_g/\lambda_r}_{\lambda_g } [ c_r(\bm x_1 ) ] } } , \nonumber \\ & = & b_{21 } { { { \rm d}^{z}_{\lambda_r } [ y(\bm x_1 ) ] } } + b_{22 } { { { \rm d}^{z}_{\lambda_r } [ c_b(\bm x_1 ) ] } } + b_{23 } { { { \rm d}^{z}_{\lambda_r } [ c_r(\bm x_1)]}}. \label{eqn : diff_y_4}\end{aligned}\ ] ] likewise , applying the relation of @xmath26 } } = { { { \rm d}^{z}_{\lambda_r } [ u(\bm x)]}}$ ] to eq.([eqn : diff_y_r3 ] ) , the following equation is derived , @xmath27}}&= & b_{31 } { { { \rm d}^{z \lambda_b/\lambda_r}_{\lambda_b } [ y(\bm x_1 ) ] } } + b_{32 } { { { \rm d}^{z \lambda_b/\lambda_r}_{\lambda_b } [ c_b(\bm x_1 ) ] } } + b_{33 } { { { \rm d}^{z \lambda_b/\lambda_r}_{\lambda_b } [ c_r(\bm x_1 ) ] } } , \nonumber \\ & = & b_{31 } { { { \rm d}^{z}_{\lambda_r } [ y(\bm x_1 ) ] } } + b_{32 } { { { \rm d}^{z}_{\lambda_r } [ c_b(\bm x_1 ) ] } } + b_{33 } { { { \rm d}^{z}_{\lambda_r } [ c_r(\bm x_1)]}}. \label{eqn : diff_y_5}\end{aligned}\ ] ] finally , we can obtain a color cgh passing through the ycbcr color space in matrix notation as follows : @xmath28 } } \\ { { { \rm d}^{z\lambda_g/\lambda_r}_{\lambda_g } [ g(\bm x_1 ) ] } } \\ { { { \rm d}^{z\lambda_b/\lambda_r}_{\lambda_b } [ b(\bm x_1 ) ] } } \end{pmatrix}= \begin{pmatrix } b_{11 } & b_{12 } & b_{13 } \\
b_{21 } & b_{22 } & b_{23 } \\
b_{31 } & b_{32 } & b_{33 } \\
\end{pmatrix } \begin{pmatrix } { { { \rm d}^{z}_{\lambda_r } [ y(\bm x_1 ) ] } } \\ { { { \rm d}^{z}_{\lambda_r } [ c_b({\bm x_1 } ) ] } } \\ { { { \rm d}^{z}_{\lambda_r } [ c_r({\bm x_1 } ) ] } } \end{pmatrix}. \label{eqn : diff_y_r_mat}\ ] ] as you can see , this conversation includes only three diffraction operators .
unfortunately , the propagation distance of the diffracted results of green and blue components change from @xmath1 to @xmath29 and @xmath30 .
when placing cghs corresponding to each component at same the location , the reconstructed rgb images from the cghs are reconstructed out of position .
we can compensate for the out of location by placing the green and blue cghs at @xmath31 and @xmath32 , respectively . since eq.([eqn : diff_y_r_mat ] ) has three diffraction operators , the calculation cost is the same as the direct calculation of the direct color cgh calculation by eqs.([eqn : rgb_cgh1])-([eqn : rgb_cgh3 ] ) . in terms of the human eye , although the negligible difference of the luminance component y is well - recognized , the difference of the other components cb and cr is not . using this property , the luminance component y is normal sampled and the chroma components are down - sampled .
the down - sampling allows us to accelerate the calculation of the color cghs .
note that we need to change the sampling pitches that are required for the diffraction operator of down - sampled @xmath33 and @xmath33 because the areas of @xmath33 and @xmath33 are smaller than @xmath34 .
for example , when down - sampling 1/4 @xmath33 and @xmath33 , we change the sampling pitches of them to 4@xmath35 where @xmath35 is the sampling pitch of @xmath34 .
let us summarize the color cgh calculation passing through the ycbcr color space .
first , we convert a rgb image to a ycbcr image using eq.([eqn : r_y ] ) .
second , we down - sample the cb and cr components .
third , we compute diffraction calculations from the components , then , we up - sample @xmath36 } } $ ] and @xmath37}}$ ] .
next , we calculate the rgb cghs using eq.([eqn : diff_y_r_mat ] ) .
last , we reconstruct the color images by placing the green and blue cghs at @xmath31 and @xmath32 , respectively .
) -([eqn : rgb_cgh3 ] ) ) and the proposed method ( eq.([eqn : diff_y_r_mat ] ) ) , respectively , width=566 ] we verify the proposed method on computer simulation .
we use three rgb color images . for the calculation condition , the propagation distance @xmath38 cm ,
the sampling pitch is 10@xmath39 , the wavelengths of rgb lights are 633 nm , 532 nm and 450 nm , respectively .
the resolution of all of the images is @xmath40 pixels .
we assume that we use a complex amplitude slm @xcite .
we use the nearest - neighbor interpolation for the down - sampling and up - sampling in the color space conversion .
the nearest - neighbor interpolation is simple interpolation , so that the interpolation is faster than other interpolation methods such as linear and cubic interpolations .
figure [ fig : img1 ] shows the reconstructed images of fruits from cghs , which are calculated by direct color cgh calculation ( eqs.([eqn : rgb_cgh1])-([eqn : rgb_cgh3 ] ) ) and the proposed method ( eq.([eqn : diff_y_r_mat ] ) ) , respectively .
the notations of `` r : g : b=1:1:1 '' , `` r : g : b=4:1:1 '' mean a color image is not down - sampled and is down - sampled by 1/4 of the green and blue components , respectively .
likewise , the notation of `` y : cb : cr=4:1:1 '' mean that a color image is down - sampled by 1/4 of cb and cr components , respectively .
the insects show magnified images of a part of the reconstructed images . the reconstructed images of `` r : g : b=4:1:1 '' and `` r : g : b=8:1:1 '' are blurred because the green and blue components affect the luminance of the image .
in contrast , the reconstructed images of `` y : cb : cr=4:1:1 '' and `` y : cb : cr=8:1:1 '' maintain the sharpness of the texture because cb and cr components do not seriously affect the luminance however , the brightness is little decreased , compared with other reconstructed images . )
-([eqn : rgb_cgh3 ] ) ) and the proposed method ( eq.([eqn : diff_y_r_mat ] ) ) , respectively , width=566 ] ) -([eqn : rgb_cgh3 ] ) ) and the proposed method ( eq.([eqn : diff_y_r_mat ] ) ) , respectively , width=566 ] figures [ fig : img2 ] and [ fig : img3 ] show the reconstructed images of mandrill and `` tifanny '' from cghs , which are calculated by direct color cgh calculation ( eqs.([eqn : rgb_cgh1])-([eqn : rgb_cgh3 ] ) ) and the proposed method ( eq.([eqn : diff_y_r_mat ] ) ) , respectively .
the reconstructed images of `` r : g : b=4:1:1 '' and `` r : g : b=8:1:1 '' are blurred .
in contrast , the reconstructed images of `` y : cb : cr=4:1:1 '' and `` y : cb : cr=8:1:1 '' maintain the sharpness of the texture .
the calculation times of `` r : g : b=1:1:1 '' , `` r : g : b=4:1:1 '' and `` y : cb : cr=4:1:1 '' are about 660 ms , 270 ms and 396 ms , respectively .
the calculation times of `` r : g : b=8:1:1 '' and `` y : cb : cr=8:1:1 '' are about 260 ms and 400 ms , respectively .
therefore , the proposed method can maintain the sharpness of the reconstructed image and accelerate the calculation speed .
we proposed a calculation reduction method for color cghs using color space conversion . the proposed method succeeded in maintaining the sharpness of the reconstructed image and accelerating the calculation speed .
the proposed method will be acceptable not only for ycbcr color space but also other color spaces such as yiq , yuv and so forth .
in addition , the proposed method will be useful for color digital holography to accelerate the numerical reconstruction .
the proposed method applies only to a pure amplitude image or weak phase image , so in future study we will attempt to develop a calculation reduction method for color cghs with full complex amplitude .
this work is supported by japan society for the promotion of science ( jsps ) kakenhi ( grant - in - aid for scientific research ( c ) 25330125 ) 2013 , and kakenhi ( grant - in - aid for scientific research ( a ) 25240015 ) 2013 .
99 h. yoshikawa , t. yamaguchi , and r. kitayama , `` real - time generation of full color image hologram with compact distance look - up table , '' osa topical meeting on digital holography and three - dimensional imaging 2009 , dwc4 ( 2009 ) .
h. nakayama , n. takada , y. ichihashi , s. awazu , t. shimobaba , n. masuda and t. ito , `` real - time color electroholography using multiple graphics processing units and multiple high - definition liquid - crystal display panels , '' appl .
opt . * 49 * , 59935996 ( 2010 ) .
t. ito , t. shimobaba , h. godo , and m. horiuchi , `` holographic reconstruction with a 10-@xmath41 m pixel - pitch reflective liquid - crystal display by use of a light - emitting diode reference light , '' opt . lett . * 27 * , 14061408 ( 2002 ) .
m. makowski , m. sypek , i. ducin , a. fajst , a. siemion , j. suszek , and a. kolodziejczyk , `` experimental evaluation of a full - color compact lensless holographic display , '' opt .
express * 17 * , 2084020846 ( 2009 ) .
t. shimobaba , a. shiraki , y. ichihashi , n. masuda and t. ito,``interactive color electroholography using the fpga technology and time division switching method , '' ieice electron .
express * 5 * , 271277 ( 2008 ) .
m. oikawa , t. shimobaba , t. yoda , h. nakayama , a. shiraki , n. masuda , and t. ito,``time - division color electroholography using one - chip rgb led and synchronizing controller , '' opt .
express * 19 * 1200812013 ( 2011 ) . | we report a calculation reduction method for color computer - generated holograms ( cghs ) using color space conversion .
color cghs are generally calculated on rgb space . in this paper
, we calculate color cghs in other color spaces : for example , ycbcr color space . in ycbcr color space , a rgb image is converted to the luminance component ( y ) , blue - difference chroma ( cb ) and red - difference chroma ( cr ) components . in terms of the human eye , although the negligible difference of the luminance component is well - recognized , the difference of the other components is not . in this method ,
the luminance component is normal sampled and the chroma components are down - sampled .
the down - sampling allows us to accelerate the calculation of the color cghs .
we compute diffraction calculations from the components , and then we convert the diffracted results in ycbcr color space to rgb color space . _
keywords _ : computer - generated hologram , digital holography , color holography , color space conversion , ycbcr color space |
one of the leading methods for growing semiconductor quantum dots is via the controlled coarsening of a film of a material that is strained with respect to the substrate on which it is grown.@xcite this ( `` self - assembled '' ) coarsening / roughening is a result of lattice - mismatch - induced strains .
the dots are often capped by the substrate material , thus extending the strain around the dot to all angular directions .
not surprisingly , the interpretation of the electronic structure of such dots is profoundly affected by their strain profile .
thus , in order to calculate or interpret the measured electronic structure , one has first to calculate or measure the position dependent strain tensor @xmath2 .
the three basic approaches to calculating such strains are : _ ( i ) harmonic continuum elasticity _ : here , one uses classical elasticity@xcite within the harmonic approximation . for a cubic system ,
the strain energy per atom , @xmath3 , is @xmath4 where @xmath5 is the equilibrium volume , @xmath6 are cubic elastic constants and @xmath2 is the strain tensor .
we illustrate the predictions of harmonic continuum elasticity for a 2d film , since this is going to be used as a test case . in the absence of shear strain ( @xmath7 ) , for a film coherently grown on a substrate with parallel lattice constant @xmath8 , the strain components are @xmath9 where @xmath10 is the equilibrium lattice constant of the unstrained material and @xmath11 is the perpendicular lattice constant of the strained film .
the equilibrium value of this @xmath12axis is determined from @xmath13 , yielding , @xmath14\;\epsilon_{||}(a_s ) , \label { econst}\ ] ] where the `` epitaxial strain reduction factor '' for orientation * g * of the @xmath12axis is @xmath15 and @xmath16 is the elastic anisotropy , @xmath17 is the bulk modulus and @xmath18 is a purely geometric factor given in ref.[@xcite ] . for principal directions , @xmath19 and @xmath20
. equations ( [ st])([rf ] ) are used routinely to predict tetragonal distortions of strained films @xcite .
the harmonic continuum elasticity method has been recently applied to pyramidal quantum dots by grundmann _
et al._@xcite and by pryor _
et al._@xcite _ ( ii ) atomistic elasticity _ : here , one avoids a continuum description and describes the strain _ energy _ in terms of few - body potentials between actual atoms @xmath21 where @xmath22 is a two - body term , @xmath23 is a three - body function of the bond angle , @xmath24 .
the functional form of these terms is taken to be strain - independent .
the strain is determined by minimizing @xmath25 with respect to atomic positions @xmath26 .
like the continuum elasticity approach , only the cubic elastic constants are used as input@xcite . however , unlike the ce approach , here ( a ) optical phonon modes can be described,@xcite ( b ) harmonicity is not assumed , and ( c ) the atomic level symmetry is retained . the last point is illustrated in fig . 1 that shows a regular pyramid representing the quantum dots from experiments of grundmann _ et al . _
@xcite in a continuum representation , the strain is equal on the @xmath27 and @xmath28 facets , while in an atomistic description these two facets can have different strain if the pyramid is made of a zincblend material .
atomistic elasticity has been widely used to determine strain in alloys,@xcite superlattices@xcite and dots,@xcite where @xmath22 and @xmath23 of the eq .
( 5 ) are taken from keating s valence force field @xcite ( vff ) model . _
( iii ) atomistic quantum - mechanical approach _ : here one does not have to assume any model for inter - atomic interactions as in the atomic elasticity .
instead , one explicitly computes the total electron and nuclear energy @xmath29 $ ] for each atomic configuration @xmath30 directly from a quantum - mechanical schrdinger equation .
atomic symmetry is retained and harmonicity is not assumed .
this approach has been used for _ small _ ( @xmath31 atom ) wires and clusters @xcite but it impractical for @xmath32 dots ( @xmath33 atoms ) .
the three approaches to the calculation of strain harmonic continuum elasticity , ( anharmonic ) atomistic elasticity , and the atomistic quantum mechanical approached have been recently compared for inas / gaas strained superlattices.@xcite however , no comparison exists for 0-dimensional quantum dots .
here we perform parallel calculations for the strain @xmath34 of a pyramidal inas dots ( fig .
1 ) surrounded by gaas using the two approaches that are practical for large dots : continuum elasticity and atomistic elasticity .
we find that : ( i ) the strain profiles obtained via continuum elasticity are in qualitative agreements with those found by atomistic elasticity ; ( ii ) the atomistic elasticity produces different strains on the two facets ( @xmath27 or @xmath35 ) of the zincblend pyramidal dots ( see fig . 1 ) , corresponding to the physical c@xmath36 symmetry , while the continuum elasticity approximates this as c@xmath37 symmetric strain .
( iii ) the quantitative discrepancy resides mostly inside the dots , while the difference in the barrier region is smaller ; these differences are traced back to the fact that the strain lies outside the domain of validity for the linear elasticity .
we illustrate this point by contrasting the predicted @xmath38 ratio of coherent 2d films , as obtained by harmonic continuum elasticity [ eq .
( 3 ) ] and atomistic elasticity .
differences are noticeable already for 1 % biaxial strain , whereas the controlled - coarsening ( `` self - assembled '' ) growth method for quantum dots needs to deal with larger mismatches ( 7 % for inas / gaas and inp / gap ) .
finally , the consequences on the electronic structure of the different strain profiles obtained for dots using ce and ae are illustrated .
in the ce approximation the strain is determined by minimizing the elastic energy given in eq . 1 . to account for the lattice mismatch we assume the coordinates
are fixed to the barrier material , and treat the island as expanded barrier material ( with different elastic constants ) .
this is accomplished by the modification @xmath39 where @xmath40 and @xmath41 are the unstrained lattice constants for the barrier and island material respectively , and @xmath42 and @xmath43 are the elastic constants for the island material .
a piece of island material with no external forces acting on it will have its energy minimum shifted to @xmath44 this fictitious strain corresponds to unstrained island material and must be subtracted .
the corrected strain is still computed with derivatives in the barrier s coordinates and must be converted to the the island coordinates through multiplication by @xmath45 .
thus , the physical strain is given by @xmath46 where @xmath47 is the strain computed directly from the displacement @xmath48 which minimizes @xmath3 .
a numerical solution requires some kind of discretization .
we define the displacements @xmath48 on a cubic grid , thereby maintaining the cubic symmetry of the crystal .
the strain is expressed in terms of forwards or backwards differences by @xmath49 where @xmath50 is the lattice vector in the @xmath51 direction .
symmetric differences @xmath52 are undesirable since they give unphysical low energy configurations which oscillate with period @xmath53 .
for example , a displacement @xmath54 has @xmath55 when constructed using symmetric differences .
the oscillatory solutions can not be simply discarded since they mix with the physical ones .
non - symmetric derivatives are also problematic since a particular choice will single out a direction in space .
the solution is to average @xmath3 over all permutation of @xmath56 on each of the three difference operators .
that is , we take @xmath57 .
physically this corresponds to taking the energy density at each site to be the average of the energy densities from each adjoining octant .
the elastic energy is a quadratic function of the displacements , which is easily minimized using the conjugate gradient algorithm . for the barrier material
the strains are computed directly using differences ( now there is no impediment to using symmetric differences ) . in the island material
we then apply the correction in eq .
( [ correction ] ) . in the vff model ,
the strain energy is expressed as a functional of atomic positions , @xmath58 , as @xmath59 ^ 2\nonumber\\ & & + { 1\over 2}\sum_i \sum_{j , k > j}^{nn } { 3\beta_{i , jk}\over 8d_{ij}^0d_{ij}^0 } [ ( \br_j-\br_i)\cdot(\br_k-\br_i)-\cos{\theta_0}d_{ij}^0d_{ij}^0]^2 .
\label { vffe}\end{aligned}\ ] ] here , @xmath60 denotes the ideal bond length between atoms @xmath51 and @xmath61 , and @xmath62 is the ideal bond angle . for the zincblend structure , @xmath63 .
the local - environment - dependent coefficients , @xmath64 and @xmath65 , are fitted to the elastic constants of bulk materials@xcite .
the long - range coulomb interactions of ref .
[ @xcite ] are neglected which causes a slight deviation from the measured bulk properties @xcite . in this case , the elastic constants of a pure bulk zincblend material are given as @xmath66 where @xmath67 is interatomic bond length .
because eq . ( [ econst ] ) contains only two free parameters , it is impossible to fit three arbitrary elastic constants . nonetheless , for zincblend materials @xmath68 and @xmath69 may be chosen so the @xmath70 s fit within a few percent of the measured values .
table i gives the elastic constants of bulk gaas and inas calculated from eq .
( [ econst ] ) using @xmath68 s and @xmath69 s of ref .
[ @xcite ] .
the elastic constants obtained differ a bit from the experimental values ( since coulomb corrections to eq .
( [ econst ] ) are neglected ) , but we will use them consistently in both our continuum elasticity and atomistic elasticity studies . for purposes of comparison ,
table i also contains the experimental elastic parameters .
note that the vff method with the standard parameterization of eq.(16 ) reproduces well the values and trends in the formation enthalpies of strained gap / inp structures , as obtained from first - principles@xcite .
the relaxed atomic configuration is obtained by conjugate gradient minimization @xcite of @xmath25 with respect to the atomic positions . at each minimization step , the atoms are displaced along the conjugate direction @xmath71 by a finite increment @xmath72 , as @xmath73 .
a line minimization of @xmath25 along the conjugate gradient direction to find @xmath72 that minimizes @xmath25 is done by taking advantage of the fact that @xmath25 is a fourth - order polynomial that depends on only the relative positions , @xmath74 , of each atom : @xmath75 & = & e_{ae}[\{\br_i-\br_j\}]\nonumber\\ & & + \lambda e^{(1)}[\{\br_i-\br_j\},\{\bh_i-\bh_j\ } ] + \lambda^2 e^{(2)}[\{\br_i-\br_j\},\{\bh_i-\bh_j\}]\nonumber\\ & & + \lambda^3 e^{(3)}[\{\br_i-\br_j\},\{\bh_i-\bh_j\ } ] + \lambda^4 e^{(4)}[\{\br_i-\br_j\},\{\bh_i-\bh_j\ } ] .
\label{ecg1}\end{aligned}\ ] ] the energy terms , @xmath76 and @xmath77 , are also simple polynomials of @xmath78 and @xmath79 . the increment @xmath72 minimizing the elastic energy is then obtained by solving exactly @xmath80\over \partial \lambda } = e^{(1)}+2\lambda e^{(2)}+3\lambda^2 e^{(3 ) } + 4\lambda^3 e^4 = 0 . \label{ecg2}\ ] ] figure 2 illustrates how the local strain is calculated . after the atomic positions are relaxed by minimizing @xmath25 , the local strain tensor @xmath81 at a cation site ( for cation - mixed systems ) is calculated by considering a tetrahedron formed by four nearest neighboring anions .
the distorted tetrahedron edges , @xmath82 and @xmath83 , are related to the ideal tetrahedron edges @xmath84 and @xmath85 via @xmath86 the ideal tetrahedron edges are @xmath87a/2,[0\bar11]a/2,[\bar110]a/2\}$ ] , where @xmath88 denotes the equilibrium lattice constant of the cation , i.e. , @xmath89 for ga atoms and @xmath90 for in atoms .
the local strain , @xmath81 , is then calculated by a matrix inversion as @xmath91 where @xmath92 is the unit matrix .
figure 3 shows @xmath93 and @xmath94 as obtained by continuum elasticity ( dashed lines ) and by atomistic elasticity ( solid lines ) as a function of the position from the pyramidal center along the @xmath95 $ ] direction at a height @xmath96 from the base ( see fig . 1 ) .
the corresponding differences in strains,@xmath97 , are given as the solid lines in figure 4 .
we note that the grid points of the continuum elasticity calculation are chosen to be commensurate with the cation positions of the ideal gaas zincblend structure for consistent comparisons of the two approaches .
the largest differences occur around the interfaces between the dot and the cap .
a significant discrepancy is also found inside the quantum dot where the inas experience large compressive strains : @xmath98 of the continuum elasticity is found to be more compressive than that of the atomistic elasticity , while the @xmath99 of the ce is more tensile .
a similar comparison is given in fig .
5 , but this time the position vector is along the @xmath100 $ ] direction , starting from the substrate , going through the wetting layer into the pyramidal tip and then into the capping layer . again
, the discrepancy is largest around the interfaces , while the strains in the barrier ( gaas substrate and capping layer ) agree within 0.5% .
figure 4 illustrates the extent to which the continuum elasticity description misses the correct atomic symmetry . in a pyramid made of zincblend materials on the @xmath101 substrates , the @xmath27 and @xmath35 facets are symmetrically inequivalent ( fig . 1 ) .
indeed , the atomistic calculation produces different strains .
the dashed lines in fig .
4 show the difference @xmath102)-\eps_{ij}^{ae}([\bar110])$ ] for these two directions .
we see that the anisotropy is pronounced at the interfaces . for the atomistic elasticity calculation
, we construct the pyramidal structure to have an in atom at the pyramidal tip . this tip in atom
has : ( i ) two - as atoms that belong to the inas dot and lie along the @xmath95 $ ] direction ; and ( ii ) the other two - as atoms that belong to the gaas capping layer and lie along the @xmath103 $ ] direction . considering only the local strain of the tip atom , one expects larger compressive @xmath98 and @xmath104 along the @xmath95 $ ] direction than along the @xmath103 $ ] direction , based on the atomic configuration . by the same token , the atoms at the @xmath27 interfaces experience larger compressive @xmath98 and @xmath104 , than those at the @xmath35 interfaces , for this particular choice of the pyramidal geometry . in the continuum description , @xmath105)=\eps_{ij}^{ce}([\bar110])$ ] and this effect is missing .
we know that the continuum and atomistic models , starting from the same input elastic constants , must agree in the limit of small strain and a large system . to study the rate at which the two methods diverge with increasing strain , we consider the simple case of biaxial strain . as eqs .
( [ st])([rf ] ) show , for a 2-dimensional film that is constrained on a @xmath101 substrate , continuum elasticity predicts @xmath106_{ce } = -{2c_{12}\over c_{11}}. \label{biaxa}\ ] ] figure 6(a ) compares this result with that obtained via atomistic elasticity , as a function of the relative film / substrate mismatch @xmath107 .
similarly , for the @xmath108 strain , @xmath109 and the corresponding comparison of the continuum and atomistic elasticity is shown in fig .
we see that the discrepancy rises linearly , reaching 4 % for a lattice mismatch of 7 % , characteristic of inas / gaas .
this difference is comparable to that found between ce and ae around the interfaces of the quantum dots ( figs .
thus , the discrepancy simply reflects the departure from the linearity regime of the continuum elasticity .
the existence of different strain magnitudes and even symmetries in a continuum elasticity _ vs. _ atomistic elasticity descriptions can affect the ensuing electronic structure of the quantum dot . most notably , the real point group symmetry of the square pyramid is @xmath0 , but continuum elasticity spuriously produces a higher @xmath1 symmetry .
regarding the quantitative effects , there are different levels of approximation for coupling the strain to the electronic structure .
the most general and accurate electronic structure approach is atomistic ( _ e.g. _ , pseudopotentials , tight - binding ) .
there , the full set of atomic positions affects the electronic structure . in more approximate electronic structure approaches , such as continuum effective - mass ,
only some aspects of the full , position - dependent - strain tensor , @xmath110 is `` felt '' by the electronic structure . in these approaches
, one considers strain - modified potential wells as barriers .
since experiments typically measure electronic energies rather than strains , it is instructive to examine these effects . assuming decoupled conduction and valence bands the strain - modified confinement potential of the conduction - band state is @xmath111 . \label{elcp}\ ] ] here , @xmath112 is the energy of the conduction - band minimum of the bulk material at and @xmath113 , the deformation potential of the conduction - band under hydrostatic deformation .
the `` strain '' hamiltonian of the valence states @xcite is @xmath114\nonumber\\ & & -\sqrt{3}d \left[\left(\begin{array}{c}0\\-1\\0\end{array } \begin{array}{c}-1\\0\\0\end{array } \begin{array}{c}0\\0\\0\end{array}\right)\eps_{xy } + \left(\begin{array}{c}0\\0\\0\end{array } \begin{array}{c}0\\0\\-1\end{array } \begin{array}{c}0\\-1\\0\end{array}\right)\eps_{yz } + \left(\begin{array}{c}0\\0\\-1\end{array } \begin{array}{c}0\\0\\0\end{array } \begin{array}{c}-1\\0\\0\end{array}\right)\eps_{zx } \right ] , \label{hh}\end{aligned}\ ] ] where @xmath115 is the hydrostatic deformation of the valence states and @xmath116 and @xmath117 are uniaxial deformation potentials for @xmath101 strain and @xmath118 strain , respectively .
the effective confinement potentials of the valence states are obtained by diagonalizing the strain hamiltonian coupled with the spin - orbit hamiltonian @xcite . along the @xmath119 $ ] direction ( @xmath120axis ) through the pyramidal tip , the shear strains ( off - diagonal terms of the strain tensor ) are zero and @xmath121 , and thus the effective confinement potentials can be simplified as @xmath122 \nonumber\\ e_{lh } & = & e_v^0 + a_v { \rm tr}(\eps)-{1\over 6}[\delta^{so}+\delta^s(\eps ) ] + { 1\over 2}\sqrt{(\delta^{so}+\delta^s(\eps))^2- { 8\over 3}\delta^{so}\delta^s(\eps ) } , \label{holecp}\end{aligned}\ ] ] where @xmath123 is the spin - orbit splitting and @xmath124 $ ] .
figure 7 shows the effective confinement potentials of the conduction and valence - band states along the @xmath120axis through the tip of the pyramid .
the strain profiles obtained by the continuum elasticity and atomistic elasticity are used for the calculation with the same material parameters given in table ii @xcite .
again , the largest difference in confinement potentials is found at the interfaces at about 100 mev for the conduction band and 200mev for the valence band .
the average difference of the confinement potentials inside the dot is about 20 mev for the conduction - band state .
although the differences in the strain - modified - confinement potentials are small , the band edge states are expected to show different characteristics depending upon which strain profile is used for the electronic structure calculation .
we compare the strain distribution of the pyramidal inas dot grown on a gaas substrate calculated using continuum elasticity and atomistic elasticity . we find a significant difference in the strain around the dot interfaces and inside the dot , while the difference in the barrier ( gaas substrate and capping layer ) is very small .
the difference between the two results is attributed to the large strain outside the linearity regime of ce , and to the loss of the correct atomic symmetry by the ce .
.ideal bond lengths ( @xmath125 ) , elastic constants , and force constants ( @xmath68 and @xmath69 ) of bulk gaas and inas .
elastic constants of valence force field method are evaluated by eq .
( 8) using @xmath68 and @xmath69 given below . [ cols="<,^,^,^,^,^,^ " , ] | the electronic structure of interfaces between lattice - mismatched semiconductor is sensitive to the strain .
we compare two approaches for calculating such inhomogeneous strain
continuum elasticity ( ce , treated as a finite difference problem ) and atomistic elasticity ( ae ) . while for _ small _ strain the two methods must agree , for the large strains that exist between lattice - mismatched iii - v semiconductors ( e.g. , 7 % for inas / gaas outside the linearity regime of ce ) there are discrepancies .
we compare the strain profile obtained by both approaches ( including the approximation of the correct @xmath0 symmetry by the @xmath1 symmetry in the ce method ) , when applied to @xmath0-symmetric inas pyramidal dots capped by gaas .
# 1 , # 2 , # 3 , 1#4#5#6 _ # 1 _ * # 2 * , # 3 ( 1#4#5#6 ) |
how do dust grains , known to permeate disks surrounding young stars , assemble into planets ?
one proposed stage of growth involves gravitational instability of a dust - rich layer at the disk midplane ( safronov 1969 ; goldreich & ward 1973 ) .
if dust grains are free to settle vertically out of gas , midplane dust densities eventually exceed roche densities , and dust particles can begin to aggregate by self - gravity .
an objection to this means of forming planetesimals is that vertical velocities of gas may be too high for dust to settle ( weidenschilling 1980 ) . even apart from turbulence intrinsic to gas ( e.g. ,
turbulence driven by the magneto - rotational instability ) , dust layers that are too vertically thin can suffer from the kelvin - helmholtz instability ( khi ) .
the khi threatens to manifest because dust - rich gas at the midplane rotates at a rate different from that of dust - poor gas at higher altitude .
dust - poor gas experiences greater acceleration from a background radial pressure gradient , and so its rotation velocity in centrifugal balance must deviate more strongly from the keplerian value .
the deviation is smaller for gas laden with dust , since dust adds inertia but contributes no pressure .
usually it is assumed that the background radial pressure gradient @xmath3 so that the vertical shear in the rotation velocity @xmath4 . whether the disk is kh - unstable is commonly assessed using the richardson number ( chandrasekhar 1961 , page 491 ) : @xmath5 where @xmath6 is the vertical gravitational acceleration and @xmath7 is the total mass density . in the limit that perturbations
are incompressible , the numerator is the square of the brunt - visl frequency of buoyant oscillations , while the denominator is the square of the vertical shearing rate .
the richardson number measures the amount of work required to overturn fluid elements that are originally in hydrostatic equilibrium , against the amount of free kinetic energy available in the background shear .
for purely cartesian flow ( no rotational forces ) , @xmath8 ( miles 1961 ; howard 1961 ; chandrasekhar 1961 ; drazin & reid 1981 ; li et al .
despite the fact that ( [ onequarter ] ) is formally not a sufficient criterion , laboratory experiments bear out its usefulness ( tritton 1988 ) . in the case of our disk , close to the midplane , @xmath9 arises from the tidal gravity of the star , @xmath10 measures distance from the midplane , @xmath11 is the keplerian angular frequency , and @xmath12 , where @xmath6 and @xmath13 denote gas and dust , respectively .
moreover , @xmath14 within the thin dust layers of interest to us ( see [ init ] ) .
thus throughout this paper @xmath15 as dust settles vertically , @xmath16 increases ; the dust surface density @xmath17 is assumed constant while the characteristic scale height @xmath18 of dust decreases . in the limit that @xmath19 , @xmath20 .
criterion ( [ onequarter ] ) for instability therefore places a lower bound on @xmath18 and a corresponding upper bound on @xmath21 .
unfortunately , for conventional solar nebular parameters in particular for a height - integrated solar metallicity of @xmath22the maximum of @xmath21 falls short of the roche density by 12 orders of magnitude ( e.g. , sekiya 1998 ; youdin & shu 2002 ) .
but is it appropriate to apply criterion ( [ onequarter ] ) to a circumstellar disk where the flow is not cartesian ?
one might guess that rotational forces merely modify @xmath23 by a factor of order unity , since the coriolis force and the radial shear ( @xmath24 ) operate on a timescale @xmath25the same timescale characterizing vertical brunt - visl oscillations and the vertical shear at @xmath26 .
yet when coriolis forces are introduced , as in the analysis of gmez & ostriker ( 2005 , hereafter go ) , the stability properties of the flow change dramatically . numerical simulations by go , performed under the assumption that dust is perfectly entrained in gas , reveal no well - defined threshold for instability .
unstable modes are detected for @xmath0 as high as 4 , with growth rates that diminish with increasing @xmath0 but that show no sign of vanishing ( see their figures 9 and 10 ) .
similar results obtain in numerical simulations by johansen , henning , & klahr ( 2006 , hereafter jhk ) , who relax the assumption of perfect coupling between dust and gas but who , like go , retain only coriolis forces and ignore radial shear . for their `` rocks '' and `` pebbles '' with momentum stopping times 1050 @xmath27 shorter than @xmath25
, the dust distribution evolves to one where @xmath0 increases from @xmath21 at the midplane to @xmath28 at higher altitude ( jhk , their figures 4 and 5 ) .
here we conduct numerical simulations that account for the full complement of disk - related forces .
we include not only the coriolis force but also the centrifugal force and radial tidal gravity . the latter two forces combine to produce radial shear in rotational equilibrium .
thus we investigate the stability of doubly shearing flows : vertically shearing dust layers in near - keplerian differential rotation .
we expect our results to differ from those of go and jhk .
the modes emphasized by go have growth rates @xmath29 at @xmath30 .
keplerian differential rotation , characterized by a strain rate of @xmath31 , should shear such modes apart before they have time to amplify .
indeed this is the conclusion of ishitsu & sekiya ( 2003 , hereafter is ) , who numerically integrate the linearized equations of motion , including the full suite of disk - related forces , to compute the factors by which modes amplify .
they find that amplification factors are sufficiently limited by radial shear that midplane dust - to - gas density ratios might reach values as high as @xmath22 ( still too low for gravitational instability , unfortunately ) .
they restrict , however , their linear analysis to odd - parity modes for which the vertical velocity @xmath32 .
gmez & ostriker ( 2005 ) establish that even - parity modes grow faster and overturn the dust layer more effectively .
our simulations do not pre - select for either type of mode .
another difference between our work and is is that we concentrate on @xmath33 flows , whereas is employ a simple , analytically tractable dust distribution . of course , @xmath34 is merely a plausible condition to which the dust - gas mixture might relax .
one of the goals of this study is to assess whether the richardson number is a good predictor of stability even when all disk - related forces are accounted for . finally , the analysis of is is linear , while our numerical simulations enable access to nonlinear phenomena .
our numerical simulations , like those of go , assume that dust particles are perfectly entrained in gas , i.e. , we assume that grain momentum stopping times @xmath35 .
this approximation is valid for small particles , e.g. , having sizes @xmath36 m at disk radius @xmath37 au in a minimum - mass nebula ( e.g. , weidenschilling 1977 ) .
the hydrodynamics code and initial conditions are described in [ method ] .
results are presented in [ res ] ; there we determine @xmath23 for our doubly shearing flows .
a summary is provided in
[ sum ] .
while our entire study is rooted in the literature on the kelvin - helmholtz instability and its impact on planetesimal formation , and as such revolves around the richardson criterion , it was brought to our attention by the referee that the khi may not be the only instability afflicting the dusty midplane .
baroclinic instabilities , studied in disks by cabot ( 1984 ) , knobloch & spruit ( 1986 ) , and arlt & urpin ( 2004 ) may also be relevant .
these authors study baroclinity specifically in the context of a non - zero vertical shear .
more generally , a baroclinic flow is one whose isobaric surfaces do not coincide with its isodensity surfaces .
our flows are baroclinic because of molecular weight gradients : dust adds to the density but not to the pressure .
baroclinic instability can strike even when the richardson number is large .
we will connect our findings to the baroclinic instability when appropriate ; indeed , in
[ cori ] we propose that the instability discovered by go in their simulations of non - radially - shearing disks is , in fact , the baroclinic instability discovered by cabot ( 1984 ) . and we will see in [ varsig ] that the richardson criterion alone does not determine stability under arbitrary conditions .
gas and pressureless dust are two fluids obeying separate momentum equations . in an inertial frame , @xmath38 where @xmath39 is velocity , @xmath40 is time , and @xmath41 is the background potential .
terms containing @xmath42 , the momentum stopping time for dust in gas , account for how the two fluids interact frictionally . in this paper , we work exclusively in the tight coupling limit where particles and gas have negligible relative velocity , @xmath43 .
this corresponds to the case where dust particles are so small i.e .
, their surface area to mass ratios are so large that they become entrained in the gas flow over short times . by computing the difference and appropriately weighted sum of ( [ one ] ) and ( [ two ] ) ( see youdin & goodman 2005 ) , and taking the limit @xmath44 , we find @xmath45 where @xmath46 .
furthermore , @xmath47 where @xmath48 is the gravitational constant , @xmath49 is the central stellar mass at the origin , @xmath50 is the internal energy density of gas , and @xmath51 and @xmath10 are the cylindrical radius and height .
the gas obeys a polytropic equation of state @xmath52 with constant @xmath53 and @xmath54 .
finally , dust and gas obey separate continuity relations @xmath55 equations ( [ four])([nine ] ) are the ones we solve in this paper , in the shearing box approximation ( goldreich & lynden - bell 1965 ; hawley , gammie , & balbus 1995 ; [ init ] ) .
they contain the same content as equations ( 2)(6 ) of go . in the tightly coupled limit ,
a fluid parcel does not change its dust - to - gas ratio in a lagrangian sense : @xmath56 , where @xmath57 is the convective derivative .
particles can not slip away from gas , and so we can not expect our simulated flows to relax , from arbitrary initial conditions , into a unique , steady state of marginal stability ( if such a state actually exists ) . in reality , such relaxation occurs over the time it takes dust to settle vertically relative to gas .
vertical settling times are measured in myrs for micron - sized particles ( or @xmath2@xmath58 yr for centimeter - sized particles ) in a minimum - mass nebula .
nevertheless , our equations do permit dust - rich parcels to settle toward the midplane if their weight can not be supported in vertical hydrostatic equilibrium .
we are able to evolve dust - gas mixtures over dynamical times , measured in orbital periods , provided @xmath42 is still shorter .
our simulations can thus determine whether a given set of equilibrium conditions is dynamically stable .
we now specify these equilibrium initial conditions .
call @xmath59 and @xmath60 the radius and azimuth of a test particle moving on a circular orbit , where @xmath61 .
shifting to axes centered on the test particle and rotating at @xmath62 , we trade the usual cylindrical coordinates @xmath63 for their shearing sheet counterparts @xmath64 : @xmath65 , @xmath66 , @xmath67 , @xmath68 , and @xmath69 . in this rotating frame , the momentum equation ( [ four ] ) reads [ abc ] @xmath70 where @xmath71 , @xmath72 , @xmath73 , and curvature terms of order @xmath74 are dropped , following the shearing sheet approximation .
we introduce the variable @xmath75 in ( [ a ] ) because in our simulations we experiment with @xmath76 ( zero radial shear ) to connect to go . for the full problem with a point - mass potential , @xmath77 . in rotational equilibrium ,
equation ( [ a ] ) for radial momentum balance gives @xmath78 initially , @xmath79 reflects how the background nebular pressure changes radially on scales of @xmath51 .
we express @xmath80 in terms of a model input parameter , @xmath81 : @xmath82 the velocity @xmath81 represents the difference between the keplerian rotation rate and the sub - keplerian rotation rate of pressure - supported gas ; it measures the maximum possible difference in rotational velocity between the dust - rich midplane and dust - free gas at higher altitude . for typical nebular parameters ,
it is approximately @xmath83 m / s and nearly constant with @xmath51 . further defining the local dust - to - gas ratio @xmath84 , we rewrite ( [ shear ] ) as @xmath85 in each of our simulations , we hold @xmath81 fixed for simplicity . then the vertical shear , @xmath86 , is non - zero only when @xmath87 is non - zero . because @xmath88 varies rapidly with @xmath10 for the vertically thin dust layers of interest to us , approximating @xmath81 as constant introduces negligible error .
all our simulations investigate the stability of constant @xmath0 flows .
the condition @xmath89 yields @xmath90 : @xmath91 which integrates to @xmath92^{1/2 } - 1 \label{mu}\ ] ] where @xmath93 is a model input parameter and @xmath94 is a characteristic dust height . equation ( [ mu ] ) implies that dust extends to a maximum height @xmath95 finally , vertical hydrostatic equilibrium gives @xmath96 : @xmath97 which , using ( [ mu ] ) , can be solved analytically for @xmath98 .
the expressions , which differ for @xmath99 and @xmath100 , are not especially illuminating and so we omit them here . self - gravity is ignored .
its neglect is justified for densities less than the roche density , @xmath101 .
equivalently , for standard nebular parameters , we are restricted to @xmath102 .
sekiya ( 1998 ; see also youdin & shu 2002 ) relaxes this condition and accounts for vertical self - gravity in computing the richardson number .
figure [ dens_1 ] shows profiles for @xmath103 and @xmath104 , for normalized ( code ) parameters of @xmath105 , @xmath106 , @xmath107 , @xmath108 , @xmath109 , @xmath110 , and @xmath111 , where @xmath112 is the sound speed at the midplane .
these parameter choices imply a height - integrated dust - to - gas surface density ratio ( height - integrated metallicity ) of @xmath113 , the nominal solar value .
we vary @xmath0 and @xmath114 from simulation to simulation .
note that our choice of @xmath115 is a factor of 2 smaller than that of go because we prefer to use gas temperatures characterizing passive disks ( see , e.g. , chiang et al .
2001 ) rather than the hotter temperatures of the hayashi ( 1981 ) nebula .
disk gas must be fairly passive , i.e. , not heated by turbulent dissipation , for dust to settle into the thin layers of interest here ( see appendix b of youdin & chiang 2004 for estimates of the degree of turbulence permitted ) .
a factor of 2 decrease in @xmath116 produces a factor of 2 increase in the central concentration @xmath114 , at fixed @xmath0 and fixed @xmath117 .
figure [ dens_0.125 ] is analogous to figure 1 , except that @xmath118 and @xmath119 ( so that again @xmath120 ) .
figure [ dens_sigma05 ] is analogous to figure 2 , except that @xmath121 , so that @xmath122 ( the height - integrated metallicity is super - solar by a factor of 5 ) . to summarize this subsection ,
our equilibrium initial conditions specify @xmath123 through ( [ vy ] ) , @xmath124 through ( [ mu ] ) , and @xmath98 through analytic solution of ( [ hydro ] ) .
initially , @xmath125 , and @xmath98 and @xmath21 are constant with @xmath126 and @xmath127 .
the primary input parameters are @xmath0 and @xmath114 ; we hold @xmath115 fixed for all runs .
we employ the hydrodynamics code zeus ( stone & norman 1992 ) , a version of which was kindly given to us by dr .
bryan johnson . because we are interested in flows that shear both in radius ( @xmath128 ) and height ( @xmath86 ) , the simulations are necessarily three - dimensional .
we adopt the usual shearing box ( e.g. , hawley , gammie , & balbus 1995 ) , with shearing periodic boundary conditions in @xmath126 , periodic boundary conditions in @xmath127 , and closed boundary conditions in @xmath10 .
three modifications are made to the code .
the first is the addition of another fluid , dust , which is transported according to its own continuity equation ( [ nine ] ) , but otherwise has the same velocity as that of gas .
the second modification is to the pressure gradient source term : @xmath129 , in accordance with ( [ abc ] ) .
the final change is to introduce an extra source term for the `` large - scale '' radial pressure gradient , distinct from the just mentioned `` local '' pressure gradient .
that is , equation ( [ a ] ) for the @xmath126-momentum is revised to read @xmath130 the extra source term @xmath80 is given by ( [ dpdr0 ] ) .
it accounts for how the pressure changes over radial lengthscales of @xmath51 , and is held fixed for each simulation .
its inclusion is necessary to have dust - rich gas rotate faster than dust - poor gas initially , i.e. , to drive the vertical shear .
but because our simulations use a shearing box in other words , because they are local we can not follow the evolution of this large - scale gradient .
a self - consistent treatment would require us to model the disk globally .
our use of an external , constant , radial pressure gradient is the same device employed by go .
the error introduced is negligible insofar as changes in pressure over lengthscales of @xmath51 occur over timescales much longer than the dynamical timescales which concern us here . in short
, the modified code solves equations ( [ eos])([nine ] ) , ( [ b ] ) , ( [ c ] ) , and ( [ amod ] ) .
initial conditions for the code are given above in [ init ] .
the equilibrium state is perturbed by adding a random velocity to each grid cell .
the perturbation velocity points in a random direction from cell to cell . for our main set of runs ,
the magnitude of the perturbation in each cell is drawn at random from a uniform distribution that extends from zero to @xmath131 ( for comparison , the maximum vertical shearing velocity is about @xmath132 ) .
several auxiliary runs use much smaller or more spatially limited initial perturbations ; these are described in
[ sivp ] and produce the same outcomes as our main simulations .
our standard box dimensions are @xmath133 and @xmath134 , and the corresponding number of grid cells is @xmath135 .
the duration of each simulation is at least @xmath136 .
run parameters are contained in tables 1 and 2 , and are justified in
[ res ] . including the von neumann & richtmyer ( 1950 ) artificial viscosity
makes little difference to our results not surprisingly since our flows are highly subsonic ( @xmath115 ) .
therefore the simulations reported in this paper have no artificially imposed viscosity apart from the unavoidable numerical kind .
to test our code , we repeat the experiments of go . the coriolis acceleration is retained but the radial shear is suppressed by setting @xmath137 in the code . effectively , this converts shearing boundary conditions in @xmath126 to periodic boundary conditions .
it also eliminates the contribution , @xmath138 , to @xmath139 from the centrifugal force and tidal gravity .
since the resultant flows have no structure in @xmath126 , we reduce @xmath140 to 2 and @xmath141 to @xmath142 .
table [ tab_cor ] lists the run parameters .
the main parameter varied from run to run is @xmath0 ; the central concentration @xmath114 is adjusted in tandem to keep @xmath22 fixed for all runs ( this restriction to solar metallicity is relaxed for the fully shearing runs of [ cor+shear ] ) .
following go , we diagnose the flow by fourier analyzing @xmath143 in the @xmath127-direction at fixed @xmath126 , @xmath10 , and @xmath40 .
we then inspect how the fourier amplitudes change with time .
we are interested in those modes whose @xmath127-wavelengths @xmath144 are comparable to the dust layer thickness @xmath145 , as shorter wavelength modes can not overturn the layer and longer wavelength modes grow more slowly .
we are able to measure the exponential growth rates @xmath146 of several modes whose @xmath144 s range from @xmath28/15 to @xmath28/3 of @xmath145 .
imply that the only modes present are those for which the simulation box fits an integral number of wavelengths .
the standard box length is @xmath147 .
] figure [ f4 ] displays the fourier amplitudes versus time for the @xmath148 mode , in simulations of varying @xmath0 .
while the growth rate @xmath146 of this mode decreases with increasing @xmath0 , growth is still reliably detected for @xmath0 as high as 16 , a value @xmath149 greater than the canonical threshold of 1/4 .
our measured growth rates are similar to those reported by go ( see their figure 10 ) . runs c1a
c1f test the sensitivity of our results to simulation box size and spatial resolution .
doubling the box height @xmath150 to @xmath151 at fixed resolution produces negligible change in measured growth rates ( c1a vs. c1 ) . doubling the vertical resolution by doubling @xmath152 at fixed @xmath150 increases mode growth rates by @xmath210% ( c1b vs. c1 ) .
changes in @xmath146 of order 10% are also produced by varying the azimuthal parameters @xmath153 and @xmath154 by factors of 2 ( c1c c1f ) .
we conclude that our standard box size @xmath155,@xmath156 and resolution @xmath157,@xmath158 adequately balance the need for accuracy with the need for computational speed .
table [ tab_cor ] and figure [ f4 ] suggest that when only the coriolis force is included , and radial shear is omitted , dust layers characterized by surprisingly large values for @xmath0 will eventually overturn and mix .
this conclusion is supported by figure [ f5 ] , which displays snapshots of a high - resolution @xmath159 run for which @xmath160 . despite @xmath0 being 16 @xmath27 greater than the traditional critical value of @xmath161 , the dust layer is clearly unstable .
the fastest growing mode has @xmath162 and overturns the layer by @xmath163 .
all these results agree with those of go .
we suspect that the instability discussed in this section , and simulated first by go , is a baroclinic instability afflicting @xmath76 disks .
cabot ( 1984 ) finds that baroclinic , @xmath76 disks are linearly unstable to non - axisymmetric perturbations ; that the `` baroclinic instability draws its energy from radial excursions by fluid elements '' ( radial excursions made possible by the coriolis force ) ; and that maximum growth rates @xmath146 are of order the maximum vertical shearing frequency @xmath164 ( see his table 1 and the discussion following his equation 22b ) . for our dusty layers , @xmath165 ;
support for the inverse square - root dependence on @xmath0 can indeed be found in figure [ f4 ] .
, the maximum vertical shear is @xmath166 , where @xmath167 is the hydrostatic disk thickness and @xmath168 is a number typically of order 0.1 .
the smallness of @xmath168 arises because two effects compete and nearly cancel : both the radial gravity @xmath169 and the radial pressure acceleration @xmath170 decrease in magnitude with increasing height . ]
knobloch & spruit ( 1986 ) extend the work of cabot ( 1984 ) by restoring the effects of keplerian shear to baroclinic disks .
they find that radial shear tends to stabilize the flow to non - axisymmetric disturbances . for their particular choice for the form of @xmath171 , the linear baroclinic instability is defeated when baroclinity is small ( @xmath172 ) and when the disk is strongly stable to thermal convection in the vertical direction .
these stability requirements are usually satisfied in disks without gradients in mean molecular weight ( such disks are their main concern ) .
knobloch & spruit ( 1986 ) speculate that instability occurs when baroclinity is large . in the context of their disks ,
large baroclinity demands variations in specific entropy on short radial length scales @xmath173 , the vertical height of disk gas .
rapid radial variations are also required for instability by arlt & urpin ( 2004 ) , who study axisymmetric perturbations .
but large baroclinity can also be achieved by having a strong molecular weight gradient in the vertical direction , as exists in a highly settled dust layer .
we turn now to simulating such highly baroclinic , radially shearing flows .
ccccccccccccc c1a@xmath174 & 2 & 0.25 & 64 & 8 & 64 & 8 & 0.25 & 0.903 & 0.01 & 20 & 8/3 & 0.334 + c1b & & & 64 & 8 & 64 & 4 & & & & & 8/3 & 0.371 + c1c & & & 64 & 16/3 & 32 & 4 & & & & & 8/3 & 0.352 + c1d & & & 32 & 8/3 & & & & & & & 8/3 & 0.368 + c1e & & & 32 & 3 & & & & & & & 3 & 0.339 + c1f & & & 64 & 7 & & & & & & & 7/3 & 0.336 + c1 & & & 64 & 8 & 32 & 4 & & & & 40 & 8/3 & 0.332 + & & & & & & & & & & & 8/5 & 0.303 + & & & & & & & & & & & 8/7 & 0.342 + c2 & & & & & & & 1 & 0.475 & & & 8/3 & na@xmath175 + & & & & & & & & & & & 8/5 & 0.165 + & & & & & & & & & & & 8/7 & 0.203 + c3 & & & & & & & 1.581 & 0.391 & & & 8/3 & na + & & & & & & & & & & & 8/5 & 0.148 + & & & & & & & & & & & 8/7 & 0.171 + c4 & & & & & & & 2.5 & 0.324 & & & 8/3 & na + & & & & & & & & & & & 8/5 & 0.088 + & & & & & & & & & & & 8/7 & 0.142 + c5 & & & & & & & 3.952 & 0.270 & & & 8/3 & na + & & & & & & & & & & & 8/5 & 0.084 + & & & & & & & & & & & 8/7 & 0.125 + c6 & & & & & & & 6.248 & 0.226 & & & 8/3 & na + & & & & & & & & & & & 8/5 & na + & & & & & & & & & & & 8/7 & 0.093 + c7 & & & & & & & 9.878 & 0.190 & & & 8/3 & na + & & & & & & & & & & & 8/5 & na + & & & & & & & & & & & 8/7 & 0.078 + c8 & & & & & & & 15.62 & 0.161 & & & 8/3 & na + & & & & & & & & & & & 8/5 & na + & & & & & & & & & & & 8/7 & 0.062 + [ tab_cor ] having found simulation parameters @xmath176 that produce realistic results for @xmath76 , we use those same parameters for our radially shearing @xmath77 simulations .
since the radial wavenumber of a disturbance grows at rate @xmath177 , our resolution in @xmath126 should be at least comparable to that in @xmath127 .
our standard run parameters are @xmath178 and @xmath179 .
table [ tab_shear ] lists the various experiments , all of which start with @xmath180 constant flow as described in [ init ] . whether the dust layer turns over within the run duration is indicated in the table .
ccccccccccccc s1 & 64 & 8 & 64 & 8 & 32 & 4 & 0.25 & 0.903 & 0.01 & @xmath181 & 40 & n@xmath174 + s2 & & & & & & & 0.125 & 1.31 & & @xmath181 & 100 & n + s2a & 128 & & 128 & & 64 & & & & & @xmath181 & 20 & n + s3 & 64 & 8 & 64 & 8 & 32 & 4 & 0.0625 & 1.989 & & @xmath181 & 50 & y + s3a & & & & & & & & & & @xmath182 & 70 & y + s4 & & & & & & & 0.03125 & 3.252 & &
@xmath181 & 20 & y + s4a & & & & & & & & & & @xmath182 & 40 & y + s5 & & & & & & & 0.25 & 3.57 & 0.03 & @xmath181 & 20 & n + s6 & & & & & 48@xmath175 & & 0.125 & 6.65 & & @xmath181 & 20 & n + s7 & & & & & 64@xmath175 & & 0.0625 & 14.7 & & @xmath181 & 20 & y + s8@xmath183 & & & & & 64@xmath175 & & 0.125 & 23.3 & 0.05 & @xmath181 & 20 & n + s9 & & & & & 32 & & 0.125 & 0.174 & 0.001 & @xmath181 & 70 & n + s10 & & & & & & & 0.0625 & 0.226 & & @xmath181 & 70 & n + s11 & & & & & & & 0.03125 & 0.296 & & @xmath181 & 70 & n + s12 & & & & & & & 0.01563 & 0.393 & & @xmath181 & 70 & y + [ tab_shear ] results for @xmath77 differ dramatically from those for @xmath76 . whereas instability characterizes all values of @xmath0 for @xmath76
, we find that @xmath0 must fall below a critical value , @xmath184 , for the layer to overturn when @xmath77 and when @xmath120 ( see [ varsig ] for experiments that vary @xmath1 ) .
figures [ f6 ] and [ f7 ] demonstrate this point : the flow for @xmath185 eventually mixes whereas that for @xmath118 keeps the dust layer intact for as long as @xmath186 .
the same fourier analysis of [ cori ] , applied to the latter run , reveals no growth of any mode . repeating the @xmath118
run at higher resolution@xmath187confirms these results ( run s2a ) . naturally , reducing the initial velocity perturbations increases the time required for the layer to overturn .
the dependence is very slight ; lowering the maximum perturbation velocity from @xmath131 to @xmath188 increases the time to overturn by @xmath189@xmath190 ( runs s3a and s4a ) .
( nonetheless , eliminating the initial perturbations altogether produces no evolution in the shearing box whatsoever . )
the instability grows fastest at `` co - rotation , '' where @xmath191 , and spreads radially inward and outward from that location .
this behavior is evident in figure [ f8 ] , and we observe it in all our unstable runs . upon removing all initial velocity perturbations in a narrow annulus surrounding co - rotation , and leaving the perturbations in place everywhere else , we find that the flow outside co - rotation still overturns , though more slowly
. eventually the instability spreads to co - rotation and mixes dust uniformly throughout the entire box .
this experiment assures us that the instability is not an artifact of @xmath192 . at present
we attribute all this behavior to the fact that our fixed - amplitude initial velocity perturbations are fractionally larger near co - rotation ; i.e. , the initial perturbation @xmath193 is a greater fraction of @xmath123 near co - rotation than elsewhere .
it would therefore be natural to expect the instability to manifest itself most quickly at co - rotation . for solar or super - solar values of @xmath1940.05 ,
the value of @xmath23 changes little from 0.1 , as runs s5s8 of table [ tab_shear ] attest .
however , @xmath23 changes substantially for sub - solar metallicities it decreases to @xmath20.02 for @xmath195 ( runs s9s12 ) .
this result indicates that the richardson number alone is an inadequate predictor of instability under general circumstances .
still , because sub - solar metallicities seem less relevant for planet formation , we have not pursued the question of what should replace the richardson criterion , and we content ourselves with citing @xmath196 with the understanding that this result applies only for solar and moderately super - solar metallicities .
values of @xmath114 and @xmath1 at fixed @xmath197 are displayed in figure [ f9 ] . in a minimum - mass nebula
for which @xmath22 , @xmath198 corresponds to a toomre @xmath199 \approx 25 $ ] , independent of disk radius if @xmath200 , as is approximately the case in standard nebular models . under these conditions ,
our neglect of the dust layer s self - gravity is well justified . only when @xmath201 ( @xmath202 ; see run s8 ) does dust self - gravity become important ( sekiya 1998 ) .
we can not say whether the observed turnover of the dust layer in a fully shearing disk is better ascribed to the kelvin - helmholtz instability ( as studied by , e.g. , chandrasekhar 1961 ) or to the baroclinic instability ( as studied by cabot 1984 , knobloch & spruit 1986 , and arlt & urpin 2004 ) .
both instabilities rely on the vertical shear . for @xmath203 , the maximum vertical shearing frequency @xmath2@xmath204
exceeds , by factors of a few , both the maximum brunt - visl frequency @xmath2@xmath11 ( thereby satisfying the traditional criterion for the khi ) and the rotation frequency @xmath11 ( thereby satisfying the criterion for `` large baroclinity , '' as defined by knobloch & spruit 1986 ) .
an analytic criterion for the baroclinic instability , relevant for non - axisymmetric perturbations , that is valid near the midplane of a radially shearing disk and for arbitrary values of the vertical shear , is not known ( knobloch & spruit 1986 ) .
we have performed numerical simulations of flows near the dust - rich midplanes of protoplanetary disks .
dust particles are assumed small enough that they are perfectly entrained in gas , and disk self - gravity is neglected .
the midplane flow is doubly shearing ; the rotational velocity varies with height according to the vertical gradient in dust density , and it also varies with radius according to the usual keplerian shear .
the simulations are necessarily three - dimensional and performed in a shearing box . despite the complications introduced by rotation and tidal gravity
, the richardson criterion for the onset of the kelvin - helmholtz instability still proves useful for solar to moderately super - solar height - integrated metallicities .
dust layers characterized by constant richardson number and @xmath2050.05 overturn and mix if @xmath206 , but remain intact at larger @xmath0 .
this result contrasts with the situation when only the coriolis force is accounted for and the radial shear is suppressed : in that case , unstable modes persist for @xmath0 as large as 16 .
but because these `` coriolis - only '' ( and likely baroclinic ; cabot 1984 ) modes at large @xmath0 grow at rates that are substantially slower than the kepler strain rate of @xmath31 , they are stabilized and rendered impotent by the kepler shear . in hindsight ,
our guess ( [ intro ] ) that the critical @xmath0 dividing stability from instability might only change by a factor of order unity proved correct , though only for solar to moderately super - solar metallicities : we find that @xmath23 for our doubly shearing flows is about half that of the traditional value of 1/4 , if @xmath2050.05 .
this is sensible insofar as @xmath196 gives a vertical shearing frequency @xmath86 that is a few times faster than any of the other frequencies of the problem : the brunt - visl frequency for vertical oscillations ( this reaches a maximum of @xmath2@xmath11 at @xmath26 ) , the kepler strain rate ( @xmath31 ) , and the coriolis turning frequency ( @xmath207 ) .
deeper insights , including an understanding of why @xmath208 for @xmath209 , might be gained by study of the geophysical literature , where doubly shearing , rotating flows are commonplace ( see , e.g. , chapter 7 of pedlosky 1979 , which discusses baroclinic instabilities ) .
we conclude that for a standard , height - integrated , solar metallicity of @xmath194 , the dust density at the midplane can be at most @xmath2@xmath210 greater than the gas density . if we assume a gas density appropriate to a minimum - mass disk , then the maximum total density of gas and dust is still too low , by a factor of @xmath225 , for the marginally stable layer to become gravitationally unstable .
the literature discusses two remedies .
the first is to find ways of enhancing @xmath1 by factors of 310 , thereby increasing the dust - to - gas ratio at the midplane by factors of @xmath230 at fixed @xmath198 ( sekiya 1998 ; youdin & shu 2002 ; see also our figures [ dens_sigma05 ] and [ f9 ] , and run s8 in table 2 ) .
possible means of increasing the height - integrated metallicity include ( i ) decreasing @xmath211 through photoevaporation of gas , ( ii ) increasing @xmath17 by radial drift and pile - up of particles ( youdin & shu 2002 ; youdin & chiang 2004 ) or ( iii ) increasing @xmath17 by radiation blow - back of grains into the rim of a transitional disk ( chiang & murray - clay 2007 ) .
gas giants that form in such a metal - rich environment might be expected to have metallicities enhanced above the solar value by similar factors of 310 .
indeed some hot jupiters have remarkably metal - rich interiors ( sato et al .
alternatively , one can relax the assumption of perfect coupling between dust and gas , and exploit drag instabilities that result from finite momentum stopping times and the backreaction of dust on gas ( youdin & goodman 2005 ; johansen et al .
2007 ) .
we thank anders johansen , bryan johnson , eve ostriker , jack wisdom , and andrew youdin for helpful discussions .
we owe to jeremy goodman an insightful referee s report that brought to light the baroclinic instability .
this work was supported by nsf grant ast-0507805 . | gravitational instability of a vertically thin , dusty sheet near the midplane of a protoplanetary disk has long been proposed as a way of forming planetesimals . before roche densities
can be achieved , however , the dust - rich layer , sandwiched from above and below by more slowly rotating dust - poor gas , threatens to overturn and mix by the kelvin - helmholtz instability ( khi ) .
whether such a threat is real has never been demonstrated : the richardson criterion for the khi is derived for 2-d cartesian shear flow and does not account for rotational forces .
here we present 3-d numerical simulations of gas - dust mixtures in a shearing box , accounting for the full suite of disk - related forces : the coriolis and centrifugal forces , and radial tidal gravity .
dust particles are assumed small enough to be perfectly entrained in gas ; the two fluids share the same velocity field but obey separate continuity equations .
we find that the richardson number @xmath0 does not alone determine stability .
the critical value of @xmath0 below which the dust layer overturns and mixes depends on the height - integrated metallicity @xmath1 ( surface density ratio of dust to gas ) .
nevertheless , for @xmath1 between one and five times solar , the critical @xmath0 is nearly constant at @xmath20.1 .
keplerian radial shear stabilizes those modes that would otherwise disrupt the layer at large @xmath0 .
if the height - integrated metallicity is at least @xmath25 times greater than the solar value of 0.01 , then midplane dust densities can approach roche densities .
such an environment might be expected to produce gas giant planets having similarly super - solar metallicities . |
in recent endeavors , explorations of high - energy particle physics have largely been conducted in accelerator experiments , and for good reason .
accelerator laboratories provide controlled , high - luminosity environments in which very precise levels of measurement can be reached . despite these advantages , astrophysics experiments have also revealed a great deal of particle physics beginning with anderson s discovery of the positron in 1932 , then the muon , the pion etc .
, predating accelerator experiments , and continuing to the observation of neutrino masses and mixings .
it is clear that astrophysics has much to offer in studying the fundamental aspects of particle physics .
particle physics has entered an exciting era .
the standard model ( sm ) of the strong and electroweak interactions has been experimentally verified to high precision while the mechanisms for electroweak symmetry breaking and mass generation remain largely unknown .
theoretical arguments and indirect experimental evidence imply the existence of new physics near the electroweak scale below a few tev @xcite .
the leading candidates of theoretical models beyond the standard model include weak scale supersymmetry ( susy ) @xcite , strongly interacting dynamics @xcite , and low - scale string or quantum gravity @xcite .
it is encouraging that all of the above scenarios often lead to observable signatures in next generation colliders such as the cern large hadron collider ( lhc ) and an @xmath0 linear collider .
the field of high - energy neutrino astronomy finds itself in a position to contribute to two very different areas of science : astronomy and particle physics @xcite . on the one hand
, the next generation of neutrino telescopes may reveal the origins of the highest - energy cosmic rays , help us understand the progenitors of gamma - ray bursts , and provide other insights into some of the greatest outstanding astrophysical puzzles . on the other hand , very high - energy cosmic neutrinos
present a unique opportunity to study the interactions of elementary particles at energies comparable to and beyond those obtainable in current or planned colliders .
this is the main advantage of such experiments over traditional collider experiments .
currently , the highest energy achieved in collider experiments is at fermilab s tevatron , with @xmath1 tev .
this center - of - mass energy roughly corresponds to a pev neutrino striking a nucleon at rest , @xmath2 .
even the lhc at cern will only reach energies which correspond to 100 pev cosmic neutrinos .
it is certainly plausible , as we will discuss , that there is a neutrino flux at energies well beyond 1 eev .
even crude measurements of neutrino cross sections at extremely high energies would provide powerful tests of fundamental physics at and beyond a scale of @xmath3 tev .
additionally , sources of high - energy neutrinos may be observed from distances of hundreds or thousands of megaparsecs , providing baselines for tests of neutrino oscillations or decays which could not be carried out using accelerator , atmospheric or solar neutrinos .
the experimental status of high - energy neutrino astronomy is developing rapidly .
current technologies such as the amanda - ii @xcite and rice @xcite experiments at the south pole have proven successful , but with too little sensitivity to reach many of the most interesting physics goals .
several new experiments are soon to enter the field fortunately .
icecube @xcite will expand the effective area of amanda - ii by more than a factor of 20 while also improving both the angular and energy resolution .
antares @xcite , in the mediterranean , will use a similar technique , but with sensitivity to neutrino induced muons of lower energy ( down to 10 gev ) .
radio techniques will be employed in the balloon - based anita @xcite experiment which has it s first flight scheduled in the next year or two .
high - energy cosmic ray experiments , such as the pierre auger observatory @xcite and space - based observatories such as owl @xcite or euso @xcite , will also be sensitive to ultrahigh - energy cosmic neutrinos . other proposals , such as using acoustic techniques @xcite or natural salt domes as a cerenkov medium @xcite , or expanding icecube into a multi - kilometer ultrahigh - energy experiment @xcite , have been discussed as well . with the development of these many and varied techniques , a new window into fundamental particle physics
will be opened @xcite , possibly predating the lhc experiments .
this article is a mini review on the potential of high - energy neutrino astronomy in studying particle physics beyond the standard model .
we first present the current theoretical predictions of the high - energy cosmic neutrino flux in sec .
we then discuss methods for exploring particle physics via high - energy neutrino astronomy in sec .
[ part ] , paying particular attention to future neutrino telescopes and air shower observatories .
we summarize in sec .
[ theory ] the predicted experimental signatures from various new physics scenarios , including models with low - scale quantum gravity , low - scale string resonances , black holes and @xmath4-branes , electroweak instantons , and supersymmetry .
we draw our conclusions in sec .
[ conclude ] .
just as the performance of an accelerator experiment crucially depends on its luminosity , the particle physics reach of high - energy neutrino astronomy will depend on the incoming flux of cosmic neutrinos . here
we will briefly review some of the arguments for various neutrino fluxes from cosmic accelerators .
the spectrum of cosmic rays has been well measured up to energies near @xmath5 ev ( @xmath6 tev in the lab frame ) where the experiments become limited by poor statistics .
the spectrum consists of a series of power laws which change at energies called to as the `` knee '' and `` ankle '' ( see fig . [ rays ] ) .
the standard process believed to be responsible for the observed spectrum of cosmic rays is the acceleration of charged particles via second - order fermi acceleration @xcite . in fermi s original paper on the subject
, he proposed that cosmic rays were accelerated by reflecting off of time - varying magnetic fields associated with galactic clouds moving with randomly distributed velocities .
although this process will statically accelerate charged particles , it does so very slowly , and is generally not capable of even countering the effects of energy losses by ionization and other processes .
if instead we consider a compact region of dense plasma , however , the random motion of the matter and associated magnetic fields can be sufficient to accelerate charged cosmic rays to very high energies . although there is no strong evidence as of yet
, it is likely that supernova remnants accelerate most of the cosmic rays up to the knee in the spectrum , occurring around @xmath7 ev , by this mechanism .
a generic feature of fermi acceleration is a power - law spectrum , @xmath8 , where @xmath9 @xcite .
the maximum energy to which a cosmic ray source may accelerate particles can be estimated by a simple argument .
first , we assume that to accelerate a proton to a given energy in a given magnetic field , the size of the accelerator must be larger than the gyroradius of the particle s orbit : @xmath10 this condition yields a maximum energy of @xmath11 where @xmath12 is the lorentz factor of the cosmic accelerator . to produce cosmic rays with energies near the highest observed ( @xmath13 ev ) ,
very compact objects are required . for the most compact objects
, we can consider @xmath14 , or the schwartzchild radius of the object .
for such a source , we find a maximum energy of @xmath15 with only micro - gauss galactic magnetic fields , we must turn to extragalactic sources to accelerate cosmic rays to energies above the eev - scale .
extragalactic sources potentially capable of accelerating particles to such energies include the relativistic jets of active galactic nuclei ( agn ) or gamma - ray bursts ( grb ) .
as protons are accelerated to very high energies in such sources , they may undergo photo - meson interactions with the surrounding radiation fields . in such interactions , both charged and neutral pions are produced
. these pions then decay producing neutrinos and gamma - rays , respectively .
this process essentially guarantees the existence of accompanying neutrinos and gamma - rays given the observed cosmic ray flux .
alternatively , bounds can be placed on the cosmic neutrino flux by relating it to the cosmic ray spectrum . using this method , waxman and bahcall
have placed a upper bound for each neutrino flavor @xcite , @xmath16 this assumes that the sources in question are optically thin , or transparent , to protons .
if sources are optically thick to protons , howbever , the bound on the neutrino flux can only be based on gamma - ray observations by egret and is thus weaker by a factor of about 40 @xcite .
furthermore , if some sources were truly `` hidden '' , meaning neither nucleons nor photons could escape , no upper bound could be made on the corresponding neutrino flux . for a further discussion of these and similar arguments ,
see ref .
@xcite .
in addition to the neutrino production from cosmic ray interactions in or near cosmic accelerators , ultrahigh - energy cosmic rays produce neutrinos during propagation over cosmological distances @xcite .
protons of energy above a few times @xmath17 ev can scatter off of cosmic microwave background ( cmb ) photons with a center - of - mass energy roughly equal to the resonance corresponding to the mass of the @xmath18-hadron ( 1.232 gev ) . again , both charged and neutral pions can be produced this way , yielding neutrinos and gamma - rays .
the neutrino flux corresponding to this process is called the `` cosmogenic neutrino flux '' .
unlike the flux of neutrinos from cosmic accelerators , the spectrum of cosmogenic neutrinos depends only on the spectrum of ultrahigh - energy protons and the distribution of their sources , and thus can be reliably calculated @xcite .
the cosmogenic neutrino flux is often thought of as a `` guaranteed '' source of ultrahigh - energy neutrinos , assuming that the cosmic ray primaries at the highest observed energies are protons and not heavy nuclei @xcite .
there are ways in which neutrino fluxes larger than those described here could be produced .
for example , in models of non - accelerator cosmic ray origins , _
i.e. _ , models in which the highest - energy cosmic rays are produced in the decay or annihilations of superheavy objects . in such models , the resulting neutrino flux is not constrained by the arguments shown here .
although such scenarios are certainly interesting , we will not consider them in this article .
for further discussion on cosmic neutrino fluxes and their constraints , see ref . @xcite . and
dotted for @xmath19 ) @xcite .
note that after neutrino oscillations are considered , these fluxes will contain all three neutrino flavors ( @xmath20 ) in equal quantities .
the figure was taken from ref .
@xcite.,width=336 ] throughout the remainder of this article , we will primarily consider two representative choices for the cosmic neutrino spectrum .
the first is a flux equal to the bound set by waxman and bahcall , which we call the waxman - bahcall flux of eq .
( [ wbbound ] ) .
the second is the cosmogenic neutrino flux , as calculated in ref .
these are each shown in figure [ fluxes ] .
to identify potential signatures of new physics in high - energy neutrino interactions , one must first understand the phenomenology predicted by the standard model . in particular , the features of charged and neutral current interactions between high - energy neutrinos and target nuclei .
the standard model predicts the cross sections for neutrino - nucleon interactions , up to uncertainties in parton distribution functions at extremely small values of the momentum fraction @xmath21 , to energies beyond those probed by any planned neutrino telescopes @xcite . in this section
, we will describe the experimental features predicted by the standard model in neutrino telescopes and air shower experiments .
neutrino telescopes are essentially arrays of detectors distributed over a large volume of a cerenkov medium , such as water or ice
. these detectors may be sensitive to optical cerenkov radiation , as are amanda - ii , icecube and antares , or radio , as is rice .
we will focus on optical cerenkov detectors here , although the treatment of shower detection is quite easily generalized to include radio .
muons produced in the charged current interactions of muon neutrinos can travel several kilometers through a detector medium producing a track " of cerenkov light which can be observed and accurately reconstructed by neutrino telescopes . the rate of muon events observed in a large volume neutrino telescope is given by @xmath22 where @xmath23 is the zenith angle of an event ( @xmath24 is vertically downgoing ) , @xmath25 is avogadro s number , @xmath26 is the flux of muon neutrinos ( per unit energy , per unit time , per solid angle ) , @xmath27 is the differential neutrino - nucleon cross section ( where @xmath28 is defined such that @xmath29 ) , @xmath30 is the survival probability of a neutrino travelling through the earth , @xmath31 is the effective area of the detector ( @xmath32 km@xmath33 for icecube ) , @xmath34 is the length of time observed and @xmath35 is either the muon range or the length of material ( _ i.e. _ ice ) between the detector and the earth s surface , whichever is smaller .
the muon range is defined as the distance a muon propagates in the medium surrounding the detector before falling below a cutoff energy .
the muon range is given by @xcite @xmath36 where @xmath37 is the minimum muon energy required to produce an event .
this value is selected to reduce the number of background events while retaining as many signal events as possible . in optical cerenkov
neutrino telescopes , muons with energy as low as @xmath38 gev can be observed , although cuts well above this energy are often imposed when searching for high - energy neutrinos . in ice , @xmath39 and @xmath40 @xcite . for a pev muon in ice , and a 100 gev muon energy threshold ,
the range is approximately 1.7 km . for muons with energies of 10 pev , 100 pev or 1 eev
, the range increases to 7 , 13 and 18 km , respectively .
thus for very energetic muon neutrinos , the target volume of the experiment becomes a long cylinder , rather than a box .
this is particularly relevant for neutrinos coming from a direction near the horizon . unlike in accelerator experiments ,
the flux ( or luminosity ) of a cosmic neutrino beam may be unknown .
therefore , simply counting the number of events will not provide sufficient information to measure a neutrino cross section . instead
, information from the angular and energy distributions of events must be used @xcite . for a cross section of about @xmath41
mb , a particle s interaction length as it travels through the earth is equal to the earth s diameter .
this cross section is reached near @xmath42tev according to the standard model prediction .
thus , as the neutrino - nucleon cross section is increased from its standard model value , the effect of absorption in the earth becomes more pronounced and fewer of the observed events will come from neutrinos travelling through the earth . a crude way to measure the neutrino - nucleon cross section could
, therefore , be a comparison of the upgoing to downgoing ( or earth - skimming , etc . )
events in a high - energy neutrino telescope .
a more sophisticated analysis of the angular distribution of events as a function of energy would be more useful , however .
in addition to muon tracks , neutrino telescopes are sensitive to electromagnetic and hadronic showers .
these events can be produced by all three neutrino flavors in neutral current interactions , or in some charged current interactions .
for example , electromagnetic showers are produced in the charged current interaction of an electron neutrino .
the rate of shower events is calculated in an expression similar to eq .
( [ muonsrate ] ) , but the muon range , @xmath43 , together with the effective area , @xmath31 , are replaced by the effective volume of the detector .
also , the shower s energy is given by @xmath44 for neutral current events and @xmath45 for electron neutrino charged current events .
the minimum energy a shower must have to be observed by an optical cerenkov neutrino telescope is on the order of a few tev . for radio detectors ,
the shower threshold is much higher , in the pev to eev range .
finally , very large volume neutrino telescopes , such as icecube , are also capable of observing events uniquely associated with tau neutrinos .
tau neutrinos which interact via charged current in the detector medium produce a shower and a charged tau lepton .
below a few pev , the tau lepton s lifetime is sufficiently short that it decays producing a second shower essentially spatially coincident with the first one .
such an event is indistinguishable from a single shower . at higher energies , however , the tau lifetime can be long enough to distinguish these two showers .
for example , at 10 pev , a tau travels , on average , about 500 meters before decaying .
if both showers occur within the detector volume , such an event is called a `` double bang '' , and is a clear signature of a tau neutrino @xcite .
if the first of these showers occurs outside of the detector , with only the second shower being observed , the event is called a `` lollipop '' . here the observation of the shower with a minimum ionizing track ( produced by the tau ) constitutes the candy and the stick of the `` lollipop '' , respectively . again , this is a clear signature of a tau neutrino .
standard cosmic accelerators produce neutrinos via charged pion decay ( see section [ flux ] ) .
pion decays produce flavors of neutrinos in the ratio @xmath46 . over the long baselines
such neutrinos travel before reaching earth , neutrino oscillations modify this ratio to @xmath47 , or nearly equal quantities of all three flavors . considering only standard model neutrino interactions , these incoming flavor ratios can be translated to ratios of observed muon tracks , electromagnetic and hadronic showers and tau unique events @xcite . by measuring
the ratios of these event types observed in icecube , the presence of interactions beyond the standard model may be tested .
very high - energy cosmic neutrinos can occasionally interact with particles in the earth s atmosphere producing extended air showers observable in high - energy cosmic ray experiments such as agasa , hires or the next generation pierre auger observatory @xcite . although the characteristics of neutrino induced showers do , in principle , differ from those initiated by hadronic cosmic rays @xcite , significantly more hadronic showers are expected , thus making showers from neutrino primaries difficult to conclusively identify .
primary particles which have a near - horizontal trajectory can provide an opportunity to distinguish neutrinos from hadronic events , however .
in contrast to hadronic cosmic rays ( which interact in the top of the atmosphere ) , neutrino primaries have considerably smaller cross sections and thus interact with nearly equal probability throughout the atmosphere .
if a shower is observed which was initiated deep inside of the atmosphere , it can be associated with a neutrino ( or other weakly interacting particle ) primary .
for there to be sufficient column depth to make this distinction ( typically 3000 to 4000 g / cm@xmath33 is required ) , only primaries within about @xmath49 of the horizon can be considered @xcite .
the class of cosmic ray events which can be associated with neutrino primaries are called `` deeply penetrating , quasi - horizontal showers '' . to calculate
the rate of neutrino induced deeply penetrating , quasi - horizontal showers in an air shower experiment , one must estimate the acceptance to neutrino detection .
this quantity is essentially the effective target mass multiplied by the accessible solid angle .
it is often given in units of km@xmath50 water equivalent steradians ( km@xmath50 we sr ) , where , for example , 1 km@xmath50 water equivalent would be the target mass contained in one cubic kilometer of water or ice .
@xcite estimates the acceptance of the agasa experiment to deeply penetrating , quasi - horizontal showers to be 0.05 km@xmath50 we sr at @xmath51 gev , rising to 1.0 km@xmath50 we sr at @xmath52 gev and above .
they estimate the acceptance of auger to be a factor of 20 , 20 and 50 larger than agasa at @xmath51 , @xmath52 and @xmath53 gev , respectively .
these estimates consider showers within 15@xmath54 of the horizon and with a maximum height of 15 kilometers . for a discussion of the hires acceptance ,
see ref .
@xcite .
the number of neutrino events observed as deeply penetrating , quasi - horizontal showers is given by @xmath55 where @xmath25 is avogadro s number , @xmath26 is the flux of neutrinos ( per unit energy , per unit time , per solid angle ) , @xmath56 is the neutrino - nucleon scattering cross section , @xmath57 is the acceptance of the detector and @xmath34 is the length of time observed .
this expression assumes that roughly all of the neutrino s energy goes into the produced shower .
if this is not the case , such as with neutral current interactions , a differential cross section should be used as in eq .
( [ muonsrate ] ) and the acceptance be written as a function of the shower energy rather than neutrino energy . at the very high energies at which air shower experiments are most effective ( 0.1 eev and higher ) , a reasonable and conservative flux of neutrinos to consider
is the cosmogenic flux ( see section [ flux ] )
. this flux peaks at about 0.1 eev , but is substantial at 1 eev and above .
if we insert this flux , an experimental acceptance and the neutrino - nucleon cross section , we can predict the number of neutrino induced deeply penetrating , quasi - horizontal events which would be observed in an air shower experiment .
in addition to deeply penetrating quasi - horizontal showers , it may be possible to identify showers produced by earth - skimming tau neutrinos using the fluorescence detectors of the auger experiment @xcite .
earth - skimming , ultrahigh - energy tau neutrinos produce tau leptons in charged current interactions . since at ultrahigh - energies the tau decay length is comparable to its interaction length , a shower produced in the tau decay can be observed as an extended air shower if the tau is produced not too deep beneath the earth s surface . the rates for this class of events
is expected to be rather small , however , and we will not study this signature in further detail here .
in high - energy cosmic neutrino experiments , the new physics enters the neutrino - nucleon scattering cross section , @xmath58 , as in eqs .
( [ nint ] ) and ( [ nair ] ) .
it is given by @xmath59 where @xmath60 is the scattering cross section for the neutrino - parton subprocesses , which reflects the fundamental dynamics of neutrino interactions .
the sum is over all contributing partons , @xmath61 , and @xmath62 are the parton distribution functions .
the cross sections are insensitive to choice of momentum transfer , @xmath63 .
the cross sections are also insensitive to uncertainties in the parton distribution functions at low @xmath21 if there is a tev scale threshold . for instance , the highest - energy neutrino fluxes , which are large at @xmath64 , probe @xmath65 , within the range of validity of the parton distribution functions that we take as cteq5 @xcite .
there are many interesting scenarios in which neutrino - nucleon interactions would be substantially enhanced over the standard model prediction at high energies . in this section ,
we summarize the literature on these models , describing how their signatures may be observed in high - energy cosmic neutrino experiments . the most dramatic proposal ,
perhaps , is that of tev scale gravity with the existence of extra spatial dimensions .
the motivation for such a scenario can be traced to the understanding of the large hierarchy between the electroweak scale , below about @xmath66(1 tev ) , and the planck scale , @xmath67 gev .
if there are @xmath68 extra spatial dimensions , then the 4-dimensional planck scale is related to the @xmath69-dimensional planck scale , @xmath70 , by @xmath71 where @xmath72 is the volume of the extra dimensions if they are flat with a compactification scale , @xmath73 @xcite ; and @xmath74 if the @xmath75 extra dimension is warped " with a curvature @xmath76 @xcite .
an immediate implication of this scenario would be to naturally understand the largeness of the planck scale in comparison with the electroweak scale .
namely it can be interpreted as due to the large volume of the extra dimensions while the fundamental @xmath69-dimensional planck scale , @xmath70 , may be low , possibly at the tev scale .
tev ) and kk graviton masses ( 500 gev@xmath77 tev ) .
dotted lines represent the prediction of the standard model.,width=384 ] in models with large and flat extra dimensions , often called the add scenario @xcite , the fundamental planck scale is assumed to be on the order of a tev and a large number of kaluza - klein ( kk ) graviton states of mass @xmath78 of sub - ev become accessible .
their couplings to the standard model particles at an energy @xmath79 are enhanced to @xmath80 , and thus potentially large effects on high - energy processes may occur @xcite .
high - energy neutrinos can exchange these kaluza - klein gravitons with quarks or gluons in target nucleons resulting in an enhancement . for a scale of gravity near 1 tev
, neutrinos above pev energies begin to interact largely by the effects of new physics . in the randall - sundrum scenario , an anti - de sitter dimension with a non - factorizable warped geometry
is introduced @xcite .
again , kaluza - klein gravitons become accessible at the tev scale @xcite enhancing the neutral current neutrino interaction rate above this scale . the effect on the neutral current neutrino - nucleon cross section in these scenarios
is shown in fig .
[ gravitysigma ] @xcite .
the solid lines represent models with large extra dimensions ( add ) for the quantum gravity scale taken as 1 tev .
the three curves correspond to different choices for the unitarization scheme of the partial wave amplitudes proposed in @xcite , as compared to some other calculations @xcite .
short dashed lines represent the randall - sundrum model with a varied ads scale ( @xmath81 tev ) and kk graviton masses ( 500 gev@xmath77 tev ) .
dotted lines represent the prediction of the standard model . in the scenarios considered here , neutrino telescopes expect to observe more neutral current events per charged current event than predicted by the standard model .
this effect sets in right above the threshold near @xmath70 , providing a clear indication for new physics .
furthermore , the ratio of downgoing to upgoing events will be enhanced over the standard model prediction as more neutrinos are absorbed as they propagate though the earth @xcite .
the behavior of the energy spectrum due to these effects is shown in fig . [ two ] .
events in icecube .
the left and right panels show the upgoing and downgoing events , respectively .
the waxman - bahcall flux has been used ( see section [ flux ] ) . in each panel ,
the solid line is the event rate when the neutrino - nucleon cross section @xmath82 is used ; the dashed line uses @xmath83 with a kk graviton mass @xmath84 gev and @xmath85 tev ; the dotted line corresponds to @xmath86 alone .
the figure is a modified version from ref .
@xcite.,width=259 ] at energies above the compactication scale , @xmath78 , the extra dimensions and kk effects may become observable if @xmath87 is not too high as discussed in the previous section . at even higher energies near the string scale , @xmath88 ,
string effects dominate over gravitational effects based on string perturbation arguments @xcite .
the string scale may be related to the @xmath69-dimensional gravity scale by @xmath89 with @xmath90 , depending on the string coupling and compactification of the extra dimensions .
large effects from string resonances may be produced when @xmath91 .
cross sections via tev string resonances considering @xmath92 and @xmath93 contributions ( dashed curved ) @xcite and the @xmath92 contribution only ( dot - dashed curve ) .
the string scale is taken to be @xmath88=1 tev and the chan - paton factor used is @xmath94 .
also plotted is the standard model neutral current prediction ( solid curve ) .
, width=384 ] general tree - level open - string amplitudes can be constructed based on the veneziano amplitude @xcite , @xmath95 where @xmath96 is the string tension .
this amplitude develops simple poles at @xmath97 with @xmath98 leading to resonances in the matrix elements .
the physical effects of these resonances have been explored @xcite , including their signatures in cosmic neutrino experiments @xcite .
we present the neutrino - nucleon cross sections due to veneziano amplitude resonances in fig .
[ stringsigma ] @xcite .
the solid curve shows the prediction for the standard model neutral current process , while the dashed and dot - dashed curves represent string excitations with and without the gluon contribution , respectively .
we see that neutrino - gluon scattering can be the dominant process , 5 to 10 times larger than the neutrino - quark induced processes .
it is interesting to note that even for processes that vanish in the standard model at tree - level , there can still be substantial stringy contributions to their amplitudes at high energies . generally speaking ,
the energy - dependence of the cross sections for the string resonances are weaker than those for kk states .
the event rates expected in icecube and auger have been calculated for these models @xcite . with the standard model interactions ,
only about 0.2 ( 0.7 ) shower events per year are expected in the experiment from a cosmogenic ( wb ) neutrino flux , these rates can be enhanced by a factor of 5 to 6 due to string excitations with @xmath99 tev or a factor of about 1.5 to 1.7 with @xmath100 tev . ) of 1 tev and a minimum black hole mass of 1 tev ( solid ) , a scale of 1 tev and a minimum black hole mass of 3 tev ( long dashes ) , a scale of 2 tev and a minimum black hole mass of 2 tev ( short dashes ) , and a scale of 2 tev and a minimum black hole mass of 6 tev ( dot - dashes ) . the dotted line represents the prediction of the standard model .
the number of extra dimensions was assumed to be six .
the figure was taken from ref .
@xcite.,width=364 ] and 3 , respectively .
all models shown have @xmath101 and @xmath102 tev .
the waxman - bahcall and cosmogenic neutrino fluxes were used in the left and right frames , respectively ( see section [ flux ] ) .
@xmath103 corresponds to a horizontal event while positive and negative values correspond to downgoing and upgoing muons , respectively .
while the enhanced cross section dramatically increases the downgoing event rate , the rate of upgoing events is suppressed due to absorption in the earth .
the figure was taken from ref .
@xcite.,width=384 ] at trans - planckian energies , @xmath104 , it has been argued that black hole production will be the leading process @xcite .
this is because the energy - dependence of the black hole production cross section grows faster than for sub - planckian processes and the number of non - perturbative states grows faster than the perturbative string states .
the cross section for black hole production can be naively estimated by the geometric description , @xmath105 where @xmath106 is the schwartzchild radius of a black hole formed with a mass equal to the center - of - mass energy of the collision . in @xmath107 dimensions ,
the schwartzchild radius of a black hole of mass @xmath108 is given by @xmath109^{1/(1+n ) } \bigg[\frac{2^n \pi^{(n-3)/2}\,\gamma(\frac{3+n}{2})}{2+n}\bigg]^{1/(1+n)}.\ ] ] although some studies support the validity of the geometric cross section argument @xcite , it is possible that a substantial fraction of the total energy will be radiated away in the form of gravitational waves , reducing the mass of any black hole which may be formed , and reducing the corresponding cross section @xcite .
although the lowest mass possible for the black hole creation is approximately the fundamental planck scale , @xmath87 , the effective center - of - mass energy should be several times larger for the semi - classical argument to hold valid . to parameterize this effect , we introduce the quantity , @xmath110 .
in addition to the ambiguity of the value of @xmath111 , other uncertainties can possibly arise in the estimation of the cross section @xcite , which we do not consider further here . in fig .
[ bhcrossfig ] , we show the cross section for black hole production in neutrino - nucleon interactions for different choices of @xmath70 and @xmath111 . as a result of the sum over all partons and the lack of suppression from small perturbative couplings ,
the black hole cross section may exceed standard model interaction rates by two or more orders of magnitude .
the cross sections corresponding to neutrino interaction lengths equal to the horizontal and vertical depths of icecube position are also given in fig .
[ bhcrossfig ] by the horizontal dotted lines .
we see that for the geometric cross section , @xmath112 , and neutrino energies @xmath113 where the cosmogenic flux peaks , black hole production increases the probability of conversion in down - going neutrinos without increasing the cross section so much that vertical neutrinos would be shadowed before reaching the detector . we therefore expect significantly enhanced rates in neutrino telescopes .
the energy - dependence of the black hole production cross section is stronger than for the other processes we have discussed so far , thus confirming the argument that black hole production would likely be the dominant effect of low scale gravity at higher energies .
black holes decay via hawking evaporation almost instantly ( with a lifetime of the order @xmath114 ) .
the hawking radiation follows a thermal distribution with temperature @xmath115 with an average multiplicity of particles radiated of approximately @xmath116 .
naively , the particles are radiated in numbers proportional to their degrees of freedom , _
@xmath117 hadronic , @xmath118 to charged leptons , etc .
assume that the signals in a neutrino telescope produced in black hole decays follow these ratios @xcite .
the 10% of hawking radiation that produces charged leptons generates equal numbers of muons , taus and electrons which can be observed as muon tracks , tau unique events and electomagnetic showers , respectively .
the full 75% of hawking radiation which goes into hadronic modes result in hadronic showers .
this is in constrast to the ratios of event types predicted for standard model interactions .
taking into account the degrees of freedom corresponding to each channel and the factors effecting the probability of detection ( _ i.e. _ muon range , etc . )
, the ratios of muons to taus to showers can be predicted for a particular black hole production model .
for example , for a model with @xmath119 tev and @xmath120 , about twice as many showers are expected than muon tracks ( considering a @xmath121 flux ) .
in contrast , the standard model prediction is about 20% more muons than showers @xcite . by combining flavor ratio measurements with angular and energy distributions , large volume neutrino telescopes such as icecube
will be capable of searching for evidence of black hole production in models with a fundamental planck scale up to 1 to 2 tev .
the angular distributions of muon tracks above 500 tev in a kilometer - scale neutrino telescope , such as icecube , in models of black hole production are shown in fig .
[ angle ] .
while the enhanced cross section dramatically increases the downgoing event rate ( @xmath122 ) over the standard model prediction ( dotted curves ) , the rate of upgoing events is suppressed due to absorption in the earth .
air shower experiments , unlike neutrino telescopes , do not have the ability to observe muon tracks or identify tau unique events .
they are , however , very sensitive to eev scale cosmic neutrinos and are thus capable of placing valuable limits on models of black hole production .
currently , the strongest such limit comes from the agasa air shower experiment .
agasa has reported the observation of 1 neutrino - like ( deeply penetrating , quasi - horizontal ) event , and predict a background to this signal of 1.7 events from misidentifying hadronic primaries @xcite . at the 95% confidence level
, this places an upper limit of 3.5 black hole events .
this can be directly translated into a limit on the fundamental planck scale , @xmath70 . for values of @xmath111 in the range of @xmath123 ,
agasa can place a lower limit on @xmath70 of 1.0 to 1.4 tev @xcite , a limit which is competitive to the strongest bounds from collider experiments @xcite .
auger , with considerably higher acceptance to these events , is expected to improve this sensitivity to @xmath124 tev for @xmath125 @xcite . -brane production for @xmath126 ( left ) , @xmath127 ( right ) , @xmath128 , and @xmath129 from bottom to top .
the standard model cross section , @xmath130 , ( dotted ) is also shown .
the figure was taken from ref .
@xcite.,title="fig:",width=288 ] -brane production for @xmath126 ( left ) , @xmath127 ( right ) , @xmath128 , and @xmath129 from bottom to top .
the standard model cross section , @xmath130 , ( dotted ) is also shown .
the figure was taken from ref .
@xcite.,title="fig:",width=288 ] @xmath4-branes are @xmath4-dimensional , spatially extended solutions of gravitational theories .
the existence of such objects is a generic prediction of theories with extra dimensions .
if the fundamental scale of gravity is of the order of a tev , then it is reasonable to expect that in addition to black holes ( a spherically symmetric @xmath131-brane ) , higher dimensional states may also be generated in high - energy collisions @xcite .
the cross section for @xmath4-brane production is argued to be geometrical , similar to that for black hole production , except that it may have a lower threshold near the quantum gravity scale . if the @xmath4-brane wraps only around the small ( compact ) dimensions , the cross section for @xmath4-brane production can be comparable to , or even larger than , the cross section for black hole production @xcite . if the p - brane wraps around large dimensions as well , their production will be suppressed by powers of @xmath132 @xcite .
typical cross sections for @xmath4-brane production in @xmath133 collisions are presented in fig .
[ fig : sigma ] @xcite , that could be higher than that of the black hole production by orders of magnitude .
unlike with the standard hawking radiation picture for black hole evaporation , the decay of @xmath4-branes is not well understood .
@xmath4-branes may decay into branes of lower dimension .
alternatively , they may decay directly into a combination of brane and bulk particles . below the energy threshold for @xmath4-brane or black hole production , lighter states , called `` string balls ''
may also be produced @xcite .
we do not study these objects further here , since we consider our presentation already quite representative for the conservative scenario as in the string resonances in sec .
[ sr ] , and for the more optimistic scheme in this section .
standard model electroweak instantons represent tunnelling transitions between topologically inequivalent vacua , leading to baryon plus lepton number ( @xmath134 ) violating processes .
such processes are exponentially suppressed below the so - called `` sphaleron '' energy , @xmath135 tev . above this scale , however , such process may be unsuppressed and the corresponding cross sections can be quite large @xcite , potentially resulting in enhanced neutrino scattering signals in cosmic neutrino experiments .
the neutrino - nucleon cross section corresponding to electroweak instanton induced interactions is difficult to reliably calculate .
one approach to this problem is to use a perturbative method in close analogy to qcd @xcite .
alternatively , this calculation has been performed using a generalized semi - classical approach @xcite . in this second approach , these interactions remain suppressed to much higher energies up to about 30 times the sphaleron energy .
the estimated neutrino - nucleon cross sections corresponding to these two approaches are shown in fig .
[ instantonsigma ] @xcite . from the standpoint of neutrino phenomenology
, it is important to note the extremely rapid increase of the neutrino - nucleon cross section demonstrated in these models .
this is in contrast to the more gradual growth predicted for the cross sections for black hole production , kaluza - klein exchanges , etc . below the energy thresholds for such interactions , the standard model predictions are accurate . at energies
roughly a factor of ten higher , the cross section becomes sufficiently large that the earth efficiently absorbs the incoming neutrino flux , as indicated above the horizontal dotted lines ( horizontal and downward depths ) in fig .
[ instantonsigma ] .
thus in a neutrino telescope , a sharp enhancement in a fairly narrow range of energies is predicted for these models , as depicted in fig .
[ instantonenergy ] . although spectacular , the ability of planned experiments to observe such features is limited , however .
even with the more optimistic of the models considered here , an experiment such as icecube is expected to see only on the order of one event per year from instanton induced processes @xcite .
future experiments with very large volume will be required to further probe such models .
another interesting characteristic feature of instanton - induced processes is the large multiplicity of final state particles and the violation of @xmath134 .
the basic operators involving quark and lepton fields are of the form @xmath136 @xcite , where @xmath137 is the number of fermion generations .
it has been argued that the processes involving multiple gauge bosons and higgs bosons , such as @xmath138 , can be significantly enhanced @xcite .
a typical neutrino - induced event could thus be @xmath139 with both quarks and leptons of all three generations involved simultaneously in the primary production , this type of events should look quite unique .
it is difficult to predict how such events would appear to the icecube detector , however , given the fact that the particles will be highly collimated and difficult to separate .
supersymmetry ( susy ) remains a leading candidate for physics beyond the standard model .
although weak - scale susy is only weakly coupled to the standard model and generally would not lead to substantially enhanced neutrino scattering cross sections , certain charged particles produced by cosmic neutrinos may be long - lived and may provide observable signatures .
this scenario could be naturally realized when the gravitino is the stable lightest supersymmetric particle ( lsp ) , and a charged slepton ( such as stau ) is the next - to - lightest supersymmetric particle ( nlsp ) @xcite .
interactions of high - energy neutrinos may be able to produce pairs of sparticles which rapidly decay to charged slepton nlsps which can only decay further into states including a gravitino .
with only highly suppressed couplings allowing this decay , the nlsp stau can be sufficiently long lived to be potentially observable in a large volume neutrino telescope such as icecube @xcite .
cross section in the scenario of stau - nlsp .
curves are shown for squarks of mass 250 , 600 and 900 gev .
a stau mass of 250 gev was used .
the chargino mass used was also 250 gev .
the top curve corresponds to the standard model charged current interactions .
the figure was taken from ref .
@xcite.,width=288 ] sparticle pair production in neutrino - nucleon interactions is dominated by the @xmath140-channel chargino exchange , resulting in a slepton and a squark .
the squark then quickly decays into a slepton nlsp .
the cross section for this process is rather small , however , typically two to three orders of magnitude below the standard model processes in the energy range well above the kinematic threshold as seen fig .
[ fig : stau ] .
the key observation here is that sleptons produced in neutrino interactions travel through the earth , losing energy via ionization processes and radiation . due to their much greater mass
, sleptons lose far less energy than muons produced in standard model charged current interactions .
the `` slepton range '' can extend to hundreds or thousands of kilometers , thus in part making up for the low cross section for their production .
the two sparticles produced in these interactions travel from their point - of - origin separated by an angle of @xmath141 . therefore , the signature of this process consists of a pair of cerenkov tracks , separated by a distance @xmath142 , where @xmath143 is the distance between the detector and the sparticles point - of - origin . considering a pev neutrino , for example , two sleptons separated by @xmath144 could be produced . after travelling @xmath145 kilometers , their tracks
would be separated by @xmath146 meters , which could be potentially distinguished in a neutrino telescope .
this `` double track '' signature would provide a method of distinguishing sparticle tracks from ordinary muon tracks .
typical models with stau - nlps predict only a one- or a few double - track events per year in icecube @xcite .
larger volume detectors , such as extensions of icecube , may be needed to further explore this possibility . as a final remark on the possibility of observing supersymmetric particles in high - energy cosmic - ray interactions , neutrino experiments may be able to identify particles which are part of the cosmic ray spectrum , such as in cosmic ray models of top - down origin @xcite .
for instance , if a neutralino is the lsp , it will interact with nucleons in a manner which somewhat resembles a neutrino neutral current interaction .
thus without very large fluxes of high - energy cosmic neutralinos , it would be very difficult to distinguish any such particles from neutrinos .
neutralinos can have considerably smaller cross sections with nucleons than neutrinos , however , allowing them to travel through the earth at energies at which neutrinos will be efficiently absorbed @xcite .
thus ultrahigh - energy , neutralino - induced showers provide a low background signal in the direction of the earth .
future space - based air shower experiments , such as owl or euso , may be sensitive to this signature in some scenarios @xcite .
in this article , we have reviewed the ability of neutrino telescopes and air shower experiments to study particle physics with high - energy cosmic neutrinos .
the main advantage of such experiments over traditional collider experiments is the higher energy at which interactions can be studied .
several of the experiments we describe in this article could expect multiple events per year at energies above 1 eev , corresponding to about 40 tev in the center - of - mass frame of a neutrino - nucleon collision .
clearly collider experiments have advantages over astroparticle techniques as well .
most notably , the high luminosities and well - controlled conditions of collider experiments are luxuries astronomers do not often enjoy . together ,
these advantages and disadvantages determine the areas of particle physics in which neutrino astronomy can be most useful .
in particular , models in which dramatic deviations from the standard model occur at energies beyond the reach of colliders can often be tested in such experiments . in this article , we have reviewed several of such scenarios , summarized as follows : * kaluza - klein gravitons in low - scale quantum gravity scenarios : sensitive near the threshold , presumably @xmath147 tev or @xmath148 pev .
* tev string resonances in low string - scale scenarios : near and above the string scale presumably @xmath149 tev , thus @xmath150 pev .
* @xmath4-brane production in low - scale quantum gravity scenarios : near and above the quantum gravity scale , presumably @xmath151 tev , thus @xmath150 pev . * black hole production in low - scale quantum gravity scenarios : likely the dominant signature at trans - planckian energies @xmath152 tev , thus @xmath153 pev .
* electroweak instanton induced processes : above the sphaleron energy @xmath154 tev , thus @xmath155 pev . * supersymmetry with a charged slepton as a long - lived nlsp .
although we did not discuss them in this article , other probes of exotic particle physics are possible using neutrino astronomy .
these include searches for particle dark matter and neutrinos associated with models of top - down cosmic ray origin
. although astronomy , and not particle physics , is the primary objective of neutrino telescopes and air shower experiments , upcoming experiments such as icecube , auger , anita , owl and euso will each study interactions at energies well beyond the reach of colliders and provide complementary probes to the traditional techniques used to study our universe at the smallest scales .
we would like to thank jaime alvarez - muniz for his assistance with the figures and jonathan feng for helpful comments .
th is supported in part by doe grant no .
de - fg02 - 95er40896 and the wisconsin alumni research foundation .
dh is supported by the leverhulme trust .
for a recent review on the status of the standard model and beyond , see _
e. g. _ , g. altarelli , nucl .
instrum .
a * 518 * , 1 ( 2004 ) [ arxiv : hep - ph/0306055 ] . for an introductory review on supersymmetry ,
e. g. _ , s. p. martin , arxiv : hep - ph/9709356 .
for a recent review on strong dynamics in the electroweak sector , see _
e. g. _ , c. t. hill and e. h. simmons , phys .
rept . * 381 * , 235 ( 2003 ) [ arxiv : hep - ph/0203079 ] . n. arkani - hamed , s. dimopoulos and g. dvali , phys
* b429 * , 263 ( 1998 ) ; phys . rev . *
d59 * , 086004 ( 1999 ) ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. dvali , phys . lett . *
b436 * , 257 ( 1998 ) .
l. randall and r. sundrum , phys .
lett . * 83 * , 3370 ( 1999 ) . for recent reviews of high - energy neutrino astronomy , see _ e. g. _ , j. g. learned and k. mannheim , ann .
nucl . part .
sci . * 50 * , 679 ( 2000 ) ; f. halzen and d. hooper , rept .
phys . * 65 * , 1025 ( 2002 ) [ arxiv : astro - ph/0204527 ] ; t. j. weiler , int .
j. mod .
phys . a * 18 * ( 2003 ) 4065 [ arxiv : astro - ph/0304180 ] .
e. andres _ et al .
_ , nature * 410 * , 441 ( 2001 ) .
i. kravchenko , arxiv : astro - ph/0306408 ; i. kravchenko _ et al .
_ , astropart .
phys . * 20 * , 195 ( 2003 ) [ arxiv : astro - ph/0206371 ] .
j. ahrens _ et al .
_ [ the icecube collaboration ] , nucl .
suppl . *
118 * , 388 ( 2003 ) [ arxiv : astro - ph/0209556 ] ; ahrens , j. , _ et al . _ , 2003 , to be published in particle astrophysics , astro - ph/0305196
; icecube preliminary design document , available at : http://icecube.wisc.edu / science / sci - tech - docs/. f. blanc _ et al . _
[ antares collaboration ] , _ presented by l. thompson on behalf of the antares collaboration , to appear in the proceedings of 28th international cosmic ray conferences ( icrc 2003 ) , tsukuba , japan , 31 jul - 7 aug 2003 . _
u. f. katz [ the antares collaboration ] , arxiv : astro - ph/0310736 .
anita proposal , p. gorham , principle investigator ; anita home page : http://www.ps.uci.edu/ anita / index.html .
j. blumer [ auger collaboration ] , _ prepared for 28th international cosmic ray conferences ( icrc 2003 ) , tsukuba , japan , 31 jul - 7 aug 2003 _ ; s. coutu , x. bertou and p. billoir [ auger collaboration ] , _ prepared for 23rd johns hopkins workshop on current problems in particle theory : neutrinos in the new millennium , baltimore , maryland , 10 - 12 jun 1999 _ ; k. s. capelle , j. w. cronin , g. parente and e. zas , astropart .
phys . *
8 * , 321 ( 1998 ) [ arxiv : astro - ph/9801313 ] .
f. w. stecker , j. f. krizmanic , l. m. barbier , e. loh , j. w. mitchell , p. sokolsky and r. e. streitmatter , arxiv : astro - ph/0408162 ; s. i. dutta , m. h. reno and i. sarcevic , phys . rev .
d * 66 * , 033002 ( 2002 ) [ arxiv : hep - ph/0204218 ] .
l. scarsi , nuovo cim . * 24c * , 471 ( 2001 ) ; d. b. cline and f. w. stecker , arxiv : astro - ph/0003459 .
l. g. dedenko , i. m. zheleznykh , s. k. karaevsky , a. a. mironovich , v. d. svet and a. v. furduev , bull .
* 61 * ( 1997 ) 469 [ izv . ross .
* 61 * ( 1997 ) 593 ] ; g. riccobene , l. cosentino , m. musumeci , g. pavan and f. speziale [ nemo collaboration ] , nucl .
instrum . meth .
a * 518 * , 220 ( 2004 ) .
gorham , d. saltzberg , a. odian , d. williams , d. besson , g. frichter and s. tantawi , nucl .
instrum .
a * 490 * , 476 ( 2002 ) [ arxiv : hep - ex/0108027 ] .
f. halzen and d. hooper , jcap/008a/1003 , arxiv : astro - ph/0310152 .
e. waxman and j. n. bahcall , phys .
d * 59 * , 023002 ( 1999 ) [ arxiv : hep - ph/9807282 ] ; j. n. bahcall and e. waxman , phys .
d * 64 * , 023002 ( 2001 ) [ arxiv : hep - ph/9902383 ] ; p. s. coppi and f. a. aharonian , astrophys .
j. * 487 * , l9 ( 1997 ) [ arxiv : astro - ph/9610176 ] .
z. fodor , s. d. katz , a. ringwald and h. tu , jcap * 0311 * , 015 ( 2003 ) [ arxiv : hep - ph/0309171 ] .
d. hooper , a. taylor and s. sarkar , arxiv : astro - ph/0407618 ; m. ave , busa , a. olinto , a. watson and t. yamamoto , in progress .
o. e. kalashev , v. a. kuzmin , d. v. semikoz and g. sigl , phys .
d * 66 * , 063004 ( 2002 ) [ arxiv : hep - ph/0205050 ] .
r. engel , d. seckel and t. stanev , phys .
d * 64 * , 093010 ( 2001 ) [ arxiv : astro - ph/0101216 ] . j. alvarez - muniz , j. l. feng , f. halzen , t. han and d. hooper , phys .
d * 65 * , 124015 ( 2002 ) [ arxiv : hep - ph/0202081 ] .
r. gandhi , c. quigg , m. h. reno and i. sarcevic , phys .
* d58 * ( 1998 ) 093009 [ arxiv : hep - ph/9807264 ] ; r. gandhi , c. quigg , m. h. reno and i. sarcevic , astropart . phys . * 5 * , 81 ( 1996 ) [ arxiv : hep - ph/9512364 ] .
s. i. dutta , m. h. reno , i. sarcevic and d. seckel , phys .
d * 63 * , 094020 ( 2001 ) [ arxiv : hep - ph/0012350 ] . c. tyler , a. olinto , and g. sigl , phys .
rev . * d63 * , 055001 ( 2001);d .
hooper , phys .
d * 65 * , 097303 ( 2002 ) [ arxiv : hep - ph/0203239 ] ; a. kusenko and t. j. weiler , phys .
lett . * 88 * , 161101 ( 2002 ) [ arxiv : hep - ph/0106071 ] ; p. jain , s. kar , d. w. mckay , s. panda and j. p. ralston , phys .
d * 66 * , 065018 ( 2002 ) [ arxiv : hep - ph/0205052 ] .
j. f. beacom , n. f. bell , d. hooper , s. pakvasa and t. j. weiler , arxiv : hep - ph/0307025 ; j. f. beacom , n. f. bell , d. hooper , s. pakvasa and t. j. weiler , phys .
lett . * 90 * , 181301 ( 2003 ) [ arxiv : hep - ph/0211305 ] . for a review of ultrahigh - energy air shower experiments , see _
e. g. _ , l. anchordoqui , t. paul , s. reucroft and j. swain , int .
j. mod .
a * 18 * , 2229 ( 2003 ) [ arxiv : hep - ph/0206072 ] .
a. jain , p. jain , d. w. mckay and j. p. ralston , hep - ph/0011310 ; l. anchordoqui , h. goldberg , t. mccauley , t. paul , s. reucroft and j. swain , phys .
d * 63 * , 124009 ( 2001 ) [ arxiv : hep - ph/0011097 ] .
l. a. anchordoqui , j. l. feng , h. goldberg and a. d. shapere , phys .
d * 65 * , 124027 ( 2002 ) [ arxiv : hep - ph/0112247 ] ; j. l. feng and a. d. shapere , phys .
* 88 * , 021303 ( 2002 ) [ arxiv : hep - ph/0109106 ] ; c. aramo , a. insolia , a. leonardi , g. miele , l. perrone , o. pisanti and d. v. semikoz , astro - ph/0407638 ; j. l. feng , p. fisher , f. wilczek and t. m. yu , phys . rev .
lett . * 88 * , 161102 ( 2002 ) [ arxiv : hep - ph/0105067 ] ; x. bertou , p. billoir , o. deligny , c. lachaud and a. letessier - selvon , astropart .
* 17 * , 183 ( 2002 ) [ arxiv : astro - ph/0104452 ] .
h. l. lai _ et al .
_ [ cteq collaboration ] , eur .
j. c * 12 * , 375 ( 2000 ) [ hep - ph/9903282 ] .
g. guidice , r. rattazzi and j. wells , nucl . phys . *
b544 * , 3 ( 1999 ) ; t. han , j.d .
lykken and r. zhang , phys . rev .
* d59 * , 105006 ( 1999 ) ; e.a .
mirabelli , m. perelstein and m.e .
peskin , phys .
* 82 * , 2236 ( 1999 ) ; j. hewett , phys .
* 82 * , 4765 ( 1999 ) ; t. rizzo , phys . rev . * d59 * , 115010 ( 1999 ) .
s. nussinov and r. shrock , phys .
d * 64 * , 047702 ( 2001 ) [ arxiv : hep - ph/0103043 ] ; g. domokos and s. kovesi - domokos , phys .
lett . * 82 * , 1366 ( 1999 ) [ arxiv : hep - ph/9812260 ] ; p. jain , d. w. mckay , s. panda and j. p. ralston , phys .
b * 484 * , 267 ( 2000 ) [ arxiv : hep - ph/0001031 ] ; c. tyler , a. v. olinto and g. sigl , phys .
d * 63 * , 055001 ( 2001 ) [ arxiv : hep - ph/0002257 ] .
j. polchinski , string theory , vol .
i and ii , cambridge univ . press , cambridge ( 1998 ) .
g. shiu and s.h .
tye , phys .
* d58 * , 106007 ( 1998 ) ; l.e .
ibanez , r. rabadan , and a.m. uranga , nucl .
* b542 * , 112 ( 1999 ) ; i. antoniadis , c. bachas , and e. dudas , nucl .
* b560 * , 93 ( 1999 ) ; k. benakli , phys .
rev . * d60 * , 104002 ( 1999 ) ; e. accomando , i. antoniadis and k. benakli , nucl .
* b579 * , 3 ( 2000 ) .
j. j. friess , t. han and d. hooper , phys .
b * 547 * , 31 ( 2002 ) [ arxiv : hep - ph/0204112 ] .
d. amati , m. ciafaloni and g. veneziano , phys .
b * 197 * , 81 ( 1987 ) ; d. amati , m. ciafaloni and g. veneziano , phys .
b * 289 * , 87 ( 1992 ) ; p. c. argyres , s. dimopoulos and j. march - russell , phys .
b * 441 * , 96 ( 1998 ) [ arxiv : hep - th/9808138 ] ; t. banks and w. fischler , arxiv : hep - th/9906038 .
for a recent review on black holes in theories with extra dimensions , see _
e.g. _ , p. kanti , arxiv : hep - ph/0402168 .
p. d. death and p. n. payne , phys .
d * 46 * , 658 ( 1992 ) ; phys .
d * 46 * , 675 ( 1992 ) ; phys .
d * 46 * , 694 ( 1992 ) ; d. m. eardley and s. b. giddings , phys . rev .
d * 66 * , 044011 ( 2002 ) [ arxiv : gr - qc/0201034 ] ; s. n. solodukhin , phys .
b * 533 * , 153 ( 2002 ) [ arxiv : hep - ph/0201248 ] .
l. a. anchordoqui , j. l. feng , h. goldberg and a. d. shapere , arxiv : hep - ph/0311365 ; h. yoshino and y. nambu , phys .
d * 67 * , 024009 ( 2003 ) [ arxiv : gr - qc/0209003 ] .
e. j. ahn , m. cavaglia and a. v. olinto , arxiv : hep - ph/0312249 .
p. kanti and j. march - russell , phys .
d * 66 * , 024023 ( 2002 ) [ arxiv : hep - ph/0203223 ] ; phys .
d * 67 * , 104019 ( 2003 ) [ arxiv : hep - ph/0212199 ] .
t. han , g. d. kribs and b. mcelrath , phys .
lett . * 90 * , 031601 ( 2003 ) [ arxiv : hep - ph/0207003 ] .
m. kowalski , a. ringwald and h. tu , phys .
b * 529 * , 1 ( 2002 ) [ arxiv : hep - ph/0201139 ] .
for a recent review on @xmath4-brane production , see _
e. g. _ , m. cavaglia , int .
j. mod .
a * 18 * , 1843 ( 2003 ) [ arxiv : hep - ph/0210296 ] .
l. a. anchordoqui , j. l. feng and h. goldberg , phys .
b * 535 * , 302 ( 2002 ) [ arxiv : hep - ph/0202124 ] .
s. dimopoulos and r. emparan , phys .
b * 526 * , 393 ( 2002 ) [ arxiv : hep - ph/0108060 ] .
g. t hooft , phys .
lett . * 37 * , 8 ( 1976 ) ; g. t hooft , phys .
d * 14 * , 3432 ( 1976 ) [ erratum - ibid .
d * 18 * , 2199 ( 1978 ) ] ; a. ringwald , nucl .
b * 330 * , 1 ( 1990 ) ; v. v. khoze and a. ringwald , phys .
b * 259 * , 106 ( 1991 ) .
f. bezrukov , d. levkov , c. rebbi , v. rubakov and p. tinyakov , phys .
b * 574 * , 75 ( 2003 ) [ arxiv : hep - ph/0305300 ] ; f. bezrukov , d. levkov , c. rebbi , v. rubakov and p. tinyakov , phys .
d * 68 * , 036005 ( 2003 ) [ arxiv : hep - ph/0304180 ] .
p. arnold and l. mclerran , phys .
* d37 * , 1020 ( 1988 ) ; a. ringwald , nucl .
b330 * , 1 ( 1990 ) ; o. espinosa , nucl . phys . *
b343 * , 310 ( 1990 ) ; v.a .
khoze and a. ringwald , phys . lett . *
b259 * , 106 ( 1991 ) .
i. albuquerque , g. burdman and z. chacko , phys .
lett . *
221802 ( 2004 ) [ arxiv : hep - ph/0312197 ] . j. l. feng , a. rajaraman and f. takayama , phys .
lett . * 91 * , 011302 ( 2003 ) [ arxiv : hep - ph/0302215 ] ; j. l. feng , s. su and f. takayama , arxiv : hep - ph/0404231 . c. barbot , m. drees , f. halzen and d. hooper , phys .
b * 563 * , 132 ( 2003 ) [ arxiv : hep - ph/0207133 ] . | we discuss the prospects for high - energy neutrino astronomy to study particle physics in the energy regime comparable to and beyond that obtainable at the current and planned colliders .
we describe the various signatures of high - energy cosmic neutrinos expected in both neutrino telescopes and air shower experiments and discuss these measurements within the context of theoretical models with a quantum gravity or string scale near a tev , supersymmetry and scenarios with interactions induced by electroweak instantons .
we attempt to access the particle physics reach of these experiments .
madph-04 - 1392 + hep - ph/0408348 + + + + + |
with the discovery of the first inorganic spin - peierls compound cugeo@xmath0 @xcite it has become possible to investigate the physics of the spin - peierls transition in quasi one - dimensional spin chains with high precision . of particular interest has been the study of the magnetic excitation spectrum by neutron scattering @xcite and raman scattering experiments @xcite .
it has been observed @xcite , that the spin - peierls gap , i.e. the gap to triplet excitations out of the singlet ground - state , has a temperature dependence which is difficult to explain with existing theories of the spin - peierls transition @xcite .
existing theories of the spin - peierls transition of spin-1/2 heisenberg chains are based on the mapping of the spin chain to an interacting gas of spin - less fermions via the jordan - wigner transformation @xcite .
the chains are coupled to a three - dimensional phonon mode and the transition occurs when the respective phonon frequency becomes soft . in the spin - peierls state
the unit - cell doubles due to the alternating displacements of the magnetic ions .
the effective hamiltonian for the magnetic excitations along the chains in the dimerized state is then @xcite @xmath2 . \label{h}\ ] ] the nearest neighbor ( nn ) exchange has alternating strength @xmath3 due to the doubling of the unit cell below the spin - peierls transition temperature @xmath4 .
also included in ( [ h ] ) is a frustrating next - nearest - neighbor ( nnn ) term @xmath5 .
the ground - state phase diagram of ( [ h ] ) has been mapped out by a density matrix renormalization group study @xcite .
there is a gap everywhere in the phase - diagram except on the line @xmath6 and @xmath7 , with @xmath8 .
a vividly debated question of key interest is therefore the actual magnitude of the strength of the frustration @xmath9 .
for cugeo@xmath0 the parameters have been estimated as @xmath10 and @xmath11 @xcite .
a consequence of existing theories of the spin - peierls transition is the occurrence of a soft phonon mode above @xmath4 , which has not yet been observed for cugeo@xmath0 @xcite .
it is therefore of interest to pursue the possibility of a purely electronically driven spin - peierls transition , as it could occur for @xmath12 .
in such a scenario the phonons would just follow the pre - formed electronic dimerization , i.e. the formation of nn singlet pairs . here
we propose a simple molecular - field type theory for the spin - peierls transition by mapping the low - energy solitonic excitations of dimerized spin chains on an effective ising model .
we emphasize that this procedure is valid both for the phonon - driven and for the electronically - driven spin - peierls transition .
we present results of the mean - field theory in comparison with experiments on cugeo@xmath0 .
we find that a straightforward determination of the frustration parameter @xmath13 is not conclusive and that cugeo@xmath0 is close to a first - order phase transition .
we also note that the famous @xmath14 scaling of the triplet gap @xcite is a simple consequence of the linear binding potential between solitons and anti - solitons in dimerized spin chains .
a strictly one - dimensional system shows no long - range order at finite temperatures due to the solitonic excitations .
the appropriate solitons in a spin - peierls state are domain - walls in between two different dimer coverings , see fig .
[ dimerchain ] . in this picture
a dimer consists of a nn pair of spins in singlet state . a single soliton in an otherwise ( dimer- ) ordered chain is in reality a complicated object @xcite .
it is spatially extended @xcite and has a spin degree of freedom together with a dispersion , which in the majumdar - gosh model ( @xmath15 @xcite ) has the form @xmath16 .
it has been shown @xcite , that the solitons yield an accurate description of the magnetic susceptibility and hence of the hilbert space .
the dispersion of the solitons might be approximated , in general , by @xmath17 where @xmath18 is the gap to solitonic excitations , @xmath19 the exchange integral , @xmath20 a constant of order of unity , and @xmath21 is the wavenumber in units of inverse lattice spacings .
only the low - energy solitons are effective in destroying the long - range dimer order , and we approximate here ( [ dispersion ] ) by a constant @xmath22 .
we will furthermore not discuss the consequence of the spatial extent of the solitons and we will include their spin - degree of freedom only when determining the susceptibility .
we take one of the two possible dimer configurations of the linear chain as the reference state and label every `` good '' dimer by @xmath23 and every bond between two `` wrong '' dimers by @xmath24 ; a bond between a soliton and a wrong dimer is also labeled @xmath24 as shown in fig . [ dimerchain ] .
every possible soliton configuration is therefore mapped to a ising - spin configuration living on every second bond . a similar mapping has been used recently by mostovoy and khomskii @xcite .
= 0.48 we now consider a two - dimensional array of chains and define by @xmath25 the ising variable on the @xmath26 chain , where the site index @xmath27 runs over every second bond of a chain of length @xmath28 .
including a coupling @xmath29 between dimers on nn chains we have @xmath30 as the effective ising hamiltonian .
the inter - chain coupling @xmath29 might be induced by the inter - chain exchange coupling , @xmath31 . in this case
. an indirect coupling of the chains via phonons is also conceivable @xcite .
it has been shown recently , that a molecular - field decoupling of inter - chain interactions in quasi one - dimensional systems is a good approximation in the strongly anisotropic limit @xcite .
we may therefore decouple the chains via @xmath33 where the @xmath34 denotes the thermodynamic expectation value .
translational invariance perpendicular to the chains yields @xmath35 and the hamiltonian for a single chain ( the number of ising variables is @xmath36 ) becomes in this mean - field approximation @xmath37 where we have set @xmath38 .
this is just the ising hamiltonian for a ferromagnetic spin chain in an external magnetic field @xmath39 .
alternatively we can interpret ( [ ising ] ) as an effective model for domain walls in a spin - peierls state with a linear binding potential @xmath40 ( with @xmath41 being the distance in sites , not bonds ) between a soliton and an anti - soliton .
so far we have neglected the spin - phonon coupling @xmath42 .
a finite value for the spin - singlet order parameter @xmath43 leads through @xmath44 to a finite lattice dimerization and consequently to a modulation of the exchange constant , @xmath45 , in ( [ h ] ) , with @xmath46 .
the spin - phonon coupling therefore adds a term @xmath47 to the confining potential @xmath39 . on a mean - field level
this corresponds to a rescaling of the coupling constant @xmath48 with appropriate constants @xmath49 and @xmath50 .
the partition function @xmath51 can be obtained from ( [ ising ] ) by the transfer matrix method . the free energy @xmath52 is given in the thermodynamic limit ( @xmath53 ) by @xmath54 where @xmath55 with @xmath56 being the inverse temperature . from the free energy all physical quantities can be derived .
we have two relevant excitation energies , the gap to single soliton excitations , given by @xmath18 in ( [ ising ] ) , and the gap to triplet excitations , @xmath57 , as measured in a neutron scattering experiment @xcite .
let us now discuss the relation of @xmath58 to @xmath59 .
a triplet excitation can dissolve into a soliton / anti - soliton pair .
the linear binding energy @xmath40 in between a soliton and an anti - soliton in ( [ ising ] ) leads to a confinement of soliton / anti - soliton pairs @xcite . in ( [ ising ] )
we have neglected the kinetic energy of the single solitons ( [ dispersion ] ) , which is of order @xmath60 . the energy levels of a particle in a linear confining potential are well known @xcite .
the lowest eigenstate has , in the limit @xmath61 , the energy @xmath62 with @xmath63 .
( [ delta_n ] ) gives then the gap to triplet excitations as a function of soliton energy @xmath18 and the strength of the confining potential @xmath39 .
the mean extension of a soliton / anti - soliton pair scales like @xmath64 .
for @xmath65 we have @xmath66 and eq .
( [ delta_n ] ) takes then the form @xmath67 where @xmath68 is the @xmath69-dependent spin - wave velocity @xcite .
it is known @xcite , that the gap scales for @xmath65 like @xmath14 implying that @xmath70 for @xmath71 .
comparison with numerical results @xcite leads to @xmath72 .
the functional form of the dependence of the soliton excitation energy @xmath18 on @xmath73 , or alternatively on the dimerization order parameter @xmath43 is not known at present .
it might be extracted for @xmath7 , in principle , from a sub - leading scaling analysis of the excitation energy @xmath74 but this has not yet been done .
we have therefore decided to assume the functional form @xmath75 where @xmath76 is the zero - temperature value of @xmath18 and where @xmath77 is the soliton energy in the disordered phase , i.e. for @xmath78 .
( [ e_s ] ) is , in the spirit of a landau functional ( [ landau ] ) , the simplest form consistent with the symmetry of @xmath79 .
we may expand the free energy ( [ free ] ) in powers of @xmath79 , @xmath80 depending on the parameters we may have either a second - order phase transition with @xmath81 or a first - order phase transition with @xmath82 . in the first case the transition temperature @xmath83 is given by @xmath84 as @xmath85 with @xmath86 .
this transcendental equation takes a simple form in some limiting cases : @xmath87 for illustration we present in fig .
[ phases ] ( a ) @xmath83 as a function of the inter - chain coupling constant @xmath29 for @xmath88 .
( 0.49,0.80 ) = 0.47(0,0.40 ) = 0.47(0,0 ) ( 0,0.42)(a ) ( 0,0.02)(b ) for @xmath89 we obtain a first - order phase transition .
this is the case for values of @xmath76 larger than a certain critical value of the soliton energy @xmath90 , which is determined from @xmath91 as @xmath92 for a fixed transition temperature @xmath93 and using eq .
( [ t_sp ] ) , @xmath94 and the corresponding inter - chain coupling energy @xmath29 can be calculated as a function of @xmath95 .
the resulting phase diagram is given in fig.[phases](b ) .
the numerical results presented throughout this paper are obtained within the second - order regime , indicated by the crosses in fig .
[ phases](b ) .
the effective hamiltonian eq .
( [ ising ] ) contains three free parameters , namely @xmath96 , @xmath97 and @xmath98 .
we examine two scenarios .
the first is the case of @xmath99 with @xmath100 .
the second is the case of @xmath101 with @xmath102 corresponding to a gap in the disordered phase of @xmath103 @xcite . for each case
we consider a range of @xmath76 ( see fig.[phases](b ) ) within the second - order regime , @xmath104 , @xmath105 @xmath106 the largest value of @xmath76 for each @xmath77 corresponds to @xmath107 , compare fig .
[ phases](b ) .
the experimental transition temperature @xmath108 of cugeo@xmath0 then determines the coupling constant @xmath109 @xcite .
the order parameter @xmath43 is determined self - consistently as a function of temperature by setting the derivative of the free energy with respect to @xmath43 to zero , @xmath110 in fig .
[ distortion ] we show the results for @xmath111 as a function of temperature for the parameters given in ( [ values1 ] ) and ( [ values2 ] ) .
we have also plotted the measured intensity of an additional super - lattice peak @xcite , which is proportional to the square of the lattice dimerization .
we have normalized the experimental data such that agreement is obtained in the low - temperature regime .
the comparison between theory and the data for cugeo@xmath0 does not lead to a determination of the frustration @xmath13 but indicates closeness to a first - order phase transition , as can be deducted from the closeness of the experimental points in fig .
[ distortion ] to the critical curves where @xmath112 , compare fig .
[ phases](b ) .
( 0.49,0.84 ) = 0.48(0,0.42 ) = 0.48(0,0 ) ( 0,0.44)(a ) ( 0,0.02)(b )
the entropy @xmath113 and the specific heat @xmath114 are obtained from the free energy ( [ free ] ) via @xmath115 in the disordered phase the entropy is ( for @xmath116 ) temperature independent with a value of @xmath117 per site , which is only half of the expected value for a spin @xmath118 chain .
this is due to the fact that we did neglect up to now the spin - degrees of freedom of the domain - walls . in fig .
[ cv ] we present results for @xmath119 for the parameters given by ( [ values1 ] ) and ( [ values2 ] ) .
( 0.49,0.84 ) = 0.48(0,0.42 ) = 0.48(0,0 ) ( 0,0.44)(a ) ( 0,0.02)(b ) for small soliton excitation energies the results are typically mean - field like . for values of @xmath76 approaching the limit of the second - order phase regime the specific heat is strongly enhanced near @xmath120 .
it will diverge as the transition becomes first - order .
note that the specific heat is linear in temperature for @xmath121 , in the limit @xmath122 and that the jump in the specific heat diverges as @xmath123 like @xmath124 .
right at @xmath125 the specific heat diverges like @xmath126 . note that a similar divergence @xmath127 has been reported in an early measurement for cugeo@xmath0 @xcite though the exact value of the specific heat critical exponent for cugeo@xmath0 is still controversial @xcite . in the inset of fig .
[ cv ] we present the measured magnetic contribution to the specific heat of cugeo@xmath0 @xcite .
a straightforward comparison with the results of the mean - field theory is not possible since all the entropy of the effective ising chain is released in a mean - field approach .
this corresponds to half of the entropy of the spin chain .
experimentally only @xmath128 of the magnetic entropy is released at @xmath120 @xcite , since the exchange constant @xmath129 and the measured specific heat is consequently smaller in magnitude than our mean - field result .
the neglect of the soliton dispersion relation ( [ dispersion ] ) is , on a microscopic level , the reason for this discrepancy between theory and experiment .
a qualitative comparison is nevertheless possible and favors an @xmath76 close to the first - order phase transition . up to know we did not take the spin - degree of freedom of the solitons into account , as they just contribute a constant factor to the partition function in the paramagnetic case . as we have no magnetic interactions between the spins of different solitons in our model we can evaluate the magnetic susceptibility simply by curie s law @xmath130 where @xmath131 is the bohr magneton , @xmath132 the measured @xcite g - factor of the @xmath133ion , @xmath134 and @xmath135 the density of thermally activated solitons per site .
@xmath135 is obtained differentiating the free energy with respect to @xmath136 : @xmath137 above @xmath120 this expression reduces to @xmath138 the mean number of solitons attains @xmath139 per site , as the temperature goes to infinity , corresponding to half a soliton per dimer .
the results for the magnetic susceptibility are shown in fig .
[ chi ] . the fast drop of @xmath140 below @xmath120 for larger values of the soliton excitations energies is again reminiscent of the experimental data for cugeo@xmath0 @xcite .
our susceptibility rises though much higher at @xmath83 than the experimental data which is a direct consequence of the neglected soliton dispersion ( [ dispersion ] ) . in the limit of large temperatures
@xmath141 the theoretical curve drops to @xmath139 of the experimental points as a consequence of the aforementioned soliton density ( [ n_t ] ) ( at @xmath142 it has dropped to about @xmath118 of the magnitude of the experimental data ) .
note that for the curves plotted as solid lines in fig .
[ distortion ] , fig .
[ cv ] , and fig .
[ chi ] , which are closest to the experimental data within the parameters chosen by us , entropy , specific heat , and susceptibilty are overestimated by about the same factor of five .
( 0.49,0.84 ) = 0.48(0,0.42 ) = 0.48(0,0 ) ( 0,0.44)(a ) ( 0,0.02)(b )
the gap to triplet excitations is given by eq .
( [ delta_n ] ) , @xmath143 with @xmath144 .
a straightforward application of ( [ triplet_gap ] ) would yield , compared with experiment , a much too large zero - temperature gap @xmath145 .
this is so since the order parameter @xmath146 takes the value one in the molecular - field approximation , while it is much smaller than unity for cugeo@xmath0 .
we have therefore decided to use the parameter @xmath147 in ( [ triplet_gap ] ) to fix @xmath148 to the experimentally observed value @xmath149 .
the results are given in fig .
[ gap ] , together with the measured gap for cugeo@xmath0 @xcite .
( 0.49,0.84 ) = 0.48(0,0.42 ) = 0.48(0,0 ) ( 0,0.44)(a ) ( 0,0.02)(b )
we have discussed a simple mean - field theory for spin - peierls transitions applicable both for phonon - driven ( @xmath150 ) and for magnetically driven ( @xmath151 ) spin - peierls transitions .
we have applied the approach to cugeo@xmath0 and found that it is not possible to determine uniquely from the experimentally measured temperature dependence of the order - parameter the magnitude of the frustration parameter @xmath69 .
the theory allows both for a first - order and a second - order spin - peierls transition , depending on the parameters of the model .
we find that the parameters which fit experiments best indicate that cugeo@xmath0 is close to a first - order phase transition .
we would like to thank p. lemmens and b. bchner for discussions and j. stolze for pointing out a straightforward derivation of the @xmath152 scaling law . the support of the dfg is gratefully acknowledged .
k. fabricius , a. klmper , u. lw , b. bchner , and t. lorenz , _
`` reexamination of the microscopic couplings of the quasi one - dimensional antiferromagnet cugeo@xmath0 '' _ , sissa - preprint cond - mat/9705036 .
v. kiryukhin , b. keimer , j.p . hill , and a. vigilante , phys .
76 * , 4608 ( 1996 ) ; j. zang , s. chakravarty and a.r .
bishop , sissa - preprint cond - mat/9702185 ; a. dobry and j.a .
riera , sissa - preprint cond - mat/9704090 .
the complete sets of parameter values @xmath157 are in [ mev ] : ( 0,0,1.22 ) , ( 0.1,0,1.31 ) , ( 0.2,0,1.39 ) , ( 0.277,0,1.45 ) for @xmath150 and ( 0.15,0.15,1.05 ) , ( 0.3,0.15,1.2 ) , ( 0.4,0.15,1.28 ) , ( 0.474,0.15,1.33 ) for @xmath151 . | we present a theory for the spin - peierls transition in cugeo@xmath0 .
we map the elementary excitations of the dimerized chain ( solitons ) on an effective ising model .
inter - chain coupling ( or phonons ) then introduces a linear binding potential between a pair of soliton and anti - soliton , leading to a finite transition temperature .
we evaluate , as a function of temperature , the order parameter , the singlet - triplet gap , the specific heat , and the susceptibility and compare with experimental data on cugeo@xmath0 .
we find that cugeo@xmath0 is close to a first - order phase transition .
we point out , that the famous scaling law @xmath1 of the triplet gap is a simple consequence of the linear binding potential between pairs of solitons and anti - solitons in dimerized spin chains . |
in this work we will describe the properties of a still oversimplified submonolayer thin film growth model which includes five important physical ingredients for these systems : \(1 ) _
deposition_. we will assume that atoms are deposited at randomly - chosen positions of the surface at a flux @xmath0 per unit surface per unit time .
atoms that happen to fall on the islands that are formed on the surface are accomodated at their proper pyramidal position ( see below ) .
\(2 ) _ diffusion_. isolated adatoms can move in a random direction by one diameter , or one lattice spacing , which we will take as our unit length .
we denote by @xmath1 the characteristic time between diffusion steps and @xmath2 the diffusion coefficient ( the atomic size is taken as the length unit ) .
\(3 ) _ evaporation_. isolated adatoms can evaporate off the surface at a constant rate .
we denote by @xmath3 the mean lifetime of a free adatom on the surface .
it is also useful to define @xmath4 , the mean adatom diffusion length before desorption \(4 ) _ aggregation . _
if two adatoms come to occupy neighboring sites , they stick irreversibly and form an island . as more
adatoms are captured , the island develops as a pyramid ( see below ) .
islands are assumed to be immobile and do not evaporate .
\(5 ) _ defect trapping _ in some parts of the paper , we introduce a concentration @xmath5 of `` defects '' on the surface . these defects , which are randomly distributed on the surface , act as perfect traps for the monomers .
therefore , a monomer which happens to occupy a defect remains there forever and serves as a nucleation center for island growth . in the following , we call _ particles _ or _ adatoms _ the isolated atoms ( or monomers ) that are deposited on the surface , and _
islands _ a set of connected particles ( thus excluding the monomers ) . some remarks on the assumptions of this simple model regarding its connection to the experiments are now addressed .
\(a ) _ island morphology_we force the islands to assume a pyramidal shape by immediately moving an adatom that touches an island to the desired position ( fig .
[ pyramide ] ) .
the pyramidal shape is adopted because we wish to understand experiments carried out by one of us ( mm ) on the system pd / nacl @xcite where the islands take approximately such a shape .
we note that this assumption does not affect crucially the growth : it should not affect the growth exponents which are only determined by the fact that islands are three - dimensional ( i.e. their mass increases as their radius to the third power , see below ) .
\(b ) _ island diffusion_we neglect in this model the possibility for dimers , trimers or larger islands to diffuse on the substrate .
island diffusion has been observed in experiments @xcite and molecular - dynamics simulations @xcite .
the effects of 2d island diffusion on the growth of thin films _ without evaporation _ have been addressed in refs .
before going into the details of the calculations and their confirmation by computer simulations , we present a simple picture of the growth mechanisms of the submonolayer structures under consideration .
we are interested mainly on two points : the time evolution of the island concentration on the substrate and the island concentration at saturation as a function of the growth parameters : flux , diffusion and evaporation times and defect concentration .
the qualitative evolution of the system is essentially as follows .
the system initiates as a clean empty surface .
monomers are then deposited at a constant rate on the surface and are allowed to diffuse and/or desorb ( evaporate ) .
when two monomers meet , they aggregate irreversibly to form a static island ( an island is also created when a monomer is trapped by a defect ) : this is island _ nucleation_.
as more of these encounters occur , the island density increases with time . competing with this nucleation process
, monomers are also captured by islands which then become larger . at some point , islands are so large that they quickly capture the free monomers , which reduces the chances of two monomers meeting to nucleate a new island
. therefore , the number of islands grows less rapidly .
moreover , when islands become large , they start touching ( coalescing ) .
these two effects lead to a saturation in the number of islands .
interestingly , the saturation is attained when the surface coverage reaches a value close to .15 , independently of the parameter values .
this is equivalent to saying that saturation is reached when the mean island radius @xmath6 is a fixed fraction of the island - island distance ( the coverage @xmath7 is given by @xmath8 where @xmath9 is the mean island - island distance ) .
a more detailed discussion of the evolution of the systems can be done by distinguishing several cases according to the relative importance of the different phenomena : diffusion , evaporation , defect concentration .
it is useful to define a typical length scale for each of these processes : @xmath10 is a typical distance between islands when evaporation and defect concentration are not important ( see below ) , @xmath4 is , as defined above , the adatom diffusion length before desorption and @xmath11 is the typical distance between defects . now , depending on the relative values of these three length scales , growth will be dominated by different mechanisms which we turn on to describe in more detail . note that @xmath12 corresponds to the typical island - island distance at the _ saturation _ time ,
i.e. it is not the actual island - island distance ( for example at the beginning of film growth ) .
therefore our following qualitative discussion is only approximated , and the more precise calculations of later sections are necessary to justify it . by `` dirty''substrate , we mean that the island concentration is affected by the defect concentration .
we will show that this is true only if @xmath13 is much larger than @xmath14 , i.e. the defect concentration is high enough , even in the absence of evaporation ( if evaporation is present , it can only decrease the number of islands , therefore increasing the relative importance of defect concentration ) .
the simplest case is when the defect concentration is very high , namely if @xmath14 is much smaller than _ both _ @xmath15 and @xmath13 . in this case ,
island nucleation is completely dominated by the trapping of adatoms by defects , which leads to two main effects : first , the island concentration reaches its saturation value rapidly ( roughly in a time @xmath16 ) ; second , this saturation value is @xmath17 , i.e. all defects are occupied by islands but there are no islands created elsewhere .
note that this situation corresponds to a `` low '' temperature case , when @xmath15 is large enough . at higher temperatures
, one could have @xmath18 ( but still @xmath19 to remain in the `` dirty '' substrate case ) . in this case
, the monomer concentration is dominated by evaporation and island nucleation still occurs on defects .
the saturation density is still equal to the defect concentration but the kinetics is different : the time needed to reach saturation is roughly @xmath20 .
here we study the cases for which the island concentration is not affected by the presence of defects , when the substrate is `` clean '' enough .
this is true when @xmath13 is much smaller than @xmath14 , irrespective of the @xmath15 value . in principle , three cases can then arise , depending on the relative value of @xmath15 against these two lengths
. our calculations will show ( section [ scal ] ) that there are only two asymptotic regimes : complete condensation and high evaporation . in the complete condensation ( cc )
case , adatoms do not evaporate from the surface and island growth proceeds mainly by capture of the diffusing adatoms . on the contrary , in the high evaporation limit ,
the number of adatoms is limited by evaporation and the most important growth mechanism for islands is that of direct impingement of adatoms on top of them ( the contribution from the adatoms diffusing on the surface is negligible ) .
the first case corresponds to @xmath21 , independently of the relative ordering of @xmath22 and @xmath15 .
then , adatoms almost never evaporate before aggregating ( after this , they are safe since islands do not evaporate ) .
the monomer density rapidly grows , leading to a rapid increase of island density by monomer - monomer encounter on the surface .
this goes on until the islands occupy a significant fraction of the surface , roughly 1% .
then , islands capture rapidly the monomers , whose density decreases . as a consequence
, it becomes less probable to create more islands , and we see that their number increases more slowly . when the coverage reaches a value close to 15% , coalescence will start to decrease the number of islands .
the maximum number of islands @xmath23 is thus reached for coverages around 15% . the second case corresponds to the opposite situation : @xmath24 , with still @xmath25 .
this happens when @xmath3 is small , for example at high temperatures . in this regime ,
evaporation significantly alters the growth dynamics .
the main point is that now the monomer density becomes roughly a _ constant _ ( @xmath26 ) , since it is now mainly determined by the balancing of deposition and evaporation .
then , the number of islands increases linearly with time ( the island creation rate is roughly proportional to the square monomer concentration ) .
we also notice that only a small fraction ( 1/100 ) of the monomers do effectively remain on the substrate , as shown by the low sticking coefficient value at early times ( the sticking coefficient is the ratio of particles on the substrate ( the coverage ) over the the total number of particles sent on the surface ( ft ) ) .
this can be understood by noting that an island of radius @xmath6 grows by capturing only the monomers that are deposited within its `` capture zone '' ( the circle of radius @xmath27 centered on island s center ) .
the other monomers evaporate before reaching the islands . as in the case of complete condensation ,
when the islands occupy a significant fraction of the surface , they capture rapidly the monomers .
this has two effects : the monomer density starts to decrease , and the sticking coefficient starts to increase .
shortly after , the island density saturates and starts to decrease because of island - island coalescence .
note that one could have in principle @xmath28 , i.e. the adatoms evaporate before diffusing .
this situation , although apparently uncommon , is not physically impossible and it also allows us to test our predictions over a larger range of parameters .
we note that , in contrast to what is observed for two dimensional ( 2d ) islands @xcite , particles deposited on top of islands significantly contribute to island growth . in presence of strong evaporation , this is actually the dominant mechanism for island growth , whereas for the 2d case this happens only in some special regimes @xcite .
we present here the summary of our results in the different limits described above .
these results are derived in detail in sections [ scal ] and [ equ ] . for each regime , we give in the order the saturation island density @xmath23 , the thickness at saturation @xmath29 ( i.e. the thickness when the island density first reaches its saturation value ) , the thickness at coalescence @xmath30 ( i.e. the thickness when the island density starts to decrease due to island - island coalescence ) , and the scaling kinetics of the mean radius as a function of time before the saturation island density is reached .
we recall that @xmath10 and @xmath31 . *
clean substrate * high evaporation : @xmath32 @xmath33^{2/3}$ ] @xmath34^{-1/3}$ ] @xmath35 low evaporation : @xmath36 or @xmath37 @xmath38 @xmath39 @xmath40 * dirty substrate * high evaporation : @xmath41 @xmath42 @xmath43}$ ] @xmath44 @xmath45 low evaporation : @xmath46 or @xmath47 @xmath42 @xmath48 @xmath44 @xmath49 for @xmath50 , i.e. before saturation @xmath51 between saturation and coalescence ( @xmath52 ) , see section [ equ ] . in this section
we present simple scaling arguments that allow to find the dependence of the maximum island density @xmath53 as a function of the deposition parameters ( flux f , diffusion time @xmath1 and evaporation time @xmath3 ) .
these arguments were originally formulated in @xcite for the special case of growth of 2d islands by atom deposition on a high - symmetry terrace , neglecting evaporation . here , the argument is extended to the case of non - negligible evaporation .
we recall that the atomic size is taken as the length unit .
the first stage of the argument requires the determination of the nucleation rate per unit surface and time , @xmath54 .
a nucleation event takes place when an adatom meets a critical island of size @xmath55 .
we call @xmath56 the density of critical nuclei ( clusters of size @xmath57 ) and , following refs
. @xcite , we assume that @xmath56 satisfies walton s relation @xmath58 .
thus , @xmath59 where @xmath2 is the adatom diffusion constant .
the term @xmath60 accounts for direct impingement of atoms onto critical islands and the second term for critical island growth by monomer diffusive attachement .
another , independent equation can be written down to relate the nucleation rate and the stable ( @xmath61 ) island density @xmath62 .
it states that in the area @xmath63 occupied by an island , only one ( on average ) nucleation event takes place , _ during the time @xmath64 needed for the growing islands to come into contact_. thus , @xmath65 the time @xmath64 is readily computed by knowing the growth velocity of an island , which in turn requires the knowledge of the adatom density .
we consider in the following three situations of interest for this paper . . the adatom density results in this case from a balance between deposition at a rate @xmath0 and capture by the stable islands at a rate @xmath66 , so that @xmath67 the rate of growth of the volume of an island of linear size @xmath6 is diffusion - limited in this case , so that @xmath68 and @xmath69 . at @xmath70 , @xmath71 , and thus @xmath72 using ( [ nuc1 ] ) and ( [ n1 ] )
, one finds @xmath73^{s^*+1}\ ] ] ( here , the direct impingement term is negligible ) . from ( [ nuc2 ] ) and ( [ tc5 ] ) , @xmath74 finally , ( [ rel4 ] ) and ( [ nucc7 ] ) yield @xcite @xmath75 .
[ strong ] strong evaporation means the adatoms are more likely to disappear due to desorption than to be captured by an island . in other words ,
the adatom diffusion length before desorption , @xmath4 , is shorter than the average island - island distance , @xmath76 . in this case , the adatom density results from a balance between deposition and desorption at a rate @xmath77 , so that @xmath78 a 3-d island grows by two mechanisms in the case of strong evaporation : ( i ) by capture of the adatoms falling on the surface at a distance smaller than @xmath15 ; ( ii ) by direct capture of all adatoms falling on it .
thus , @xmath79 .
solving for @xmath6 with the condition @xmath80 at @xmath81 , one gets @xmath82 . at @xmath70 , @xmath83 , which means that direct capture always dominates .
it follows @xmath84 , or @xmath85\;.\ ] ] using ( [ nuc1 ] ) and ( [ n2 ] ) , one finds @xmath86 from ( [ nuc2 ] ) and ( [ tc6 ] ) , @xmath87 finally , ( [ rel5 ] ) and ( [ nucc8 ] ) yield @xmath88 we discuss the preceding results in section [ discussion ] .
we will show in this section that it is possible to derive interpolation formulae describing the crossovers between the no - evaporation and the strong - evaporation regimes . to do this
, we will use burton , cabrera and frank s theory of step flow @xcite .
the adatom density will be computed on a terrace bounded by two parallel steps , at a distance @xmath89 .
they may be the steps in the orderd array of a vicinal surface ; in this case , we will let @xmath90 .
or they may represent the edges of a big enough island ; in the latter case we will let @xmath91 .
the adatom density obeys the equation @xmath92 the time @xmath1 is the average lifetime of an adatom .
since adatoms disappear either by capture by islands , or by desorption , we will alternatively let @xmath93 or @xmath94 . in both cases , the notation @xmath95 will be used . in the quasi - stationary approximation @xcite , @xmath96 , and choosing the origin @xmath97 at the terrace centre , equation ( [ bcf ] )
can be solved with the conditions @xmath98 at boundary steps .
the solution reads @xmath99 \;.\ ] ] this formula will be needed to compute the nucleation rate ( [ nuc1 ] ) .
the latter is an average quantity , independent of @xmath100 .
we let thus @xmath97 in ( [ sol1 ] ) , since the higher nucleation probability is at the terrace centre , given the symmetry of our problem .
one finds @xmath101 = 2f\tau \frac{\sinh^2(\kappa \ell/4)}{\cosh(\kappa \ell/2)}\;,\ ] ] where we used the identity @xmath102 .
the next task is the determination of the island density . to this end
, it suffices to consider the total island density @xmath62 .
its time variation is simple : @xmath62 increases each time a new island is nucleated , so @xmath103 . on the other hand
, @xmath62 decreases when two islands touch and coalesce . following the authorities @xcite , we write @xmath104 , where @xmath105 is the average area of an island of linear size @xmath6 .
this means that coalescence results from binary encounters of immobile islands , whose area increases at a rate @xmath106 .
collecting @xmath107 and @xmath108 yields @xmath109 at stationarity , which is what we are interested in , @xmath110 .
note that , by definition , coalescence takes place when @xmath111 .
one can thus write @xmath112 the final task concerns therefore the evaluation of the growth rate of an island .
this can be done by noting that the mass @xmath113 of an island increases either by direct capture of atoms from the beam , or by diffusion of adatoms on the surface .
since the surface diffusion current of adatoms is @xmath114 , assuming circular ( @xmath115 ) or hemispherical ( @xmath116 ) islands of radius @xmath6 one can write @xmath117 the result can be found in a number of papers @xcite .
it makes use of bessel functions , which are not easy to manipulate .
approximate results , which have the merit of reproducing the correct limiting regimes ( see below ) , and of allowing analytical expressions to be written down , will be used .
the adatom density is still given by eq.([sol2 ] ) , and the nucleation rate by eq.([nuc1 ] ) .
the growth rate of the ( projected ) area of a 3-d island has two contributions , as for 2-d islands : a diffusion - limited contribution , which is given by @xmath118 and a direct capture contribution , @xmath119 a useful interpolation formula between ( [ gr1 ] ) and ( [ gr2 ] ) is @xmath120 finally , ( [ nuc2 ] ) and ( [ da3 ] ) yield @xmath121^{s^*+1}\approx f\ell_s\left(1+\frac{x_s}{\ell_s}\right)\tanh(\kappa\ell_s/2 ) n^2 \;,\ ] ] or , @xmath122^{s^*+1 } \tanh(\kappa n^{-1/2}/2)(1+x_sn^{1/2 } ) n^{3/2 } \;.\ ] ] multiplying both sides by @xmath123 yields @xmath124 where @xmath125^{s^*+1 } \tanh\left(\frac{1}{2x}\right)(1+x ) x^{3}\;.\ ] ] letting @xmath126 , formula ( [ sche14 ] ) can be cast in the form @xmath127 where @xmath128^{2/3}$ ] .
inverting @xmath129 yields the crossover scaling function @xmath130 the function @xmath131 has the following properties : @xmath132 in this section we study the growth in presence of defects using rate equations @xcite .
we will consider that the islands are semi - spherical `` droplets '' as a result of `` fast '' large scale reacomodation of the monomers forming the island .
in addition to evaporation , we will also consider the effects due to the presence of point defects on the surface .
we will assume that these defects serve as perfect nucleation points , and so , as the concentration of defects increases , the system passes from homogeneous to heterogeneous nucleation .
we can write the evolution of the density @xmath133 of monomers on the surface as : @xmath134 the first term on the rhs denotes the flux of monomers onto the surface .
the second term represents the effect of evaporation .
the third term is due to the possibility of losing monomers by effect of direct impingement of a deposited monomers right beside a monomer still on the surface to form an island .
the next terms represent the loss of monomers by aggregation with other monomers , nucleation on defects and aggregation with islands respectively .
the factors @xmath135 and @xmath136 are the `` cross sections '' for encounters and are detailed below .
the number @xmath62 of islands will be given by : @xmath137 where the first term represents the formation of islands due to direct impingement of deposited monomers onto other monomers or defects , and the second term accounts for the formation of islands by the encounter of two monomers diffusing on the surface , or a monomer on a defect .
it should be noted that since this description yields scaling results , some numerical factors are omitted .
the concentration of `` free defects '' varies as @xmath138 the total island mass density ( @xmath139 ) changes as @xmath140 + \sigma_i n + \sigma_o c_{def } + fr^2n , \eqno(5)\ ] ] where the direct impingement term is proportional to the area of the islands .
the typical island radius @xmath6 will be given by @xmath141 and the actual island coverage will then be @xmath142 the expressions for the diffusive cross sections in the presence of evaporation were calculated in @xcite .
@xmath143 where @xmath4 is the typical distance traveled by a diffusing particle on the surface before it desorbs . and : @xmath144 these mean - field cross sections only depend on the radius of the sphere touching the surface , and not on the height of the island , therefore they are the same for 2d or 3d particles of same radius . as the most we can expect to obtain from this description are the scaling behaviors , we focus only on the extreme limiting cases of the system described by the rate equations . for the case of negligible evaporation and high enough initial concentration of defects ,
we expect the equations to reduce to @xmath145 @xmath146 @xmath147 adding the second and last equations , we have : @xmath148 for high enough initial concentration of defects @xmath5 , we can write @xmath149 @xmath150 @xmath151 the island density at saturation is therefore equal to the initial defect concentration @xmath5 . to find what is `` high enough '' for the defect concentration
, we note that nucleation events start to reduce the number of islands roughly when the island size is of the order of the distance between islands , or equivalently , when the coverage reaches a constant value .
since the deposited mass grows as @xmath152 ( evaporation is negligible ) , the surface coverage will be given by @xmath153 thus , the coalescence time is @xmath154 and the cross over to the `` clean system '' occurs when @xmath155 it should be stressed that in this defect dominated , or `` dirty '' , regime the island number density saturates far before coalescence , indeed the saturation time @xmath156 can be estimated by the characteristic time for defect ocupation : @xmath157 .
after @xmath156 , the island density remains essentially constant and the islands grow ( the typical island radius can be easily shown to grow as @xmath51 ) until coalescence . at high evaporation rates the aggregation of mass on the surface is dominated by `` direct impingement '' of particles on the islands , and the concentration of monomers on the surface is determined by the balance between deposition and evaporation . under these conditions , at high initial defect concentration we will have : @xmath158 @xmath159 @xmath160 these equations can be solved immediately , from which we get : @xmath161ft]}\ ] ] and @xmath162 ft ] } ] + f^2\tau_e[x_s^2 + 1 ] t,\ ] ] where we have substituted @xmath163 . from these expressions
it is evident that the time at which n saturates is @xmath164)$ ] , after which time , the island density reaches the value @xmath165 ( this , of course , assuming that the homogeneous contributions are negligible throughout the evolution of the system ) .
it is then straight forward to find that the mass deposited on the system grows as @xmath166 and the typical island radius @xmath45 .
the coverage increases as @xmath167 , from which we can estimate the coalescence time @xmath168 .
while it can be imagined that these times could occur in the wrong order , we have checked that for this to be the case , an initial concentration of defects larger than one would be required . by comparing the maximum value of the subdominant homogeneous term with @xmath5 , we can determine that this high evaporation dirty regime is attained when @xmath169)^{2/3}$ ] . finally ,
for completeness , we sketch the derivation of the results for the `` clean '' substrate with negligible evaporation . under these circumstances , the monomer density on the surface is determined by the balance between deposition and diffusive capture by the islands on the surface .
thus , the monomer and island densities are given by @xmath170 from these equations we find that the island density is @xmath171 .
also , as we are considering the case of negligible evaporation , the mass deposited on the surface will be @xmath172 . from these quantities
we can infer the behavior of the typical island radius to be given by @xmath173 , and the saturation and coalescence times to be @xmath174 , at which times the maximum value of n is @xmath175 , as obtained in the previous section .
in the following paragraphs , we test the assumptions and predictions of the analysis given in the preceding sections . we also show results that are not attainable from this mean - field calculations , namely the island size distributions .
our computer simulations generate sub - monolayer structures using the four processes included in our model ( see the introduction ) . here
we take @xmath176 as the time scale of our problem .
the monomer diffusion coefficient is then given by @xmath177 .
we use triangular lattices ( six directions for diffusion ) of sizes up to @xmath178 with periodic boundary conditions to limit finite size effects .
the program actually consists of a repeated loop . at each loop , we calculate two quantities @xmath179 and @xmath180 that give the respective probabilities of the three different processes which could happen : depositing a particle ( deposition ) , moving a particle ( diffusion ) or removing a particle from the surface ( evaporation ) . more precisely , at each loop we throw a random number p ( @xmath181 ) and compare it to @xmath182 and @xmath183 .
if @xmath184 , we deposit a particle ; if @xmath185 , we remove a monomer , otherwise we just move a randomly chosen monomer .
after each of these possibilities , we check whether an aggregation has taken place and go to the next loop ( for more details , see @xcite ) . before looking in detail into
the different regimes predicted in section [ equ ] , we summarize our simulation results in fig .
[ crossover ] .
we show there _ all _ our data for @xmath53 as a function of the parameters .
our scaling analysis predicts that the data should fall into a single curve , given by equation [ eqcross ] .
we see that the data remarkably confirms our analysis , over more than 30 orders of magnitude .
this gives us confidence on our entire approach and its predicted exponents , which we now turn on to check in more detail .
we now check that the results summarized in section [ summa ] are correct . figure [ nmaxfnoscal ] shows the evolution of the maximum island density as a function of the flux for different evaporation times .
each of these curves is different from the others , since they correspond to different evaporation times .
however , according to our preceding analysis , they should all present a transition from the low evaporation regime to the high evaporation regime .
this can be detected by a change of slope , from @xmath186 in the high evaporation regime ( solid line ) to @xmath187 in the low one ( dashed line ) .
of course , this regime change does not occur for all the curves at the same value of the flux , since the parameter that determines that change is not the flux but rather @xmath188 .
figure [ nmaxfnoscal ] shows that the results of section [ summa ] accurately describes the behaviour of our model , at least concerning the flux evolution of the maximum island density .
we now turn to the other variable , the evaporation time .
we show in figure [ nmaxtenoscal ] the dependence of the maximum island density on @xmath3 ( @xmath176 ) .
we notice that for high enough evaporation times , the island density tends to become roughly constant , as predicted by our calculations . for lower values of @xmath3
, @xmath53 changes rapidly .
we predict ( section [ summa ] ) two regimes : for @xmath189 , we expect @xmath190 , while for @xmath191 , we expect @xmath192 .
this last regime is clearly seen for the curves obtained for fluxes @xmath193 and @xmath194 ( squares and diamonds respectively , the slope 2/3 is given by the solid line ) .
the first regime is difficult to see for two reasons .
first , we need @xmath195 _ and _ strong evaporation , i.e. @xmath196 .
this means a very low island density , meaning very long computing times and large lattices .
second , the crossovers with the two other regimes ( exponents 2/3 and 0 ) tend to mask the exponent 4/3 . taking a lower value for the flux ( @xmath197 , filled circles ) , we can see that the slope in this intermediate regime is larger than 2/3 .
our treatment predicts two limiting regimes for the power - law ( @xmath198 ) evolution of the mean island radius as a function of time : @xmath199 in cases of strong evaporation and @xmath200 for complete condensation ( we only treat here the case of `` clean '' substrates " ) .
[ rt]a shows that we observe indeed an exponent very close to 2/9=0.22 when evaporation is negligible ( @xmath201 , @xmath202 ) , while the exponent is close to 1 in the opposite case ( @xmath203 , @xmath204 ) , see fig .
of course , intermediate cases can arise in experiments and the exponent is between these two extreme values , with values around 0.5 - 0.6 as shown in fig .
[ rt]c ( @xmath205 ) .
note that we have defined here the radius as @xmath206 where m is the total mass present on the substrate and n the island density , but we have checked that similar exponents are measured if one defines the mean radius as @xmath207 where @xmath208 is the surface coverage . here
we investigate how the different microscopic mechanisms can affect the growth kinetics .
this can be an important help for experimentalists seeking information on which processes are actually present in their experiments @xcite .
[ cinetns ] confirms our analytical analysis and shows that evaporation or the presence of surface defects can significantly alter the time evolution of island density .
if defects are present , monomers will be trapped by them at the very beginning of the growth and the number of islands equates rapidly the number of defects , whatever the diffusivity of the atoms .
if evaporation is present , the opposite effect is observed : since many atoms do not contribute to the growth ( they evaporate before reaching an island ) , the saturation is reached for very high thicknesses ( typically @xmath209 ) .
island size distributions have proven very useful as a tool for experimentalists to distinguish between different growth mechanisms @xcite . by _
size _ of an island , we mean the surface it occupies on the substrate . for the `` three dimensional '' particles studied here ,
their projected surface is the easiest quantity to measure by microscopy .
note that the projected surface for a given mass depends on the precise shape of the islands , which is assumed here to be pyramidal ( close to a half - sphere ) .
size distributions are normalized by the mean island size in the following way : one defines @xmath210 as the probability that a randomly chosen island has a surface @xmath211 when the average surface per island is @xmath212 , where @xmath213 stands for the number of islands of surface @xmath211 , @xmath214 is the total number of islands and @xmath7 for the coverage of the surface .
it has been shown @xcite that by normalizing the probabilities and plotting @xmath215 against @xmath216 , one obtains a `` universal '' size distribution independent of the coverage , the flux or the substrate temperature for a large range of their values .
these size distributions can be obtained from the simulations @xcite .
fig [ distrib3d ] shows the evolution of the _ rescaled _ island size distributions for three dimensional islands ( pyramids ) in presence of evaporation .
we recall that size means here the projected surface of the island , a quantity which can be measured easily by electronic microscopy .
we note the same trends as for 2d islands @xcite .
it is clear that the distributions are significantly affected by the evaporation , smaller islands becoming more numerous when evaporation increases .
this trend can be qualitatively understood by noting that new islands are created continuously when evaporation is present , while nucleation rapidly becomes negligible in the complete condensation regime .
the reason is that islands are created ( spatially ) homogeneously in the last case , because the positions of the islands are correlated ( through monomer diffusion ) , leaving virtually no room for further nucleation once a small portion of the surface is covered ( @xmath217 ) . in the limit of strong evaporation
, islands are nucleated randomly on the surface , the fluctuations leaving large regions of the surface uncovered .
these large regions can host new islands even for relatively large coverages , which explains that there is a large proportion of small ( @xmath218 ) islands in this regime .
fig [ distribdef ] shows the evolution of the _ rescaled _ island size distributions for pyramidal islands nucleating _ on defects_. two main differences can be noted .
first , the histograms are significantly narrower than in the preceding case , as had already been noted in experimental studies @xcite .
this can be understood by noting that all islands are nucleated at almost the same time ( at the very beginning of growth ) .
the second point is that the size distributions are sensitive to the actual coverage of the substrate , in contrast with previous cases .
in other words , there is no perfect rescaling of the data obtained at different coverages , even if rescaling for different fluxes or diffusion times has been checked .
other authors have analyzed similar mean - field rate equations to find the growth dynamics and maximum island density in the presence of evaporation @xcite .
they have also obtained different regimes identified roughly in the same way as in the present work .
we think that there is some ambiguity in their definition of the different regimes .
a regime - meaning a single relation between @xmath53 and the deposition parameters - should be defined only by the values of f , @xmath3 and @xmath1 , as in section [ summa ] .
this is what is required from the experimental side : given some values of the parameters , what will happen on the surface ?
instead , previous works have introduced other parameters , such as the coalescence coverage , the capture cross sections , or even the island density itself , in the characterization of the regimes . besides this general remark
, we note that there is some disagreement about the different regimes between the various authors .
stoyanov and kashchiev @xcite find two regimes which correspond to our complete condensation and strong evaporation ( with @xmath195 ) cases .
we have added here the case @xmath219 .
venables et al .
@xcite find three different regimes .
the two extreme regimes coincide with stoyanov s and ours , but their intermediate regime ( which predicts @xmath220 is not observed in our simulations : this is particularly clear in fig .
[ nmaxtenoscal ] which shows no @xmath221 dependence : instead , the intermediate regime shows an exponent greater than 2/3 , as predicted by eq . [ nmaxstrev ] ) .
actually , venables s intermediate regime seems physically strange because there is no dependence on the diffusion coefficient of the adatoms .
the time evolution of the mean radius has also been studied previously ( for a recent review , see @xcite and references therein ) . here
, we have shown that the exponent of the radius versus time power - law is 1 when evaporation is important ( islands grow only by direct impingement ) , 1/3 when condensation is almost complete _ and _ the number of islands is constant ( for example when their concentration reaches the defect concentration for `` dirty '' substrates ) or 2/9 for clean substrates and complete condensation ( here the number of islands is never really constant since nucleation does not stop until coalescence ) .
in previous studies @xcite , it has been predicted that the radius shows a power - law dependence with an exponent 1/3 for complete condensation growth and 1 in the case of strong evaporation .
intermediate values were found thanks to the numerical resolution of mean - field equations of island growth @xcite .
the experimental values are rather in the range 0.21 - 0.30 for the pd / mgo(100 ) system @xcite and around 0.3 in other systems . while the intermediate exponents ( between 1/3 and 1 ) can be explained as pertaining to the intermediate regime ( see section [ sectionrt ] ) ,
the values lower than 1/3 are more difficult to explain . in our analysis , the exponents close to 2/9=0.22 can simply be explained ( in the complete condensation regime and for clean substrates ) by the fact that the island density is not constant ( as assumed in previous studies ) .
in principle , figure [ nmaxfnoscal ] allows to determine the value of the microscopic parameters ( diffusion , evaporation ) if the saturation island density is known .
the problem is : does this island density correspond to the defect concentration of the surface or to homogeneous nucleation ?
is evaporation present in our experiments and what is the magnitude of @xmath3 ?
the first question can be answered by looking at the density evolution with the flux .
as already explained , if this leaves unaffected the island density , nucleation is occuring on defects .
a similar test can be performed by changing the substrate temperature , but there is the nagging possibility that this changes the defect concentration on the surface .
it is also possible to study the kinetics of island nucleation , i.e. look at the island concentration as a function of thickness or coverage . the presence of defects can be detected by the fact that the maximum island density is reached at very low coverages ( typically less than 1% , see fig .
[ cinetns ] ) .
one should be careful however to check that all the islands , even those containing a few atoms , are visible in the microscope images .
the second question is more delicate .
first , one should check whether atomic reevaporation is important . in principle , this can be done by measuring the sticking coefficient , i.e. the amount of matter present on the surface as a function of the matter brought by the beam . if possible , this measure leaves no ambiguity . otherwise , the kinetics of island creation is helpful .
if the saturation is reached at low thicknesses ( @xmath222 ) , this means that evaporation is not important .
another way of detecting atom evaporation is by studying the evolution of the saturation island density with the flux : the exponent goes from 0.29 to 0.66 ( fig .
[ nmaxfnoscal ] ) .
suppose now that one finds that evaporation is indeed important : before being able to use fig .
[ nmaxfnoscal ] , one has to know the precise value of @xmath3 , and this is not easily achieved .
for example , one could try to measure the sticking coefficient or the quantity of matter needed to reach saturation to obtain an estimation of the evaporation . intuitively , the more evaporation is important , the more matter we need to reach the saturation density .
unfortunately , this strategy , although useful for growth of 2d islands @xcite is not so straightforward here .
the reason is that in the limit of strong evaporation ( section [ strong ] ) , one has @xmath223 , thus bringing no independent information on the parameters .
the same is true for the sticking coefficient , which is a _ constant _ , i.e. independent of the value of @xmath3 or the normalized flux .
this counterintuitive result can be understood by noting that in this limit , islands only grow by direct impingement of atoms within them .
fortunately , in many experimental situations the limit of high evaporation is not reached and we `` benefit '' from ( mathematical ) crossover regimes where these quantities do depend on the precise values of @xmath3 .
[ coll3d ] gives the evolutions of @xmath224 and @xmath29 for different values of @xmath3 and f in this crossover region .
then , a measurement of @xmath224 or @xmath29 can shed light on experimental the value of @xmath3 . for more details on interpretation of experimental data ,
we refer the reader to a review paper to appear @xcite .
we have presented a comprehensive theoretical analysis of the growth of three dimensional structures on surfaces by atom evaporation .
the study has been carried out by combining a simple scaling analysis , a more rigorous rate equations approach and computer simulations , with the main scope of helping experimentalists to analyze their data .
the scaling analysis can give very simply the growth exponents in the `` clean '' substrate case , in the limiting regimes of high and low evaporation as well as in the crossover between these .
the rate equations confirm this analysis and predict the growth on `` dirty '' substrates , i.e. surfaces containing perfect traps for adatoms .
the two approaches were compared to monte - carlo computer simulations and very good agreement was found .
in addition to the analytical predictions , computer simulations allowed to predict for the first time important growth characteristics such as the island size distributions and intermediate regimes which are difficult to study analytically . this is particularly interesting for the interpretation of experimental data from the the sticking coefficient and the saturation island density ( see section [ expdata ] ) .
h. rder , e. hahn , h. brune , j .-
bucher , and k. kern , nature * 366 * , 141 ( 1993 ) ; h. brune , c. romainczyk , h. rder , and k. kern , nature * 369 * , 469 ( 1994 ) ; h. brune , h. rder , c. boragno , and k. kern , phys .
* 73 * , 1955 ( 1994 ) for recent experimental results , see for example _ evolution of epitaxial structure and morphology _ , ed .
a. zangwill , d. jesson , d. chambliss and r. clarke , or _
disordered materials and interfaces - fractals , structure and dynamics _ , ed .
stanley , h.z .
cummins , d.j .
durian and d.l .
johnson , both proceedings of the materials research society , fall meeting , boston , december 1995 , volumes * 399 * and * 407 * respectively . to help experimentalists , a software simulating the submonolayer regime of thin - film growth in the various growth hypothesis ( 2d or 3d islands , reversible aggregation , mobile islands ) is presently being developed and will soon be available by simple request to one of us ( pj ) .
p. jensen , h. larralde and a. pimpinelli , phys .
b * 55 * , 2556 ( 1997 ) : note that in this paper , a mistake was made in the normalization of the island size distributions ( fig . 9 ) .
this error is corrected in @xcite | we perform a comprehensive study of the formation of three dimensional ( pyramidal ) structures in a large range of conditions , including the possible _ evaporation _ of adatoms from the surface and the presence of _ surface defects_. we compare our computer simulations to theoretical calculations of the growth and find good agreement between them .
this work clarifies precedent studies of three dimensional growth and predicts the island size distributions obtained in the different regimes .
finally , we show how our analysis can be used to interpret experimental data . how can one grow _ useful _ thin films or nanostructures from atomic beams ?
the usual and most effective way is certainly by a carefully controlled and inspired trial and error method . theoreticians dream that another possibility may exist : by knowing the detailed atomic mechanisms that govern thin film growth , one could in principle tailor the morphologies to the desired application .
fortunately , there are many other justifications to the study of these atomic mechanisms : for example the understanding in terms of atomic mechanisms of growth experiments carried under controlled conditions , where great care is taken to avoid complications ( contamination , surface defects , etc .
see for example @xcite ) . to be able to interpret more complex situations
demands the study of models including many atomic processes @xcite .
the first models only included deposition , diffusion of the adatoms and their irreversible aggregation to form flat islands @xcite .
ratsch et al .
@xcite improved these models by including reversible aggregation in order to reproduce the formation of compact islands . in this paper
we get closer to new experimental situations by including two new ingredients : the growth of _ three - dimensional _ islands and the presence of _ surface defects _ which act as traps for the monomers , in presence of adatom evaporation .
evaporation , i.e. the possibility of desorption of adatoms from the surface , is a feature that should be observed for any system at high enough temperatures . in this sense , it is a phenomenon that is as general as the rest of the ingredients of recent models of film growth , and , as we have shown in the case of two dimensional ( 2d ) growth @xcite , is capable of completely changing the quantitative behaviour of the system .
moreover , evaporation is present in a number of experimental situations @xcite , where 3d islands are built .
we should add that thin film growth models which include evaporation have already been studied using a mathematical analysis of rate equations @xcite .
computer simulations of such models have , to our knowledge , never been carried out .
the point is that computer simulations represent an `` exact '' way of reproducing the growth , in the sense that they avoid the mean - field approximations of rate - equations approaches @xcite .
we have shown previously @xcite that the mean - field equations could lead to wrong predictions in the case of 2d growth : is this also the case in 3d growth ?
three - dimensional ( 3d ) structures are often observed in the growth of heteroepitaxial and non - epitaxial films
. a simple explanation of the formation of such structures , based in the bonding energies between adsorbed atoms and adsorbed atoms and the substrate can be found in @xcite .
we note that such thermodynamic arguments are not always trustworthy , since kinetics play an essential role in determining the growth morphologies @xcite .
we will not consider this point here and will just simulate 3d growth in a schematic way , forcing the adatoms to build pyramidal islands as they aggregate on the surface .
a more realistic model should reproduce the geometric structure of the islands as a result of the different relevant energies ( edge diffusion around the island , barrier for down and up - hill diffusion , ) .
this is beyond the scope of the present paper where we only want to investigate the _ consequences _ of the formation of 3d structures on the growth characteristics ( mainly saturation island density and island size distribution as a function of the growth parameters ) .
the possible influence of surface defects has been stressed repeatedly @xcite .
we want to clarify the manifestations of defects on the growth and to check simple mathematical analysis of the growth by computer simulations .
we study here the _
first _ stages of the growth , roughly until the number of islands on the substrate saturates .
the reason is that it is in this stage that such a simplified model can be of some help to experimentalists who want to understand the microscopic processes present in their experiments .
these fundamental microscopic processes are most easily detected in the first stages of the growth , since in the subsequent stages additional processes can be involved ( additional diffusion paths , interlayer transport , geometrical details of the lattice ) .
the paper is organized as follows .
section i briefly presents the model and discusses some of its approximations .
then , in section ii , we study the growth of 3d islands , first by a simple scaling analysis in the absence of defects , then by a more rigorous mathematical mean - field treatment , where we also include the influence of surface defects which act as perfect traps on the surface . in section iii ,
computer simulations are used to test these analytical predictions and to calculate the island size distributions in the different cases .
after a discussion ( section iv ) where we compare our analysis to precedent studies , we show in section v how experimental results can be analyzed using these results . |
modelling the mass distribution of m31 is a classical problem that has seen many past iterations ( e.g. , einasto 1972 ; kent , huchra & stauffer 1989 ; klypin , zhoa & somerville 2002 ; widrow , perrett & suyu 2003 ; geehan et al.2006 ) .
recently , a spitzer 3.6 @xmath0 m mosaic image of m31 has become available ( barmby et al .
this allows for a more accurate determination of the baryonic mass distribution in m31 than has been determined previously .
accurate mass modelling of m31 is an important tool for both exploring our ideas of cosmology , and for understanding the dynamics of newly discovered local group dwarf galaxies ( e.g. , majewski et al . 2007 ; chapman et al .
2007 ) .
because m31 can be studied in exquisite detail , it provides a crucial testing ground for our ideas of galaxy formation ( kent 1989 ; evans & wilkinson 2000 ) .
a problem of particular relevance for cosmological constant plus cold dark matter ( @xmath3cdm ) cosmology is the tully - fisher zero - point problem , which refers to the fact that standard models can not reproduce the relation between galaxy luminosity and circular velocity ( tully & fisher 1977 ) without over - producing the number density of galaxies at fixed luminosity ( e.g. gonzalez et al .
2000 ; cole et al . 2000 ; benson et al . 2003 ; yang , mo & van den bosch 2003 ) .
another problem of related concern is the cusp / concentration problem namely that dark - matter dominated galaxy rotation curves seem to rise more slowly than predicted in @xmath3cdm ( moore 1994 ; flores & primack 1994 ; moore et al . 1999 ; van den bosch & swaters 2001 ; blais - ouellette , amram & carignan 2001 ; alam , bullock & weinberg 2002 ; swaters et al . 2003 ; simon et al . 2005 ;
dutton et al .
2005 ; kuzio de naray et al . 2006 ) .
it remains to be seen whether these problems point to some unaccounted for process in galaxy formation ( e.g. dutton et al .
2007 ) or to some new physics of cosmological relevance ( kaplinghat , knox & turner 2000 ; zentner & bullock 2002 ; kaplinghat 2005 ; cembranos et al . 2005 ; strigari , kaplinghat & bullock 2007 ; gnedin et al . 2007 ) .
recently , dutton et al .
( 2007 ) and gnedin et al .
( 2007 ) revisited the tully - fisher problem using two large well - defined samples of disc - dominated ( late - type ) galaxies .
they both took as a starting point the `` standard '' model of disc formation , which assumes that initial navarro , frenk & white ( 1997 ; hereafter nfw ) dark haloes respond to disc formation via adiabatic contraction ( ac ) ( blumenthal et al .
1986 ; gnedin et al .
2004 ; sellwood & mcgaugh 2005 ; choi et al .
both groups conclude that the tully - fisher zero point can not be explained if initial nfw haloes have concentrations as high as those expected for @xmath3cdm with @xmath7 ( with @xmath8 for m31-size haloes , e.g. bullock et al .
2001a , macci et al . 2006 ) .
both sets of authors agree that the problem could be alleviated if ac did not operate ( e.g. somerville & primack 1999 ) and dutton et al .
( 2007 ) go on to advocate a model where disc formation induces an _ expansion _ in the underlying halo density structure .
in contrast , gnedin et al .
( 2007 ) argue that ac is a fundamental prediction in galaxy formation and instead suggest that the cosmology be changed to favor lower halo concentrations ( e.g. zentner & bullock 2002 ) or that the imf is lighter ( with a stellar mass - to - light ratio @xmath9 lower by 0.15 dex ) than the standard kroupa ( 2001 ) assumption . given the fundamental issues at hand , the question of whether ac occurs in nature is of significant interest .
theory certainly favours the idea that haloes contract . indeed
, halo contraction must occur when the infall of baryons is smooth and adiabatic ( e.g. blumenthal et al .
1986 ; ryden & gunn 1987 ; sellwood & mcgaugh 2005 ) and simulations suggest that dark haloes will contract even when galaxies or galaxy clusters form quickly from an irregular collapse ( barnes 1987 ; flores et al .
1993 ; jesseit , naab & burkert 2002 ; gnedin et al .
2004 ; sellwood & mcgaugh 2005 ; choi et al .
2006 ; weinberg et al .
2008 ) . on the other hand
, there is very little observational evidence that halo contraction actually occurs in nature .
for example , zappacosta et al . ( 2006 ) performed a detailed xmm study of the radio - quiet galaxy cluster abell 2589 and conclude that an nfw halo + ac model can not explain the data but that a pure nfw halo provides a remarkable fit down to @xmath10 of the halo s virial radius .
similarly , an investigation of seven elliptical galaxies with _ chandra _ by the same group ( humphrey et al .
2006 ) finds that ac degrades the mass profile fits significantly unless strong deviations from a kroupa imf are allowed .
also , kassin , de jong & pogge ( 2006a ) and kassin , de jong & weiner ( 2006b ) found that the rotation curves of 32 out of 34 of the bright spiral galaxies they study are better fit without adiabatic contraction .
klypin et al . (
2002 ) have shown that the observed rotation curve of m31 is best explained with a halo model that includes ac . however , other recent m31 mass modelling papers ( e.g. , widrow et al . 2003 ; geehan et al .
2006 ) have not directly addressed the question of ac .
furthermore , m31 is essentially the only well - studied galaxy that has been shown to require ac ( klypin et al . 2002 ) .
most other dynamical studies of galaxies tend to disfavour ac ( e.g. , kassin et al .
2006a , b ) . given the availability of a spitzer 3.6 @xmath0 m image , it is now appropriate to revisit this question and test the conclusion that m31 requires ac . in what follows
we show that the rotation curve of m31 can only be explained without ac if an unusually high nfw concentration , @xmath11 , is adopted .
such a concentration is well outside the range expected in @xmath3cdm haloes , and therefore this model is disfavoured .
moreover , the best - fit ac model has a slightly high nfw concentration , @xmath4 , for typical , @xmath12 , @xmath3cdm haloes ( bullock et al .
2001a , macci et al . 2006 ) , but it is within 1.5@xmath13 of the expected value for an m31 size halo .
we speculate on the implications of these results in the conclusion section .
our approach is unique compared to other recently published mass profiles derived from modelling the rotation curve of m31 . while our method bears most similarity to
that adopted by klypin et al .
( 2002 ) , we use an updated baryonic mass component , which is derived using a 3.6 @xmath0 m spitzer image .
geehan et al .
( 2006 ) used a pure nfw halo profile , and did not consider ac .
tempel et al .
( 2007 ) also use the 3.6 @xmath0 m spitzer image , yet once again , ac is not considered .
lastly , widrow et al .
( 2003 ) used a different functional form for the halo density profile ( see evans 1993 ) .
the evans profile is based upon an isothermal distribution function . unlike the nfw profile
, the evans profile was not derived by fitting models to dark matter haloes produced in @xmath3cdm simulations .
widrow et al .
( 2003 ) then go on to fit an nfw model to their best - fitting evans ( 1993 ) model .
however , we have chosen to use the cosmologically motivated nfw halo density profile model directly . throughout this paper
we assume a flat @xmath3cdm cosmology with @xmath14 and a hubble constant of @xmath15 km s@xmath16 mpc@xmath16 .
we relate virial masses ( @xmath17 ) and radii ( @xmath18 ) assuming a virial over - density relative to _ average _ of @xmath19 kpc . ] . at times
we quote baryon fractions relative to the universal value with @xmath20 .
we adopt a distance of 784 kpc to m31 ( holland 1998 ) . as a result ,
an angular distance of @xmath21 is equivalent to a distance of 228 pc .
we make use of the spitzer / irac 3.6 @xmath0 m image of m31 ( barmby et al . 2006 ) .
the irac observations of m31 were taken as part of _ spitzer _ general observer program 3126 in 2005 august .
fifteen astronomical observation requests ( aors ) were used to map a region approximately @xmath22 on the sky .
the central @xmath23 was covered by three aors , each having two 12 s frames per position .
the outer regions were covered by two aors , each with two dithered 30 s frames per position .
the pixel scale of irac at 3.6 @xmath0 m is 1.221 arcsec , with a point spread function ( psf ) that has a fwhm @xmath241.2 arcsec .
the 3.6 @xmath0 m filter is 1.2 @xmath0 m broad , spanning the wavelengths 2.9654.165 @xmath0 m . for a description of the irac instrument see fazio et al .
( 2004 ) .
the spitzer data are preferred over 2mass @xmath25-band data for a few reasons .
the regular 2mass @xmath25-band image is simply not deep enough to obtain an accurate bulge / disc decomposition .
furthermore , this same 2mass image suffers from over subtraction of the sky by about 0.5 magnitudes ( t. jarrett , private communication ) .
a 6@xmath26 2mass @xmath25-band image is also available ( where the data were observed with an integration time 6 times longer than the regular 2mass survey ) , but it is only about 1 magnitude deeper and also hampered by uncertain sky background .
the spitzer image has a larger field of view , better depth , and a more accurate determination of the sky level .
we also use the @xmath1 colour profile of walterbos & kennicutt ( 1987 ) , in order to determine the stellar mass - to - light ratio as a function of radius .
we have corrected this @xmath1 profile has for galactic foreground extinction , which is @xmath27 mags in the direction of m31 ( schlegel , finkbeiner & davis 1998 ) .
this is converted to a correction for the @xmath1 colour by using the transformation , @xmath28 from schlegel et al .
( 1998 ) , i.e. @xmath29 for m31 . as a comparison
the amount of extinction at 3.6 @xmath0 m is 1.5 per cent of that at @xmath30 , and about 34 per cent of that in the @xmath25 band .
we adopt the rotation curve data from several sources . in one case ( m1 )
we adopt the h@xmath31 rotation data out to 25 kpc from rubin & ford ( 1970 ) @xmath326583 emission feature , for the region within 3 kpc . however , these velocities are not tabulated in their paper , and so we do not include them here . ] and extend the rotation curve to 35 kpc using hi data from carignan et al .
( 2006 ) . in another case
( m2 ) we use the co rotational velocities from loinard , allen & lequeux ( 1995 ) and the hi from brinks & burton ( 1984 ) to construct an observed rotation curve out to a 30 kpc radius .
this was the rotation curve adopted by klypin et al .
( 2002 ) in their m31 model .
we adopt m1 as our fiducial case here because rubin & ford published errors on their observed velocities .
however , as we show below , both cases yield consistent results . in both cases
we take into account the presence of a @xmath33 supermassive black hole ( bender et al .
2005 ) and the hi gas distribution ( carignan et al .
our goal is to determine the best possible mass model for m31 .
we perform a bulge - disc decomposition in order to estimate the baryonic contribution to the rotation curve .
we then determine several possible mass models .
the best fit model is determined by minimizing the reduced-@xmath34 in a fit to the observed rotation curve .
we first extract the surface brightness profile of m31 using the spitzer 3.6 @xmath0 m image and the iraf ellipse routine , which fits ellipses to an image using an iterative method described by jedrzejewski ( 1987 ) . in order to mask out foreground stars and satellite galaxies sextractor (
bertin & arnouts 1996 ) was used .
an inclination correction was then applied to the surface brightness profile ( see de jong 1996 ; seigar & james 1998a ) as follows @xmath35 where @xmath36 is the surface brightness when viewed at some inclination , @xmath37 , @xmath0 is the corrected surface brightness , @xmath38 is the major axis , @xmath39 is the minor axis and @xmath40 is a factor dependent on whether the galaxy is optically thick or thin ; if @xmath41 then the galaxy is optically thin ; if @xmath42 then the galaxy is optically thick .
since we are using infrared data , the correction for inclination is small , and in many cases completely ignored ( see e.g.seigar & james 1998a ; de jong 1996 ) .
however , graham ( 2001 ) showed that @xmath43 is a good typical value to use for the near - infrared @xmath25-band . if one adopts a simple reddening law , where extinction falls as the square of wavelength , then the value of @xmath44 is appropriate at 3.6 @xmath0 m , and we adopt this value here .
the resulting surface brightness profile ( figure 1 ) reaches a surface brightness of @xmath45 mag arcsec@xmath46 at a radius of @xmath47 .
this is very similar to the surface brightness found by barmby et al .
furthermore , the pne counts of merrett et al .
( 2006 ) predict an r band surface brightness of @xmath48 mag arcsec@xmath46 . at these radii
the metallicity of m31 is @xmath49 ( brewer , richer & crabtree 1995 ; worthey et al .
2005 ) . using the maraston ( 2005 ) population synthesis codes
, one can see that the expected @xmath50[3.6 ] color at this radius is @xmath50[3.6]@xmath242.5 , very similar to our measured surface brightness . from this surface brightness profile
, we perform a one - dimensional bulge - disc decomposition , which employs the srsic model for the bulge component and an exponential law for the disc component ( e.g. andredakis et al . 1995 ; seigar & james 1998a ; khosroshahi , wadadekar & kembhavi 2000 ; donofrio 2001 ; graham 2001b ; mllenhoff & heidt 2001 ; see also graham & driver 2005 for a review ) .
the srsic ( 1963 , 1968 ) @xmath51 model is most commonly expressed as a surface brightness profile , such that @xmath52\right\}},\ ] ] where @xmath53 is the surface brightness at the effective radius @xmath54 that encloses half of the total light from the model ( ciotti 1991 ; caon , capaccioli & donofrio 1993 ) .
the constant @xmath55 is defined in terms of the parameter @xmath56 , which describes the overall shape of the light profile .
when @xmath57 the srsic model is equivalent to a de vaucouleurs ( 1948 , 1959 ) @xmath58 model and when @xmath59 it is equivalent to an exponential model .
the parameter @xmath55 has been approximated by @xmath60 , for @xmath61 ( capaccioli 1989 ; prugniel & simien 1997 ) . the exponential model for the disc surface brightness profile can be written as follows , @xmath62 where @xmath63 is the disc central surface brightness and @xmath64 is the disc scale length .
our bulge - disc decomposition ignores the inner @xmath65 of m31 , which is dominated by an independent nuclear feature ( light , danielson & schwarschild 1974 ) .
the results of the bulge - disc decomposition can be seen in figure [ sb ] .
note that although there appears to be a slight break in the disc surface brightness profile at @xmath244000@xmath66 , the residuals are still very small .
this break may appear due to the presence of spiral structure , evident at similar radii in the 24 @xmath0 m of m31 .
observational properties of m31 are derived from this bulge - disc decomposition and these are listed in table 1 .
[ bdtab ] [ cols="<,<",options="header " , ] with a disc scalelength of @xmath67 and a surface brightness profile extending to a radius of @xmath47 , we are tracing the disc profile out to @xmath243.5 scalelengths . assuming that the exponential disc profile extrapolates to infinity , our profile includes 86.1 per cent of the total disc light .
below , where we assign a mass to the disc , we extrapolate the disc profile to 6 scalelengths , which would contain 98.2 per cent of the disc light .
since ibata et al .
( 2005 ) have shown that the disc of m31 extends to very large radii ( @xmath2470 kpc ) with the same scalelength , our above calculation assumes no disc truncation . both the 10 kpc ring and the 2.6 kpc bar ( beaton et al .
2007 ) show up as small perturbations in the surface brightness profile ( fig .
1 ) at radii of 46.67@xmath68 and 26.67@xmath68 respectively .
we now assign masses to the disc and bulge of m31 .
the stellar mass - to - light ratio in the @xmath25-band is a well calibrated quantity ( see bell et al .
2003 ) which depends on @xmath1 colour .
however , a similar calibration does not exist for the spitzer 3.6 @xmath0 m waveband .
an alternative method would be to estimate the @xmath25-band surface brightness profile , given the 3.6 @xmath0 m profile , and since we know the @xmath1 colour profile , we have chosen to determine the @xmath69[3.6 ] colour dependence on @xmath1 colour using both the bruzual & charlot ( 2003 ; hereafter bc03 ) stellar population synthesis models and the maraston ( 2005 ; hereafter m05 ) stellar population synthesis models .
the bc03 models use the stellar library of pickles ( 1998 ) , which includes stars in a wide range of spectral types ( 05-m10 ) and luminosity classes ( i - v ) and three metallicity groups ( metal poor , metal rich and solar ) .
the pickles ( 1998 ) library has wavelength coverage from 0.1 @xmath0 m to 2.5 @xmath0 m .
basel 3.1 spectra ( see bc03 tables 2 and 3 ) are then used to extend the pickles ( 1998 ) spectra into the infrared from 2.5 @xmath0 m to 160 @xmath0 m .
the spectra of m giant stars are the only ones not observed in the pickles ( 1998 ) library . as a result the bc03 models use synthetic m giant spectra calculated by fluks et al .
these spectra are extended into the infrared using the model atmospheres of schultheis et al .
( 1999 ) , which cover a spectral range from 0.5 @xmath0 m to 10 @xmath0 m , and 10 equally spaced stellar temperatures in the range @xmath70 k. the main difference between the bc03 and the m05 models is the way in which thermally - pulsing agb ( tp - agb ) stars are treated .
the m05 models provide a better treatment of tp - agb stars than the bc03 models . furthermore , maraston et al .
( 2006 ) find that in general the m05 models provide better fits to observations of infrared spectra of galaxies than the bc03 models .
they indicate systematically lower ages and , on average , masses that are 60 per cent lower for the stellar population samples in these galaxies .
figure [ synphot ] shows the result of both the bc03 ( _ left panel _ ) and m05 ( _ right panel _ ) models . in both cases ,
the dotted lines in figure [ synphot ] represent the @xmath1 colour of the inner and outer disc of m31 ( walterbos & kennicutt 1987 and figure [ colour ] ) after correction for galactic foreground extinction .
since the m05 models provide an updated treatment of tp - agb stars and generally provide better fits to observations than the bc03 models , from here on we concentrate on the m05 results .
however , it is interesting to note that the @xmath69[3.6 ] colour index predicted by the bc03 and m05 models differs by @xmath710.03 mag ( see figure 2 ) .
lines of constant metallicity ( solid lines ) and age ( dashed lines ) are shown .
the typical metallicity in the inner few kpc of m31 is @xmath242.4 @xmath72 , whereas in the outer disc this falls to @xmath240.4 @xmath72 ( brewer , richer & crabtree 1995 ; worthey et al .
2005 ) . from this metallicity gradient and the m05 models ( figure [ synphot ]
_ right panel _ )
it can be seen that the @xmath1 colour gradient in m31 is almost entirely due to changes in metallicity , and that the @xmath69[3.6 ] colour difference is @xmath73 mag .
furthermore , we can now use the @xmath1 gradient ( figure [ colour ] ) and the metallicity gradient of m31 to estimate the @xmath69[3.6 ] colour as a function of radius .
we find that the inner region of m31 has a colour @xmath69[3.6]@xmath74 , and the outer disc ( at 26 kpc ) has a colour @xmath69[3.6]@xmath75 .
given the @xmath1 colour profile and the k - band mass - to - light ratio calibration from bell et al .
( 2003 ) we calculate a central mass - to - light ratio , @xmath76 and a mass - to - light ratio at 26 kpc , @xmath77 .
from the @xmath69[3.6 ] colours determined in our stellar population synthesis analysis , the above mass - to - light ratios are equivalent to 3.6 @xmath0 m mass - to - light ratios of @xmath78 in the centre and @xmath79 at 26 kpc ( equivalent to a @xmath80 gradient 0.13 kpc@xmath16 ) . in our analysis , we prefer to keep the central @xmath81 value a free parameter , and as such we allow it to vary in the range @xmath82 .
this is a standard approach to adopt in mass modelling of galaxies ( see humphrey et al .
2006 ; zappacosta et al .
2006 ) and as discussed at the end of section 3.2 and shown in table 2 , when the @xmath83 is left as a free parameter , the best - fit value found is @xmath84 , i.e. consistent with the value of @xmath83 derived here .
we use our derived value for the 3.6 @xmath0 m disc luminosity of @xmath85 @xmath86 .
we then use the derived disc and bulge light profiles @xmath87 to determine the stellar mass contribution to the rotation curve , @xmath88 . a concern in using the 3.6 @xmath0 m spitzer waveband to determine the underlying stellar mass , is the effect of emission from hot dust in this waveband , although this is probably only important in or near hii regions . in order to place some constraint on this ,
we have chosen to explore the emission from dust in the the near - infared @xmath89-band at 2.2 @xmath0 m .
using near - infrared spectroscopy at 2.2 @xmath0 m it has been shown that hot dust can account for up to 30 per cent of the continuum light observed at this wavelength in areas of active star formation , i.e. , spiral arms ( james & seigar 1999 ) . when averaged over the entire disc of a galaxy , this reduces to a @xmath712 per cent effect , if one assumes that spiral arms can be up to 12@xmath90 in width . at 3.6 @xmath0 m
this would therefore result in @xmath713 per cent of emitted light from dust .
another concern for the 3.6 @xmath0 m waveband , would be the contribution from the polycyclic aromatic hydrocarbon ( pah ) emission feature at 3.3 @xmath0 m .
however , iso spectra covering the wavelength range 5.15 to 16.5 @xmath0 m have shown that pah emission in m31 seems to be very weak ( e.g. , cesarsky et al .
1998 ) . furthermore , in a survey of actively star forming galaxies ,
helou et al . ( 2000 ) found that the 3.3 @xmath0 m feature was also very weak when they analysed the average @xmath91 @xmath0 m spectrum of 45 galaxies .
the contribution of the pah feature to the 3.6 @xmath0 m spitzer waveband , is therefore not a major concern . in the following section , where we model the dark matter halo , we have included a bulge surface brightness model using a deprojected srsic function
the function we use is from trujillo et al .
( 2002 ) who show that the deprojected mass density profile is given by , @xmath92 where @xmath93 is the effective radius , @xmath94 is the central intensity , @xmath95 is the stellar mass - to - light ratio , @xmath96 , @xmath97 and @xmath98 is the @xmath99-order modified bessel function of the third kind .
trujillo et al . ( 2002 ) give values for @xmath100 , @xmath101 , @xmath102 , @xmath103 and @xmath104 for different values of the srsic parameter , @xmath56 . in our case , where @xmath105 , we adopt values of @xmath106 , @xmath107 , @xmath108 , @xmath109 and @xmath110 .
finally , the constant @xmath111 depends on the three - dimensional spatial orientation of the bulge . in our case , we assume the bulge is intrinsically spherical , and therefore @xmath112 .
it should be noted that this version of the deprojected srsic function ( equation [ dep ] ) is an approximation , but it is good enough for our purposes .
a range of allowed dark matter halo masses and density profiles are now explored , using two models for disc galaxy formation . in the first we assume that the dark matter halo surrounding m31 has not responded significantly to the formation of the disc ,
i.e. , adiabatic contraction does not occur ( the `` no - ac '' model ) . in this case
the dark matter contribution to the rotation curve is described by a density profile similar to those found in dissipationless dark matter simulations ( the nfw profile ) : @xmath113 where @xmath114 is a characteristic `` inner '' radius , and @xmath115 is a corresponding inner density .
this is a two parameter function and is completely specified by choosing two independent parameters , e.g. , the virial mass @xmath17 ( or virial radius @xmath18 ) and concentration @xmath116 ( see bullock et al .
2001a for a discussion ) .
similarly , given a virial mass @xmath17 and the dark matter circular velocity at any radius , the halo concentration @xmath117 is completely determined . in the second class of models
we adopt the scenario of adiabatic contraction ( ac ) and specifically use the original scheme discussed by blumenthal et al .
( 1986 , hereafter b86 ; see also bullock et al . 2001b and
pizagno et al . 2005 ) .
we also investigate the slightly revised scheme presented in gnedin et al .
( 2004 , hereafter g04 ) . as we discuss below
, the g04 scheme reproduces the outer slope of the observed rotation curve of m31 , but needs a very high concentration to do so .
[ modtab ] lccccc parameter & m1 & m2 & m1+m2 & m1 & m1 + & b86 & b86 & b86 & g04 & no - ac + shear & 0.51 & 0.50 & 0.51 & 0.54 & 0.50 + + @xmath118 ( km s@xmath16 ) & 275@xmath1195 & 275@xmath1195 & 275@xmath1195 & 275@xmath1195 & 275@xmath1195 + + disc luminosity , @xmath120 ( @xmath86 ) & @xmath121 & @xmath122 & @xmath123 & @xmath121 & @xmath122 + + disc scale length , @xmath64 ( kpc ) & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + + central disc mass - to - light ratio , & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + @xmath80 ( 3.6 @xmath0 m solar units ) + + mass - to - light ratio gradient , & 0.013 & 0.013 & 0.013 & 0.013 & 0.013 + @xmath124 ( kpc@xmath16 ) + + initial nfw concentration , @xmath117 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + + bulge - to - disc ratio , @xmath125 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + + virial mass , @xmath17 ( @xmath126 ) & @xmath127 & @xmath128 & @xmath129 & @xmath130 & @xmath131 + + dark matter concentration , & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% + @xmath132 , @xmath133 kpc + + total mass concentration , & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% + @xmath134 , @xmath133 kpc + + stellar baryon fraction , @xmath135 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% + + @xmath34 & 38.75 & 40.63 & 80.31 & 46.50 & 41.58 + + degrees of freedom , @xmath100 & 31 & 27 & 64 & 31 & 31
+ for each ac algorithm ( b86 and g04 ) and for the no - ac model we generate a grid of final rotation curves .
we vary the baryonic contribution to the rotation curve by allowing the bulge - disc decomposition parameters for @xmath64 and @xmath136 to range over their @xmath137 values ( see table 1 and the previous section ) .
the central mass - to - light ratio of the galaxy is allowed to vary over the range @xmath138 . from this
we determine the bulge and disc contribution to the rotation curve . for the bulge
a simplified spherical model is used . for the disc the potential
is derived assuming zero thickness and modified bessel functions as described by binney & tremaine ( 1987 ) . for each set of baryonic parameters
we generate a range of dark halo models with total circular velocities at @xmath139 disc scalelength ( @xmath118 ) that span the best - fit value for m31 ( @xmath140 km s@xmath16 ) quoted by rubin & ford ( 1970 ) by a wide margin : @xmath141 km s@xmath16 . in practice
we achieve each @xmath118 value in our models by allowing the initial halo concentration to vary over @xmath142 of its expected range @xmath143 ( bullock et al .
2001a ; wechsler et al .
2002 ; macci et al .
2006 ) and then setting the halo virial mass @xmath17 necessary to reproduce the desired value of @xmath118 . finally , we determine the implied fraction of the mass in the system in the form of stars compared to that expected from the universal baryon fraction , @xmath144 and demand that @xmath145 obeys @xmath146 . from this grid of choices
we derive best - fitting dark halo parameters by minimizing the reduced-@xmath34 for our two choices of rotation curve data .
these results are summarized in table 2 . for the most part ,
when minimizing @xmath34 , we assume that the errors on the observational data ( i.e. , the rotation velocities ) are not correlated .
instead we assume that they are gaussian in nature . however , this is not the case for the carignan et al .
( 2006 ) hi data ( see figure [ rcurve ] top row and bottom right beyond 25 kpc ) , where the errors obviously are correlated . in order to account for the strange behaviour of the errors on these hi rotation velocities we have treated all of this data as an average rotation velocity of 226.3 km s@xmath16 at 31.16 kpc .
this is an idealized simplification but allows a quantitative estimate of the relative goodness of fit between models .
furthermore , although error bars are not available for the `` m2 '' curve , we assume 10 per cent errors in the model fitting , as this is similar to the size of the errors quoted for the rubin & ford ( 1970 ) h@xmath31 rotation curve .
we therefore quote a @xmath34 for the `` m2 '' curve in table 2 that is consistent with 10 per cent errors .
we then determine how consistent all the derived rotation curves are with each other , by adopting the formal best fitting rotation curve as our fiducial mass model .
for our `` m1 '' curve , the number of degrees of freedom in the fit is @xmath147 .
the number of data points we fit to is @xmath148 , and 6 parameters are determined from the fit , the nfw concentration @xmath117 , the halo virial mass @xmath149 , the stellar mass - to - light ratio @xmath80 , the disc luminosity @xmath136 , the disc scalelength @xmath64 and the bulge - to - disc ratio @xmath125 , thus leaving the degrees of freedom in the fit as @xmath147 .
the m1 rotation curve from rubin & ford ( 1970 ) is shown by the solid squares with error bars in figure [ rcurve ] ( _ top left , top right , middle right , bottom left _ ) along with our best - fitting rotation curve models ( solid lines ) .
the contribution from stellar mass ( short - dashed ) and dark matter ( long - dash ) in each case is shown along with the rather minor contributions from neutral gas ( dotted ) and the central black hole ( dot - dashed ) .
overlaid in the _ top left _
panel of figure [ rcurve ] is a model that includes adiabatic contraction ( overall reduced-@xmath150 ) .
the dark halo in this case _ has undergone _ adiabatic contraction according to the b86 prescription .
our best - fitting parameters are summarized in table 2 under `` m1 b86 '' .
we determine the halo of m31 to have a total ( dark+baryonic ) virial mass of @xmath151 and an initial concentration @xmath152 .
our derived virial mass is almost a factor of 2 smaller than that quoted by klypin et al .
we also find a concentration that is higher than both klypin et al .
( 2002 ) , who find @xmath153 and widrow et al .
( 2003 ) , who find @xmath154 . in the case of widrow
et al . ( 2003 ) it should be noted that they initially fit an isothermal model given by evans ( 1993 ) .
they then fit an nfw profile to this model . given that the isothermal model favors a constant density core
, the derived nfw concentration will be articifially low and fitting an nfw model in a more direct manner may result in a higher concentration .
the model rotation curve fits the data adequately within the uncertainties .
the innermost data point from the observational rotation curve is significantly lower than the model .
this can be explained by the recent strong evidence for a bar , which dominates the near - infrared light out to a radius of @xmath242.6 kpc , in m31 ( athanassoula & beaton 2006 ; beaton et al .
this bar has been used to explain the minimum seen in the observed rotation curve at a radius of @xmath155 kpc ( athanassoula & beaton 2006 ) .
this bar may also affect the outer rotation velocities as well , although probably by less than 10 per cent .
this would therefore be within the errors associated with the observed data .
indeed , shifting the outer rotation velocities systematically upward by @xmath2410 per cent changes the best - fit models by an amount that remains within the quoted error range of the original fit ( see table 2 ) .
it should also be noted that by a radius of @xmath156 kpc the rotational velocity has dropped to @xmath157 km s@xmath16 , which is similar to the rotational velocities seen at radii of @xmath158 kpc in the hi rotation curve ( carignan et al .
2006 ; open triangles ) . the best - fitting model without adiabatic contraction is shown by the solid line in the _ middle right _ panel of figure [ rcurve ] .
this , the no - ac model , was determined in exactly the same way as our best - fit b86 ac model .
as listed in table 2 , the no - ac assumption requires a much higher nfw concentration , @xmath159 and a lower virial mass , @xmath160 .
this model also produces a reasonable fit , with a reduced-@xmath161 , but it should be noted that the required concentration is , at best , only marginally acceptable within a @xmath3cdm context .
this would correspond to a @xmath162 outlier from the expected concentration distribution of m31-sized haloes ( e.g. , bullock et al .
2001a ) . in figure [ chi2 ]
we plot the nfw concentration , @xmath117 , as a function of the virial mass , @xmath17 , for both the `` m1 b86 '' case ( _ solid line _ ) and the `` m1 no - ac '' case ( _ dashed line _ ) .
the contours represent the 1@xmath13 , 2@xmath13 and 3@xmath13 confidence levels for the concentration and virial mass for both models . from figure [ chi2 ]
, one can see that the `` m1 no - ac '' model is inconsistent with any reasonable concentration of @xmath163 at the 5@xmath13 level at least .
however , since it is possible to produce a reasonable fit with a pure nfw model , albeit with a very large concentration , we do not rule out our `` m1 no - ac '' case , although it is clearly not consistent with standard @xmath3cdm cosmology .
the confidence contours shown in figure [ chi2 ] have also allowed us to estimate errors on the derived best - fit masses and concentrations .
the 1@xmath13 , 2@xmath13 and 3@xmath13 contours are drawn as the boundaries where @xmath34 increases by 1.0 , 2.71 and 6.63 relative to its minimum value . for our best - fit ` m1 b86 ' model , @xmath34=38.75 .
there is also a ` m1 b86 ' model for which the @xmath34=39.75 , i.e. 1@xmath13 higher than the best fit . for this model ,
the concentration is @xmath117=21.1 and the virial mass is @xmath164 .
there is a change in the concentration of @xmath165=1.1 and a change in the virial mass of @xmath166 at the 1@xmath13 level , i.e. when the value of @xmath34 is increased by 1 .
the values of @xmath165 and @xmath167 are therefore adopted as our 1@xmath13 error on @xmath117 and @xmath17 respectively , and these are the errors that we report in table 2 . an intermediate result is shown in the _ top right _ panel of figure [ rcurve ] . here
we use the ac model of g04 , which produces less contraction than the b86 prescription .
if the initial nfw concentration is held fixed at the same value as the best - fitting b86 case , i.e. , @xmath4 , then the g04 model overestimates the observed rotation curve at @xmath168 kpc .
the g04 model can only reproduce this outer slope when the initial halo has @xmath169 . with this high concentration
the reduced-@xmath34 of the g04 model is 1.50 .
this model can not be ruled out , but ( as in the pure nfw case ) its halo concentration is higher than expected in @xmath3cdm ( bullock et al .
2001a , macci et al . 2006 ) . therefore , from here on
, we chose to focus on the b86 ac case , but provide the best - fitting model results for the g04 case in table 2 .
for completeness , we also perform a fit to the same observed rotation curve that was used by klypin et al .
( 2002 ) in their m31 mass decomposition .
our best fitting mass model is shown in the _ middle left _ panel of figure [ rcurve ] ( `` m2 b86 '' in table 2 ) .
here we have adopted the b86 ac model . the initial nfw concentration in this case ( @xmath170 ) is higher than that the value of @xmath153 favored by klypin et al .
( 2002 ) and lower than our `` m1 b86 '' case . however , it is not a large difference , and at face value our `` m1 b86 '' and `` m2 b86 '' models are consistent with each other . as we discuss below , the difference between our result and that of klypin et al .
( 2002 ) is driven mainly by our updated baryonic model , which seems to provide a less concentrated bulge .
for the m2 case , we calculate a @xmath34 , assuming 10 per cent errors . as a final consistency check
, we also perform a fit of the b86 ac model to a combination of the m1 and m2 data ( _ bottom left _ of figure [ rcurve ] ; `` m1+m2 b86 '' in table 2 ) .
this produces a result consistent with `` m1 b86 '' and `` m2 b86 '' .
figure [ mass ] shows the enclosed mass as a function of radius for our fiducial model ( m1 b86 ) .
the fraction of mass contained within the central 10 kpc ( the `` central mass concentration '' hereafter , as originally defined by seigar et al .
2006 ) is @xmath171 per cent . in absolute terms , the mass contained within 10 kpc is @xmath172 .
this is about a factor of 2 larger than that calculated by rubin & ford ( 1970 ) , but it should be noted that they did not take into account non - circular orbits due to the presence of a bar , which we now know exists in m31 .
this would have the effect of increasing the mass contained within the bar region .
the mass calculated within a 35 kpc radius from the hi rotation curve by carignan et al .
( 2006 ) is @xmath173 , similar to the mass derived from our best fit model within a 35 kpc radius of @xmath174 . within a 31 kpc radius
we find a mass of @xmath175 , which is similar to the mass of @xmath176 within the same radius found using kinematic data of planetary nebulae ( evans & wilkinson 2000 ) .
the total implied stellar baryon fraction of this model is @xmath177 per cent .
other mass models of m31 ( klypin et al . 2002 ;
geehan et al . 2006 ; widrow et al .
2003 ) have used bulge velocity dispersions as inputs into their models .
for example , klypin et al . (
2002 ) used the line - of - sight velocity dispersion of m31 from kormendy & bender ( 1999 ) to constrain mass models of m31 within the bulge region . in the bulge region , the velocity dispersion , @xmath13 is dominant ( c. 150 km s@xmath16 ) at least out to twice the effective radius , i.e. @xmath1783.96 kpc .
geehan et al .
( 2006 ) assume a hernquist ( 1990 ) model for the bulge , and compare the expected velocity dispersion profile from jeans equation modelling with the observations of kormendy ( 1988 ) .
our approach differs from this .
we have chosen to assume a srsic law for the bulge and we then perform a mass - to - light conversion . in our approach , we essentially ignore any kinematics within the inner few kpc , because in that region the kinematics are likely to be perturbed by the bar or affected by streaming motions along the bar .
instead , we assume that stellar light traces stellar mass .
figure 4 shows that the bar kinematics are strongest within the inner @xmath246 kpc . beyond 10 kpc ,
the bar kinematics are not likely to affect the disc kinematics .
furthermore , we assume that the bulge is spherically symmetric ( an assumption also made in the papers by klypin , geehan and widrow ) .
the bulge in our model has a mass of @xmath179 @xmath126 .
it is reassuring to note that the jeans modelling of geehan et al .
( 2006 ) derives a bulge mass of @xmath180 @xmath126 , very similar to our bulge mass .
we note that the bulge component in our models is less concentrated than that derived by some other groups .
klypin et al . (
2002 ) have a more concentrated bulge , but that is probably due to a combination of the fact that they use a hernquist model ( equivalent to a srsic index of @xmath57 ) for the m31 bulge and optical data to determine the bulge characteristics .
geehan et al .
( 2006 ) also use a hernquist model .
this type of model will naturally produce a more concentrated bulge than a srsic model with index @xmath181 .
the use of the @xmath80 ratios from bell & de jong ( 2001 ) also gives klypin et al .
( 2002 ) an overall more massive bulge .
the more recent models by tamm et al .
( 2007 ) and tempel et al .
( 2007 ) make use of a model that is similar to a srsic law , but they find an index of @xmath182 compared to our index of @xmath181 . as a result , their bulge model is also naturally more concentrated than ours .
it is also interesting to note that replacing our srsic bulge with a bulge profile modelled using a hernquist or exponential law does not affect the overall dark matter halo profile by more than 10 per cent .
this may be because the overall bulge mass does not change ; it is just the way in which the bulge mass is distrbuted that changes .
the issue of turbulent gas motions is very important for low - mass galaxies with small rotation velocities ( e.g. , valenzuela et al .
2007 ) , as this requires a correction for pressure support or asymmetric drift in the mass model .
the correction to the total mass is @xmath183 .
since the h@xmath31 and hi rotation curves are measured from gas , it is important to understand what effect the turbulent gas motions may have on our mass model .
since , m31 is a galaxy with a high rotation velocity ( @xmath184 km s@xmath16 ) and a turbulent velocity dispersion in galaxy discs is typically @xmath185 km s@xmath16 ( i.e. around 10 per cent of the rotation velocity ) this effect is small . indeed ,
if we assume that the scatter of the h@xmath31 rubin & ford ( 1970 ) rotation curve around our best fitting model is due to turbulent gas motions , then we can derive an estimate for their typical value in m31 .
we can do this because the rubin & ford ( 1970 ) rotation curve is derived from h@xmath31 spectroscopic observations of individual hii regions in m31 . in this case
, we have to assume that the errors on the rubin & ford ( 1970 ) rotation curve are uncorrelated , and we derive a turbulent velocity dispersion of @xmath186 km s@xmath16 . adding this in quadrature to the maximum rotation velocity from the rubin & ford ( 1970 ) , results in a mass difference of @xmath240.4 per cent for the _ total _ ( baryonic + halo ) galaxy mass . as mentioned above
, the preferred total ( dark+baryonic ) virial mass for m31 is @xmath187 .
this is similar to the estimate of @xmath188 derived from satellite kinematics ( ct et al .
2000 ) and the total mass of @xmath189 from kinematics of the andromeda stream ( fardal et al . 2006 ; geehan et al .
it is also close to the lower limit of @xmath190 derived from kinematics of halo stars ( chapman et al .
our favored virial mass is about a factor of 2 smaller than the value favored by the klypin et al .
( 2002 ) analysis of the m31 rotation curve .
this can be attributed to three differences in our analysis .
first , klypin et al . ( 2002 ) use the hi rotation curve of brinks & burton ( 1984 ) to find their best fitting model . in this study
we use the h@xmath31 rotation velocities of rubin & ford ( 1970 ) .
however , we find that this is not the major driver .
we find that the model rotation velocities are consistent with those from the more recent hi study of carignan et al .
( 2006 ) , and when we fit to the same observed rotation curve as klypin et al .
( 2002 ) , we find that our best - fit model is consistent with our best - fit model to the rubin & ford ( 1970 ) curve .
furthermore , if we try and model a combination of data from rubin & ford ( 1970 ) and klypin et al .
( 2002 ) we also find a best - fitting rotation curve consistent with our fiducial case . more importantly , klypin et al .
( 2002 ) used the @xmath80 ratios of bell & de jong ( 2002 ) to determine the stellar mass from their bulge - disc decomposition . in this study
we allow @xmath80 to vary freely in the range @xmath191 , and find a best - fitting value close to the expected 3.6 @xmath0 m mass - to - light ratios ( see section [ bruzual ] ) .
possibly the most significant difference is that klypin et al .
( 2002 ) perform a bulge - disc decomposition using an @xmath192 band image , whereas we use a spitzer 3.6 @xmath0 m image . using a 3.6 @xmath0 m image
provides a better trace of the stellar mass ( e.g. , barmby et al .
2006 and references therein ) .
overall , we should therefore have a more accurate estimate of the baryonic mass profile in m31 .
indeed , if we adopt the baryonic mass profile of klypin et al .
( 2002 ) and try to fit either to the hi or to the h@xmath31 rotation curve , we find a best fitting model that is consistent with their result .
typically , the best - fitting @xmath80 values we find are in the range 0.951.15 , depending on the model used .
the lowest value of @xmath193 is difficult to reconcile with the expected value , which is @xmath194 , for a galaxy with the central @xmath1 colour of m31 .
this expected value was determined by combining the bell et al .
( 2003 ) @xmath25-band @xmath80 with the @xmath69[3.6 ] colours derived from our stellar population synthesis described in section 3.1 ( figure [ synphot ] ) .
the lowest value we find ( @xmath193 ) is for our no - ac model .
we also find that this model is difficult to explain in the context of @xmath3cdm cosmology , due to its large concentration .
our m1 b86 model reproduced the highest value of @xmath195 that we find .
this is consistent with the expected range of @xmath80 values .
in this paper we have derived new mass models for m31 , based on an updated baryonic contribution to the rotation curve . the baryonic contribution is derived using a spitzer 3.6 @xmath0 m image .
we find that both a pure nfw and an adiabatically contracted nfw profile can produce reasonable fits to the observed rotation curve of m31 .
however , the best - fit concentration of m31 for a pure nfw model is @xmath196 , which is disfavoured when compared to the expected value of @xmath197 ( wechsler et al .
2002 ; maccio et al . 2006 ) in a @xmath3cdm cosmology ( our best fitting concentration is a @xmath198 outlier ) .
the best fit produced by the ac recipe of gnedin et al .
( 2004 ) brings down the concentration to @xmath169 , or @xmath199 , and is thus marginally consistent with @xmath3cdm as a @xmath243.7@xmath13 outlier .
if one accepts @xmath3cdm cosmology , then the most consistent fit to the observed rotation curve seems to be that produced by the ac recipe of blumenthal et al .
( 1986 ) . in this case
the concentration , @xmath200 , and this is consistent with the expectations of @xmath3cdm , to within 1.5@xmath13 . for all of the types of model we use , the derived best - fits are in good agreement with the mass distributions derived from various other observational methods .
our estimate of the halo virial mass lies in the range @xmath201 ( for the m1 no - ac case ) and @xmath202 ( for the m1 g04 case ) .
this is in close agreement with the virial mass of @xmath203 found via kinematics of satellite galaxies ( ct et al .
2000 ) and the andromeda stream ( fardal et al . 2006 ; geehan et al .
2006 ) , and the lower limit of @xmath204 found from kinematics of halo stars ( chapman et al .
our m1 b86 model , which is in closest agreement with the expectations of @xmath3cdm , lies in the middle of the range with a virial mass of @xmath205 .
another recent m31 mass model by tamm , tempel & tenjes ( 2007 ) and tempel , tamm & tenjes ( 2007 ) also uses spitzer 3.6 @xmath0 m imaging to determine the total mass of m31 , and they find a mass of @xmath206 for a pure nfw model , i.e.a slightly higher mass than for our best fit .
however , they did not look at the question of ac . also , their bulge seems to have a more concentrated density profile ( see figure 6 in tempel et al . 2007 ) , which is probably driven by the use of rotation velocities inside 5 kpc .
this may be driving the differences between our model and the model presented in tempel et al .
we have been cautious not to include rotation data within the inner few kpc in our model , since it is unlcear how non - circular motions induced in this barred region will affect the overall mass model .
it is interesting that , while it has been shown that adiabatically contracted halo models do not generally fit the observed rotation curves of galaxies ( e.g. dutton et al .
2005 , 2007 ; kassin et al .
2006a , b ; pizagno et al .
2005 ) , the observed rotation curve of a well - studied galaxy such as m31 can be well described by a halo that has undergone adiabatic contraction .
indeed , one either has to adopt _
( i ) _ a pure nfw profile with an unphysically large concentration , within the context of @xmath3cdm , or _
( ii ) _ an adiabatically contracted nfw profile and a more reasonable concentration .
such a scenario is in agreement with the results of klypin et al .
( 2002 ) , who claim that an adiabatically contracted halo is favoured for m31 . much of the power in this analysis came from including the normalization _ and _ outer slope of m31 s rotation curve in our fits ( where many similar studies of ac attempt to fit to just the rotation velocity at 2.2 disc scalelengths , @xmath118 ) .
this highlights the usefulness of extended rotation curve data in constraining general models of galaxy formation .
having rotation curve data out to large radii is essential for determining dark matter density profiles .
studies that test the applicability of adiabatic contraction often make use of large samples of galaxies that are dominated by late - type disc - dominated galaxies , with small bulges or no bulge at all ( e.g. pizagno et al .
2005 ; dutton et al . 2005 , 2007 ;
gnedin et al .
2006 ; kassin et al .
2006a , b ) .
these `` bulgeless '' galaxies also tend to have very flat ( or even rising ) rotation curves , and it is therefore easy to explain these galaxy rotation curves without any need for adiabatic contraction .
however , we suggest that early - type disc galaxies ( e.g. hubble types sa sb ) , which have bulge - dominated centres , and rotation curves that fall - off due to the presence of a large bulge , may require adiabatic contraction .
although we have studied very few galaxies so far ( m31 in this paper and two other galaxies in seigar et al .
2006 ) our results indicate that in cases where rotation curves have a fall - off ( two out of three cases ) , adiabatic contraction is favoured
. this may be true for all galaxies with large bulges .
if so , it may favour a secular ( rather than merger - driven ) origin for these bulge - dominated systems , as the gradual accumulation of central mass increases the likelihood that ac will operate .
it should also be noted that in our best - fit model , while it is consistent with the current hi data from carignan et al .
( 2006 ) , it is unclear if the rotation curve is declining or remaining flat past 30 kpc , due to the large error bars .
the most important requirement for future kinematical studies of galaxies would be improved hi measurements at large radii .
this will enable us to put better constraints on the halo properties of m31 and other spiral galaxies .
support for this work was provided by nasa through grant number hst - ar-10685.01-a from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
mss acknowledges partial support from a gary mccue fellowship through the center for cosmology at uc irvine .
mss also acknowledges the support of the college of science and mathematics at the university of arkansas at little rock .
research by ajb is supported in part by nsf grant ast-0548198 .
research by jsb is supported by nsf grants ast-0507916 and ast-0607377 .
this work is based in part on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
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 .
the authors wish to thank tom jarrett , luis ho , stephane courteau and steve majewski for their discussions .
we also wish to express gratitude to pauline barmby , for supplying us with the most recent mosaic of the spitzer 3.6 @xmath0 m image of m31 .
finally , the authors wish to thank the anonymous referees for comments which improved the content of this paper .
99 alam s. m. k. , bullock j. s. , weinberg d. h. , 2002 , apj , 572 , 34 andredakis y. c. , peletier r. f. , balcells m. , 1995 , mnras , 275 , 874 athanassoula e. , beaton r. l. , 2006 , mnras , 370 , 1499 barmby p. , et al .
, 2006 , apj , 650 , l45 barnes j. e. , 1987 , in _ nearly normal galaxies : from the planck time to the present _ , ed .
s. m. faber et al .
( new york : springer verlag ) , p 154 beaton r. l. , et al . , 2007 , apj , 658 , l91 bell e. f. , de jong r. s. , 2001 , apj , 550 , 212 bell e. f. , mcintosh d. h. , katz n. , weinberg m. d. , 2003 , apj , 585 , 117 bender r. , et al .
, 2005 , apj , 631 , 280 benson a. j. , bower r. g. , frenk c. s. , lacey c. g. , baugh c. m. , cole s. , 2003 , apj , 599 , 38 bertin e. , arnouts s. , 1996 , a&as , 117 , 393 binney j. , tremaine s. , 1987 , _ galactic dynamics _ ,
( princeton : princeton university press ) blais - ouellette s. , amram p. , carignan c. , 2001 , aj , 121 , 1952 blumenthal g. r. , faber s. m. , flores r. , primack j. r. , 1986 , apj , 301 , 27 ( b86 ) braun r. , 1991 , apj , 372 , 54 brewer j. p. , richer , h. b. , crabtree d. r. , 1995 , aj , 109 , 2480 brinks e. , burton w. b. , 1984 , a&a , 141 , 195 bruzual g. , charlot s. , 2003 , mnras , 344 , 1000 ( bc03 ) bullock j. s. , dekel a. , kolatt t. s. , kravtsov a. v. , klypin a. a. , porciani c. , primack j. r. 2001a , apj , 555 , 240 bullock j. s. , kolatt t. s. , sigad y. , somerville r. s. , kravtsov a. v. , klypin a. a. , primack j. r. , dekel a. , 2001b mnras , 321 , 559 caon n. , capaccioli m. , donofrio m. , 1993 , mnras , 265 , 1013 capaccioli m. , 1989 , in _ the world of galaxies _ , ed .
h. g. corwin , l. bottinelli ( berlin : springer - verlag ) , p 208 carignan c. , chemin l. , huchtmeier w. k. , lockman , f. j. , 2006 , apj , 641 , l109 cembranos j. a. , feng , j. l. , rajaraman a. , takayama f. , 2005 , phys .
lett . , 95 , 181301 cesarsky d. , lequeux j. , pagani l. , ryter c. , loinard l. , sauvage m. , 1998 , a&a , 337 , l35 chapman s. c. , ibata r. , lewis g. f. , ferguson a. m. n. , irwin m. , mcconnachie a. , tanvir n. , 2006 , apj , 653 , 255 chapman s. c. , et al .
, 2007 , apj , 622 , l79 choi j .- h . , lu y. , mo h. j. , weinberg m. d. , 2006 , mnras , 372 1869 cole , s. , lacey , c. g. , baugh c. m. , frenk c. s. , 2000 , mnras , 319 , 168 ct p. , mateo m. , sargent w. l. w. , olszewski e. w. , 2000 , apj , 537 , l91 de jong r. s. , 1996 , a&as , 118 , 557 de vaucouleurs g. , 1948 , ann .
, 11 , 247 de vaucouleurs g. , 1959 , handbuch der physik , 53 , 275 & 311 de vaucouleurs g. , de vaucouleurs a. , corwin h. g. , buta r. j. , paturel g. , fouqu p. , 1991
, _ the third reference catalog of bright galaxies _ , ( new york : springer ) ( rc3 ) donofrio m. , 2001 , mnras , 326 , 1517 dutton a. a. , courteau s. , de jong r. , carignan , c. , 2005 , apj , 619 , 218 dutton a. a. , van den bosch f. c. , dekel a. , courteau s. , 2007 , apj , 654 , 27 einasto j. , 1972 , in _ external galaxies and quasi - stellar objects _ , iau symp . 144 , ed .
d. e. evans ( dordrecht : reidel ) , p 37 evans n. w. , wilkinson m. i. , 2000 , mnras , 316 , 929 fardal m. a. , babul a. , geehan j. j. , gahathakurta p. , 2006 ,
mnras , 366 , 1012 fazio g. , et al . , 2004 , apjs , 154 , 10 flores r. a. , primack j. r. , 1994 , apj , 427 , l1 flores r. , primack j. r. , blumenthal g. r. , faber s. m. , 1993 , apj , 412 , 443 fluks m. a. , plez b. , the p. s. , de winter d. , westerlund b. e. , steenman h. c. , 1994 , a&as , 105 , 311 geehan j. j. , fardal m. a. , babul a. , gahathakurta p. , 2006 ,
mnras , 366 , 996 gnedin o. y. , kravstov a. v. , klypin a. a. , nagai d. , 2004 , apj , 616 , 16 ( g04 ) gnedin o. y. , weinberg d. h. , pizagno j. , prada f. , rix , h .- w .
, 2007 , apj , 671 , 1115 gonzalez a. h. , williams k. a. , bullock j. s. , kolatt t. s. , primack j. r. , 2000 , apj , 528 , 145 gordon k. d. , et al . , 2006 ,
apj , 638 , l87 ciotti l. , 1991 , a&a , 249 , 99 graham a. w. , 2001a , mnras , 326 , 543 graham a. w. , 2001b , aj , 121 , 820 graham a. w. , driver s. , 2005 , pasa , 22 , 118 helou g. , lu n. y. werner m. w. , malhotra s. , silbermann n. , 2000 , apj , 532 , l21 hernquist l. , 1990 , apj , 356 , 359 holland s. , 1998 , aj , 115 , 1916 humphrey p. j. , buote d. a. , gastaldello f. , zappacosta l. , bullock j. s. , brighenti f. , mathews w. g. , 2006 , apj , 646 , 899 ibata r. , chapman s. , ferguson a. m. n. , lewis g. , irwin m. , tanvir n. , 2005 , apj , 634 , 287 james p. a. , seigar m. s. , 1999 , a&a , 350 , 791 jedrzejewski r. i. , 1987 , mnras , 226 , 747 jesseit r. , naab t. , burkert a. , 2002 , apj , 571 , l89 kaplinghat m. , knox l. , turner m. s. , 2000 , phys .
, 85 , 3335 kaplinghat m. , 2005 , phys .
d , 72 , 063510 kassin s. a. , de jong r. s. , pogge r. w. , 2006a , apjs , 162 , 80 kassin s. a. , de jong r. s. , weiner r. j. , 2006b , apj , 643 , 804 kent s. m. , 1989 , aj , 97 , 1614 kent s. m. , huchra j. p. , stauffer j. , 1989 , aj , 98 , 2080 khosroshahi h. g. , wadadekar y. , kembhavi a. , 2000 , apj , 533 , 162 klypin a. , zhao h. , somerville r. s. , 2002 , apj , 573 , 597 kormendy j. , 1988 , apj , 325 , 128 kormendy j. , bender r. , 1999 , apj , 522 , 772 kroupa p. , 2001 ,
mnras , 322 , 231 kuzio de naray r. , mcgaugh s. s. , de blok w. j. g. , bosma a. , 2006 , apjs , 165 , 461 light e. s. , danielson r. e. , schwarschild , m. , 1974 , apj , 194 , 257 loinard l. , allen r. j. , lequeux j. , 1995 , a&a , 301 , 68 macci a. v. , moore b. , stadel j. , diemand j. , 2006 , mnras , 366 , 1529 majewski s. r. , et al . , 2007 , apj , 670 , l9 maraston c. , 2005 , mnras , 362 , 799 ( m05 ) maraston c. , daddi e. , renzini a. , cimatti a. , dickinson m. , papovich c. , pasquali a. , pirzkal n. , 2006 , apj , 652 , 85 merrett h. r. , et al .
, 2006 , mnras , 369 , 120 mllenhoff c. , heidt j. , 2001 , a&a , 368 , 16 moore b. , 1994 , nature , 370 , 629 moore b. , quinn t. , governato f. , stadel j. , lake g. , 1999 , mnras , 310 , 1147 navarro j. f. , frenk c. s. , white , s. d. m. , 1997 , apj , 462 , 563 ( nfw ) pickles a. j. , 1998 , pasp , 110 , 863 pizagno j. , et al . , 2005 , apj , 633 , 844 prugniel p. , simien f. , 1997 , a&a , 321 , 111 rodriguez - fernandez n. j. , braine j. , brouillet n. , combes f. , 2006 , a&a , 453 , 77 rubin v. c. , ford w. k. , 1970 , apj , 159 , 379 ryden b. s. , gunn j. e. , 1987 , apj , 318 , 15 schlegel d. j. , finkbeiner d. p. , davis m. , 1998 , apj , 500 , 525 schultheis m. , aringer b. , jorgensen u. g. , lebzelter t. , plez b. , 1999 , in hron j. , hfner s. , eds , abstract of the 2nd austrian iso workshop : atmospheres of m , s , and c giants , p. 93
seigar m. s. , james p. a. , 1998a , mnras , 299 , 672 seigar m. s. , james p. a. , 1998b , mnras , 299 , 685 seigar m. s. , block d. l. , puerari i. , 2004 , in _ penetrating bars through masks of cosmic dust _ , ed .
d. l. block et al .
( dordrecht : springer ) , p 155 seigar m. s. , 2005 , mnras , 361 , l20 seigar m. s. , block d. l. , puerari i. , chorney n. e. , james p. a. , 2005 , mnras , 359 , 1065 seigar m. s. , bullock j. s. , barth a. j. , ho l. c. , 2006 , apj , 645 , 1012 sellwood j. a. , mcgaugh s. s. , 2005 , apj , 634 , 70 srsic j .-
l . , 1963 , boletin de la asociacion argentina de astronomia , 6 , 41 srsic j .-
l . , 1968 , atlas de galaxias australes ( cordoba : observatorio astronomico ) simon j. d. , bolatto a. d. , leroy a. , blitz l. , gates e. l. , 2005 , apj , 621 , 757 somerville r. s. , primack j. r. , 1999 , mnras , 310 , 1087 strigari l. e. , kaplinghat m. , bullock j. s. , 2007 , phys .
lett . , 75 , 061303 swaters r. a. , madore b. f. , van den bosch f. c. , balcells m. , 2003 , apj , 583 , 732 tamm a. , tempel e. , tenjes p. , 2007 ,
mnras , submitted ( astro - ph/0707.4375 ) tempel e. , tamm a. , tenjes p. , 2007 ,
mnras , submitted ( astro - ph/0707.4374 ) trujillo i. , ramos a. a. , rubio - martn j. a. , graham a. w. , aguerri j. a. l. , cepa j. , gutirrez , c. m. , 2002 , mnras , 333 , 510 tully r. b. , fisher j. r. , 1977 , a&a , 54 , 661 valenzuela o. , rhee g. , klypin a. , governato f. , stinson g. , quinn t. , wadsley j. , 2007 , apj , 657 , 773 van den bosch f. c. , swaters r. a. , 2001 , mnras , 325 , 1017 walterbos r. a. m. , kennicutt r. c. , 1987 , a&as , 69 , 311 wechsler r. h. , bullock j. s. , primack j. r. , kravtsov a. y. , dekel a. , 2002 , apj , 568 , 52 weinberg d. h. , colombi s. , dav r. , katz n. , 2008 , apj , 678 , 6 widrow l. m. , perrett k. m. , suyu s. h. , 2003 , apj , 588 , 311 worthey g. , espaa a. , macarthur l. a. , courteau s. , 2005 , apj , 631 , 820 yang x. , mo h. j. , van den bosch f. c. , 2003 , mnras , 339 , 1057 zappacosta l. , buote d. a. , gastaldello f. , humphrey p. j. , bullock j. , brighenti f. , mathews w. , 2006 , apj , 650 , 777 zentner a. r. , bullock j. s. , 2002 , phys . rev .
d , 66 , 043003
because m31 is a highly inclined galaxy , it is difficult to learn how tightly wound its spiral arm pattern is from imaging data .
however , seigar , block & puerari ( 2004 ) and seigar et al .
( 2005 , 2006 ) have shown that there is a very tight correlation between the spiral arm pitch angle ( i.e. how tightly wound the spiral arms are ) and rotation curve shear ( measured at a 10 kpc radius ) .
therefore the spiral arm pitch angle can be estimated from the shear measured for the m31 rotation curve .
rotation curve shear is defined as @xmath207 where @xmath208 is the first oort constant , @xmath209 is the angular velocity and @xmath210 is the rotational velocity at radius @xmath192 .
the shear rate depends upon the shape of the rotation curve . for a rotation curve that remains flat the shear rate , @xmath211 , for a falling rotation curve the shear rate , @xmath212 and for a continually rising rotation curve the shear rate , @xmath213 . following the method described by seigar et al .
( 2006 ) we measure the shear at a 10 kpc radius .
for m31 , 10 kpc is equivalent to @xmath214 .
the shear is measured using the prescription described by seigar ( 2005 ) and seigar et al .
( 2004 , 2005 , 2006 ) . in this method
an average slope is fit to the outer part of the rotation curve ( i.e. , past the solid - body rotation regime ) , and the shear is then calculated at the given radius . for m31 , at a radius of 10 kpc , the shear is @xmath215 .
there is little scatter in the correlation between shear rate and pitch angle , and the line of best fit is given by @xmath216 where @xmath217 is the spiral arm pitch angle in degrees and @xmath218 is the shear . from this the pitch angle for m31
is @xmath6 .
( it should be noted that the shear at 10 kpc determined for the m1 b86 model in section 2.2 is @xmath219 and for m1 g05 is @xmath220 .
these estimates are consistent with the shear measured here .
they result in pitch angles of @xmath221 and @xmath222 respectively , both of which are consistent with the pitch angle derived above ) . assuming a logarithmic spiral ( see e.g. , seigar & james 1998b for the equations that define a logarithmic spiral ) , the spiral arm pattern of m31 is calculated .
figure [ spiral ] shows the predicted spiral arm pattern of m31 for a face - on view ( _ left _ ) and with an inclination angle , @xmath223 , and a position angle @xmath224 ( de vaucouleurs et al .
1991 ) appropriate for m31 ( _ centre _ ) .
it should be noted that the position and inclination angles are not fitted parameters , they are taken from the literature .
the spiral arm pitch angle calculated here is larger than the hi spiral arm pitch angle of 16@xmath90 calculated by braun ( 1991 ) , although it is possible that the spiral arm pattern in the neutral gas may be a few degrees tighter than that observed in the underlying stellar distribution .
a pitch angle of 16@xmath90 would be a @xmath242@xmath13 outlier in the spiral arm pitch angle versus rotation curve shear correlation when the shear , @xmath225 ( see figure 3 of seigar et al .
if this point was added to the correlation , it would be one of the larger outliers , yet the correlation would still be strong . for comparison
the 24 @xmath0 m image is shown in the right panel of figure [ spiral ] .
it is important to note that the spiral arm pattern shown in figure [ spiral ] does not take into account the disc scale height and how broad the arms are .
these plots are intended to illustrate show how tightly wound the arm pattern would be if viewed from face - on . when viewed close to edge - on ( as is the case for m31 ) ,
the scale height of the disc and breadth of the spiral arms make it virtually impossible to determine the spiral arm pitch angle directly . | we present an updated mass model for m31 that makes use of a spitzer 3.6 @xmath0 m image , a mass - to - light ratio gradient based on the galaxy s @xmath1 colour profile , and observed rotation curve data from a variety of sources .
we examine cases where the dark matter follows a pure nfw profile and where an initial nfw halo contracts adiabatically in response to the formation of the galaxy .
we find that both of these scenarios can produce a reasonable fit to the observed rotation curve data .
however , a pure nfw model requires a concentration @xmath2 that is well outside the range predicted in @xmath3cdm cosmology and is therefore disfavoured .
an adiabatically contracted nfw halo favors an initial concentration @xmath4 and virial mass @xmath5 , and this is in line with the cosmological expectations for a galaxy of the size of m31 .
the best - fit mass is consistent with published estimates from andromeda stream kinematics , satellite galaxy radial velocities , and planetary nebulae studies .
finally , using the known linear correlation between rotation curve shear and spiral arm pitch angle , we show that the stellar spiral arm pitch angle of m31 ( which can not be deduced from imaging data due to the galaxy s inclination ) is @xmath6 .
[ firstpage ] dark matter galaxies : fundamental parameters galaxies : haloes galaxies : individual ( m31 ) galaxies : spiral galaxies : structure |
higgs production has been shown to be sensitive to the polarization of gluons , even in collisions between unpolarized protons such as at the lhc @xcite .
gluons with nonzero transverse momentum with respect to the proton momentum can be linearly polarized @xcite , which affects for instance the transverse momentum distribution of produced higgs bosons .
although the amount of polarization is currently unknown , it is known that it is at the very least perturbatively generated @xcite and , therefore , nonzero .
there are also strong indications that at small momentum fractions @xmath1 the linear polarization becomes maximal @xcite .
if sufficiently large , the polarization offers a new tool to analyze higgs couplings to the various standard model particles into which it can decay @xcite . to describe the effects of this gluon polarization , the formalism of transverse momentum dependent ( tmd ) parton distribution functions ( or tmds , for short )
is natural to consider , cf .
e.g. @xcite .
this has been studied for the particular case of higgs production in refs .
@xcite and including the effects of tmd evolution in refs .
it turns out that at the higgs mass scale @xmath2 of about 125 gev , the effects of linear gluon polarization are not expected to be large , at the few percent level most likely .
moreover , the effects are largest at small values of the transverse momentum of the higgs , i.e. a few gev , where the cross section is difficult to measure . in the present paper
we consider an alternative offered by the production of a higgs in association with an additional jet , which has been widely studied without including gluon polarization , e.g. refs .
the effect of gluon polarization shows up in the transverse momentum distribution of the higgs plus jet pair , where the pair transverse momentum can be of the order of a few gev , while the separate transverse momenta of the higgs and the jet can be substantially larger .
the invariant mass of the higgs plus jet system will be even larger than @xmath2 , but the advantage is that a range of scales is now accessible , as opposed to the very narrow range around @xmath2 accessible in higgs production . in principle , this range of scales offers a way to map out the tmd evolution , although the feasibility in practice remains to be seen .
experimentally the limiting factors are the resolution of the transverse momentum of both the higgs and the jet and how well the jet direction coincides with that of the fragmenting parton . at cms
the jet transverse momentum resolution at 10 gev is typically 1.5 gev and at 100 gev it is 8 gev @xcite .
the ultimately achievable resolution on the higgs transverse momentum is not clear , but it is likely multiple gev . on an event - by - event basis
the deviation of the jet axis as obtained by jet finding algorithms from the direction of the fragmenting parton , can also be as large as a few gev in transverse momentum @xcite .
altogether these uncertainties in the pair transverse momentum can be substantial and the goal of obtaining several bins in the region up to say 10 gev will be quite challenging .
a numerical study using a monte carlo simulation will have to be done to study the actual feasibility , but that is beyond the scope of this paper . here
we focus on the cross section expressions and on the differences of the higgs plus jet process to higgs production and to some other similar processes , pointing out the advantages it in principle has to offer .
in this paper we present the relevant expressions for higgs plus jet production in leading order and study the impact of the gluon polarization in two models for the gluon distributions involved .
both models have the advantage that they allow to obtain analytic expressions , but we will mostly present numerical results to show the qualitative differences between the two cases more clearly .
we also present results for angular distributions , which have the advantage of singling out specific contributions .
although measurements of angular distributions generally require large statistics , probing a nonzero result may nevertheless be possible when integrating over transverse momenta up to some maximum value as suggested in ref .
we study the process @xmath3 where the four - momenta of the particles are given within round brackets , and the higgs boson and jet in the final state are produced with momenta that have components in the plane orthogonal to the direction of the initial protons that are almost back to back . to leading order in perturbative qcd
the reaction proceeds via the partonic subprocesses @xmath4 with parton @xmath5 fragmenting into the observed jet .
specifically , the following channels can contribute : @xmath6 , @xmath7 and @xmath8 @xcite .
the corresponding feynman diagrams are depicted in fig .
[ fig : fd ] . in the calculation of the scattering amplitudes ,
we take the quark masses to be zero , except for the top quark mass @xmath9 .
therefore the higgs boson can couple to gluons only via a top quark loop .
we consider the limit @xmath10 in which this coupling can be approximated by a point interaction
. the corresponding feynman rules of the effective lagrangian can be found , for example , in ref .
furthermore , we perform a lightcone decomposition of the two incoming hadronic momenta , @xmath11 and @xmath12 , in terms of the light - like vectors @xmath13 and @xmath14 , which satisfy the relations @xmath15 and @xmath16 : @xmath17 the partonic momenta @xmath18 and @xmath19 can be expressed in terms of the lightcone momentum fractions ( @xmath20 , @xmath21 ) and the intrinsic transverse momenta ( @xmath22 , @xmath23 ) , as follows @xmath24 using @xmath13 and @xmath14 the lightcone components of any vector @xmath25 are defined as @xmath26 , while @xmath27 refers to the components of @xmath25 orthogonal to the proton momenta @xmath11 and @xmath12 .
moreover , one has @xmath28 .
therefore in eq .
( [ partondecompositions ] ) , if we neglect the proton mass , @xmath29 and @xmath30 .
we assume that , at sufficiently high energies , tmd factorization @xcite holds for the process in eq .
( [ eq : proc ] ) , hence its cross section is given by the convolution of one soft , partonic correlator for each proton and a hard part , @xmath31}_a(x_a { , } \bm p_{a { { \scriptscriptstyle t } } } ) \phi^{[u]}_b(x_b { , } \bm p_{b { { \scriptscriptstyle t } } } ) \left|{\cal m}^{a b \to h c } ( p_a , p_b ; k_h , k_{\rm j})\right|^2\right \}\ , , \label{crosssec}\end{aligned}\ ] ] with @xmath32 being the total energy squared in the hadronic center - of - mass frame . the sum in eq .
( [ crosssec ] ) runs over all the partons that take part in the reaction , the appropriate trace is taken over dirac and lorentz indices , and @xmath33 denotes the amplitude for the process @xmath34 .
the parton correlators @xmath35}_{a , b}$ ] describe the hadron @xmath36 parton transitions .
they can be parameterized in terms of tmds and are defined in terms of qcd operators on the lightfront ( lf ) : @xmath37 , where @xmath38 for parton @xmath39 with momentum @xmath40 and @xmath41 for parton @xmath42 with momentum @xmath43 .
specifically , at leading twist the quark correlator for an unpolarized hadron can be written as @xcite @xmath44}}(x { , } \bm p_{{\scriptscriptstyle t } } ) = { \int}\frac{{\mathrm{d}}(\xi{\cdot}p)\,{\mathrm{d}}^2\xi_{{\scriptscriptstyle t}}}{(2\pi)^3}\ e^{ip\cdot\xi}\ , \langle p |\,\overline\psi_j(0)\,u_{[0,\xi]}\ , \psi_i(\xi)\,|p\rangle\,\big\rfloor_{\text{lf } } = \frac{1}{2}\ , \bigg \{\,f_1^q(x{,}\bm{p}_{{\scriptscriptstyle t}}^2)\;\slash p_{ij } + i h_1^{\perp\,q}(x{,}\bm{p}_{{\scriptscriptstyle t}}^2)\;\frac{[\slash p_{{\scriptscriptstyle t } } , \slash p]_{ij}}{2 m_p}\bigg \ } , \end{aligned}\ ] ] where @xmath45}$ ] is the process dependent gauge link connecting the two quark fields , which renders the correlator gauge invariant . furthermore , @xmath46 is the tmd describing unpolarized quarks inside an unpolarized hadron , and @xmath47 , commonly referred to as the boer - mulders function , is the time - reversal ( @xmath48 ) odd distribution of transversely polarized quarks inside an unpolarized hadron @xcite .
analogously , for an antiquark one has @xmath49}}(x { , } \bm p_{{\scriptscriptstyle t } } ) = -{\int}\frac{{\mathrm{d}}(\xi{\cdot}p)\,{\mathrm{d}}^2\xi_{{\scriptscriptstyle t}}}{(2\pi)^3}\ e^{-ip\cdot\xi}\ , \langle p |\,\overline\psi_j(0)\,u_{[0,\xi]}\ , \psi_i(\xi)\,|p\rangle\,\big\rfloor_{\text{lf } } = \frac{1}{2}\ , \bigg \{\,f_1^{\bar q}(x{,}\bm{p}_{{\scriptscriptstyle t}}^2)\;\slash p_{ij } + i h_1^{\perp\,\bar q}(x{,}\bm{p}_{{\scriptscriptstyle t}}^2)\;\frac{[\slash p_{{\scriptscriptstyle t } } , \slash p]_{ij}}{2 m_p}\bigg \}\ , .
\end{aligned}\ ] ] the definition of the gluon correlator in terms of the gluon field strength @xmath50 has been given for the first time in ref .
@xcite . for an unpolarized hadron ,
using the naming convention of ref .
@xcite , one has @xmath51\,\mu\nu}(x,\bm p_{{\scriptscriptstyle t } } ) & = & \frac{n_\rho\,n_\sigma}{(p{\cdot}n)^2 } { \int}\frac{{\mathrm{d}}(\xi{\cdot}p)\,{\mathrm{d}}^2\xi_{{\scriptscriptstyle t}}}{(2\pi)^3}\ e^{ip\cdot\xi}\ , \langle p|\,\operatorname{tr}\big[\,f^{\mu\rho}(0)\,u_{[0,\xi]}\ , f^{\nu\sigma}(\xi)\,u^{\prime}_{[\xi,0]}\,\big ] \,|p \rangle\,\big\rfloor_{\text{lf } } \nonumber \\ & = & -\frac{1}{2x}\,\bigg \{g_{{\scriptscriptstyle t}}^{\mu\nu}\,f_1^g ( x,\bm p_{{\scriptscriptstyle t}}^2 ) -\bigg(\frac{p_{{\scriptscriptstyle
t}}^\mu p_{{\scriptscriptstyle t}}^\nu}{m_p^2}\ , { + } \,g_{{\scriptscriptstyle t}}^{\mu\nu}\frac{\bm p_{{\scriptscriptstyle t}}^2}{2m_p^2}\bigg ) \;h_1^{\perp\,g } ( x,\bm p_{{\scriptscriptstyle t}}^2 ) \bigg \ } , \label{phig}\end{aligned}\ ] ] with the transverse tensor @xmath52 defined as @xmath53 .
the function @xmath54 is the unpolarized gluon distribution and @xmath55 the distribution of linearly polarized gluons , which satisfies the model - independent positivity bound @xcite , @xmath56 valid for all values of @xmath1 and @xmath57 .
in contrast to @xmath58 , @xmath59 is @xmath48-even , hence it can be nonzero also in absence of initial and/or final state interactions .
however , as any other tmd , @xmath59 can in principle receive contributions from these interactions , which can render it process - dependent and even hamper its extraction for processes where factorization does not hold , such as dijet production in hadron - hadron collisions @xcite .
the cross section is calculated using eq .
( [ crosssec ] ) , in which we insert the parameterizations of the correlators in eqs .
( [ quarkcorr])-([gluoncorr ] ) and the explicit expressions of the amplitudes @xmath60 . from the @xmath61 function in eq .
( [ crosssec ] ) , with @xmath62 , the lightcone momentum fractions @xmath63 can be expressed in terms of the rapidities ( @xmath64 , @xmath65 ) and transverse momenta ( @xmath66 , @xmath67 ) of the produced higgs boson and jet , respectively : @xmath68 valid with corrections of order @xmath69 and where we have introduced the transverse mass @xmath70 . if we neglect terms suppressed by powers of @xmath71 , the final result has the form @xmath72\ , , \label{eq : cs}\ ] ] where we have introduced the sum and difference of the final transverse momenta , @xmath73 and @xmath74 .
moreover , @xmath25 is the vacuum expectation value and @xmath75 is the azimuthal angle between @xmath76 and @xmath77 , namely @xmath78 .
the functions @xmath79 , @xmath80 , and @xmath81 contain convolutions of the various tmds and , besides @xmath82 , they depend also on the mandelstam variables @xmath83 , @xmath84 , @xmath85 for the partonic subprocesses in eq .
( [ eq : subproc ] ) , which satisfy the relations @xmath86 with @xmath87 .
the explicit expressions for @xmath79 , @xmath80 and @xmath81 are provided in the following three subsections .
the term @xmath79 in eq .
( [ eq : cs ] ) is given by the sum of contributions from the relevant partonic subprocesses , i.e. @xmath88 where @xmath89\,,\label{eq : afgg}\\ { \cal a}_f^{gq \to hq } & = & \frac{1}{2 n_c } \left \ { - \frac{\hat s^2 + \hat u^2}{\hat t } \,{\cal c}[f_1^g\,f_1^q ] - \frac{\hat s^2 + \hat t^2}{\hat u } \
, { \cal c}[f_1^q\,f_1^g ] \right \}\,,\\ { \cal a}_f^{q\bar q \to hg } & = & \frac{n_c^2 - 1 } { 2 n_c^2 } \ , \frac{\hat t^2 + \hat u^2}{\hat s } \ , { \cal c}[f_1^q\,f_1^{\bar q } ] \label{eq : afqq}\ , ,
\\ { \cal a}_h^{gg\to hg } & = & \frac{n_c}{n_c^2 - 1}\,m_h^4\ , \frac{\hat s}{\hat t \hat u } \,{\cal c}[w_0^{hh}\,h_1^{\perp\,g}h_1^{\perp\,g } ] = \frac{1}{9}\,\frac{n_c}{n_c^2 - 1}\,\frac{m_h^4}{\bm k_\perp^2 } \,{\cal c}[w_0^{hh}\,h_1^{\perp\,g}h_1^{\perp\,g}]\ , .
\label{eq : ahgg}\end{aligned}\ ] ] in eqs .
( [ eq : afgg])-([eq : afqq ] ) , @xmath90 is the number of colors and we have introduced the convolutions of tmds @xmath91 & \equiv & \int d^{2}\bm p_{a{{\scriptscriptstyle t}}}\int d^{2}\bm p_{b{{\scriptscriptstyle t}}}\ , \delta^{2}(\bm p_{a{{\scriptscriptstyle t}}}+\bm p_{b{{\scriptscriptstyle t}}}-\bm q_{{{\scriptscriptstyle t } } } ) \ , w(\bm p_{a{{\scriptscriptstyle t}}},\bm p_{b{{\scriptscriptstyle t}}})\ , f(x_{a},\bm p_{a{{\scriptscriptstyle t}}}^{2})\ , f(x_{b},\bm p_{b{{\scriptscriptstyle t}}}^{2})\,,\label{eq : conv}\end{aligned}\ ] ] with the transverse weight @xmath92 given by @xmath93\ , .
\label{eq : w0}\ ] ] the expressions in eqs .
( [ eq : afgg])-([eq : afqq ] ) are in full agreement with the unpolarized partonic cross sections calculated for the first time in ref .
the term in eq .
( [ eq : ahgg ] ) , due to the presence of linearly polarized gluons inside an unpolarized proton , is a new result , similar to the modifications of the transverse momentum distribution of higgs bosons @xcite and ( pseudo)scalar quarkonia @xcite inclusively produced in hadronic collisions . similarly to eq .
( [ eq : a ] ) , the term @xmath80 in eq .
( [ eq : cs ] ) can be written as @xmath95 where @xmath96 \right \ } + ( x_a \leftrightarrow x_b , \hat t \leftrightarrow \hat u ) \
, , \label{eq : bgg}\\ { \cal b}^{gq\to hq } & = & \frac{1}{2 n_c}\ , \left \ { \frac{\hat
s \hat u}{\hat t}\ , { \cal c}[w_2^{fh } \,f_1^q\,h_1^{\perp\,g } ] + \frac{\hat s \hat t}{\hat u}\ , { \cal c}[w_2^{hf } \
, h_1^{\perp\,g}\,f_1^q ] \right \}\ , , \label{eq : bgq}\\ { \cal b}^{q\bar q\to h g } & = & \frac{n_c^2 - 1}{2 n_c}\ , \frac{\hat t \hat u}{\hat s}\ , 2\,{\cal c}[w_2^{hh } \,h_1^{\perp\ , q}\,h_1^{\perp \,\bar q } ] \,,\end{aligned}\ ] ] and the transverse weights read @xmath97 \,,\\ w_2^{hf}\ , & = & \ , \frac{1}{m_p^2}\,\left [ 2\ , \frac{(\bm q_{{\scriptscriptstyle t}}\cdot \bm p_{a{{\scriptscriptstyle t}}})^2}{\bm q_{{\scriptscriptstyle t}}^2}-\bm p_{a{{\scriptscriptstyle t}}}^2 \right ] \,,\\
w_2^{hh } & = & \frac{1}{m_p^2}\,\left [ 2\,\frac{(\bm q_{{{\scriptscriptstyle t } } } \cdot \bm p_{a{{\scriptscriptstyle t}}})\cdot ( \bm q_{{{\scriptscriptstyle t } } } \cdot \bm p_{b{{\scriptscriptstyle t}}})}{\bm q_{{{\scriptscriptstyle t}}}^2 } -\bm p_{a{{\scriptscriptstyle t}}}\cdot \bm p_{b{{\scriptscriptstyle t } } } \right]\,.\end{aligned}\ ] ] a @xmath0 double boer - mulders ( quark ) contribution with transverse weight @xmath98 has been found for the drell - yan process as well , and it is expected to lead to a violation of the lam - tung relation @xcite .
a similar asymmetry has been predicted for photon+jet @xcite and dijet production @xcite in proton - proton collision .
a @xmath0 modulation due to the convolution of unpolarized and polarized gluon distributions as in eq .
( [ eq : bgg ] ) has been predicted also for the inclusive hadroproduction of diphotons @xcite , dijets @xcite , heavy quark pairs @xcite , and @xmath99+photon pairs @xcite .
the only channel that contributes to the @xmath101 modulation of the cross section is @xmath6 .
therefore we can write @xmath102 = \frac{1}{9}\,\frac{n_c}{n_c^2 - 1}\,\bm k_\perp^2 \,{\cal c}[w_4^{hh}\,h_1^{\perp\,g}h_1^{\perp\,g}]\,,\end{aligned}\ ] ] with @xmath103 ^ 2 - \bm p_{a{{\scriptscriptstyle t}}}^2 \bm p_{b{{\scriptscriptstyle t}}}^2
\right \}~. \end{aligned}\ ] ] an analogous @xmath104 modulation has been predicted for the first time for the inclusive hadroproduction dijets @xcite , and subsequently also for reactions with diphotons @xcite , heavy quark pairs @xcite , and @xmath99+photon pairs @xcite in the final state .
the ( normalized ) cross section for the process @xmath105 , differential in @xmath106 and @xmath75 , is defined as @xmath107 where , restricting now only to the subprocess @xmath6 , @xmath108 is given by @xmath109\ , .
\label{eq : csgg}\ ] ] by substituting into eq .
( [ eq : qtdist ] ) the expression for @xmath108 in eq .
( [ eq : csgg ] ) , one obtains @xmath110\ , , \label{eq : csgg-2}\end{aligned}\ ] ] with @xmath111}{\int_0^{{q^2_{{{\scriptscriptstyle t}}\rm max } } } { \mathrm{d}}\bm q^2_{{\scriptscriptstyle t}}\ , { \cal c}[f_1^g \ , f_1^g ] } \ , , \label{eq : sigma0}\ ] ] and @xmath112 } { { \cal c}[f_1^g \ ,
f_1^g ] } \ , , \label{eq : r0 } \\
r_2(\bm q_{{\scriptscriptstyle t}}^2 ) & = & \frac { \hat t ^2(\hat t + \hat u)^2 -2 m_h^2 \hat u^2 ( \hat t + \hat u ) + m_h^4 ( \hat t^2 + \hat u^2 ) } { m_h^8
+ \hat s^4 + \hat t^4 + \hat u^4 } \ , \frac{{\cal c}[w_2^{fh } \,f_1^g\,h_1^{\perp\,g}]}{{\cal c}[f_1^g \ , f_1^g ] } + ( x_a \leftrightarrow x_b , \hat t \leftrightarrow \hat u ) \ , , \label{eq : r2}\\ r_4(\bm q_{{\scriptscriptstyle t}}^2 ) & = & \frac{\hat t^2 \hat u^2}{m_h^8+\hat s^4 + \hat t^4 + \hat u^4 } \,\frac{{\cal c}[w_4^{hh}\,h_1^{\perp\,g}h_1^{\perp\,g}]}{{\cal c}[f_1^g \ , f_1^g ] } \ , .
\label{eq : r4}\end{aligned}\ ] ] it is possible to single out the different terms @xmath113 , @xmath114 , @xmath115 in eq .
( [ eq : csgg-2 ] ) by defining the observables @xcite @xmath116 such that the average values of @xmath117 are given by the integrals of @xmath118 over @xmath82 , @xmath119 we will comment on the value of @xmath120 in the next section .
it can be easily shown that @xmath121\ , , \label{eq : qt } \\
\langle \cos 2\phi \rangle_{q_{{\scriptscriptstyle t } } } & = & \frac{1}{2}\ , \sigma_0(\bm q_{{{\scriptscriptstyle t}}}^2)\ , r_2(\bm q_{{{\scriptscriptstyle t}}}^2 ) \ , , \\
4\phi \rangle_{q_{{\scriptscriptstyle t } } } & = & \frac{1}{2}\ , \sigma_0(\bm q_{{{\scriptscriptstyle t}}}^2)\ , r_4(\bm q_{{{\scriptscriptstyle t}}}^2 ) \,.\end{aligned}\ ] ] in the next section we provide numerical estimates for these observables in the specific configuration in which the higgs boson and the jet have the same rapidities ( @xmath122 ) . in this particular kinematic domain , the expressions in eqs .
( [ eq : r0])-([eq : r4 ] ) reduce to @xmath123 } { { \cal c}[f_1^g \ , f_1^g ] } \ , , \label{eq : r0 - 2}\\ r_2(\bm q_{{\scriptscriptstyle t}}^2 ) & = & \frac{\bm k_\perp^2 ( 2 \,\bm k_\perp^2 + m_h^2)}{9 \,\bm k^4_\perp + 8\ , \bm k_\perp^2\ , m_h^2 + m_h^4}\ , \frac { { \cal c}[w_2^{fh } \,f_1^{g}\ , h_1^{\perp\,g } ] + { \cal c}[w_2^{hf } \,h_1^{\perp\,g}\,f_1^g ] } { { \cal c}[f_1^g\,f_1^g]}\ , , \\ r_4(\bm q_{{\scriptscriptstyle
t}}^2 ) & = & \frac{1}{2}\,\frac{\bm k_\perp^4}{9 \,\bm k^4_\perp + 8\ , \bm k_\perp^2\ , m_h^2 + m_h^4 } \,\frac { { \cal c}[w_4^{hh } \,h_1^{\perp\,g}\ , h_1^{\perp\,g}]}{{\cal c}[f_1^g\,f_1^g]}\ , .
\label{eq : r4 - 2}\end{aligned}\ ] ]
in order to quantify the effects of gluon polarization on the observables defined in the previous section , we assume that the unpolarized gluon tmd is _ approximately _ a gaussian at small transverse momentum , but has the proper power law fall - off at large transverse momentum @xcite : @xmath124 where @xmath125 is the gluon distribution integrated over the transverse momentum squared @xmath126 and we choose @xmath127 gev@xmath128 . in order to show the maximal effects , for the distribution of linearly polarized gluons we take @xmath129 with @xmath130 given by eq .
( [ eq : f1gdb ] ) . in this case
the bound in eq .
( [ eq : bound ] ) is saturated for every value of @xmath1 and @xmath131 . in analogy to eq .
( [ eq : f1gdb ] ) , we also consider the following model from ref .
@xcite , @xmath132 with @xmath133 and @xmath134 , for which the bound is saturated only in the limit @xmath135 . defined in eq .
( [ eq : r0 ] ) for the subprocess @xmath6 and @xmath122 , see also eq .
( [ eq : r0 - 2 ] ) .
the unpolarized distribution @xmath130 is taken as in eq .
( [ eq : f1gdb ] ) , while @xmath136 either saturates its positivity bound in eq .
( [ eq : h1pg ] ) ( left panel ) , or it is given by eq .
( [ eq : h1pgdb ] ) ( right panel ) .
the blue band indicates the range for @xmath137 .
_ , title="fig : " ] defined in eq .
( [ eq : r0 ] ) for the subprocess @xmath6 and @xmath122 , see also eq .
( [ eq : r0 - 2 ] ) .
the unpolarized distribution @xmath130 is taken as in eq .
( [ eq : f1gdb ] ) , while @xmath136 either saturates its positivity bound in eq .
( [ eq : h1pg ] ) ( left panel ) , or it is given by eq .
( [ eq : h1pgdb ] ) ( right panel ) .
the blue band indicates the range for @xmath137 .
_ , title="fig : " ] it is convenient to consider the fourier transforms of the above functions in eqs .
( [ eq : f1gdb ] ) and ( [ eq : h1pgdb ] ) : @xmath138 in this way the relevant convolutions can be expressed as : @xmath139 = f_1^g(x_a ) f_1^g(x_b ) \int_0^\infty \frac{db}{2\pi}b j_0(b|\bm q_{{\scriptscriptstyle t}}| ) k_0(b / r)^2 , \\ & & \mathcal{c}\left[w_0^{hh}\,h_1^{\perp\,g}h_1^{\perp\,g } \right ] = \frac{c^2}{8 } f_1^g(x_a ) f_1^g(x_b ) \int_0^\infty \frac{db}{2\pi}b j_0(b|\bm q_{{\scriptscriptstyle t}}| ) \frac{b^2}{r_h^2 } k_1(b / r_h)^2 , \\ & & \mathcal{c}\left[w_2^{fh}\,f_1^{g}h_1^{\perp\,g } \right ] = - \frac{c}{2 } f_1^g(x_a ) f_1^g(x_b ) \int_0^\infty\frac{db}{2\pi}b j_2(b|\bm q_{{\scriptscriptstyle t}}| ) \frac{b}{r_h } k_0(b / r ) k_1(b
/ r_h ) , \\ & & \mathcal{c}\left[w_4^{hh}\,h_1^{\perp\,g}h_1^{\perp\,g } \right ] = \frac{c^2}{8 } f_1^g(x_a ) f_1^g(x_b ) \int_0^\infty \frac{db}{2\pi}b j_4(b|\bm q_{{\scriptscriptstyle t}}| ) \frac{b^2}{r_h^2 } k_1(b / r_h)^2 .\end{aligned}\ ] ] by substitution into eqs .
( [ eq : r0 - 2])-([eq : r4 - 2 ] ) , one finds @xmath140 results for the upper bound of @xmath141 in eq .
( [ eq : r0 - 2 ] ) are shown in fig .
[ fig : r0_db ] , where we use the unpolarized tmd distribution in eq .
( [ eq : f1gdb ] ) , while @xmath136 is given by eq .
( [ eq : h1pg ] ) in the left panel and by eq .
( [ eq : h1pgdb ] ) in the right panel .
the results are presented for two different choices of @xmath142 : @xmath143 and @xmath144 gev .
the corresponding results for the transverse momentum distribution defined in eq .
( [ eq : qt ] ) , with @xmath145 given in eq .
( [ eq : sigma0 ] ) and @xmath146 , are depicted in fig .
[ fig : qt_db ] .
the choice of @xmath147 is motivated by the requirement of tmd factorization that @xmath148 , where @xmath149 denotes the hard scale . in the present case
we have two hard scales : @xmath2 and @xmath150 .
the kinematics considered here is strictly speaking the back - to - back correlation region where @xmath151 . by integrating up to @xmath146 ,
one however also includes configurations @xmath152 in which @xmath153 and the jet are not approximately back to back in the lab frame .
this situation is not included in the calculation of @xmath154 scattering processes presented here .
however , for the model where the tmd has a power - law tail , the recoil against a third particle emitted into the final state in @xmath155 processes , is mimicked to some extent .
differently put , the tail of the tmd is sufficiently hard to produce large-@xmath156 pairs .
this is the reason why we extend the integration to @xmath146 . for the numerical results it does not make too much of a difference . in the gaussian model considered in the next subsection , the tail of the tmds is too suppressed to mimick the contribution from @xmath155 processes , hence , in that case we will restrict to @xmath157 to emphasize the proper region of validity . as a last comment on this point ,
sometimes the angular distribution of pair production processes are considered in the rest frame of the pair , for instance the collins - soper frame @xcite . in that case
the relative magnitude of @xmath158 w.r.t .
@xmath159 is not automatically apparent . in the case of higgs plus
jet , the center of mass energy of the pair is generally much larger than @xmath158 , while @xmath158 can be smaller or larger than @xmath159 .
if one restricts to @xmath154 scattering processes , one should realize that the region @xmath152 is not properly described , but at best mimicked by including the perturbative tails of the tmds .
, as defined in eq .
( [ eq : qtdist ] ) , for the subprocess @xmath6 , @xmath146 , and @xmath122 .
the tmds are the same as in fig .
[ fig : r0_db ] .
the solid line indicates the distribution in absence of linear polarization .
the blue band indicates the range for @xmath137.__,title="fig:",width=264 ] , as defined in eq .
( [ eq : qtdist ] ) , for the subprocess @xmath6 , @xmath146 , and @xmath122 .
the tmds are the same as in fig .
[ fig : r0_db ] .
the solid line indicates the distribution in absence of linear polarization .
the blue band indicates the range for @xmath137.__,title="fig:",width=264 ] our estimates for @xmath160 and @xmath161 are presented in figs .
[ fig : cos2phi_db ] and [ fig : cos4phi_db ] , respectively , with @xmath162 and @xmath144 gev . as before ,
for @xmath130 we have adopted the _
ansatz _ in eq .
( [ eq : f1gdb ] ) , while @xmath136 is given either by eq .
( [ eq : h1pg ] ) , in the left panels , or by eq .
( [ eq : h1pgdb ] ) , in the right panels .
moreover , we have chosen again @xmath146 .
asymmetries for the process @xmath163 , defined in eq .
( [ eq : cosnphiqt ] ) , as a function of the transverse momentum @xmath164 of the higgs plus jet pair , under the same conditions as in fig .
[ fig : qt_db ] .
the blue band indicates the range for @xmath165.__,title="fig:",width=264 ] asymmetries for the process @xmath163 , defined in eq .
( [ eq : cosnphiqt ] ) , as a function of the transverse momentum @xmath164 of the higgs plus jet pair , under the same conditions as in fig .
[ fig : qt_db ] .
the blue band indicates the range for @xmath165.__,title="fig:",width=264 ] asymmetries for the process @xmath166 , defined in eq .
( [ eq : cosnphiqt ] ) , as a function of the transverse momentum @xmath164 of the higgs plus jet pair , under the same conditions as in fig .
[ fig : qt_db ] .
the blue band indicates the range for @xmath165.__,title="fig:",width=264 ] asymmetries for the process @xmath166 , defined in eq .
( [ eq : cosnphiqt ] ) , as a function of the transverse momentum @xmath164 of the higgs plus jet pair , under the same conditions as in fig .
[ fig : qt_db ] .
the blue band indicates the range for @xmath165.__,title="fig:",width=264 ] although we have plotted its absolute value , we point out that @xmath160 is the only observable , among the ones discussed here , that is sensitive to the sign of the polarized gluon distribution , and it is expected to be negative if @xmath167 .
since the magnitudes of @xmath160 and @xmath161 turn out be very small , it will be easier to measure the integral of these observables over @xmath106 , up to @xmath168 , as defined in eq .
( [ eq : cosnphi ] ) . in both models for @xmath136
, we find that @xmath169 when @xmath170 gev , while its value is about 0.5% when @xmath162 gev .
we find that @xmath171 is about 0.2% at @xmath170 gev and completely negligible at @xmath150 = 10 gev .
these numbers are for @xmath146 in both numerator and denominator . for comparison , we now consider a gaussian model for the tmd distributions , which is widely adopted in many phenomenological studies at lower hard scales .
we assume that the unpolarized tmd gluon distribution has the following gaussian form @xcite , @xmath172\ , , \label{eq : f1g}\ ] ] with a rather large @xmath173 = 7 gev@xmath174 to effectively include the broadening effects due to multiple gluon emissions .
the polarized distribution is chosen to saturate the positivity bound , as in eq .
( [ eq : h1pg ] ) , in order to see the maximal effects allowed .
our results for the transverse momentum distribution of the higgs plus jet pair , the @xmath175 and @xmath176 asymmetries are shown in fig .
[ fig : gauss ] . for the gaussian model we fix @xmath157 as explained above .
we find that @xmath177 and @xmath178 at @xmath170 gev , very similar to the values obtained with the previous model , despite the considerable differences . at @xmath179 gev , @xmath180 and @xmath181 . as can be seen in the above numerical results , the gaussian model exhibits a double node in @xmath141 , whereas the gaussian+tail model does not . in ref .
@xcite it was noted that the tmd convolution involving the weight @xmath92 exhibits a double node independent of the form of the tmd @xmath182 .
more explicitly , the following integrals vanish : @xmath183=0 $ ] for @xmath184 and @xmath185 , which is important to mention is not due to the angular integration .
however , this does not imply that the actual distribution @xmath141 exhibits two nodes , because the expression in eq .
( [ eq : r0 ] ) in terms of tmd convolutions only holds for @xmath186 for some hard scale @xmath149 . in general , @xmath187 \neq 0 $ ] .
addition of order @xmath188 terms that are dominant at large @xmath189 and that can be cast into the same convolution form @xcite , allows to extend the integration region to all @xmath164 .
however , this need not lead to a vanishing @xmath190-weighted integral . in order for linearly polarized gluons to not affect the @xmath164-integrated cross section , there should always be one node at least , but it may well be outside the tmd region , i.e. in the region @xmath191 .
this is why models for @xmath182 can lead to an @xmath141 distribution exhibiting any number of nodes in the tmd region .
as the gaussian model has no significant contributions outside the tmd region , it does have to display two nodes in the tmd region , as we confirm it does .
for the measurement of the effects discussed here , the jet transverse momentum resolution is an important factor .
as the resolution scale is probably on the multiple gev level and the effects in the models are largest for @xmath164 values in the few gev region , it may be hard to experimentally probe the region of interest here .
we note however that no experimental knowledge is available on the shape of the @xmath192 distribution , hence , there are no constraints available to bound or indicate the typical width of the distribution . in the models we have made specific assumptions ( @xmath127 gev@xmath128 in the gaussian+tail model and @xmath173 = 7 gev@xmath174 in the gaussian model ) which need not correspond to the actual distribution .
the latter could be significantly broader than expected from intrinsic transverse momentum effects , due to multiple gluon emissions , just like nonperturbative effects can affect the @xmath193-boson production distribution at transverse momentum values well above a few gev .
we therefore caution the reader not to take the @xmath164 ranges in the figures too literally .
the models are intended to illustrate what kind of features can arise qualitative from linearly polarized gluons .
the magnitudes of the asymmetries do give an indication of the maximal effects one might expect .
next we comment on possible effects from the color flow in the process .
the analysis presented here ignored the effects of initial and final state interactions . in the process @xmath194 , the gauge link structure of @xmath195\,\mu\nu}(x,\bm p_{{\scriptscriptstyle t}})$ ] in eq .
( [ phig ] ) is given by two infinite staple - like gauge links that both run to minus lightcone infinity .
it is denoted as @xmath196}$ ] in ref .
given that we restrict to @xmath48-even distributions , this is equal to @xmath197}$ ] . in the process of higgs plus jet production there is a more complicated gauge link structure , which has not been considered yet for all subprocesses
. the gauge link structure of @xmath198 will be the same as that of @xmath199 given in ref .
@xcite , but @xmath200 has no analogue considered before . in any case , we expect that the @xmath201}(x , \bm{p}_{{\scriptscriptstyle t}}^2)$ ] distribution(s ) probed in @xmath202 are different from the one in @xmath203 .
it is not clear at present how large quantitatively the differences between these distributions are in practice , but it should be kept in mind that @xmath48-even distributions do not require initial or final state interactions to be nonzero , unlike for instance sivers functions .
the gauge link dependence need not be the dominant dependence therefore . as such , there is no reason to assume that they are very large or very small .
this simply remains to be seen .
given that there are sufficiently many processes sensitive to the linear gluon polarization , the size of the initial and/or final state interactions can , at least in principle , be extracted from experiment . as a final comment , we point out that there are at present no indications that factorization will be broken in the process @xmath204 due to color entanglement effects like in @xmath205 @xcite .
in this paper we have considered linear gluon polarization effects in higgs plus jet production .
compared to similar processes where a color singlet state plus a jet is produced , the case of higgs production is special .
its large mass allows for an unsuppressed angular independent contribution in the region @xmath206 , cf .
( [ eq : r0 - 2 ] ) .
this is not the case , for instance , in heavy quarkonium plus jet production , where in addition issues related to color octet contributions ( even if suppressed ) may lead to complications regarding the factorization @xcite .
other color singlet state plus jet production processes , such as on- or offshell ( drell - yan @xcite ) photon plus jet production , or @xmath207 or @xmath193 boson plus jet production , are not sensitive to @xmath208 , but only to @xmath209 .
this also applies to higgs production in association with a photon , @xmath207 or @xmath193 boson .
the @xmath210 and @xmath211 processes have been investigated in great detail already , but without inclusion of polarization effects ( of quarks in this case ) @xcite .
we point out that to probe @xmath209 the best processes probably are still the drell - yan process as suggested in ref . @xcite or photon plus jet production @xcite .
moreover , we note that the process higgs plus @xmath207 boson receives a contribution from @xmath209 only proportional to the very small higgs coupling of the light quarks inside the proton and can therefore rather be viewed as a cross - check process .
linear gluon polarization effects in other color singlet pair production processes have been studied , i.e. diphoton production in ref .
@xcite and @xmath99 plus photon production in ref .
@xcite , but as mentioned those studies were done in the collins - soper frame , where the restriction @xmath212 is not imposed or automatically respected .
the numerical results obtained in those studies do not show a clear advantage over higgs plus jet production , giving similar or smaller values for the angular asymmetries .
the lower energy and the better transverse momentum resolution could provide a big advantage though .
when it comes to probing the linear polarization of the gluons , higgs plus jet production has some additional features compared to inclusive higgs production .
one is the possibility of probing a continuous range of hard scales , allowing , at least in principle , for a study of tmd evolution .
this is not possible in higgs production , where the hard scale is fixed to be @xmath2 .
another is that , because the higgs - jet system can be in various angular momentum states , angular distributions can be probed that are not accessible in inclusive higgs production where the recoiling jet is not observed .
thanks to this , effects involving only one initial linearly polarized gluon can be probed through the @xmath0 distribution , as in @xcite .
this work was supported by the european community under the `` ideas '' program qwork ( contract 320389 ) .
we thank paolo gunnellini , jean - philippe lansberg , hannes jung and pierre van mechelen for useful comments and information .
s. catani and m. grazzini , nucl .
b * 845 * , 297 ( 2011 ) .
p. sun , b. -w .
xiao , and f. yuan , phys .
d * 84 * , 094005 ( 2011 ) .
d. boer , w. j. den dunnen , c. pisano , m. schlegel , and w. vogelsang , phys .
lett . * 108 * , 032002 ( 2012 ) .
d. boer , w. j. den dunnen , c. pisano , and m. schlegel , phys .
lett . * 111 * , 032002 ( 2013 ) .
j. mulders and j. rodrigues , phys .
d * 63 * , 094021 ( 2001 ) .
p. m. nadolsky , c. balazs , e. l. berger , and c. -p .
yuan , phys .
d * 76 * , 013008 ( 2007 ) .
a. metz and j. zhou , phys .
d * 84 * , 051503 ( 2011 ) .
f. dominguez , j. -w .
qiu , b. -w .
xiao , and f. yuan , phys . rev .
d * 85 * , 045003 ( 2012 ) .
a. schfer and j. zhou , phys .
d * 85 * , 114004 ( 2012 ) .
j. collins , _ foundations of perturbative qcd _ ( cambridge university press , 2011 ) .
m. g. echevarra , a. idilbi and i. scimemi , jhep * 1207 * , 002 ( 2012 ) .
d. boer and w. j. den dunnen , nucl .
b * 886 * , 421 ( 2014 ) .
m. g. echevarra , t. kasemets , p.
j. mulders , and c. pisano , nikhef-2014 - 36 , _ to be submitted_. u. baur and e. w. n. glover , nucl .
b * 339 * , 38 ( 1990 ) .
d. graudenz , m. spira , and p. m. zerwas , phys .
lett . * 70 * , 1372 ( 1993 ) .
s. abdullin , m. dubinin , v. ilyin , d. kovalenko , v. savrin , and n. stepanov , phys .
b * 431 * , 410 ( 1998 ) .
d. de florian , m. grazzini , and z. kunszt , phys .
lett . * 82 * , 5209 ( 1999 ) .
a. belyaev , r. guedes , s. moretti , and r. santos , pos dis * 2010 * , 207 ( 2010 ) .
t. t. jouttenus , i. w. stewart , f. j. tackmann , and w. j. waalewijn , phys .
d * 88 * , 054031 ( 2013 ) .
x. liu and f. petriello , phys .
d * 87 * , 094027 ( 2013 ) .
p. sun , c .- p . yuan , and f. yuan , arxiv:1409.4121 [ hep - ph ] .
s. dawson , i. m. lewis , and m. zeng , phys .
d * 90 * , 093007 ( 2014 ) .
v. khachatryan _ et al . _
[ cms collaboration ] , phys .
d * 90 * , 032005 ( 2014 ) .
[ cms collaboration ] , cms - pas - jme-08 - 002 .
w. j. den dunnen , j. p. lansberg , c. pisano , and m. schlegel , phys . rev .
lett . * 112 * , 212001 ( 2014 ) .
r. k. ellis , i. hinchliffe , m. soldate , and j. j. van der bij , nucl .
b * 297 * , 221 ( 1988 ) .
s. dawson , nucl .
b * 359 * , 283 ( 1991 ) .
r. p. kauffman , s. v. desai , and d. risal , phys .
d * 55 * , 4005 ( 1997 ) [ erratum - ibid .
d * 58 * , 119901 ( 1998 ) ] .
a. bacchetta , m. diehl , k. goeke , a. metz , p. j. mulders , and m. schlegel , jhep * 0702 * , 093 ( 2007 ) .
d. boer and p. j. mulders , phys .
d * 57 * , 5780 ( 1998 ) .
s. meissner , a. metz and k. goeke , phys .
d * 76 * , 034002 ( 2007 ) .
d. boer , p. j. mulders , and c. pisano , phys .
d * 80 * , 094017 ( 2009 ) .
m. g. a. buffing and p. j. mulders , jhep
* 1107 * , 065 ( 2011 ) . t.c .
rogers and p.j .
mulders , phys .
d * 81 * , 094006 ( 2010 ) .
d. boer and c. pisano , phys .
d * 86 * , 094007 ( 2012 ) .
d. boer , phys .
d * 60 * , 014012 ( 1999 ) .
d. boer , s. j. brodsky , and d. s. hwang , phys .
d * 67 * , 054003 ( 2003 ) .
d. boer , p.
j. mulders , and c. pisano , phys .
b * 660 * , 360 ( 2008 ) .
z. lu and i. schmidt , phys .
d * 78 * , 034041 ( 2008 ) .
j. w. qiu , m. schlegel and w. vogelsang , phys .
lett . * 107 * , 062001 ( 2011 ) .
d. boer , s. j. brodsky , p. j. mulders , and c. pisano , phys .
lett . * 106 * , 132001 ( 2011 ) . c. pisano , d. boer , s. j. brodsky , m. g. a. buffing , and p. j. mulders , jhep * 1310 * , 024 ( 2013 ) . c. j. bomhof and p. j. mulders , nucl .
b * 795 * , 409 ( 2008 ) .
j. p. ma , j. x. wang and s. zhao , phys .
b * 737 * , 103 ( 2014 ) .
s. dooling , f. hautmann and h. jung , phys .
b * 736 * , 293 ( 2014 )
. t. han and s. willenbrock , phys .
b * 273 * , 167 ( 1991 ) .
k. hagiwara , r. szalapski , and d. zeppenfeld , phys .
b * 318 * , 155 ( 1993 ) .
s. chatrchyan _ et al . _
[ cms collaboration ] , jhep * 1211 * , 088 ( 2012 ) .
j. ellis , v. sanz , and t. you , eur .
j. c * 73 * , 2507 ( 2013 ) .
r. m. godbole , d. j. miller , k. a. mohan , and c. d. white , arxiv:1409.5449 [ hep - ph ] .
r. v. harlander , a. kulesza , v. theeuwes , and t. zirke , jhep * 1411 * , 082 ( 2014 ) . | in this paper we consider higgs plus jet production as a process that is sensitive to the linear polarization of gluons inside the unpolarized protons of the lhc .
the leading order expressions for the transverse momentum distribution of the higgs plus jet pair are provided in terms of transverse momentum dependent quark and gluon distributions .
this includes both angular independent and azimuthal angular dependent contributions , presented directly in the laboratory frame . lacking experimental constraints on the linearly polarized gluon distribution , we study its effects on higgs plus jet production using two different models to illustrate the generic features and maximal effects .
it is found that the @xmath0 distribution may be the most promising observable , as it is driven by only one initial linearly polarized gluon .
the potential advantages of the higgs plus jet process compared to other processes sensitive to the linear polarization of gluons are discussed . |
the length scale of baryon acoustic oscillation ( bao ) provides a `` standard ruler '' in the universe , making it one of the most powerful probes of cosmic expansion history .
this characteristic scale is the distance traveled by sound waves in the photon - baryon plasma before recombination , and has been detected both in the cosmic microwave background ( cmb ) at @xmath3 @xcite and in galaxy surveys at @xmath4 @xcite .
recently , much effort has been expended to establish feasible schemes of measuring the bao features accurately at a variety of redshifts .
it is commonly believed that a combined analysis of bao measurements and other independent observables can substantially improve the current constraints on the equation of state of dark energy and its time dependence ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
since emission in the 21 cm hyperfine transition can be safely assumed to be proportional to the cosmic density of neutral hydrogen , measurement of intensity fluctuations in the cumulative 21 cm emission appears as a very promising technique to trace large scale structure ( lss ) over a wide range of redshifts . in this manner
, bao signatures may be detected with small errors and used to gain cosmological insight ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
because the structure of interest is certainly above the non - linear scale and the 21 cm radiation is sensitive to neutral hydrogen regardless of whether it is part of a resolved object , fine angular resolution is not necessary for such experiments .
thus the blind 21 cm surveys could be more efficient than traditional galaxy surveys at high redshifts where individual sources are hard to be resolved in sufficient numbers .
moreover , the 21 cm experiments have the advantage of simultaneously exploring a large redshift range , benefiting from the broad frequency bandwidth allowed by radio telescopes .
furthermore , the 21 cm emission line can provide a direct and precise measure of the redshift , which is unlikely to be true in galaxy surveys .
a considerable survey volume is necessary for 21 cm measurements to achieve the precision required to detect large - scale features .
cylindrical radio telescopes with a wide field of view have been proposed to perform such surveys in the future @xcite , and near term measurements using existing telescopes are discussed by @xcite .
however , all of the most important lss surveys have been taken with optical but not radio telescopes .
baos are weak features imprinted in the 21 cm cosmological signal which has a strength of @xmath5 @xmath6k near @xmath7 .
in addition to the extremely faint signals themselves , foreground contamination seems formidable in low - frequency experiments .
the galactic and extragalactic foreground sources exceed the cosmological signal by at least four orders of magnitude @xcite .
radio techniques have so far yielded direct detections of the 21 cm emission only to @xmath8 @xcite , indicating a significant uncertainty in future neutral hydrogen surveys .
the ability to constrain the nature of dark energy represents the ability to detect bao wiggles in the 21 cm power spectrum .
keys to these detections lie not only in experimental strategies but also in data processing techniques . in this paper , we discuss the detectability of baos in future hydrogen surveys around @xmath9 . simulations of the 21 cm interferometric measurements are described , and a probable method of foreground subtraction is also considered .
we present predictions for possible systematic and statistical errors in the bao scale estimates over an important parameter range , showing the range of behaviors we might expect in future observations .
we note that the foreground strategy in the post - reionization era is very different from that in the epoch of reionization ( eor ) . at high redshifts ( @xmath10 ) ,
most of the neutral hydrogen is distributed in intergalactic space , nearly separated from galaxies .
hence the extragalactic point sources can be treated as just foreground contaminants .
however , the neutral hydrogen is locked in the overdense regions such as gas - rich galaxies at medium redshifts , indicating that the redshifted 21 cm line will be emitted from the extragalactic point sources , like the continuum .
galaxies therefore play two completely different roles here : foreground sources as well as cosmological signal emitters .
when we consider the foreground subtraction technique at @xmath7 , it must be kept in mind that only the continuum of galaxies should be cleaned out .
efficient operations should not remove foregrounds at the sacrifice of desired signals .
we can neither simply discard the contaminated pixels in sky maps as suggested in cmb observations @xcite nor remove point sources with the prior sky models as discussed in reionization experiments @xcite .
recently , analyses of radio observations at higher frequencies ( e.g. , 610 mhz ) have provided some possible ways to separate the post - reionization 21 cm signal from astrophysical foregrounds ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . and our work can be used as a starting point for modeling bao detections in blind 21 cm surveys up to @xmath11 .
this paper is organized as follows . in section 2
, we introduce the proper simulations of the foreground model as well as the radio interferometric measurements .
our foreground removal technique is also described .
next , we apply the subtraction method to the simulated data cube in section 3 , and further estimate the sensitivities on measurements of the 21 cm power spectrum .
we analyze results from a series of simulations in order to understand how the measurement uncertainties depend on the data reduction and instrumental configurations . finally , we summarize our main conclusions in section 4 . throughout the paper
we adopt a concordance cosmology of @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 and @xmath17 , as revealed by the wmap three - year observations @xcite .
we employ monte - carlo simulations presented by @xcite and references therein to generate the brightness temperature sky maps at relevant frequencies , which incorporate contributions from three main components : ( 1)galactic synchrotron and free - free emission ; ( 2)galaxy clusters ; and ( 3)extragalactic discrete sources such as star - forming galaxies and agns . random variations of morphological and spectroscopic parameters as well as the evolution of radio halos in galaxy clusters are also considered .
the outline of these foreground simulations has been given in @xcite , and more details can be found in the original papers .
all the simulations are performed on @xmath18 grids over a field - of - view of @xmath19 .
our foreground box is kindly provided by @xcite .
since most bao experiments are designed to provide low angular resolutions , we further arrange the sky maps onto grids of @xmath20 pixels , corresponding to the angular resolution of 5 arcminutes . along the line of sight
, we use the frequency range extending from 500 mhz to 900 mhz with a resolution of 1 mhz . in figure 1 , the simulated foreground maps with high and low angular resolutions are shown separately . to calculate the sensitivity of a bao measurement
, we need to specify both the instrumental configurations and the observing strategy .
many factors contribute to the instrumental response of a radio interferometer . among the most important to consider are the field of view , angular resolution , collecting area , antenna distribution , and bandwidth @xcite . in this work ,
the smallest scale we consider is set to be the third bao peak which locates at @xmath21 degree angular scale around @xmath9 @xcite . and
thus a radio instrument of modest size , typically with diameter @xmath22 m , can be used to resolve the bao features .
moreover , bao surveys should be performed in various pointing directions in order to observe a large volume of the universe which is crucial to obtain small statistical errors on measurements of the lss power spectrum .
hence the characteristic observing time scales with the collecting area , the instantaneous field of view ( _ i.e. _ , fov ) and the number of pointings . obviously , a large fov @xmath23 is important to raise the survey efficiency . here
, we introduce an interferometer consisting of @xmath24 dishes with diameter @xmath25 m. all the dishes are nearly coplanar and organized in circles with different diameters of @xmath26 m. for @xmath27 m , the dishes are spaced roughly 1 m from each other , otherwise , the dishes are spaced roughly 3 m from each other . in the left panel of figure 2
, we show the array layout for the 487 dishes with a maximum baseline of 150 m. the corresponding _
uv _ sampling density at @xmath28 mhz is plotted in the right panel .
one may note that , when the dishes are limited in quantity , the dense - to - sparse arrangement could promise complete _ uv _ coverage in the central area , but also provide some long baselines for good angular resolutions .
lr @xmath29 & 0.0005 + array diameter , d ( m ) & 100 + number of dishes , @xmath30 & 330 + system temperature , @xmath31 ( k ) & 50 + bandwidth , b ( mhz ) & 100 + central frequency , @xmath32 ( mhz ) & 710 + spectral resolution , @xmath33 ( mhz ) & 1 + order of polynomial fit , n & 3 + sky coverage , @xmath34 & 0.5 + how shall we carry out the large - volume survey ? generally speaking , a larger sky coverage requires a longer observation time .
however , if we could observe several independent fields of view simultaneously , the total survey volume will be extended without prolonging the observing time .
here we consider a 21 cm survey with 10 separate interferometers , each situated at a different location and observing a different patch of the sky .
compared to a single interferometer , the cost of building the proposed telescope will no doubt increase .
and we need to find a balance between price and efficiency . for computational simplicity ,
our reference measurement is chosen to be a deep observation of a single target field centered on the north celestial pole ( ncp ) .
the effect of the earth rotation synthesis is not taken into account in the reference observation .
we assume that observations pointing in different directions can be performed in the same way as our reference measurement . for the fiducial model .
top panel : the dash - dotted line and dotted line correspond to the input foregrounds and detector noise , respectively .
the solid line represents the expected 21 cm signal , and the dashed line shows the post - subtraction noise in the cleaned sky map , including residual foregrounds and detector noise .
the vertical lines indicate locations of the first three bao peaks .
bottom panel : variance in the estimate of 21 cm power spectrum for an integration time of 1 year . ] in order to understand how the instrumental parameters impact the future bao measurements , our simulations are passed through the observational pipeline described in @xcite . in
what follows , we briefly introduce its scheme .
since the 21 cm signal is very faint compared to foregrounds and its low - frequency properties are highly uncertain , we do not include it in our sky model .
besides , the effect of foreground removal on the cosmological signal is outside the scope of this work and hence ignored . the original image cube is first established on the basis of the brightness temperature distribution of foreground sources .
also , the primary beam is simply assumed to be unity within the fov . at each frequency channel
, the sky image is related to visibilities via the two dimensional fourier transform .
subsequently , we estimate the noise visibilities with one year integration .
the rms noise per visibility per frequency channel can be written as @xcite @xmath35 in which @xmath36 and @xmath37 are the effective area and the beam solid angle of a single dish respectively , @xmath33 is the bandwidth of a single frequency channel , and @xmath38 is the total integration time for sampling a given @xmath39 location . for the reference observation , we assume the system temperature to be @xmath40 k according to current technology .
in addition , we approximate the integration time @xmath41 , where @xmath42 s is the accumulation duration for each visibility measurement , and @xmath43 is the number of independent samples in that pixel @xcite . because the thermal noise is random , we draw complex visibilities from gaussian distributions with zero mean and rms described above . finally , the real - world sampling is carefully performed in the _ uv _ plane .
the contribution of any single visibility measurement is applied to only one grid cell .
we further normalize the baseline distribution to ensure that each pixel in the sampled part of the _ uv _ plane has the same weight . with the uniform weighting , we artificially emphasize the information contained in long baselines and increase the effective resolution of the derived sky maps .
following all these simulations , we generate our image cube and visibility cube representing realistic observations .
figure 3 shows the dirty map and the corresponding _
uv _ map at @xmath28 mhz .
additionally , the frequency spectra in a random image and _ uv _ pixel are also plotted separately in the bottom row . as illustrated in these two panels , the frequency coherence of foregrounds
will be destroyed in real space owing to the `` mode - mixing '' effect @xcite .
it is generally agreed that the intergalactic medium ( igm ) contains the majority of baryons in the universe , and is completely ionized by uv radiation beyond @xmath44 @xcite .
nevertheless , the damped lyman series absorption seen in quasar optical spectra indicates that a significant fraction by mass of neutral hydrogen does reside in the interstellar medium ( ism ) of galaxies @xcite , making it possible to use the neutral hydrogen as a tracer of the underlying large - scale structure in the post - reionization epoch . as we mentioned above ,
most of the neutral hydrogen is outside and inside of galaxies before and after reionization respectively , indicating that new developments in methods of isolating the 21 cm signal from foregrounds will be required in post - reionization experiments . in such observations , emissions of galaxies
consist of two components : the radio continuum as well as the redshifted 21 cm line .
the former is relatively featureless , however , the latter is expected to fluctuate rapidly in the frequency and angular directions .
both these radiations are emitted from extragalactic point sources , but have fundamentally different qualities .
in addition , the instrumental response changes with observing frequency , and thereby creates different sidelobe patterns across the measured sky maps , inducing the `` mode - mixing '' effect which will destroy the foreground smoothness in frequency space .
the frequency decoherence has been shown clearly in the left hand column of figure 3 . over the last decade
, much effort has been made in exploring possible methodologies for foreground subtraction ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the most widely discussed technique concentrates on the spectral fitting along each line - of - sight ( los ) , dealing with the removal of diffuse emission and unresolved point sources under the assumption that bright point sources have been cleaned out with high accuracy and precision .
owing to the `` mode - mixing '' effect , excision of point sources is always first carried out in a cleaning process @xcite .
however , for the proposed neutral hydrogen surveys , the extragalactic point sources should be treated not only as foreground contaminants but also as signal emitters .
this means that the point sources should not be removed , except for their continuum .
@xcite introduced a foreground removal technique in fourier space , which handles the confusion - level contaminants and the bright point sources as equivalent .
they fit the visibility spectrum using a weighted low - order polynomial along the los . and
any special treatments to point sources have been sidestepped . in this method ,
the inverse - variance weighting scheme protects the frequency coherence of foregrounds , and most importantly , the polynomial function guarantees that only the spectrally smooth component of the point sources radiation will be subtracted .
one can therefore hope that most of the 21 cm signal emitted from the point sources will be preserved well in the polynomial - subtracted residuals .
furthermore , a key feature of this cleaning method is that the algorithm is blind , indicating that it is not necessary to resolve point sources in future observations .
one can thus get a good start because the low - frequency bao surveys will be performed with limited resolutions . here , this formalism is employed to remove the radio foreground contamination . refer to the original paper for more details of the cleaning method .
in this work , we aim to study the feasibility of detecting baos in the 21 cm intensity fluctuations after reionization , using a radio interferometric array .
the measurement uncertainties depend on many parameters , and we will mainly consider the impacts of telescope design , observation strategy and foreground subtraction techniques . to get a sense of the precision which we are able to achieve ,
we first describe a fiducial observation in section 3.1 , and further quantify its sensitivity in section 3.2 .
a detailed account of the extent to which the various parameters matter is represented in section 3.3 .
table 1 lists the parameter values adopted in our fiducial model .
most of these parameters reflect the current technology , and some specifications may be subject to change because of the developing experiments . under the fiducial model ,
we perform the simulations described in 2.2 , and then use the proposed foreground cleaning method on the visibility data cube .
figure 4 illustrates the ability to remove the foreground contamination with instrumental response .
the top row of the figure shows the simulated visibility cube .
the left panel is a map from a single frequency channel , and the right panel plots a @xmath45-_v _ plane slicing along the frequency direction .
moreover , the bottom row shows the same cuts , but in the residual data cube following the polynomial subtraction .
qualitatively , the residual visibilities have been suppressed to a level of order @xmath46 @xmath6k .
this represents the fiducial scenario that can be achieved by our foreground cleaning method to the simulated interferometric measurements . in principle
, the amplitude of the residuals could be made arbitrarily low by increasing the order of the polynomial used in the spectral fitting .
however , the process of foreground subtraction may accidently remove the spectrally smooth component of the cosmological signal , and hence lead to suppression in the final estimate of power spectrum primarily on large scales .
these artifacts have been discussed in previous work ( e.g. * ? ? ?
? * ; * ? ? ?
we should thereby perform the polynomial fit with order as low as possible .
after foreground subtraction , we are left with a residual data cube containing the fitting errors and the thermal noise . with this data set in hand
, we can compute the analytic approximations of measurement uncertainties in a simple way that is often accurate enough for studying the experiment design .
@xcite has shown that the noise in a pixelized map could be modeled as a random field with an angular power spectrum given by @xmath47 where we take the pixel solid angle to be @xmath48 , and @xmath49 is the standard deviation for a gaussian beam function in the experiment . the most straightforward way of estimating the rms error in the map is to specify @xmath50 . for the observation we simulate here
, @xmath50 is defined as the standard deviation of the temperature fluctuations in the residual sky map , which has contributions from the residual foregrounds and other systematic effects .
this approximation only strictly holds for a gaussian distribution , and the foreground residuals in the sky map may not be a gaussian field , but will be spatially correlated .
however , we ignore these and use this relationship here since it allows considerable simplification .
the estimate of the 21 cm power spectrum could therefore be described as a normal distribution with a variance @xcite @xmath51.\end{aligned}\ ] ] thus the simple estimate of the accuracy with which the expected power spectrum could be measured directly depends on the sampling variance , cosmological signal and pixel noise . in our simulations , we never create the 21 cm signal .
instead we base our estimate of the 21 cm power spectrum on an analytical approach @xmath52 in which @xmath53 and @xmath54 represent the matter power spectrum and bias factor respectively .
the mean brightness temperature due to the 21 cm line can be written as @xcite @xmath55^{-1/2 } \left[\frac{1+z}{2.5}\right]^{1/2}.\end{aligned}\ ] ] also , the angular power spectrum @xmath56 is derived using the standard limber s equation under the flat - sky approximation @xcite . in figure 5
, we show @xmath56 , @xmath57 and @xmath58 for the fiducial model .
the dashed line in the top panel represents the angular power spectrum of the intensity fluctuations in the foreground subtracted sky map , compared to the input foregrounds ( dash - dotted line ) , detector noise ( dotted line ) and the expected 21 cm signal ( solid line ) .
the vertical lines indicate the locations of the first three bao peaks .
one can see that the huge contaminants will be effectively removed with obviously smaller residuals which are less than one part in @xmath59 of original model .
therefore the first two peaks of bao wiggles could be measured with required precision . however , for @xmath60 the noise level increases rapidly . to a certain extent , this is induced by the fast growth of the thermal noise at small scales . in the bottom panel ,
the variance @xmath58 changes with @xmath61 : it decreases at the first and then increases .
its minimum value appears at @xmath62 . from equation 3
, it is evident that the sampling variance as well as the cosmological signal and pixel noise will together determine the quantity of @xmath58 .
their effects at different scales will be studied in the next section . above we investigated errors in the estimates of bao scales under our fiducial model , including the effects of foreground contamination and instrumental response .
now we use a set of simulations with various parameters to investigate how well we could expect to measure the acoustic scales in future 21 cm surveys . in each case , we examine only one of these parameters and keep all others fixed .
observations of quasar absorption spectra indicate that the intergalactic medium is highly ionized and a significant amount of neutral hydrogen could be found only in galaxies in the post - reionization universe . in this work , there is no need to describe the hi distribution , but rather we make a quantitative estimate of the total neutral hydrogen content and study its evolution at moderate redshifts .
our knowledge of the neutral hydrogen content is derived mainly from gas absorption in the hi hyperfine transition .
these observations find that the comoving hi density in galaxies is almost constant with a density parameter of @xmath63 beyond @xmath9 , only doubling over the redshift range @xmath64 @xcite .
moreover , the proposed hi survey traces galaxies but not the underlying matter distribution , which over - weights the overdense regions and thus introduces an inherent bias .
since this bias factor is as yet fairly unconstrained observationally , we ignore its time and scale dependences and assume a constant value for it ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. this would be accurate enough for our purposes in the present work .
based on these , we use the halo model to calculate the mass power spectrum , and adopt @xmath65 for the pessimistic , fiducial and optimistic predictions in the theoretical model respectively .
figure 6 investigates these three scenarios .
it appears to be true that the bao measurements would benefit from increasing the intensity of 21 cm emission which is proportional to the cosmic density of neutral hydrogen in the universe . and
in the bottom - right panel , the variances of power spectrum estimates increase with the amount of neutral hydrogen especially on large scales . in a radio interferometric experiment , different distributions of dishes give different instrumental responses .
we consider three cases here , in which dishes within diameters of 60 , 100 , 150 m are involved respectively and the total numbers of dishes are 226 , 330 , 487 correspondingly .
results are shown in figure 7 . in the top panels and the bottom - left panel ,
it is clear that the foreground subtraction can be done well using all these arrays , and there is no significant difference in performance at @xmath66 .
nevertheless , one may note that the noise in the residual map ( dashed lines ) grows rapidly at large @xmath61 due to the exponential factor in the noise term .
meanwhile , the thermal noise ( dotted lines ) will increase faster for lower angular resolutions . in the bottom - right panel , for small values of @xmath61 , @xmath58 is limited by the sample variance together with the cosmological signal , and thus decreasing the beam size @xmath49 decreases @xmath58 only slightly .
but for @xmath67 , the reduction in @xmath58 is dramatic , greatly expanding the available range of scales .
for this reason , it is suggested that a bao experiment should make measurements at resolution @xmath68 arcminutes in order to resolve the higher harmonics of bao features in the extracted power spectrum .
moreover , the smaller beam experiment will provide information over a wider range of modes .
now we focus on the effect of increasing the field of view .
the observed fraction of the sky is defined by @xmath34 . in figure 8
, we compare the possible errors in bao measurements with @xmath69 . comparing the top panels to the bottom - left one
, we do not find that the residual noise level ( dashed lines ) has reduced obviously .
the most significant effect appears in the bottom - right panel , where the variance @xmath58 seems to be proportional to the factor @xmath70 .
the errors expected from the survey of a larger covering area appear smaller at most scales .
it indicates that an experiment with poorer sky coverage will be less precise at every value of @xmath61 , since the increase in the number of available modes decrease the sample variance errors .
the figure also makes it clear that we are unable to extract meaningful information with our current methods until @xmath34 is large enough . for the bao detections
, it would be important to cover multiple sky areas in a fixed observing time , and hence reduce the large - scale errors which are dominated by sample variance . in this section ,
we examine three parameters relating to the foreground subtraction algorithm : frequency resolution , bandwidth of polynomial fit and order of the fitting polynomial .
the results are plotted in figure 9 , figure 10 and figure 11 respectively . from figure 9 ,
one can see that the lower frequency resolutions will make little difference in the measured power spectrum , indicating that the measurement uncertainties are not sensitive to the frequency resolution .
this coincides with the conclusion presented in @xcite which illustrated that the proposed foreground cleaning method would provide the residual data with similar quality even performed with various frequency resolutions . in figure 10
, we investigate the effect of changing the frequency range over which the foreground spectra are fitted with smooth functions . at first sight
, a smaller frequency bandwidth seems favorable for the bao detections .
after foreground cleaning , the overall level of noise ( dashed lines in the top panels and the bottom - left panel ) would increase with the bandwidth , and hence enlarge the variance @xmath58 especially at @xmath71 .
however , we caution that the fewer frequency channels will degrade the signal - to - noise ratio in each band , furthermore , the cosmological signal on smaller scales may be accidentally removed during foreground subtraction .
one should therefore approximate the frequency spectrum of foreground contaminants over a frequency range that is as broad as possible , with the constraint that the produced residuals lie below the expected level of cosmic signal . for real observations ,
a choice of the frequency range would be made in the context of a specific experiment .
compared to the two parameters we mentioned above , the order of the polynomial fit will influence the results to the maximum extent , which is clearly shown in figure 11 . as we see in the top - left panel , there are insufficient degrees of freedom to remove radio foregrounds efficiently , and thus the rms noise ( dashed line ) in the residual maps may exceed the expected cosmological signal ( solid line ) , inducing the destruction of measured bao signatures . and obviously , the residual noise ( dashed lines in the top - right and bottom - left panels ) could be reduced significantly by increasing the order of polynomial used in foreground fits .
similarly , the bottom - right panel represents dramatic reductions in the detection variance .
nevertheless , as emphasized by previous work , some of the cosmological signal may be mistaken if n is too high ( e.g. * ? ? ?
* ; * ? ? ?
in this work , we obtain believable measurements of the first two bao peaks using a third - order polynomial to fit the real and imaginary parts of the visibility spectrum separately at each point in the _ uv _ plane .
recently , it was realized that the redshifted 21 cm line emitted from neutral hydrogen can be used as a tracer of the large - scale structure even after reionization .
also , detection of baos on the inferred 21 cm power spectrum can be used to explore the geometry of the universe especially in the dark energy dominated regime . in this paper
, we quantified the likelihood of measuring bao features near @xmath72 with a set of ground - based radio interferometers .
simulations of the 21 cm interferometric measurements have been described , taking into account the detector noise , the foreground contamination and the frequency - dependent instrumental response .
different from galaxy redshift surveys in optical and radio wavebands , we developed a possible observing strategy without individual object detections .
the foreground cleaning technology has also been studied , which is very different from both the cmb and eor experiments . moreover , we investigated what the accuracy of bao measurements is likely to be available in future 21 cm surveys , and further discussed how the measurement uncertainties would change if different assumptions are made for key configurations of the radio telescope as well as some potential parameters used in the foreground subtraction method .
the blind surveys of the redshifted 21 cm emission line have proven useful in offering information for acoustic scales at moderate redshifts , providing an important complement to the existing galaxy surveys . in the fiducial scenario
, we considered a pixel - by - pixel method to subtract continuum foregrounds from simulated visibilities , and found that a three - order polynomial can be used to well approximate the smooth spectrum . and
in the cleaned sky maps , the detector noise together with the residual foregrounds will contribute to the pixel noise .
just as we saw in figure 5 , this noise term ( dashed line ) could be suppressed to a level of order @xmath46 @xmath6k at @xmath73 after a whole year integration .
the imprints of the first two bao peaks are therefore possible to be resolved on the 21 cm angular power spectrum .
although bao experiments may not require fine resolutions , we found that an angular resolution of @xmath46 arcminutes is necessary in order to identify the higher harmonics of the bao feature on the extracted power spectrum .
this is because the residual noise would increase exponentially with the angular resolution , preventing significant detections on large values of @xmath61 .
we note that , in the fiducial scenario , all the antennas point toward the ncp and continuously observe a fixed sky patch . therefore , the integration time can be approximated by the observing time
. however , the total observing time should be longer if we observe a different part of the sky , depending not only on the location of the interferometer but also on the sky coordinates of the patch being observed . making a simplifying assumption that a single field of view can be observed for 6 - 8 hours in each day @xcite , the required integration time ( 1 year )
would be possibly achieved over a period of 3 - 4 years .
in addition , the precision of bao measurements is proportional to the survey volume . fortunately , next generation radio telescopes will benefit from the rapid rate of increase of field - of - view and collecting area
. then the survey speed can be greatly accelerated , which is likely to be orders of magnitude higher than that achievable by optical spectrographs .
generally speaking , each radio interferometer should produce an instantaneous fov of @xmath1 @xmath2 , and possess a collection area on the order of @xmath74 @xmath75 .
moreover , several tens of separate pointings could be used to map out a significant fraction ( _ e.g. _ @xmath76 ) of the sky within a limited observation time .
we investigated here the role of blind 21 cm surveys in bao measurements .
our results suggest that radio interferometers have the potential to become very competitive facilities for performing cosmologically interesting tests .
while it is a challenging experiment to achieve constraints on the order of @xmath77 in the bao measurements , these calculations expand our understanding of the effects of instrument design , observing strategy and data analysis techniques , offering a new angle for the future generation of neutral hydrogen surveys .
ansari , r. , campagne , j. e. , colom , p. , et al .
2011 , arxiv:1108.1474 ansari , r. , le goff , j .-
m . , magneville , c. , et al .
2008 , arxiv:0807.3614 becker , r. h. , fan , x. , white , r. l. , et al .
2001 , aj , 122 , 2850 bernardi , g. , mitchell , d. a. , ord , s. m. , et al . 2011 , , 413 , 411 bharadwaj , s. , & ali , s. s. 2005 , , 356 , 1519 bowman , j .
d. , morales , m. f. , & hewitt , j. n. 2006 , , 638 , 20 bowman , j .
d. , morales , m. f. , & hewitt , j. n. 2009 , , 695 , 183 chang , t .- c . , pen , u .- l . , peterson , j. b. , & mcdonald , p. 2008 , phys .
lett , 100 , 091303 chang , t .- c . , pen , u .- l . ,
bandura , k. , & peterson , j. b. 2010 , nature , 466 , 463 chapman , e. , abdalla , f. b. , harker , g. , et al 2012 , arxiv:1201.2190 datta , a. , bhatnagar , s. , & carilli , c. l. 2009 , , 703 , 1851 datta , a. , bowman , j. d. , & carilli , c. l. 2010 , , 724 , 526 eisenstein , d. j. , zehavi , i. , hogg , d. w. , et al . 2005 , , 633 , 560 fan , x. , strauss , m. a. , becker , r. h. , et al . 2006 ,
aj , 132 , 117 furlanetto , s. r. , & briggs , f. h. 2004 , new astron .
, 48 , 1039 furlanetto , s. r. , oh , s. p. , & briggs , f. h. 2006 , phys .
, 433 , 181 gleser , l. , nusser , a. , & benson , a. j. 2008 , , 391 , 383 ghosh , a. , bharadwaj , s. , ali , s. s. , & chengalur , j. n. 2011 , , 411 , 2426 ghosh , a. , bharadwaj , s. , ali , s. s. , & chengalur , j. n. 2011 , , 418 , 2584 harker , g. , zaroubi , s. , bernardi , g. , et al .
2009 , , 397 , 1138 harker , g. , zaroubi , s. , bernardi , g. , et al .
2010 , , 405 , 2492 hinshaw , g. , weiland , j. l. , hill , r. s. , et al .
2009 , , 180 , 225 hu , w. , & haiman , z. 2003 , phys .
d , 68 , 063004 kaiser , n. 1992 , , 388 , 272 knox , l. 1995 , phys .
d , 52 , 4307 lah , p. , chengalur , j. n. , briggs , f. h. , et al .
2007 , , 376 , 1357 liu , a. , tegmark , m. , & zaldarriaga , m. 2009 , , 394 , 1575 liu , a. , tegmark , m. , bowman , j. , hewitt , j. , & zaldarriaga , m. 2009 , , 398 , 401 liu , a. , tegmark , m. 2011 , phys .
d , 83 , 103006 mao , x .- c . , &
2008 , , 673 , l107 mao , x .- c .
2012 , , 744 , 29 masui , k. w. , mcdonald , p. , & pen , u .- l .
2010 , phys .
d , 81 , 103527 matteo , t. di , perna , r. , abel , t. , & rees , m. j. 2002 , , 564 , 576 mcquinn , m. , zahn , o. , zaldarriaga , m. , et al .
2006 , , 653 , 815 morales , m. f. , bowman , j .
d. & hewitt , j. 2006 , , 648 , 767 oh , s. p. , & mack , k. j. 2003 , , 346,871 petrovic , n. , & oh , s. p. 2011 , mnras , 413 , 2103 proux , c. , irwin , m. , mcmahon , r. g. , & storrie - lombardi , l. j. 2001 , apsss , 277 , 551p proux , c. , dessauges - zavadsky , m. , dodorico , s. , sun kim , t. , & mcmahon , r. g. 2005 , , 363 , 479 peterson , j. b. , bandura , k. , & pen , u .- l . 2006 , arxiv:0606104 pindor , b. , wyithe , s. b. , mitchell , d. a. , et al .
2011 , pasa , 28 , 46 prochaska , j. x. , wolfe , a. m. 2009 , , 696 , 1543 rhook , k. j. , geil , p. m. , & wyithe , j. s. b. 2009 , , 392 , 1388 rohlfs , k. , & wilson , t. l. 2004 , tools of radio astronomy ( 4th ed . ; berlin : springer ) samushia , l. , & ratra , b. 2009 , , 703 , 1904 santos , m. g. , cooray , a. , & knox , l. 2005 , , 625 , 575 sarkar , t. g. , & bharadwaj , s. 2011 , arxiv:1112.0745 seo , h .- j . , & eisenstein , d. j. 2003 , , 598 , 720 seo , h .- j . ,
dodelson , s. , marriner , j. , et al .
2010 , , 721 , 164 spergel , d. n. , bean , r. , dor , o. , et al .
2007 , , 170 , 377 storrie - lombardi , l. j. , & wolfe , a. 2000 , , 543 , 552s tegmark , m. , & oliveira - costa , a. d. 1998 , , 500 , l83 wang , x. , tegmark , m. , santos , m. , & knox , l. 2006 , , 650 , 529 wang , j. , xu , h. , gu , j. , et al .
2010 , , 723 , 620 wang , y. , xu , l. , & gui , y. 2011 , phys .
d , 84 , 063513 wolfe , a. m. , gawiser , e. , prochaska , j. x. 2005 , ara&a , 43 , 861 wyithe , j. s. b. , loeb , a. , & geil , p. m. 2008 , , 383 , 1195 wyithe , j. s. b. , & loeb , a. 2009 , , 397 , 1926 zaldarriaga , m. , furlanetto , s. r. , & hernquist , l. 2004 , , 608 , 622 zhan , h. , wang , l. , pinto , p. , & tyson , j. a. 2008 , , 675 , l1 zhang , t .- j . , yuan , q. , & lan , t. 2009 , new astron . | after reionization , emission in the 21 cm hyperfine transition provides a direct probe of neutral hydrogen distributed in galaxies .
different from galaxy redshift surveys , observation of baryon acoustic oscillations in the cumulative 21 cm emission may offer an attractive method for constraining dark energy properties at moderate redshifts .
keys to this program are techniques to extract the faint cosmological signal from various contaminants , such as detector noise and continuum foregrounds . in this paper , we investigate the possible systematic and statistical errors in the acoustic scale estimates using ground - based radio interferometers . based on the simulated 21 cm interferometric measurements , we analyze the performance of a fourier - space , light - of - sight algorithm in subtracting foregrounds , and further study the observing strategy as a function of instrumental configurations .
measurement uncertainties are presented from a suite of simulations with a variety of parameters , in order to have an estimate of what behaviors will be accessible in the future generation of hydrogen surveys .
we find that 10 separate interferometers , each of which contains @xmath0 dishes , observes an independent patch of the sky and produces an instantaneous field - of - view of @xmath1 @xmath2 , can be used to make a significant detection of acoustic features over a period of a few years .
compared to optical surveys , the broad bandwidth , wide field - of - view and multi - beam observation are all unprecedented capabilities of low - frequency radio experiments . |
most of the analyses concerning neutrino oscillations have so far been done within the framework of the theory of two flavor neutrino oscillation . in these scenarios , the observed deficit of solar electron neutrinos
is usually explained by means of @xmath9 oscillations , whereas the lack of atmospheric muon neutrinos is interpreted as a consequence of @xmath10 oscillations .
the main advantage of the two flavor oscillation scenario is that the oscillation probability depends on only two parameters , the mass squared difference @xmath11 and the mixing angle @xmath12 .
recently , it was pointed out by several authors , that in some cases the interpretation of the experimental data in terms of two flavor oscillations might yield misleading and sometimes even wrong results ( see _ e.g. _ ref.@xcite ) and that a three flavor oscillation description is to be favored .
unfortunately , the oscillation probabilities depend in this case on five parameters , two mass squared differences and three mixing angles , which makes it quite hard to deal with the dependence of the probability functions on these parameters .
the most popular three flavor neutrino oscillation scenario is the so called bimaximal mixing scenario @xcite , where two of the mixing angles are maximal , _
i.e. _ , they have values of about @xmath13 , whereas the remaining one is restricted by the chooz experiment to be rather small at least for certain parameter ranges @xcite .
the three flavor oscillation scenario can then be shown to decouple into two distinct scenarios involving two flavors , in which case it is of course much easier to perform a fit to the experimental data .
many authors @xcite use the chooz result to simplify the oscillation probability formulas for three flavors and then perform a fit of the two decoupled two flavor neutrino oscillation scenarios to the experimental data . in this paper , we will present a numerical fit within a three flavor neutrino mixing scenario to the zenith angle distribution of 850 day atmospheric neutrino super - kamiokande data and 708 day solar neutrino super - kamiokande data . however , we will not make any assumptions about the parameters on which the oscillation probability formulas depend , _
i.e. _ , we fit the scenario in the most general case and for a large parameter range to the experimental data .
the paper is organized as follows . in sec .
[ kamiotheory ] , we summarize the basic features of three flavor neutrino oscillation theory . in sec .
[ kamioexp ] , we present the choice of experimental data and in sec .
[ kamiomini ] , we discuss the minimization procedure . in sec . [ kamioresults ] , we give the obtained solutions of the minimization problem , which are interpreted in sec .
[ kamiointerpretation ] .
the fits corresponding to the solutions are discussed and compared in sec .
[ kamiofits ] and then these solutions are tested for stability in sec .
[ kamiostability ] . finally , sec .
[ kamioconclusions ] contains a summary and our conclusions .
in three flavor neutrino oscillation theory one assumes the neutrino states with definite flavor @xmath14 , @xmath15 , to be linear superpositions of states with definite mass @xmath16 , @xmath17 , @xmath18 the unitary mixing matrix @xmath19 is called cabibbo kobayashi maskawa ( ckm ) matrix and can generally be parameterized by three mixing angles @xmath20 , where @xmath17 , and three cp - violating phases .
one of the latter were recently claimed to be observable by next generation neutrino detectors @xcite and could lead to physically very interesting consequences .
the other two can be shown to cause no physical effects . however , for this analysis these phases are too small to be important and we will neglect them in what follows .
this implies @xmath21 and therefore we can write the ckm matrix in its ( real ) standard parameterization @xcite as @xmath22 where @xmath23 and @xmath24 . the probability for detection of a neutrino of flavor @xmath25 in a beam of neutrinos consisting exclusively of flavor @xmath26 at the source of the beam is for three flavors given by @xmath27 where @xmath28 is kronecker s delta , @xmath29 is the distance from the source to the detector , @xmath30 is the neutrino energy , and @xmath31 is the difference between the squares of the masses corresponding to the mass eigenstates @xmath16 and @xmath32 . since @xmath33 the mass squared differences are not linearly independent , and it is convenient to choose @xmath34 from which follows that @xmath35 so far we have considered the neutrino states to be plane waves , _
i.e. _ , to have a definite momentum .
since the neutrino state is neither produced nor detected with a definite momentum or propagation length , one has to average over @xmath36 as well as other uncertainties in production and detection .
we will here follow closely ref .
@xcite and assume that these uncertainties to be described by a gaussian average @xmath37 where @xmath38 and @xmath39 . inserting formula ( 3 ) for @xmath40 yields @xmath41 , \label{probability2}\ ] ] where @xmath42 is given by @xmath43 and therefore related to the sensitivity of the experiment .
@xmath44 should here be inserted in meters , whereas @xmath45 should be inserted in mev .
the parameter @xmath46 is the so called damping factor and in accordance with ref . @xcite we will choose @xmath47 where @xmath48 and @xmath49 are the uncertainties in neutrino energy and propagation length , respectively . for large values of @xmath46 @xmath50 and the oscillation term vanishes .
the transition probability then becomes a constant dependent on the values of the mixing angles ; the oscillation is said to become `` washed out '' . to proceed , one now has to determine the values of the parameters @xmath51 and @xmath52 for the experiments under consideration .
we are going to consider three types of super - kamiokande data , multi - gev @xmath53-like atmospheric neutrino data , multi - gev @xmath54-like atmospheric neutrino data , and solar neutrino data .
the probabilities were taken from ref .
@xcite for 14 data points in an energy range from 6 mev to 12 mev .
the probability is almost constant for all 14 data points .
the path length for solar neutrinos is given by the distance from the sun to earth ( @xmath55 m ) and the uncertainty of the path length is here assumed to be negligible compared to @xmath29 , _
i.e. _ , @xmath56 the damping factor then becomes @xmath57 where @xmath58 is the energy resolution of the experiment .
the atmospheric neutrino data are divided into atmospheric multi - gev @xmath53-like events and atmospheric multi - gev @xmath54-like events . in both cases ,
the flux measurements are performed for five bins , _
i.e. _ , for five different values @xmath59 of the zenith angle @xmath12 , which is defined as the angle between the direction of the measurement and the axis through experiment and the middle of the earth
. therefore , @xmath60 corresponds to straight upward going neutrinos , whereas for @xmath61 the neutrinos are produced in the atmosphere right above the detector .
the experimental data used are 850 day super - kamiokande data and monte carlo no - oscillation predictions for 40 years generated lifetime for the five bins .
the corresponding errors were taken from ref .
@xcite .
the results of the super - kamiokande experiment are usually presented by means of the `` ratio of the ratios '' @xmath62 where @xmath63 is the flux of neutrino flavor @xmath26 . in summer 1998 , the super - kamiokande collaboration published experimental values for this `` ratio of the ratios '' of @xmath64 for sub - gev events and @xmath65 for multi - gev events @xcite , which was the first direct observation of neutrino oscillations .
however , to be able to fit both , @xmath53-like and @xmath54-like data to the measurements , we will use here the directly measured disappearance rates @xmath66 for @xmath53-like events and @xmath67 for @xmath54-like events .
the main disadvantage of this is the uncertainty in the expected neutrino flux , which was claimed to be up to 20@xmath68 @xcite , due to the unknown flux of cosmic particles hitting the atmosphere . to handle
this one can introduce an overall flux uncertainty factor @xmath69 with values between 0.8 and 1.2 . the actual neutrino flux for the neutrino flavor @xmath26 is then given by @xmath70 .
thus , in absence of neutrino oscillations , the ratio between the theoretical and the measured fluxes would just be @xmath71 , whereas otherwise one obtains @xmath72 where @xmath73 is the transition probability from @xmath26 to @xmath25 , see sec .
[ kamiotheory ] . in
what follows , we are going to neglect the @xmath74 contribution due to too small production cross sections .
the uncertainty in the flux of cosmic rays is then an overall factor common to the ratios @xmath75 , where @xmath76 , of the both remaining neutrino flavors . denoting @xmath77 , one obtains @xmath78 for the electron neutrino ratio and @xmath79 for the muon neutrino ratio .
the value of @xmath80 is theoretically well determined to be about @xmath81 at @xmath82 gev @xcite .
the energy of the measured multi - gev neutrinos is assumed to be @xmath83 gev with negligible uncertainty compared to @xmath30 and we set @xmath84 . the path length for a neutrino that was detected in a bin corresponding to the zenith angle @xmath12 is given by @xmath85 as is easily obtained from geometrical considerations .
here @xmath86 is the radius of the earth and @xmath87 is the typical altitude of the neutrino production point in the atmosphere , which we assumed to be @xmath88 km .
the uncertainty in path length is mainly determined by @xmath89 and therefore @xmath90 the damping factor @xmath46 is thus a function of @xmath12 @xmath91
according to the last section , we have 24 data points , which we can use for the fit to the parameters of the theory : the five bins for @xmath53-like events , the five bins for @xmath54-like events , and the 14 data points for the solar neutrinos .
the function used to obtain the parameters is given by @xmath92 ^ 2 + \sum_{i=1}^5 { \rm w}_{\mu , i}\frac{1}{r^2_{\mu , i}(exp ) } \left[r_{\mu , i}-r_{\mu , i}(exp)\right]^2\nonumber\\ & + & \sum_{i=1}^{14}{\rm w}_{sun , i}\frac{1}{p^2_{sun , i}(exp ) } \left[p_{sun , i}-p_{sun , i}(exp)\right]^2\bigg\ } , \label{minfunction}\end{aligned}\ ] ] where the first two sums run over the five bins and the third one over the 14 data points for the solar neutrinos .
the weights have been taken as @xmath93 the mixing angles were constrained to the interval @xmath94 $ ] , whereas for the mass squared differences we assumed one of them to be in the range @xmath95 and the other one in @xmath96 . since these large parameter ranges yield no numerically stable solution , we minimized function ( [ minfunction ] ) for 45 parameter ranges corresponding to all combinations of values for the mass squared differences between @xmath97 for the small mass squared difference and @xmath98 for the larger one . following ref .
@xcite , we chose the following strategy to deal with the flux uncertainty factor @xmath99 : during the search for minima of function ( [ minfunction ] ) , we set it equal to unity , _ i.e. _ , we neglected the flux uncertainty .
the solutions obtained in this way were then tested for their stability with respect to variations of @xmath99 , see sec .
[ kamiostability ] . to minimize the function @xmath100 in eq .
( [ minfunction ] ) we generated @xmath101 random values for each of the five parameters in the corresponding range such that we obtained @xmath102 values of the minimization function .
after that we picked the parameter set that yielded the smallest value for the function and repeated this procedure @xmath103 times .
each of the obtained @xmath104 parameter sets then served as starting values for a deterministic minimization procedure , using a sequential quadratic programming method such that one obtains again @xmath104 parameter sets as an end result .
the solutions to the minimization problem can best be obtained by considering tables like table [ masstable ] , where the best point values for the two mass squared differences for each of the investigated mass regions are shown .
one can see , that for most of the mass regions at least one of the two mass squared differences was obtained on one of the region boundaries . in this case
, we assumed that the value of the corresponding mass squared difference converges towards a better minimum in one of the neighbor regions .
thus , a given parameter region was only considered to contain a solution to the minimization problem , if both values of the mass squared differences were obtained within the boundaries of the region .
note , that the values obtained for @xmath105 are almost randomly distributed in the last column of table [ masstable ] , _ i.e. _ , in the range @xmath106 .
this is due to a too large value of the damping factor @xmath46 in eq .
( [ probability2 ] ) such that the minimization function becomes independent of this parameter . only @xmath107 and the mixing angles are in this case fitted to the experimental data . to keep overview over the various considered parameter regions one can consider table [ masstable2 ] . here
a region , where a value for @xmath108 was obtained within the boundaries , is marked with an `` @xmath109 '' , whereas a region with a corresponding value for @xmath107 is marked with an `` @xmath110 '' .
one obtains in this way eight regions , marked `` @xmath111 '' in table [ masstable2 ] , where the mass squared differences are within the region boundaries .
seven of these regions are contained in the first two rows of table [ masstable2 ] , whereas the eighth one was obtained within the range @xmath112 , @xmath113 .
the parameter values obtained from the minimization procedure with the mass squared differences constrained to these regions correspond to the possible solutions of the minimization problem .
one notices that the regions marked with `` @xmath111 '' split table [ masstable2 ] in three distinct areas , corresponding to three types of solutions , as we will see .
the first of these interesting areas is the one containing the parameter regions in the upper left corner of table [ masstable2 ] , _ i.e. _ , the nine regions in the range @xmath114 , @xmath115 . in this range , there are all together five possible solutions , _
i.e. _ , mass regions , where the mass squared differences were obtained within the region boundaries . to pick the best of the possible solutions for this type , consider table [ minimizationtable ] , where the best point values for the minimization function ( [ minfunction ] ) are shown .
the smallest value of the minimization function was obtained in the region corresponding to @xmath116 , @xmath117 . in all surrounding regions
, we obtained values for the minimization function , which are larger , such that there seems indeed to be a minimum in the region specified above .
thus , the corresponding parameter set is a solution of the minimization problem , and we will denote it by _ solution 1 _ in what follows . considering the rest of table [ minimizationtable ] , one can see that this solution corresponds to the smallest value of the minimization function in the whole investigated parameter range , which means that it provides the best fit to the experimental data , as will be seen in sec .
[ kamiofits ] .
note that the value of the minimization function in the parameter range containing solution 1 is very close to the one obtained in the region @xmath118 .
furthermore , the values for the mass squared differences obtained within these two regions are rather close to each other , and one can not really distinguish , if these two possible solutions correspond to the same minimum or to two distinct minima very close to each other .
the second interesting area of table [ masstable2 ] contains six parameter regions in the range @xmath119 , @xmath120 . here a second type of solution is obtained in the region @xmath121 , @xmath122 , _
i.e. _ , in the last column , eighth row of table [ masstable2 ] .
we will denote this solution by _ solution 2 _ in what follows .
considering now table [ minimizationtable ] , one can see that in one of the neighbor regions we obtained a smaller value for the minimization function .
one has therefore only a `` local '' minimum inside the region containing solution 2 , and the values of the minimization function in the surrounding regions do not converge that obviously towards this region , as they did in the case of solution 1 .
however , in all the surrounding regions at least one of the mass squared differences was obtained on one of the region boundaries , which is why we will consider the region @xmath123 , @xmath122 to contain the second solution of the minimization problem . comparing the corresponding value of the minimization function to the one of solution 1
, one sees that this second solution provides a less exact fit to the experimental data . finally , the third interesting area is situated in the upper right corner of table [ masstable2 ] , in the range @xmath124 , @xmath125 . here
two regions with both mass squared differences within the region boundaries are obtained .
consideration of table [ minimizationtable ] tells us , that the smallest value for the minimization function was obtained in the parameter region @xmath126 , @xmath127 .
thus , we obtain a third type of solution of the minimization problem , denoted by _ solution 3 _ in what follows . as in the case of solution 2 , there are neighbor regions with smaller values for the minimization function , but they both have at least one of the mass squared differences on one of the region boundaries . note that the value of the minimization function is about five times larger for solution 3 than the one corresponding to solution 1 , and roughly three times larger than the one corresponding to solution 2 .
the parameter values for the three solutions obtained from the analysis in the last section are shown in table [ solutions ] . here the average value obtained from the @xmath103 minimizations are depicted , together with the corresponding standard deviations as well as the best point values .
solution 1 actually corresponds to the most common three flavor neutrino oscillation scenario , with one of the mass squared differences in the `` just - so '' domain ( @xmath128 ) and the other one in the range of possible solutions of the atmospheric neutrino puzzle ( @xmath129 ) . as for our solution , in this common scenario the values of the mixing angles are normally a set with a small value for @xmath130 , and values around @xmath13 for the other two mixing angles , @xmath131 and @xmath132 , _
i.e. _ , one has bimaximal mixing .
the three flavor oscillation scenario can in this case be shown to decouple into two independent oscillation scenarios involving two flavors @xcite .
the two flavor oscillation scenario with the small mass squared difference is then commonly used to describe the solar electron neutrino deficit in terms of @xmath133-@xmath134 oscillations , whereas the scenario with the larger mass squared difference is assumed to describe the atmospheric muon neutrino deficit in terms of @xmath134-@xmath135 oscillations .
however , this approximation is only exact in the case @xmath136 , and the nonzero value of this angle causes a mixing of these two decoupled two flavor scenarios .
this influences in turn the values obtained for the mass squared differences , which might explain the deviation of our result for the large mass parameter @xmath137 from the super - kamiokande result @xmath138 @xcite , which was obtained in the framework of a two flavor oscillation scenario .
note that we obtained this bimaximal solution without putting any restrictions on the parameters .
furthermore , the value we obtained for the second mixing angle @xmath5 implies that @xmath139 , which is in accordance with the upper bound obtained from the chooz experiment @xmath140 @xcite .
a disadvantage of solution 1 is the rather large standard deviation of the value obtained for @xmath107 , see table [ solutions ] .
solution 2 , on the other hand , is characterized by a value for the minimization function , which is twice as large as the one corresponding to solution 1 , but all parameters , except for the large mass squared difference , were obtained with very small standard deviations .
as discussed earlier , the minimization function is in this mass range independent of @xmath105 .
this means , that one obtains random numbers for this parameter , which explains the large standard deviation of @xmath105 . only the mixing angles and @xmath141 are in this case fitted to the experimental data . here
the small mass squared difference is lying in the range of possible solutions to the atmospheric neutrino puzzle , whereas the large one is compatible with the results of the liquid scintillator neutrino detector ( lsnd ) experiment @xcite , which yielded a value for the mass squared difference in the ev - range within the framework of a two flavor oscillation scenario .
however , one can see from table [ solutions ] , that the value obtained for the second mixing angle is maximal , _
i.e. _ , @xmath142 and , unlike for the first solution , it is therefore in this case not possible to reduce the three flavor oscillation scenario to two oscillation scenarios involving two flavors
. finally , solution 3 has the major disadvantage of a rather large value of the minimization function compared to the other two solutions , as discussed in the last section .
the remarkably large standard deviation of the large mass squared difference can be explained as in the case of the solution 2 , but solution 3 also shows a rather large standard deviation of the third mixing angle @xmath132 .
this solution is actually an `` intersection '' of the first two solutions , with one of the mass squared differences in the `` just - so '' range and the other one in the range obtained by lsnd .
all three mixing angles have values around @xmath143 , where @xmath17 , which means that it is , like in the case of solution 2 , not possible to decouple the three flavor oscillation scenario into a pair of two flavor oscillation scenarios .
figures [ diagramatmos_e ] - [ diagramsolar ] show the results of the fit .
the best point probabilities obtained from the three solutions are depicted as well as compared to the corresponding experimental data including error bars .
let us first consider the results of the fit to the atmospheric electron neutrino data , shown in fig .
[ diagramatmos_e ] . here
solution 1 and solution 3 provide the best fits to the experimental data .
solution 2 shows larger deviations , especially for the two bins with the smallest zenith angles , _
i.e. _ , the largest values of @xmath144 .
the neutrinos measured in these two bins come from right above the detector , which means that the background of cosmic particles is not shielded by the earth .
the experimental data corresponding to these zenith angles have accordingly the largest experimental deviations , and from eqs .
( [ eweight ] ) - ( [ sunweight ] ) one can see , that these bins therefore correspond to the smallest weights .
this explains why the fit is less exact for these zenith angles .
the fact that one obtains a zenith angle independent fit for solution 3 can be understood from consideration of the oscillation lengths @xmath145 where @xmath30 is again the neutrino energy and the @xmath146 are the mass squared differences .
this solution has one of the mass squared differences in the `` just - so '' range and the other one in the ev - range .
the oscillation length corresponding to the small mass squared difference is then far too long to make it possible for the experiment to see any variations in the oscillation probability , whereas in the case of the large mass squared difference the oscillations wash out before the neutrinos reach the detector , due to a too large value of the damping factor @xmath46 in eq .
( [ probability2 ] ) . considering next the fit to the atmospheric muon neutrino data , shown in fig .
[ diagramatmos_mu ] , one clearly sees that here solution 1 provides the best fit .
solution 2 has again larger deviations for the two bins with the smallest zenith angle , which as before can be explained by the fact that these bins have the largest experimental errors .
solution 3 provides again a constant fit to the experimental data , for the same reasons as the ones discussed above .
it appears that this solution is definitely not capable of explaining the zenith angle dependence of the atmospheric muon neutrino flux measured by the super - kamiokande detector , and therefore this solution can be ruled out .
finally , all three solutions provide rather good fits to the solar neutrino data , see fig .
[ diagramsolar ] .
solution 1 and solution 3 are seen to yield a constant oscillation probability up to an energy of about @xmath147 . above that energy
the oscillation length becomes of the same order of magnitude as the sun - earth distance such that the oscillation probability obtained from the two solutions starts to vary with energy .
solution 2 yields a constant fit to the solar neutrino data , which can be understood , having in mind that this solution is characterized by one mass parameter in the range @xmath121 and the other one in the ev - range . in both cases ,
the corresponding oscillation lengths are much shorter than the sun - earth distance , and all oscillatory effects wash out before the neutrinos reach the earth .
to summarize , only the solution 1 provides a good fit to all three types of experimental data , whereas solution 3 can be definitely ruled out by the zenith angle dependence of the atmospheric muon neutrino data measured by super - kamiokande .
it remains now to test the stability of the obtained solutions with respect to variation of the flux uncertainty factor introduced in sec .
[ kamiomini ] .
as mentioned before , we performed the minimization procedure setting the flux uncertainty factor @xmath99 introduced in eq .
( [ disappearencerate_e ] ) and eq .
( [ disappearencerate_mu ] ) equal to one .
but this factor can vary between @xmath148 and therefore one has to test the solutions obtained in the last section for stability with respect to variations of this flux uncertainty factor .
we saw in the last section , that the solution we denoted by solution 3 in table [ solutions ] can be ruled out by the zenith angle dependence of the atmospheric muon neutrino flux measured by super - kamiokande .
we will therefore only test the stability of solution 1 and solution 2 . to do this
, we applied basically the same minimization procedure as the one described in sec .
[ kamiomini ] , but this time for nine equidistant values of the flux uncertainty factor between @xmath149 and @xmath150 .
again , the angles were restricted to the intervals @xmath151 , where @xmath17 . in the case of solution 1 ,
the small mass squared difference was constrained to @xmath126 . to choose an interval for the large mass squared difference ,
one has to recall that in one of the neighboring regions of solution 1 , we obtained parameter values , which are very close to those corresponding to this solution , as we mentioned in sec .
[ kamioresults ] .
this could mean both that the two parameter sets correspond to two distinct minima or to the same one .
we will here follow the latter assumption , _ i.e. _ , we will assume that there is one minimum of the function ( [ minfunction ] ) somewhere between these two parameter regions . to avoid obtaining
the mass squared differences on the boundary between the two regions , we widened the interval for @xmath105 to @xmath152 . in the case of solution 2 , we restricted both mass parameters to the same intervals as in sec .
[ kamiomini ] , _ i.e. _ , to @xmath153 and @xmath154 .
figure [ stabdiagram1 ] shows the best point values for the five parameters , which were obtained for the different values of the flux uncertainty factor @xmath99 in the case of solution 1 .
all three mixing angles show the largest deviations from the value obtained for @xmath155 for the smallest values of the flux uncertainty factor , whereas for @xmath156 they become almost stable .
somewhat surprising is the high stability of the small mass squared difference , since it was obtained from the minimization procedure with large standard deviations .
the large mass squared difference , finally , grows almost linearly with the flux uncertainty factor , but remains within a region around @xmath157 .
the dependence of the parameters corresponding to solution 2 on the flux uncertainty factor is shown in fig .
[ stabdiagram2 ] .
the mixing angles are clearly seen to be more stable than in the case of solution 1 , they remain within ranges of @xmath158 or less around their values at @xmath159 .
the small mass squared difference , on the other hand , varies much more than the one corresponding to the solution 1 .
finally , the large mass squared difference shows a random distribution , as expected , since the minimization function is independent of this parameter in the mass region corresponding to solution 2 , as pointed out in sec .
[ kamioresults ] .
we have fitted the five parameters of a three flavor neutrino oscillation scenario to experimental values obtained by the super - kamiokande collaboration for atmospheric and solar neutrinos . to obtain numerically stable solutions , we divided a large region for the mass squared difference parameters into 45 smaller regions and performed the fit for each of these regions .
a mass region was considered to contain a solution of the minimization problem , if the best point values for mass squared differences were obtained from the fit within the region boundaries .
furthermore , the best point value of the minimization function was supposed to go through a minimum in such a region if compared to the values obtained in the surrounding regions . in this way ,
three types of solutions of the minimization problem were obtained , out of which the best one corresponds to the most common three flavor neutrino oscillation scenario , with one mass parameter in the `` just - so '' range and the other one in the interval @xmath160 . as in this common scenario , the first and the third mixing angle of this solution were obtained to be maximal , @xmath4 and @xmath6 , respectively , whereas the second one was obtained to be small , @xmath5 , which is below the chooz upper bound .
this solution corresponds to the global minimum of the minimization function in the considered parameter range , and it was the only one that provides a good fit to all three types of data considered . as a disadvantage
the value of the large mass squared difference was obtained with a rather large standard deviation .
the second solution is characterized by a value of the small mass squared difference in the range , which contains possible solutions of the atmospheric neutrino problem , _
i.e. _ , between @xmath161 and @xmath162 but by a large mass squared difference in the lsnd range , _
i.e. _ , between 1 @xmath163 and 10 @xmath163 . here
the mixing angles @xmath131 and @xmath130 were obtained to be maximal , whereas for the third one there was obtained a smaller value , @xmath164 .
this solution provides a comparably worse fit to the atmospheric electron neutrino data , which is the reason for the larger value of the minimization function corresponding to that solution .
the third solution corresponds to a small mass squared difference in the range @xmath153 and a larger one in the lsnd range . here
all three mixing angles were obtained quite close to each other , @xmath165 , the latter one with a rather large standard deviation .
this solution provides good fits to the atmospheric electron neutrino data and the solar neutrino data .
however , the oscillation probabilities for the atmospheric muon neutrino data corresponding to this solution show no zenith angle dependence such that this solution can be ruled out by the results obtained by super - kamiokande .
support for this work was provided by the engineer ernst johnson foundation ( t.o . ) .
we would like to thank hkan snellman and jonny lundell for useful discussions .
furthermore , we are grateful to kate scholberg from the super - kamiokande collaboration for providing us with the experimental data .
petcov , talk given at the _ xviith international workshop of weak interactions and neutrinos ( win 99 ) _ , cape town , south africa , 1999 ( to be published ) , hep - ph/9907216 ; w.m .
alberico and s.m .
bilenky , hep - ph/9905254 ; m. jeabek and y. sumino , phys .
b * 457 * , 139 ( 1999 ) , hep - ph/9904382 .
v. berezinsky , g. fiorentini , and m. lissia , astroparticle physics ( to be published ) , hep - ph/9904225 ; y. totsuka , talk given at the _ xixth texas symposium on relativistic astrophysics _ , paris , france , 1998 ; k. inoue , talk given at the _ viiith international workshop on neutrino telescopes _ , venice , italy , 1999 .
super - kamiokande and kamiokande collaborations , t. kajita , in _ proceedings of the xviiith international conference on neutrino physics and astrophysics ( neutrino 98 ) _ , takayama , japan , 1998 , nucl .
. suppl . * 77 * , 123 ( 1999 ) , hep - ex/9810001 . | we have analyzed the most recent available super - kamiokande data in a three flavor neutrino oscillation model .
we have here neglected possible matter effects and performed a fit to atmospheric and solar super - kamiokande data .
we have investigated a large parameter range , where the mixing angles were restricted to @xmath0 , @xmath1 , and the mass squared differences were taken to be in the intervals @xmath2 and @xmath3 .
this yielded a best solution characterized by the parameter values @xmath4 , @xmath5 , @xmath6 , @xmath7 , and @xmath8 , which shows that the analyzed experimental data speak in favor of a bimaximal mixing scenario with one of the mass squared differences in the `` just - so '' domain and the other one in the range capable of providing a solution to the atmospheric neutrino problem . |
one of the major efforts in current research is to determine a more accurate age for the milky way as an independent test of modern theories of cosmology .
metal - poor globular clusters ( gcs ) are important in this effort as they are among the oldest objects in the galaxy .
a comparison of gc color - magnitude diagrams ( cmds ) to theoretical isochrones can be used to determine gc distances and ages .
stellar evolution models were first developed for solar - metallicity stars but are now pushing towards calibration at lower metallicities .
the calibration of these models requires one to define the location of the main - sequence ( ms ) as a function of metallicity via ms - fitting ; however , to date there is only one star , hd25329 , with [ fe / h ] @xmath6 -1.4 dex and with an _ hipparcos _ parallax @xcite and chemical abundance determination from high resolution spectroscopy @xcite accurate enough for use in ms - fitting .
this lack of calibration stars introduces a great amount of uncertainty in the distance and age determinations for metal - poor gcs which we we aim to improve with our study .
one of the key features for an accurate ms - fitting calibration is accurate distances and abundances for very low metallicity ms stars .
accurate parallaxes for metal - poor stars are now available in the first _ gaia _
data release , which use the tycho - gaia astrometric solution ( tgas ) @xcite to obtain parallaxes with typical uncertainties of 0.25 mas ; however , even more accurate stellar parallaxes have been measured with the hubble space telescope s ( hst ) fine guidance sensor ( fgs ) for nine very metal - poor dwarf stars ( @xmath7 ) and are presented in @xcite .
our goal is to obtain accurate overall metallicities of the same very metal - poor ms stars with hst parallaxes to be used in ms - fitting .
we were motivated to use the robust technique of a line - by - line , differential abundance analysis , ultimately relative to the sun , as it should provide abundances that are independent of loggf values .
the iterative nature of this method was expected to provide excitation temperatures independent of line equivalent widths ( ews ) and ionization balance for ionized and neutral species of the same elements . however , in the process we uncovered issues ( e.g. non - ionization equilibrium and low spectroscopic gravities ) that a 1d lte analysis is incapable of resolving .
this paper is organized as follows . in @xmath82
we present our target selection criteria and spectroscopic observations , in @xmath83 we present the abundance analysis based on the ews of fe and @xmath2-element spectral lines and stellar model atmospheres .
the differential abundance results are provided in @xmath84 along with an error analysis and a comparison to the dartmouth stellar evolution program isochrones ( * ? ? ?
* hereafter dsep ) .
finally , the significance of our findings is discussed in @xmath85 .
the release of the _ hipparcos _ catalog @xcite led to a renewed interest in using ms - fitting to determine distances to gcs ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
these studies generally found that the gcs were somewhat more distant , and hence younger than found in previous work .
the location of the ms in a color - magnitude diagram is a sensitive function of the composition of the stars , and these main sequence fitting studies were hampered by the lack of very metal - poor stars ( @xmath9}\la-1.4$]dex ) which had accurate _
_ parallaxes .
the original _ hipparocs _ parallaxes were reanalyzed by @xcite leading to smaller uncertainties in the parallaxes . unfortunately , even with this improved data , only one very metal - poor star ( hd25329 ) was found to have a parallax measured to better than 10% accuracy , suitable for ms - fitting . having only one comparison star at low metallicity results in large uncertainties in distance determinations and therefore ages of metal - poor gcs .
to alleviate this situation , we were awarded time on hst to obtain fgs parallaxes of nine very metal - poor stars with the goal of extending the range of metallicities below at least [ fe / h ] @xmath10dex for stars with well - determined parallaxes . to select stars for the hst program , single , un - evolved ( i.e. main sequence ) stars that were very metal - poor
were selected from the lists complied by @xcite , @xcite and @xcite . based upon the increased number of stars with reliable parallaxes , and the fact that the hst parallaxes would have substantially smaller uncertainties than the _ hipparcos _ parallaxes
, we anticipated that the hst parallaxes would result in an improvement in the distance scale to metal - poor gcs by a factor of three compared to previous work .
long term radial velocity monitoring from @xcite was used to ensure the target stars are not members of multiple star systems . to ensure that the stars are on the main - sequence
, we required that the target stars be faint ( @xmath11 ) and/or have a large surface gravity ( @xmath12 dex cm s@xmath13 ( henceforth , dex ) ) and also be fairly red in color ( @xmath14 ) , all characteristics of ms stars .
these specific cut - offs in absolute magnitude , surface gravity and color were based upon an examination of the properties of metal - poor ms stars in the dsep theoretical isochrones . since the _ hipparcos _ parallaxes have relatively large uncertainties for these stars , surface gravity and color constraints helped to eliminate evolved stars ( sub - giants and giant branch stars ) from the sample .
high resolution spectroscopy of the nine target stars were obtained between 2008 and 2012 using the magellan inamori kyocera echelle ( mike ) double spectrograph on the 6.5 meter magellan ii clay telescope , at las campanas observatory , and the high resolution echelle spectrometer ( hires ) on the twin telescopes at the keck observatory .
our mike spectra roughly cover the range of 33755000 on the blue side and 48509160 on the red side ; however , small changes in the choice of grating tilt between runs can lead to the inclusion or exclusion of an order .
the mike blue - channel spectra actually extend beyond 5000 , but beyond 5000 the spectra are not useful for abundance analysis , due to the presence of spectral ghosts ; these ghosts result from internal reflections within the dichroic optical element .
no spectral ghosts are detectable in the red - channel spectra .
for all mike observations we employed a [email protected] arc second slit , with 2@xmath151 binning , resulting in a measured spectral resolving power of @xmath16=48,000 for the red side and @xmath16=55,000 for the blue side . the hires spectra consist of data from three ccds in the plane of this single spectrograph ; for our setup chips 1 through 3
cover the spectral range 41155600 , 56507175 , and 72358695 respectively .
our hires observations employed a slit width of 0.574 arc seconds and 2@xmath151 binning of the ccd pixels , which resulted in a measured average resolving power of r=70,500 ; however , significant degradation of the resolving power was seen from the center to the edges of the spectral orders .
we note that the resolving power we have measured , for both spectrographs , is higher than cited in the manuals ; this was likely a consequence of the good seeing at the time of observation .
we measured the per pixel s / n from the rms scatter of the blaze peak near 67306750 using the iraf _ splot _ routine ; typically , our s / n values greatly exceed 100:1 , permitting reliable measurement of ews down to the 5m level .
radial velocities ( rvs ) of the target stars were determined by cross - correlating our spectra with the kurucz solar spectrum , using the iraf _ fxcor _ routine ; the measurement uncertainties ranged from 0.6 to 0.9 km / s .
heliocentric corrections to the measured velocities were obtained using the iraf _ rvcorrect _ routine .
a log of the spectroscopic observations along with hst f606w magnitudes and parallaxes appears in table [ table : obs ] .
[ cols= " > , < , > , > , > , > , > " , ]
we performed model atmosphere abundance analysis of our program stars in an iterative procedure , where the stellar temperatures ( t@xmath17 ) , [ fe / h ] , [ @xmath2/fe ] and microturbulent velocities were determined from the ews of lines measured from the high s / n spectra .
we used these t@xmath17 , [ fe / h ] and [ @xmath2/fe ] values , together with published photometric data , to determine the logg indicated by the dsep theoretical isochrones .
a check on the adopted isochrone logg is obtained from the [ fe / h ] determined from fe ii lines , which are sensitive to the model atmosphere gravity . in order to reduce systematic errors on [ fe / h ] and [ @xmath2/fe ] due to inaccurate line oscillator strengths ( loggf values )
, we employed a line - by - line differential abundance analysis , relative to the same lines in the solar spectrum , similar to the method employed by @xcite .
this method may also provide some mitigation for a number of systematic errors , due to incomplete input physics in the analysis ( e.g. blends , non - lte effects , 3d hydrodynamical effects ) .
this differential analysis is facilitated by the fact that the gravities of the halo dwarfs are close to the solar value , while their temperatures are not too far from the solar t@xmath17 of 5777k ( e.g. * ? ? ?
* ) ; however , irradiance measurements of @xcite indicate 5772 k for the sun .
although we determined spectroscopic model atmosphere effective temperatures , we also checked for consistency with temperatures indicated by photometric colors and color - t@xmath17 relations .
the photometric values also provided initial t@xmath17 estimates for the iterative abundance analysis .
since unsaturated lines , on the linear portion of the curve of growth , have the greatest sensitivity to abundance , an ideal differential abundance analysis would employ only such weak , unsaturated , lines in both the sun and the metal - poor program stars .
typically , unsaturated lines in solar - temperature dwarf stars have an equivalent width below @xmath183050m .
however , lines that are 3050m in the sun are in the 14m range for our metal - poor dwarf stars . the need to measure the ew of such weak lines to high precision motivated the acquisition of very high s / n spectra for our metal - poor halo dwarf sample .
ultimately however , the differential abundance scatter obtained was higher than we desired and the number of 14m lines was insufficient to adequately constrain the atmosphere parameters and final [ fe / h ] for our program stars . to address this problem
, we chose to employ a standard star , with similar t@xmath17 and logg to the sun , but intermediate [ fe / h ] , in order to increase the number of unsaturated lines in the differential analysis . fortunately , we had a high s / n spectrum of the nearby metal - poor dwarf , hip66815 , suitable for this purpose .
we also elected to include some stronger lines , which required an estimate of the microturbulent velocity .
one should note that the method of spectroscopic determination of stellar atmosphere parameters is not universally accepted .
@xcite caution against use of excitation equilibrium for determining stellar surface temperatures due to nlte effects and instead opt to use infrared flux method temperatures .
specifically , they find that their differential approach , relative to the sun , is only capable of removing abundance trends as a function of excitation potential ( ep ) for stars with [ fe / h]@xmath19 dex and state that , for more metal - poor stars , the remaining trend is likely due to uncertainties associated with van der waals damping .
there are important differences between @xcite and this study that are able to address these issues .
@xcite choose an upper limit for their ew measurements of 180 m as an attempt to minimize the influence of damped lines on their results . as can be seen by the results of their differential analysis for metal - poor stars ,
this upper limit is still too high and their results are in fact influenced by the effects of damping . on the other hand , we use an upper limit of only 85 m in our standard star , hip66815 , and our metal - poor ms stars , which corresponds to an upper limit of 120 130 m in the sun .
the authors also only use lines with ep@xmath20 ev and see that when they include lines with lower ep they are able to remove the trend in abundance with ep in their differential method .
we use lines covering the full range of eps and , coupled with our more stringent ew upper limit , are able to resolve the issues presented by @xcite and remove abundance trends with excitation potential in our line - by - line differential abundance method .
we manually measured the ews of fe and @xmath2-element spectral lines for each of the nine target metal - poor stars using the iraf _ splot _ routine for gaussian fits to the line profiles ; typical ew uncertainties were found to be @xmath213% , based on rms scatter around the best - fit profiles . several of the lines appeared on adjacent spectral orders in which case the final value used in the analysis was the average of the two measurements .
these repeated lines also provided estimates of our ew measurement uncertainty .
weak lines were specifically chosen because they are sufficiently sensitive to elemental abundance while also being insensitive to line damping or broadening mechanisms .
for both fe lines and lines of @xmath2-elements , a lower limit of @xmath22 was established due to the uncertainty in the placement of the continuum level ; an upper limit of 85 m was enforced to ensure the line behaved well in the linear portion of the curve of growth and wing damping did not affect the measurements .
a total of 85 fe lines were taken from the studies of @xcite and @xcite supplemented with lines identified by @xcite with line parameters from the kurucz online database .
a total of 36 @xmath2-element lines taken from @xcite were found to be suitable in the range of ew values .
a final list of the measured ews for fe and @xmath2-elements are provided in tables [ table : felines ] and [ table : alphalines ] , respectively .
the line ews for our standard stars , the sun and hip66815 , are listed separately in table [ table : ew66815 ] .
l r r r r r r r r r r r ion & wavelength & ep & + & ( ) & ( ev ) & 46120 & 54639 & 80679 & 87062 & 87788 & 98492 & 103269 & 106924 & 108200 + fe i & 4447.72 & 2.22 & ... & ... & 55.7 & 53.2 & ... & 81.0 & 72.2 & 60.1 & ... + fe i & 4469.38 & 3.65 & ... & ... & 20.3 & 33.0 & ... & 56.6 & 42.0 & 27.7 & ... + fe i & 4489.74 & 0.12 & ... & ... & 41.7 & 27.7 & ... & 59.7 & 51.9 & 44.7 & ... + fe i & 4525.14 & 3.60 & ... & ... & 15.4 & 29.2 & ... & 55.2 & 37.3 & 23.5 & ... + fe i & 4736.77 & 3.21 & ... & ... & 30.2 & 40.4 & ... & 63.9 & 54.2 & 39.9 & ... + fe i & 4938.81 & 2.88 & ... & 30.4 & 28.6 & 36.0 & 12.7 & 62.7 & 49.9 & 39.0 & 61.9 + fe i & 4939.69 & 0.86 & ... & 41.0 & 34.7 & 25.7 & 30.1 & 56.0 & 46.7 & 40.3 & 54.3 + fe i & 5072.08 & 4.28 & ... & ... & ... & 10.6 & ... & 23.0 & 23.0 & ... & ... + fe i & 5079.74 & 0.99 & ... & 36.4 & 34.1 & 23.8 & 22.3 & 58.3 & 45.3 & 40.0 & 54.7 + fe i & 5083.34 & 0.96 & ... & 53.0 & 47.0 & 38.2 & 36.6 & 68.7 & 59.2 & 53.3 & 65.7 + this table is published in its entirety in the electronic edition of the _ astrophysical journal . _ a portion is shown here for guidance regarding its form and content .
l r r r r r r r r r r r ion & wavelength & ep & + & ( ) & ( ev ) & 46120 & 54639 & 80679 & 87062 & 87788 & 98492 & 103269 & 106924 & 108200 + cai & 4283.01 & 1.89 & ... & ... & 64.7 & 63.3 & 53.3 & 90.8 & ... & 74.1 & ... + cai & 4318.65 & 1.90 & ... & ... & 51.0 & 64.6 & 30.5 & 81.8 & 69.6 & 62.5 & ... +
cai & 4425.44 & 1.88 & ... & ... & 50.3 & 57.4 & 28.9 & 79.9 & 74.8 & 60.4 & ... + cai & 4435.68 & 1.89 & ... & ... & 39.2 & 51.1 & 19.7 & 74.5 & 61.6 & 52.3 & ... + cai & 4454.78 & 1.90 & ... & ... & ... & ... & 65.3 & 114.7 & ... & ... & ... + cai & 5265.56 & 2.52 & 33.0 & 31.7 & 24.8 & 40.3 & 14.6 & 63.9 & 46.5 & 38.6 & 56.7 + cai & 5349.47 & 2.71 & 12.9 & 13.2 & 12.7 & 23.7 & 7.3 & 43.4 & 27.4 & 19.4 & 28.9 + cai & 5512.98 & 2.93 & 8.7 & 5.4 & 5.6 & 13.2 & ... & 29.0 & 14.6 & 12.7 & 14.6 + cai & 5581.97 & 2.52 & 12.4 & 14.7 & 12.7 & 21.3 & ... & 36.5 & 26.5 & 18.3 & 32.0 + cai & 5588.75 & 2.53 & 51.2 & 58.4 & 52.1 & 52.5 & 28.7 & 78.6 & 72.0 & 63.6 & 85.6 + this table is published in its entirety in the electronic edition of the _ astrophysical journal . _
a portion is shown here for guidance regarding its form and content .
r r r r ion & wavelength & ep & + & ( ) & ( ev ) & hip66815 & sun + o i & 7771.944 & 9.146 & 59.3 & 72.6 + o i & 7774.166 & 9.146 & 43.4 & 63.0 + o i & 7775.388 & 9.146 & 37.2 & 49.0 + na i & 6154.226 & 2.102 & 9.6 & 42.0 + na i & 6160.747 & 2.104 & 16.8 & 55.4 + mg i & 4730.029 & 4.346 & 25.1 & 69.6 + mg i & 5711.088 & 4.346 & 67.4 & 112.5 + mg i & 6318.717 & 5.108 & 17.6 & 39.4 + mg i & 6319.237 & 5.108 & 8.9 & 25.5 + mg i & 6319.495 & 5.108 & 3.1 & 11.5 + mg i & 8213.013 & 5.753 & 68.0 & 125.2 + mg i & 8712.690 & 5.932 & 10.2 & 58.9 + mg i & 8717.820 & 5.933 & 68.0 & 103.0 + this table is published in its entirety in the electronic edition of the _ astrophysical journal .
_ a portion is shown here for guidance regarding its form and content . the abundance indicated from the ew of each absorption line
was computed using the _ abfind _ driver of the 2014 version of the spectrum synthesis program moog @xcite .
this analysis required the use of stellar atmospheres generated from the kurucz grid of lte models , with updated opacity distributions from castelli & kurucz ( 2004 , henceforth kurucz models ) .
for the sun we employed the scaled - solar , odfnew , kurucz model . since lines in the spectra of our metal - poor halo stars indicated @xmath2-element ( e.g. , o , mg , si , ca , ti ) enhancements , near @xmath230.4 dex , typical for their low metallicity , ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , we employed for those the alpha - enhanced , aodfnew , kurucz models .
we note that the kurucz metallicity parameter , indicated by [ m / h ] , is equal to the adopted [ fe / h ] ( and not the mass fraction of metals ) .
for odfnew models the solar abundance distribution was scaled by [ m / h ] , while for aodfnew models the same scaling was applied , but the alpha - elements ( o , mg , si , s , ca , ti ) were increased by an additional @xmath230.4 dex . the model atmosphere effective temperature , t@xmath17 , for each program star was selected so that the differential abundance ( ultimately relative to the sun ) derived from our fe i lines was independent of line ep . since this effective temperature estimate relies on the saha - boltzmann excitation of fe i we call it the excitation temperature , or t@xmath24 .
the use of excitation temperatures offers the advantage that they are independent of reddening .
furthermore , the average differential fe i abundance , relative to the sun , provides our choice of the model atmosphere metallicity . whilst the final stellar model temperatures employed in our abundance analysis were determined from excitation equilibrium of fe i , the initial guess for effective temperatures
was based on the ( v@xmath0k ) photometric colors for our stars , listed in the simbad database @xcite .
clearly , it is of interest to compare the excitation temperatures , based on the spectra , with the photometric t@xmath17 values , based on calibrations of broad - band colors .
although the majority of the target stars are not significantly affected by reddening , it was noticed that hip87062 and hip98492 showed strong interstellar na i ( 5890.0 , 5895.9 ) lines requiring a reddening correction to the photometric colors .
of the remaining program stars , hip108200 showed some weak na i interstellar absorption .
@xcite have calibrated a relation between interstellar na i ews and reddening that is most sensitive between @xmath25 with an accuracy of 0.05 mag to 0.15 mag depending on the separation of the stellar and interstellar lines , ew values , and quality of the measurement . as our data are of such high resolution it should not be the case that the stellar and interstellar lines would be blended and the largest reddening found in our stars is still rather low compared to the range covered by @xcite we take 0.05 mag as the reddening uncertainty of our three reddened stars .
the na i ews and corresponding reddening factors are provided in table [ table : red ] .
l r r r hipid & na i m ( d1 ) & na i m ( d2 ) & @xmath26 + 46120 & 12.0 & 13.0 & 0.00 + 54639 & 11.2 & 7.6 : & 0.00 + 80679 & 0.0 & 0.0 & 0.00 + 87062 & 184.0 & 159.0 & 0.06 + 87788 & 0.0 & 0.0 & 0.00 + 98492 & 271.1 & 172.8 & 0.11 + 103269 & 0.0 & 0.0 & 0.00 + 106924 & 0.0 & 0.0 & 0.00 + 108200 & 31.6 & 17.5 & 0.02 + these reddening values were applied where necessary to the photometric colors prior to determining the effective temperatures of the stars using color - temperature relations .
we use the @xcite corrections to convert e(b - v ) to e(v - k ) ; as a result , the uncertainty of t@xmath27 for the stars without any reddening is @xmath1830 k. the uncertainty of t@xmath27 for the three stars affected by reddening are listed individually in table [ table : atmpar ] . because of the metallicity - dependence of the photometric color - temperature relations ( redder colors occur at higher metallicity but also with lower temperature ) , it is necessary to determine both t@xmath17 and [ fe / h ] iteratively , until both are consistent with the colors and spectra .
l c c l |c c c c| c
c c c & & & & & & + hipid & v & ( v - k)@xmath28 & t@xmath29 ( k ) & t@xmath24 ( k ) & logg ( dex ) & @xmath30 ( km / s ) & & t@xmath24 ( k ) & logg ( dex ) & @xmath30 ( km / s ) + 46120 & 10.10 & 1.64 & 5533 & 5100 & 4.64 & 1.30 & & 5250 & 4.20 & 1.60 + 54639 & 11.41 & 2.06 & 4923 & 4850 & 4.72 & 1.40 & & 4900 & 4.30 & 1.70 + 80679 & 11.29 & 1.89 & 5146 & 4950 & 4.73 & 1.70 & & 4950 & 4.40 & 1.30 + 87062 & 10.59 & 1.65 & @xmath31 & 5800 & 4.42 & 1.70 & & 6000 & 4.20 & 1.70 + 87788 & 11.29 & 1.83 & 5284 & 4900 & 4.71 & 1.70 & & 5000 & 4.30 & 1.60 + 98492 & 11.57 & 1.74 & @xmath32 & 5500 & 4.17 & 1.40 & & 5500 & 3.80 & 1.40 + 103269 & 10.33 & 1.72 & 5410 & 5300 & 4.64 & 1.50 & & 5325 & 4.20 & 1.50 + 106924 & 10.39 & 1.82 & 5254 & 5125 & 4.66 & 1.70 & & 5150 & 4.20 & 1.50 + 108200 & 11.03 & 1.87 & @xmath33 & 5175 & 4.68 & 1.75 & & 5175 & 4.60 & 1.50 + the photometric t@xmath17 calibration by ( * ? ? ? * henceforth rm05 ) for our six stars with zero reddening are , on average , @xmath18220 k hotter than the spectroscopic t@xmath24 , as indicated in column 2 of table [ table : atmpar ] .
because of this large temperature difference , if we employ the photometric v@xmath0k t@xmath17 for the model atmospheres we find a steep slope of fe i abundance with line ep , leading to significant uncertainty of fe i abundance . the photometric color t@xmath17 calibrations of ( * ? ? ?
* henceforth cas10 ) result in even hotter stellar temperatures than rm05 , some 415k higher than our spectroscopic t@xmath24 , and give steeper slopes of fe i abundance with ep . figure [ figure : comparetemps ] shows the [ fe i / h ] abundance versus ep plot for hip54639 , which shows the _ smallest _ difference between photometric and excitation temperatures , for three different model atmosphere temperatures : the cas10 color - temperature , the rm05 color - temperature , and the spectroscopic t@xmath24 .
the plots in figure [ figure : comparetemps ] demonstrate an enormous disagreement between temperatures based on photometric colors and those derived from the excitation of fe i lines .
given that the 1@xmath34 uncertainty in the determination of the spectroscopic temperatures is @xmath1825 k , one might ask whether the uncertainty in the reddening alone could account for this difference . a 0.01mag change in @xmath35 would result in a 12k temperature shift .
therefore , with the average discrepancy between the spectroscopic and photometric temperatures , reddening corrections of @xmath180.1mag would need to be applied to account for the difference .
these values are too high given the small uncertainty in the reddening correction . to be consistent ,
we have taken the excitation temperature as the final temperature used throughout the rest of this analysis .
along the same lines of determining a spectroscopic excitation temperature , we find spectroscopic microturbulent velocity parameters , @xmath30 , and surface gravities of our program stars .
for the microturbulent velocity , we followed the usual procedure ( e.g. * ? ? ?
* ; * ? ? ?
* ) of demanding that fe i line abundances are independent of ew .
our reliance on weak , unsaturated , lines for the fe i abundances of our program stars provides relative insensitivity to uncertainties on the exact value of @xmath30 . for the surface gravities
, we make an initial estimate based on 12.0 gyr dsep isochrones with a helium fraction y@xmath36 , @xmath37 = + 0.4 $ ] , and initial [ fe / h ] = @xmath02.0 and adjust this initial estimate until ionization equilibrium between [ fei / h ] and [ feii / h ] is achieved . at the low metallicity of our program
stars the [ @xmath2/fe ] ratios are expected to be enhanced by approximately @xmath230.4 dex ( e.g. * ? ? ?
* ; * ? ? ?
for this reason , we have employed the alpha - enhanced , aodfnew , kurucz model atmospheres in our abundance analysis . however , in order to check that our stars actually possess halo - like @xmath2-enhancements , we have measured the ews of a number of lines from @xmath2-elements , and from those derive [ @xmath2/fe ] abundance ratios in a line - by - line differential analysis . in order to use standard star hip66815 in our line - by - line differential abundance analysis , we first need to determine its atmosphere parameters , [ fe i / h ] , [ fe ii / h ] and [ @xmath2/fe ] values . for our 1d lte differential abundance analysis of hip66815
we require physical parameters for the solar atmosphere , including the microturbulent velocity . here
we adopt the solar t@xmath17 of 5777k and logg of 4.4377dex given by @xcite .
we have obtained the microturbulent velocity for the sun in an absolute abundance analysis , using the fe i line ews and loggf values of @xcite and @xcite , as listed in @xcite . while differences in the loggf scales of the holweger and blackwell groups gave slightly different solar fe abundances , within each loggf scale the fe abundance trend with ep was flat , indicating agreement with the solar effective temperature of 5777k .
a least squares fit of fe i ews versus absolute solar abundances , derived using the asplund lines and holweger loggf values , gave zero slope for a microturbulent velocity parameter of 0.99km / s , with a formal 1@xmath34 uncertainty of @xmath380.07km / s . unfortunately , systematic errors made it impossible to simply include solar fe i abundances found from the lines with blackwell loggf values . therefore , we have adopted a solar microturbulent velocity parameter rounded to 1.0km / s .
our initial physical atmosphere parameters for hip66815 were based on the _ hipparcos _ parallax of [email protected] mas @xcite , v=8.83 , and 2mass v@xmath0k and v@xmath0j colors of 1.44 and 1.10 , respectively , suggesting t@xmath17=5864 k and logg=4.50 dex using the rm05 color - temperature relations , and assuming a mass near 0.76 m@xmath39 , inferred from a 12 gyr dsep isochrone , appropriate for an old metal - poor halo star .
notably , the color - temperature calibration of cas10 would have indicated a higher temperature of 5974 k , for hip66815 .
we have also determined an excitation temperature for hip66815 by finding the model atmosphere yielding fe i abundances independent of line ep .
we only consider lines in hip66815 with ew less than 85m in order to avoid systematic errors arising from difficulties in measuring damped wings of the same lines in the sun , typically near 120130m .
accordingly , with @xmath2-enhanced kurucz model atmospheres and 81 fe i lines , we found an excitation temperature of 5822k , with a 1@xmath34 uncertainty of 31k , and a microturbulent velocity parameter of 1.33 km / s ( see figures [ fig : ewab66815 ] and [ fig : epab66815 ] ) .
these parameters are relatively close to those found by @xcite , who obtained t@xmath17=5812 k , logg=4.41 dex , [ fe / h][email protected] dex , and @xmath30=1.13 km / s .
this star was also analyzed by @xcite and @xcite .
@xcite derived an excitation temperature , based on laboratory gf values and found t@xmath17=5875k , with a 1@xmath34 uncertainty of 40k ; a spectroscopic logg of 4.50 dex , was chosen by @xcite in order to force ionization equilibrium of iron , [ fe i / h]=[fe ii / h ] , again based on laboratory gf values .
@xcite claimed to employ the @xcite atmosphere parameters and computed [ fe / h][email protected] dex , but reported [ fe / h][email protected] dex for the @xcite value ; this conflicts with the on - line data of @xcite , which indicates @xmath9}=-$]0.5dex .
we suspect a typographical error in the on - line table entry for [ fe / h ] in @xcite .
@xcite found a microturbulent velocity of 0.95km / s for hip66815 , lower than that found here .
we are encouraged by the good agreement between the excitation temperature and photometric temperatures , indicated by v@xmath0k and v@xmath0j photometry , based on the rm05 temperature - color relations .
therefore , we adopt the average of these three temperatures ( from v@xmath0k , v@xmath0j , and t@xmath24 ) , at t@xmath17=5850 k and @xmath40=24 k. we reject the photometric temperature indicated by the cas10 calibration , as this is 152 k higher than our excitation temperature , nearly 5@xmath34 beyond our estimated uncertainties .
such a high temperature would impose a strong slope of fe i abundance with line ep and make it difficult to determine [ fe / h ] .
interestingly , @xcite found that the cas10 t@xmath17 values were hotter than direct measurements by 131 k in a comparison of 69 dwarf stars . on the other hand ,
a disagreement is implied with @xcite , who claim no significant shift between their color - temperature scale and excitation temperatures .
the final adopted t@xmath17 of 5850k and @xmath9}=-0.72$]dex , consistent with the spectra , when combined with the 2mass k - band magnitude and _ hipparcos _ parallax are best fit with an 11 gyr dsep isochrone and @xmath41dex .
for an alpha - enhanced kurucz model for hip66815 , with t@xmath17=5850 k , logg=4.50 dex , and @xmath9}=-0.75$]dex , the 81 fe i line ews give [ fe / h][email protected] dex , with 1@xmath34 dispersion of 0.055 dex , providing a formal random error on the mean of 0.006 dex . if a temperature of 5822 k is employed the [ fe / h ] is lowered by an additional 0.02 dex .
based on 11 fe ii lines this model for hip66815 gives an average [ fe / h][email protected] dex , with 1@xmath34 dispersion of 0.032 dex , indicating an error on the mean of 0.010 dex .
thus , our analysis of hip66815 has not achieved ionization balance .
various factors may have contributed to the 0.14 dex difference between [ fe i / h ] and [ fe ii / h ] : an over - estimated model atmosphere electron number density ( n@xmath42 ) ; an inexplicably low logg , by @xmath180.38 dex , which could be due to a mass 60% lower than indicated by the isochrones , or to an under - estimated luminosity , due to a @xmath1855% error in the measured parallax , or an actual stellar radius 55% larger than the dsep isochrones ; a strong non - lte over - ionization of iron , leading to low [ fe i / h ] ; or an under - estimated t@xmath17 by @xmath18150 k , perhaps due to non - lte or 3d effects , also leading to low [ fe i / h ] . of these possibilities
, an overestimated n@xmath42 could result from an inappropriate [ @xmath2/fe ] model atmosphere parameter . in table [
table : ab66815 ] , we show the results of our differential chemical abundance analysis of hip66815 , relative to the sun , based on the ews listed in table [ table : ew66815 ] .
our measurements indicate [ @xmath2/fe ] ratios for hip66815 that are relatively low for its metallicity , near @xmath230.10 dex for the electron donors mg , si , ca and ti , but @xmath230.28 dex for o. this is much less than the @xmath230.4 dex enhancements employed in the kurucz model atmospheres . thus , a solar composition model atmosphere is more appropriate for hip66815 than an alpha - enhanced model ; use of the solar composition model atmosphere for hip66815 gives [ fe i / h][email protected] ( rms dispersion 0.055 dex ) and [ fe ii / h][email protected] ( rms dispersion 0.030 dex ) . interpolating between models for [ @xmath2/fe][email protected] dex , which is appropriate for the electron donor alpha - elements , we obtain [ fe ii / h][email protected] dex , as the best estimate for hip66815 ; thus , the fe ionization imbalance is reduced to 0.10 dex .
l r r r ion & [ x / fe ] & @xmath34 & n@xmath43 + fe i & @xmath00.72 & 0.06 & 81 + fe ii & @xmath00.59 & 0.03 & 11 + o i & @xmath230.28 & 0.07 & 3 + na i & @xmath00.05 & 0.04 & 2 + mg i & @xmath230.12 & 0.18 & 8 + al i & @xmath00.03 & 0.04 & 5 + si i & @xmath230.10 & 0.03 & 9 + ca i & @xmath230.09 & 0.05 & 7 + ti i & @xmath230.09 & 0.12 & 4 + ti ii & @xmath230.09 & ... & 1 + use of the marcs model atmospheres @xcite gives the same abundances as the kurucz model , with differences less than 0.01 dex .
while hip66815 appears to be relatively @xmath2-deficient for its [ fe / h ] , [ o / mg ] versus [ o / h ] and [ mg / h ] are completely consistent with normal trends throughout the milky way components ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
regarding over - ionization of fe i by non - lte effects in hip66815 , the calculations of @xcite show this is very small , near @xmath180.03 dex , but negligible in a differential abundance of hip66815 relative to the sun .
furthermore , the @xcite calculations also showed a negligible non - lte effect ( less than 10 k ) on the derived excitation temperature .
negligible non - lte differential corrections for fe abundances in hip66815 are supported by the review of @xcite and the discussion in section [ sec : nlte ] .
interestingly , the non - lte calculations of @xcite showed no significant errors in spectroscopically derived 1d - lte teff and logg for dwarf stars similar to hip66815 .
the 1d - lte teff and logg should give an answer very close to the truth ( @xmath44 k , @xmath45dex ) .
in the same paper , @xcite found an analysis of the 1d spatially averaged 3d ( henceforth @xmath43d@xmath5 ) solar model of @xcite indicated non - lte corrections for fe i lines always less than @xmath180.02 dex ( i.e. negligible ) .
similar @xmath43d@xmath5 models are discussed further in section [ sec:3dmodels ] .
notwithstanding uncertainties regarding 3d model atmospheres , none of the effects listed above appears capable of explaining the 0.10 dex fe ionization imbalance . however , it is possible that a combination of systematic effects might be the cause of the imbalance .
for example , a 30 k , 1@xmath34 , increase in t@xmath17 reduces @xmath46fe by 0.03dex ; a 2@xmath34 decrease in the parallax of hip66815 accounts for a further reduction of 0.04 dex ; and a 0.020.03 dex non - lte differential fe over - ionization effect is not unreasonable . to aid in understanding the roles of systematic and random errors on the fe ionization balance
, it would be useful to consider fe i and fe ii lines for all our program stars .
concerns about increased 3d effects at low metallicity would also be addressed with [ fe / h ] measured from fe ii lines . until the ionization imbalance is resolved we have two metallicity scales for hip66815 : 81 fe i lines favor [ fe / h][email protected] dex ( rms 0.06 dex ) , while 11 fe ii lines favor [ fe / h][email protected] dex ( rms 0.03 dex ) .
similar to the analysis of hip66815 relative the sun we performed a line - by - line differential abundance analysis of our program stars relative to hip66815 , using the ews listed in tables [ table : felines ] and [ table : alphalines ] and the atmosphere parameters listed in table [ table : atmpar ] .
the abundance ratios were normalized to the solar scale using our hip66815 chemical composition , given in table [ table : ab66815 ] , with the final lte [ fe / h ] results listed in table [ table : feresults ] .
we remind the reader that @xmath2-enhanced , aodfnew , kurucz models were employed for both the program stars and the standard , hip66815 , in the differential analysis .
table [ table : feresults ] lists our lte [ fe i / h ] and [ fe ii / h ] results . as will be discussed in section [ sec : ioneq ] , the use of isochrone gravities resulted in inconsistent fe i and fe ii abundances for all of our stars . in order to approach ionization equilibrium , where fe i and fe ii abundances are equal , it was necessary to use a logg that was , on average , 0.4dex lower than that predicted by the dsep isochrones .
because of this , in the following discussion we provide atmospheric parameters and abundances both with and without ionization equilibrium in table [ table : atmpar ] . in tables [
table : feresults ] and [ table : aresults ] , the subscript ie indicates results for which the model atmosphere logg was found by forcing ionization equilibrium .
table [ table : aresults ] lists our lte [ x / fe ] ratios for alpha - elements , o , mg and ca . due to the low metallicity
, we were unable to detect o i lines in most of the program stars , and many of the mg i lines .
typically , three ca i lines were seen in our program stars ; however , for our most metal - poor star , hip87788 , only one ca i line was detected . for our most @xmath2-enhanced star , hip98492 , we measured five ca i lines .
thus , our measured [ @xmath2/fe ] ratios are strongly weighted by ca i lines and our measured [ ca / fe ] ratios ( with a mean [ ca / fe ] = 0.34 dex ) are consistent with the enhanced [ @xmath2/fe ] ratios of the mw halo ( * ? ? ? * and references therein ) .
t r c c c c c c c c c & & & + hipid & fe i & fe i@xmath47 & @xmath34 & n & & fe ii & fe ii@xmath47 & @xmath34 & n + 46120 & -2.33 & -2.22 & 0.13 & 40 & & -1.87 & -2.18 & 0.28 & 3 + 54639 & -2.48 & -2.50 & 0.15 & 48 & & -2.14 & -2.46 & 0.28 & 5 + 80679 & -2.49 & -2.50 & 0.08 & 52 & & -2.30 & -2.46 & 0.21 & 12 + 87062 & -1.70 & -1.56 & 0.15 & 52 & & -1.37 & -1.54 & 0.18 & 15 + 87788 & -2.69 & -2.66 & 0.10 & 32 & & -2.42 & -2.60 & 0.24 & 7 + 98492 & -1.40 & -1.40 & 0.08 & 69 & & -1.17 & -1.35 & 0.13 & 6 + 103269 & -1.85 & -1.83 & 0.07 & 66 & & -1.64 & -1.84 & 0.20 & 16 + 106924 & -2.22 & -2.23 & 0.09 & 60 & & -1.86 & -2.25 & 0.21 & 11 + 108200 & -1.82 & -1.83 & 0.12 & 47 & & -1.75 & -1.79 & 0.12 & 5 + r c c c c c c c c c c c c c c & & & & & + hipid & o & o@xmath47 & @xmath34 & n & & mg & mg@xmath47 & @xmath34 & n & & ca & ca@xmath47 & @xmath34 & n + 46120 & ... & ... & ... & ... & & 0.12 & 0.08 & n / a & 1 & & 0.41 & 0.41 & 0.15 & 3 + 54639 & ... & ... & ... & ... & & ... & ... & ... & ... & & 0.27 & 0.33 & 0.12 & 3 + 80679 & ... & ... & ... & ... & & ... & ... & ... & ... & & 0.35 & 0.38 & 0.13 & 3 + 87062 & 0.01 & 0.11 & 0.02 & 3 & & 0.54 & 0.47 & 0.50 & 3 & & 0.26 & 0.21 & 0.03 & 3 + 87788 & ... & ... & ... & ... & & ... & ... & ... & ... & & 0.78 & 0.79 & n / a & 1 + 98492 & 0.69 & 0.71 & 0.08 & 3 & & 0.39 & 0.41 & n / a & 1 & & 0.19 & 0.20 & 0.05 & 5 + 103269 & 0.68 & 0.69 & n / a & 1 & & 0.15 & 0.14 & n / a & 1 & & 0.23 & 0.22 & 0.08 & 5 + 106924 & ... & ... & ... & ... & & 0.28 & 0.27 & n / a & 1 & & 0.40 & 0.40 & 0.05 & 3 + 108200 & ... & ... & ... & ... & & 0.06 & 0.05 & n / a & 1 & & 0.12 & 0.14 & 0.06 & 3 + as can be seen from table [ table : feresults ] , our isochrone logg atmospheres gave [ fe i / h ] abundance results significantly lower than [ fe ii / h ] for all of our stars ; on average , the fe i abundances were 0.27 dex lower than those from fe ii .
this exceeds the ionization imbalance for hip66815 of 0.10 dex .
if the model atmospheres and spectral synthesis programs replicate stellar atmospheres perfectly then the fe i and fe ii abundances should be the same .
a solution to the ionization imbalance might be sought in non - lte corrections , since non - lte effects often over - ionize the minority neutral species , like fe i , but leave the dominant , singly ionized species , fe ii , relatively unaffected , and result in a departure from ionization equilibrium ( e.g. * ? ? ?
* ) . however , the non - lte corrections for fe i lines in our program stars are negligible , of order 0.01 dex , based on calculations from the inspect program , v1.0 ( see @xcite and section [ sec : nlte ] ) . in lte model atmospheres ,
the electron number density , @xmath48 , controls the strength of lines from the dominant ionization state ( e.g. , fe ii in our stars ) , and thus determines whether ionization equilibrium is obtained .
while n@xmath42 is strongly sensitive to the surface gravity , it is also sensitive to other effects , like over - all metallicity , the abundance of low - ionization species , such as @xmath2-elements , and temperature .
typically , fe ii lines are stronger in lower surface gravity models whilst fe i line strengths do not change much with logg ; because of this , it has become common practice to choose a spectroscopic gravity by finding the model atmosphere that results in ionization equilibrium . in this analysis
we have modified the model atmosphere gravity in order to approach ionization equilibrium .
this has resulted in an average decrease in logg of 0.4 dex from the isochrone values . in lte ionization equilibrium ,
our fe ii abundances are still , on average , 0.03 dex higher than the fe i abundances , quite similar to the scale of the non - lte abundance corrections .
the non - lte abundance corrections are discussed in greater detail , for various elements , in the next section . here ,
we consider the potential corrections , due to non - lte effects , to be applied to our differential lte abundances . because our very metal - poor program stars lack strong line blanketing at blue and uv wavelengths , the path length of such photons is large .
thus , uv photons from hot regions of the stellar atmosphere can travel to , and be absorbed in , much cooler regions , and thereby affect the excitation and ionization populations of atomic species . to compute synthetic spectra in non -
lte requires a detailed consideration of radiative and collisional excitation and de - excitation rates for each level of the atom or ion , and consideration of radiation from the whole atmosphere .
this is computationally much more taxing than the simple lte situation where the total population of a species is derived from one atomic transition using the saha - boltzmann equation .
in addition , it is unfortunate for the non - lte case that collisional rates are normally not well known ; in particular , the rate of collision with hydrogen atoms is parameterized with a scaling factor applied to the formula of @xcite .
since our differential , line - by - line , abundance analysis was performed relative to the solar spectrum , only the differential non - lte corrections between the sun and program stars need to be applied to our lte abundances .
we first consider non - lte effects on fe because we employ fe i and fe ii line abundances to constrain our model atmosphere parameters .
we are encouraged by the extensive 1d non - lte investigation of @xcite showing non - lte corrections to the fe i and fe ii abundances for our program stars are negligible , with no significant effect on the spectroscopically derived atmospheric parameters .
specific non - lte calculations by bergemann ( 2011 unpublished ; but see also @xcite ) , showed @[email protected] dex non - lte abundance corrections for fe i lines used here ; however , differential to the sun , the non - lte fe i corrections are + 0.02 to + 0.03 dex higher than our lte results . on the other hand , the non - lte web tool inspect program , v1.0 ( see @xcite ) , indicates non - lte corrections for fe i lines used in our program stars of only + 0.01 dex . this magnitude for the non - lte effect on fe i abundances in our stars is supported by the review of non - lte by @xcite .
non - lte corrections for the o i infrared triplet lines used in this analysis ( at 7772 , 7774 , and 7775 ) have been computed for an extensive grid of fgk dwarf model atmospheres by @xcite .
the results are similar to previous calculations ( e.g. , * ? ? ?
* ; * ? ? ?
these non - lte corrections increase strongly with temperature and with decreasing log g. typically , the lower levels of the o i triplet lines are over - populated relative to lte in our stars , so the lte abundances are too high .
however , since the non - lte correction is larger for the sun than for our , relatively cooler program stars , net positive corrections to the lte [ o / h ] results are required for our program stars .
for the three program stars where we were able to measure o i triplet line ews , we have interpolated the grid of non - lte corrections of @xcite to the appropriate atmosphere parameters ; we also interpolated the grid of non - lte o i corrections for the sun .
consequently , the non - lte oxygen abundance corrections for hip87062 , hip98492 , and hip103269 are @xmath00.07 , @xmath00.13 , and @xmath00.05dex , respectively . for differential abundance
relative to the sun the solar non - lte correction , at @xmath00.18 dex , must be subtracted .
thus , the [ o / h ] ratios of our three program stars should be increased by @xmath230.11 , @xmath230.05 , and @xmath230.13 dex , respectively .
the non - lte corrected [ o / fe ] ratios must also include the differential non - lte effect on the fe abundance , of approximately @xmath230.02 dex .
thus , the lte [ o / fe ] ratios for hip87062 , hip98492 , and hip103269 should be increased by @xmath230.09 , @xmath230.03 and @xmath230.11 dex respectively . for our mg i lines we consult @xcite , who computed non - lte corrections for a range of stellar parameters , encompassing the parameters similar to those of the stars studied here .
these corrections are sensitive to temperature , with larger corrections in hotter stars . for the mg
i lines employed here , the @xcite absolute non - lte corrections are all near , or below , 0.10 dex ; and , the non - lte correction for the stars near 5200 k is 0.07 dex .
the coolest star in @xcite , with a temperature of 5070 k has a non - lte mg i correction of only 0.03 dex .
relative to the sun , the mg i non - lte effect in our metal - poor dwarf stars is reduced due to the solar non - lte correction .
thus , our hotter program stars have a mean non - lte mg i differential correction of 0.07 dex , while for stars at 5200 k the differential correction is only 0.02 dex . for non - lte corrections to our ca i lte abundances we used the extensive compilation of @xcite , whose computations cover most of the parameter space of our program stars and includes 16 ca i lines , including good overlap with the lines employed here . while the solar non - lte correction is small , near 0.02dex , corrections for our program stars are all near 0.14dex .
thus , a differential non - lte correction of + 0.12dex should be added to our ca i abundance values . in order to estimate systematic abundance errors , due to uncertainties in atmosphere parameters
, we performed an error analysis in which we varied the atmospheric parameters by conservative quantities ( @xmath49k ; @xmath50dex ; @xmath51 0.1 km / s ; [ m / h ] @xmath38 0.1 dex ) and recomputed abundances in order to determine the effect each parameter may have on the final abundance ratios .
we show the effects on three stars in table [ table : errors ] listed as the difference of the original abundance from the abundance with the modified parameters .
c c r r r r r r
r r r r r hipid & & & & & & & & + + & & -50 k & + 50 k & & -0.4 dex & + 0.4 dex & & -0.1 km / s & + 0.1 km / s & & -0.1 dex & + 0.1 dex + & [ fe i / h ] & -0.06 & 0.04 & & -0.05 & 0.04 & & 0.01 & -0.02 & & 0.01 & -0.01 + hip87788 & [ fe ii / h ] & 0.02 & -0.01 & & -0.19 & 0.18 & & 0.01 & -0.01 & & -0.01 & 0.02 + & [ ca / fe ] & 0.03 & -0.04 & & -0.01 & 0.03 & & 0.01 & @xmath52 & & 0.01 & @xmath52 + & [ fe i / h ] & -0.05 & 0.05 & & -0.02 & 0.02 & & 0.01 & -0.01 & & @xmath52 & @xmath52 + & [ fe ii / h ] & 0.02 & 0.02 & & -0.16 & 0.17 & & 0.01 & -0.01 & & -0.01 & 0.02 + hip103269 & [ o / fe ] & 0.04 & -0.04 & & -0.02 & @xmath52 & & -0.01 & 0.01 & & 0.01 & -0.02 + & [ mg / fe ] & 0.03 & -0.02 & & 0.01 & @xmath52 & & @xmath52 & 0.01 & & 0.01 & @xmath52 + & [ ca / fe ] & 0.02 & -0.02 & & 0.01 & -0.01 & & -0.01 & 0.01 & & 0.01 & @xmath52 + & [ fe i / h ] & -0.05 & 0.04 & & -0.04 & 0.04 & & @xmath52 & -0.01 & & @xmath52 & -0.01 + & [ fe ii / h ] & 0.01 & -0.01 & & -0.19 & 0.18 & & 0.01 & -0.01 & & -0.01 & 0.01 + hip106924 & [ mg / fe ] & 0.05 & -0.01 & & 0.02 & @xmath52 & & @xmath52 & 0.01 & & 0.01 & 0.01 + & [ ca / fe ] & 0.01 & -0.01 & & 0.02 & -0.02 & & @xmath52 & 0.01 & & @xmath52 & @xmath52 + as indicated in section [ sec : stellarparameters ] for star hip54639 , a 1@xmath34 50k effective temperature uncertainty in the error analysis is conservative for the stars in our sample . by using the spectroscopic temperatures and requiring ionization equilibrium to determine surface gravities we find the spectroscopic surface gravity to be 0.4 dex lower than that predicted from the isochrones and assume a conservative offset of 0.4 dex as our maximum logg uncertainty .
we also assume a possible microturbulent velocity uncertainty of 0.05 km / s .
the results provided in table [ table : errors ] may be used to estimate the systematic abundance uncertainties .
we add in quadrature the average uncertainty from each parameter in table [ table : errors ] to determine the average systematic uncertainty .
the actual uncertainties are likely to be smaller due to the covariance among parameters @xcite .
we find @xmath53 dex and @xmath54 dex for [ fe i / h ] and [ fe ii / h ] , respectively .
the size of the @xmath40 uncertainty on [ fe ii / h ] is due mainly to the uncertainty incurred by the reduction in logg by @xmath180.4 dex to obtain ionization equilibrium .
the @xmath2-element abundance uncertainties are @xmath55 dex , @xmath56 dex , and @xmath57 dex , where the high uncertainty for [ mg / fe ] stems from the lack of available line measurements .
the stellar abundances for each star , including the total @xmath40 uncertainties , are provided in table [ table : ieresults ] .
the uncertainties are found by adding quadrature the average systematic uncertainty and the rms uncertainty of the abundance for a given star . for stellar abundances that are measured from a single line ,
the rms abundance uncertainty is the average of rms abundance uncertainties for that species from stars with multiple lines measured ( including hip66815 ) .
l c c c c c hipid & [ fei / h ] & [ feii / h ] & [ o / fe ] & [ mg / fe ] & [ ca / fe ] + 46120 & @xmath58 & @xmath59 & ... & @xmath60 & @xmath61 + 54639 & @xmath62 & @xmath63 & ... & ... & @xmath64 + 80679 & @xmath62 & @xmath65 & ... & ... & @xmath66 + 87062 & @xmath67 & @xmath68 & @xmath69 & @xmath70 & @xmath71 + 87788 & @xmath72 & @xmath73 & ... & ... & @xmath74 + 98492 & @xmath75 & @xmath76 & @xmath77 & @xmath78 & @xmath79 + 103269 & @xmath80 & @xmath81 & @xmath82 & @xmath83 & @xmath84 + 106924 & @xmath85 & @xmath86 & ... & @xmath87 & @xmath88 + 108200 & @xmath80 & @xmath89 & ... & @xmath90 & @xmath91 + although 1d stellar atmosphere models have improved over the years , for example by including updated atomic and molecular line list data , opacities , and a variety of chemical compositions ( e.g. * ? ? ?
* ; * ? ? ?
* ) , there are still issues with their assumption of full radiative equilibrium for dwarf metal - poor stars @xcite and their dependence on an adjustable microturbulent velocity parameter inherent in the underlying mixing - length theory . here
, we use the stagger grid of @xmath43d@xmath5 hydrodynamic model atmospheres @xcite to explore the effects of 3d model atmospheres on our derived abundances .
we remind the reader that the @xmath43d@xmath5 models are horizontally averaged versions of the full-3d hydrodynamic models , as described by @xcite . for the sun , we employed the fe i lines used in the full 3d analysis of @xcite with the ews measured here from the kurucz solar atlas , and the @xcite @xmath43d@xmath5 solar model .
this analysis gave a solar fe i abundance of 7.48 dex , higher than the 3d @xcite value of 7.44 dex .
interestingly , the kurucz solar model gives an fe abundance of 7.44 dex from the same lines , in agreement with the full 3d results .
the line - by - line ratio of the individual @xcite fe i 3d abundances to those computed with the @xcite @xmath43d@xmath5 solar model showed a strong slope with line ep , as seen in figure [ fig-3d3dsolarfei ] .
this suggests that the @xmath43d@xmath5 model is @xmath18150k hotter than the full 3d model .
note that the slope in figure [ fig-3d3dsolarfei ] can not be due to errors in loggf values , as these cancel - out in the @xmath43d@xmath5 to 3d ratio .
unfortunately , the slope with ep means that the derived fe i abundance , relative to the full 3d model , depends on the ep of the line employed and adds significant uncertainty to abundance measurements .
if we employ the @xmath43d@xmath5 excitation equilibrium to determine the model t@xmath17 for our program stars we must consider that the solar @xmath43d@xmath5 excitation temperature is cooler than the actual solar t@xmath17 by 150k . such a cool solar @xmath43d@xmath5 model results in lower fe i abundances , by 0.08dex , and an absolute mean fe i abundance of 7.36dex , with an rms scatter about the mean of 0.06 dex , for the lines used in the @xcite analysis
clearly , the horizontally averaged @xmath43d@xmath5 solar model atmosphere @xcite fails to reproduce the full 3d fe i abundances results of @xcite .
we note that this is can not be due to differences in saturation between the @xmath43d@xmath5 and 3d models , because the trend in figure [ fig-3d3dsolarfei ] is unchanged when only weak lines are considered and there is no clear trend in a plot of abundance differences with ew . in order to estimate the effect of the @xmath43d@xmath5 models on our results , we have computed line - by - line differential spectroscopic t@xmath24 , fe i and fe ii abundances using the @xmath43d@xmath5 models for our standard star , hip66815 , and one representative program star , hip103269 , which has good s / n and is slightly more metal - rich than many of the other program stars , so its lines are stronger , making the comparison of the effects of kurucz versus @xmath92 models easier to identify .
a comparison of these @xmath92 results with those computed with kurucz models is presented in table [ table:3dkurfeh ] .
presently , the uncertainties in table [ table:3dkurfeh ] include random errors only and so should be combined , in quadrature , with the systematic errors discussed in section [ sec : errors ] . for our standard star , hip66815 ,
the line - by - line differential abundance , relative to the sun , using the @xcite @xmath43d@xmath5 models , gave fe i line excitation equilibrium for a model with t@xmath17=5970k ; this is 120k hotter than the excitation temperature derived using the kurucz models .
notably , this higher temperature agrees well with the photometric v@xmath0k temperature derived using the @xcite calibration . our @xmath43d@xmath5 abundance analysis of hip66815 gave [ fe / h][email protected] dex from 84 fe i lines , but @xmath9}=-0.59$]dex from 13 fe ii lines .
the scatter of both fe i and fe ii lines , at 1@xmath34=0.060 dex , indicates an error on the mean of 0.017 and 0.007 dex , respectively .
thus , the @xmath43d@xmath5 model for hip 66815 , with the appropriate photometric gravity , did not provide ionization equilibrium .
compared to the hip66815 results from the kurucz models , the @xmath43d@xmath5 models gave t@xmath17 hotter by 120k , [ fe
i / h ] higher by 0.06 dex , and [ fe ii / h ] higher by 0.03 dex .
neither grid achieved ionization equilibrium for the isochrone gravities , although the @xmath43d@xmath5 models gave slightly closer fe i and fe ii abundances . for our program star , hip103269 ,
the @xmath43d@xmath5 model suggested an excitation temperature of 5580k compared to 5300k for the kurucz model .
thus , the difference in excitation temperature between the two grids of model atmospheres appears to increase toward lower metallicity .
for both hip66815 and hip103269 the kurucz model excitation temperatures are close to the rm05 photometric values , while the @xmath43d@xmath5 model excitation temperatures are close to the @xcite scale .
we found the fe abundance of hip103269 using the @xmath43d@xmath5 model of [ fe i / h][email protected] dex and @xmath93}=-1.64$]dex .
thus , we again find that neither model grid obtains fe ionization equilibrium using the gravities indicated by the dsep isochrones , although the situation is slightly better for the @xmath43d@xmath5 models .
furthermore , the @xmath43d@xmath5 [ fe i / h ] scale is higher than that of the kurucz grid by @xmath180.07 dex .
r c c c |c c c hipid & t@xmath24 ( kurucz ) & [ fei / h ] & [ feii / h ] & t@xmath24 ( @xmath43d@xmath5 ) & [ fei / h ] & [ feii / h ] + 66815 & 5850 & @[email protected] & @[email protected] & 5970 & @[email protected] & @[email protected] + 103269 & 5300 & @[email protected] & @[email protected] & 5580 & @[email protected] & @[email protected] + while the @xmath43d@xmath5 models of @xcite provide abundances slightly closer to ionization equilibrium , the range of t@xmath17 , logg and [ m / h ] are not sufficient to properly include the parameters of our program stars ; in particular , extra models with logg=5.00 at low metallicities would be most helpful .
our calculations for table [ table:3dkurfeh ] show that lte abundance analysis using @xmath43d@xmath5 model atmospheres gives hotter spectroscopic excitation temperatures than the kurucz models , and that this t@xmath24 difference increases towards progressively more metal - poor stars .
this is consistent with the results of @xcite , who concluded that lte analysis of metal - poor stars using the @xmath43d@xmath5 models leads to over - estimated t@xmath24 values .
this resulted from large non - lte corrections , increasing with decreasing line ep , in the cool regions of the upper atmospheres of the low - metallicity @xmath43d@xmath5 models .
conversely , @xcite found small non - lte corrections , roughly constant with line ep , when using the 1d marcs models of @xcite . based on the work of @xcite
, we conclude that the @xmath43d@xmath5 models should only be used with non - lte abundance analyses . for lte abundance analysis
the 1d results are preferred over @xmath43d@xmath5-lte results ; however , this does not mean that 1d lte results are any closer to the truth than @xmath43d@xmath5 models with full non - lte analysis .
several previous studies have attempted to determine the abundances of our target stars in their analyses of the chemical composition of metal - poor dwarf stars in the milky way . in performing this literature comparison , which we present in table [ table : litcomp ] , we specifically used the atmosphere parameters and abundances found when we attempted to achieve ionization equilibrium . lcrccc comparison & group & star & @xmath94 ( k ) & @xmath46logg ( dex ) & @xmath46[fe / h ] dex + @xcite & 2 & 46120 & -200 & -0.3 & -0.13 + @xcite & 1 & 103269 & 25 & -0.4 & -0.13 + @xcite & 2 & 103269 & -85 & -0.38 & -0.08 + & & 106924 & -241 & -0.43 & -0.22 + @xcite & 2 & 87062 & 400 & -0.3 & 0.11 + & & 103269 & 25 & -0.4 & -0.13 + @xcite & 1 & 87062 & 245 & 0.0 & -0.09 + & & 103269 & 200 & -0.09 & -0.06 + & & 106924 & 83 & -0.19 & -0.28 + @xcite & 1 & 87788 & 25 & 0.02 & 0.11 + & & 108200 & -151 & -0.4 & -0.21 + @xcite & 1 & 87062 & -392 & -0.19 & -0.28 + @xcite & 1 & 103269 & 200 & -0.09 & -0.11 + @xcite & 1 & 87062 & -216 & -0.61 & -0.17 + @xcite & 2 & 87788 & -599 & -0.7 & -0.49 + & & 103269 & -562 & -0.8 & -0.34 + & & 106924 & -566 & -0.8 & -0.41 + & & 108200 & -204 & -0.4 & -0.13 + @xcite & 1 & 80679 & -30 & 0.0 & 0.12 ( [ feii / h ] ) + & & 108200 & 85 & 0.30 & -0.24 ( [ feii / h ] ) + we find a median difference of -0.13 dex between the abundances determined in this study and those from previous studies @xcite .
it is important to note that the direct comparisons made in table [ table : litcomp ] are complicated by the fact that the studies do not use the same methodology nor do they make use of a consistent line list . as discussed in @xcite ,
group - to - group discrepancies could be minimized by using a homogeneous method , line list , and atmospheric parameters when performing an abundance analysis . as such
, we have separated our comparisons based on methodology , specifically , the method in which the effective temperature is determined .
group 1 indicates studies that used the spectroscopic technique of finding @xmath95 by minimizing the trend in fei abundance with ep while group 2 indicates studies that used photometric colors to determine @xmath96 .
recently , @xcite showed that this choice in methodology for determining stellar temperature creates a significant difference when comparing stellar atmosphere parameters and abundances ; where essentially no difference is found between @xcite and other spectroscopic @xmath95 studies ( @xmath97 k , @xmath46 logg= -0.24@xmath98 and @xmath99 } = -0.13\pm0.22 $ ] dex from 80 stars ) , much larger and statistically significant differences are found when comparing to studies that use photometric temperatures ( @xmath100 k , @xmath46logg@xmath101 and @xmath99 } = -0.21\pm0.18 $ ] dex from 110 stars ) .
we find similar results to @xcite in our comparisons to group 1 studies which use spectroscopic @xmath95 with with a mean @xmath102 k , @xmath46 logg@xmath103 and @xmath99}=-0.14\pm0.11 $ ] dex .
we also find larger differences between our study and those in using photometric temperatures in group 2 with a mean @xmath104 k , @xmath46 logg@xmath105 and @xmath99}=-0.20\pm0.17 $ ] dex
. there are several important implications of this work that stem from the overall improvement of accuracy in low metallicity stellar models .
the accuracy of these models can be tested using metal - poor ms stars with accurate abundances and parallax measurements .
we briefly discuss the level of agreement between our target star observations and the model isochrones from dsep , but a more detailed analysis can be found in @xcite .
eight of the stars used in this study also have accurate _ hst _ parallaxes provided in @xcite making them suitable candidates for ms - fitting .
we performed a reduced @xmath106 analysis in v and i similar to that done in the native hst filters by @xcite in which the location of each star in a cmd is compared to a 12 gyr isochrone with the given star s metallicity . the choice of age for the comparison isochrones should not affect the ms - fitting as the location of the ms is independent of age for a given metallicity .
we find the median deviation of the isochrones from the target star observations to be @xmath107 for isochrones constructed using the @xcite color - temperature relation .
the metallicity range of these stars may seem too large to be used in ms - fitting ; however , it is important to note that , compared to the alternative of having only a single star from which to base our calibrations , this is a significant improvement . in figure [
fig : targetisochrones ] we show the absolute magnitude and reddening corrected colors of these eight stars in a cmd compared to 12 gyr dsep isochrones for their respective metallicities ; we exclude hip80679 as it did not have an hst parallax measurement at the time this analysis was performed .
hip98492 and hip87062 have been labeled in the figure as they are obvious outliers .
the low isochrone logg we find for hip98492 along with a bright absolute magnitude ( @xmath108 mag ) suggests that this star is actually a sub - giant , while the overly red color of hip87062 suggests that either we have underestimated the reddening by at least 0.1 mag or that hip87062 is an unresolved binary that was not filtered out properly . on the other hand
, the agreement between the observed and predicted colors and magnitudes of the remaining six stars in figure [ fig : targetisochrones ] suggests that , although we may be uncertain about the underlying atmospheric parameters , the dsep models provide a good representation of these metal - poor ms stars and can be used with confidence in the ms - fitting of globular clusters .
we performed a detailed , line - by - line , differential chemical abundance analysis , ultimately relative to the sun , for nine metal - poor dwarf stars , near @xmath9}=-2.0$]dex . a virtue of this method is the results are independent of line gf values .
additionally , use of this method should lessen the abundance uncertainty due to effects which may not be included in 1d lte models ( e.g. unidentified blends , non - lte , 3d and chromospheres ) .
the spectrum synthesis program used here , moog @xcite , includes reliable van der waals damping constants from @xcite , enabling calculation of synthetic line profiles for damped lines .
however , we found it difficult to distinguish between subtle damped wings and weak blending features for lines larger than @xmath18130 @xmath109 in the intermediate comparison star , hip66815 , and the sun and therefore chose not to consider lines with ews larger than this . for a robust abundance analysis we targeted unsaturated lines , on the linear portion of the curve of growth , where the ew is directly proportional to the number of absorbers .
because of the large metallicity difference between the sun and the target stars , line - by - line differences for unsaturated lines could only be accomplished using an intermediate standard .
we chose the metal - poor ms star hip66815 , with [ fe / h]@xmath18 - 0.7 dex , as our standard .
however , we note an additional standard , near [ fe / h]@xmath18 - 1.5 dex , would have been helpful ; certainly , there would have been more unsaturated lines with which to compute abundance differences . for our model atmosphere parameters we chose to consider both spectroscopic and photometric indicators
we ultimately chose the model atmosphere t@xmath110 that gave a flat trend of differential abundance with ep , @xmath111 , of fe i lines .
we employed two sets of gravities , the first of which is given by the dartmouth stellar evolution code , for a 12 gyr isochrone with [ fe / h]=@xmath02 , and [ @xmath2/fe]=+0.4 dex , the second being the surface gravity that gave ionization equilibrium .
the average non - lte corrected [ ca / fe ] ratio for our program stars is @xmath230.33dex , which is close to values seen in halo stars , near @xmath230.35 to @xmath230.40 dex ( e.g. * ? ? ?
* ; * ? ? ?
if we remove the highest [ ca / fe ] ratio , [ ca / fe ] = 0.79 dex for hip87788 , then we find an average [ ca / fe ] = 0.26 dex .
since the differential [ ca / fe ] non - lte correction is estimated to be @xmath112 dex , then the non - lte [ ca / fe ] for the 8 stars is @xmath113 dex , consistent with the mw halo ( * ? ? ?
* and references therein ) .
we conclude that our use of the aodfnew , @xmath2-enhanced , kurucz model atmospheres is reasonably consistent with the spectra .
we note that hip108200 has the lowest non - lte corrected [ ca / fe ] ratio , at @xmath230.14 dex ; thus , it is possible that this star belongs to the category of @xmath2-poor halo stars identified by @xcite .
further chemical composition studies of hip108200 and hip87788 would be useful to check these findings .
we have found in this analysis that the spectroscopic gravities derived for these stars are smaller than those calculated from isochrones using a photometric color - temperature relation by @xmath114 dex . because @xcite provides hst parallaxes for eight of these stars we can use these to calculate the gravities as another comparison . using these parallaxes along with the apparent magnitudes and reddenings we can find the absolute magnitudes of the stars to be used in the following relation : @xmath115 @xmath116 where the bolometric corrections ( bc ) are interpolated from the table 3 of @xcite assuming a mass of 0.8m@xmath117 , m@xmath118 , @xmath119k , @xmath120dex . in doing
so we find the spectroscopic logg to be lower than those determined using stellar parallaxes by @xmath121 dex .
this result is similar to that found in @xcite , suggesting that the determination of low logg values might simply be an artifact of determining stellar atmosphere parameters spectroscopically , possibly due to systematic error in the 1d model atmospheres .
if one takes the spectroscopic t@xmath110 and logg as correct , then the disagreement with the theoretical isochrones would indicate a very large error in the isochrone stellar radii . however , the comparison between observations and theoretical dsep isochrones , in color - magnitude space , is reasonably good , with only small color shifts , of order 0.02 mag , required to bring them into agreement , supporting the isochrone results . when ionization equilibrium is not enforced the fe ii line abundances are 0.1 to 0.3 dex higher than our fe i abundances .
while an increase in t@xmath27 of @xmath18300 k could bring the average of these fe i and fe ii abundances into agreement , and would approximately satisfy photometric color - temperature relations ( alonso et al . 1999 ; ramirez & gonzalez 2005 ; casagrande et al . 2010 ) , such an increase in t@xmath27 would result in a strong slope of abundance with ep and a large dispersion on the mean fe abundance .
we also note that the error analysis of table [ table : errors ] shows that a temperature increase of only 150k for a typical program star , hip103269 , would result in a decrease in the [ ca / fe ] ratio by 0.2 dex , inconsistent with the normal halo [ @xmath2/fe ] ratio .
thus , the [ @xmath2/fe ] ratios do not favor an increase of model atmosphere t@xmath17 to resolve the fe ionization equilibrium problem .
because of this , we do not believe that our spectroscopically derived @xmath95 are the source of our ionization imbalance nor can the difference be explained by non - lte over - ionization of fei because the non - lte corrections are quite small for our stars , in the range 0.01 to 0.03 dex ( e.g. * ? ? ?
* ) , at least in 1d .
we explored the possibility that 3d hydrodynamical model atmospheres may bring the isochrones and abundance analysis results into better agreement . indeed , absolute analysis experiments ( e.g. * ? ? ?
* ) of @xmath43d@xmath5 atmospheres ( e.g. * ? ? ?
* ) , give hotter temperatures , based on ep , than the corresponding 1d kurucz models . in our analysis of the solar fe lines using the @xcite @xmath43d@xmath5 atmospheres we compared the @xmath43d@xmath5 abundances with the full 3d results of @xcite and found a trend with ep showing
the @xmath43d@xmath5 excitation temperature is 150k cooler than the actual solar t@xmath17 .
our @xmath43d@xmath5 analysis of hip66815 and hip103269 showed the @xmath43d@xmath5 models gave hotter excitation temperatures than the kurucz models , ranging from 100k to 300k going from solar metallicity to [ fe / h]@xmath18@xmath02 .
the [ fe i / h ] metallicity scale of the @xmath43d@xmath5 models is 0.07 dex higher than the kurucz model grid .
unfortunately , the use of @xmath43d@xmath5 models did not lead to fe ionization equilibrium , although the higher temperatures alleviated part of the problem .
it is clear that 1d and @xmath43d@xmath5 hydrodynamical models combined with separate 1d non - lte effects do not yet account for the atmospheres of real metal - poor ms stars .
we note , however , that the non - lte treatment is more appropriately performed with fully 3d atmospheres , not the averaged @xmath43d@xmath5 investigated by @xcite .
such averaged models are really 1d , where ionizing photons must travel from great depths to distort the excitation and ionization of atoms in cooler regions in upper layers . in the case of fully 3d non - lte
, hot rising bubbles could emit ionizing photons into horizontally adjacent cool regions , over much smaller distances than for the 1d case , possibly enhancing any non - lte effects .
such a situation might lead to the ionization and excitation distortions apparent in this work .
the improved accuracy we were hoping to achieve in this work led us to perform a detailed differential abundance analysis on nine metal - poor dwarf stars , but ultimately highlighted issues regarding the logg comparison to isochrones . given that separate non - lte and @xmath43d@xmath5 calculations do not predict the corrections needed to resolve the logg problem it is still possible that fully 3d hydrodynamics plus non - lte may be at the root of the problem .
however , with [ fe / h ] values known to better than 0.1 dex and photometric observations that match theoretical isochrones within a small window of uncertainty we can now look forward to determining improved distances and ages of the oldest mw gcs . with gaia projected to provide even more accurate parallaxes in the coming years , it will only be a matter of time before we can not only solve the logg problem but have an extensive set of metal - poor stars to use in future calibrations .
support for this work ( proposal number go-11704 & 12320 ) was provided by nasa through a grant from the space telescope science institute which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 .
this material is based upon work supported by the national science foundation graduate research fellowship under grant no .
any opinion , findings , and conclusions or recommendations expressed in this material are those of the authors(s ) and do not necessarily reflect the views of the national science foundation .
the authors also 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 .
alonso , a. , arribas , s. & martinez - roger , c. 1999 , , 140 , 261 amarsi , a.m. , asplund , m. , collet , r. , & leenaarts , j. 2016 , , 455 , 3735 asplund , m. , nordlund , , trampedach , r. , & stein , r. f. 2000 , 359 , 743 athay , r. g. , & lites , b. w. , 1972 , , 176 , 809 barklem , p. s. , piskunov , n. , & omara , b. j. 2000 , , 142 , 467 bensby , t. , feltzing , s. , lundstrm , i. , & ilyin , i. 2005 , , 433 , 185 bergemann , m. , lind , k. , collet , r. , magic , z. , asplund , m. 2012 , 427 , 27 bergemann , m. , & nordlander , t. 2014 , arxiv:1403.3088b beveridge , r. c. , & sneden , c. 1994 , , 108 , 285 blackwell , d. e. , lynas - gray , a. e. & smith , g. 1995 , , 296 , 217 blackwell , d. e. , smith , g. & lynas - gray , a. e. 1995 , , 303 , 575 boesgaard , a. m. , rich , j. , levesque , e. m. , bowler , b. p. 2011 , 743 , 140 carretta , e. , gratton , r. g. , & sneden , c. 2000 , 356 , 238 casagrande , l. , ramirez , i. , melendez , j. , bessell , m. , asplund , m. 2010 , 512 , a54 casagrande , l. , schrnrich , r. , apslund , m. , cassisi , s. , ramrez , i. , melndez , j. , bensby , t. , & feltzing , s. 2011 , 530 , a138 castelli , f. , & kurucz , r.l .
2004 , unpublished , ( arxiv : astro - ph/0405087 ) chaboyer , b. , demarque , p , kernan , p.j . & kraus , l.m .
1998 , , 494 96 chaboyer , b. , mcarthur , b.e . , omalley , e. , benedict , g.f . ,
feiden , g.a . ,
harrison , t.e . , mcwilliam , a. , nelan , e.p . , patterson , r.j . , and sarajedini , a. 2017 , 835 , 152 charbonnel , c. , & primas , f. 2005 , 442 , 961 cohen j.g . ,
christlieb , n. , mcwilliam , a. , shectman , s. , thompson , i. et al .
2004 , , 612 , 1107 collet r. , magic z. , asplund m. , 2011 , j. phys . conf .
ser . , 328 , 012003 conti , p. s. , greenstein , j. l. , spinrad , h. , wallerstein , g. , vardya , m. s. 1967 , , 148 , 105 cox a.n .
in``allen ' 's astrophysical quantities , 4th ed .
publisher : new york : aip press ; springer , 2000 . edited by arthur n. cox .
drawin h.w .
1968 , z. physik 211 , 404 drawin h.w .
1969 , z. physik 225 , 483 deemin , t. j. 1961 , , 123 , 237 dotter , a. , chaboyer , b. , jevremovic , d. , kostov , v. , baron , e. , ferguson , j. w. 2008 , 178 , 89 fabbian , d. , asplund , m. , barklem , p. s. , carlsson , m. , kiselman , d. 2009 , 500 , 1221 flower , p. j. 1996 , , 469 , 355 fulbright , j. p. , 2000 , 120 , 1841 fulbright , j. p. , mcwilliam , a. , & rich , r. m. 2007 , 661 , 1152 gaia collaboration 2016 ( arxiv:1609.04153 ) garz , t. , holweger , h. , kock , m. , & richter , j. 1969 , , 2 , 446 gehren , t. , shi , j. r. , zhang , h. w. , zhao , g. , & korn , a. j. 2006 , , 451 , 1065 gratton , r. g. , et al .
1997 , , 491 , 749 gratton , r. g. , carretta , e. , claudi , r. , lucatello , s. , barbieri , m. 2003 , 404 , 187 grundahl , f. , stetson , p. b. , & andersen , m. i. 2002 , , 395 , 481 gustaffson , b. , edvardsson , b. , eriksson , k. , jorgensen , u. g. , norlund , a. , & plez , b. 2008 , , 486 , 951 heiter , u. , jofre , p. , gustafsson , b. , korn , a.j . , soubiran , c. , & thevenin , f. 2015 , , 582 , a49 hinkel , n. r. , young , p. a. , pagano , m. d. , et al .
2016 , , 226 , 4 hog , e. , fabricius , c. , makarov , v. v. , urban , s. , corbin , t. , wycoff , g. , bastian , u. , schwekendiek , p. , wicenec , a. 2000 , , 355 , 27 holweger , h. , kock , m. & bard , a. 1995 , , 296 , 233 huang , y. , liu , x .- w . , yuan , h .- b . ,
xiang , m .- s . ,
chen , b .- q .
, & zhang , h .- w . , 2015 , , 454 , 2863 ishigaki , m. , chibo , m. , & aoki , w. 2010 , 62 , 143 ishigaki , m. , chibo , m. , & aoki , w. 2012 , 753 , 64 koch , a. & mcwilliam , a. 2008 , 135 , 1551 koch , a. & mcwilliam , a. 2011 , 142 , 63 kopp , g. & lean , j.l .
2011 geophysical research letters , 38 , l01706 latham , d. w. , stefanik , r. p. , torres , g. , davis , r. j. , mazeh , t. , carney , b. w. , laird , j. b. , morse , j. a. 2002 , 124 , 1144 lind , k. , bergemann , m. , & asplund , m. 2012 , , 427 , 50 lindegren , l. , et al .
2017 , doi:10.1051/0004 - 6361/201628714 ; arxiv:1609.04303 magic , z. , collet , r. , asplund , m. , trampedach , r. , hayek , w. , chiavassa , a. , stein , r. f. , nordlund , 2013 , 557 , 26 mashonkina , l. , korn , a. j. , przybilla , n. 2007 , , 461 , 261 mashonkina , l. 2014 , in `` setting the scene for gaia and lamost '' , proceedings of the international astronomical union , iau symposium , volume 298 , pp .
355 - 365 mcwilliam , a. & rich , r. m. , 1994 , 91 , 749 mcwilliam , a. , preston , g. w. , sneden , c. , searle , l. 1995 , 109 , 2757 mcwilliam , a. , araa , 35 , 503 mcwilliam , a. , matteucci , f. , ballero , s. , rich , r. m. , fulbright , j. p. , cascutti , g. 2008 , 136 , 367 mermilliod , j.c .
1986 , catalogue of eggen s ubv data , 0 michalik , d. , lindengren , l. & hobbs , d. 2015 , 574 , 115 moore , c. e. , minnaert , m. g. , & houtgast , j. 1966 , the solar spectrum , 2935 to 8770 (nbsmonog .
61 , washington , dc : nbs ) munari , u. & zwitter , t. 1997 , , 318 , 269 nissen , p.e . , & schuster , w.j . , 1997 , 326 , 751 nissen , p.e . , & schuster , w.j . ,
2010 , 511 , l10 perryman , m.a.c . , &
esa 1997 , the hipparcos and tycho catalogues .
astrometric and photometric star catalogues derived from the esa hipparcos space astrometry mission , esa sp , 1200 petigura , e. a. & marcy , g. w. 2011 , , 735 , 41 pont , f. , mayor , m. , turon , c. , & vandenberg , d. a. 1998 , , 329 , 87 ramirez , i. & melendez , j. 2005 , , 626 , 465 reid , i. n. 1997 , , 114 , 161 roederer , i. u. , preston , g. w. , thompson , i. b. , et al .
2014 , , 147 , 136 ryan , s. g. , & deliyannis , c. p. 1998 , 500 , 398 sandage , a. 1964 , , 139 , 442 sitnova , t.m . ,
mashonkina , l.i . , & ryabchikova , t.a .
2013 , astron .
lett . , 39 , 126 .
sitnova , t. , mashonkina , l. , chen , y. , et al .
2015 , , 808 , 148 sneden , c. 1973 , 184 , 839 sneden , c. 2004 , , 75 , 267 sobeck , j. s. , ivans , i. i. , simmerer , j. a. , sneden , c. , hoeflich , p. , fulbright , j. p. , & kraft , r. p , 2006 , 131 , 2949 sousa , s. g. , santos , n. c. , israelian , g. , lovis , c. , mayor , m. , silva , p. b. , udry , s. 2011 , 526 , a99 struve , o. , & elvey , c. t. 1934 , 79 , 409 wallerstein , g. 1962 , 6 , 407 winkler , h. 1997 , 287 , 481 valenti , j. a. , fischer , d. a. 2005 , 159 , 141 vandenberg , d. a. & clem , j. l. 2003 , , 126 , 778 van leeuwen , f. 2007 , hipparcos , the new reduction of the raw data , astrophys .
space sci . lib .
( springer ) , 350 zhao , g. , & gehren , t. 2000 , , 362 , 1077 | we have performed a differential , line - by - line , chemical abundance analysis , ultimately relative to the sun , of nine very metal - poor main sequence halo stars , near [ fe / h]=@xmath02dex .
our abundances range from @xmath1}\leq-1.40$]dex with conservative uncertainties of 0.07dex .
we find an average [ @xmath2/fe]@xmath3dex , typical of the milky way . while our spectroscopic atmosphere parameters provide good agreement with hst parallaxes , there is significant disagreement with temperature and gravity parameters indicated by observed colors and theoretical isochrones .
although a systematic underestimate of the stellar temperature by a few hundred degrees could explain this difference , it is not supported by current effective temperature studies and would create large uncertainties in the abundance determinations .
both 1d and @xmath43d@xmath5 hydrodynamical models combined with separate 1d non - lte effects do not yet account for the atmospheres of real metal - poor ms stars , but a fully 3d non - lte treatment may be able to explain the ionization imbalance found in this work . |
in the past , insurance companies did not have a large amount of information about their customers so everyone would pay similar prices , despite the large variations in driving habits .
recently , technological advances have created ways to observe human driving behavior .
taking advantage of this technology , some us - based insurance companies @xcite now offer consumers vehicle insurance policies with the option of installing a tracking device into their vehicle .
this allows the insurance companies to charge less to safe drivers , in an attempt to attract those people as customers .
however , unlike a truck driver who is being monitored while doing a job , these consumers will be monitored during all of their daily activities , public and private .
although some of these monitoring devices are based upon gps information and offer no privacy protections , ( such as onstar @xcite ) , many other devices and insurance programs are advertised as being privacy - preserving .
for instance , as an alternative to gps data some of these devices log speed information instead . in this paper , we demonstrate that logging this timestamped speed data is also not privacy - preserving , despite the insurance companies claims .
the privacy problem introduced in this paper is significant because the data collection by insurance companies is an `` always on '' activity .
data may be logged forever , thus , deducing location data from speed traces represents a huge breach of privacy for drivers using these programs . even if insurance companies are not currently obtaining location traces from their data today , this does not guarantee it will not present problems in the future . because the data is not considered location data it is probably not being treated as such , and may eventually be obtained by any number of antagonists that do know how to process the data traces to obtain location information .
for example , law enforcement agencies may request the information as they have done with other data sets , e.g. electronic toll records @xcite . breaching user privacy with just a starting point and
speed information is a difficult task , otherwise insurance companies would not claim that the information being gathered is privacy preserving .
indeed , matching a single speed trace to all of the roads in a country or state is extremely difficult , but having the home address of a person , as the insurance companies do , provides a starting location for some paths simplifying the problem .
number of path choices within a given distance of a starting location . the pattern of growth on a log - linear scale shows that the number of paths increases exponentially , making exploration of all possible paths infeasible .
within one mile there are over @xmath0 paths diverging from a grocery store and over @xmath1 paths from a residential area .
locations central to major transportation routes , like a grocery store , will see a higher increase in paths compared to a residential location .
residential roads are also more likely to dead - end : in this example the number of paths from the residence actually falls at the beginning because only one path does not dead - end immediately . ] reproducing an exact driving path from just speed data and a starting location is challenging .
speed data does not indicate if a person is turning at an intersection or merely stopping at a stop sign or red light , so multiple alternative pathways exist that match some of the speed data
. of course multiple routes can be explored , but even within just a few minute s drive of a person s home in our data set there may be thousands of unique paths that the person could have taken , so exploring every single possible path is infeasible .
the growth of possible paths within a distance of just one mile of a starting location is illustrated by figure [ fig : path_growth ] .
there may only be a few tens or hundreds of roads within a mile of a starting location , but there are many unique ways to travel along those roads . at one mile , there are over 100,000 possible paths the driver could have taken when the trip starts from a grocery store next to several highways .
in addition to the explosion of possible paths the driver could have taken , the problem is made hard by the fact that speed data is ambiguous ; that is to say it does not uniquely match just a single path .
for example , the speed of a vehicle making a right turn can also match the speed of a vehicle making a left , or the speed of a vehicle that decelerates because of a pot hole .
there is no certain way to tell which way a car goes at an intersection with just speed data , meaning that a segment of speed data may match multiple paths .
since there are so many possible paths and speed data is ambiguous we must score paths according to some error metric and seek out the final path with the least error .
however , finding the optimal path from an exponentially growing search space is not easy .
simple solutions , such as greedily following a single path , may end in failure without even finding any solution ; for instance when the path chosen by a greedy algorithm suddenly dead ends . later in the paper
we discuss several other approaches that do not work .
this paper makes two major contributions . 1 .
we present a novel algorithm _ elastic pathing _ to extract location traces from speedometer data and starting locations .
applying this algorithm and to real - world traces , we show how the data collected by many insurance companies is not privacy preserving despite their claims .
_ to best of our knowledge , our work is the first to demonstrate how people s locations can be discovered based only on timestamped speed data and knowledge of a single starting location . _ we will also show that even without predicting path endpoints with 100% accuracy long - term data collection allows an antagonist to eventually identify physical regions that are frequently visited .
the combination of our algorithm with analysis of long - term driving traces can obtain private information from what is currently considered privacy - preserving data . as minor contributions ,
we have collected real - world traces over several months that serve as a testing set for our approach , we offer suggestions for future work that may improve path prediction further , and provide a discussion of privacy - preserving alternatives to the collection of speedometer data that do not require cryptography to implement .
in several countries , including countries in north america , europe and japan , some insurance companies @xcite have introduced `` usage - based '' insurance programs . in addition to basing insurance premiums on the characteristics of the car and history and knowledge about the drivers , usage - based insurance programs also take into account how and when the car is driven .
the incentive to participate by insurance policy holders is a potential discount to their premiums for good driving behavior .
in addition to providing potential premium discounts , some companies also provide their customers statistical data , such as overviews of driving habits and information about how much they drive per day or per week . in the past
most of these programs used to track drivers locations with gps @xcite , but currently there are several programs that claim to be privacy preserving because they use other kinds of data . in this paper
, we focus on a more recent type of a program , which is enabled by collecting speedometer among other data directly from the car .
the idea is that by not collecting gps data and instead collecting other driving information , such as speedometer data , these approaches are more privacy - preserving since the insurance companies are not collecting drivers locations .
there are several insurance companies in the us that use this method .
for example , the snapshot device provided by progressive and the drivewise device provided by allstate will record vehicle speed data while not collecting gps locations @xcite . the data in these programs is collected by connecting a device ( provided by the insurance company ) to the car diagnostic port obd - ii @xcite . since 1996 , obd - ii has been made mandatory for all cars sold in the united states @xcite .
the european union has also made the european equivalent eobd @xcite mandatory for both petrol and diesel engined cars in europe .
this standard interface allows for collecting comprehensive amounts of diagnostic information .
the data that is collected varies with each insurance company , but in the cases discussed above , it includes time and speedometer data .
the devices will regularly upload this data to the insurance company s servers via a mobile data link .
the driving habits that some insurance companies state that they are interested in include `` how often you make hard brakes , how many miles you drive each day and how often you drive between midnight and 4 a.m. '' there is also a very important subtlety here . for example , progressive claims they do not collect the data about when people are speeding ( which some people in the public interpret as meaning that speed data is not collected at all ) .
instead , they _ do _ collect speed data , but it follows from the logic that since they do not know your location , they do not know if are you speeding or not . for our analysis of the privacy concerns of these programs we need to know the sampling rate of the data collection devices , but the companies do not disclose on their web sites how often do they sample e.g. speedometer data .
however , e.g. allstate @xcite includes the following information `` hard braking events are recorded when your vehicle decelerates more than 8 mph in one second ( 11.7 @xmath2 ) [ and ] extreme braking events are recorded when your vehicle decelerates more than 10 mph in one second ( 14.6 @xmath2 ) . ''
we also know that companies will sample data much faster than they need to . in the united states federal vehicular safety regulations @xcite mandate that vehicles traveling at 20 mph must be able to brake to a full stop in 20 feet at a deceleration rate of 21 feet / second .
since we know what kind of events the companies are interested in and the maximum performance of vehicles on the road , we can simply estimate the bounds on sampling rates based upon the amount of information required to detect these events .
the events that are to be detected occur over a time interval of just one second ( @xmath3 ) . using the nyquist sampling theorem , we know that we must sample at twice this rate to detect these features , a rate of two samples per second .
this is a lower bound on any feature detection , and companies may sample at significantly higher rates to detect other features .
consider that a driver that tailgates very closely will often tap their own brakes for very short time durations . if insurance companies wished to detect these events they would need to sample at a rate that was twice the speed of these brief taps . to summarize the privacy issues ,
the insurance companies using speedometer based measurements are claiming that they can not find people s locations based on their collected data .
our work shows this is not true
. we will show in further sections that a lot private location information is indeed obtainable , contrary to the claim of 100% privacy protection .
further , by finding out correct paths on the roads , the insurance companies could also deduce subtle issues such as were the drivers speeding , despite their claims to the contrary .
we will show that our elastic pathing algorithm is able to do more than just identify the endpoint of a trip it also eliminates a large number of locations where the trip could not have ended .
this is also a serious privacy concern because it allows to identify changes in regular routines very easily .
for example , if a person usually drives for groceries after work but the predicted path goes in a different direction , then it is highly likely the driver is breaking their routine .
indeed , previous research has shown that location ( derived from gps data ) can be automatically analyzed to produce profiles of driver s behavior , social activities and work activities @xcite . a comprehensive discussion on the the consequences of breaches to location privacy are given by blumberg and eckersley @xcite .
they discuss how location databases can be used by others to ask questions such as `` did you go to an anti - war rally on tuesday ? '' , `` did you see an aids counselor ? '' , `` have you been checking into a motel at lunchtimes ? '' , `` why was your secretary with you ? '' , or `` which church do you attend ?
which mosque ? which gay bars ? ''
a lot of work has concentrated on anonymizing or obfuscating @xcite location traces because location traces contain a great deal of behavioral information that people consider private @xcite .
krumm @xcite has written an overview of computational location privacy techniques , and zang and bolot @xcite have recently questioned the possibility of releasing privacy - preserving cell phone records while still maintaining research utility in those records . accordingly , predicting future mobility patterns and paths from human mobility traces is a well - explored topic , including the following results .
based on analysis of location data , such as mobile phone cell tower locations , researchers have shown 1 ) the nature of individual mobility patterns is bounded , and that humans visit only a few locations most of the time ( e.g. , just two @xcite ) ; 2 ) that there is high level of predictability for future and current locations ( e.g. @xcite most trivially , `` 70% of the time the most visited location coincides with the user s actual location '' @xcite ) ; and 3 ) that mobility patterns can also be used to predict social network ties ( e.g , @xcite ) . in a more applied domain , gps traces have been used to learn transportation modes ( e.g. walking , in a bus , or driving ) @xcite , for predicting turns , route , and destination @xcite , for predicting family routines ( e.g. picking up or dropping of children to activities ) @xcite , probabilistic models describing when people are home or away @xcite , and recommending friends and locations @xcite . however
, none of the above work can be used to discover locations based solely on speedometer measurements and a user s starting location .
related work in the field of dead reckoning does suggest that speedometer traces should have some level of information that can be extracted .
dead reckoning works by using speed or movement data and a known location to deduce a movement path .
dead reckoning has typically been used for map building with mobile robots @xcite or as an addition to global navigation satellite systems ( gnss ) such as gps @xcite . when supplementing gnss data , odometer data for dead reckoning might only be used when gnss information is unavailable , such as when a vehicle passes through a tunnel or an area with many tall buildings . however , this kind of dead reckoning can not work without frequent location ground truths as there is no perfect method to match speed data to turns in a road .
map building with robots is interesting because there are often no exact ground truths , only location estimates .
golfarelli et al .
@xcite describe a method that can assemble maps from arbitrary distance estimates to and from landmarks .
the problem of deducing a traveled path from only speed data and a starting point is , in some ways , the reverse of this map building problem : we already have the map , but have difficulty identifying landmarks ( in this case turns ) from just speed data .
ubiquitous computing devices help people during their daily lives in different contexts , for instance by monitoring our activity and fitness levels or interesting events social that we have attended .
however , the ubiquitous nature of these monitoring devices has security and privacy ramifications that are not always considered .
for example , a lack of sufficient protection in the nike+ipod sport kit allowed other people to wirelessly monitor and track users during their daily activities @xcite .
although these kinds of devices clearly appeal to consumers , the potential loss of privacy is a growing concern . in the specific area of usage - based insurance ,
the privacy concerns of these insurance programs have been studied before by troncoso et al .
however , this work dates back to schemes which would send raw location ( gps ) coordinates to either insurance providers or brokers , and troncoso et al .
proposed a cryptographic scheme pripayd @xcite to address the problem .
our work shows that speedometer - based solutions , which were not considered by troncoso et al .
, are not privacy - preserving either .
finally , most closely related to our work are side - channel attacks that use accelerometer data from smartphones @xcite .
projects such as accomplice @xcite and autowitness @xcite have used accelerometers and gyroscopes for localization of drivers .
however , the information from the smartphone can be used to actually detect when turns occur .
in contrast , we have only a time series of speed data available .
while this speed data might indicate when a vehicle stops or slows down at an intersection , unlike accelerometer data , it does not indicate if any turn is taken .
to test the hypothesis that a driving route can be reconstructed from a known starting point and a speed trace we needed to collect a large body of testing data .
we require both a ground truth of gps data and a sample set of speed data .
this data can be collected with the approach that the insurance companies use , by connecting a device to the onboard diagnostics standard obd - ii connector .
devices connected to the car this way will receive accurate information about the car s speed directly from the vehicle s computer and can pair these speed values with gps information to provide an accurate ground truth . to log this information we used a cell phone with gps and a bluetooth enabled odb - ii diagnostic device .
we have also logged a much higher volume of gps - only data , from which we can reconstruct the speed trace . using the speedometer data from the odb - ii device
is straightforward because raw speed data is immediately available .
however , collecting data using the odb - ii device required obtaining the devices and was more cumbersome ; in the rush of daily activities it was easier to use just gps data gathered from a smartphone .
this data requires a small amount of processing to obtain speed values from latitude / longitude pairs .
this was a two step process , as outlined below .
first , we used the haversine formula ( see appendix ) to approximate distances from the raw gps coordinates .
next we simply divide this value by the time interval ( in fractions of an hour ) to get an instantaneous speed value for each time interval . each speed value
is then given a time stamp and converted into the same format as the speedometer traces .
we note that we did not see differences in the accuracy of our algorithm between the two approaches of data collection .
we had seven volunteers collecting data .
five volunteers collected gps - only data over the period of several months , and two volunteers gathered both gps and speedometer data with an odb - ii device over a period of one month .
there were 240 data traces that we used for pathing , totaling nearly 1200 miles ( nearly 2000 km ) of driving data collected , with a median trip length of 4.65 miles ( 7.5 km ) , minimum trip distance of 0.38 miles ( 0.62 km ) , and maximum trip length of 9.96 miles ( 16.0 km ) .
these traces comprise 46 unique destinations , with more than half of these destinations visited multiple times by individual drivers .
we believe this represents a diverse testing set ; if our algorithm works well on this data , with all of its variety ( e.g. the differences between driver s habits ) , then it should perform similarly in many other driving scenarios .
in addition to the differences between driving habits of participating drivers , there are also differences in their vehicles
. examples of vehicles we used in the study are small sedans , sports utility vehicles , and a pickup truck .
we feel that this represents a good sample of driving habits and vehicle types , so a pathing algorithm that works properly on all of these traces will work correctly across the driving public as a whole . additionally , the insurance companies know the make and model of the vehicle they are insuring and might be able to use additional information , such as the turning radius of particular vehicles , that we do not use
. map data was collected from the openstreetmap ( osm ) project @xcite using their open rest api .
we parsed this data into node adjacency lists , turn restrictions , and other road characteristics ( e.g. speed , directionality ) and stored the data in a local sql database for rapid querying .
the open street map data is divisible into two main types : _ nodes _ and _ ways_. nodes are numbered latitude / longitude pairs .
they may document significant map features like bends in roads or intersections .
nodes may consist of multiple attributes but they always include a latitude , longitude and a unique node i d .
ways are collections of node ids and some context tags that identify features of the road that the group of nodes describe .
when ways cross the nodes at the intersection belong to both ways , indicating that a turn is possible .
turn restrictions are indicated by the way - ness of the road ( one - way or two - way ) and additional meta - information , such as `` no u - turn '' or `` only right turn '' .
in this section , we describe the algorithm we use to recreate a person s driving path given data trace with speed / time tuples .
our approach relies upon two features of vehicular travel .
first , there are physical limitations to a vehicle s turning radius at high speeds , so a vehicle that travels a particular path must travel at speed at or below the maximum speed possible .
second , people will only stop when they need to , e.g. at traffic lights and stop signs , until they reach their final destination .
this second assumption is not always true ; a person may need to stop because of actions of other vehicles , road construction , accidents , etc .
however , when we continually observe a driver only a small number of traces will have extraordinary events .
we use data from open street maps ( osm ) to identify possible routes and locations of turns and we find the route that best fits the given speed data .
osm data is in the form of nodes that are associated with different roads so a path will be a sequence of nodes .
each node has a physical location , given by a latitude and longitude , allowing us to determine distances between segments of road and the turning angle of a path of nodes .
each sample from the data trace corresponds to some amount of distance traveled from one node to the next along a path .
the osm data provides adjacency data for each node , so once a node is reached , we must choose which node to begin moving towards .
for instance , a four - way intersection will have a single node at the intersection and four adjacent nodes .
since we reach the intersection from one direction we have three possible next nodes from which to choose .
any pathing algorithm must follow some basic steps .
first , as discussed previously and illustrated in figure [ fig : path_growth ] , the growth of paths is too great to explore all possible paths .
therefore , any algorithm must have some methods of comparing paths and choosing `` better '' paths that more closely match the given speed trace .
this also means that the algorithm must keep a list of possible paths , while also either limiting growth of the number of paths or removing paths as they become too plentiful .
another problem to overcome is that no path perfectly matches the speed data , because drivers swerve around objects in the road , or take turns more widely or sharply than we expect .
increasingly the estimation of a vehicle s progress along a path will become incorrect and mismatches might make even the correct path seem impossible .
for example , if the progress along a segment of road is ahead of the vehicle s actual position then the vehicle might be going too quickly to make a turn , but if we somehow corrected the distance traveled to account for some of these distance estimation errors then we would find that the speed traces lines up perfectly with the turn .
thus , any algorithm must also correct paths as it explores them , trying to take into account these variations in the travel distance
. one could think of many ways to attempt this pathing , and we thought of and explored several approaches that do not work before finding an approach that gave surprisingly good results to be harmful for people s privacy . before describing our novel approach
, we will first document the failed approaches to provide more background information to the community .
if deducing a driving route from speed data and a starting location was easy then insurance companies would not claim that their data collection was privacy preserving .
although such path reconstruction may seem simple in concept , it is actually very difficult .
the first major difficulty is detecting turns with only speed data .
when someone breaks and then accelerates , did they make a turn or did they slow down because the person in front of them was turning ? if a person comes to a complete stop and then accelerates , did they go straight or turn
? when a person stops at an intersection are they the car closest to the stop light , or are they behind ten other cars ?
the speed data alone does not give enough information , and a single incorrectly identified turn will result in an error possibly as large as the entire trip distance .
even if a turn is identified from the speed data , if the predicted position of the car is off by a few meters then the wrong intersecting road may be chosen as the destination of a turn . in principle , turns could be predicted probabilistically based upon past driver behavior @xcite . however , we assume that we do not already know information about an individual driver , so we can not build a model of their specific driving
. we could consider using a model developed using many other drivers in the same area @xcite , but that has the disadvantage of predicting the most common turns and paths rather than the unique path of an individual .
estimating that a person probably shops at a supermarket near them or that they probably work at the largest employer in their town is not necessarily a considerable breach of their privacy .
instead , we identify specific locations that a particular individual visits , and are especially interested in those places that are unique to the individual and not common to many people .
since the heart of this pathing problem is the lack of turn information one approach is to treat this as a classification problem . here
, we examine the speed traces to extract features : stop , right turn , left turn , or straight road .
if classification would work well , we can assign probabilities to different choices at each intersection and eventually find the highest probability path . however , road conditions and driving habits are very diverse . with our data
set we were unable to classify turn features with high enough accuracy to build correct paths .
the sheer volume of paths to explore meant that unless the correct path was extremely different from all of the incorrect paths , the algorithm would find some incorrect paths that looked just as good , or better , than the correct path . another approach we tried was to switch from a probabilistic model to a model where we identified time windows where turns could occur and did not follow any paths that would make turns outside of these windows .
we expected this to reduce the number of paths we needed to explore , perhaps leaving just a few paths at completion . however , because of the accumulation of distance estimation errors , turns in the speed data do not align perfectly with the map data so we need to widen the time intervals where turns were allowed . as the time windows widen the number of accepted paths also increased , until we had too many possible paths to distinguish a single best fit .
we also tried a dynamic programming approach , keeping track of the score of the best possible path at each node of the ways in the osm data . however , there is a subtle flaw in this approach .
whatever error metric is used for each path , that error will slowly increase as the path grows in size due to the slight discrepancies in estimated and real travel distance .
if there are two possible paths to reach a node then this dynamic programming approach will tend to favor the shorter path that reaches that node .
however , if we explored the path further we might find that the longer path with a worse score at the previous node better matches the road going forward but with this dynamic programming approach we would have already dropped that path because it had the worse score .
this error would be unrecoverable .
the approach that works must keep track of the best path in terms of time ; if two paths have the same error score but one path has traveled through more of the speed samples then that path is better than the other ( up to that point ) because it is accumulating less error per unit time .
next , we describe our successful approach we call
_ elastic pathing_. this algorithm is based around the observation that as we attempt to match the speed data to a path we stretch or compress the predicted distance moved .
however , after reconciling differences between a section of road and the speed trace we must `` pin '' the path at that point ( which we call a landmark ) because any movement would break apart the newly reconciled speed / time trace and distance data . for instance , if the speed data goes to 0 indicating a stop where there is no intersection we might pull the path forward by some distance to reach an intersection .
after this has been done the path from this point and earlier can not be moved since it aligns with a feature in the road .
we call this approach _ elastic pathing _ because of the stretching and compressing of the speed trace to fit the road is conceptually similar to stretching a piece of elastic along a path while pinning it into place at different points . to help understanding the algorithm
, we introduce a small set of definitions .
* _ calculated distance _ the distance a vehicle traveled calculated from the speed and time values in the speed trace . * _ predicted distance _ the distance along a possible route on the road at a certain time in the speed trace . * _ error _ the difference between calculated distance and predicted distance of a possible route . * _ feature _ a vehicle stop in the speed trace or an intersection in the road . * _ landmark _ a place where the speed trace and road data match , but would become unmatched if any stretching or compression of the predicted distance moved occurs .
this includes the previously mentioned feature , but also any other place where a mismatch between speed and road data can occur .
we can compute the fit of a path by the amount of stretching and compressing that needed to be done to make the speed data match the path . the more stretching and compressing along a path the worse its score . at each iteration of the algorithm
we sort all of the partial paths by their current scores and then explore the path with the best score . that path advances until it reaches a feature that requires a `` pin '' operation .
at this point there may be multiple ways to advance the path so several new paths may be created .
each new path s score is adjusted to reflect the stretching or compression of the distance traveled from the last pinned landmark and the algorithm proceeds to the next iteration .
the algorithm follows with the best score at each iteration , so the first path to finish can not be worse than any other path .
the pseudo - code for this algorithm is given in algorithm [ alg : pathing ] .
the most complicated operation done in the _ elasticpathing _ algorithm is _ gotobranch _ function , whose pseudo code appears in algorithm [ alg : gotobranch ] .
the _ gotobranch _
function advances a path until it comes to a feature on the map and returns every possible branching from that feature .
there are two features : an intersection in the road of the path and a zero speed section of the speed samples . when a path reaches an intersection it explores every possible direction .
if the path is going at a speed where it can move along the curve in the road then a landmark is set at that location with the _ pin _ function .
if the curve in the road is too great for the current speed then the distance moved from the last landmark is compressed with the _ compressa _ function and stretched with the _ stretcha _ function .
thus , when there is a mismatch between the speed samples and road data there are two ways to resolve it and thus two new possible paths .
a similar situation occurs when the speed data indicates that the vehicle has come to a stop .
if the path is already at an intersection then the landmark is set with the _ pin _ function .
otherwise the two solutions are found with the _ compressb _ and _ stretchb _ functions . rather than compressing or stretching the speed trace as in the previous compress and stretch functions ,
these functions compress or stretch the route data from the last landmark until an intersection is found to match the zero speed segment .
after any landmarks are set , the ratio of stretching or compression from the last landmark increases the error of each path .
this means that a preexisting path may have less error than the current paths and should be explored instead , so all possible branches are returned to the elastic pathing function , the paths are placed in sorted order by their error , and pathing continues with the new best path .
there are several details that we do not show in the pseudo - code .
for instance , when we check if we are at an intersection we allow space for the number of lanes in the intersecting road and offset for another car in front of our vehicle . determining the maximum speed of a turn
is done by assuming the maximum allowed incline in the road ( 7% ) @xcite , the turn radius of the road based upon the number of lanes in the road and typical lane widths @xcite , and the turn angle equation @xcite .
the best possible case , a path from a shopping store to a home with no error .
the ground truth is indicated with a solid line while the nodes along the predicted are shown as dots . in this case
the predicted and actual paths match perfectly . ]
we generated results for 240 traces comprising 46 unique destinations of 7 different individuals .
of these destinations , 28 were visited more than once and 10 locations were visited more than 5 times .
the median trip length was 4.65 miles ( 7.5 km ) with a minimum trip distance of 0.38 miles ( 0.62 km ) and maximum trip length of 9.96 miles ( 16.0 km ) .
the 2009 national household travel survey nhts found the average drive distance is 9.62 miles on weekdays and 10.03 miles on weekends @xcite , so our drivers had short , but not unreasonable commutes .
traces were taken at many different times of day and in varying road conditions over several months , and we believe that these traces are representative of common driving behavior . a ruby implementation of the elastic - pathing algorithm processed all of the traces in under 30 minutes on a 2-core 2.2ghz machine , with an average running time of less than ten seconds per trace .
even without an implementation in a high - performance language , the algorithm is able to process traces far faster than they are generated . since data analysis
could essentially be done in real - time as a vehicle s speed trace is being collected , the limiting factor on the time delay between data recording and path prediction is probably the time overhead of collecting , transmitting , and preprocessing ( e.g. matching a trace to a specific driver and their known locations ) speed traces .
this means that it may be entirely feasible to do near real - time tracking of drivers with just speed data , although the tracking will not be 100% accurate .
the cdf of the distance from the pathing algorithm s endpoint compared to the ground truth s endpoint . in 10% of traces
the algorithm found an endpoint less than 250 meters from its actual location and 30% of the traces were within 800 meters of the actual end point . with thousands of driving traces gathered per individual per year this gives an insurance company a very detailed view of an individual s lifestyle and habits .
results did vary with the individual ; the top 30% of one driver s traces were all within 500 meters while that same percentage were only within 1300 meters for another driver . ]
the cdf of errors is shown in figure [ fig : cdf ] . in 10% of the traces ,
the pathing algorithm found an endpoint that was within 250 meters of the actual endpoint and 30% of the traces were within 800 meters of the actual endpoint .
this distance is easily accurate enough to tell when a person goes home , what areas a person drives through , and can identify many destinations of a driver . since the size of a parking lot in a movie theater or shopping center is comparable to these error values , this level of detail is enough to gain an in - depth view of an individual s lifestyle and habits .
when we consider that individuals often take multiple trips to the same locations it is clear that , over time , every location that is commonly visited will be identified . the average error of endpoints relative to trip distance .
this figure shows that there is no noticeable decrease in the performance of the elastic pathing algorithm with trip distance .
manual analysis of our algorithm s performance shows that prediction accuracy is more strongly influenced by the route taken and traffic encountered than by the trip distance .
the expected error from guessing is simply the distance from the starting point to a random location along the circumference of a circle centered at the starting point and with radius equal to the trip distance . ]
the relative inaccuracy of our approach does not go up with trip distance , as shown in figure [ fig : relative_error ] .
the percent of predicted endpoints within 250 meters of the actual endpoint also does not decrease with distance in our data set , with trips as long as 10.5 miles still having endpoints correctly predicted to within 250 meters .
our approach clearly does much better than random guessing and almost always correctly identifies at least the general direction of travel .
the direction of travel is not a very serious privacy concern though , so we will use another filtering step to screen out `` noise '' from incorrectly predicted points .
we can further take advantage of multiple available traces for each driver to screen out points that may be noise .
we note that erroneous predictions are unlikely to repeatedly find the same wrong location and instead usually make mistakes in different locations .
thus , we can use the frequency that locations are predicted as a way to remove poor predictions and focus only on common destinations , as shown in figure [ fig : endpoint_circles ] .
the first row of the figure , figure [ fig : endpoint_circles](a , b ) , shows predicted endpoints while the second row , figure [ fig : endpoint_circles](c , d ) , shows ground truths .
the first column , figure [ fig : endpoint_circles](a , c ) , shows all endpoints , while the second column , figure [ fig : endpoint_circles](b , d ) , shows only the endpoints that occurred with high frequency in the traces .
although the initial endpoint predictions ( figure [ fig : endpoint_circles]a ) are very noisy , filtering the endpoints by frequency in figure [ fig : endpoint_circles]b creates a collection of endpoints that is fairly close to the ground truth in figure [ fig : endpoint_circles]d .
we also noted that some individuals were more easily tracked than others .
although the top 10% of traces for traces from all individuals had a maximum error of 250 meters , for one individual the top 10% had a maximum error error of only 164 meters , and the top 30% of those individual s traces still had errors of less than 500 meters .
these individuals actually commuted along the same major road and lived in similar areas , so the differing performance of the pathing algorithm on these two individuals is likely to stem from differences in their personal driving characteristics and the driving characteristics of their vehicles than from the areas where they drove .
the practical meaning of this is that learning the common locations and habits of some individuals might be much easier than for others .
the pathing algorithm is able to do more than just identify the endpoint of a trip it also eliminates a large number of locations where the trip could not have ended .
this is also a serious privacy concern because it allows an antagonist to identify changes in regular routines very easily .
for example , if a person usually goes home after work but the predicted path goes in a different direction then it is highly likely the person is breaking their routine .
when the algorithm has enough information and a route is very unique for instance the spacing between intersections on the road travelled does not match the spacing in any other road the algorithm can do very well .
this is illustrated in figure [ fig : best_case ] .
however , this case is not normal , and at the very end when the algorithm chooses the correct left turn it may just as well have chosen to turn right or gone straight . however , even if the algorithm made the wrong choice it would still be close to 100 meters from the actual destination .
* we believe the probability of error is closely tied to a few factors : * * homogeneity of roads .
if every road seems the same ( same speed , similar intersection intervals , etc ) , especially as in areas built in a grid , then they can not be distinguished . * * traces with only slow speeds . without high speeds to rule out turns and constrain paths to a few major roads the correct path is indistinguishable from any other . * * unpredictable stops from traffic , construction , etc .
the largest barrier to improved performance is distinguishing one road from another .
if two roads have similar features ( e.g. same speed , similarly spaced intersections ) then there is little way for the algorithm to distinguish between the two .
this is also affected by the traffic pattern of the trip ; if a vehicle stops at every intersection because of red lights then there is a great deal of information about the spacing of those intersections .
if , however , the vehicle stops at no lights then there is no information about the spacing of intersection on the current road .
with more prior information we may be able to do better for instance , if we knew the average waiting time at different lights then the waiting time at an intersection might distinguish one road from another even if a vehicle only stops at one or two lights
. this additional information , such as the average wait time at certain lights , or the average traffic speeds of different roads during different times of day , may already be available to insurance companies and this information is not difficult to gather .
if the algorithm had this information available it could score different turning choices with higher accuracy , leading to improved results .
the implications of our results are that , given just timestamped speed traces and a starting location , the privacy of a driver will be eroded as more of their speed traces are collected .
this is even though our approach does not yet predict traces and trip endpoints with 100% accuracy .
driving habits are often regular and locations are visited multiple times , so an antagonist gets many chances to correctly identify traveled paths . as a person
s driving traces are collected and analyzed , a set of possible destinations will be built for that person .
a set of thousands of possible destination will be reduced to a handful .
incorrect endpoint predictions are noisy and unlikely to repeat , so over time the endpoints that are predicted the most are increasingly likely to be correct predictions of a person s common destinations .
thus , we believe that the large body of data already collected and still being collected would allow an antagonist to deduce many details about a person s regular routine and irregular events . starting from a person s home address
an antagonist could identify work locations , commonly visited stores , locations associated with regular hobbies , etc .
additionally , any breaks in routines would be very obvious because the details that could be obtained from a speed trace , such as distance traveled , duration of trip , and predicted endpoint of the trip , would all change .
it is possible for companies to adjust their sampling rates to reduce the potential for privacy breaches , however , there is no guarantee that someone could not improve upon our algorithm and obtain results better than ours even with a lower sampling rate .
perhaps the best way to improve the privacy protection of these systems is to do data processing in the car before sending it to a remote server , so that only the end results are ever transmitted and stored .
although it is tempting for a company to capture as much information as possible in case that data becomes useful later , this also presents unnecessary privacy risks since . in the case of auto insurance companies
, a speed trace is not vital to this system , as the real goal is the detection of unsafe driving habits , which can be detected with a more privacy preserving set of driving features .
there are alternatives to collecting speedometer readings which should have much better privacy protection .
in fact , some usage - based insurance programs are already offering more privacy - preserving products .
companies such as gmac insurance , milemeter and corona direct offer programs that only require the mileage information from vehicles @xcite .
payd coverage provided by hollard insurance company collects mileage information and driver style information @xcite .
companies like paygo systems collect driving zone information instead of fine - grained location , besides basic information including mileage and minutes of use @xcite .
the best practice for such monitoring systems is to only record relevant data .
for example , devices could check speed data but only send a notification when an even of interest was detected . in general , this approach could be applied to other data collection methods , including gps .
an example path where the algorithm correctly follows most of a path but predicts the wrong turn when the speed trace runs out of features . ]
our main limitation is the difficulty distinguishing between similar roads . when a vehicle stops at an intersection and then accelerates slowly a turn or straight path are both possible , so our algorithm relies upon later features in the speed trace to determine which of the possible routes was correct .
at the ends of paths there are simply too few features to make this distinction . at the beginning of paths
this may simply cause there to be too many possible paths to sort through .
these problems are exacerbated in residential areas , where repeated , similar intersections are common .
we notice that travel through residential areas tends to occur at either ends of paths rather than during the middle , because traveling through a residential area is unlikely to be the fastest route to a destination .
fixing mistakes in residential areas by incorporating additional knowledge about the area may greatly improve our algorithm s success since we had little difficulty reconstructing paths along major roads , but tend to make mistakes once the path leaves major roads .
an example of this is in figure [ fig : follow_highway ] .
there are two ways these errors could be addressed .
first , the street addresses of customers are available so the algorithm could check for paths to a person s known home address at the end of predicted paths .
second , at the beginnings of paths drivers move from local roads to larger arteries , so the algorithm could also attempt to simply take the shortest paths to major roads if there are too many ambiguous turn choices at the beginning of a path . when accepted for publication , we will open source our work , which contains tools to get the needed map data from open street maps , and implementations of our algorithms .
we will also include a subset of the collected traces that has been explicitly authorized to be made publicly available on crawdad or similar archival sites .
we have demonstrated an attack against user location privacy that uses speed data from driving traces and an initial starting location .
this problem seems simple in concept , but is difficult in execution largely because turn detection with speed data is often ambiguous .
for instance , if someone breaks and then accelerates they could have turned or they could have slowed down because the person in front of them was turning .
we have presented the design and implementation of a novel elastic pathing algorithm and used a large corpus of real - world traces to show its accuracy .
our trials showed that the algorithm is fast , and could also be used for real - time tracking of people just based on their speed data .
our results show that with a large corpus of data an entity can errode individuals privacy and gain information about typical routes and destinations . in 20% of individual traces
we predicted the correct trip end point to within 500 meters .
this means that a location visited daily can be identified in about a week ; locations visited on a weekly basis could be identified with slightly more than a month of data .
variations in daily routines are easier to identify than endpoints .
even when the algorithm does not correctly find the proper end point it does correctly follow much of the actual travelled path . if that predicted path breaks from a normal routine , such as going home after work , then a malicious entity may be able to either deduce where the person could have gone with some foreknowledge of the person s habits or take advantage of their altered routine . in time
, we expect that improvements will be made to our initial approach .
the data , once collected , does not go away and a continually improving path prediction algorithm may plumb more private information from these traces .
these traces should be treated as any other private information instead of being handled as already anonymized data .
our work is not directed against any company or organization experiences have shown that even well - intentioned uses of data can result in losses of privacy so we wish to highlight potential dangers of this type of data collection .
usage - based insurance has been introduced in at least 39 states of the united states , and more are approving the approach because it has been claimed to be privacy - preserving .
our work shows that speed - based data collection erode user privacy of policy holders over time .
we do not claim that insurance companies are violating policy holder privacy in this way .
however , the general principle of any privacy preserving data collection is to collect only the data that is necessary for a particular application , and no more .
this material is based upon work supported by the national science foundation under grant numbers 1228777 and 1211079 .
any opinions , findings , and conclusions or recommendations expressed in this material are those of the author(s ) and do not necessarily reflect the views of the national science foundation .
@xmath4 m. golfarelli , d. maio , and s. rizzi . elastic correction of dead - reckoning errors in map building . in _ intelligent robots and systems , 1998 .
proceedings . ,
1998 ieee / rsj international conference on _ , volume 2 , pages 905911 vol.2 , 1998 .
e. krakiwsky , c. harris , and r. wong .
a kalman filter for integrating dead reckoning , map matching and gps positioning . in _ position location and navigation symposium , 1988 .
navigation into the 21st century .
ieee plans 88 .
, ieee _ , pages 3946 , 1988 .
l. zhao , w. y. ochieng , m. a. quddus , and r. b. noland .
an extended kalman filter algorithm for integrating gps and low cost dead reckoning system data for vehicle performance and emissions monitoring . | today people increasingly have the opportunity to opt - in to `` usage - based '' automotive insurance programs for reducing insurance premiums . in these programs ,
participants install devices in their vehicles that monitor their driving behavior , which raises some privacy concerns .
some devices collect fine - grained speed data to monitor driving habits .
companies that use these devices claim that their approach is privacy - preserving because speedometer measurements do not have physical locations .
however , we show that with knowledge of the user s home location , as the insurance companies have , speed data is sufficient to discover driving routes and destinations when trip data is collected over a period of weeks . to demonstrate the real - world applicability of our approach we applied our algorithm ,
_ elastic pathing _ , to data collected over hundreds of driving trips occurring over several months . with this data and our approach
, we were able to predict trip destinations to within 250 meters of ground truth in 10% of the traces and within 500 meters in 20% of the traces .
this result , combined with the amount of speed data that is being collected by insurance companies , constitutes a substantial breach of privacy because a person s regular driving pattern can be deduced with repeated examples of the same paths with just a few weeks of monitoring .
[ privacy ] |
there is now a consensus that , in well - ionized accretion disks , magnetorotational instabilities ( mri ) create turbulence that provides the ` anomalous viscosity ' required to drive accretion @xcite .
this removes a major uncertainty that afflicted previous theoretical models for black hole accretion , and opens the possibility of using numerical simulations to study directly the structure and variability of the accretion flow .
questions that we might hope to address include : * the magnetohydrodynamics ( mhd ) of the flow as it crosses @xmath1 , the marginally stable circular orbit .
recent work has suggested that mhd effects interior to @xmath1 ( in the ` plunging ' region ) could invalidate existing models of black hole accretion , with consequences that include an increase in the predicted radiative efficiency of thin disk accretion @xcite .
this suggestion remains controversial @xcite .
* the predicted variability in emission from the inner disk . * the structure of the disk magnetic field and the rate of transport of magnetic flux through @xmath1 , which together determine the efficiency of the blandford - znajek mechanism for extracting spin energy of the black hole @xcite .
this article presents results from simplified mhd simulations@xcite of black hole accretion that focus on the first of these questions .
we outline the numerical approach , and discuss our results and how they compare with those of other groups @xcite . our conclusion from work to date
is that mhd effects in the plunging region _ can _ have an important influence on the dynamics of the inner disk flow , provided that the magnetic fields in the disk are already moderately strong .
ideally , we would like to simulate a large volume of disk ( to allow for global effects@xcite and ease worries about treatment of the boundaries ) , for a long time period ( to average out fluctuations ) , at high resolution ( to resolve the most unstable scales of the magnetorotational instability ) , with the most realistic physics possible .
unfortunately , we ca nt , so compromises are needed .
our approach has been to simulate the simplest disk model that we believe includes the essential physical effects , while aiming for the highest resolution near and inside the last stable orbit .
.summary of the simulations discussed in this article .
all runs have the same sound speed , @xmath4 ( evaluated at @xmath1 ) , spatial domain @xmath5 , @xmath6 , @xmath7 , and resolution @xmath8 , @xmath9 , @xmath10 .
the time units are such that the orbital period at the last stable orbit is @xmath11 . [
cols="<,<,<,^,^ " , ] we use the zeus mhd code @xcite to solve the equations of ideal mhd within a restricted ` wedge ' of disk in cylindrical @xmath12 geometry .
the equation of state is isothermal , and a paczynski - wiita potential @xcite is used to model the effect of a last stable orbit within the newtonian hydrocode . to further simplify the problem
, we ignore the vertical component of gravity and consider an unstratified disk model .
the boundary conditions are set to outflow at the radial boundaries , are periodic in azimuth , and are either periodic or reflecting in @xmath13 .
the simulations begin with a stable , approximately gaussian surface density profile outside @xmath1 , which is threaded with a weak magnetic field .
this initial seed field has a constant ratio of thermal to magnetic energy @xmath14 , and is either vertical or azimuthal .
we evolve this setup , which is immediately unstable to the mri , until a significant fraction of the mass has been accreted , and plot results from timeslices towards the end of the runs when the magnetic fields have reached a saturated state .
table [ table1 ] summarizes the parameters of three simulations , which are improved versions of those previously reported @xcite .
the simulated disks are ` thin ' in the sense that pressure gradients are negligibly small in the disk outside @xmath1 .
specifically , the ratio of sound speed to orbital velocity , which in a stratified disk is approximately equal to @xmath15 , is @xmath16 at the last stable orbit .
we have not ( yet ) attempted the more difficult task of simulating very thin disks , which have longer viscous timescales , and caution against extrapolating our conclusions into that regime . following a suggestion by charles gammie ( personal communication ) , we investigated whether varying the magnetic field strength in the _ disk _ ( i.e. at @xmath17 ) led to changes in the dynamics of the flow within the plunging region . to vary the field strength in the simulations , we make use of the fact that the saturation level of the mri can be altered by varying the net flux of seed fields in the initial conditions .
local simulations@xcite show that the resultant shakura - sunyaev @xmath18 parameter@xcite is approximately , @xmath19 where @xmath20 and @xmath21 are the alfven speeds for initial conditions with uniform vertical and azimuthal seed fields , and @xmath22 is the sound speed . for our simulations , we find that the range of initial conditions ( and vertical boundary conditions ) shown in table [ table1 ] leads to a variation in @xmath18 between @xmath23 and @xmath24 in the disk . of course , this is a numerical trick .
the _ true _ value of @xmath18 in the disk immediately outside @xmath1 will probably depend upon details of the disk physics ( for example , the relative importance of gas and radiation pressure ) , and may vary with time .
figure 1 illustrates the geometry of the simulations .
all mhd disk simulations look broadly similar , and these are no exception . we obtain a pattern of surface density fluctuations that are strongly sheared by the differential rotation , and disk magnetic fields that are predominantly azimuthal .
a map of the ratio of magnetic to thermal energy , also shown in the figure , displays clearly the predicted increase@xcite in the relative importance of magnetic fields near and interior to the marginally stable orbit . determining the influence of these fields upon the dynamics of the flow in the inner disk and in the plunging region is the main goal of the calculations .
figure [ fig2 ] shows the magnetic torque as a function of radius in the three simulations , expressed as the magnetic contribution to the equivalent shakura - sunyaev @xmath18 parameter , @xmath25 there are also hydrodynamic ( reynolds ) stresses , which are somewhat harder to measure@xcite , but which are found to be substantially smaller than the magnetic stresses in local simulations .
the three choices for the initial flux and vertical boundary conditions lead to large variations in the saturation level of the mri and associated @xmath26 .
the run with an initially azimuthal field produces an @xmath27 , while the run with a relatively strong initial vertical flux leads to a disk @xmath26 that exceeds 0.1 .
this increase is qualitatively in agreement with local simulations@xcite . for the run with the strongest field ,
the magnetic field energy density in the disk near @xmath1 is on average near equipartition with the thermal energy @xmath28 .
there are large fluctuations with time , however , including brief periods where the magnetic energy substantially exceeds the thermal energy . in all runs , the relative importance of magnetic fields compared to the thermal energy increases within @xmath1 .
figure [ fig2 ] also shows how the torques influence the dynamics of the flow .
we plot the specific angular momentum of the flow @xmath29 as a function of radius .
hydrodynamic models for thin disk accretion onto black holes obtain @xmath30 within the last stable orbit , corresponding to a zero - torque boundary condition for the disk@xcite
. a non - zero @xmath31 implies transport of angular momentum ( and implicitly energy ) into the disk from within the plunging region .
we find that @xmath31 in the plunging region correlates with @xmath26 in the disk .
weak disk fields lead to a small ( but significant ) decline in @xmath29 within the last stable orbit , while strong fields lead to a steeply declining specific angular momentum profile at small radii .
the former behavior is similar to that seen in our earlier simulations@xcite ( which also had rather weak fields ) , while the latter is comparable to the results obtained by hawley@xcite and hawley and krolik@xcite .
their global simulations , which are substantially more ambitious than ours in terms of the included physics and spatial domain , did indeed generate relatively strong magnetic fields .
the limitations of the current simulations are myriad and obvious .
we are accutely aware that effects in the disk corona could be important@xcite , and that there is more to general relativity than a pseudo - newtonian potential@xcite .
nonetheless , we believe that some conclusions can be drawn from existing work : * unstratified global simulations confirm that magnetic torques persist within the last stable orbit , but suggest that if @xmath3 their influence on the dynamics of the flow is relatively modest@xcite . by modest we mean that the gradient of the specific angular momentum , @xmath31 , is non - zero but small at and inside @xmath1 . if these simulations reflect reality , existing models of black hole accretion would be a pretty good approximation for thin disks@xcite . *
a gross violation of the zero - torque boundary condition at @xmath1 is also possible@xcite .
this would increase the radiative efficiency of thin disk accretion above the usual @xmath32 , and have other consequences@xcite .
we believe that these strong effects _ only occur _ if @xmath2 in the disk
. this would be consistent with existing simulations@xcite , and with the results presented here . for observations
, these results suggest that the radiative efficiency of thin disk accretion may vary , both between systems , and in an individual system if the magnetic field strength in the disk varies with time .
we thank the developers of zeus and zeus - mp for making these codes available as community resources . p.j.a .
thanks jila for hospitality during the course of part of this work .
c.s.r . acknowledges support from hubble fellowship grant hf-01113.01 - 98a .
this grant was awarded by the space telescope science institute , operated by aura for nasa under contract nas 5 - 26555 .
c.s.r . also thanks support from the nsf under grants ast 98 - 76887 and ast 95 - 29170 .
j.c . was supported by nasa / atp grant nags 5 - 7723 . | we discuss the results of three - dimensional magnetohydrodynamic simulations , using a pseudo - newtonian potential , of thin disk @xmath0 accretion onto black holes .
we find ( i ) that magnetic stresses persist within @xmath1 , the marginally stable orbit , and ( ii ) that the importance of those stresses for the dynamics of the flow depends upon the strength of magnetic fields in the disk outside @xmath1 .
strong disk magnetic fields ( @xmath2 ) lead to a gross violation of the zero - torque boundary condition at @xmath1 , while weaker fields ( @xmath3 ) produce results more akin to traditional models for thin disk accretion onto black holes .
fluctuations in the magnetic field strength in the disk could lead to changes in the radiative efficiency of the flow on short timescales . |
there is a standard list of fundamental open questions beyond the standard model : + @xmath1 what is the origin of particle masses ?
+ are they indeed due to a higgs boson , as hypothesized within the standard model ?
if so , is the higgs boson accompanied by some other physics such as supersymmetry ? if not , what replaces the higgs boson ? + @xmath1 why are there so many types of matter particles ?
+ related to this question is the mixing of the different flavours of quarks and leptons , and the mechanism for cp violation .
this matter - antimatter difference is thought to be linked to cosmological baryogenesis .
however , the standard model can not , by itself , explain the cosmological matter - antimatter asymmetry , adding urgency to the search for flavour and cp violation beyond the standard model .
+ @xmath1 are the fundamental forces unified ?
+ if so , in the simplest models this unification occurs only at some very high energy @xmath2 gev .
physics at this scale may be probed via neutrino physics , or possibly less directly at accelerators via measurements of particle masses and couplings , and looking for unification relations between them .
+ @xmath1 what is the quantum theory of gravity ? + the best candidate for such a theory may be ( super)string theory , which generously predicts extra space - time dimensions as well as supersymmetry , but at what energy scale ?
the good news is that all of these fundamental open questions will be addressed by the lhc : its energy should be ample for resolving the problem of mass , including the questions whether there is a higgs boson and/or supersymmetry , a dedicated experiment will be examining matter - antimatter differences , models of unification could be probed via measurements of sparticle masses and couplings , and string theory might be probed via supersymmetry breaking , extra dimensions or even black hole production and decay .
accordingly , most of this talk will be concerned with lhc physics , accompanied by some asides about linear collider ( lc ) physics .
supersymmetry may play a role in answering most of the open questions , for example by stabilizing the scale of electroweak symmetry breaking or by aiding the unification of the gauge coupling , and it also seems to play an essential role in string theory .
therefore , i have invested many or my personal efforts in supersymmetry , and apologize for giving pride of place to it in the rest of this talk .
several different higgs production mechanisms will be important at the lhc , including gluon - gluon and @xmath3 fusion , and production in association with gauge bosons or heavy quarks .
several decay modes will also be important , including @xmath4 , four leptons , @xmath5 , @xmath3 , @xmath6 , etc .
accordingly the search for the higgs boson at the lhc will require the combination of many different signatures and hence excellent understanding of all components of atlas and cms . as seen in fig .
[ fig : higgs ] , once both detectors have accumulated and analyzed a couple of hundred pb@xmath7 of data , they may be able to start excluding certain ranges of higgs masses at the 95 % confidence level @xcite . with @xmath8 fb@xmath7 of analyzed data , they might be able to exclude a standard model higgs boson over the entire mass range up to 1 tev , or they might be able to establish a five-@xmath9 signal if the higgs boson weighs between @xmath10 and 500 gev .
several fb@xmath7 , corresponding to several months at one tenth of the design luminosity , would be needed to discover the higgs boson over the entire mass range . with some luck
, it may also be possible to determine the spin of the higgs boson @xcite .
for example , if it is light and is seen to decay into @xmath4 , it can not have spin one .
if it has higher mass and is seen to decay into four leptons , their angular correlations will enable @xmath11 states other than @xmath12 to be excluded with high significance .
analogous studies could be made at a lc , also using angular distibutions and the threshold behaviour for @xmath13 @xcite . as seen in fig .
[ fig : couplings ] , the lhc may also make a first analysis of the proportionality of the higgs couplings to particle masses @xcite .
if @xmath14 gev , it may be able to measure the couplings to @xmath5 , @xmath6 , @xmath3 , @xmath15 and @xmath16 each with an accuracy @xmath17 % .
much more precise measurements could be made subsequently with a lc @xcite . as well as its importance for particle physics
, the discovery of a higgs boson would also be very important for cosmology .
an elementary higgs boson would have caused a phase transition in the early universe when it was @xmath18 s old , and might have generated the matter in the universe via electroweak barygenesis .
further back in the history of the universe , a related inflaton might have expanded the universe exponentially when it was @xmath19 s old .
coming back to the present , naively the higgs boson of the standard model would contribute a factor @xmath20 too much to the present - day dark energy , apparently requiring some ` miraculously ' fine - tuned cancellation .
cosmologists should be as interested as particle physicists in the dnouement of the higgs saga .
for the time being , the lhc and a possible subsequent lc will be our only direct windows on this physics .
my personal favourite candidate for new physics beyond the higgs boson is supersymmetry , for several reasons .
true , it is intrinsically beautiful and ( almost ) an essential ingredient in string theory , etc . , but these are not the reasons that motivate me to expect that it may appear at the lhc . there are four specific reasons why one might expect supersymmetry to appear around the tev scale , and hence be accessible to the lhc .
one is the naturalness or hierarchy problem @xcite , another is the unification of the gauge couplings @xcite , another is the supersymmetric prediction of a light higgs boson as preferred by the precision electroweak data @xcite , and another is that many supersymmetric models predict the existence of cold dark matter with a density comparable to that required by astrophysics and cosmology @xcite . in the following ,
i concentrate on the minimal supersymmetric extension of the standard model ( mssm ) .
in addition to the coupling @xmath21 between its two gauge multiplets , and the ratio @xmath22 of their vacuum expectation values , the mssm has many apparently arbitrary soft supersymmetry - breaking parameters , including scalar masses @xmath23 , gaugino masses @xmath24 , trilinear soft couplings @xmath25 and a bilinear soft coupling @xmath26 .
it is commonly assumed that the @xmath27 and @xmath25 are each universal at some input gut scale , a framework known as the constrained mssm ( cmssm ) .
this is not the same as minimal supergravity ( msugra ) , which fixes in addition the gravitino mass : @xmath28 , and imposes @xmath29 .
we will see later some potential implications of these extra conditions .
there are direct limits on sparticle masses from their absences at lep and the tevatron , and indirect constraints from the lep lower limit @xmath30 gev and from @xmath31 physics , including in particular measurements of @xmath32 decay .
one possible indication of new physics at the tev scale may be provided by the bnl measurement of the anomalous magnetic moment of the muon @xcite , that seems to exhibit a three-@xmath9 discrepancy with the standard model , though this is still somewhat controversial .
the strongest constraint on ( one combination of ) supersymmetric model parameters is provided by the density of cold dark matter : @xmath33 , assuming that it is mainly composed of the lightest neutralino @xmath34 .
this not the only possibility : presumably the lsp should have neither strong nor electromagnetic interactions , but there are other candidates that also have these properties .
the supersymmetric partners of the neutrinos have been excluded by a combination of lep and direct dark matter searches , but the lsp might be the spartner of some particle beyond the standard model , such as the gravitino . in a minimal supersymmetric model with universal soft supersymmetry - breaking parameters , one may consider these constraints in the @xmath35 plane .
as shown in the left panel of fig .
[ fig : cmssm ] , where the lsp is assumed to be the lightest neutralino , the resulting allowed regions include a narrow ( pale turquoise ) strip near the boundary beyond which @xmath36 ( dark brown shading ) where @xmath37 gev , and another strip ( not shown ) where @xmath38 tev near the boundary beyond which electroweak symmetry breaking is no longer possible ( this region is disfavoured by @xmath39 , indicated by pink shading ) . the fact that these allowed strips extend to relatively large values of @xmath24 and @xmath23 implies that even the lightest visible sparticles may not be very light , as seen in the right panel of fig .
[ fig : cmssm ] . we generated a large sample of possible cmssm scenarios ( the red symbols ) ,
of which a subsample ( shown in darker blue ) provide the right amount of cold dark matter @xcite .
also shown ( in paler green ) are the scenarios that would be detectable at the lhc .
we see that most ( but not all ) of the dark matter scenarios should be visible at the lhc : one may be hopeful , but there is no guarantee that the lhc will discover supersymmetry .
on the other hand , direct astrophysical dark matter search experiments ( very pale yellow ) may have less chances in the foreseeable future .
an lc with centre - of - mass energy @xmath40 tev would be needed to see sparticles in all the cmssm dark matter scenarios we sampled .
can one estimate the scale of supersymmetry on the basis of present data @xcite ?
the precision electroweak measurements of @xmath41 and @xmath42 both have some sensitivity to @xmath24 through radiative corrections , and @xmath41 slightly prefers smaller values of @xmath24 , though this trend is hardly significant .
the agreement of @xmath32 with the standard model offers no encouragement to enthusiasts for light supersymmetry , and other @xmath31-decay observables such as @xmath43 and @xmath44 do not yet provide much information about the possible scale of supersymmetry breaking . on the other hand ,
the disagreement between the experimental value of @xmath39 with the theoretical value calculated in the standard model using low - energy @xmath0 data could be explained by light supersymmetry , as shown already in the left panel of fig .
[ fig : cmssm ] . as shown in the left panel of fig .
[ fig : ehoww ] @xcite , a global fit to precision electroweak and @xmath31-decay observables indicates a preference for relatively small values of @xmath24 .
this is due predominantly to @xmath39 , but there is some support from the measurements of @xmath41 .
correspondingly , the most likely value for the mass of the lightest supersymmetric higgs boson is only slightly above the lep lower limit , as seen in the right panel of fig .
[ fig : ehoww ] .
the classic signature of sparticle production is the appearance of events with missing energy - momentum carried away by invisible dark matter particles .
studies indicate that such a signature should be observable at the lhc above instrumental and standard model backgrounds @xcite . in many supersymmetric scenarios ,
the lhc would produce a pair of squarks and/or gluinos primarily , and these would subsequently decay into a cascade of secondary lighter sparticles , at the end of each decay chain would appear the lsp , namely the lightest neutralino @xmath34 in the cmssm framework considered so far .
studies indicate that in many benchmark scenarios a number of different sparticle species could be observed at the lhc and their masses measured quite accurately @xcite .
a lc would produce democratically all the sparticle species that are kinematically accessible , and therefore would be particularly interesting for producing sparticles without strong interactions , such as sleptons , charginos and neutralinos , which are expected to be lighter than squarks and gluinos in cmssm scenarios
. even a limited amount of lhc luminosity would enable the lhc to see squarks or gluinos weighing a tev or more .
as shown in the right panel of fig .
[ fig : higgs ] , it is estimated @xcite that 0.1 fb@xmath7 of lhc luminosity would be sufficient to observe a gluino with a mass of 1.2 gev at the five-@xmath9 level , or to exclude a gluino weighing @xmath45 tev .
the discovery and exclusion reaches would extend to about 2.2 and 2.5 tev , respectively with 10 fb@xmath7 of lhc luminosity . as also shown in the right panel of fig .
[ fig : higgs ] , assuming universal input sparticle masses at the gut scale , the corresponding thresholds for sparticle pair production at a linear @xmath0 collider would be 0.5 ( 0.6 ) or 0.8 ( 1.0 ) tev @xcite . hence ,
for example , if the lhc discovers the gluino with 0.1 fb@xmath7 , one may expect that sparticle pair production would be accessible to a linear collider with centre - of - mass energy of 0.5 tev , whereas if the lhc does not discover the gluino even with 10 fb@xmath7 , the @xmath0 sparticle pair production threshold may be above 1 tev .
at least in such a simple model , the lhc will tell us how much energy a linear collider would need to find supersymmetry @xcite .
if nature is kind , and sparticles not only exist but also are quite light , it will be possible to test directly unification of the gauge couplings and universality of the soft supersymmetry - breaking scalar masses with high precision , in particular by comparing measurements at the lhc and the ilc @xcite . however , there is no guarantee that any sparticles will be light enough to be produced at the ilc , even with a centre - of - mass energy of 1 tev .
for this reason , and also because even if there are some light sparticles the heavier ones will be produced only at higher energies , a high - energy lc would be advantageous , which is why cern and its collaborators are developing the clic technology that should be capable of reaching 3 tev in the centre of mass @xcite .
so far , i have concentrated on the cmssm , assuming that all the soft supersymmetry - breaking scalar masses are universal at the gut scale . however
, this may not be the case : in particular , there is no good theoretical or phenomenological reason why the higgs scalar masses should be universal . in models with non - universal higgs masses ( nuhm ) ,
there are two additional degrees of freedom , and the higgs mixing parameter @xmath21 and the pseudoscalar higgs mass @xmath46 may be treated as free parameters @xcite .
we have studied @xcite whether the nuhm framework could accommodate a pseudoscalar supersymmetric higgs boson as light as the value not yet excluded by direct searches at the tevatron .
recall that the cdf experiment saw a two-@xmath9 excess in the @xmath5 spectrum that could be explained by supersymmetric higgs bosons weighing @xmath47 gev @xcite , whereas the d0 experiment saw no such excess @xcite .
we found @xcite that @xmath48 gev could indeed be accommodated within the nuhm , if @xmath49 gev , @xmath50 gev , @xmath51 gev and @xmath52 tev , as seen in fig .
[ fig : ehow5 ] .
this would be quite an extreme scenario , but it would have the merit of being testable in the near future . in this corner of the nuhm parameter space , @xmath53 and the cold dark matter scattering rates would all be very close to the present experimental limits .
the lsp should have no strong or electromagnetic interactions , otherwise it would bind to conventional matter and be detectable as some anomalous heavy nucleus . apart from the lightest neutralino @xmath34 ,
possible weakly - interacting scandidates include the sneutrino , which is excluded by lep and direct searches in simple models , and the gravitino @xmath54 , which has attracted relatively little detailed attention until recently . it would be a nightmare for astrophysical detection , but a bonanza for the lhc @xcite , as we now discuss .
the next - to - lightest sparticle ( nlsp ) would have a very long lifetime in models with gravity - mediated supersymmetry breaking , due to the weak gravitational strength of the interactions responsible for its decays .
for example , if the nlsp is the lighter stau slepton @xmath55 one has @xmath56 leading to a lifetime that could be measured in hours , days , weeks , months or even years !
the stau is not the only candidate for the nlsp .
other generic possibilities include the lightest neutralino @xmath34 , a sneutrino @xcite , or even the lighter stop squark @xcite . in the case of minimal supergravity ( msugra )
, one finds that a gravitino lsp combined with a stau nlsp is as generic as the conventional neutralino lsp scenario , as seen by the light ( yellow ) shaded region in the left panel of fig .
[ fig : gdm ] .
all of these possibilities are constrained by astrophysics and cosmology , particularly through limits on the decays of the metastable nlsp @xcite , and its bound states if it is charged @xcite .
these effects may even improve the agreement between cosmological nucleosynthesis calculations and the observed lithium abundances , as shown in the pink ( darker ) shaded region in the right panel of fig .
[ fig : gdm ] for one particular non - msugra scenario @xcite .
this scenario could be probed only with a higher - energy lc @xcite .
we recently examined how a stau nlsp scenario could be explored at the lhc @xcite .
we found that the normal experimental triggers on jets and energetic muons and electrons would select the events containing staus @xcite .
it would then be possible to identify the stau tracks in the events with quite high efficiency , and measure the stau mass very accurately via a combination of momentum and time - of - flight measurements , as seen in fig .
[ fig : staumass ] . the stau could be combined with jets in the event to reconstruct the masses of heavier sparticles in the supersymmetric decay cascades , as in neutralino lsp scenarios , as shown in fig [ fig : recontau ] .
very slow - moving staus might be stopped either in the detector material or in the rock surrounding the cavern @xcite .
one could use the muon system to locate the stau s impact point on the cavern wall with an uncertainty @xmath8 cm , and its impact angle with an accuracy @xmath57 .
one might then be able to bore into the cavern wall and remove a core from the rock , which one could then store while waiting for the stau to decay . before boring , one would have to wait for a shut - down of the lhc , since the radioactivity in the cavern is quite high while it is operating .
however , it is planned to stop the accelerator for a couple of days every month , so this strategy might work if the lifetime exceeds about @xmath58 s !
supersymmetry is certainly not the only possibility for new physics at the lhc .
another possibility offered by string theory is that there might be large extra dimensions .
when string theory was originally proposed as a ` theory of everything ' , it was imagined that all the extra dimensions would be curled up on length scales comparable to the planck length @xmath59 cm . however
, then it was realized that string unification could be achieved more easily if one of these dimensions was somewhat smaller than the gut scale @xcite , and a number of scenarios with much larger extra dimensions have been considered .
for example , an extra dimension of size @xmath8 tev@xmath7 could help break supersymmetry @xcite and/or the electroweak gauge symmetry , an extra dimension of micron size could help rewrite the hierarchy problem @xcite , and even infinite extra dimensions are allowed if they are warped appropriately @xcite . in many of these scenarios
, there are potential signals to be found at the lhc , such as kaluza - klein excitations of gravitons , or missing energy ` leaking ' into an extra dimension
. the most spectacular possibility would occur if gravity becomes strong at the tev energy scale , in which case microscopic black holes might be produced at the lhc
. these would be very unstable , decaying rapidly via hawking radiation into multiple jets , leptons and photons , that would be easily detectable @xcite and distinguishable from supersymmetry and other extra - dimensional scenarios @xcite .
the origin of mass is the most pressing problem in particle physics , and requires a solution within the lhc energy range .
will it be a simple higgs boson ? and/or
the lhc will tell !
there are many speculative ideas for other possible new physics beyond the standard model , such as grand unification , strings , and extra dimensions .
the lhc is also capable of probing many of these speculations including , as the last two examples show , novel ideas that were undreamt of when the lhc and its experiments were designed .
we do not know what the lhc will find , but we can be sure that its discoveries will set the agenda for possible future projects , such as a linear @xmath0 collider .
a. blondel , l. camilleri , a. ceccucci , j. ellis , m. lindroos , m. mangano and g. rolandi , _ physics opportunities with future proton accelerators at cern _ , arxiv : hep - ph/0609102 . s. y. choi , d. j. miller , m. m. muhlleitner and p. m. zerwas , phys . lett .
b * 553 * ( 2003 ) 61 [ arxiv : hep - ph/0210077 ] .
c. p. buszello , i. fleck , p. marquard and j. j. van der bij , eur .
j. c * 32 * ( 2004 ) 209 [ arxiv : hep - ph/0212396 ] .
e. accomando _ et al .
_ , _ workshop on cp studies and non - standard higgs physics _ , arxiv : hep - ph/0608079 .
d. j. miller , s. y. choi , b. eberle , m. m. muhlleitner and p. m. zerwas , phys .
b * 505 * ( 2001 ) 149 [ arxiv : hep - ph/0102023 ] .
g. weiglein _ et al . _
[ lhc / lc study group ] , phys .
* 426 * ( 2006 ) 47 [ arxiv : hep - ph/0410364 ] ; a. djouadi , j. lykken , k. monig , y. okada , m. j. oreglia and s. yamashita _ et al .
_ , _ international linear collider reference design report volume 2 : physics at the ilc _ , arxiv:0709.1893 [ hep - ph ] .
l. maiani , _ all you need to know about the higgs boson _ , proceedings of the gif - sur - yvette summer school on particle physics , 1979 , pp.1 - 52 ; g. t hooft , in _ recent developments in gauge theories _ , proceedings of the nato advanced study institute , cargse , 1979 , eds .
g. t hooft et al .
( plenum press , ny , 1980 ) ; e. witten , phys .
b * 105 * ( 1981 ) 267 .
j. r. ellis , s. kelley and d. v. nanopoulos , phys .
b * 249 * ( 1990 ) 441 and phys .
b * 260 * ( 1991 ) 131 ; u. amaldi , w. de boer and h. furstenau , phys .
b * 260 * ( 1991 ) 447 ; c. giunti , c. w. kim and u. w. lee , mod . phys
a * 6 * ( 1991 ) 1745 .
m. w. grunewald , arxiv:0709.3744 [ hep - ex ] .
j. r. ellis , j. s. hagelin , d. v. nanopoulos , k. a. olive and m. srednicki , nucl .
b * 238 * ( 1984 ) 453 .
h. n. brown _ et al .
_ [ muon g-2 collaboration ] , phys .
* 86 * ( 2001 ) 2227 [ arxiv : hep - ex/0102017 ] .
j. r. ellis , k. a. olive , y. santoso and v. c. spanos , phys .
b * 603 * ( 2004 ) 51 [ arxiv : hep - ph/0408118 ] .
j. r. ellis , s. heinemeyer , k. a. olive , a. m. weber and g. weiglein , arxiv:0706.0652 [ hep - ph ] .
i. hinchliffe , f. e. paige , m. d. shapiro , j. soderqvist and w. yao , phys .
d * 55 * ( 1997 ) 5520 [ arxiv : hep - ph/9610544 ] .
m. battaglia , a. de roeck , j. r. ellis , f. gianotti , k. a. olive and l. pape , eur .
j. c * 33 * ( 2004 ) 273 [ arxiv : hep - ph/0306219 ] ; a. de roeck , j. r. ellis , f. gianotti , f. moortgat , k. a. olive and l. pape , arxiv : hep - ph/0508198 .
this discussion is being extended in j. r. ellis , k. a. olive and p. sandick , _ in preparation_. p. bechtle , k. desch , w. porod and p. wienemann , eur .
j. c * 46 * ( 2006 ) 533 [ arxiv : hep - ph/0511006 ] .
e. accomando _ et al .
_ [ clic physics working group ] , _ physics at the clic multi - tev linear collider _ , arxiv : hep - ph/0412251 .
j. ellis , t. falk , k. olive and y. santoso , _ nucl .
phys . _ * b 652 * ( 2003 ) 259 , hep - ph/0210205 .
j. r. ellis , s. heinemeyer , k. a. olive and g. weiglein , phys .
b * 653 * ( 2007 ) 292 [ arxiv:0706.0977 [ hep - ph ] .
cdf collaboration , cdf note 8676 , see : + http://www-cdf.fnal.gov/~aa/mssm_htt_1fb/note/cdf8676.pdf .
a. de roeck , j. r. ellis , f. gianotti , f. moortgat , k. a. olive and l. pape , eur .
j. c * 49 * ( 2007 ) 1041 [ arxiv : hep - ph/0508198 ] .
j. r. ellis , a. r. raklev and o. k. oye , jhep * 0610 * ( 2006 ) 061 [ arxiv : hep - ph/0607261 ] .
j. r. ellis , a. r. raklev and o. k. oye , _ in preparation_. l. covi and s. kraml , arxiv : hep - ph/0703130 . j. l. diaz - cruz , j. r. ellis , k. a. olive and y. santoso , jhep * 0705 * ( 2007 ) 003 [ arxiv : hep - ph/0701229 ] .
r. h. cyburt , j. r. ellis , b. d. fields and k. a. olive , phys .
d * 67 * ( 2003 ) 103521 [ arxiv : astro - ph/0211258 ] .
m. pospelov , phys .
* 98 * ( 2007 ) 231301 [ arxiv : hep - ph/0605215 ] .
r. h. cyburt , j. r. ellis , b. d. fields , k. a. olive and v. c. spanos , jcap * 0611 * ( 2006 ) 014 [ arxiv : astro - ph/0608562 ] .
o. cakir , i. t. cakir , j. r. ellis and z. kirca , arxiv : hep - ph/0703121 .
j. l. feng and b. t. smith , phys .
d * 71 * ( 2005 ) 015004 [ erratum - ibid .
d * 71 * ( 2005 ) 0109904 ] [ arxiv : hep - ph/0409278 ] .
k. hamaguchi , y. kuno , t. nakaya and m. m. nojiri , phys .
d * 70 * ( 2004 ) 115007 [ arxiv : hep - ph/0409248 ] .
e. witten , nucl .
b * 471 * ( 1996 ) 135 [ arxiv : hep - th/9602070 ] .
i. antoniadis , phys .
b * 246 * ( 1990 ) 377 .
i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 436 * ( 1998 ) 257 [ arxiv : hep - ph/9804398 ] .
l. randall and r. sundrum , phys .
* 83 * ( 1999 ) 3370 [ arxiv : hep - ph/9905221 ] . c. m. harris , m. j. palmer , m. a. parker , p. richardson , a. sabetfakhri and b. r. webber , jhep * 0505 * ( 2005 ) 053 [ arxiv : hep - ph/0411022 ] .
j. m. smillie and b. r. webber , jhep * 0510 * ( 2005 ) 069 [ arxiv : hep - ph/0507170 ] . | many of the open questions beyond the standard model will be addressed by the lhc , including the origin of mass , supersymmetry , dark matter and the possibility of large extra dimensions .
a linear @xmath0 collider ( lc ) with sufficient centre - of - mass energy would add considerable value to the capabilities of the lhc .
address = theoretical physics division , physics department , cern , ch 1211 geneva 23 , switzerland , [email protected] cern - ph - th/2007 - 174 |
an @xmath0-step uniform random walk in @xmath1 starts at the origin and consists of @xmath0 independent steps of length @xmath2 , each of which is taken into a uniformly random direction .
in other words , each step corresponds to a random vector uniformly distributed on the unit sphere .
the study of such walks originates with pearson @xcite , who was interested in planar walks , which he looked at @xcite as migrations of , for instance , mosquitos moving a step after each breeding cycle .
random walks in three dimensions ( known as random flights ) seem first to have been studied in extenso by rayleigh @xcite , and higher dimensions are touched upon in @xcite .
this paper is a companion to @xcite and @xcite , which studied the analytic and number theoretic behaviour of short uniform random walks in the plane ( five steps or less ) . in this work
we revisit the same issues in higher dimensions . somewhat to our surprise , we are able to provide complete extensions for most of the central results in the culminating paper @xcite . throughout the paper ,
@xmath0 and @xmath3 are the positive integers corresponding to the number of steps and the dimension of the random walk we are considering .
moreover , we denote with @xmath4 the half - integer @xmath5 it turns out that most results are more naturally expressed in terms of this parameter @xmath4 , and so we denote , for instance , with @xmath6 the probability density function of the distance to the origin after @xmath0 random unit steps in @xmath1 . in section [ ssec :
den ] , we mostly follow the account in @xcite and develop the basic bessel integral representations for these densities beginning with theorems [ thm : bi ] and [ thm : pn : besseld ] , which are central to our analysis . in particular , a brief discussion of the ( elementary ) case of odd dimensions is included in section [ ssec : odd ] . in section [ ssec : mom ] , we turn to general results on the associated moment functions @xmath7 in particular , we derive in theorem [ thm : w : even ] a formula for the even moments @xmath8 as a multiple sum over the product of multinomial coefficients . as a consequence ,
we observe another interpretation of the catalan numbers as the even moments of the distance after two random steps in four dimensions , and realize , more generally , in example [ eg : w : values4d ] the moments in four dimensions in terms of powers of the narayana triangular matrix .
we shall see that dimensions two and four are privileged in that all even moments are integral only in those two dimensions . in section [ sec : mom ]
, we turn to a detailed analysis of the moments of short step walks : from two - step walks ( [ ssec : mom2 ] ) through five steps ( [ ssec : mom5 ] ) .
for instance , we show in , theorem [ thm : w3:gf ] and example [ eg : w4:ogf ] that the ordinary generating functions of the even moments for @xmath9 can be expressed in terms of hypergeometric functions .
we are also able to give closed forms for all odd moments for less than five steps . in section [ sec : den ] , we perform a corresponding analysis of the densities @xmath6 : from two - step walks ( [ ssec : den2 ] ) through five steps ( [ ssec : den5 ] ) .
one especially striking result for @xmath10 , shown in corollary [ cor : fe3 ] , is the following functional equation for the probability density function @xmath11 . for @xmath12 , and each half - integer @xmath13 , the function @xmath14 satisfies the functional equation @xmath15 finally , in section [ sec : conc ] , we make some concluding remarks and leave several open questions .
as much as possible , we keep our notation consistent with that in @xcite , and especially @xcite , to which we refer for details of how to exploit the mellin transform and similar matters .
random walks in higher dimensions are also briefly discussed in ( * ? ? ?
* chapter 4 ) .
in particular , wan gives evaluations in arbitrary dimensions for the second moments , which we consider and generalize in example [ eg : w : smallk ] , as well as for two - step walks .
for the benefit of the reader , we briefly summarize the account given in ( * ? ? ?
* chapter 2.2 ) of how to determine the probability density function @xmath6 of an @xmath0-step random walk in @xmath3 dimensions .
the reader interested in further details and corresponding results for more general random walks , for instance , with varying step sizes , is referred to @xcite . throughout the paper ,
the _ normalized bessel function of the first kind _ is @xmath16 with this normalization , we have @xmath17 and @xmath18 as @xmath19 on the real line .
note also that @xmath20 , which in part explains why analysis in 3-space is so simple . more generally , all half - integer order @xmath21 are elementary .
[ thm : bi ] the probability density function of the distance to the origin in @xmath22 dimensions after @xmath23 steps is , for @xmath24 , @xmath25 where , as introduced in , @xmath26 . let @xmath27 be a random vector , which is uniformly distributed on the unit sphere in @xmath1 .
that is , @xmath27 describes the displacement of a single step in our random walk .
then the fourier transform of its induced probability measure @xmath28 is @xmath29 this is a special case of the famous formula ( * ? ? ? * ( 2.30 ) ) , @xmath30 with @xmath31 , for integrals of orthogonally invariant functions .
note that the position @xmath32 of a random walk after @xmath0 unit steps in @xmath1 is distributed like the sum of @xmath0 independent copies of @xmath27 .
the fourier transform of @xmath33 therefore is @xmath34 we are now able to obtain the probability density function @xmath35 of the position after @xmath0 unit steps in @xmath1 via the inversion relation @xmath36 combining , and , we find @xmath37 since the surface area of the unit sphere in @xmath1 is @xmath38 , the density functions for the position and distance are related by @xmath39 whence we arrive at the formula .
the probability densities @xmath11 of the distance to the origin after three random steps in dimensions @xmath40 are depicted in figure [ fig : p3 ] . with the exception of the planar case , which has a logarithmic singularity at @xmath41 , these functions are at least continuous in the interval @xmath42 $ ] , on which they are supported
. their precise regularity is provided by corollary [ cor : pn : reg ] . for comparison , the probability densities of four - step walks in dimensions
@xmath40 are plotted in figure [ fig : p4 ] , and corresponding plots for five steps are provided by figure [ fig : p5 ] .
@xmath11 for @xmath43,width=340 ] @xmath44 for @xmath43,width=340 ] observe how the density functions center and spike as the dimensions increase . indeed , this is a general phenomenon and the distributions described by the densities @xmath6 approach a dirac distribution centered at @xmath45 .
the intuition behind this observation is as follows : as the dimension @xmath3 increases , the directions of each of the @xmath0 random steps increasingly tend to be close to orthogonal to each other .
that is , given an incoming direction , the direction of the next step will probably belong to a hyperplane which is almost orthogonal to this direction .
pythagoras theorem therefore predicts that the distance after @xmath0 steps is about @xmath45 .
a precise asymptotic result , which confirms this intuition , is given in example [ eg : p : diminf ] .
integrating the bessel integral representation for the probability density functions @xmath6 , we obtain a corresponding bessel integral representation for the cumulative distribution functions , @xmath46 of the distance to the origin after @xmath0 steps in @xmath3 dimensions .
suppose @xmath22 and @xmath23 .
then , for @xmath24 , @xmath47 a famous result of kluyver @xcite asserts that , for @xmath48 , @xmath49 that is , after @xmath0 unit steps in the plane , the probability to be within one unit of the starting point is @xmath50 .
this is nearly immediate from .
an elementary proof of this remarkable result was given recently by bernardi @xcite .
it is natural to wonder whether there exists an extension of this result to higher dimensions .
clearly , these probabilities decrease as the dimension increases . 1 . in the case of two steps
, we have @xmath51 which , in the case of integers @xmath13 and @xmath41 , reduces to @xmath52 alternatively , @xmath53 in particular , in dimensions @xmath54 , @xmath55 and @xmath56 , @xmath57 and it is obvious from that all the probabilities @xmath58 are of the form @xmath59 for some rational factor @xmath60 .
2 . in the case of three steps , we find @xmath61 indeed , for integers @xmath13 , we have the general formula @xmath62 in the limit we arrive at the improbable evaluation @xmath63 since @xmath64 goes to zero as the dimension goes to infinity .
the case of @xmath0-step walks , with @xmath65 , is much less accessible @xcite , and it would be interesting to obtain a more complete extension of kluyver s result to higher dimensions .
[ eg : p : diminf]asymptotically , for @xmath66 as @xmath67 , @xmath68 it may thus be derived from the bessel integral representation that asymptotically , as the dimension goes to infinity , @xmath69 , where @xmath70 for @xmath71 and @xmath72 for @xmath73 .
the probability density @xmath74 describes a scaled chi distribution with @xmath75 degrees of freedom .
its average is @xmath76 which converges to @xmath45 as @xmath67 . more generally , for the @xmath77th moment , with @xmath78 , @xmath79 as @xmath67 , and it is straightforward to compute further terms of this asymptotic expansion .
the fact that the @xmath77th moment approaches @xmath80 for large dimensions , of course , reflects the observation from figures [ fig : p3 ] , [ fig : p4 ] and [ fig : p5 ] that the probability densities @xmath81 center and spike at approximately @xmath45 .
these bessel integral representations also allow us to deduce the regularity of the density functions .
[ cor : pn : reg]the density @xmath6 is @xmath82 times continuously differentiable for @xmath24 if @xmath83 it follows from , that @xmath84 observe that this integral converges absolutely if @xmath85 , in which case @xmath6 is continuous .
repeatedly differentiating under the integral sign as long as is permitted , we conclude that @xmath6 is @xmath82 times continuously differentiable if @xmath86 and it only remains to solve for @xmath0 , respectively @xmath82 . in the case
@xmath87 , this implies that @xmath88 is in @xmath89 for @xmath90 , in @xmath91 for @xmath92 , in @xmath93 for @xmath94 , and so on .
indeed , @xmath95 has a logarithmic singularity at @xmath41 , and @xmath96 as well as @xmath97 are not differentiable at @xmath98 and @xmath99 , respectively .
let @xmath65 and @xmath22 such that @xmath100 .
then , @xmath101 where the derivative is understood to be taken from the right . starting with
, we differentiate @xmath102 many times , and compare with @xmath103 in the case @xmath87 , this reduces to @xmath104 , which was crucial in @xcite for explicitly evaluating @xmath105 . as long as @xmath6 is sufficiently differentiable at @xmath41 , we can further relate the value @xmath106 , occuring in , to corresponding values of derivatives .
let us illustrate this by showing that @xmath107 for all @xmath108 , @xmath109 such that @xmath110 .
in that case , we may , as in corollary [ cor : pn : reg ] , differentiate under the integral sign to obtain @xmath111 on the other hand , integrating by parts , we find @xmath112 combining these , we arrive at .
the densities of an @xmath0-step walk can be related to the densities of an @xmath113-step walk by the following generalization of ( * ? ? ?
* ( 2.70 ) ) .
[ thm : pn : rec]for @xmath24 and @xmath114 , the function @xmath115 satisfies @xmath116 recall from that the probability density of the position after @xmath0 steps in @xmath1 is @xmath117 since the steps of our walks are uniformly distributed vectors on the unit sphere in @xmath1 , we have @xmath118 where @xmath119 denotes the normalized surface measure of the unit sphere . after introducing @xmath3-dimensional spherical polar coordinates , as detailed , for instance , in @xcite , and a straightforward change of variables we arrive at .
finally , a computationally more accessible expression for the densities @xmath120 is given by the following generalization of a formula , which was derived by broadhurst @xcite in the case of two dimensions , that is , @xmath121 .
note that is the special case @xmath122 in .
[ thm : pn : besseld]let @xmath123 and @xmath22 . for any nonnegative integer @xmath124 , and @xmath125 , @xmath126 as in @xcite
, we proceed by induction on @xmath124 . the case @xmath127 is .
suppose that is known for some @xmath124 .
by the bessel function identity @xmath128 we find that , for any smooth function @xmath129 which is sufficiently small at @xmath130 and @xmath131 , @xmath132 g ( t ) { \mathrm{d}}t\\ & = & - \frac{1}{x } \int_0^{\infty } ( t x)^{\alpha + 1 } j_{\alpha } ( t x ) \left [ \frac{{\mathrm{d}}}{{\mathrm{d}}t } g ( t ) \right ] { \mathrm{d}}t\\ & = & \frac{1}{x^2 } \int_0^{\infty } ( t x)^{\alpha + 2 } j_{\alpha } ( t x ) \left [ - \frac{1}{t } \frac{{\mathrm{d}}}{{\mathrm{d}}t } g ( t ) \right ] { \mathrm{d}}t .
\end{aligned}\ ] ] in the second step , we used integration by parts and assumed that @xmath129 is such that @xmath133 vanishes as @xmath134 and @xmath135 . in the present case , @xmath136 and @xmath137
in order to complete the proof of , it only remains to demonstrate that indeed vanishes as required .
as @xmath134 , we have @xmath138 while @xmath139 , because @xmath140 is an even function .
hence , for @xmath141 , the term indeed vanishes as @xmath134 . on the other hand , as @xmath142 , we have @xmath143 by .
moreover , and imply that @xmath144 , as @xmath135 , and therefore that @xmath145 .
since @xmath146 , it follows that also vanishes as @xmath135 .
we note that the @xmath147 in the integrand of may be expressed as a finite sum of products of bessel functions .
this is made explicit in remark [ rk : j : d : k ] . in the case of odd dimension
@xmath3 , the bessel functions in section [ sec : basic ] have half - integer index @xmath4 and are therefore elementary , so that the situation is fairly well understood since rayleigh @xcite .
in particular , the probability density functions @xmath148 are piecewise polynomial in odd dimensions .
this is made explicit by the following theorem , which is obtained in @xcite , translated to our notation .
[ thm : pn : odd ] assume that the dimension @xmath149 is an odd number . then , for @xmath150 , @xmath151 where @xmath152 is the piecewise polynomial obtained from convoluting @xmath153$,}\\ 0 , & \text{otherwise } , \end{array } \right.\ ] ] @xmath154 times with itself . in the case
@xmath10 and @xmath155 , we have @xmath156 and @xmath157 as well as @xmath158 and , hence , @xmath159 we thus find that , for @xmath160 , @xmath161 since @xmath162 , evaluation of the densities in three dimensions can also be approached using the tools provided by @xcite .
similarly , we obtain , for instance , @xmath163 we note that theorem [ thm : pn : odd ] can be usefully implemented in a computer algebra system such as _ maple _ or _
mathematica_. by integrating , we are able to symbolically compute the corresponding moments , as introduced in , as @xmath164 which has a removable singularity @xmath165 and poles at @xmath166 and @xmath167 .
likewise , in five dimensions , @xmath168 with poles at @xmath169 and removable singularities at the other apparent poles .
thus , even this elementary evaluation has subtle structure .
theorem [ thm : pn : besseld ] allows us to prove a corresponding bessel integral representation for the _ moment function _
@xmath170 of the distance to the origin after @xmath0 random steps .
the following result generalizes ( * ? ? ?
* theorem 1 ) from two to arbitrary dimensions .
[ thm : w : bessel]let @xmath123 and @xmath22 . for any nonnegative integer @xmath124
, @xmath171 provided that @xmath172 .
using theorem [ thm : pn : besseld ] , we have @xmath173 interchanging the order of integration and substituting @xmath174 , we obtain @xmath175 \int_0^{\infty } z^{\nu +
s - k + 1 } j_{\nu + k } ( z ) { \mathrm{d}}z { \mathrm{d}}t.\ ] ] the inner integral may be evaluated using the standard bessel integral evaluation @xmath176 which holds for @xmath177 and @xmath4 such that @xmath178 and @xmath179 .
we conclude that @xmath180 \frac{2^{\nu + s - k + 1 } \gamma \left ( \frac{s}{2 } + \nu + 1 \right)}{\gamma \left ( k - \frac{s}{2 } \right ) } { \mathrm{d}}t , \end{aligned}\ ] ] which is the desired result . for the evaluation of the intermediate bessel integral
, we assumed @xmath181 and @xmath182 , and so holds for all @xmath77 in this non - empty strip provided that the original integral converges . using the asymptotic bounds from the proof of theorem [ thm : pn : besseld ]
, we note that the integral converges absolutely in the strip @xmath183 .
analytic continuation therefore implies that holds for all @xmath77 in this strip .
we deduce the following from , with @xmath122 .
this extends ( * ? ? ?
* proposition 2.4 ) .
in particular , we observe that , for @xmath184 , the first pole of @xmath185 occurs at @xmath186 .
[ cor : w : res]let @xmath187 . in the half - plane @xmath188 ,
the moment functions @xmath185 are analytic apart from simple poles at @xmath189 for integers @xmath82 such that @xmath190 .
the residues of these poles are @xmath191 by equation , with @xmath122 , we have @xmath192 valid for @xmath77 in the strip @xmath193 , in which the integral converges absolutely . in the region of interest , the only poles are contributed by the factor @xmath194 , which , as a function in @xmath77 , has simple poles at @xmath189 , for @xmath195 , with residue @xmath196 .
note that the value for the residue of @xmath197 at @xmath198 agrees with @xmath199 from provided that these values are finite .
[ eg : w3:poles]the moment functions @xmath200 are plotted in figure [ fig : w3 ] for @xmath201 .
the first pole of @xmath200 occurs at @xmath198 , is simple and has residue @xmath202 as follows from the bessel integral formula . in order to record some more general properties of the pole structure of @xmath203 , we use the fact that @xmath200 satisfies a functional equation , , which relates the three terms @xmath200 , @xmath204 and @xmath205 . by reversing this functional equation
, we find that the residues @xmath206 satisfy the recursion @xmath207 with @xmath208 and @xmath209 . observe that , when @xmath121 , the recursion for these residues is essentially the same as the recurrence for the corresponding even moments ( with @xmath210 replaced by @xmath211 ) . as recorded in (
* proposition 2.4 ) , this lead to @xmath212 define , likewise , the numbers @xmath213 by @xmath214 in analogy with , we find that @xmath215 solves the three - term recurrence @xmath216 with @xmath217 and @xmath218 . for small dimensions , initial values for @xmath213 are given by @xmath219 note the increasingly irregular behaviour of @xmath213 as @xmath3 increases . in particular , in dimension @xmath56 , we find @xmath220 , which signifies the disappearance of the perhaps expected pole of @xmath221 at @xmath222
. @xmath200 on @xmath223 $ ] for @xmath201.,title="fig:",width=170 ] @xmath200 on @xmath223 $ ] for @xmath201.,title="fig:",width=170 ] @xmath200 on @xmath223 $ ] for @xmath201.,title="fig:",width=170 ] [ eg : w4:poles]in the case @xmath224 and @xmath225 , that is @xmath226 , the moment function @xmath227 has a simple pole at @xmath228 with residue @xmath229 and is otherwise analytic in the half - plane @xmath230 . at @xmath231 , on the other hand , @xmath227 has a double pole . indeed , analyzing the functional equations that arise from and
, we derive that @xmath232 where in the last equality we used the known value from the planar case ( * ? ? ?
* example 4.3 ) .
similarly , we obtain @xmath233 in the higher - dimensional case , @xmath234 has poles at @xmath235 for @xmath236 , which are initially simple but turn into poles of order ( up to ) @xmath237 beginning at @xmath238 .
the approach indicated in examples [ eg : w3:poles ] and [ eg : w4:poles ] enables us to determine , at least in principle , the pole structure of the moment functions @xmath197 in each case .
we do not pursue such a more detailed analysis herein .
we next follow the approach of @xcite to obtain a summatory expression for the even moments from theorem [ thm : pn : besseld ] .
[ thm : w : even]the even moments of an @xmath0-step random walk in dimension @xmath3 are given by @xmath239 replacing @xmath124 by @xmath240 in and setting @xmath241 , we obtain @xmath242_{x = 0 } .
\label{eq : w : even : jn } \end{aligned}\ ] ] observe that , at the level of formal power series , we have @xmath243_{x = 0 } = k!a_k .\ ] ] recall from the series @xmath244 to conclude that @xmath245 which is equivalent to the claimed formula .
[ rk : j : d : k]proceeding as in the proof of theorem [ thm : w : even ] , we observe that @xmath246 and hence @xmath247 if applied to , this ( finite ) expansion , together with @xmath17 , makes the conclusion of theorem [ thm : w : even ] apparent . in conjunction with the asymptotics of @xmath248
, we conclude , as in the proof of theorem [ thm : pn : besseld ] , that the integrand in is @xmath249 as @xmath142 , so that each additional derivative improves the order at @xmath250 by @xmath2 at the expense of increasing the size of the coefficients .
we note , inter alia , that may , alternatively , be expressed as @xmath251 which yields a fine alternative generating function for the even moments .
[ eg : w : smallk]in the case @xmath252 , theorem [ thm : w : even ] immediately implies that the second moment of an @xmath0-step random walk in any dimension is @xmath253 this was proved in ( * ? ? ?
* theorem 4.2 ) using a multi - dimensional integral representation and hyper - spherical coordinates .
similarly , we find that @xmath254 more generally , theorem [ thm : w : even ] shows that @xmath8 is a polynomial of degree @xmath124 in @xmath0 , with coefficients that are rational functions in @xmath4 .
for instance , @xmath255 and so on .
using the explicit expression of the even moments of an @xmath0-step random walk in dimension @xmath3 , we derive the following convolution relation .
[ cor : w : iter ] for positive integers @xmath256 , half - integer @xmath4 and nonnegative integer @xmath124 we have @xmath257 note the special case @xmath258 , that is @xmath259 which allows us to relate the moments of an @xmath0-step walk to the moments of an @xmath260-step walk .
[ eg : w : values4d]corollary [ cor : w : iter ] provides an efficient way to compute even moments of random walks in any dimension . for illustration , and because they are integral , we record the moments of an @xmath0-step walk in two and four dimensions for @xmath261 .
@xmath262 for @xmath263 , these are central binomial coefficients , see , while , for @xmath264 , these are apry - like sequences , see and .
likewise , the initial even moments for four dimensions are as follows .
@xmath265 observe that the first terms are as determined in example [ eg : w : smallk ] .
in the two - step case in four dimensions , we find that the even moments are the _ catalan numbers _
@xmath266 , that is @xmath267 this adds another interpretation to the impressive array of quantities that are given by the catalan numbers .
it is a special property of the random walks in two and four dimensions that all even moments are positive integers ( compare , for instance , ) .
this is obvious for two dimensions from theorem [ thm : w : even ] which , in fact , demonstrates that the moments @xmath268 count _ abelian squares _
on the other hand , to show that the four - dimensional moments @xmath269 are always integral , it suffices to recursively apply and to note that the factors @xmath270 are integers for all nonnegative integers @xmath271 and @xmath124 .
the numbers are known as _
narayana numbers _ and occur in various counting problems ; see , for instance , ( * ? ? ?
* problem 6.36 ) .
[ eg : w : matpow ] the recursion for the moments @xmath8 is equivalent to the following interpretation of the moments as row sums of the @xmath0th power of certain triangular matrices . indeed , for given @xmath4 , let @xmath272 be the infinite lower triangular matrix with entries @xmath273 for row indices @xmath274 and column indices @xmath275 .
then the row sums of @xmath276 are given by the moments @xmath277 , @xmath278 .
for instance , in the case @xmath225 , @xmath279 with the row sums @xmath280 and @xmath281 corresponding to the moments @xmath282 and @xmath283 as given in example [ eg : w : values4d ] .
observe that , since the first column of @xmath284 is composed of @xmath2 s , the sequence of moments @xmath269 can also be directly read off from the first column of @xmath285 .
the matrix @xmath286 is known as the _ narayana triangle _ or the _ catalan triangle _
* ` a001263 ` ) .
let us note another point of view on the appearance of the narayana triangle in the context of random walks .
let @xmath27 be a random vector , which is uniformly distributed on the unit sphere in @xmath1 .
if @xmath287 is the angle between @xmath27 and a fixed axis , then @xmath288 has the probability density ( * ? ? ?
* ( 2 ) ) @xmath289 .\ ] ] denote with @xmath290 the random variable describing the distance to the origin after @xmath0 unit steps .
then @xmath291 is related to @xmath290 via ( * ? ? ?
* ( 9 ) ) @xmath292 writing @xmath293 $ ] for the expected value of a random variable @xmath294 , we therefore have @xmath295 = \mathbb{e } [ ( 1 + 2 \lambda r_n + r_n^2)^k ] .\ ] ] in terms of the generalized _ narayana polynomials _ @xmath296 , \label{eq : narayana : poly}\ ] ] which were introduced in ( * ? ? ? * ( 6.2 ) ) for @xmath297 , we obtain @xmath298 .\ ] ] this recurrence identity on the moments can then be expressed in matrix form by defining a matrix @xmath299 as in example [ eg : w : matpow ] .
we note , moreover , that expressing the narayana polynomials in terms of the gegenbauer polynomials @xmath300 , as demonstrated in ( * ? ? ?
* theorem 6.3 ) , we deduce the expression of the even moments as @xmath301,\ ] ] which is a variation of . to contrast with the integral even moments in 2 and 4 dimensions in example [ eg : w : values4d ] , we record a few initial even moments in 6 dimensions .
@xmath302 it may be concluded from , with @xmath303 , that the entries of the matrix @xmath304 satisfy @xmath305 .
this implies that the even moments in dimension 6 are rational numbers whose denominators are powers of @xmath306 .
similar observations apply in all dimensions but we do not pursue this theme further here .
it follows from theorem [ thm : w : even ] that the general expression of the even moments for a @xmath237-step walk in @xmath3 dimensions is given by @xmath307 note that in the special case of @xmath121 , that is , dimension @xmath237 , this clearly reduces to the central binomial coefficient . in dimension
@xmath54 , as noted in example [ eg : w : values4d ] , the even moments are the catalan numbers .
we note that the generating function for the two - step even moments is @xmath308 which reduces to the known generating functions for @xmath309 .
equation is an immediate consequence of rewriting as @xmath310 .
in fact , also holds true when @xmath124 takes complex values .
this was proved in ( * ? ? ?
* theorem 4.3 ) using a multi - dimensional integral representation and hyper - spherical coordinates .
we offer an alternative proof based on theorem [ thm : w : bessel ] .
[ thm : w2]for all complex @xmath77 and half - integer @xmath13 , @xmath311 the case @xmath122 of in theorem [ thm : w : bessel ] gives @xmath192 provided that @xmath312 . using that , for @xmath313 , @xmath314 the claimed formula then follows from the duplication formula for the gamma function and analytic continuation .
the nonnegative even moments for a @xmath306-step walk in @xmath3 dimensions are @xmath315 we apply corollary [ cor : w : iter ] with @xmath316 and @xmath258 .
using that @xmath317 and that an evaluation of @xmath318 is given by , we obtain @xmath319 expressing the factorials as binomial coefficients yields . in the case
@xmath87 , or @xmath121 , the moments of a @xmath306-step walk reduce to the apry - like numbers @xmath320 ( * ? ? ? * ( 3.2 ) & ( 3.4 ) ) show that the generating function for this sequence is @xmath321 [ eg : w3:4d]in the case @xmath226 , or @xmath225 , the moments , whose initial values are recorded in example [ eg : w : values4d ] , are sequence ( * ? ? ?
* ` a103370 ` ) .
the oeis also records a hypergeometric form of the generating function ( as the linear combination of a hypergeometric function and its derivative ) , added by mark van hoeij . on using linear transformations of hypergeometric functions , we have more simply that @xmath322 which we are able to generalize .
example [ eg : w3:4d ] suggests that a nice formula for the generating function for the moments @xmath323 exists for all even dimensions .
indeed , we have the following result .
[ thm : w3:gf]for integers @xmath13 and @xmath324 , we have @xmath325 where @xmath326 is a polynomial ( that is , @xmath327 is the principal part of the hypergeometric term on the right - hand side ) . for integers @xmath13 ,
define the rational numbers @xmath328 by @xmath329 for @xmath330 , and @xmath331 for @xmath332 .
writing the sum for @xmath323 in hypergeometric form , we obtain , for integers @xmath333 , the representation @xmath334 in order to prove the claimed generating function it therefore suffices to show the ( more precise ) claim @xmath335 for instance , this predicts that @xmath336 .
taking the real part is only necessary for @xmath337 while , for @xmath338 , the @xmath339 is terminating .
we note , as will be demonstrated later in the proof , that the right - hand side of vanishes for @xmath332 , that is , the @xmath339 takes purely imaginary values then . the holonomic systems approach @xcite , implemented in the _ mathematica _ package ` holonomicfunctions ` , which accompanies koutschan s thesis @xcite
, shows that the coefficients @xmath328 , defined by , satisfy the recursive relation @xmath340 for all integers @xmath124 and all integers @xmath13 .
we already know that holds for @xmath121 and @xmath341 .
verifying , by using @xmath342 and letting @xmath343 to see that the real part vanishes ( for any choice of analytic continuation to @xmath41 ) , that holds for @xmath121 and @xmath344 , we conclude from that is indeed true for @xmath121 and all integers @xmath124 . as in the case of
, we find that the coefficients @xmath345 further satisfy the dimensional relations @xmath346 and @xmath347 for all integers @xmath124 and all integers @xmath348 .
the three relations , , , together with the case @xmath121 as boundary values , completely determine the coefficients @xmath328 for all integers @xmath124 and @xmath348 .
since we already verified the case @xmath121 , it only remains to demonstrate that the right - hand side of satisfies the same recursive relations .
another application of ` holonomicfunctions ` finds that the @xmath339 on right - hand side of , and hence its real part , indeed satisfy , , for the required ( real ) values of @xmath4 and @xmath124 . for the convenience of the reader , and because we will frequently use it in the following , we state carlson s theorem next @xcite .
recall that a function @xmath349 is of _ exponential type _ in a region if @xmath350 for some constants @xmath351 and @xmath352 .
[ thm : carlson]let @xmath353 be analytic in the right half - plane @xmath354 and of exponential type with the additional requirement that @xmath355 for some @xmath356 on the imaginary axis @xmath357 .
if @xmath358 for @xmath274 , then @xmath359 identically . [ eg : rec3]applying creative telescoping to the binomial sum , we derive that the moments @xmath323 satisfy the recursion @xmath360 observe that @xmath197 is analytic for @xmath361 and is bounded in that half - plane by @xmath362 ( because , in any dimension , the distance after @xmath0 random steps is bounded by @xmath0 ) .
it therefore follows from carlson s theorem , as detailed in ( * ? ? ?
* theorem 4 ) , that the recursion remains valid for complex values of @xmath124 .
the next result expresses the complex moments @xmath363 in terms of a meijer @xmath364-function and extends ( * ? ? ?
* theorem 2.7 ) .
[ thm : w3:meijer]for all complex @xmath77 and dimensions @xmath22 , @xmath365 if @xmath366 , then @xmath367 this simple integral is a consequence of the fact that @xmath368 combined with theorem [ thm : w : bessel ] with @xmath369 and @xmath122 .
similarly , for @xmath370 , the mellin transform of @xmath371 is given by .
applying parseval s formula to these two mellin transforms , we obtain , for @xmath372 , @xmath373 the claim then follows , by analytic continuation , from theorem [ thm : w : bessel ] with @xmath10 and @xmath122 .
we note that , as in @xcite , this meijer @xmath364-function expression can be expressed as a sum of hypergeometric functions by slater s theorem @xcite .
this is made explicit in .
equation gives a hypergeometric expression for the even moments of a @xmath306-step random walk .
it was noticed in @xcite that , in the case of planar walks , the _ real part _ of this hypergeometric expression provides an evaluation of the odd moments .
these odd moments are much harder to obtain , and it was first proved in @xcite , based on this observation , that the average distance of a planar @xmath306-step random walk is @xmath374 where @xmath375 in the sequel , we generalize these results from two to arbitrary even dimensions . in particular , as explained in example [ eg : w3:odd ] , we establish the transcendental nature of the odd moments of @xmath306-step walks in all even dimensions by showing that they are all rational linear combinations of @xmath376 and @xmath377 .
[ thm : w3:re]suppose that @xmath3 is even , that is , @xmath4 is an integer . for all odd integers
@xmath378 , @xmath379 the case @xmath121 is proved in ( * ? ? ?
* theorem 6 ) .
we will prove the general case by induction on @xmath4 .
it is routine to verify that the hypergeometric function @xmath380 which , for even @xmath77 , agrees with @xmath200 by , satisfies the contiguity relation @xmath381 on the other hand , it follows from theorem [ thm : w3:meijer ] and slater s theorem @xcite that @xmath382 in that form , it is again a routine application of the holonomic systems approach @xcite to derive that @xmath383 since this relation matches the relation satisfied by @xmath384 , and hence @xmath385 when @xmath4 and @xmath77 are real , the general case follows inductively from the base case @xmath121 .
[ rk : w3:meijer : hyp]the coefficients of the hypergeometric functions in can be expressed as @xmath386 and @xmath387 , respectively . here
, the factor @xmath388 is @xmath389 for half - integers @xmath4 , and @xmath390 for integers @xmath4 .
[ eg : w3:odd]the planar case of theorem [ thm : w3:re ] was used in @xcite to prove that the average distance to the origin after three random steps in the plane is given by .
it is a consequence of , extended to complex @xmath124 , that all planar odd moments are @xmath391-linear combinations of @xmath392 , defined in , and @xmath393 .
the dimensional recursion used in the proof of theorem [ thm : w3:re ] shows that this observation extends to all even dimensions .
for instance , @xmath394 moreover , the average distance to the origin after three random steps in four dimensions is @xmath395 with similar evaluations in six or higher even dimensions
. theorem [ thm : w3:re ] does not hold in odd dimensions , in which the involved quantities can be evaluated in elementary terms .
for instance , in the case of dimension @xmath306 , @xmath396 while for integer @xmath397 @xmath398 only agrees for even @xmath77 .
[ eg : w3:der]by differentiating the hypergeometric representation of @xmath363 in , it was shown in ( * ? ? ?
* examples 6.2 and 6.6 ) that @xmath399 where the derivatives are with respect to @xmath77 .
it follows from differentiating , extended to complex @xmath124 , that all derivatives @xmath400 lie in the @xmath391-linear span of @xmath2 , @xmath401 and @xmath402 .
then , differentiating , we find that , indeed , for all integers @xmath13 , the derivatives @xmath403 can likewise be expressed as @xmath404 with rational numbers @xmath405 .
moreover , in the case of @xmath406 , a slightly more careful analysis reveals that @xmath407 .
while we omit the details , we note that this can be seen , for instance , by evaluating @xmath408 and then deriving , in analogy with , a functional equation relating @xmath409 , @xmath410 and @xmath411 . in four and six dimensions ,
we obtain , for example , @xmath412 a special motivation for considering these derivative values is that , in the case of two dimensions , @xmath413 is the ( logarithmic ) mahler measure of the multivariate polynomial @xmath414 ; see ( * ? ? ?
* example 6.6 ) or ( * ? ? ?
* section 4 ) .
the second derivative @xmath415 , interpreted there as a higher mahler measure , is evaluated in ( * ? ? ?
* theorem 4.4 ) in the form @xmath416 where @xmath417 denotes the @xmath0th _ log - sine integral _
@xmath418 for alternative expressions of the log - sine integral in in terms of other polylogarithmic constants , we refer to @xcite .
we have not been able to obtain an equally natural log - sine evaluation of @xmath419 for integers @xmath420 .
we may , however , derive @xmath421 by the methods of @xcite , we arrive at @xmath422 we may now integrate by parts and obtain @xmath423 moreover , @xmath424 where @xmath425 is given by ( * ? ? ?
* ` a025565 ` ) of the oeis and counts the number of number of udu - free paths of @xmath154 upsteps ( u ) and @xmath154 downsteps ( d ) " with recursion @xmath426 ( alternatively , @xmath427 with @xmath428 the _ motzkin numbers _ given in ` a001006 ` . )
solving for the generating function @xmath429 of @xmath425 and considering @xmath430 we finally arrive at @xmath431 as in example [ eg : w3:der ] , all second derivatives @xmath432 in even dimensions may then be expressed in terms of @xmath433 and @xmath434 as well as the constants in example [ eg : w3:der ] .
for instance , @xmath435 and so on .
the nonnegative even moments for a @xmath54-step walk in @xmath3 dimensions are @xmath436 we apply corollary [ cor : w : iter ] with @xmath316 and @xmath437 , to obtain @xmath438 using the evaluation of @xmath318 given by then yields .
[ eg : domb ] the binomial sums in generalize the _
domb numbers _ , also known as the _ diamond lattice _ numbers ( * ? ? ?
* ` a002895 ` ) , @xmath439 for @xmath440 , which have played an important role in dimension @xmath237 .
their ordinary generating function , @xmath441 was determined in @xcite .
the final equation follows from clausen s product formula . in four dimensions , the recursive relation combined with yields @xmath442 where @xmath266 are
the catalan numbers , as in , and @xmath443 are the narayana numbers , as in example [ eg : domb ] .
after developing some further properties of the moments , we illustrate in example [ eg : w4:ogf ] that the ordinary generating function for the even moments @xmath444 can be expressed in terms of hypergeometric functions whenever the dimension is even . for @xmath440
, we have @xmath445 we note that this hypergeometric function is _ well - poised _ @xcite . applying creative telescoping to the binomial sum , we derive that the moments @xmath446 satisfy the recursion @xmath447 as in example [ eg : rec3 ] , @xmath234 is analytic , exponentially bounded for @xmath361 and bounded on vertical lines .
hence , carlson s theorem [ thm : carlson ] applies to show that the recursion extends to complex @xmath124 .
the following result is the counterpart of theorem [ thm : w3:meijer ] and extends ( * ? ? ?
* theorem 2.8 ) .
[ thm : w4:meijer]for all complex @xmath77 with @xmath448 and dimensions @xmath22 , @xmath449 the proof is obtained along the lines of the proof of theorem [ thm : w3:meijer ] .
this time , parseval s formula is applied to the product @xmath450 .
we note that , as in the case of for three steps , the meijer @xmath364-function in theorem [ thm : w4:meijer ] can be expressed as a sum of hypergeometric functions , namely @xmath451 by slater s theorem @xcite .
the coefficients @xmath452 and @xmath453 of the two hypergeometric functions are @xmath454 where the coefficients @xmath455 and @xmath456 are as in the three - step case , see remark [ rk : w3:meijer : hyp ] .
it was shown in ( * ? ? ?
* section 2.3 & 3.1 ) that the average distance to the origin after four random steps in the plane , as well as all its odd moments , can be evaluated in terms of the elliptic integrals @xmath457 we note that each of the @xmath458 hypergeometric functions may alternatively be expressed as the sum of two @xmath459 hypergeometric functions .
then , in two dimensions , @xmath460 and it follows from , extended to complex values , that all odd moments are indeed rational linear combinations of @xmath376 and @xmath461 . in order to generalize this observation to higher dimensions , we claim that @xmath462 this counterpart of can be proved using the holonomic systems approach @xcite applied to the hypergeometric form .
we conclude that , in any even dimension , the odd moments lie in the @xmath391-span of the constants @xmath376 and @xmath461 , which arose in the planar case . for instance , we find that the average distance after four random steps in four dimensions is @xmath463 and so on . in continuation of example [ eg : w3:der ] , we recall from ( * ? ? ? * examples 6.2 and 6.6 ) that @xmath464 where , again , the derivatives are with respect to @xmath77 .
these evaluations may , for instance , be obtained from differentiating the hypergeometric expression .
proceeding as in example [ eg : w3:der ] , we differentiate both , extended to complex values , and , we conclude that , for all integers @xmath333 , @xmath465 with rational numbers @xmath405 . again , we find that @xmath407 in the case @xmath122 . in four and six dimensions ,
we obtain , for example , @xmath466 in analogy with the case of three - step walks , the derivative @xmath467 is particularly interesting because it is the mahler measure of the multivariate polynomial @xmath468 . in ( * ? ? ?
* ( 6.5 ) , ( 6.10 ) ) evaluations for the second derivatives @xmath469 and @xmath470 are given in terms of polylogarithmic constants .
it follows from these evaluations and the functional equation that the derivative values @xmath471 all lie in the @xmath391-linear span of @xmath472 where @xmath473 is the polylogarithm of order @xmath0 .
indeed , we realize from the dimensional recursion that the same is true for the values @xmath474 in all even dimensions . for example , @xmath475 the number of ( presumed ) transcendental constants can be somewhat reduced when working in terms of kummer - type polylogarithms , as illustrated in ( * ? ? ?
* theorem 4.7 ) .
[ eg : w4:ogf]in we noted that the ordinary generating function of the moments @xmath476 has a concise hypergeometric expression .
it is natural to wonder if this result extends to higher dimensions .
combining the recursive relations and , we are able to derive that the ordinary generating function of the moments @xmath444 , when complemented with an appropriate principal part ( as in theorem [ thm : w3:gf ] ) , can be obtained from the corresponding generating function of @xmath477 as well as its first two derivatives . because the precise relationship is not too pleasant , we only record the simplified generating function , @xmath478,\ ] ] that we obtain in dimension @xmath54 . here ,
@xmath479 which , by the differentiation formula ( * ? ? ?
* ( 16.3.1 ) ) @xmath480 can be expressed as generalized hypergeometric functions .
it would be nice if it was possible to make the general case as explicit as we did in theorem [ thm : w3:gf ] for three - step walks , but we have not succeeded in doing so . as in the planar case , as well as in many related problems ,
it is much harder to obtain explicit results in the case of five or more steps .
this is reflected , for instance , in the fact that an application of corollary [ cor : w : iter ] with @xmath482 and @xmath437 , and appealing to , results in a double ( and not single ) sum of hypergeometric terms .
the nonnegative even moments for a @xmath481-step walk in @xmath3 dimensions are @xmath483 [ eg : w5:rec]as in the case of three and four steps , we can apply creative telescoping to the binomial sum to derive a recursion for the moments @xmath484 .
in contrast to the three - term recursions and , we now obtain a four - term recursion , namely @xmath485 where @xmath486 as in example [ eg : rec3 ] , @xmath487 is analytic and suitably bounded for @xmath361 , so that we may conclude from carlson s theorem [ thm : carlson ] that the recursion extends to complex @xmath124 .
[ eg : w5:rec : dim]creative telescoping , applied to the binomial sum for @xmath488 , allows us to derive the following more involved counterpart of the dimensional recursions , for three and four steps .
@xmath489 where @xmath490 this recursion is first obtained for nonnegative even integers @xmath77 , and then extended to complex values using carlson s theorem [ thm : carlson ] .
we find an explicit formula for the probability density @xmath491 of the distance to the origin after two random steps in @xmath1 by computing the bessel integral .
an equivalent formula is given in ( * ? ? ?
* corollary 4.2 ) , which exploited the fact that the probability density is essentially the inverse mellin transform of the moments which are evaluated in theorem [ thm : w2 ] .
the probability density function of the distance to the origin in @xmath492 dimensions after @xmath237 steps is , for @xmath493 , @xmath494 it follows from that @xmath495 from ( * ? ? ?
* chapter 13.46 ) , we have the integral evaluation @xmath496 assuming that @xmath497 and that @xmath498 are the sides of a triangle of area @xmath499 . in our case , @xmath500 and therefore @xmath501 which is equivalent to .
note that reflects the general fact , discussed in section [ ssec : odd ] , that the densities @xmath6 are piecewise polynomial in odd dimensions . for comparison with the case of @xmath306 steps , we record that the density @xmath502 satisfies the following functional equation
: if @xmath503 , then @xmath504 [ rk : p2:fun]the probability density in two dimensions , that is @xmath505 is readily identified as the distribution of @xmath506 with @xmath287 uniformly distributed on @xmath507 $ ] . in other words ,
if @xmath508 is uniformly distributed on the sphere of radius @xmath237 in @xmath509 , then @xmath510 describes the distribution of @xmath511 .
it is then natural to wonder whether this observation extends to higher dimensions . indeed , if @xmath512 is uniformly distributed on the sphere of radius @xmath237 in @xmath513 , where @xmath27 is partitioned into @xmath514 , then @xmath491 describes the distribution of @xmath511 .
details are left to the interested reader . in this stochastic interpretation ,
the invariance of under @xmath515 is a reflection of the fact that @xmath511 and @xmath516 share the same distribution .
it was shown in @xcite that the density @xmath517 of the distance to the origin after three random steps in the plane has the closed form @xmath518 valid on the interval @xmath519 .
we next generalize this hypergeometric expression to arbitrary dimensions . in order to do so
, we need to establish the behaviour of @xmath520 at the end points . to begin with ,
we use information on the pole structure of the moments @xmath521 to deduce the asymptotic behaviour of @xmath520 as @xmath522 . [
prop : p3:asy0 ] for positive half - integer @xmath4 and @xmath523 , @xmath524 as shown in corollary [ cor : w : res ] , the moments @xmath363 are analytic for @xmath525 and the first pole has residue @xmath526 where the last equality is another special case of .
the asymptotic behaviour then follows because @xmath527 is the mellin transform of @xmath11 . on the other hand , to obtain the asymptotic behaviour of @xmath520 as @xmath528 , we have to work a bit harder than in the case of the behaviour as @xmath523 . [
prop : p3:asy3 ] for positive half - integer @xmath4 as @xmath528 , @xmath529 using theorem [ thm : pn : rec ] together with the fact that , for @xmath530 $ ] , @xmath531 we find @xmath532 observe that the upper bound of integration is @xmath2 if @xmath533 $ ] , and @xmath534 for @xmath535 $ ] . on @xmath536
$ ] , after substituting @xmath537 , we therefore find @xmath538 note that the polynomial in the integrand factors into @xmath539 times a factor which approaches @xmath540 as @xmath528 and @xmath541 .
hence , as @xmath528 , @xmath542 we now relate this integral to the incomplete beta function to find @xmath543 putting these together , we conclude that holds .
we are now in a position to generalize to higher dimensions . for any half - integer @xmath13 and @xmath544 $ ] ,
we have @xmath545 we observe that both sides of satisfy the differential equation @xmath546 , where the differential operator @xmath547 is given by @xmath548 where @xmath549 .
note that , in the case of @xmath550 , this is a consequence of the functional equation resulting from .
the indices of this differential equation at @xmath551 are @xmath130 and @xmath552 .
it follows from proposition [ prop : p3:asy0 ] that @xmath553 is the unique solution @xmath554 , on @xmath555 , such that @xmath556 as @xmath523 .
since this property is satisfied by the right - hand side of as well , it follows that holds for all @xmath557 .
similarly , to show for all @xmath558 , we use the fact that the differential equation has indices @xmath130 and @xmath552 at @xmath559 as well . in light of proposition [ prop : p3:asy3 ] , @xmath560 is the unique solution @xmath554 , on @xmath561 , such that , as @xmath562 , @xmath563 again , it is routine to verify that this property is also satisfied by the right - hand side of .
hence , it follows that holds for all @xmath558 . as a consequence
, we have the following functional equation for the probability density function @xmath520 . the role of the involution @xmath564 for @xmath237 steps ,
see , is now played by the involution @xmath565 .
[ cor : fe3 ] for any half - integer @xmath13 and @xmath544 $ ] , the function @xmath566 satisfies the functional equation @xmath567 the hypergeometric right - hand side of clearly satisfies the functional equation
. it would be very interesting to have a probabilistic interpretation of this functional equation satisfied by the densities @xmath520 .
we note the following relation between the functional equations for the two- and three - step case in and , respectively .
namely , if @xmath568 then the relation @xmath569 translates into @xmath570 .
it is natural to wonder whether this observation might help explain the functional equation for three - step densities .
[ eg : p3:dimrec]as a consequence of the contiguity relations satisfied by hypergeometric functions , we derive from that the densities @xmath520 satisfy the dimensional recursion @xmath571 where @xmath348 .
it follows from the hypergeometric formula that the densities @xmath520 , for @xmath109 , take the special values @xmath572 in particular , @xmath573 in odd dimensions , and @xmath574 in even dimensions .
[ fe : diff ] from and the functional equation of corollary [ cor : fe3 ] , for @xmath575 , we learn that @xmath576 but learn nothing about @xmath577 .
it is shown in @xcite that , in the planar case , the probability density of the distance to the origin after four steps admits the hypergeometric closed form @xmath578 in this section we obtain a higher - dimensional analog of this evaluation by demonstrating that hypergeometric formulae can be given for the densities @xmath579 in all even dimensions .
since the mellin transform of the density @xmath580 is given by the corresponding probability moments @xmath581 , the recursion for the moments translates into a differential equation for the density .
we refer to @xcite for details .
we thus find that @xmath582 is annihilated by the differential operator @xmath583 where @xmath549 .
we conclude that , in the planar case , @xmath96 satisfies the differential equation @xmath584 which agrees ( up to a typo there ) with ( * ? ? ?
* ( 2.7 ) ) .
the differential equation is the symmetric square of a second order differential equation and , moreover , admits modular parametrization ( * ? ? ?
* remark 4.11 ) .
these ingredients ultimately lead to the hypergeometric closed form . the differential equation associated with , on the other hand , is a symmetric square of a second order differential equation only in the cases @xmath121 and @xmath585 . in three dimensions , that is when @xmath585 , the density is @xmath586 for @xmath587 $ ] .
this is equivalent to and may be derived directly from or from theorem [ thm : pn : odd ] .
basic mellin calculus connects the asymptotic behaviour of @xmath81 as @xmath523 with the nature of the poles of @xmath197 in the left half - plane .
for instance , from the explicit information in example [ eg : w4:poles ] on the poles of @xmath227 at @xmath228 and @xmath588 , we conclude that @xmath589 as @xmath523 .
the following result connects the @xmath54-step density in @xmath3 dimensions with the corresponding density in @xmath590 dimensions .
in particular , using and as base cases , this provides a way to obtain explicit formulas for the densities @xmath44 in all dimensions .
[ thm : p4:rec]for @xmath591 and any half - integer @xmath13 , @xmath592 where @xmath593 we have already observed in and that the moments @xmath594 satisfy a functional equation connecting @xmath595 , @xmath596 , @xmath597 , as well as a functional equation relating @xmath598 , @xmath596 , @xmath599 .
the usual mellin calculus translates the first of these two into the third - order differential equation recorded in , while the second is translated into a more complicated equation involving seven terms ranging from @xmath44 to @xmath600 . with assistance of the _ mathematica _ package ` holonomicfunctions ` , which accompanies koutschan s thesis @xcite , we compute a grbner basis for the ideal that these two relations generate and use it to find the claimed relation involving @xmath601 as well as @xmath602 and its first two derivatives .
[ eg : p41 ] by combining with theorem [ thm : p4:rec ] and , we conclude , for instance , that the @xmath54-step density in four dimensions , for @xmath603 , can be hypergeometrically represented as @xmath604 , \label{eq : p41}\ ] ] with @xmath605 and @xmath606 as well as @xmath607 as in , taking the real part of the hypergeometric functions provides a formula for @xmath608 which is valid for @xmath609 as well .
the above hypergeometric formula also provides a modular parametrization ( in a suitably generalized sense ) for @xmath608 .
this is a consequence of @xmath610 which is proved in ( * ? ? ?
* ( 4.16 ) ) and which involves the modular function @xmath611 and the weight @xmath237 modular form @xmath612 differentiating this modular parametrization of @xmath613 , we find that @xmath614 can be expressed in terms of modular quantities such as @xmath615 .
then inductively we obtain the like result for higher even dimensions . motivated by ( * ?
* corollary 4.8 ) , which proves that @xmath616 or @xmath617 , we discover that , for integers @xmath13 , @xmath618 is a rational combination of the moments @xmath619 and @xmath620 .
for instance , @xmath621 and @xmath622 to deduce the first of these from equation takes a little care as @xmath623 is not differentiable at @xmath237 and one must take the limit from the left in ; likewise for higher dimensions ( note that @xmath624 ) .
theorem [ thm : p4:rec ] specializes to @xmath625 .
\end{aligned}\ ] ] on the other hand , specializing a corresponding relation among @xmath626 , @xmath601 , @xmath627 and @xmath628 , which one obtains as in the proof of theorem [ thm : p4:rec ] , we find @xmath629 the @xmath481-step densities for dimensions up to @xmath630 are plotted in figure [ fig : p5 ] .
@xmath631 for @xmath43,width=340 ] a peculiar feature of the planar density is its striking ( approximate ) linearity on the initial interval @xmath632 $ ] .
this phenomenon was already observed by pearson @xcite , who commented on @xmath633 , between @xmath551 and @xmath41 , as follows : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` the graphical construction , however carefully reinvestigated , did not permit of our considering the curve to be anything but a _ straight
_ line@xmath634 even if it is not absolutely true , it exemplifies the extraordinary power of such integrals of @xmath635 products [ that is , ] to give extremely close approximations to such simple forms as horizontal lines . ''
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ pearson s comment was revisited by fettis @xcite , who rigorously established the nonlinearity . in (
* theorem 5.2 ) , it is shown that the density satisfies a certain fourth - order differential equation , and that , for small @xmath24 , @xmath636 where @xmath637 with initial conditions @xmath638 and @xmath639 numerically , we thus find that @xmath640 which reflects the approximate linearity of @xmath97 for small @xmath641 . by
, the residue @xmath642 can be expressed as @xmath643 .
the modularity of @xmath644 in the planar case , combined with the _ chowla selberg formula _
@xcite , then permits us to obtain the explicit formula ( * ? ? ?
* theorem 5.1 ) @xmath645 moreover , high - precision numerical calculations lead to the conjectural evaluation ( * ? ? ? * ( 5.3 ) ) @xmath646 and the recursion then implies that all the coefficients @xmath647 in can be expressed in terms of @xmath648 .
[ thm : r51 ] the conjectural relation is true .
as noted after corollary [ cor : w : res ] , we have @xmath649 on the other hand , applying the dimensional recursion for @xmath650 with @xmath225 and using the values @xmath651 and @xmath652 , we obtain @xmath653 since @xmath654 has simple poles only ( * ? ? ?
* example 2.5 ) . as noted in ( * ? ? ? * ( 6.2 ) ) , it is a consequence of the functional equation that @xmath655 combining the last three equations , we arrive at @xmath656 equation is therefore equivalent to @xmath657 the equality can now be deduced ( at least in principle ) from the hypergeometric formula for @xmath608 , made explicit in example [ eg : p41 ] , the modular parametrization of @xmath613 as well as the chowla selberg formula @xcite . in conclusion , we know that @xmath97 has a taylor expansion at @xmath551 , which converges and gives its values in the interval @xmath632 $ ]
. moreover , we have a recursive description of the taylor coefficients and know that they are all @xmath391-linear combinations of @xmath648 in and @xmath658 .
all of these statements carry over to the 5-step densities @xmath631 in any even dimension .
since the details are unwieldy , we only sketch why this is so .
recall that the moments @xmath650 satisfy the recursive relations and .
indeed , there is a third relation which connects @xmath650 , @xmath659 , @xmath660 , @xmath661 . as in the proof of theorem [ thm : p4:rec ]
, the mellin transform translates these three recursive relations into ( complicated ) differential relations for the densities @xmath631 .
assisted , once more , by koutschan s package ` holonomicfunctions ` @xcite , we compute a grbner basis for the ideal that these three differential relations generate . from there
, we find that there exists , in analogy with theorem [ thm : p4:rec ] , a relation @xmath662 with @xmath663 polynomials of degrees @xmath664 in @xmath641 ( with coefficients that are rational functions in @xmath4 ) .
we therefore conclude inductively that , for integers @xmath4 , the density @xmath631 has a taylor expansion at @xmath551 whose taylor coefficients lie in the @xmath391-span of @xmath648 in and @xmath658 .
it remains an open challenge , including in the planar case , to obtain a more explicit description of @xmath631 .
we have shown that quite delicate results are possible for densities and moments of walks in arbitrary dimensions , especially for two , three and four steps .
we find it interesting that induction between dimensions provided methods to show theorem [ thm : r51 ] , a result in the plane that we previously could not establish @xcite .
we also should emphasize the crucial role played by computer experimentation and by computer algebra .
one stumbling block is that currently _
mathematica _ , and to a lesser degree _ maple _ , struggle with computing various of the bessel integrals to more than a few digits thus requiring considerable extra computational effort or ingenuity . *
the even moments @xmath665 associated to a random walk in two dimensions have combinatorial significance .
they count _ abelian squares _
@xcite of length @xmath666 over an alphabet with @xmath0 letters ( i.e. , strings @xmath667 of length @xmath666 from an alphabet with @xmath0 letters such that @xmath668 is a permutation of @xmath641 ) .
as observed in example [ eg : w : values4d ] , the even moments @xmath269 are positive integers as well and we have expressed them in terms of powers of the narayana triangular matrix , whose entries count certain lattice paths . does that give rise to an interpretation of the even four - dimensional moments themselves counting similar combinatorially interesting objects ? * as discussed in example [ eg : domb ] , in the case @xmath121 , the moments @xmath446 are the domb numbers , for which a clean hypergeometric generating function is known .
referring to example [ eg : w4:ogf ] , we wonder if it is possible to give a compact explicit hypergeometric expression for the generating function of the even moments @xmath444 , valid for all even dimensions , as we did in theorem [ thm : w3:gf ] for three - step walks .
* verrill has exhibited @xcite an explicit recursion in @xmath124 of the even moments @xmath665 in the plane .
combined with a result of djakov and mityagin @xcite , proved more directly and combinatorially by zagier ( * ? ? ? * appendix a ) , these recursions yielded insight into the general structure of the densities @xmath669 .
for instance , as shown in ( * ? ? ? * theorem 2.4 ) , it follows that these densities are real analytic except at @xmath130 and the positive integers @xmath670 it would be interesting to obtain similar results for any dimension .
* by exhibiting recursions relating different dimensions , we have shown that the odd moments of the distances after three and four random steps in any dimension can all be expressed in terms of the constants arising in the planar case .
is it possible to evaluate these odd moments in a closed ( hypergeometric ) form which reflects this observation ? * in the plane , various other fragmentary modular results are ( conjecturally ) known for five and six step walks , see ( * ? ? ?
* ( 6.11 ) , ( 6.12 ) ) for representations of @xmath671 and @xmath672 , conjectured by rodriguez - villegas @xcite , as well as a discussion of their relation to _
mahler measures_. are more comprehensive results possible ?
the third author would like to thank j. m. borwein for his invitation to visit the carma center in september 2014 ; this research was initiated at this occasion .
thanks are due to ghislain mckay and corey sinnamon who explored the even moments during visiting student fellowships at carma in early 2015 .
d. boyd , d. lind , f. rodriguez villegas , and c. deninger . the many aspects of mahler s measure .
final report of 2003 banff workshop .
available at http://www.birs.ca/workshops/2003/03w5035/report03w5035.pdf , 2003 . | we study arithmetic properties of short uniform random walks in arbitrary dimensions , with a focus on explicit ( hypergeometric ) evaluations of the moment functions and probability densities in the case of up to five steps . somewhat to our surprise , we are able to provide complete extensions to arbitrary dimensions for most of the central results known in the two - dimensional case .
oldtitletitle titleoldtitle @oldtitletitle title@oldtitle |
supernovae ( sne ) , hypernovae and a few detected sne associated with long gamma - ray bursts ( grbs ) serve as important cornerstones of several major branches in astrophysics and cosmology .
physical mechanisms and outcomes for such violent explosions of massive stars have been actively pursued for decades .
hydrodynamics and magnetohydrodynamics ( mhd ) together with simplifying approximations and increasingly sophisticated microphysics have been invoked to model various aspects of sne in both analytic treatments and numerical simulations .
as nuclear fuels eventually become insufficient in the stellar core , the process of core - collapse sne signaling the demise of massive progenitors ( e.g. red and blue giants ) may be conceptually divided into three stages of core collapse , rebound shock and neutrino heating ( e.g. burrows et al . 1995 ; janka & mller 1996 ) .
the fortuitous detection of neutrinos from sn1987a ( hirata et al .
1987 ; bionta et al .
1987 ; koshiba 2009 private communications ) , bolsters such a scenario framework in part or as a whole .
optical observations before sn1987a revealed its progenitor as a blue giant star in a mass range of @xmath9 ( e.g. arnett et al . 1989 ) . at the time of sn1987a explosion , twenty neutrinos in the energy range of @xmath10 mev were intercepted within @xmath11 s , confirming the occurrence of neutronization .
the timescale of neutrino emissions was consistent with the prediction for neutrino trapping inside an extremely dense collapsed core .
the total neutrino flux was consistent with energetic neutrinos carrying off the binding energy during the core neutronization ( chevalier 2009 private communications ) , even though no signals of a neutron star ( e.g. a pulsar ) or a black hole are detected ( mccray 2009 private communications ) . in spite of extensive research on analytic and numerical studies of sne over several decades ( e.g. goldreich & weber 1980 gw hearafter ; yahil 1983 ; bruenn 1985 ; bruenn 1989a , b ; herant et al .
1995 ; janka & m@xmath12ller 1995 , 1996 ; fryer & warren 2002 , 2004 ; blondin , mezzacappa & demarino 2003 ; blondin & mezzacappa 2006 ; burrows et al .
2006 , 2007a , b ; lou & wang 2006 , 2007 ; wang & lou 2007 , 2008 ; lou & cao 2008 ; hu & lou 2009 ) , a few major issues in the sn model development remain to be explored ( see burrows et al .
2007 for a recent review ) . among these theoretical challenges ,
the dynamics of core - collapse stage inside the progenitor and the possibility of convective instabilities during this phase are the main thrust of this paper .
the simplest hydrodynamic model to describe a core collapse is a one - dimensional radial contraction with spherical symmetry under the self - gravity . in analytical model analyses ,
approximations to the equation of state ( eos ) for gas medium are necessarily introduced .
it can be shown in statistical mechanics ( e.g. huang 1987 ) that a relativistic hot fermi gas with a temperature much lower than the fermi energy^{1/3}c=30(\rho_{14}y_e)^{1/3}$ ] mev where @xmath13 is the number of electrons per baryon , @xmath14 is the planck constant , @xmath15 is the mass density , @xmath16 is the proton mass , @xmath17 is the speed of light , @xmath18 is the mass density in unit of @xmath19 .
] can be modelled by a simple @xmath0 polytropic eos with @xmath20 being the polytropic index ( i.e. the rest mass of a single particle @xmath21 the kinetic energy of a particle @xmath21 the fermi energy ) .
this approximate eos also gains support in numerical simulations ( e.g. bethe et al .
1979 ; hillebrandt , nomoto & wolff 1984 ; shen et al . 1998 ) . for instance
, bethe et al . ( 1979 ) concluded that as the neutrino trapping occurs , relativistic electrons , high - energy photons and neutrinos mainly contribute to the total pressure within a collapsing stellar core under gravity .
as nuclei start to feel " each other at a high density reaching up to @xmath22 , the stiffness of nuclear matter adjusts the polytropic index @xmath20 to @xmath23 .
there are two kinds of polytropic approximations .
one is the conventional polytropic eos with @xmath24 ( @xmath25 and @xmath15 are pressure and mass density , respectively ) where @xmath26 remains constant in space and time .
the other is a general polytropic eos where @xmath26 remains constant along streamlines , i.e. @xmath27 where @xmath28 is the bulk flow velocity .
the former is actually a special case of the latter , although we treat them separately .
it is easy to prove that for a relativistic hot fermi gas with a temperature much lower than the fermi energy , the specific entropy of the material is a function of @xmath29 ( see footnote 1 ) . in other words ,
the former eos is interpreted as a constant specific entropy while the latter bears a variable distribution of specific entropy .
the latter general polytropic eos obeys the conservation of specific entropy along streamlines .
the gas dynamics involves the competition between self - gravity and pressure gradient force .
such models give rise to various self - similar solutions by which the flow system partially loses its memory of initial and boundary conditions ( e.g. larson 1969 ; penston 1969 ; shu 1977 ; cheng 1978 ; gw ; yahil 1983 ; suto & silk 1988 ; lou & shen 2004 ; yu et al .
2006 ; lou & wang 2006 , 2007 ; wang & lou 2007 , 2008 ; lou & cao 2008 ; hu & lou 2009 ) . for core - collapse sne
, gw studied the homologous stellar core collapse of a @xmath0 conventional polytropic gas .
they concluded that starting from a static core , if the pressure is reduced by no more than @xmath30 , the central core will evolve into a homologously collapsing phase .
this fraction is much less than @xmath31 as indicated by bethe et al .
yahil ( 1983 ) extended gw analysis to @xmath32 conventional polytropic cases and noted the existence of an outer envelope moving inwards with a supersonic speed .
one important difference between gw and yahil ( 1983 ) is that solutions of the former have outer boundaries with zero mass density there while those of the latter extend to infinity ( lou & cao 2008 ) . with a similarity transformation ( fatuzzo et al . 2004 ) , lou & cao ( 2008 ) extended homologous core collapse to general polytropic cases
this is a substantial theoretical development of the model framework because several studies ( e.g. bethe et al .
1979 ; bruenn 1985 , 1989b ; burrows et al . 2006
; hillebrandt et al . 1984 ; janka & m@xmath33ller 1995 , 1996 ; shen et al . 1998 ) on eos with microphysics during sne do suggest variable specific entropy depending on physical conditions , including density , temperature and metallicity .
meanwhile , extensive numerical simulations show that spherically symmetric models can not initiate sn explosions with an energy of @xmath34 ( e.g. janka & m@xmath33ller 1995 , 1996 ; kitaura , janka & hillebrandt 2006 ) partly because the sn explosion energy appears insufficient and partly because instabilities occur in multi - dimensional simulations .
the role of instabilities and symmetry breaking in sne has now been emphasized ( e.g. burrows 2000 , 2006 ) . along this line ,
various instabilities were proposed ( e.g. goldreich , lai & sahrling 1996 ; lai 2000 ; lai & goldreich 2000 ; murphy , burrows & heger 2004 ; blondin et al .
2003 ; blondin & mezzacappa 2006 ) and several mechanisms may provide seed fluctuations before and during sn explosions ( e.g. bazan & arnett 1998 ; meakin & arnett 2006 , 2007a , b ) . prior to the onset of a core collapse , the so - called @xmath35-mechanism " ( e.g. goldreich et al . 1996 ; murphy et al . 2004 ) may lead to g@xmath1mode overstabilities in the progenitor , due to the overreaction of nuclear processes against perturbations .
lai & goldreich ( 2000 ) found an instability in the outer supersonic envelope during the collapsing phase .
they performed both analytical and numerical irrotational perturbation analysis using the conventional polytropic eos for collapsing solutions including ewcs of shu ( 1977 ) and post - collapse solution of yahil ( 1983 ) , and found instabilities in supersonic regions . in their derivation
, the perturbed flow was assumed irrotational ( i.e. without vorticity ) and thus g@xmath1modes should have been excluded . in an example , they initiated their calculation with a g@xmath1mode perturbation ; this initial condition appears to contradict the constraints of their analytical derivations and numerical analysis .
after the emergence of a rebound shock , several instabilities have been suggested .
intense convective motions may be sustained outside the neutrino sphere ( e.g. herant , benz & colgate 1992 ; herant et al .
standing accretion shock instability ( hereafter sasi ; e.g. foglizzo 2001 ) appears around @xmath36 ms after the core bounce in some simulations ( e.g. blondin & mezzacappa 2006 ; blondin et al .
burrows et al .
( 2006 , 2007a , b ) proposed that @xmath7 g@xmath1modes at @xmath37 ms after the stellar core bounce may serve as an agent to extract the gravitational energy for the kinetic energy of sne .
the roles of these instabilities are still hotly debated and a successful sn explosion requires further explorations .
no significant core instabilities were reported for the core - collapse phase .
linear stability analyses were performed by gw , lai ( 2000 ) and lai & goldreich ( 2000 ) for certain dynamic flows .
gw perturbed their homologously collapsing solutions and concluded that this collapse is stable for acoustic p@xmath1modes with the g@xmath1modes being neutral convective modes .
lai ( 2000 ) extended this acoustic stability analysis to solutions of yahil ( 1983 ) and found no unstable modes for @xmath38 cases in numerical explorations .
lai & goldreich ( 2000 ) studied the stability of collapsing core and claimed that the core remains stable in subsonic regions while the envelope becomes unstable in supersonic regions .
all these stable core statements relies on the assumption of a conventional polytropic eos .
conventional polytropic gas flows and perturbations correspond to a constant specific entropy and make g@xmath1modes just neutral convective modes .
our main theme is to examine stability properties of a stellar core collapse with a variable specific entropy distribution in a homologously collapsing model ( lou & cao 2008 ) . by numerical explorations ,
we classify various perturbation modes including p@xmath1modes and g@xmath1modes ( e.g. cowling 1941 ) , some of which are oscillatory while others grow with time in power laws . as
the hydrostatic equilibrium is a limiting case of our model , we find connection and evolution between perturbations in progenitors of hydrostatic case and dynamic collapsing stage .
the most interesting result is that some stable g@xmath1modes in the static case become unstable in the dynamically collapsing stage . in particular
, the stability of each mode now becomes sensitive to the self - similar evolution of specific entropy .
this instability occurs during the core - collapse phase , neither before the core collapse nor after the core bounce .
we speculate implications of such core instabilities during the collapse phase .
for example , the @xmath7 unstable g@xmath1mode may lead to the kick velocity of a pulsar .
sn explosions involve a chain of physical processes .
the spherical symmetry may be destroyed by a series of instabilities in the progenitor during the entire sn explosion .
for example , overstable g@xmath1modes of goldreich et al .
( 1996 ) and murphy et al .
( 2004 ) may provide seed g@xmath1mode perturbations during the core collapse .
perturbations of our model may connect to further instabilities after the core bounce and the emergence of an outgoing shock .
our model results are highly suggestive and can be tested numerically .
this paper is structured as follows .
section 2 describes general polytropic solutions of spherically symmetric homologous core collapse , as a generalization of gw results .
homogeneous ordinary differential equations ( odes ) for three - dimensional ( 3d ) general polytropic perturbations are obtained in section 3 .
numerical results are analyzed in section 4 .
we consider several aspects of sne in section 5 and conclude in section 6 .
mathematical details are summarized in appendices a@xmath1c for the convenience of reference .
before starting the time - dependent 3d general polytropic perturbation analysis , we first briefly summarize homologous core collapse solutions with spherical symmetry of lou & cao ( 2008 ) , making some notational adjustments in our model development for the convenience of comparison with gw results .
the nonlinear partial differential equations ( pdes ) for ideal hydrodynamics are conservations of momentum and mass , poisson equation for the gravitational field and a general polytropic eos , viz .
@xmath39 where @xmath40 , @xmath41 , @xmath42 and @xmath43 are bulk flow velocity , gas pressure , mass density and gravitational potential of the flow system , respectively and @xmath44 @xmath45(g s@xmath46 ) is the universal gravitational constant , and @xmath0 is the polytropic index for a relativistically hot gas .
it is known that these ideal nonlinear hydrodynamic pdes are invariant under the time reversal operation , @xmath47 this property enables us to use an outflow solution to also describe a collapse process , which is very important to understand homologous core collapse in terms of these expressions for the time reversal invariance . to generalize the analysis of gw model of spherical symmetry , we introduce the following time - dependent spatial scale factor @xmath48 , @xmath49 where time - dependent @xmath50 and constant coefficient @xmath51 are the values of @xmath15 and @xmath26 at the core centre of a massive progenitor star ( thus the subscript @xmath52 ) . the dimensional vector radius @xmath53 is scaled to a dimensionless vector radius @xmath54 . consistently , flow variables of @xmath55 , @xmath56 , @xmath57 and @xmath58 are assumed to take on the following forms of @xmath59 where @xmath60 is prescribed in the form of @xmath61 with a constant @xmath51 .
the central mass density @xmath50 is proportional to @xmath62 and varies with @xmath63 . substituting expressions ( [ eq8])@xmath1([eq11 ] ) into nonlinear pdes @xmath64 under spherical symmetry with @xmath65 , we reduce these nonlinear pdes to a set of coupled nonlinear odes .
first , pdes ( [ g1 ] ) and ( [ g4 ] ) are automatically satisfied , the latter of which means that @xmath66 can be of an arbitrary form .
for example , @xmath67 brings our general polytropic eos back to the conventional polytropic eos with a constant coefficient @xmath68 studied by gw . momentum equation ( [ g1 ] ) then leads to @xmath69\ , \end{aligned}\ ] ] where the left - hand side ( lhs ) depends only on @xmath63 while the right - hand side ( rhs ) depends only on @xmath70 . for consistency
, we therefore need to set both sides equal to a constant @xmath71 and obtain two separate nonlinear odes , viz .
@xmath72 ode ( [ 13 ] ) indicates that the spatial scale factor @xmath48 is either a constant independent of @xmath63 with @xmath73 or a power law of @xmath63 being proportional to @xmath74 . substituting equation ( [ 14 ] ) into poisson equation ( [ g3 ] ) , we derive an ode for @xmath75 , viz .
@xmath76 + 4f^3=4\lambda\ , \label{equf}\end{aligned}\ ] ] which gives a profile of mass density @xmath15 by eq ( [ eq9 ] ) at time @xmath63 and radius @xmath77 because @xmath70 is the independent similarity variable combining @xmath63 and @xmath77 together .
once @xmath75 is known , other variables can all be readily derived .
the special case of @xmath73 leads to the limit of general polytropic lane - emden equation , we refer to the resulting equilibrium equation as the general polytropic lane - emden equation . ]
( e.g. eddington 1926 ; chandrasekhar 1939 ) and @xmath78 describes outflows or collapses by the time reversal operation . for a necessary check , @xmath67 in ode ( [ equf ] ) reduces to equation ( 16 ) of gw precisely as expected .
the ` boundary conditions ' for second - order nonlinear ode ( [ equf ] ) are as follows .
the radial gradient of pressure should vanish at the centre , i.e. @xmath79 where the prime @xmath80 indicates the first derivative in terms of the self - similar independent variable @xmath70 . in order to obtain a physically sensible solution of @xmath75 related to the mass density by algebraic expression ( [ eq9 ] ) , we require an outer boundary @xmath81 which is the smallest value of solutions @xmath82 .
the reason is , if no solution is found for @xmath82 at a finite @xmath83 , it means that the system extends to infinity .
therefore , the dimensional velocity @xmath84 will diverge towards extremely large radii .
such a divergent flow velocity is unacceptable in realistic astrophysical gas systems .
hence , these boundary conditions determine a continuous range of @xmath85 values and the maximum acceptable value of @xmath85 , denoted by @xmath86 hereafter , corresponding to a solution @xmath75 where @xmath87 also at the outer boundary @xmath88 . by numerical explorations , solution @xmath75 for @xmath89 does not go to zero at a finite @xmath70 but oscillate with decreasing amplitude with increasing @xmath70 .
different profiles of @xmath66 will lead to different values of @xmath86 .
for the special case of @xmath67 , the @xmath85 range is @xmath90 which was first determined by gw and also confirmed by lou & cao ( 2008 ) with a corresponding @xmath91 .
the local polytropic sound speed is defined by @xmath92^{1/2}\ .\end{aligned}\ ] ] in the theory of stellar oscillations , it is required that the sound speed at the centre approaches a finite value ( e.g. unno et al .
we impose the same condition for a dynamic collapse . hence , @xmath93 and therefore @xmath94 .
once @xmath75 is obtained , it is straightforward to calculate the total enclosed mass of the collapsing core and the ratio between the mean mass density and the central mass density in terms of @xmath95 and @xmath81 .
the total enclosed mass @xmath96 of the collapsing core is given by @xmath97\nonumber\\ \!\!\!\!\!&=&\!\!\!\!\!\frac{4}{3}\pi x_b^3\left(\frac{\kappa_c}{\pi g}\right)^{3/2}\frac{\bar{\rho}}{\rho_c}\ , \end{aligned}\ ] ] where the ratio between the mean mass density @xmath98 and the central mass density @xmath99 is readily identified with @xmath100 the mean mass density @xmath98 varies with @xmath63 because @xmath101 decreases with @xmath63 for a homologous stellar core collapse .
for @xmath102 , we have @xmath87 and therefore @xmath103 .
we now consider 3d general polytropic perturbations to the background self - similar hydrodynamic collapse described in the previous section in spherical polar coordinates @xmath104 .
flow variables including small perturbations are assumed to bear the following forms of @xmath105\ , \label{p2}\\ p=\frac{\kappa_c^3}{(\pi g)^2}a(t)^{-4}g(x)f^4(x)[1 + \beta_1(x,\ \theta,\ \varphi)\tau(t)]\ , \label{p3}\\ \phi=\frac{4}{3}\left(\frac{\kappa_c^3}{\pi g}\right)^{1/2}a(t)^{-1}[\psi(x)+\psi_1(x,\
\theta,\ \varphi)\tau(t)]\ , \label{p4}\end{aligned}\ ] ] where the subscript @xmath106 indicates associations with perturbation terms which are small compared to the background dynamic flow variables and the free - fall timescale @xmath107 is itself time - dependent and is written in the specific form of @xmath108 in expressions ( [ p1])@xmath1([p4 ] ) above , @xmath109 is a time - dependent factor for perturbations and is assumed to bear the form of @xmath110\qquad\quad { \rm for\ \lambda>0}\ , \label{grow}\end{aligned}\ ] ] where the value of index parameter @xmath111 indicates either increase or decrease as well as oscillations of perturbations relative to the self - similar dynamic background flow .
substituting expressions ( [ p1])@xmath1([p4 ] ) into nonlinear pdes ( [ g1])@xmath1([g4 ] ) with the standard linearization procedure , we obtain equations governing 3d linear perturbations , viz .
@xmath112-\nabla\psi_1\ , \label{pf1}\\ ff_1 + 3{\bf w}_1\cdot\nabla f+f\nabla\cdot{\bf w}_1=0\ , \label{pf2}\\ \nabla^2\psi_1=3f^3f_1\ , \label{pf3}\\ g\left(\beta_1-\frac{4}{3}f_1\right)+{\bf w}_1\cdot\nabla g=0\ , \label{pf4}\end{aligned}\ ] ] for vector momentum equation , mass conservation , poisson equation , and specific entropy conservation along streamlines , respectively , where the velocity perturbation @xmath113 and @xmath114 $ ] .
therefore we have two values @xmath115 for a given @xmath85 and @xmath116 .
our definition of @xmath116 here has an opposite sign difference as compared to gw definition immediately after their equation ( 27 ) . for @xmath117
, we have a complex conjugate pair for @xmath111 corresponding to oscillations , while for @xmath118 , we have two real values of @xmath111 corresponding to different perturbation growth rates for upper plus and lower minus signs in @xmath111 . the angular components of @xmath119 , @xmath120 and @xmath121 can be readily separated out from the above perturbation equations by spherical harmonics for spherical harmonics @xmath122 to distinguish from eigenvalue parameter @xmath114 $ ]
. the spherical harmonics is defined by @xmath123^{1/2}p_l^\mathfrak{m } ( \cos\theta)\exp(i\mathfrak{m}\varphi)\ , \\ \mathfrak{m}=-l,\ -(l-1),\ \ldots,\
l-1,\ l\ , \end{aligned}\ ] ] where the associate legendre polynomial @xmath124 is defined as @xmath125 ( e.g. gupta 1978 ) . ] and @xmath126 takes the specific form of @xmath127 by this form of velocity perturbation , the radial component of vorticity perturbation is zero , while the @xmath128 and @xmath129 components of vorticity perturbation do not vanish in general
. this allows the possible presence of g@xmath1mode perturbations as well as convective motions and is distinctly different from irrotational velocity perturbations of gw and lai & goldreich ( 2000 ) . in the following analysis , @xmath120 , @xmath121 and @xmath119
only describe the radial variations of respective perturbation variables with the understanding that the relevant angular parts involving the spherical harmonics @xmath130 have been separated out . using equation ( [ pf4 ] ) and the angular ( i.e. transverse ) components of equation ( [ pf1 ] ) to eliminate @xmath121 and @xmath120 in the other three equations
, one finally arrives at the following fourth - order system of homogeneous linear odes for 3d general polytropic perturbations , viz .
@xmath131w_r+\frac{3m}{4g}\frac{dg}{dx}xw_t\qquad\nonumber\\ -m \frac{d}{dx}\left(xw_t\right)+\frac{3}{4g}\frac{dg}{dx}\psi_1=0\ , \label{pff2}\\ \frac{f}{x^2}\frac{d}{dx}\left(x^2w_r\right ) -\frac{l(l+1)f}{x}w_t+3\frac{df}{dx}w_r\qquad\qquad\qquad\nonumber\\ -\frac{1}{g } \left(mxw_t-\frac{3}{4}f\frac{dg}{dx}w_r+\psi_1\right)=0\ .\label{pff3}\end{aligned}\ ] ] because of the background spherical symmetry , 3d general polytropic perturbations are degenerate with respect to the azimuthal degree @xmath132 as expected ( note that @xmath132 and @xmath116 are two distinctly different parameters in our notations ) .
regular boundary conditions at both the centre and outer boundary are required to keep perturbations physically sensible .
they are prescribed as follows .
@xmath133 boundary conditions at @xmath134 are imposed in order to avoid singularity in perturbation solutions at the centre .
the other boundary condition at the moving radius of a collapsing core is to require a zero lagrangian pressure perturbation there , i.e. @xmath135 . for the special case of @xmath67 corresponding to a conventional polytropic gas , these perturbation equations automatically reduce to those of gw as expected .
note that our euler equation is written in a vector form because the curl of the velocity perturbation field will not vanish for a general polytropic gas and thus the stream function approach of gw ( whose gradient represents the velocity perturbation field ) is not sufficiently inclusive especially in view of possible g@xmath1mode oscillations and convective instabilities .
however , for a conventional polytropic gas of constant @xmath26 or @xmath67 , a stream function can be defined and thus gw expressed euler equations in a scaler form by using such a stream function .
they perturb the stream function instead of the velocity field directly .
more specifically , if the perturbation stream function takes the form of @xmath136 , then @xmath137 gives @xmath138 this illustrates gw result being a special subcase of our more general polytropic model description . in other words ,
gw consider only perturbed potential flows without vorticity perturbations ; this approach suffices for purely acoustic oscillations .
likewise , the irrotational perturbation flows of lai & goldreich ( 2000 ) should retain acoustic p@xmath1modes but exclude gravity g@xmath1modes and convective motions . in our perturbation approach ,
vorticity perturbations are present and all possible oscillations for p@xmath1modes , f@xmath1modes , and g@xmath1modes are included in the model consideration .
several solution properties of this eigenvalue perturbation problem can be demonstrated .
for example , the eigenfunctions of different eigenvalues @xmath116 are mutually orthogonal ( see appendix a ) . in appendix b , the eigenvalues and eigenfunctions can also be written in terms of the variational principle ( e.g. chandrasekhar 1964 ) .
in particular , we demonstrate in appendix c that the total energy criterion of chandrasekhar ( 1939 ) is not sufficient to guarantee the stability of an equilibrium configuration in view of the possible onset of convective instabilities for a variable specific entropy distribution .
in this section , 3d perturbation solutions are divided into several classes analogous to the classification schemes of cowling ( 1941 ) , cox ( 1976 ) and unno et al .
( 1979 ) for global stellar oscillations of static spherical stars .
non - radial oscillation modes of a static spherical star have been separated into different branches according to their respective dominating restoring forces and their frequency ranges ( e.g. cowling 1941 ) .
the p@xmath1modes correspond to acoustic oscillation modes in which pressure force is the major restoring force .
gravity modifies such trapped sound waves in several ways .
the characteristics of such p@xmath1modes are : ( 1 ) the peaks of perturbation functions tend to concentrate towards the outer envelope with increasing degree @xmath139 ( i.e. larger @xmath139 values ) .
( 2 ) their mode frequencies are relatively high , compared with eigenfrequencies of other modes such as f@xmath1modes and g@xmath1modes . ( 3 ) the more the number of radial nodes in eigenfunctions , the higher the p@xmath1mode eigenfrequencies and the more longitudinal the oscillations are .
the radial components of perturbed velocity dominate the oscillation in high - degree ( i.e. @xmath140 ) modes .
another type of oscillatory modes is the so - called internal g@xmath1modes in which gravitational restoring force takes the dominant role .
in contrast to p@xmath1modes , the g@xmath1mode characteristics are : ( 1 ) the maxima of perturbation eigenfunctions bury deeply in the stellar interior .
( 2 ) their frequencies are relatively low . ( 3 ) the frequency goes lower with the increase of the number of radial nodes ( the well - known anti - sturmian property , e.g. lou 1995 ) . as a limiting case , the frequency will approach zero and the perturbation becomes nearly horizontal . between p@xmath1modes and g@xmath1modes
, there exist the transitional f@xmath1modes which have no nodes in both the mass density perturbation and the radial component of velocity perturbation .
they are essentially surface modes in that perturbations have evanescent behaviours beneath the surface layer .
when the perturbation degree @xmath139 becomes very large , the perturbation concentrates around the surface layer .
this mode is closely related to the so - called lamb waves ( lamb 1932 ; lou 1990 , 1991 ) which propagate in the horizontal direction and vanish in the vertical direction .
these f@xmath1modes may also be regarded as the lowest - order p@xmath1modes more specifically , g@xmath1modes can be further divided into two kinds , namely g@xmath2modes and g@xmath3modes ( reading g@xmath1plus modes and g@xmath1minus modes , respectively ) , according to whether the eigenvalue @xmath116 is less or greater than zero . for global 3d perturbations in static stars , the former class of modes is stable while the latter class is unstable ; the stability property of these modes ( especially the g@xmath2modes ) is modified in background dynamic collapses as discussed presently .
the existence of such two classes of g@xmath1modes depend on the square of the so - called brunt - v@xmath5is@xmath5l@xmath5 buoyancy frequency @xmath141 , defined explicitly by @xmath142 where @xmath143 is the magnitude of the local gravitational acceleration and @xmath20 is the polytropic index of perturbations .
if @xmath6 is positive everywhere , eigenvalues of @xmath116 are always negative and only g@xmath2modes occur .
if @xmath6 is negative everywhere , eigenvalues of @xmath116 are always positive and only unstable g@xmath3modes occur .
if @xmath6 is positive in a certain part of the system and negative in another part , both types of g@xmath1modes may occur ( e.g. lebovitz 1965a , b , 1966 and cox 1976 for stability properties of oscillations in static stars ) .
note that the gravitational acceleration always points towards the centre of the gas sphere ( i.e. @xmath144 by our notation ) .
therefore , the sign of @xmath6 is determined by the expression in the parenthesis of definition ( [ bvfreq ] ) . in fact
, if the sign of the expression in the parenthesis is negative , this region satisfies the schwarzschild criterion for convective instability ( e.g. lebovitz 1965a ) .
hence , the g@xmath3modes represent convectively unstable modes . in the simplest interpretation
, @xmath141 is the buoyancy frequency associated with a perturbed parcel of fluid in a convectively stable medium .
moreover , scuflaire ( 1974 ) concluded for oscillations in static stars that eigenfunctions of g@xmath3modes can be oscillatory only in convectively unstable region ( i.e. @xmath145 ) . on the other hand , eigenfunctions of g@xmath2modes
can be oscillatory only in radiative region ( i.e. @xmath146 ) .
we found similar features for perturbations in a homologously collapsing stellar core by our extensive numerical explorations . cowling ( 1941 ) showed and we readily confirm by definition ( [ bvfreq ] ) that for a conventional polytropic eos @xmath147 with @xmath26 being constant for both dynamic core collapse and perturbations , the g@xmath1modes are simply neutrally stable convective modes with @xmath148 as noted by gw .
for this reason , gw investigated stability properties of acoustic p@xmath1modes and suppressed vorticity perturbations in their model analysis by introducing a stream function for velocity perturbation .
we note in passing that the situation with two different polytropic indices for core collapse and perturbations deserves a further investigation .
in contrast , our general polytropic eos allows fairly free options of specific entropy evolution along streamlines so that different @xmath66 gives different profiles of @xmath6 .
consequently , g@xmath1modes can occur in a homologously collapsing stellar core and can modify the convective instability criterion of such a dynamic core collapse in a non - trivial manner .
the numerical schemes we use to solve this perturbation problem are described below .
first , given a proper value of @xmath85 parameter and a prescribed @xmath66 , we use an explicit fourth - order runge - kutta scheme to numerically solve nonlinear ode ( [ equf ] ) for @xmath75 and then determine all relevant background self - similar dynamic flow variable profiles , i.e. @xmath149 by first - order ode ( [ 14 ] ) , and then @xmath55 , @xmath56 , @xmath57 and @xmath58 by equations ( [ eq8])@xmath1([eq11 ] ) correspondingly . having done this , we discretize self - similar perturbation odes ( [ pff1])@xmath1([pff3 ] ) by using a proper mesh in order to cast this eigenvalue problem of odes into a matrix eigenvalue problem . because small eigenvalues are practically important , inverse iteration method is employed to compute eigenvalues of @xmath116 accurately and to determine the corresponding eigenfunctions ( e.g. wilkinson 1965 ) . for the purpose of checking , we substitute the obtained eigenvalues and eigenfunctions into linear odes ( [ pff1])@xmath1([pff3 ] ) to compute the residues which are sufficiently small .
this verification confirms the validity of our numerical method . for double check ,
the runge - kutta shooting method is also applied ( i.e. starting numerical integrations near both ends towards the centre ) to verify the eigenvalues and eigensolutions for perturbations .
more specifically , this perturbation problem involves undesirable diverging solutions towards both ends .
we use the numerical results obtained from the matrix inverse iteration method at two respective points sufficiently close both ends and integrate towards each other to meet at a mid - point .
this avoids the numerical difficulty of diverging solutions towards both ends at @xmath150 and @xmath151 .
given a specified profile of @xmath66 for the specific entropy distribution with @xmath152 at @xmath150 and a perturbation degree @xmath139 value , numerical computations for 3d general polytropic perturbations lead to families of eigenvalue @xmath153 curves .
one important feature for these curves as revealed by our numerical explorations is that none of these families of eigenvalue curves intersects with the @xmath154 line , i.e. each eigenvalue branch @xmath153 remains either above or below @xmath154 line .
physically , this indicates that when @xmath85 increases from zero to the maximum @xmath86 , no oscillations initially belonging to g@xmath3mode regime jump into g@xmath2mode regime across the demarcation @xmath154 . this is a general empirical conclusion on the basis of our very extensive numerical explorations .
the physically relevant form of @xmath66 , i.e. the radial evolution of specific entropy along streamlines , requires a comprehensive understanding of nuclear processes inside the high - density stellar core under consideration . in order to effectively illustrate essential features of our model through numerical explorations , several possible trial distributions of specific entropy which have relatively simple analytic forms of @xmath66
are prescribed .
it should be emphasized that our model analysis regarding g@xmath1modes does carry more general and important implications for the stellar core collapse . in the following ,
the conventional polytropic results of gw model are first carefully examined by our approach and then three other types of @xmath66 are prescribed to explore stability properties of p@xmath1modes , f@xmath1modes and g@xmath1modes . for the three lowest - order @xmath7 p@xmath1modes versus @xmath85 variation in a conventional polytropic stellar core ( i.e. constant @xmath26 or @xmath67 ) in a homologous collapse .
the upper most one is an artificial mode branch corresponding to a displacement of the coordinate origin .
the lower two branches are acoustic p@xmath1modes with zero and one node in the radial velocity perturbation , respectively .
the eigenvalues of @xmath116 converge to @xmath155 in the limit of @xmath156 .
the dashed curves represent gw results for which we have already adjusted the sign difference between eigenvalues of ours and those of gw .
our eigenvalues @xmath116 differ from those of gw significantly , even though the corresponding eigenfunctions agree very well with those of gw .
, scaledwidth=50.0% ] for the three lowest - order @xmath8 p@xmath1modes versus @xmath85 variation in a conventional polytropic stellar core ( i.e. constant @xmath26 or @xmath67 ) in a homologous collapse .
they are all eigenvalues for acoustic p@xmath1modes with zero , one and two nodes in the radial velocity perturbation , respectively and converge to @xmath155 in the limit of @xmath156 .
the dashed curves represent gw eigenvalues for which we have already adjusted the sign difference between eigenvalues of ours and gw s .
again , our eigenvalues @xmath116 differ from those of gw significantly , while the corresponding eigenfunctions agree very well with those of gw . ,
scaledwidth=50.0% ] mode eigenfunction profiles of our @xmath157 ( solid curves ) which is equal to @xmath158 ( dashed curves ) of gw for @xmath8 and @xmath73 ( i.e. the static conventional lane - emden sphere ) are illustrated here for comparison .
the eigenfunctions of gw are displayed by dashed curves ; they are not readily distinguishable , because the agreements of theirs and ours are extremely close .
we suspect that the eigenvalues of gw as shown in figs .
[ fig1 ] and [ fig2 ] are in systematic errors .
, scaledwidth=50.0% ] p@xmath1mode eigenfunctions with two radial nodes in the transverse velocity perturbation for different values of @xmath85 .
these nodes concentrate towards the surface of the dynamic stellar core as @xmath85 approaches @xmath86 .
the dashed curves show gw results taken from their figure 3 .
our eigenfunctions agree well with those of gw but with a considerable differences in the corresponding eigenvalues @xmath116 as displayed in fig .
, scaledwidth=50.0% ] in order to compare with gw results for acoustic perturbations in a conventional polytropic gas with @xmath0 , we simply set @xmath67 and let all terms containing derivatives of @xmath66 vanish ; this simplifies perturbation odes ( [ pff1])@xmath1([pff3 ] ) considerably .
comparing our results with those of gw , the eigenvalue curves @xmath153 for the same oscillation p@xmath1modes do not coincide with each other well .
we have carefully checked our results in several ways and suspect systematic computational errors in the determination of eigenvalues)@xmath1([pff1 ] ) to verify the results .
meanwhile , given an initial value from the calculated eigenfunctions , we use an explicit fourth - order runge - kutta scheme to solve odes ( [ pff1])@xmath1([pff1 ] ) with the calculated eigenvalue again to double - check the correctness of the results .
in this verification , regions near @xmath150 and the outer boundary @xmath151 are excluded to avoid diverging solutions there .
we then calculate and compare the counterpart results shown in the figures of gw .
after that , we find excellent coincidence in eigenfunctions but systematic errors in eigenvalues between ours and gw results .
] by gw . figs .
[ fig1 ] and [ fig2 ] show the first three branches for eigenvalues of @xmath116 versus @xmath85 for @xmath7 and @xmath8 p@xmath1modes ( the branch with @xmath159 for @xmath7 is an exception ) . for visual comparisons ,
gw eigenvalues are also displayed by dashed curves in figs [ fig1 ] and [ fig2 ] .
we have specifically compared the eigenfunctions of ours with those of gw in figs .
[ fig3 ] and [ fig4 ] . within numerical errors ,
the corresponding eigenfunctions agree with each other very well .
we have carefully examined our computational procedures and programs and make sure that the first three eigenvalues correspond to the first three lowest - order p@xmath1mode oscillations with @xmath160 , @xmath106 and @xmath161 nodes along the radial direction .
we suspect systematic errors in the computational results of gw in terms of their eigenvalues . figs .
[ fig3 ] and [ fig4 ] show several eigenfunctions for different eigenvalues of @xmath116 with specified @xmath85 values for the background dynamic collapse . in figs .
[ fig3 ] and [ fig4 ] , dashed curves are those of gw results .
the coincidence of their curves and ours are very good .
[ fig3 ] shows that similar to classical non - radial oscillations of a static spherical star , the @xmath162th lowest eigenvalue of a p@xmath1mode correspond to an eigenfunction with @xmath163 number of nodes .
[ fig4 ] shows shape variations of eigenfunctions as @xmath85 varies in succession .
the positive @xmath153 curve in the @xmath7 case is an artificial one whose eigenfunction represents a movement of the coordinate origin .
another exceptional curve noted by gw is for the case of @xmath164 , representing a different choice of the time origin . except for these two special cases ,
eigenvalue @xmath153 curves of all families approach a common limit of @xmath155 in the limit of @xmath156 .
curves of eigenvalue @xmath153 which are always smaller than @xmath155 must have such a limit .
otherwise , an essential singularity of solutions appears for the limiting case of @xmath102 ( see gw for details ) .
note that eigenfunctions of acoustic p@xmath1modes become concentrated towards the surface as @xmath85 approaches @xmath86 .
this singular behaviour of @xmath86 should be regarded as an artifact of the special mathematical character of this limiting solution . in reality
, the lagrangian pressure perturbation vanishes at the free surface .
a finite sound speed at the surface ensures that the normal modes are regular and the eigenvalues are isolated . only belongs to a purely mathematical singularity and makes little sense in physics , in the following discussion , we present common features for general values of @xmath85 without any special attention towards this limiting case . ] with the physical requirement of @xmath93 , we consider three types of variable @xmath66 for numerical explorations .
the first type is an increasing function of @xmath70 , the second type is a decreasing function of @xmath70 and the last type has both increasing and decreasing regions with increasing @xmath70 .
referring to definition ( [ bvfreq ] ) of @xmath6 for the buoyancy frequency squared which serves as the criterion for the existence of various g@xmath1modes , we find that the sign of @xmath6 closely follows the sign of the first derivative of @xmath66 .
thus in order to illustrate g@xmath1modes under various situations , such three types of @xmath66 are chosen as representatives .
more specifically , the three forms of @xmath66 are prescribed as even functions of @xmath70 and are given by @xmath165 where @xmath166 , @xmath167 , and @xmath168 are three parameters calibrating the values of respective derivatives and thus the degree of variation for the dimensionless @xmath66 . for various oscillation modes of several lowest orders versus the angular spherical harmonic degree @xmath139
are shown here .
three classes of modes , namely p@xmath1modes , f@xmath1modes and g@xmath2modes , are identified according to the values of @xmath116 and the structures of their corresponding eigenfunctions .
note that the specific entropy @xmath66 is an increasing function of @xmath70 in the form of expression ( [ dg1 ] ) with @xmath169 and parameter @xmath170 is adopted .
the g@xmath3modes do not exist in this case . , scaledwidth=50.0% ] for various oscillation modes of several lowest orders versus the angular spherical harmonic degree @xmath139 are shown .
three classes of modes , namely p@xmath1modes , f@xmath1mode and g@xmath3modes are identified according to the values of @xmath116 and the structures of their corresponding eigenfunctions .
only eigenvalues @xmath116 of g@xmath3modes are greater than zero . in this case , the specific entropy @xmath66 is a decreasing function of @xmath70 in the form of expression ( [ dg2 ] ) with @xmath171 and parameter @xmath170 is adopted .
no g@xmath2modes appear in this case . , scaledwidth=50.0% ] parameter for various oscillation modes of several lowest orders versus the angular spherical harmonic degree @xmath139 are shown .
four classes of perturbation modes , i.e. p@xmath1modes , f@xmath1modes , g@xmath2modes and g@xmath3modes , are identified according to the values of @xmath116 .
only eigenvalues of @xmath116 for convectively unstable g@xmath3modes are greater than zero . here
, @xmath66 takes the form of expression ( [ dg3 ] ) with @xmath172 , neither purely increasing nor purely decreasing , and parameter @xmath170 is adopted . ,
scaledwidth=50.0% ] .
it is clear that g@xmath2modes and g@xmath3modes are strictly divided by the dashed line @xmath154 .
the g@xmath2 and g@xmath3modes are distinguished according to the sign of @xmath116 ; the former corresponds to negative eigenvalues @xmath116 while the latter corresponds to positive eigenvalues @xmath116 .
however , the criterion for convective stability is @xmath173 in a dynamically collapsing stellar core rather than @xmath174 for a static background .
it is therefore possible for g@xmath2modes to become convectively unstable . , scaledwidth=50.0% ] of @xmath8 p@xmath1modes , f@xmath1modes , g@xmath2mode and g@xmath3mode are shown in this case of @xmath175 . ,
scaledwidth=50.0% ] mode for a @xmath66 in the form of expression ( [ dg3 ] ) are shown .
relevant parameters are @xmath170 and @xmath8 and the form of @xmath66 is shown in the figure .
the horizontal component @xmath176 ( dashed curve ) is larger than the radial component @xmath177 ( solid curve ) . here
, @xmath178 is shown by the dash - dotted curve .
the peaks of eigenfunctions concentrate in the region of @xmath179 corresponding to @xmath145 .
, scaledwidth=50.0% ] modes for a @xmath66 in the form of expression ( [ dg3 ] ) are shown .
relevant parameters are @xmath170 and @xmath8 and the form of @xmath66 is given in the figure .
the horizontal component @xmath176 ( dashed curve ) is larger than the radial component @xmath177 ( solid curve ) .
the specific entropy @xmath178 is shown by dash - dotted curve .
the peaks of eigenfunctions concentrate in the central region , distinctive from g@xmath3modes and p@xmath1modes .
, scaledwidth=50.0% ] modes are shown with the eigenvalue @xmath116 exceeding @xmath180 .
relevant parameters and results are shown in the figure .
it is clear that @xmath181 which corresponds to an unstable mode .
the dash - dotted curve is @xmath178 .
the concentration of amplitude peaks within the region of @xmath182 , i.e. @xmath183 , and the negative value of eigenvalue @xmath116 both show that it is a g@xmath2mode.,scaledwidth=50.0% ] several features of non - radial oscillations are contained in the spectrum of eigenvalues @xmath116 versus spherical harmonic degree @xmath139 for a given value of @xmath85 and a prescribed form of @xmath66 . figs .
[ increase ] , [ decrease ] and [ varying ] schematically present the mode spectra for each type of specific entropy @xmath66 in equations ( [ dg1 ] ) , ( [ dg2 ] ) and ( [ dg3 ] ) , respectively .
parameters @xmath184 , @xmath185 and @xmath172 are adopted for our numerical computations .
[ amplification ] zooms in the neighborhood of the ordinate in fig . [ varying ] .
each asterisk @xmath186 in the four figures represents computed eigenvalues @xmath116
( i.e. the abscissa ) for two or three lowest order modes versus the given spherical harmonic degree @xmath139 .
as the background dynamic core collapse is spherically symmetric , 3d perturbations are degenerate with respect to the azimuthal degree @xmath132 ( see footnote 2 ) as expected .
[ lambda_of_m ] further illustrates the variation of @xmath116 versus @xmath85 and reveals that no curve of @xmath153 goes across the line of @xmath154 . by these figures , it is clear to see the variation of eigenvalues with changing @xmath139 .
the absolute values of @xmath116 for p@xmath1modes are largest .
the higher the order of a p@xmath1mode , the larger the absolute value of the corresponding @xmath116 .
moreover , we find that @xmath116 is always smaller than @xmath180 , which lead to an imaginary part in the power index of time - dependent term @xmath109 for perturbations .
eigenvalues of g@xmath1modes , in contrast , lie in the regime of smaller absolute value @xmath116 , being further divided into two sub - modes according to the signs of @xmath6 . figs .
[ increase ] , [ decrease ] and [ varying ] demonstrate the criteria for the existence of g@xmath3modes and g@xmath2modes by @xmath6 , the square of the brunt - v@xmath5is@xmath4l@xmath5 buoyancy frequency @xmath141 . in fig .
[ increase ] where the form of @xmath66 keeps @xmath6 positive , only g@xmath2modes exist and they are thus calculated ; g@xmath3modes do not exist as confirmed by our numerical explorations . in fig .
[ decrease ] where @xmath6 is negative , only unstable g@xmath3modes exist and they are thus calculated ; g@xmath2modes do not exist as confirmed by our numerical explorations . in fig . [ varying ] where the sign of @xmath6 varies in the region ,
both types of g@xmath1modes exist and they are calculated by examples .
[ amplification ] clearly shows the difference between g@xmath3modes and g@xmath2modes in terms of eigenvalues of @xmath116 .
recalling the time - dependent factor @xmath109 by definitions ( [ timedepend ] ) and ( [ grow ] ) and the expression of parameter @xmath187 , we find that the criterion for convective stability changes with @xmath85 value of a dynamic core collapse .
as discussed presently , the first term @xmath188 is interpreted as the gas compression effect . therefore for @xmath181 of two real values of @xmath189 , one mode decays with time @xmath63 while the other mode diverges with time @xmath63 . for @xmath190 ,
both modes oscillate with time @xmath63 .
consequently , the criterion for convective stability should be @xmath173 rather than simply @xmath174 .
the most interesting physical consequence is the appearance of unstable g@xmath2modes of sufficiently high orders
[ example1 ] , [ example2 ] and [ example3 ] demonstrate concrete examples of unstable @xmath191modes , stable @xmath192modes and unstable @xmath192modes , respectively .
the form of specific entropy @xmath66 is described by expression ( [ dg3 ] ) for @xmath193 .
the unique f@xmath1modes ( i.e. lamb modes ) , shown in these figures and treated as the lowest p@xmath1modes , exist only for @xmath194 .
they separate p@xmath1modes from g@xmath1modes in the sense that the eigenvalues @xmath116 fall between the eigenvalues of p@xmath1modes and those of g@xmath1modes .
an important feature is that the eigenfunctions of f@xmath1modes contain no nodes in the radial component of velocity perturbation .
non - radial oscillations of p@xmath1modes , f@xmath1modes and g@xmath1modes exist for @xmath195 ( in particular f@xmath1modes requires @xmath194 ) .
oscillations for @xmath164 modes are purely radial acoustic oscillations as @xmath196 .
since the horizontal component vanishes , one can infer from the primary direction of perturbation of different modes that only p@xmath1modes exist .
our numerical results confirm this inference for a self - similar dynamic core collapse in that only stable p@xmath1modes exist for purely radial oscillations .
moreover , several characteristic features of g@xmath1modes are displayed in figs .
[ example1 ] , [ example2 ] and [ example3 ] .
these perturbations are primarily horizontal as the horizontal component of perturbation velocity is usually larger than the radial component of perturbation velocity .
the peaks of g@xmath2modes are trapped deeply in the stellar interior of the collapsing stellar core while peaks of g@xmath3mode eigenfunctions appear in the region of @xmath145 .
both have much difference from acoustic p@xmath1modes in which the radial component of perturbation dominates and oscillation peaks of eigenfunctions appear towards the surface layer of a collapsing stellar core .
homologous collapses have been invoked here to model the dynamic phase of stellar core collapse prior to the emergence of a rebound shock and the subsequent sn explosion . in numerical simulations ( e.g. bruenn 1985 ) and theoretical studies ( gw ; lai 2000 ) , this phase was claimed to be stable .
we perform general polytropic 3d perturbation analysis on such a homologous core collapse .
analogous to stellar oscillations of a static star , 3d perturbations during this dynamic core collapse can be classified into several distinct modes .
the most interesting and important revelation is that convectively unstable g@xmath1modes appear under generic conditions .
when this happens , several physical consequences for the remnant core follow , which may alter the scenario for the breakup of spherical symmetry during the core collapse . in our model derivations and computations ,
we start from assumptions of a general polytropic gas and the self - similar collapse of such a gas sphere under self - gravity . for a homologous core collapse , we find that the general polytropic eos is automatically satisfied . a radial evolution of specific entropy @xmath66 needs to be prescribed for a complete solution of a homologous core collapse .
moreover , the instability of such a dynamic core collapse depends on the radial evolution of specific entropy .
a brief review for the radial entropy distribution in the core collapse phase deems necessary . according to bethe et al .
( 1979 ) , entropy per nucleon ( in unit of boltzmann s constant @xmath197 ) is @xmath198 at electron capture where the mean electron number per baryon @xmath13 is @xmath199 .
this entropy value varies little such that a conventional polytropic eos was adopted by gw .
nevertheless , bethe et al .
( 1979 ) noted that multiple nuclear processes indeed change this value from this range and this entropy variation serves as the conceptual basis of our model .
for example , when the breakup of @xmath200fe occurs at a mass density of @xmath201 , entropy per nucleon varies from @xmath202 to @xmath203 at a temperature of @xmath204 mev .
while a constant specific entropy may be an expedient approximation , electron capture , neutrino trapping and other physical processes certainly make the specific entropy vary with position and time .
it is this variation of specific entropy distribution that leads to g@xmath1mode convective instabilities during a stellar core collapse before the emergence of a rebound shock . during the deleptonization process of a stellar core collapse , an entropy evolution according to numerical simulations ( e.g. bruenn 1985 ; janka et al .
1995 , 1996 ; burrows et al . 2006 ) might suggest that specific entropy increases with increasing radius , as grossly described by an increasing type of @xmath66 .
bruenn ( 1985 ) displayed an increasing distribution of entropy versus the enclosed mass .
as the enclosed mass always increases with radius , the specific entropy then ascends with increasing radius .
more evidence comes from simulations for the evolution after the core bounce .
several models illustrate the increasing trend of specific entropy distribution with radius just after the core bounce ( e.g. janka et al . 1995 , 1996 ; burrows et al . 2006 ) .
if the entropy distribution varies not very much through the core bounce , then the specific entropy distribution might also be an increasing function in the core collapse stage , corresponding to an increasing trend of @xmath66 versus @xmath70 . in short , the analysis of bethe et al .
( 1979 ) indicates that entropy can vary in the core collapse phase while other numerical simulations ( e.g. bruenn 1985 , 1989a , b ; janka & m@xmath205ller 1995 , 1996 ; burrows et al .
2006 , 2007b ) suggested an increasing distribution of entropy during the phase of rebound shock .
therefore , the overall distribution of entropy might be assumed to increase with radius .
having said this , the possibility can not be ruled out that in parts of regions the entropy may decrease with radius locally .
bruenn ( 1985 ) pointed out that entropy decreases with density for mass density @xmath206 .
a turning point @xmath207 appears here , above which the entropy increases with density .
the reason is that at a certain nuclear density , the energy level of @xmath208 neutron shell in the nuclei becomes filled and the electron capture by nuclei is no longer allowed ( e.g. fuller et al .
according to this result and the fact that mass density is a decreasing function of radius , when @xmath209 at a certain stage , the specific entropy distribution may not be a monotonic increasing function of radius .
if this analysis reflects the physical reality , there are likely some regions where the specific entropy decreases with radius . besides , detailed numerical simulations ( e.g. janka et al .
2007 ) with more sophisticated eos ( e.g. hillebrandt et al . 1984 ; shen et al .
1998 ) reveal that although the overall tendency is an outward increase , the entropy distribution may have regions where the entropy dips slightly .
the eos and the specific entropy distribution remain open questions .
no definitive evidence requires necessarily a constant or a monotonically increasing specific entropy distribution with increasing radius during the core collapse phase . by our computations ,
variable entropy regions certainly lead to unstable g@xmath1modes with possibly rapid growth rates .
these findings bear important implications to perturbation growths during the core collapse prior to the emergence of a rebound shock and sn explosions . for the type of sne involving stellar core collapses
, a rebound shock emerges surrounding the centre because the inner core is drastically compressed and stiffened obeying the eos at nuclear density .
after such a core bounce , the central neutron - rich core may become a proto - neutron star within a mass range of @xmath210 ( e.g. rhoades & ruffini 1974 ) .
one expects that in the pre - collapse stage , there should exist a dense central core with a comparable mass collapsing inwards to form such a proto - neutron star . some numerical simulations ( e.g. woosley , heger & weaver 2002 ) give the central core masses for progenitors of various masses with different metallicities .
the core mass appears in the range of @xmath211 ( @xmath212 g is the solar mass ) . for a sudden core collapse ,
the outer layers of the progenitor may not move in immediately ( burrows 2000 ) .
as the stellar core rapidly contracts inwards , it may temporarily detach from outer shells and evolve independently . in this sense , we conceive a collapsing core under the self - gravity .
meanwhile , a quantitative definition would be desirable .
gw used the invariance of the inner core mass to consider the maximum central pressure reduction percentage from a @xmath73 configuration to a homologous core collapse .
however , gw value of @xmath213 for the central pressure reduction is much less than the result of @xmath31 in the simulation of bethe et al .
gw defined a core mass as @xmath214 where @xmath215 and @xmath216 are the enclosed mass and the value of @xmath51 for the @xmath73 initial static lane - emden core while the coefficient @xmath217 comes from the variation of @xmath218 as @xmath85 varies from @xmath160 to the maximum value @xmath86 .
gw also noted that the inner core mass in the computation of van riper ( 1978 ) is @xmath219% larger than that of their definition .
in reference to gw , we define the following inner core mass inside the iron core of the progenitor .
given a form of specific entropy evolution @xmath66 , we can determine @xmath86 and then obtain by what factor @xmath220 , the value of @xmath218 increases as @xmath85 increases from @xmath160 to @xmath86 . as
this inner enclosed mass @xmath96 is proportional to @xmath221 , if we know the value of @xmath51 for the two cases of @xmath73 and @xmath102 , the inner core mass is then defined by @xmath222 this definition has an advantage that it allows arbitrary pressure reduction which triggers the sn explosion , though the inner core may be very small if the pressure reduction is great , especially in the gw cases of conventional polytropic eos . in our general polytropic eos characterized by a @xmath66 , the inner core mass can be larger according to this definition . in our model development
, we actually use parameter @xmath51 which is the proportional coefficient between the pressure @xmath25 and @xmath223 at the centre of a collapsing core to estimate the inner core mass . according to the results shown in figure 4 of hillebrandt et al .
( 1984 ) which gave @xmath25 versus @xmath15 relations for different entropies and values of the central entropy for progenitors with different metallicities ( e.g. woosley et al .
2002 ) , we infer parameter @xmath51 to be @xmath224 cgs units . the total enclosed core mass is sensitive to the value of @xmath51 because of the power - law dependence of @xmath225 .
a form of specific entropy distribution @xmath226 which deviates slightly from a constant is chosen here as a demonstration . for @xmath227 cgs units ,
the enclosed core mass does not exceed @xmath228 ; for @xmath229 cgs units , the enclosed mass is around @xmath230 ; and for @xmath231 cgs units , the enclosed core mass can reach @xmath232 .
as expected , the central part gives the main contribution to the enclosed core mass , i.e. materials are highly concentrated around the centre .
typically , the chosen form of @xmath66 does not change the enclosed core mass very much unless it deviates from a constant significantly .
the high - frequency acoustic oscillations of p@xmath1modes and f@xmath1modes may occur in any prescribed form of @xmath66 and are stable during the homologous core collapse .
this conclusion here is more general than that of gw for a conventional polytropic stellar core collapse . in our model analysis and for convenience
, f@xmath1modes may be regarded as the lowest - order p@xmath1modes .
hereafter , we do not distinguish the lowest order p@xmath1modes and f@xmath1modes unless f@xmath1modes are necessarily emphasized . as eigenvalues @xmath116 of p@xmath1modes decrease with increasing orders ,
if the lowest order mode is stable , then all acoustic p@xmath1modes remain stable . the local analysis in the wentzel - kramers - brillouin - jeffreys ( wkbj ) approximation suggests the stability of high - order p@xmath1modes ( gw ) .
a notable feature for perturbation modes in a dynamic core collapse reveals that the time - dependent factor does not take on a fourier harmonic form @xmath233 with @xmath234 being the angular frequency of a perturbation . instead , the temporal factor acquires a power - law form , except for the limiting case of @xmath235 which consistently reduces to the description of harmonic oscillations in a static general polytropic sphere . for stable acoustic p@xmath1modes ,
all eigenvalues @xmath116 found are negative and make @xmath117 where @xmath85 features the collapsing core .
therefore , the power - law index parameter @xmath109 has both real and imaginary parts .
the real part has a specific value of @xmath236 , corresponding to the adiabatic amplification of acoustic waves due to compression of gas collapsing towards the centre as noted by gw .
the imaginary part leads to a form of @xmath237 with @xmath238 being a real number ; this represents an oscillation , but not in a conventional sinusoidal form of harmonics . for these oscillations of p@xmath1modes and f@xmath1modes during a dynamic core collapse , analyses of gw and ours indicate stability .
if one includes effects of radiative losses and diffusive processes , then some of these acoustic oscillations may become overstable , i.e. oscillatory growths in a dynamic background .
such overstable acoustic oscillations may serve as seed acoustic fluctuations for the sasi ( e.g. foglizzo 2001 ) to operate during the subsequent emergence of an outgoing rebound shock ( e.g. lou & wang 2006 , 2007 ) .
the most important results of our model analysis are that some g@xmath1modes which were used to be regarded as convectively neutral modes under the conventional polytropic approximation ( gw ) can become unstable in a dynamic background with a general polytropic core collapse under gravity .
the occurrence of g@xmath2 and g@xmath3modes depends on the sign of @xmath6 as highlighted in section 4.1 .
in addition to the power - law factor @xmath239 for oscillation amplitudes by the core collapse compression , by using the modified onset criterion of convective instabilities , i.e. @xmath240 ( see section 4.2.2 ) , we find that g@xmath3modes and sufficiently high - order g@xmath2modes are unstable during the core collapse . here ,
g@xmath3modes whose eigenvalues @xmath116 by definition are always greater than zero are unstable .
table 1 provides information of several illustrative examples of g@xmath3modes ; eigenvalues @xmath116 of g@xmath2modes approach @xmath160 when the order goes higher . as a result ,
its value must exceed @xmath180 when the order is high enough .
therefore high - order g@xmath2modes become unstable . consequently , unless for a constant @xmath66 , unstable g@xmath1modes always exist , because at least one of the two kinds of g@xmath1modes occurs for a variable distribution of specific entropy .
it should be emphasized that numerical simulations so far have not shown drastic growths of unstable convective modes found here .
part of the reason is that there exists no systematic study on the stability of the core collapse .
one would have thought that numerical truncation errors and errors in the determination of thermodynamic variables from a tabulated eos are considerable .
it would be highly desirable to further pursue this problem numerically .
a physical scenario is advanced below for g@xmath2modes .
when g@xmath3modes do not occur , unstable g@xmath2modes may give rise to convective instabilities during the core collapse .
initially , the inner core of a progenitor remains in a @xmath73 configuration for which g@xmath2modes oscillate stably .
when an insufficient nuclear energy supply triggers a reduction of core pressure , the inner core begins to collapse homologously .
sufficiently high order g@xmath2modes become unstable , starting to grow and break the spherical symmetry of the collapsing core .
such convective instability is limited during the core collapse because the growth rate does not exceed @xmath241 because g@xmath2modes are defined by @xmath174 .
if the spatial scale of the core shrinks by a factor of @xmath242 , perturbation grows by a factor of @xmath243 .
this is comparable to the compression effect of a collapsing core .
we speculate that such g@xmath1mode instabilities under favorable conditions might break up a collapsing core of high density and influence the formation of the central compact object and its companions , such as binary pulsars ( e.g. hulse & taylor 1975 ) and planets around neutron stars ( e.g. bailes , lyne & shemar 1991 ; wolszczan & frail 1992 ) .
the importance of asymmetry has been emphasized recently for the core bounce in a progenitor star prior to sn explosions .
so far , one - dimensional model can not lead to a successful sn explosion .
another consensus is that the neutrino heating mechanism alone also fails to produce a sn explosion as the energy of @xmath244 appears insufficient by one or two orders of magnitude ( e.g. kirauta et al .
2006 ; burrows et al .
in recent years , physical mechanisms involving two kinds of fluctuations have been proposed to effectively extract the available gravitational energy to power sn explosions .
one is the sasi process , which invokes acoustic fluctuations to transfer energy .
the other process relying on @xmath7 g@xmath1modes appears at several hundred microseconds after the core bounce as simulated by burrows et al .
( 2006 , 2007a , b ) .
these two energy transfer processes destroy the spherical symmetry and make sn explosions anisotropic .
regarding the origin of such oscillatory modes or fluctuations , our proposed instabilities which take place during the core collapse phase before the emergence of a rebound shock should have already destroyed the spherical symmetry .
the condition for the occurrence of such instabilities appears generic , only requiring a variable radial specific entropy distribution .
physically , the g@xmath3modes and unstable g@xmath2modes lead to convections . by definition
( [ bvfreq ] ) for the brunt - v@xmath5is@xmath5l@xmath5 frequency @xmath141 , inequality @xmath146 is equivalent to the schwarzschild criterion for convective stability in a star .
the global eigenfunction of a g@xmath3mode describes convective motions in a homologous core collapse . for a given mode
, we can calculate the time evolution of the perturbation .
table [ table1 ] gives an example of the power indices for the lowest order @xmath8 mode under the prescribed form ( [ dg3 ] ) of @xmath245 . according to table [ tb1 ]
, the power - law index varies in a wide range , allowing for both fast and slow perturbation growths , indicating the trend that instabilities go into the nonlinear regime .
.the power - law index @xmath246^{1/2}$ ] in the time - dependent factor @xmath109 ( see eqns ( [ timedepend ] ) and ( [ grow ] ) ) for the lowest order @xmath8 unstable g@xmath3modes varies with @xmath85 values for the dynamic core collapse .
there are two values of @xmath111 resulting from one eigenvalue @xmath116 ; as the plus sign corresponds to stable modes , we only list the index of unstable modes with minus signs in calculating index @xmath111 . in calculations of this example , we find that the decrease of eigenvalue @xmath116 does not exceed @xmath247 when @xmath85 increases from @xmath160 ( static core ) to the maximum value @xmath248 . the specific entropy distribution @xmath245 is adopted according to expression ( [ dg3 ] ) . [ cols="^,^",options="header " , ] apparent similarities exist in parallel between stellar oscillations of a static star and 3d general polytropic perturbations in a homologous core collapse .
the classification of different oscillatory modes are introduced in reference to stellar oscillations and the degeneracy with respect to azimuthal degree @xmath132 is expected ( see footnote 2 ) .
acoustic p@xmath1modes exist for all values of perturbation degree @xmath139 and remain stable .
the acoustic f@xmath1modes exist for @xmath194 and also remain stable .
the criterion for the existence of two different types of g@xmath1modes remains the same , by using the sign of the square of the brunt - v@xmath5is@xmath5l@xmath5 buoyancy frequency @xmath6 defined by expression ( [ bvfreq ] ) .
similar to stellar oscillations , amplitudes of g@xmath1modes concentrate around the core while those of p@xmath1modes approach the outer layer especially for high - degree @xmath139 acoustic oscillations .
more specifically , peak amplitudes of g@xmath3modes appear in the radial regions with @xmath145 .
meanwhile , notable differences from stellar oscillations also arise in our perturbation analysis . as the most important features ,
the onset criterion for convective instability is modified in reference to the well - known schwarzschild criterion .
consequently , not only g@xmath3modes but also sufficiently high order g@xmath2modes become unstable .
we provide specific examples to demonstrate this novel phenomenon that may give rise to several possibilities to inner stellar core collapses prior to the emergence of rebound shocks and sn explosions .
the time - dependent factor no longer takes the exponential form but is replaced by a power - law form .
conceptually , our perturbation analysis during the phase of stellar core collapse shows intimate connections to perturbations before the commencement of stellar core collapse and after the onset of core bounce .
physically , the origin of perturbations can be fairly natural for massive progenitor stars . in terms of stellar evolution and dynamics , stellar oscillations of p@xmath1modes , f@xmath1modes and g@xmath1modes
may well occur in massive stars , such as red or blue giants prior to an inner core collapse ( e.g. murphy et al . 2004 ) .
nuclear burning in massive stars can also provide seed perturbations ( e.g. bazan & arnett 1998 ; meakin & arnett 2006 , 2007a , b ) .
such pre - existing stellar oscillations and perturbations associated with nuclear burning serve as sources of fluctuations during the core collapse phase before the emergence of a rebound shock around the centre .
gw studied the acoustic stability of core collapse phase for a conventional polytropic gas and noted that g@xmath1modes are all convectively neutral and p@xmath1modes are all stable .
lai ( 2000 ) extended this polytropic acoustic stability analysis to dynamic solutions of yahil ( 1983 ) and concluded that unstable acoustic bar modes ( i.e. @xmath8 ) exist for @xmath249 .
the instability is caused by an insufficient pressure against perturbations due to a soft eos and may indicate star formation .
a cloud with this instability tends to deform into an ellipsoid , in which fragmentation might occur . except for this
, no other unstable modes relevant to yahil solutions were found in lai s exploration .
lai also claims that shu ( 1977 ) isothermal ewcs is unstable . in the context of type ii sne
, he stated that no destructive oscillation modes exist in the collapsing core before forming a proto - neutron star .
these conclusions are based on the assumption of a conventional polytropic gas .
the fact is that several nuclear processes in sn explosions most likely make specific entropy distribution variable .
then g@xmath1modes are no longer convectively neutral . in particular
, their stability now becomes sensitive to the evolution of specific entropy .
these unstable g@xmath1modes should play significant roles for compact remnants and sne .
lai ( 2000 ) analyzed and claimed the stability of yahil ( 1983 ) solution when @xmath38 ; the background dynamic flows differ from ours .
lai & goldreich ( 2000 ) found acoustic instability growing during dynamic flows .
however , their proposed instability only grows in the supersonic region where the outer envelope resides ( yahil 1983 ) while the core still collapses subsonically .
a main assumption for these results is the same conventional polytropic eos for background flow and perturbations .
thus g@xmath1modes are convectively neutral in their analysis and only acoustic p@xmath1modes exist .
one can assess consequences of g@xmath1mode instabilities during the phase of the core collapse prior to sne .
such g@xmath1mode instabilities can grow and evolve nonlinearly and several plausible scenarios may be speculated . in the presence of a series of core nuclear processes , including
neutronization and electron capture , fluctuations in the specific entropy during the core - collapse phase is inevitable ; in particular , such unstable g@xmath1modes , i.e. convective instabilities destroy the spherical symmetry of a collapsing core before the emergence of a rebound shock .
this is a likely mechanism of producing asphericity now seemingly necessary for sn explosions .
two major consequences of such g@xmath1mode instabilities are as follows .
( i ) the formation and proper motion of a proto - neutron star can be affected .
as the collapse becomes aspherical , the proto - neutron star may also become aspherical . under favorable conditions , we speculate that such convective instabilities might be violent enough to tear a proto - neutron star into pieces .
( ii ) such instabilities are expected to distort the shape of a rebound shock front .
compared with current numerical simulations , in which a rebound shock front is initially spherical , then becomes non - spherical and bends towards a particular direction where gas is driven out , our proposed non - spherical shock front at the beginning might give rise to considerable differences for these numerical simulations .
there are several conjectures if our proposed instabilities play a major role in sn explosions .
the @xmath7 g@xmath1mode instability may give rise to kicks of proto - neutron stars . the @xmath8 instability may split the central massive core apart and implies a possible formation of binary pulsars ; by adjusting the parameters in our model , we can have low - order @xmath8 g@xmath1mode instability in a central core whose enclosed mass is @xmath250 .
high - order and high - degree g@xmath1mode instabilities always have a tendency to tear a central core into pieces . as a consequence ,
even though core collapse occurs during which neutronization is triggered and energetic neutrinos burst out , the outgoing rebound shock may leave behind a collection of broken clumps .
we suggest this possibility because in observations no central compact object is found in some sn remnants , including sn1987a ( chevalier 1992 ; manchester 2007 ; mccray 2009 private communications ) . one important prediction of our proposed unstable g@xmath1mode convective instabilities before and during the rebound shock emergence is the resulting turbulent mixing of heavier elements in the inner layers of the core for sne .
according to the evolution theory of massive stars ( e.g. nomoto & hashimoto 1988 ; woosley & weaver 1995 ) , different heavy elements ( e.g. fe , si , o , c etc . ) lie in ordered core layers from inside out . without convective instabilities occurring before the emergence of a rebound shock , these elemental layers should be more or less kept during a sn explosion .
in this situation , boundaries between elemental shells are expected to be identifiable .
in contrast , for our proposed g@xmath1mode convective instabilities , the resulting convective turbulence with sufficiently fast growths mixes elements fe , si , o , c in different core layers .
this convective mixing mechanism of the core turbulence should bear observational consequences in detecting various heavy nuclear species and mapping their spatial distributions in sn remnants .
observational evidence , including high neutron star peculiar velocities ( e.g. lyne & lorimer 1994 ; cordes et al .
1993 ; burrows 2000 ; arzoumanian et al . 2002 ; hobbs et al .
2005 ) , the detection of geodetic precession in the binary pulsar psr 1913 + 16 ( e.g. wex et al .
2000 ) , the spin - orbit misalignment ( e.g. kaspi et al . 1996 ) , implies kick " processes by which proto - neutron stars gain considerable kinetic energies during sn explosions .
three major mechanisms have been pursued ( e.g. lai , chernoff & cordes 2001 ) . here ,
our proposed @xmath7 g@xmath1mode instabilities in a collapsing core belong to hydrodynamically driven kicks " .
we shall not dwell upon the other two , viz .
electromagnetically driven kicks " and neutrino - magnetic field driven kicks " .
the dipole modes of @xmath7 , including p@xmath1modes and both types of g@xmath1modes , are characterized by possible displacements of the central core mass as already noted in stellar oscillations ( e.g. christensen - dalsgaard 1976 ) . as such dipole modes involve the core mass motion about the equilibrium centre , the core mass gains a certain amount of kinetic energy . during the violent rebound shock breakout of a sn , this core mass movement enables the remnant compact object to further gain kinetic energy and move away from the equilibrium centre .
the overall centre of mass should remain fixed in space .
this is a physically plausible ` kick ' process .
observationally , proper velocities of nascent neutron stars are typically @xmath251 ( e.g. lyne & lorimer 1994 ) and can reach as high as @xmath252 ( e.g. cordes et al . 1993 ; burrows 2000 ) . for all perturbations with @xmath253 ,
the boundary condition at @xmath254 requires a zero velocity there . in these cases ,
the core mass and the equilibrium centre coincide .
[ fig12 ] offers an example of unstable @xmath7 g@xmath2mode .
unstable @xmath7 g@xmath1modes of low orders during the core collapse phase in a massive progenitor star may give rise to the initial ` kick ' of a nascent proto - neutron star as joint results of nonlinear collapse evolution , rebound shock and sn explosion .
here , the requirement of low - order g@xmath1modes is to enclose a sufficient amount of core mass to be kicked out .
our model results indicate that with a plausible distribution of specific entropy , low - order @xmath7 g@xmath1modes can indeed be unstable during core collapse .
low - order g@xmath1modes ensures a sufficient amount of mass inside the inner most node . with such instabilities ,
a sizable mass in the central core can be kicked away from the equilibrium centre .
we may apply this
kick " scenario to a plausible situation in which several physical variables are assigned typical values for a stellar core . from hillebrandt et al .
( 1984 ) and woosley et al .
( 2002 ) , the central coefficient @xmath51 is estimated as @xmath255 cgs unit .
the time @xmath63 is characterized by the free - fall timescale @xmath256 in the order of a few to several tens of millisecond when the mass density @xmath15 falls in the range of @xmath257 . in our model , the time @xmath63 is negative ( i.e. time reversal ) for a homologously core collapse solution ; we thus take the initial time to be @xmath258 ms . here
, @xmath259 represents a slight increase of specific entropy with increasing @xmath70 . the case of @xmath260
is close to the maximum value @xmath86 of a physically allowed @xmath85 and also leads to a central mass density of @xmath261 consistent with simulation results ( e.g. bruenn 1985 ) . , the collapsing velocity @xmath262 and the mass density @xmath15 versus the enclosed mass @xmath96 in the unit of the solar mass @xmath263 .
, scaledwidth=50.0% ] lowest order g@xmath2mode .
parameters for this oscillatory mode are displayed in the figure .
this is a convectively unstable g@xmath2mode with @xmath181 .
we can see that the centre of mass has a velocity amplitude , which contributes to the kick velocity of the core .
, scaledwidth=50.0% ] fig .
[ fig13 ] illustrates the initial dynamic core collapse of such a situation , including the radius @xmath77 , the collapsing velocity @xmath262 and the mass density @xmath15 versus the enclosed mass @xmath96 .
[ fig12 ] shows a @xmath7 dipole unstable g@xmath2mode in which the central core has maximum perturbation amplitudes .
it is estimated that the sound speed of the central core is @xmath264 where the subscript @xmath17 indicates physical variables at the centre @xmath150 . for a velocity perturbation only of a few percent of the sound speed
, it can still be several hundred kilometers per second which is grossly consistent with observed kick velocities of radio pulsars ( e.g. lyne & lorimer 1994 )
. a somewhat larger velocity perturbation amplitude ( say 10% of the core sound speed ) may also give rise to a kick velocity of @xmath265 or even higher .
we propose that this lower - order @xmath7 g@xmath1mode instability is a plausible mechanism leading to kicks of central remnant compact objects in the aftermath of sn explosions .
a few binary pulsars have been detected , including the famous psr1913 + 16 ( hulse & taylor 1975 ) .
ideas have been proposed to explain their formation ( e.g. flannery & van den heuvel 1975 ) .
our @xmath8 g@xmath1mode instabilities suggest an alternative yet plausible origin of such binary pulsars .
unstable @xmath8 g@xmath1modes of low orders developed during the stellar core collapse of a massive progenitor might break the dense core apart before and during the emergence of a rebound shock and eventually give rise to binary compact objects .
the violent sn explosion drives the two compact blobs apart .
the spins of the two blobs and the binary orbital motion pick up part of the angular momentum of the massive progenitor star .
the mass of each components in psr1913 + 16 is estimated to be @xmath266 ( e.g. flannery & van den heuvel 1975 ) . for a central @xmath51 of @xmath267 cgs unit ,
the enclosed core mass can be @xmath268 .
that would be a sufficiently massive core for a possible split into two compact objects due to nonlinear evolution of unstable @xmath8 g@xmath1modes and subsequent core rebounce and sn explosion .
we also speculate that under possible and favorable conditions , the growth of unstable high - order high - degree ( e.g. @xmath269 ) g@xmath1modes and their nonlinear evolution might lead to an eventual breakup of a central proto - neutron star after the rebound shock emergence and subsequent sn explosion . in this scenario ,
the neutronization does occur during the brief core - collapse phase and high - energy neutrinos escape after a short moment of trapping , but without forming a coherent central neutron star in the end .
the expected compact remnant is actually shredded into pieces by the nonlinear evolution of high - order and high - degree unstable convective g@xmath1modes .
sn1987a might be such an example , for which no signals from a central compact object have ever been detected ( e.g. chevalier 1992 ; manchester 2007 ; mccray 2009 private communications ) .
unstable high - order high - degree g@xmath1modes under other conditions might also lead to the formation of planets orbiting around a neutron star .
these planets can hardly be those before the sn explosion because they can seldom survive under the expansion of the progenitor giant and the strong stellar wind as well as the sn explosion .
these high - order high - degree unstable g@xmath1modes during the core collapse might help certain amount of gas concentrated in isolated blobs , leading to the possible formation of planets around a new - born neutron star .
a few neutron stars with planets have been detected observationally ( e.g. bailes et al .
1991 ; wolszczan & frail 1992 ) .
we have revealed and emphasized unstable growths of g@xmath1mode convective instabilities during the inner core collapse under self - gravity inside a massive progenitor star and speculated several physical consequences of such instabilities about central compact remnants of sne .
together with other physical processes at different stages , such g@xmath1mode convective instabilities prior to the emergence of a rebound shock can be a necessary ingredient to achieve successful sn explosions .
moreover , different modes of oscillations in our classification provide sources of perturbations during the emergence of a rebound shock , for example , the @xmath7 g@xmath1modes discussed by burrows et al .
( 2006 , 2007a , b ) may have well originated from our @xmath7 g@xmath2 and g@xmath3modes during the stellar core collapse phase before the core bounce .
nonlinear evolution of diverse initial fluctuations in various stellar core collapse conditions is expected to give rise to a diversity of possible outcomes for remnant cores of sne .
we have systematically investigated physical properties of 3d perturbations in a hydrodynamic background of self - similar core collapse with a general polytropic eos for a relativistically hot gas of @xmath0 , as studied in lou & cao ( 2008 ) . for both background and perturbations , the two values of @xmath20
are taken to be the same ; the case of two different values of @xmath20 will be explored separately .
analogous to the mode classification of stellar oscillations in a non - rotating static star ( e.g. cowling 1941 ) , our 3d general polytropic perturbations are divided into four distinct classes of modes , viz .
p@xmath1modes , f@xmath1modes , g@xmath2modes ( i.e. with @xmath174 ) and g@xmath3modes ( i.e. with @xmath270 ) , according to their eigenvalue regimes of parameter @xmath116 in equation ( [ pf1 ] ) .
stability properties of these different perturbation modes are analyzed .
similar to stellar oscillations , acoustic p@xmath1modes and f@xmath1modes remain stable for 3d general polytropic perturbations in homologous stellar core collapse .
this more general conclusion also confirms the acoustic p@xmath1mode stabilities claimed by gw although their p@xmath1mode eigenvalues appear in systematic errors .
the temporal amplification factor @xmath239 in the perturbation magnitude is associated with the background gas compression during a homologous inner core collapse .
in contrast , g@xmath3modes and sufficiently high - order g@xmath2modes are both convectively unstable modes because the onset criterion of convective instabilities is now shifted from @xmath270 for a static general polytropic sphere to @xmath240 for a collapsing general polytropic core where @xmath78 characterizes the hydrodynamic background of a homologously collapsing stellar core .
the existence of both types of g@xmath1modes depends on @xmath6 , the square of the brunt@xmath1v@xmath5is@xmath5l@xmath4 buoyancy frequency ( see definition [ bvfreq ] ) which is determined by the evolution of the specific entropy distribution @xmath66 .
meanwhile , above what value of perturbation degree @xmath139 the g@xmath1modes are unstable also depends on the specific form of @xmath66 . as an example , the lowest - order @xmath7 unstable g@xmath2mode is shown in fig .
[ fig12 ] .
the peak amplitudes of g@xmath3modes lie in regions of @xmath145 , which is analogous to those in stellar oscillations .
these unstable g@xmath1modes lead to growths of convective motions in a self - similar collapsing stellar core .
their divergent growths scale as power laws in time @xmath63 ( with @xmath271 ) while stable perturbation modes oscillate in the manner of @xmath237 . in analyzing this perturbation problem
, we also realize that the global energy criterion of chandrasekhar ( 1939 ) is not sufficient to ensure the stability of general polytropic equilibria in view of the possible occurrence of convective instabilities for variable entropy distributions ( appendix c ) .
compared to possible sources of perturbations proposed in earlier models , including the so - called @xmath272mechanism " before the onset of inner core collapse and sasi after the core bounce , our g@xmath1mode convective instabilities occur during the dynamic core collapse .
contrary to earlier theoretical notions , the pre - sn stellar core collapse phase is most likely convectively unstable due to both types of g@xmath1mode instabilities .
this is because an exactly constant specific entropy everywhere in a stellar core should be extremely rare in any realistic progenitor star ( e.g. bruenn 1985 ) .
therefore , the spherical symmetry of a self - similar collapsing stellar core should be actually broken up earlier than presumed by most previous models of sne . in our scenario , oscillations of the progenitor star
serve as the most natural source for 3d perturbations .
for some regions of @xmath179 leading to @xmath145 locally , the unstable g@xmath3modes , of which low - degree modes may have sufficiently fast growth rates , will soon dominate and destroy the spherical symmetry .
if @xmath182 everywhere , high - order unstable g@xmath2modes will grow in a self - similar collapsing core and evolve nonlinearly . in the presence of inevitable core g@xmath1mode convective instabilities , several possible consequences may follow .
most prominently , the early break - up of spherical symmetry may lead to energy concentration in particular directions and there is thus no need for a rebound shock to push against the entire outer envelope .
the low - order unstable @xmath7 g@xmath1modes may give rise to the initial kick of a remnant central compact object .
meanwhile unstable g@xmath1modes correspond to the growth of convective motions which may stir up heavier elements fe , si , o and c in different inner layers of the collapsing core inside a massive progenitor via convective turbulence so that a mixed distribution of these elements to various extents in sn remnants is expected .
this prediction of our model can be tested by nuclear abundance observations in sne .
we further suspect that the nonlinear evolution of low - order unstable @xmath8 g@xmath1mode instabilities might disintegrate the central proto - neutron star core into two blobs to form binary pulsar systems under favorable conditions .
other high - order high - degree unstable g@xmath1modes might even prevent the formation of a coherent central compact object by breaking the core into multiple pieces .
it is speculated that this might happen to supernova sn1987a without signals from a central remnant compact object .
high - degree and high - order g@xmath1mode instabilities may lead to smaller blobs of low masses which might eventually form planets around a neutron star after a sn explosion .
we thank the anonymous referee for constructive suggestions to improve the quality of the manuscript .
this research was supported in part by tsinghua centre for astrophysics ( thca ) , by the national natural science foundation of china ( nsfc ) grants 10373009 and 10533020 at tsinghua university , and by the srfdp 20050003088 and 200800030071 , the yangtze endowment and the national undergraduate innovation training project from the ministry of education at tsinghua university .
0.9 cm 99 arnett w. d. , bahcall j. n. , kirshner r. p. , woosley s. e. , 1989 , ara&a , 27 , 629 arzoumanian z. , chernoff d. f. , cordes j. m. , apj , 2002 , 568 , 289 bailes m. , lyne a. g. , shemar s. l. , 1991 , nat , 352 , 311 barus r. , janka h .-
, rampp m. , kifonidis k. , 2006 , a&a , 457 , 281 bazan g. , arnett w. d. , 1998 , apj , 496 , 316 bethe h. e. , brown g. e. , applegate j. , lattimer j. m. , 1979 , nucl .
phys . , a324 , 487 bionta r. m. , et al . , 1987 , prl , 58 , 1494 blondin j. m. , mezzacappa a. , demarino c. , 2003 , apj , 584 , 971 blondin j. m. , mezzacappa a. , 2006 , apj , 642 , 401 bruenn s. w. , 1985 , apj , 58 , 771 bruenn s. w. , 1989a , apj , 340 , 955 bruenn s. w. , 1989b , apj , 341 , 385 burrows a. , 2000 , nat , 403 , 727 burrows a. , hayes j. , fryxell b. a. , 1995 , apj , 450 , 830 burrows a. , livne e. , dessart l. , ott c. d. , murphy j. , 2006 , apj , 640 , 878 burrows a. , dessart l. , ott c. d. , livne e. , 2007a , phys .
, 442 , 23 burrows a. , dessart l. , livne e. , ott c. d. , murphy j. , 2007b , apj , 664 , 416 burrows c. j. et al .
1995 , apj , 452 , 680 chandrasekhar s. , 1939 , an introduction to the study of stellar structure , dover publications , inc .
, london chandrasekhar s. , 1964 , apj , 139 , 664 cheng a. f. , 1978 , apj , 221 , 320 chevalier r. a. , 1992a , nat , 355 , 691 chevalier r. a. , 1992b , nat , 360 , 628 christensen - dalsgaard j. , 1976 , mnras , 174 , 87 cordes j. m. , romani r. w. , lundgren s. c. , 1993 , nat , 362 , 133 cowling t. g. , 1941 , mnras , 101 , 367 cox j. p. , 1976
, ara&a , 14 , 247 deubner f .-
l . , gough d. , 1984 , ara&a , 22 , 593 eddington a. s. , 1926 , the internal constitution of the stars , cambridge university press , cambridge fatuzzo m. , adams f. c. , myers p. c. , 2004 , apj , 615 , 813 flannery b. p. , van den henvel e. p. j. , 1975 , a&a , 39 , 61 foglizzo t. , 2001 , a&a , 368 , 311 fryer c. l. , warren m. s. , 2002 , apj , 574 , l65 fryer c. l. , warren m. s. , 2004 , apj , 601 , 391 fuller g. m. , fowler w. a. , newman m. j. , 1982 , apj , 252 , 715 goldreich p. , weber s. v. , 1980 , apj , 238 , 991 gupta b. d. , 1978 , mathematical physics , vikas , ghaziabad , india hansen c. j. , kawaler s. d. , 1999 , stellar interiors , springer , new york harent m. , benz w. , colgate s. , 1992 , apj , 395 , 642 harent m. , benz w. , hix w. r. , fryer c. l. , colgate w. , 1994 , apj , 435 , 339 hillebrandt w. , nomoto k. , wolff r. g. , 1984 , a&a , 133 , 175 hirata k. et al . , 1987 , phys .
lett , 58 , 1490 hobbs g. , lorimer d. r. , lyne a. g. , kramer m. , 2005 , mnras , 360 , 974 hu r. y. , lou y .- q . , 2009 , mnras , 396 , 878 huang k. , 1987 , statistical mechanics , wiley , new york hulse r. a. , taylor j. h. , 1975 , apj , 195 , l51 janka h .-
, m@xmath33ller e. , 1995 , apj , 448 , l109 janka h .-
, m@xmath33ller e. , 1996 , a&a , 306 , 167 janka h .-
, langanke k. , marek a. , mart@xmath273nez - pinedo g. , m@xmath33ller b. , 2007 , phys . rep .
, 442 , 38 kitaura f. s. , janka h .-
, hillebrandt w. , 2006 , a&a , 450 , 345 lai d. , 2000 , apj , 540 , 946 lai d. , goldreich p. , 2000
, apj , 535 , 402 lamb h. , 1932 , hydrodynamics , 6th ed .
university cambridge press , london landau l. d. , lifshitz e. m. , 1987 , fluid mechanics , 2nd edn .
pergamon press , new york larson r. b. , 1969 , mnras , 145 , 271 lebovitz n. r. , 1965a , apj , 142 , 229 lebovitz n. r. , 1965b , apj , 142 , 1257 lebovitz n. r. , 1966 , apj , 146 , 946 lou y .- q .
, 1990 , apj , 361 , 527 lou y .- q .
, 1991 , apj , 367 , 367 lou y .- q . , 1995 , mnras , 276 , 769 lou y .- q . , 1996 , science , 272 , 521 lou y .- q .
, cao y. , 2008 , mnras , 384 , 611 lou y .- q . ,
shen y. , 2004 , mnras , 348 , 717 lou y .- q . , wang w .- g . , 2006 , mnras , 372 , 885 lou y .- q . , wang w .- g . , 2007 , mnras , 378l , 54 lyne a. g. , lorimer d. r. , 1994 , nat , 369 , 127 manchester r. n. , 2007 , supernova 1987a : 20 years after : supernovae and gamma - ray bursters .
aip conference proceedings , 937 , 134 meakin c. a. , arnett d. , 2006 , apj , 637 , l53 meakin c. a. , arnett d. , 2007a , apj , 665 , 690 meakin c. a. , arnett d. , 2007b , apj , 667 , 448 murphy j. w. , burrows a. , heger a. , 2004 , apj , 615 , 460 nomoto k. , hashimoto m. , 1988 , phys . rep . , 163 , 13 penston m. v. , 1969 , mnras , 144 , 425 phukon t. c. , sarmah b. p. , phys .
d , 55 , 4 , 1777 rhoades c. e. , ruffini r. , 1974 , phys .
, 32 , 324 scuflaire r. , 1974 , a&a , 34 , 449 shen h. , toki h. , oyamastsu k. , sumiyoshi k. , 1998 , nucl .
a , 637 , 435 shu f. , 1977 , apj , 214 , 488 suto y. , silk j. , 1988 , apj , 326 , 527 unno w. , osaki y. , ando h. , shibahashi h. , 1979 , nonradial oscillations of stars , university of tokyo press , tokyo van riper k. a. , 1978 , apj , 221 , 304 wang l. et al . , 2002 , apj , 579 , 671 wang w .- g . , lou y .- q . , 2007 , ap&ss , 311 , 363 weaver t. a. , zimmerman g. b. , woosley s. e. , 1978 , apj , 225 , 1021 wilkinson j. h. , 1965 , the algebraic eigenvalue problem , clarendon press , oxford wolszczan a. , frail d. a. , 1992 , nat , 355 , 142 woosley s. e. , weaver t. a. , 1986 , ara&a , 24 , 205 woosley s. e. , weaver t. a. , 1995 , apjs , 101 , 181 woosley s. e. , langer n. , weaver t. a. , 1993 , apj , 411 , 823 yahil a. , 1983 , apj , 265 , 1047 yu c. , lou y .- q .
, bian f. y. , wu y. , 2006 , mnras , 370 , 121
we can now prove that eigenfunctions of different eigenvalues @xmath116 are mutually orthogonal . in the following illustration ,
superscripts @xmath274 and @xmath275 are utilized to distinguish different eigenfunctions and eigenvalues .
let us proceed to evaluate the integral of @xmath276 the last term on the rhs of eq .
( [ app1 ] ) leads to @xmath277 the first integration on the rhs of eq ( [ app1 ] ) can be re - expressed as @xmath278 and the second integration on the rhs of eq ( [ app1 ] ) is equal to @xmath279 note that both @xmath280 and @xmath281 depend only on @xmath70 .
therefore , the result indicates that the integral @xmath282 is manifestly symmetric in terms of superscripts @xmath274 and @xmath275 .
consequently , the orthogonality of eigenfunctions is proved .
the variational principle ( e.g. chandrasekhar 1964 ) can also be applied to this eigenvalue problem . referring to the formula in the proof of orthogonality in appendix a
, the eigenvalue can be written as the ratio of two integrals , viz .
@xmath283 where @xmath284 using the variational principle , we will show that the eigenvalue parameter @xmath116 has a stationary property when @xmath285 and @xmath286 are evaluated in terms of the true proper solutions . according to equation ( [ defm ] ) , @xmath287 where @xmath288 and @xmath289 are the changes in @xmath285 and @xmath286 in response to the variation @xmath290 in @xmath291 .
we have @xmath292 and @xmath293 keeping in mind of equations ( [ pf1])@xmath1([pf4 ] ) , we can write the variation @xmath289 as @xmath294-f^3\nabla\psi_1 \right\ } \cdot\delta{\bf w}dv\ .\label{deltai2}\end{aligned}\ ] ] by equations ( [ deltai1 ] ) and ( [ deltai2 ] ) , it then follows that @xmath295 if @xmath296-f^3\nabla\psi_1\ , \end{aligned}\ ] ] which is precisely equation ( [ pf1 ] ) for the eigenvalue problem of our model formulation .
in the model formulation of lou & cao ( 2008 ) , the situation of @xmath73 represents the limiting static general polytropic sphere . for the conventional polytropic equation of state ( i.e. a constant specific entropy everywhere independent of time @xmath63 ) , the governing equation of the static equilibrium returns to the well - known lane - emden equation ( e.g. eddington 1926 ; chandrasekhar 1939 ) . for a general polytropic gas sphere with a variable radial distribution of the specific entropy
, we refer to the governing equation with @xmath73 as the general lane - emden equation ( see footnote 1 in the main text ) .
we emphasize that the linear stability property now depends on the radial distribution of the specific entropy .
if the specific entropy decreases with radius @xmath77 , corresponding to a decreasing @xmath66 , convectively unstable g@xmath3modes will develop and drive the static sphere out of equilibrium configuration .
the total energy that is equal to the gravitational energy plus the thermal energy together still remains zero for @xmath0 ; however , this situation under the conventional polytropic assumption is referred to as being marginally stable by chandrasekhar ( 1939 ) .
the hydrostatic equilibrium equations are @xmath297 using equation ( [ ss1 ] ) to eliminate @xmath96 in equation ( [ ss2 ] ) and assuming the radial distribution of @xmath298 , one readily arrives at the dimensional general lane - emden equation whose dimensionless form appears as equation ( [ equf ] ) with @xmath73 .
the total energy of the system can be derived as follows .
the total gravitational energy is @xmath299 and the total thermal energy is @xmath300 the total energy of a general lane - emden sphere is @xmath301 where finite pressure at the centre and zero pressure at the stellar surface are presumed .
consequently , we prove that irrespective of the radial distribution of the specific entropy , the total energy of a general polytropic lane - emden sphere remains zero for @xmath0 .
evidently , this does not guarantee the linear stability of such a equilibrium configuration .
the simple total energy criterion is not enough to judge the stability of the equilibrium configuration .
the statement of chandrasekhar ( 1939 ) is only valid because the assumption of a conventional polytropic gas excludes g@xmath1modes , leaving them as convectively neutral .
in fact , linear stability properties rely on the radial distribution of the specific entropy . for instance
, we find in this paper g@xmath3mode instability for a static general polytropic lane - emden sphere with the square of the brunt - v@xmath5is@xmath5l@xmath5 frequency @xmath145 in some parts .
we therefore conclude that simply using the polytropic index @xmath20 to judge the stability property of a static sphere is not enough for a general polytropic lane - emden sphere .
one still needs to perform a detailed linear stability analysis for such a hydrostatic equilibrium configuration . | for dyanmic background models of goldreich & weber and lou & cao , we examine three - dimensional perturbation properties of oscillations and instabilities in a general polytropic homologously collapsing stellar core of a relativistically hot medium with a polytropic index @xmath0 .
perturbation behaviours , especially internal gravity g@xmath1modes , depend on the variation of specific entropy in the collapsing core . among possible perturbations ,
we identify acoustic p@xmath1modes and surface f@xmath1modes as well as internal gravity g@xmath2modes and g@xmath3modes . as in stellar oscillations of a static star ,
we define g@xmath2 and g@xmath3modes by the sign of the brunt - v@xmath4is@xmath5l@xmath5 buoyancy frequency squared @xmath6 for a collapsing stellar core . a new criterion for the onset of convective instabilities is established for a homologous stellar core collapse .
we demonstrate that the global energy criterion of chandrasekhar is insufficient to warrant the stability of general polytropic equilibria .
we confirm the acoustic p@xmath1mode stability of goldreich & weber , even though their p@xmath1mode eigenvalues appear in systematic errors .
unstable modes include g@xmath3modes and sufficiently high - order g@xmath2modes , both corresponding to convective core instabilities .
such instabilities occur before the stellar core bounce , in contrast to instabilities in other models of supernova ( sn ) explosions . the breakdown of spherical symmetry happens earlier than expected in numerical simulations so far .
the formation and motion of the central compact object are speculated to be much affected by such g@xmath1mode instabilities . by estimates of typical parameters , unstable low - order @xmath7 g@xmath1modes may produce initial kicks of the central compact object .
other high - order and high - degree unstable g@xmath1modes may shred the nascent neutron core into pieces without an eventual compact remnant ( e.g. sn1987a ) .
formation of binary pulsars and planets around neutron stars might originate from unstable @xmath8 g@xmath1modes and high - order high - degree g@xmath1modes , respectively .
hydrodynamics instabilities stars : neutron stars : oscillations ( including pulsations ) supernovae : general waves |
bi@xmath0sr@xmath0cacu@xmath0o@xmath1 ( bi-2212 ) and bi@xmath0sr@xmath0co@xmath0o@xmath6 ( bico ) commonly have a bi@xmath0o@xmath0 block layer , in which the lattice is modulated by the excess oxygen .
although these two materials have a similar crystal structure of alternating stacks of the bi@xmath0o@xmath0 layer and the cuo@xmath0 / coo@xmath0 layer , they show quite different physical properties ; bi-2212 is known as a high - temperature superconductor , while bico is known as a thermoelectric material @xcite .
high resolution transmission electron microscope studies of the bi - based superconductors ( including bi-2212 ) have clarified that they have strongly modulated structure along @xmath2-axis and that modulation periodicity is reported about 4.8@xmath2 for pb - free bi-2212 @xcite .
this modulated structure disappears in pb - doped bi-2212 @xcite .
bico was first thought as having a structure isomorphic to that of bi-2212 .
however , it turned out to have a misfit layer structure along the @xmath2 axis , which consists of alternating stacks of the rock - salt bi@xmath0o@xmath0 layers and the hexagonal coo@xmath0 layers @xcite .
the modulation structure also exists in the pure bico although its direction is tilted by 45@xmath5 from @xmath3- and @xmath2-axes in contrast with bi-2212 .
since the cuo@xmath0 and coo@xmath0 layers are responsible for electric conduction , it is interesting how the electric properties are affected by the modulation and the misfit structure . here
, we report on the in - plane anisotropy on the resistivity and thermopower of the two compounds .
-linear extrapolation ( @xmath7 and @xmath8 ) .
( b ) : the change in conductivity @xmath9=@xmath10/@xmath11 .
, width=264 ] -linear extrapolation ( @xmath7 and @xmath8 ) .
( b ) : the change in conductivity @xmath9=@xmath10/@xmath11 .
, width=264 ]
superconducting compound bi-2212 with bi : sr=2.2:1.8 and thermoelectric material pb - free and 20% pb - doped bico ( bi@xmath12pb@xmath13sr@xmath0co@xmath0o@xmath6 ) were used in this study .
single crystals of bi-2212 and bico was grown by a traveling solvent floating zone method .
starting composition of bi-2212 and bico was bi@xmath14sr@xmath15cacu@xmath0o@xmath1 and ( bi , pb)@xmath0sr@xmath0co@xmath0o@xmath6 respectively . the growth velocity was adopted to be 0.5mm / h , and the growth atmosphere was a mixed gas flow of o@xmath0 ( 20% ) and ar ( 80% ) .
structural analysis for bico was performed by using 4-axis x - ray diffractmater ( mac science mxc ) and transmission electron microscope ( jeol jem-100cx ) .
pb - doped bico was found to monoclinic symmetry with lattice parameter of a=4.91@xmath16 , b=5.22@xmath16 , c=29.98@xmath16 , @xmath17 and @xmath1890.0@xmath19 , @xmath2092.4@xmath19 . in tem diffraction patterns
, we clearly observed a rock - salt diffraction pattern from the bio layer and a hexagonal diffraction pattern from the coo@xmath0 layer as was previously reported in ref .
the montgomery method was adopted to measure the in - plane resistivity of bi-2212 because they show rather small anisotropy between the @xmath3- and @xmath2-axes . on the other hand ,
the resistivity of bico was measured using a standard four probe method , where the current contacts were carefully mounted on the side surface ( @xmath21- or @xmath22-surface ) so that current density should be homogeneous everywhere in the sample .
the thermopower was measured using a steady - state technique from 4.2 to 300k .
figure 1 shows the temperature dependence of the @xmath3- and @xmath2- axis resitstivities ( @xmath23 and @xmath24 ) for various doping levels . the magnitudes and the slopes @xmath25/@xmath26 monotonically decrease with increasing oxygen contents @xmath27 , indicating that the carriers are actually doped with increasing @xmath27 @xcite .
@xmath24 is higher than @xmath23 , where @xmath28 is roughly expressed as a sum of @xmath29 and a t - independent residual resistivity ( @xmath30 ) at high temperature .
an underdoped sample shows a typical downward deviation from high temperature @xmath31-linear behavior below a characteristic temperature @xmath4 , which implies pseudogap formation . in fig . 2 ( a ) , we show the magnitudes of the deviation from @xmath31-linear extrapolation ( @xmath7 and @xmath8 ) for oxygen content @xmath320.23 , and 0.25 . as seen in fig .
2(a ) , @xmath7 and @xmath8 deviate each other below the pseudogap temperature @xmath4 , indicating that the decrease of @xmath24 is more rapid than that of @xmath23 .
although @xmath7 and @xmath8 are anisotropic in the form of resistivity , there is no in - plane anisotropy in the deviation in conductivity ( @xmath9=@xmath10/@xmath11 ) as seen in fig .
sr@xmath15cacu@xmath0o@xmath1 single crystal for various doping levels measured along the @xmath3- and @xmath2- axes . ,
width=264 ] figure 3 shows the temperature dependence of the thermopower in various doping levels measured along the @xmath3- and @xmath2- axes .
the magnitude of the thermopower decreases with increasing @xmath27 , also indicating that the excess oxygen causes the carrier doping . typical negative slope at high temperature was observed in all the doping levels .
in contrast with the in - plane resistivities , there is no anisotropy in the thermopower within our experimental error of 10% .
these results suggest that the modulation structure along the b - axis works as an anisotropic scattering center above @xmath4 , but does not affect the pseudogap value . in principle , thermopower does not attribute to the scattering rate . in yba@xmath0cu@xmath33o@xmath34 ( ybco ) case , since the cuo chain also attributes to the electron conduction , there exists large anisotropy in the normal - state resistivity . in superconducting state ,
angle - resolved photoemission spectroscopy ( arpes ) measurement reported a significant ( @xmath35 ) difference in the gap magnitude between @xmath36 and @xmath37 point @xcite .
however , this kind of gap anisotropy has not been reported in arpes measurement for bi-2212 . in the superconducting state in the bi-2212
, optical conductivity shows an upturn behavior at low frequencies for the @xmath2-axis polarization that is not present in the @xmath3-axis polarization @xcite .
this upturn behavior is very simlarly seen in the optical conductivity of zn doped ybco @xcite .
this indicates the presence of low - lying excitations along the @xmath2-axis , perhaps due to an additional scattering caused by the modulation structure .
thus , an anisotropy in the scattering rate exists not only in the normal state but also in the superconducting state .
[ t ] -axis resistivity @xmath29 and @xmath2-axis resistivity @xmath28 of ( bi , pb)-sr - co - o single crystal measured by the four - probe method .
, width=264 ] figure 4 shows the temperature dependence of the resistivities for the @xmath3- and @xmath2-axis directions .
pb - doped bico exhibits lower resistivities than pb - free bico , suggesting that carrier is doped through the substitution of divalent pb@xmath38 for trivalent bi@xmath39 . as for the pb - free sample@xmath40 and @xmath24 exhibit metallic behavior near room temperature and insulating behavior below 50k and the in - plane
anisotropy is relatively small . on the other hand
, the pb - doped sample shows a significant anisotropy at low temperature ; @xmath23 remains metallic down to 4.2k , whereas @xmath24 shows an upturn near 50k . - and @xmath2-axes .
, width=264 ] figure 5 shows the temperature dependence of the @xmath3- and @xmath2-axis thermopower for the pb - free and pb - doped samples .
in contrast with bi-2212 , they exhibit a large anisotropy , where the @xmath2-axis thermopower is about two times larger than the @xmath3-axis thermopower around 50k . at room temperature ,
the @xmath3-axis thermopower is larger than the @xmath2-axis thermopower by about 10@xmath41v / k .
temperature independent thermopower at high temperature is often explained in terms of an entropy per site ( the heikes formula ) @xcite , which is in principal isotropic .
thus the heikes formula is unlikely to be applicable to the thermopower of bico .
the large anisotropy in the pb - doped sample would come from the misfit structure along the @xmath2-axis . on the other hand , in the case of the resistivities of the pb - free sample
, anisotropy would be averaged by the modulation structure , whose direction is tilted by 45@xmath5 from the @xmath3- and @xmath2-axes . it was pointed out that the physical properties of the layered cobalt oxides resemble those of the ce - based compounds @xcite .
the thermopower of cecu@xmath0ge@xmath0 increases with increasing pressure . in the ce - based compound
, pressure shrinks the lattice to increase the hybridization between the ce 4@xmath42 state and the conduction band .
the large thermopower of bico would be also attributed to the increase of hybridization , which comes from chemical pressure due to the misfit structure .
in conclusion , we investigated the in - plane anisotropy on the resistivity and thermopower of the bi-2212 and bi - sr - co - o .
we found in bi-2212 that the modulation structure causes the residual resistivity on @xmath24 as @xmath24=@xmath23+@xmath43 .
however , it does not affect the pseudogap formation and the thermopower . in the case of pure bico
, the in - plane anisotropy of the resistivity is averaged by the modulation structure , whose direction is tilted by 45@xmath5 from the @xmath3- and @xmath2-axes . on the other hand ,
the misfit structure makes the resistivities and thermopowers of bico anisotropic .
the large anisotropy of bico would be attributed to the chemical pressure due to the misfit structure .
99 t. itoh , i. terasaki , jpn .
* 39 * ( 2000 ) 6658 .
y. matsui , h. maeda , y. tanaka , s. horiuchi , jpn .
* 27 * ( 1988 ) l372 .
y. matsui , a. maeda , k. uchinokura , s. takekawa , jpn . j. appl
* 29 * ( 1990 ) l273 .
h. leligny , d. grebille , o. prez , a. -c , masset , m. hervieu , c. michel , b. raveau , c. r. acad .
paris t.2 , iic ( 1999 ) 409 .
t. yamamoto , i. tsukada , k. uchinokura , m. takagi , t. tsubone , m. ichihara , k. kobayashi , jpn . j. appl
* 39 * ( 2000 ) l747 .
t. watanabe , t. fujii , a. matsuda , phys .
* 79 * , 2113 ( 1997 ) .
d. h. lu , d. l. feng , n. p. armitage , k. m. shen , a. damascelli , c. kim , f. ronning , z. -x .
shen , d. a. bonn , r. liang , w. n. hardy , a. i. rykov , s. tajima , phys .
86 ( 2001 ) 4370 .
m. a. quijada , d. b. tanner , r. j. kelley , m. onellion , h. berger , g. margaritondo , phys . rev . * b * 60 ( 1999 ) 14917 .
m. l. wang , s. tajima , a. i. rykov , k. tomimoto , phys . rev . *
b * 57 ( 1998 ) r11081 .
w. koshibae , s. maekawa , phys . rev . * b * 62 ( 2000 ) 6869 .
i. terasaki , _ materials transactions_. * 42 * ( 2001 ) 951 . | we investigated the in - plane anisotropy on the resistivity and thermopower of the bi@xmath0sr@xmath0cacu@xmath0o@xmath1 ( bi-2212 ) and bi - sr - co - o ( bico ) single crystals . in bi-2212 ,
the @xmath2-axis resistivity is higher than the @xmath3-axis resistivity , and is expressed as a sum of the @xmath3-axis resistivity and an additional residual resistivity .
a downward deviation due to pseudogap is observed below a characteristic temperature @xmath4 , which is isotropic in the form of conductivity .
these results suggest that the modulation structure along the @xmath2-axis works as an anisotropic scattering center , but does not affect the pseudogap formation . on the other hand , the anisotropy of the resistivity and the thermopower in pb - doped bico is substantial , probably owing to the misfit structure between the hexagonal coo@xmath0 layer and the rock salt bi@xmath0o@xmath0 layer
. however , the anisotropy in the resistivity in pb - free bico is very small , suggesting that the in - plane anisotropy is averaged by the modulation structure , whose direction is tilted by 45@xmath5 from the @xmath3- and @xmath2-axes . |
let @xmath11 be a finite subset of a commutative group . for any integer @xmath12 , a _ @xmath13-term arithmetic progression _ ( or @xmath13-ap , for short ) is an ( ordered ) sequence of @xmath13 elements in @xmath11 of the form @xmath14 , where @xmath15 is the _ first element _ and @xmath16 is the _
common difference_. a @xmath3-coloring of @xmath11 is a map @xmath17 .
a @xmath13-ap @xmath18 is _ monochromatic _ if @xmath19 .
let @xmath20 be the number of monochromatic @xmath13-aps in the 2-coloring @xmath21 .
a natural question is how small @xmath20 can be ? let @xmath22 be the number of all @xmath13-aps in @xmath11 .
define @xmath23 we are interested in the asymptotic value of @xmath24 as @xmath25 approaches infinity .
( this is similar to those questions on schur triples @xcite or on general patterns @xcite . ) in this paper , we consider only the cases that @xmath26 $ ] and @xmath27 . here
@xmath9=\{1,2,\ldots , n\}$ ] and @xmath5 is the cyclic group of order @xmath6 . when @xmath6 is a prime number @xmath1 , we write @xmath5 as @xmath0 for emphasis .
a @xmath13-ap is _ degenerated _ if it contains repeated terms ; it is _ non - degenerated _ otherwise .
the _ mirror image _ of a @xmath13-ap @xmath28 is another @xmath13-ap @xmath29 .
here we allow @xmath13-aps to be degenerated ; a @xmath13-ap differs from its mirror image except for @xmath30 .
in contrast , many papers require @xmath13-aps to be non - degenerated and treat a @xmath13-ap the same as its mirror image .
the two different definitions of @xmath13-aps derive two different versions of @xmath24 .
however , they are asymptotically equivalent as @xmath25 goes to infinity ; this is because the number of degenerated @xmath13-aps is only @xmath31 while the number of all aps is @xmath32 .
a @xmath13-ap @xmath18 is parametrized by a pair @xmath33 .
the parameter space of all @xmath13-aps in @xmath9 $ ] can be described as @xmath34 a @xmath13-ap @xmath18 in @xmath9 $ ] is degenerated if and only if @xmath30 . the parameter space of all @xmath13-aps in @xmath5 is simply @xmath35 .
a @xmath13-ap @xmath18 in @xmath5 is degenerated if @xmath36 for some @xmath37 . in both cases ,
the number of degenerated @xmath13-aps is @xmath31 .
random @xmath3-colorings of @xmath9 $ ] ( or @xmath5 ) give the following upper bounds . @xmath38)&\leq & 2^{1-k}+o(1);\\ \label{eq : randzn } m_k({{\mathbb z}}_n)&\leq & 2^{1-k}+o(1).\end{aligned}\ ] ] van der waerden s number @xcite @xmath39 can be used to provide a lower bound on @xmath40)$ ] .
for example , using a double counting method , one can prove @xmath40)\geq \frac{2(k-1)}{w^3}+o(1)$ ] ( see @xcite ) . a similar argument
can show @xmath41 .
these bounds are usually too weak ; stronger bounds exist for @xmath42 and @xmath43 .
the case @xmath0 is of particular interest .
the number of monochromatic @xmath44-aps in @xmath0 depends only on the size of the coloring classes , but not on the coloring itself ( see @xcite ) .
namely , if @xmath21 is a 2-coloring of @xmath0 such that the size of red class is @xmath45 , then we have @xmath46 the minimum is achieved when @xmath47 is closed to @xmath48 .
thus @xmath49 is achieved by random 2-colorings . for @xmath43 ,
wolf @xcite proved that for any sufficiently large prime number @xmath1 , we have @xmath50 this lower bound improved a previous lower bound due to cameron , cilleruelo , and serra @xcite , @xmath51 where @xmath6 is relatively prime to @xmath52 and large enough .
( cameron , cilleruelo , and serra s result actually holds for any abelian group of order @xmath6 provided @xmath53 . )
wolf s upper bound indicates that @xmath54 is not achieved by random 2-colorings .
this is a nice result .
however , the quantity is only slightly less than @xmath55 the density of monochromatic @xmath56-aps in random @xmath3-colorings .
her method for the upper bound relies heavily on the method initialized by gowers ( see @xcite ) .
the existence of such @xmath3-coloring is proved by probabilistic methods ; it is non - constructive . to get a better upper bound for @xmath57
, we introduce a construction consisting of periodic blocks . for
a fixed @xmath58 , let @xmath59 be a good @xmath3-coloring of @xmath60 with @xmath61 monochromatic @xmath13-aps .
( here @xmath59 is viewed as a 0 - 1 string of length @xmath58 . )
write @xmath62 with @xmath63 .
we consider the following periodic construction @xmath21 @xmath64 here @xmath65 is any bit - string of length @xmath66 .
if @xmath6 is divisible by @xmath58 , then @xmath65 is empty . in this case , it is easy to see that the periodic construction above gives @xmath67 monochromatic @xmath13-aps .
thus , we have @xmath68 if @xmath6 is not divisible by @xmath58 , then the computation of @xmath69 is more complicated in general .
note that the number of @xmath13-aps containing some element(s ) in @xmath65 is a lower order term as @xmath6 goes to infinity ; the value @xmath69 can be still determined asymptotically by @xmath59 .
( see lemma [ upper ] and [ le3 ] . )
two colorings @xmath21 and @xmath70 of @xmath5 are _ isomorphic _ if there is an integer @xmath71 such that @xmath72 and @xmath73 for any @xmath74 .
two colorings @xmath21 and @xmath70 of @xmath5 are _ conjugated _ if @xmath75 for any @xmath74 .
it is clear that @xmath76 , whenever @xmath21 and @xmath70 are isomorphic or conjugated to each other . to find a good coloring @xmath59 , we implement an efficient bread - first search algorithm reducing isomorphic copies . from the proof of the lower bound for @xmath77
, we need pay attention to those @xmath6 divisible by @xmath56 . using this efficient program
, we find a good @xmath3-coloring @xmath78 of @xmath79 for 4-aps , @xmath80 this coloring @xmath78 gives @xmath81 we also search the coloring of @xmath0 without any non - degenerated monochromatic @xmath56-aps . at @xmath82 ,
there is a unique coloring with this property up to isomorphisms . since @xmath83 s and @xmath84 s
are not balanced in this coloring , we search good colorings in @xmath85 instead .
we found a good @xmath3-coloring @xmath86 of @xmath85 for 4-aps , @xmath87 in this coloring , all monochromatic @xmath56-aps in @xmath86 are degenerated ; there are 22 monochromatic @xmath56-aps with @xmath30 and 20 monochromatic @xmath56-aps with @xmath88 .
this coloring @xmath86 gives @xmath89 the following theorem improves both wolf s lower bound and upper bound on @xmath54 .
our lower bound is obtained by combining wolf s elegant method and an exhaustive search .
our upper bound is proved by a novel method of analyzing the number of monochromatic @xmath13-aps in the periodic construction ( [ eq : pcon ] ) .
[ prime ] if @xmath1 is prime and large enough , then we have @xmath90 in fact , our methods naturally lead ( asymptotic ) bounds on @xmath77 for general @xmath6 .
the results depend on @xmath6 case - wisely . for simplicity ,
we split it into two theorems : one on the lower bound and the other one on the upper bound .
[ t : m4 lb ] if @xmath6 is sufficiently large , then we have @xmath91{ll } \frac{7}{96 } & \mbox { if $ n$ is not divisible by 4},\\ \vspace{-0.3cm}&\\ \frac{2}{33 } & \mbox { if $ n$ is divisible by 4 . } \\ \end{array } \right.\ ] ] here is a theorem for the upper bound on @xmath77 for general @xmath6 .
[ t : m4ub ] for @xmath6 sufficiently large , we have @xmath92{ll } \frac{17}{150}+o(1 ) < 0.1133334 & \mbox { if $ n$ is odd , } \\
\vspace{-0.3cm}&\\ \frac{8543}{72600}+o(1)<0.1176722 & \mbox { if $ n$ is even . }
\end{array } \right.\ ] ] note that theorem [ prime ] is a corollary of theorem [ t : m4 lb ] and [ t : m4ub ] .
the upper bound above is small enough to beat the bound @xmath55 reached by random 2-colorings . using inequality ( [ eq : divisible ] ) , we can get a much better bound for certain @xmath6 s . for example , @xmath93{ll } 0.09 & \mbox { if } 20\mid n , \\
0.086777 & \mbox { if } 22\mid n , \\
\end{array } \right.\ ] ] for @xmath94 , we use the periodic construction with the following good coloring of @xmath95 : + @xmath96 + @xmath97 all monochromatic @xmath98-aps in @xmath99 are degenerated ones . among them
there are 74 @xmath98-aps with @xmath30 and 72 @xmath98-aps with @xmath100 .
this coloring gives @xmath101 .
we have the following theorem .
[ t : m5ub ] if @xmath6 is sufficiently large , then we have @xmath102{ll } \frac{3629}{65712}+o(1)<0.055226 & \mbox { if $ n$ is odd , } \\
\vspace{-0.3cm}&\\ \frac{3647}{65712}++o(1)<0.0554998 & \mbox { if $ n$ is even . }
\end{array } \right.\ ] ] once again , the upper bound above is small enough to beat the bound @xmath103 , which is reached by random 2-colorings . using inequality ( [ eq : divisible ] ) , we can get a much better upper bound for certain @xmath6 . for example , we have @xmath104 the following theorem gives the best lower bounds ( for some @xmath6 s ) .
[ tzb ] we have @xmath105 in fact , we show for any @xmath106 , there is an odd integer @xmath6 with @xmath107 .
combining this result with theorem [ t : m4 lb ] , we get @xmath108 note the gap is pretty small . here
we conjecture that the upper bound is tight .
@xmath109 maybe it is true even if the condition that `` @xmath6 is not divisible by 4 '' is removed .
the periodic construction also works for @xmath40)$ ] .
we have [ l0 ] for any @xmath12 and any positive integer @xmath58 , we have @xmath110)\leq m_k({{\mathbb z}}_b).\ ] ] in particular , we have @xmath110)\leq \varliminf_{n\to\infty } m_k({{\mathbb z}}_n).\ ] ] when we consider the similar problems for @xmath9 $ ] , the @xmath13-ap and its mirror image are often not distinguished in the literature . to avoid the ambiguity
, we call a @xmath13-ap ( in @xmath9 $ ] ) with @xmath111 an _ increasing _ @xmath13-ap . for @xmath12 ,
let @xmath112 be the largest number satisfying `` for any @xmath113 , there is a sufficiently large @xmath6 such that any @xmath3-coloring of @xmath9 $ ] contains at least @xmath114 monochromatic increasing @xmath13-aps '' .
since @xmath9 $ ] has @xmath115 increasing @xmath13-aps , it is equivalent to say @xmath116).\ ] ] in 2008 , parillo , robertson , and saracino @xcite proved @xmath117 their construction was generalized by butler , costello , and graham , who @xcite proved @xmath118 and @xmath119 via an extensive computation .
both bounds beat random @xmath3-colorings . combining theorem [ tzb ] with lemma [ l0 ]
, we have @xmath120 these numerical results indicate that the periodic construction is often better than the block construction used in @xcite .
we believe the following conjecture holds . for fixed @xmath121
, we have @xmath122)= \varliminf_{n\to\infty } m_k({{\mathbb z}}_n)$ ] .
bounding @xmath123 is very different from bounding @xmath124 .
this conjecture above is not true for @xmath42 .
we have the following theorem .
[ t : z3 ] if the integer @xmath6 is large enough , then any @xmath3-coloring of @xmath5 contains at least @xmath125 monochromatic arithmetic progressions .
in particular , we have @xmath126 with the help of computer search , we found three good 2-colorings @xmath78 , @xmath86 , and @xmath99 , which are used as building blocks in constructing good 2-colorings of @xmath5 and @xmath9 $ ] . the data in table [ tab:3 ] , [ tab:4 ] , [ tab:5 ] , and [ tab:6 ] , can be easily verified by anyone with limited programming experience .
some lower bound requires nontrivial exhaustive search in the same way as cameron , cilleruelo , and serra @xcite proved the previous lower bounds .
however , those lower bounds using an exhaustive computer search are not the focus of this paper .
the organization of the paper is following . in section 2 , we will prove a necessary lemma and theorem [ t : z3 ] . in the section 3 ,
we first prove a lemma and a corollary on counting lattice points in a polygon ; then we prove theorem [ t : m4ub ] for odd @xmath6 and theorem [ t : m5ub ] . in section 4 ,
we introduce a recursive construction and then use it to prove theorem [ tzb ] and theorem [ t : m4ub ] for even @xmath6 . in the last section , we will deal with the lower bounds and prove theorem [ t : m4 lb ] .
let @xmath127 be a @xmath3-coloring of @xmath128 . the coloring @xmath21 is often viewed as a bit - string of length @xmath6 . for convenience ,
we say an element @xmath74 is _ red _ if @xmath129 and _ blue _ if @xmath130 .
the coloring @xmath21 induces a partition @xmath131 , where @xmath132 is the set of red elements while @xmath59 is the set of blue elements .
let @xmath133 be an integer and @xmath134 .
we have @xmath135 . for each @xmath136 , let @xmath137 ( or @xmath138 ) be the set of all @xmath13-aps whose @xmath139-th number is red ( or blue ) , respectively ; we have @xmath140 [ l1 ] for @xmath141 , if @xmath142 , then we have @xmath143 if @xmath144
, then we have @xmath145 * proof : * for @xmath141 , the value of @xmath146 equals the number of @xmath13-aps whose @xmath139-th and @xmath147-th terms are red .
if @xmath142 , then every ordered pair of elements ( distinct or not ) in @xmath5 can be extended into a unique @xmath13-ap whose @xmath139-th and @xmath147-th terms are the given pair .
note the number of ordered pairs of red ( and blue ) elements is exactly @xmath148 ( and @xmath149 ) , respectively .
equations ( [ eq : aij ] ) and ( [ eq : bij ] ) follow .
if @xmath150 , then every pair of elements in @xmath5 may or may not be extended into a @xmath13-ap whose @xmath139-th and @xmath147-th terms are the given pair .
let @xmath151 .
for @xmath152 , let @xmath153 be the number elements @xmath154 in @xmath5 such that @xmath154 is red and @xmath155 .
for any pair @xmath156 , the elements @xmath157 and @xmath158 are the @xmath139-th and @xmath147-th elements of an arithmetic progression if @xmath159 for some element @xmath16 in @xmath5 .
equivalently , @xmath160 . moreover , if @xmath161 , then equation ( [ eq : d ] ) has @xmath66 solutions .
it follows that @xmath162 equation ( [ eq : bij1 ] ) can be proved similarly .
@xmath163 * proof of theorem [ t : z3 ] : * observe that if we assign red and blue to each number equally likely , then the expected value of @xmath164 is @xmath165 . therefore , there is a 2-coloring @xmath21 such that @xmath166 , that is @xmath167 . for the other direction ,
let @xmath21 be any 2-coloring of @xmath5 .
we use the notations @xmath47 , @xmath137 , and @xmath138 defined in the beginning of this section .
we have the following inclusion - exclusion formula .
@xmath168 note that @xmath169 and @xmath170 by the definition of @xmath164 , we have @xmath171 .
applying lemma [ l1 ] , we have @xmath172 note that @xmath173 reaches the maximum value at @xmath174 .
we have @xmath175
. therefore @xmath176 and the lemma follows .
in this section , we will examine the number of monochromatic @xmath13-aps in the periodic construction ( [ eq : pcon ] ) .
we need a tool to count the grid points inside a polygon on the plane .
a point in @xmath177 is a _ grid point _ if both coordinates are integers .
let @xmath178 be a simple polygon whose vertices are grid points .
let @xmath179 be the area of @xmath178 , @xmath180 be the number of grid points inside @xmath178 , and @xmath181 be the number of grid points on the boundary of @xmath178 .
the classical pick s theorem @xcite states @xmath182 intuitively , if @xmath181 is a lower order term , then @xmath183 .
let @xmath184 be a simple polygon in the plane @xmath177 .
for any @xmath185 and a point @xmath186 , a new polygon @xmath187 is obtained by first scaling @xmath184 by a factor of @xmath188 and then translating by a vector @xmath186 .
we have the following lemma .
suppose @xmath184 is a simple polygon with @xmath71 vertices and circumference @xmath189 .
for any point @xmath186 and sufficiently large @xmath188 , we have @xmath190 * proof : * since @xmath184 has @xmath71 vertices , let @xmath191 be the vertices of the polygon @xmath187 . for @xmath192 , let @xmath193 be a grid point closest to @xmath194 ( if there are more than one choice , then break ties arbitrarily ) .
we have @xmath195 .
let @xmath178 be the polygon with vertices @xmath196 .
( for convenience , we write @xmath197 and @xmath198 . )
the polygon @xmath178 can be viewed as an approximation of the polygon @xmath187 ; thus @xmath178 is simple for sufficiently large @xmath188 . applying pick s theorem to @xmath178 , we have @xmath199 we observe that the number of grid points on any line segment of length @xmath200 is at most @xmath201 .
we have @xmath202 let @xmath203 be the convex region spanned by @xmath204 .
note @xmath203 is covered by four triangles @xmath205 , @xmath206 , @xmath207 , and @xmath208 exactly twice .
we have @xmath209 summing up , we get @xmath210 let @xmath211 be the set of grid points inside @xmath203 or on the line segment @xmath212 .
let @xmath213 be the convex set spanned by @xmath211 . applying pick s theorem to @xmath213 , we have @xmath214 thus @xmath215 summing up , we get @xmath216 putting together , we have @xmath217 the proof of this lemma is finished
. @xmath163 by counting grid points in @xmath218 , we get the following corollary .
since the number of grid points on the boundary of @xmath219 is always a lower order term , it does not matter whether grid points on the boundary are included or not .
in fact , in latter applications , the polygon region @xmath184 is often defined with one part of boundary included while the other part of boundary excluded .
[ lattice ] for any fixed point @xmath220 , let @xmath221 be a lattice @xmath222 and @xmath184 be a simple polygon . for @xmath223 , we have @xmath224 * proof of lemma [ l0 ] : * let @xmath59 be a `` good '' @xmath3-coloring / bit - string of @xmath60 with @xmath61 monochromatic @xmath13-aps .
any @xmath13-ap in @xmath60 can be parametrized by a pair @xmath225 satisfying @xmath226 .
let @xmath227 be the set of parameters @xmath225 such that the corresponding @xmath13-aps in @xmath60 are monochromatic .
we have @xmath228 for sufficiently large @xmath6 , we write @xmath62 with @xmath63 . consider the periodic construction @xmath229 ( see ( [ eq : pcon ] ) ) .
note that the number of @xmath13-aps containing some elements of @xmath65 is @xmath31 .
we need estimate monochromatic @xmath13-aps lying entirely in @xmath230 $ ] .
let @xmath184 be a parallelogram defined by @xmath231 the area of @xmath184 is clearly @xmath232 .
a @xmath13-ap @xmath233 in @xmath230 $ ] is monochromatic if and only if @xmath234 .
let @xmath235 be the lattice @xmath236 .
applying corollary [ lattice ] , the number of monochromatic @xmath13-aps in @xmath230 $ ] is exactly @xmath237 thus , @xmath238,c)=\frac{1}{k-1 } m_k({{\mathbb z}}_b)(bt)^2+o(b^2t).\ ] ] note that the number of @xmath13-aps in @xmath9 $ ] is @xmath239 .
taking the ratio , we have @xmath238)\leq \frac{m_k([n],c)}{ap([n])}=m_k({{\mathbb z}}_b ) + o(\frac{1}{t}).\ ] ] first taking ( upper ) limit as @xmath6 goes to infinity , we get @xmath240)\leq m_k({{\mathbb z}}_b ) .\ ] ] then taking ( lower ) limit as @xmath58 goes to infinity , we have @xmath240)\leq \varliminf_{b\to\infty}m_k({{\mathbb z}}_b).\ ] ] the proof of lemma [ l0 ] is finished .
@xmath163 it suffices to consider the case that @xmath6 is not divisible by @xmath58 .
write @xmath62 with @xmath241 .
recall the periodic construction @xmath242 here @xmath65 is any bit - string of length @xmath66 .
the number of @xmath56-aps containing some bit(s ) in @xmath65 is @xmath31 .
the major term in the number of all monochromatic @xmath56-aps depends only on @xmath59 and @xmath66 .
we divide the set of all non - degenerated @xmath56-aps in @xmath5 into eight classes @xmath243 for @xmath244 . [ cols="^,^",options="header " , ] by lemma [ le3 ] , we have @xmath245{ll } 3629/65712 + o(1 ) & \mbox { for odd } r \not=37,\\ 289/10952 + o(1 ) & r = 37,\\ 3647/65712 + o(1 ) & \mbox { for even } r \not=0.\\ \end{array } \right.\ ] ] the proof of theorem [ t : m5ub ] is finished .
recall the good coloring of @xmath85 : @xmath87 we observe that the first 11 coordinates and last 11 coordinates differ only by @xmath84 bit .
let @xmath246 , where ` @xmath247 ' could be either @xmath83 or @xmath84 . @xmath248 has the following property .
( this is because @xmath86 contains no non - degenerated monochromatic @xmath56-aps . ) * proof : * let @xmath252 be a 2-coloring / bit - string of @xmath253 which has exactly @xmath254 monochromatic @xmath56-aps . first we consider the periodic construction @xmath255 each block @xmath248 has exactly one ` @xmath247 ' ; there are @xmath188 @xmath247 s in total .
finally , we replace these @xmath247 s by the values of @xmath252 in the cyclic order .
we denote the coloring by @xmath256 .
( for example , @xmath257 . )
because of property [ b11 ] , a @xmath56-ap of @xmath256 with parameter @xmath33 is monochromatic only if @xmath258 .
the number of monochromatic @xmath56-aps is exactly @xmath259 we have @xmath260 the proof of this lemma is finished .
let @xmath252 be a 2-coloring / bit - string of @xmath253 which has exactly @xmath262 monochromatic @xmath56-aps .
we can define @xmath263 similarly .
for example , @xmath264 . note that @xmath99 contains no non - degenerated monochromatic @xmath98-aps .
we have the following property .
* proof of theorem [ t : m4ub ] for even @xmath6 : * here we assume @xmath6 is even and not divisible by @xmath271 .
let @xmath272 which is a @xmath3-coloring of @xmath273 .
write @xmath274 with @xmath275 . here
@xmath66 is even and not divisible by @xmath271 .
consider the periodic construction @xmath229 as before . for these @xmath66 , the values
@xmath276 depends only on @xmath139 but not on @xmath66 .
these values are given in table [ tab:6 ] .
* proof of theorem [ t : m4 lb ] : * given a 2-coloring @xmath21 of @xmath282 , we will establish an inequality which is similar to equation ( 4.8 ) in @xcite . for each @xmath283
, let @xmath193 be the number of @xmath56-aps with exactly @xmath139 red numbers .
we have @xmath284 note that @xmath285 ) , we get @xmath286 by the inclusion - exclusion formula , for any @xmath287 , we have @xmath288 since @xmath289 , we have @xmath290 by the symmetry of @xmath137 s and @xmath138 s , we get @xmath291 combining equations ( [ eq : ab1 ] ) and ( [ eq : ab2 ] ) and summing over @xmath292 , we get @xmath293 combining the equation above with equation ( [ eq : ab0 ] ) , we have @xmath294 lemma [ l1 ] implies @xmath295 and @xmath296 for @xmath297 .
we get @xmath298 let @xmath299 be the collection of all even - colored 4-term progressions and @xmath300 be the collection of all odd - colored 4-term progressions .
we have @xmath301 and @xmath302 .
inequality ( [ eq:6 ] ) together with @xmath303 give that @xmath304 we aim to modify the method in @xcite to find a lower bound on @xmath305 which gives a lower bound on @xmath306 .
assume @xmath227 is a 3-term progression in @xmath5 .
let @xmath307 be the number of even - colored 4-aps containing @xmath227 and @xmath308 be the number of odd - colored 4-aps containing @xmath227 .
observe that @xmath309 .
if @xmath310 is a 3-ap , then it determines a pair of integers @xmath311 such that @xmath312 and @xmath313 are two 4-aps containing @xmath227 ; the pair @xmath314 is the _ frame pair _ of @xmath227 .
we have @xmath315 where the expectation operator @xmath316 runs over all 3-aps . the following equality which ensures us to obtain a lower bound on @xmath299 .
we have @xmath317 solving for @xmath305 in equation ( [ eq:9 ] ) gives @xmath318 we have the following claim which will be proved at the end of this section .
observe that @xmath320 if and only if the frame pair of @xmath227 is monochromatic .
furthermore , @xmath321 if @xmath322 .
note that when @xmath6 is prime , each frame pair belongs to a unique 3-term progression as 4 is invertible in @xmath5 .
however , if @xmath6 is not prime , then each frame pair may belong to more than one 3-term progression or does not belong to any 3-term progressions .
we will compute the value of @xmath323 case by case according to @xmath6 modulo @xmath56 . * case 1 : * @xmath324 . in this case
, each frame pair belongs to a unique 3-term progression since 4 is invertible in @xmath5 .
we have @xmath325 equals twice of the number of monochromatic pairs in the coloring @xmath21 , that is @xmath326 .
we obtain @xmath327 by lemma [ l1 ] , we have @xmath328 and @xmath329 .
therefore , in this case , inequality ( [ eq:7 ] ) is @xmath330 it is straightforward to check that the minimum value of the right hand side of inequality ( [ eq:11 ] ) is @xmath331 and it is achieved at @xmath174 .
we have @xmath332 in this case .
* case 2 : * @xmath333 . for @xmath334 ,
let @xmath335 , @xmath336 , and @xmath337 .
a pair @xmath314 is a frame pair if and only if @xmath338 for some @xmath339 .
assume @xmath340 .
if @xmath16 is a solution for @xmath341 , then @xmath342 is another solution .
we have @xmath343 by the same argument , we have @xmath344 and @xmath345 therefore , @xmath346 is at least @xmath347 we have the following inequality @xmath348 combining inequalities ( [ eq:7 ] ) and ( [ eq:13 ] ) , we get @xmath349 note the minimum is reached at @xmath174 .
it follows @xmath350 .
* case 3 : * @xmath351 .
this method fails in this case ; which suggests that it is possible to find a good 2-coloring of @xmath282 which contains few monochromatic 4-term progressions when @xmath6 is a multiple of @xmath56 . replacing the terms on the right hand side of inequality ( [ eq:6 ] ) by the lower bounds from lemma [ l1 ] ,
we obtain @xmath352 combining with @xmath353 , we have @xmath354 the remark following the proof of theorem 4.4 in @xcite gives @xmath355 combining inequalities ( [ eq:15 ] ) and ( [ eq:16 ] ) , we get @xmath356 it implies @xmath357 .
we completed the proof of theorem [ t : m4 lb ] .
@xmath163 moreover , for each @xmath227 , @xmath360 if @xmath358
. therefore the value of @xmath361 is twice of the number of increasing 5-term progressions with coloring pattern from @xmath362 . using an exhaustive search
, one can show that for any 2-coloring of @xmath363 $ ] , there is at least one increasing 5-ap of coloring pattern in @xmath362 .
a further computation shows that any @xmath3-coloring of @xmath364 $ ] contains at least 27 increasing @xmath98-aps of coloring pattern in @xmath362 .
note that the number of increasing 5-aps in @xmath364 $ ] with @xmath365 is @xmath366 , the number of 5-aps in @xmath364 $ ] with @xmath367 is @xmath368 , etc .
the number of @xmath98-aps in @xmath364 $ ] is @xmath369 for any @xmath3-coloring of @xmath5 , the number of @xmath370-aps is exactly @xmath371 ; each of them ( degenerated or not ) contains @xmath372 @xmath98-aps of coloring pattern in @xmath362 .
each @xmath98-ap with coloring pattern in @xmath362 is counted at most @xmath373-times .
thus , the number of 5-aps with coloring pattern in @xmath362 is at least @xmath374 thus we have @xmath375 we finished the proof of the claim .
@xmath163 9 s. butler , k. costello and r. graham . finding patterns avoiding many monochromatic constellations .
_ experiment .
math_. * 19 * ( 2010 ) , 399 - 411 .
p. cameron , j. cilleruelo and o. serra . on monochromatic solutions of equations in groups .
_ rev . mat . iberoam_. * 23*(2007 ) , 385 - 395 .
b. datskovsky . on the number of monochromatic schur triples .
math_. * 31 * ( 2003),193 - 198 .
p. frankl , r. graham and v. rdl .
quantitative theorems for regular systems of equations .
_ j. combin .
theory ser .
a_. * 47 * ( 1988 ) , 246 - 261 .
a. robertson and d. zeilberger .
a 2-coloring of @xmath376 $ ] can have @xmath377 monochromatic schur triples , but not less!. _ elec .
j. combin_. * 5 * ( 1998 ) , research paper 19 . t. schoen . the number of monochromatic schur triples .
j. combinatorics_. * 20 * ( 1999 ) , 855- 866 . | this paper is motivated by a recent result of wolf @xcite on the minimum number of monochromatic 4-term arithmetic progressions ( 4-aps , for short ) in @xmath0 , where @xmath1 is a prime number .
wolf proved that there is a 2-coloring of @xmath0 with @xmath2 fewer monochromatic 4-aps than random @xmath3-colorings ; the proof is probabilistic and non - constructive . in this paper , we present an explicit and simple construction of a @xmath3-coloring with @xmath4 fewer monochromatic 4-aps than random @xmath3-colorings .
this problem leads us to consider the minimum number of monochromatic 4-aps in @xmath5 for general @xmath6 .
we obtain both lower bound and upper bound on the minimum number of monochromatic 4-aps in all 2-colorings of @xmath5 .
wolf proved that any @xmath3-coloring of @xmath0 has at least @xmath7 monochromatic 4-aps .
we improve this lower bound into @xmath8 .
our results on @xmath5 naturally apply to the similar problem on @xmath9 $ ] ( i.e. , @xmath10 ) . in 2008 , parillo , robertson , and saracino @xcite constructed a @xmath3-coloring of @xmath9 $ ] with 14.6% fewer monochromatic 3-aps than random @xmath3-colorings . in 2010 ,
butler , costello , and graham @xcite extended their methods and used an extensive computer search to construct a @xmath3-coloring of @xmath9 $ ] with 17.35% fewer monochromatic 4-aps ( and 26.8% fewer monochromatic 5-aps ) than random @xmath3-colorings .
our construction gives a @xmath3-coloring of @xmath9 $ ] with 33.33% fewer monochromatic 4-aps ( and 57.89% fewer monochromatic 5-aps ) than random 2-colorings . |
during the last two decades , there has been seen an increasing scientific and technological interest in spin related phenomena and in possible application of these in future data processing , communication and storage.@xcite quantum dots have been proposed as components in few - electron spintronics devices , such as spin filters or spin memories,@xcite and in spin - based quantum computation devices.@xcite existing fabrication techniques allow construction of semiconductor quantum rings of nanometer dimensions containing only a few electrons.@xcite nanoscopic quantum rings are sufficiently small systems to show quantum effects , and large enough to be able to trap magnetic flux in their interior , when subjected to an experimentally reachable magnetic field .
the trapping of magnetic flux quanta gives rise to interesting effects , such as persistent currents and other periodic properties related to the aharonov
bohm effect.@xcite the main focus of this paper is on the magnetic properties of the quantum rings .
the previous study by koskinen et al.@xcite for one type of quantum rings has shown that a model of localized charges and antiferromagnetic coupling of the nearest - neighbor spins , corresponding to an antiferromagnetic heisenberg model , captures the physics of systems they study .
this means , for example , that when the angular momentum of the system is increased by making the magnetic field stronger , the change of the total spin of the system follows values found using the antiferromagnetic heisenberg model . on the other hand ,
a study of a hard - wall quantum dot@xcite where strong electron - electron interaction forces the electron density to be ring - shaped , has found a ferromagnetic behavior of the system as a function of the magnetic field
. motivated by this discrepancy , we study tunable quantum rings and aim to identify the underlying physics that leads to different magnetic properties for the various quantum rings .
our results show interestingly that the same quantum ring system can show oscillation between ferromagnetic and antiferromagnetic behavior as a function of the magnetic field .
a similar control of the magnetism can be obtained by changing the width or the radius of the quantum ring .
we use an effective - mass approximation and model the semiconductor quantum rings as two - dimensional systems with the hamiltonian @xmath0 where @xmath1 is the number of electrons , @xmath2 is the vector potential of the perpendicular magnetic field @xmath3 , @xmath4 is the external confinement potential , @xmath5 is the effective electron mass , @xmath6 is the coulomb constant ( normally 1 ) , and @xmath7 is the dielectric constant .
the electrons are restricted to the cartesian @xmath8 plane , and the magnetic field is normal to the plane . in our calculations , we use the effective parameter values of gaas , namely , @xmath9 0.067@xmath10 and @xmath11 12.7 . to enhance the spin effects ,
we have left out the zeeman potential .
this can be justified since experimentally the gyromagnetic factor , and thereby also the zeeman term , can be made vanishingly small.@xcite to control the spin effects , we use confinement potentials of the form @xmath12 where @xmath13 determines the confinement strength and @xmath14 determines the strength of the gaussian perturbation .
we set @xmath15 5 mev and @xmath16 2 @xmath17 , where @xmath18 10.03 nm is the effective bohr radius of gaas . by tuning the gaussian perturbation stronger , we can make the system more and more ring - shaped .
in addition , we have used a parabolic ring potential given by @xmath19 to model a narrow quantum ring . in the exact diagonalization ( ed ) calculations ,
our single - particle states are the one - body eigenstates of the hamiltonian @xmath20 . in cases
when the confinement potential @xmath4 is parabolic , the single - particle eigenstates are the fock - darwin states.@xcite for systems with perturbed parabolic or parabolic ring potentials , we calculate the single - electron states as linear combinations of fock - darwin states , all with the same angular momentum , corresponding to the 15 lowest energy values .
the ed calculations take into account landau - level mixing , as the 25 lowest - lying eigenstates , irrespective of landau level , are included in the computational procedure .
the ground state eigenvalues of the hamiltonian matrix of the ed method were obtained using the lanczos diagonalization method , and the interaction matrix elements were calculated using numerical integration .
the principles of the ed method are described in ref . .
when the electrons in a quantum ring become sufficiently localized , charge and spin excitations separate from each other.@xcite this phenomenon is a non - perturbative effect due to the strong correlations in quantum ring systems at high magnetic fields .
the qualitative behavior of the many - particle spectrum of a quasi - one - dimensional system has been described by the lattice hamiltonian@xcite @xmath21 where the first term is a heisenberg hamiltonian that models the spin degrees of freedom , the second term is a rigid rotation of the system , and the last term includes vibrational modes of the localized electrons .
the parameters of the model are the nearest - neighbor coupling constant between the spins @xmath22 , the total moment of inertia @xmath23 , and the vibration frequency @xmath24 .
the angular momentum of the ring is given by @xmath25 and @xmath26 is the number of excitation quanta of the vibrational mode . in an antiferromagnetic system
the coupling is such that @xmath27 0 , and in a ferromagnetic system @xmath28 0 . the total angular momentum
spin pairs of the lowest eigenstates for the effective hamiltonian ( [ eq : effham ] ) can be calculated using exact diagonalization , or using group - theoretical methods.@xcite in table [ tab : af ] , we have given the angular momentum and spin states of the ferromagnetic and antiferromagnetic systems of four particles , as calculated by the exact diagonalization .
the spin sequence is @xmath1-periodic as a function of the angular momentum , where @xmath1 is the number of electrons .
one can see that for four particles , the antiferromagnetic behavior is manifested by the three consecutive ground states with spin equal to one , and one signature of the ferromagnetic model is the occurrence of ground state with spin equal to two .
one should also note that there is one ground state with total spin equal to zero for both versions of the heisenberg model .
[ tab : af ] .total angular momenta @xmath25 and their corresponding spin values @xmath29 for the ferromagnetic and antiferromagnetic ground states of the heisenberg model of a four - electron ring .
spin states corresponding to higher angular momenta are obtained by utilizing the periodicity of the spin sequence . [ cols=">,^,^,^,^,^",options="header " , ]
we start with the results for the parabolic dot that has been perturbed with the gaussian potential in the center to form a quantum ring . for a four - electron parabolic quantum dot with gaussian perturbations of different strengths @xmath14 .
the spin states are 0 ( red / middle grey ) , 1 ( green / light grey ) , and 2 ( dark grey ) . the upper integers denote the number of central vortices found in each of the different states , and the integers surrounded by a ring denote the corresponding angular momenta .
all states at zero magnetic field have angular momentum 0 .
if two subsequent states are separated by a solid line , the angular momentum increases by one when moving to the state at right , and if they are separated by a dashed line , the angular momentum increases by two . ] in fig .
[ fig : phased ] , we have plotted spin ground states as a function of the magnetic field @xmath30 for four - electron perturbed parabolic dots with perturbation strengths @xmath31 0 , 1 , 2 , and 3 ha@xmath32 ( 1 ha@xmath33 11.3036 mev ) .
for all these perturbation strengths , the radius of the quantum ring is around 2 @xmath17 . for @xmath34 ,
corresponding to a pure parabolic quantum dot , the states with @xmath35 and 18 are the only fully spin - polarized states ( dark gray regions in phase diagram ) , as also found in ref . in the lowest landau level approximation . the state with @xmath35
is called the maximum - density droplet ( mdd),@xcite corresponding to the fractional quantum hall states with filling fraction @xmath36 , and the state with @xmath37 corresponds to @xmath38.@xcite between these two states the total spin has values of zero and one , but these do not follow the predictions from the antiferromagnetic heisenberg model , most clearly seen from the fact that there are no regions with three consecutive ground states with spin equal to one .
this is not surprising , as we expect the heisenberg model to be relevant only in the limit where electrons are strongly in a ring - shaped confinement .
next we turn on the gaussian perturbation , and one can see that the spin - polarized state with @xmath35 at @xmath34 splits into several spin - polarized states with different angular momentum .
in addition , between these @xmath39 states , the total spins follow the ferromagnetic heisenberg spins given in table [ tab : af ] .
the only exception on this rule is found at @xmath40 around @xmath41 t , where @xmath42 is found instead 0 . outside these ferromagnetic regions at stronger magnetic fields , one can see antiferromagnetic behavior .
the antiferromagnetic phase can be identified by the three @xmath42 states in a row , separated by the @xmath43 state .
in addition , antiferromagnetic behavior could develop at weak magnetic fields , as at @xmath44 there are already three consecutive states with @xmath42 .
as will be shown later in this paper , when the ring is sufficiently narrow there is indeed found an antiferromagnetic region at magnetic fields lower than the first ferromagnetic region . in a previous study@xcite of a hard - wall quantum dot with a radius @xmath45 5 @xmath17 and maximum of the electron density around 3 @xmath17 ,
the spin structure was ferromagnetic for angular momenta @xmath25 between 6 and 18 .
such a magnetic structure is almost the same as that found for the perturbed parabolic system with perturbations @xmath31 1 and 2 .
the magnetic structure of the wide quantum ring is thus for angular momenta 6 @xmath46 18 similar to that of a hard - wall quantum dot . from the phase diagram shown in fig . [
fig : phased ] , one can also see that the state with angular momentum @xmath35 changes spin when the gaussian impurity is made stronger . this change and many other details of the phase diagram
can be understood by studying the vortex structure of the many - body wave function . when subjected to an external magnetic field ,
the electrons in a quantum dot are forced to rotate .
if the rotation is sufficiently strong , vortices are formed in the electron liquid.@xcite one can consider the vortices as quasiparticles that are holes in the occupied fermi sea.@xcite the vortices are seen as zero points in the conditional electron density , where the phase of the wave function changes as a multiple of 2@xmath47 for a path enclosing the node .
as shown in the fractional quantum hall effect ( fqhe ) theory,@xcite the formation of vortices is a result of quantization of the magnetic field , and each vortex is associated with an integer number of dirac flux quanta . to analyze the nodal structure of the perturbed quantum dot , we have calculated conditional wave functions,@xcite defined for an @xmath1-particle system as @xmath48 where @xmath49 is the position of the moved particle 1 , and @xmath50 is the most probable position of particle @xmath51 that are found by maximizing the total electron density @xmath52 .
the phase @xmath53 is obtained from the relation @xmath54 and the conditional electron density is defined as @xmath55 . now returning to the question related to the spin of the @xmath35 state , a starting point for this analysis
is the conditional wave function of the parabolic dot shown in fig .
[ fig : condwb](j ) . in this
, one can see a pauli vortex on top of each electron , except the probe particle on right .
the vortices are shown by the discontinuous jumps in the gray - scale from white to black when the electrons are circulated in a clockwise fashion .
the fact that the vortex number is the same as the electron number shows that the state is a finite - size example of a quantum hall state with filling fraction @xmath36 . for @xmath56 , shown in fig .
[ fig : condwb](g ) , the spin is still the same and the conditional wave function is nearly identical to the @xmath34 state .
however , for @xmath40 the spin has changed from 2 to 0 , and the conditional wave function in fig .
[ fig : condwb](d ) looks completely different .
there is still one pauli vortex on top of the left - most electron , but the two other electrons have switched spins .
in addition , there is now one vortex that is located at the center of the system .
it turns out that one needs to combine two pauli vortices in order to make one vortex at the center of the dot . from this data
, one can understand the change in the total spin as follows : the system places the vortices so that it minimizes the total energy . without the gaussian impurity at the center of the dot ,
it is energetically favorable to place the vortices on top of the electrons to reduce the coulomb repulsion . when the potential at the center of the dot is raised by the gaussian impurity , at some strength it is more favorable to reduce the probability of the electrons to be at the center of the dot by placing a vortex there .
one can do a similar analysis for the @xmath37 state , which for @xmath34 corresponds to the @xmath38 state .
the correspondence can be seen by the three vortices bound to each electron , as shown in fig .
[ fig : condwb](k ) .
the reason for the small separation of the vortices is due to long - range nature of the interaction . for @xmath56
shown in fig .
[ fig : condwb](h ) , there are already two vortices at the center of the dot , and to create these , four pauli vortices have vanished from the system . due to this , the spin has again changed , and the opposite spins have only two vortices bound to them . as a general rule ,
opposite spin electrons can have even vortices bound to them , and the antisymmetry requirement of same spin electrons forces the vortex number to be odd .
the vortex structure of the @xmath43 states is similar to that of a @xmath57 halperin state,@xcite apart from the vortices at the center of the dot .
going to @xmath40 shown in fig . [
fig : condwb](e ) , a third vortex has appeared at the center , and now again the total spin of the system has changed . finally , at @xmath44 shown in fig .
[ fig : condwb](b ) the conditional wave function is nearly identical to the @xmath40 case
. the same trend can be found at even larger values of the angular momentum , and as a final example , we show the data for the @xmath58 states in figs .
[ fig : condwb](c),(f ) , and ( i ) .
the main difference to the previous examples is that when the number of central vortices grows , also the area they occupy seems to be larger .
analyzing the vortex structure of the other states in detail enables us to label the central vortex numbers of the ground states .
these numbers are given in fig .
[ fig : phased ] for the most interesting states .
one can see that in general , the number of central vortices grows as a function of the magnetic field .
however , at each fixed gaussian impurity strength , there is a point where it is energetically favorable to add pauli vortices instead of the central vortices , and at this point the number of central vortices is constant although the angular momentum increases . at these points ,
the ferromagnetic behavior is changed to antiferromagnetic .
somewhat similar transitions are seen at the magnetic fields below the first ferromagnetic states .
it is also interesting to analyze the ground states at magnetic fields around 67 t. at @xmath34 , the ground state is the @xmath39 state with @xmath35 corresponding to @xmath36 . when the gaussian impurity is made stronger ,
there are more and more vortices at the center of the system , and the angular momentum is increased .
however , the spin of the ground state does not change .
this shows that the @xmath36 state is in some sense stable even for this small particle number when one pierces it with three fluxes at the center . by this
we mean that the vortex structure of the conditional wave function is the same apart from the central vortices , see figs .
[ fig : phased](b ) and ( j ) .
based on the analysis presented above , one would expect to find an antiferromagnetic region at low magnetic fields , if the ring is made sufficiently narrow ( see the region with @xmath59 24 t and @xmath31 3 ha@xmath32 of fig .
1 ) . to investigate this , we switch the confinement potential to @xmath60 where the confinement strength is taken to be 40 mev and the radius 8@xmath61 .
this system is a very narrow ring with approximately the same radius as the ring studied above . in fig .
[ fig : ring ] , the total angular momentum and spin of the ground state is plotted for the narrow ring .
we have varied the coulomb constant between the values @xmath62 0.1 , 1 , and 10 . in reality , the interaction strength can be changed by changing the quantum ring radius.@xcite the case with @xmath63 , shown in fig .
[ fig : ring](b ) , corresponds to the natural interaction strength , and serves as basis for the comparison with the ring used above . for this case ,
the total spin has values zero and one , and above @xmath64 t there are three states with spin one between each ground state with spin equal to zero , in agreement with the antiferromagnetic heisenberg model .
thus , as expected there is an antiferromagnetic region at low magnetic fields when the ring is sufficiently narrow .
when the coulomb interaction is made weak by setting @xmath65 , two of the three consecutive @xmath42 states are no longer ground states as shown in fig .
[ fig : ring](a ) .
now , although the spin has values 0 and 1 , the system does not behave as the antiferromagnetic model .
this shows that the localization induced by the strong interaction is a necessity for antiferromagnetism .
on the other hand , when we make the interaction stronger by setting @xmath66 , one can see that fully spin polarized ground states are found .
in addition , the magnetic field range of each ground state is nearly the same . in the magnetic field region between 6 and 10 t ,
the ferromagnetism is not as pure as the antiferromagnetism for the @xmath63 case , as now the ground states with @xmath43 are missing .
also , the system is antiferromagnetic upto @xmath67 t. we conclude that a transition between antiferromagnetic and ferromagnetic behaviour also can be controlled by the coulomb interaction strength , or physically by tuning the radius of the quantum ring .
our results indicate that both ferro- and antiferromagnetic phases are found for quantum rings in a strong magnetic field . in addition , the occurrence of the different magnetic phases are not trivial to predict , as the magnetic phase is determined by the magnetic field region , the ring width and the radius of the system .
we found that the magnetic phase changes oscillatory as a function of the different system parameters .
knowledge of the nontrivial magnetic phase structure could be very fruitful for experiments , as the observed phenomena opens the possibility for tunability of the magnetism by changing system parameters .
for example , in an experiment , a metallic electrode with adjustable voltage could be used to achieve the control we had in our calculations when using the gaussian impurity .
+ this study has been supported by the academy of finland through its centers of excellence program ( 20062011 ) .
et acknowledges financial support from the vilho , yrj , and kalle visl foundation of the finnish academy of science and letters .
we also thank henri saarikoski for useful discussions . | we have studied the spin structure of circular four - electron quantum rings using tunable confinement potentials .
the calculations were done using the exact diagonalization method .
our results indicate that ringlike systems can have oscillatory flips between ferromagnetic and antiferromagnetic behaviour as a function of the magnetic field .
furthermore , at constant external magnetic fields there were seen similar oscillatory changes between ferromagnetism and antiferromagnetism when the system parameters were changed . according to our results ,
the magnetism of quantum rings could be tuned by system parameters . |
the deployment of large groups of autonomous vehicles is rapidly becoming possible because of technological advances in networking and in miniaturization of electro - mechanical systems . in the near future large numbers of robots
will coordinate their actions through ad - hoc communication networks and will perform challenging tasks including search and recovery operations , manipulation in hazardous environments , exploration , surveillance , and environmental monitoring for pollution detection and estimation .
the potential advantages of employing teams of agents are numerous .
for instance , certain tasks are difficult , if not impossible , when performed by a single vehicle agent .
further , a group of vehicles inherently provides robustness to failures of single agents or communication links .
working prototypes of active sensing networks have already been developed ; see @xcite . in @xcite , launchable miniature mobile robots communicate through a wireless network . the vehicles are equipped with sensors for vibrations , acoustic , magnetic , and ir signals as well as an active video module ( i.e. , the camera or micro - radar is controlled via a pan - tilt unit ) .
a second system is suggested in @xcite under the name of autonomous oceanographic sampling network ; see also @xcite . in this case ,
underwater vehicles are envisioned measuring temperature , currents , and other distributed oceanographic signals .
the vehicles communicate via an acoustic local area network and coordinate their motion in response to local sensing information and to evolving global data . this mobile sensing network is meant to provide the ability to sample the environment adaptively in space and time . by identifying evolving temperature and current gradients with higher accuracy and resolution than current static sensors ,
this technology could lead to the development and validation of improved oceanographic models .
a fundamental prototype problem in this paper is that of characterizing and optimizing notions of quality - of - service provided by an adaptive sensor network in a dynamic environment . to this goal ,
we introduce a notion of _ sensor coverage _ that formalizes an optimal sensor placement problem .
this spatial resource allocation problem is the subject of a discipline called locational optimization @xcite .
locational optimization problems pervade a broad spectrum of scientific disciplines .
biologists rely on locational optimization tools to study how animals share territory and to characterize the behavior of animal groups obeying the following interaction rule : each animal establishes a region of dominance and moves toward its center .
locational optimization problems are spatial resource allocation problems ( where to place mailboxes in a city or cache servers on the internet ) and play a central role in quantization and information theory ( the design of a minimum - distortion fixed - rate vector quantizer is a locational problem ) .
other technologies affected by locational optimization include mesh and grid optimization methods , clustering analysis , data compression , and statistical pattern recognition .
because locational optimization problems are so widely studied , it is not surprising that methods are indeed available to tackle coverage problems ; see @xcite . however , most currently - available algorithms are not applicable to mobile sensing networks because they inherently assume a centralized computation for a limited size problem in a known static environment .
this is not the case in multi - vehicle networks which , instead , rely on a distributed communication and computation architecture .
although an ad - hoc wireless network provides the ability to share some information , no global omniscient leader might be present to coordinate the group .
the inherent spatially - distributed nature and limited communication capabilities of a mobile network invalidate classic approaches to algorithm design .
in this paper we design coordination algorithms implementable by a multi - vehicle network with limited sensing and communication capabilities .
our approach is related to the classic lloyd algorithm from quantization theory ; see @xcite for a reprint of the original report and @xcite for a historical overview .
we present lloyd descent algorithms that take into careful consideration all constraints on the mobile sensing network . in particular
, we design coverage algorithms that are adaptive , distributed , asynchronous , and verifiably asymptotically correct : adaptive : : : our coverage algorithms provide the network with the ability to address changing environments , sensing task , and network topology ( due to agents departures , arrivals , or failures ) . distributed : : : our coverage algorithms are distributed in the sense that the behavior of each vehicle depends only on the location of its neighbors .
also , our algorithms do not required a fixed - topology communication graph , i.e. , the neighborhood relationships do change as the network evolves .
the advantages of distributed algorithms are scalability and robustness .
asynchronous : : : our coverage algorithms are amenable to asynchronous implementation .
this means that the algorithms can be implemented in a network composed of agents evolving at different speeds , with different computation and communication capabilities .
furthermore , our algorithms do not require a global synchronization and convergence properties are preserved even if information about neighboring vehicles propagates with some delay .
an advantage of asynchronism is a minimized communication overhead .
verifiable asymptotically correct : : : our algorithms guarantees monotonic descent of the cost function encoding the sensing task .
asymptotically the evolution of the mobile sensing network is guaranteed to converge to so - called centroidal voronoi configurations that are critical points of the optimal sensor coverage problem .
let us describe in some detail what are the contributions of this paper .
section [ sec : review ] reviews certain locational optimization problems and their solutions as centroidal voronoi partitions .
section [ sec : coverage - control ] provides a continuous - time version of the classic lloyd algorithm from vector quantization and applies it to the setting of multi - vehicle networks . in discrete - time
, we propose a family of lloyd algorithms .
we carefully characterize convergence properties for both continuous and discrete - time versions ( appendix [ sec : appendix ] collects some relevant facts on descent flows ) .
we discuss a worst - case optimization problem , we investigate a simple uniform planar setting , and we present numerical results .
section [ sec : distributed - asynchronous ] presents two asynchronous distributed implementations of lloyd algorithm for ad - hoc networks with communication and sensing capabilities .
our treatment carefully accounts for the constraints imposed by the distributed nature of the vehicle network .
we present two asynchronous implementations , one based on classic results on distributed gradient flows , the other based on the structure of the coverage problem .
section [ sec : vehicle - dynamics ] considers vehicle models with more realistic dynamics .
we present two formal results on passive vehicle dynamics and on vehicles equipped with individual local controllers .
we present numerical simulations of passive vehicle models and of unicycle mobile vehicles .
next , section [ sec : geometric - patterns ] describes density functions that lead the multi - vehicle network to predetermined geometric patterns .
recent years have witnessed a large research effort focused on motion planning and coordination problems for multi - vehicle systems .
issues include geometric patterns @xcite , formation control @xcite , and conflict avoidance @xcite .
algorithms for robotic sensing tasks are presented for example in @xcite .
it is only recently , however , that truly distributed coordination laws for dynamic networks are being proposed ; e.g. , see @xcite and the conference versions of this work @xcite .
heuristic approaches to the design of interaction rules and emerging behaviors have been throughly investigated within the literature on behavior - based robotics ; see @xcite .
an example of coverage control is discussed in @xcite . along this line of research ,
algorithms have been designed for sophisticated cooperative tasks .
however , no formal results are currently available on how to design reactive control laws , ensure their correctness , and guarantee their optimality with respect to an aggregate objective .
the study of distributed algorithms is concerned with providing mathematical models , devising precise specifications for their behavior , and formally proving their correctness and complexity . via an automata - theoretic approach , the references @xcite treat distributed consensus , resource allocation , communication , and data consistency problems . from a numerical optimization viewpoint , the works in @xcite
discuss distributed asynchronous algorithms as networking algorithms , rate and flow control , and gradient descent flows .
typically , both these sets of references consider networks with fixed topology , and do not address algorithms over ad - hoc dynamically changing networks .
another common assumption is that , any time an agent communicates its location , it broadcasts it to every other agent in the network . in our setting
, this would require a non - distributed communication set - up .
in this section we describe a collection of known facts about a meaningful optimization problem .
references include the theory and applications of centroidal voronoi partitions , see @xcite , and the discipline of facility location , see @xcite . along the paper
, we interchangeably refer to the elements of the network as sensors , agents , vehicles , or robots .
let @xmath0 be a convex polytope in @xmath1 and let @xmath2 denote the euclidean distance function .
we call a map @xmath3 a _ distribution density function _ if it represents a measure of information or probability that some event take place over @xmath0 . in equivalent words
, we can consider @xmath0 to be the bounded support of the function @xmath4 .
let @xmath5 be the _ location of @xmath6 sensors _ , each moving in the space @xmath0 .
because of noise and loss of resolution , the _ sensing performance _ at point @xmath7 taken from @xmath8th sensor at the position @xmath9 degrades with the distance @xmath10 between @xmath7 and @xmath9 ; we describe this degradation with a non - decreasing differentiable function @xmath11 .
accordingly , @xmath12 provides a quantitative assessment of how poor the sensing performance is . as an example , consider @xmath6 mobile robots equipped with microphones attempting to detect , identify , and localize a sound - source .
_ how should we plan to robots motion in order to maximize the detection probability ? _ assuming the source emits a known signal , the optimal detection algorithm is a matched filter ( i.e. , convolve the known waveform with the received signal and threshold ) .
the source is detected depending on the signal - to - noise - ratio , which is inversely proportional to the distance between the microphone and the source .
various electromagnetic and sound sensors have signal - to - noise ratios inversely proportional to distance . within the context of this paper ,
a _ partition _ of @xmath0 is a collection of @xmath6 polytopes @xmath13 with disjoint interiors whose union is @xmath0 .
we say that two partitions @xmath14 and @xmath15 are equal if @xmath16 and @xmath17 only differ by a set of @xmath4-measure zero , for all @xmath18 .
we consider the task of minimizing the locational optimization function @xmath19 where we assume that the @xmath8th sensor is responsible for measurements over its `` dominance region '' @xmath16 .
note that the function @xmath20 is to be minimized with respect to both ( 1 ) the sensors location @xmath21 , and ( 2 ) the assignment of the dominance regions @xmath14 .
this problem is referred to as a facility location problem and in particular as a continuous @xmath22-median problem in @xcite .
note that if we interchange the positions of any two agents , along with their associated regions of dominance , the value of the locational optimization function @xmath20 is not affected . to eliminate this discrete redundancy
, one could take the discrete group of permutations @xmath23 with the natural action on @xmath24 , and consider @xmath25 as the configuration space for the position @xmath21 of the @xmath6 vehicles .
one can easily see that , at fixed sensors location , the optimal partition of @xmath0 is the _ voronoi partition _
@xmath26 generated by the points @xmath27 : @xmath28 we refer to @xcite for a comprehensive treatment on voronoi diagrams , and briefly present some relevant concepts .
the set of regions @xmath29 is called the voronoi diagram for the generators @xmath30 .
when the two voronoi regions @xmath31 and @xmath32 are adjacent , @xmath9 is called a _ ( voronoi ) neighbor _ of @xmath33 ( and vice - versa ) .
the set of indexes of the voronoi neighbors of @xmath9 is denoted by @xmath34 . clearly , @xmath35 if and only if @xmath36 .
we also define the @xmath37-face as @xmath38 .
voronoi diagrams can be defined with respect to various distance functions , e.g. , the @xmath39- , @xmath40- , @xmath41- , and @xmath42-norm over @xmath43 , and voronoi diagrams can be defined over riemannian manifolds ; see @xcite .
some useful facts about the euclidean setting are the following : if @xmath0 is a convex polytope in a @xmath44-dimensional euclidean space , the boundary of each @xmath31 is the union of @xmath45-dimensional convex polytopes . in
what follows , we shall write @xmath46 note that @xmath47 , \nonumber\end{aligned}\ ] ] that is , the locational optimization function can be interpreted as an expected value composed with a min operation . this is the usual way in which the problem is presented in the facility location and operations research literature @xcite .
remarkably , one can show @xcite that @xmath48 and deduce some smoothness properties of @xmath49 .
since the voronoi partition @xmath50 depends at least continuously on @xmath51 , the function @xmath49 is at least continuously differentiable .
let us recall some basic quantities associated to a region @xmath52 and a mass density function @xmath53 .
the ( generalized ) mass , centroid ( or center of mass ) , and polar moment of inertia are defined as @xmath54 additionally , by the parallel axis theorem , one can write , @xmath55 where @xmath56 is defined as the polar moment of inertia of the region @xmath57 about its centroid .
let us consider again the locational optimization problem , and suppose now we are strictly interested in the setting @xmath58 that is , we assume @xmath59 . applying the parallel axis theorem leads to simplifications for both the function @xmath49 and its partial derivative : @xmath60 it is convenient to define @xmath61 and @xmath62 . therefore , the ( not necessarily unique ) local minimum points for the location optimization function @xmath49 are _ centroids _ of their voronoi cells , i.e. , each location @xmath9 satisfies two properties simultaneously : it is the generator for the voronoi cell @xmath31 and it is its centroid @xmath63 accordingly , the critical partitions and points for @xmath20 are called _ centroidal voronoi partitions_. we will refer to a sensors configuration as a _ centroidal voronoi configuration _ if it gives rise to a centroidal voronoi partition .
this discussion provides a proof alternative to the one given in @xcite for the necessity of centroidal voronoi partitions as solutions to the continuous @xmath22-median location problem .
in this section , we describe algorithms to compute the location of sensors that minimize the cost @xmath20 , both in continuous and in discrete - time . in section [ se : continuous - lloyd ] , we propose a continuous - time version of the classic lloyd algorithm . here , both the positions and partitions evolve in continuous time , whereas lloyd algorithm for vector quantization is designed in discrete time . in section [ se : discrete - lloyd ] , we develop a family of variations of lloyd algorithm in discrete time . in both setting , we prove that the proposed algorithms are _ gradient descent flows_. assume the sensors location obeys a first order dynamical behavior described by @xmath64 consider @xmath49 a cost function to be minimized and impose that the location @xmath9 follows a gradient descent . in equivalent control theoretical terms , consider @xmath49 a lyapunov function and stabilize the multi - vehicle system to one of its local minima via dissipative control .
formally , we set @xmath65 where @xmath66 is a positive gain , and where we assume that the partition @xmath26 is continuously updated . for the closed - loop system induced by equation , the sensors location converges asymptotically to the set of critical points of @xmath67 , i.e. , the set of centroidal voronoi configurations on @xmath0 . assuming this set is finite , the sensors location converges to a centroidal voronoi configuration . under the control law
, we have @xmath68 by lasalle s principle , the sensors location converges to the largest invariant set contained in @xmath69 , which is precisely the set of centroidal voronoi configurations .
since this set is clearly invariant for , we get the stated result .
if @xmath69 consists of a finite collection of points , then @xmath70 converges to one of them , see corollary [ corollary : lasalle - finite ] . if @xmath69 is finite , and @xmath71 , then a sufficient condition that guarantees exponential convergence is that the hessian of @xmath67 be positive definite at @xmath72 .
this property is known to be an open problem , see @xcite .
note that this gradient descent is not guaranteed to find the global minimum .
for example , in the vector quantization and signal processing literature @xcite , it is known that for bimodal distribution density functions , the solution to the gradient flow reaches local minima where the number of generators allocated to the two region of maxima are not optimally partitioned .
let us consider the following class of variations of lloyd algorithm .
let @xmath73 be a continuous mapping @xmath74 verifying the following two properties : \(a ) for all @xmath75 , @xmath76 , where @xmath77 denotes the @xmath8th component of @xmath73 , \(b ) if @xmath21 is not centroidal , then there exists a @xmath78 such that @xmath79 .
property ( a ) guarantees that , if moving , the agents of the network do not increase their distance to its corresponding centroid .
property ( b ) ensures that at least one robot moves at each iteration and strictly approaches the centroid of its voronoi region .
because of this property , the fixed points of @xmath73 are the set of centroidal voronoi configurations . [ prop : discrete - lloyd ] let @xmath80 @xmath81 denote the initial sensors location .
then , the sequence @xmath82 converges to the set of centroidal voronoi configurations . if this set if finite , then @xmath83 converges to a centroidal voronoi configuration .
consider @xmath84 as an objective function for the algorithm @xmath73 .
note that @xmath85 with strict inequality if @xmath86 .
moreover , the parallel axis theorem guarantees @xmath87 as long as @xmath88 for all @xmath89 , with strict inequality if for any @xmath8 , @xmath90 .
in particular , @xmath91 , with strict inequality if @xmath92 , where @xmath93 denotes the set of centroids of the partition @xmath14 .
now , we have @xmath94 because of .
in addition , because of property ( a ) of @xmath73 , inequality yields @xmath95 and the inequality is strict if @xmath21 is not centroidal by property ( b ) of @xmath73 .
hence , @xmath67 is a descent function for the algorithm @xmath73 .
the result now follows from the global convergence theorem [ lemma : discrete - lasalle ] and proposition [ prop : discrete - lasalle - surprising ] .
lloyd algorithm in quantization theory @xcite is usually presented as follows : given the location of @xmath6 agents , @xmath96 , ( i ) construct the voronoi partition corresponding to @xmath5 ; ( ii ) compute the mass centroids of the voronoi regions found in step ( i ) .
set the new location of the agents to these centroids ; and return to step ( i ) .
lloyd algorithm can also be seen as a fixed point iteration .
consider the mappings @xmath97 for @xmath98 @xmath99 let @xmath100 be defined by @xmath101 .
clearly , @xmath102 is continuous ( indeed , @xmath103 ) , and corresponds to lloyd algorithm .
now , @xmath104 , for all @xmath105 . moreover , if @xmath21 is not centroidal , then the inequality is strict for all @xmath106 .
therefore , @xmath102 verifies properties ( a ) and ( b ) .
different sensor performance functions @xmath107 in equation correspond to different optimization problems .
provided one uses the euclidean distance in the definition of @xmath108 , the standard voronoi partition computed with respect to the euclidean metric remains the optimal partition . for arbitrary @xmath107 ,
it is not possible anymore to decompose @xmath67 into the sum of terms similar to @xmath109 and @xmath110 .
nevertheless , it is still possible to implement the gradient flow via the expression for the partial derivative .
assume the sensors location obeys a first order dynamical behavior , @xmath111 .
then , for the closed - loop system induced by the gradient law , @xmath112 , the sensors location @xmath5 converges asymptotically to the set of critical points of @xmath67 .
assuming this set is finite , the sensors location converges to a critical point .
more generally , various distance notions can be used to define locational optimization functions .
different performance function gives rise to corresponding notions of `` center of a region '' ( any notion of geometric center , mean , or average is an interesting candidate ) .
these can then be adopted in designing coverage algorithms .
we refer to @xcite for a discussion on voronoi partitions based on non - euclidean distance functions and to @xcite for a discussion on the corresponding locational optimization problems .
next , let us discuss an interesting variation of the original problem . in @xcite , minimizing the expected minimum distance function @xmath49 in equation
is referred to as the _ continuous @xmath22-median problem_. it is instructive to consider the worst - case minimum distance function , corresponding to the scenario where no information is available on the distribution density function . in other words ,
the network seeks to minimize the largest possible distance from any point in @xmath0 to any of the sensor locations , i.e. , to minimize the function @xmath113 = \max_{i\in\{1,\ldots , n\ } } \left [ \max_{q\in v_i } \|q - p_i\| \right ] \ , .
\end{aligned}\ ] ] this optimization is referred to as the _ @xmath22-center problem _ in @xcite .
one can design a strategy for the @xmath22-center problem analog to the lloyd algorithm for the @xmath22-median problem : each vehicle moves , in continuous or discrete - time , toward the center of the minimum - radius sphere enclosing the polytope . to the best of our knowledge
, no convergence proof is available in the literature for this algorithm ; e.g. , see @xcite .
we refer to @xcite for a convergence analysis of the continuous and discrete time algorithms . in what follows , we shall restrict our attention to the @xmath22-median problem and to centroidal voronoi partitions . in this section
, we investigate closed - form expression for the control laws introduced above
. assume the voronoi region @xmath31 is a convex polygon ( i.e. , a polytope in @xmath114 ) with @xmath115 vertexes labeled @xmath116 such as in figure [ fig : polygon ] .
it is convenient to define @xmath117 .
furthermore , we assume that the density function is @xmath118 . by evaluating the corresponding integrals , one can obtain the following closed - form expressions @xmath119 to present a simple formula for the polar moment of inertia , let @xmath120 and @xmath121 , for @xmath122 .
then , the polar moment of inertia of a polygon about its centroid , @xmath123 becomes @xmath124 the proof of these formulas is based on decomposing the polygon into the union of disjoint triangles .
we refer to @xcite for analog expressions over @xmath1 .
a second observation is that the voronoi polygon s vertexes can be expressed as a function of the neighboring vehicles .
the vertexes of the @xmath8th voronoi polygon which lie in the interior of @xmath0 are the circumcenters of the triangles formed by @xmath9 and any two neighbors adjacent to @xmath9 .
the circumcenter of the triangle determined by @xmath125 , @xmath126 , and @xmath127 is @xmath128 where @xmath129 is the area of the triangle , and @xmath130 .
equation for a polygon s centroid and equation for the voronoi cell s vertexes lead to a closed - form _ algebraic _ expression for the control law in equation as a function of the neighboring vehicles location .
to illustrate the performance of the continuous - time lloyd algorithm , we include some simulation results .
the algorithm is implemented in ` mathematica ` as a single centralized program . for the @xmath114 setting
, the code computes the bounded voronoi diagram using the ` mathematica ` package ` computationalgeometry ` , and computes mass , centroid , and polar moment of inertia of polygons via the numerical integration routine ` nintegrate ` .
careful attention was paid to numerical accuracy issues in the computation of the voronoi diagram and in the integration .
we illustrate the performance of the closed - loop system in figure [ fig : coverage-1 ] .
in this section we show how the lloyd gradient algorithm can be implemented in an asynchronous distributed fashion . in section [ subsec : modeling ]
we describe our model for a distributed asynchronous network of robotic agents .
next , we provide two distributed algorithms for the local computation and maintenance of the voronoi cells . finally , in section [ subsec : asynchronous - lloyd ] we propose two distributed asynchronous implementations of lloyd algorithm : the first one is based on the gradient optimization algorithms as described in @xcite and the second one relies on the special structure of the coverage problem .
we start by modeling a robotic agent that performs sensing , communication , computation , and control actions .
we are interested in the behavior of the asynchronous network resulting from the interaction of finitely many robotic agents . a theoretical framework to formalize the following concepts
is that developed in the theory of distributed algorithms ; see @xcite .
let us here introduce the notion of _ robotic agent with computation , communication , and control capabilities _ as the @xmath8th element of a network .
the @xmath8th agent has a processor with the ability of allocating continuous and discrete states and performing operations on them .
each vehicle has access to its unique identifier @xmath8 .
the @xmath8th agent occupies a location @xmath131 and it is capable of moving in space , at any time @xmath132 for any period of time @xmath133 , according to a first order dynamics of the form : @xmath134.\ ] ] the processor has access to the agent s location @xmath9 and determines the control pair @xmath135 .
the processor of the @xmath8th agent has access to a local clock @xmath136 , and a _ scheduling sequence _ ,
i.e. , an increasing sequence of times @xmath137 such that @xmath138 and @xmath139 .
the processor of the @xmath8th agent is capable of transmitting information to any other agent within a closed disk of radius @xmath140 .
we assume the communication radius @xmath141 to be a quantity controllable by the @xmath8th processor and the corresponding communication bandwidth to be limited .
we shall alternatively consider networks of _ robotic agents with computation , sensing , and control capabilities_. in this case , the processor of the @xmath8th agent has the same computation and control capabilities as before .
furthermore , we assume the processor can detect any other agent within a closed disk of radius @xmath140 .
we assume the sensing radius @xmath141 to be a quantity controllable by the processor .
a key requirement of the lloyd algorithms presented in section [ sec : coverage - control ] is that each agent must be able to compute its own voronoi cell .
to do so , each agent needs to know the relative location ( distance and bearing ) of each voronoi neighbor .
the ability of locating neighbors plays a central role in numerous algorithms for localization , media access , routing , and power control in ad - hoc wireless communication networks ; e.g. , see @xcite and references therein
. therefore , any motion control scheme might be able to obtain this information from the underlying communication layer . in
what follows , we set out to provide a distributed asynchronous algorithm for the local computation and maintenance of voronoi cells .
the algorithm is related to the synchronous scheme in @xcite and is based on basic properties of voronoi diagrams .
we present the algorithm for a robotic agent with sensing capabilities ( as well as computation and control ) .
the processor of the @xmath8th agent allocates the information it has on the position of the other agents in the state variable @xmath142 .
the objective is to determine the smallest distance @xmath141 for vehicle @xmath8 which provides sufficient information to compute the voronoi cell @xmath31 .
we start by noting that @xmath31 is a subset of the convex set @xmath143 where @xmath144 and the half planes @xmath145 are @xmath146 provided @xmath141 is twice as large as the maximum distance between @xmath9 and the vertexes of @xmath147 , all voronoi neighbors of @xmath9 are within distance @xmath141 from @xmath9 and the equality @xmath148 holds . the minimum adequate sensing radius is therefore @xmath149 .
we are now ready to state the following algorithm .
a similar algorithm can be designed for a robotic agent with communication capabilities .
the specifications go as in the previous algorithm , except for the fact that steps ` 2 : ` and ` 7 : ` are substituted by send @xmath150 within radius @xmath141 receive @xmath151 from all agents within radius @xmath141 further , we have to require each agent to perform the following event - driven task : if the @xmath8th agent receives at any time @xmath152 a `` request to reply '' message from the @xmath78th agent located at position @xmath33 , it executes send @xmath153 within radius @xmath154 we call this algorithm adjust communication radius algorithm .
next , we present an algorithm whose objective is to maintain the information about the voronoi cell of the @xmath8th agent , and detect the presence of certain events .
we consider only robotic agents with sensing capabilities .
we call an agent active if it is moving and we assume that the @xmath8th agent can determine if any agent within radius @xmath141 is active or not .
two events are of interest : ( i ) a voronoi neighbor of the @xmath8th agent becomes active and ( ii ) a new active agent becomes a voronoi neighbor of the @xmath8th agent . in both cases , we require a trigger message `` request recomputation '' to an appropriate control algorithm that we shall present in the next section . before presenting the algorithm ,
let us introduce the map @xmath155 that assigns to the state vector @xmath156 a tuple @xmath157 according to @xmath158 the algorithm is designed to run for times @xmath159 $ ] .
let us now present two versions of lloyd algorithm for the solution of the optimization problem that can be implemented by an asynchronous distributed network of robotic agents . for simplicity ,
we assume that at time @xmath160 all clocks are synchronized ( although they later can run at different speeds ) and that each agent knows at @xmath160 the exact location of every other agent .
the first algorithm is designed for robotic agents with communication capabilities , and requires the adjust communication radius algorithm ( while it does not require the monitoring algorithm ) . as a consequence of theorem 3.1 and corollary 3.1 in @xcite , we have the following result .
let @xmath161 denote the initial sensors location .
let @xmath162 be the sequence in increased order of all the scheduling sequences of the agents of the network .
assume @xmath163 .
then , there exists a sufficiently small @xmath164 such that if @xmath165 , the coverage behavior algorithm i converges to the set of critical points of @xmath67 , that is , the set of centroidal voronoi configurations .
next , we focus on distributed asynchronous implementations of lloyd algorithm that take advantage of the special structure of the coverage problem .
the following algorithm is designed for robotic agents with sensing capabilities , it requires the monitoring and the adjust sensing radius algorithms .
two advantages of this algorithm over the previous one are that there is no need for each agent to exactly go toward the centroid of its voronoi cell nor to take a small step at each stage . the control law @xmath166 in step ` 7 : ` can be defined via a saturation function .
for instance , @xmath167 @xmath168 then set @xmath169 .
resorting to the discussion in section [ se : discrete - lloyd ] on the convergence of the discrete lloyd algorithms , one can prove that the coverage behavior algorithm ii verifies properties ( a ) and ( b ) . as a consequence of proposition [ prop : discrete - lloyd ]
, we then have the following result .
let @xmath161 denote the initial sensors location .
the coverage behavior algorithm ii converges to the set of critical points of @xmath67 , that is , the set of centroidal voronoi configurations .
in this section we investigate various extensions and applications of the algorithms proposed in the previous sections .
we extend the treatment to vehicles with passive dynamics and we also consider discrete - time implementations of the algorithms for vehicles endowed with a local motion planner .
finally , we describe interesting ways of designing density functions to solve problems apparently unrelated to coverage . here ,
we consider vehicles systems described by more general linear and nonlinear dynamical models . _ coordination of vehicles with passive dynamics_. we start by considering the extension of the control design to nonlinear control systems whose dynamics is passive .
relevant examples include networks of vehicles and robots with general lagrangian dynamics , as well as spatially invariant passive linear systems .
specifically , assume that for each @xmath170 , the @xmath8th vehicle state includes the spatial variable @xmath9 , and that the vehicle s dynamics is passive with input @xmath166 , output @xmath171 and storage function @xmath172 .
furthermore , assume that the input preserving the zero dynamics manifold @xmath173 is @xmath174 . for such systems ,
we devise a proportional derivative ( pd ) control via , @xmath175 where @xmath176 and @xmath177 are scalar positive gains .
the closed - loop system induced by this control law can be analyzed with the lyapunov function @xmath178 yielding the following result .
[ le : second - order ] for passive systems , the control law achieves asymptotic convergence of the sensors location to the set of centroidal voronoi configurations . if this set is finite , then the sensors location converges to a centroidal voronoi configuration .
consider the evolution of the function @xmath179 , @xmath180 by lasalle s principle , the sensors location converges to the largest invariant set contained in @xmath181 .
given the assumption on the zero dynamics , we conclude that @xmath182 for @xmath183 , i.e. , the largest invariant set corresponds to the set of centroidal voronoi configurations .
if this set is finite , lasalle s principle also guarantees convergence to a specific centroidal voronoi configuration . in figure
[ fig : second - order ] we illustrate the performance of the control law for vehicles with second - order dynamics @xmath184 .
vehicles with second order dynamics .
the environment and gaussian density function are as in figure [ fig : coverage-1 ] .
the control gains are @xmath185 and @xmath186.,title="fig : " ] vehicles with second order dynamics .
the environment and gaussian density function are as in figure [ fig : coverage-1 ] .
the control gains are @xmath185 and @xmath186.,title="fig : " ] vehicles with second order dynamics .
the environment and gaussian density function are as in figure [ fig : coverage-1 ] .
the control gains are @xmath185 and @xmath186.,title="fig : " ] _ coordination of vehicles with local controllers_. next , consider the setting where each vehicle has an arbitrary dynamics and is endowed with a local feedback and feedforward controller .
the controller is capable of strictly decreasing the distance to any specified position in @xmath0 in a specified period of time @xmath187 .
assume the dynamics of the @xmath8th vehicle is described by @xmath188 , where @xmath189 denotes its state , and @xmath190 is such that @xmath191 .
assume also that for any @xmath192 and any @xmath193 , there exists @xmath194 such that the solution @xmath195 of @xmath196 verifies @xmath197 .
[ prop : local - feedback+feedforward ] consider the following coordination algorithm . at time @xmath198 ,
@xmath199 , each vehicle computes @xmath200 and @xmath201 ; then , for time @xmath202 , the vehicle executes @xmath203 . for this closed - loop system
, the sensors location converges to the set of centroidal voronoi configurations .
if this set is finite , then the sensors location converges to a centroidal voronoi configuration .
the proof of this result readily follows from proposition [ prop : discrete - lloyd ] , since the algorithm verifies properties ( a ) and ( b ) of section [ se : discrete - lloyd ] . as an example
, we consider a classic model of mobile wheeled dynamics , the _ unicycle model_. assume the @xmath8th vehicle has configuration @xmath204 evolving according to @xmath205 where @xmath206 are the control inputs for vehicle @xmath8 .
note that the definition of @xmath207 is unique up to the discrete action @xmath208 . given a target point @xmath209
, we use this symmetry to require the equality @xmath210 for all time @xmath211 .
should the equality be violated at some time @xmath212 , we shall redefine @xmath213 and @xmath214 as @xmath215 from time @xmath212 onwards . following the approach in @xcite ,
consider the control law @xmath216 where @xmath176 is a positive gain .
this feedback law differs from the original stabilizing strategy in @xcite only in the fact that no final angular position is preferred .
one can prove that @xmath217 is guaranteed to monotonically approach the target position @xmath209 when run over an infinite time horizon .
we illustrate the performance of the proposed algorithm in figure [ fig : coverage - mobile ] . here
we suggest the use of decentralized coverage algorithms as formation control algorithms , and we present various density functions that lead the multi - vehicle network to predetermined geometric patterns
. in particular , we present simple density functions that lead to segments , ellipses , polygons , or uniform distributions inside convex environments .
consider a planar environment , let @xmath66 be a large positive gain , and denote @xmath218 .
let @xmath219 be real numbers , consider the line @xmath220 , and define the density function @xmath221 similarly , let @xmath222 be a reference point in @xmath114 , let @xmath223 be positive scalars , consider the ellipse @xmath224 , and define the density function @xmath225 we illustrate this density function in figure [ fig : coverage - circle ] . during the simulations
, we observed that the convergence to the desired pattern was rather slow . [ cols="^ " , ] finally , define the smooth ramp function @xmath226 , and the density function @xmath227 this density function leads the multi - vehicle network to obtain a uniform distribution inside the ellipsoidal disk @xmath228 .
we illustrate this density function in figure [ fig : coverage - disk ] .
vehicles to an ellipsoidal disk .
the density function parameters are the same as in figure [ fig : coverage - circle ] , and @xmath229 , @xmath230.,title="fig : " ] vehicles to an ellipsoidal disk .
the density function parameters are the same as in figure [ fig : coverage - circle ] , and @xmath229 , @xmath230.,title="fig : " ] vehicles to an ellipsoidal disk .
the density function parameters are the same as in figure [ fig : coverage - circle ] , and @xmath229 , @xmath230.,title="fig : " ] it appears straightforward to generalize these types of density functions to the setting of arbitrary curves or shapes .
the proposed algorithms are to be contrasted with the classic approach to formation control based on rigidly encoding the desired geometric pattern .
one disadvantage of the proposed approach is the requirement for a careful numerical computation of voronoi diagrams and centroids .
we refer to @xcite for previous work on algorithms for geometric patterns , and to @xcite for formation control algorithms .
we have presented a novel approach to coordination algorithms for multi - vehicle networks . the scheme can be thought of as an interaction law between agents and as such it is implementable in a distributed asynchronous fashion .
numerous extensions appear worth pursuing .
we plan to investigate the setting of non - convex environments and non - isotropic sensors .
we are currently implementing these algorithms on a network of all - terrain vehicles .
furthermore , we plan to extend the algorithms to provide collision avoidance guarantees and to vehicle dynamics which are not locally controllable .
this work was supported by nsf grant cms-0100162 , aro grant daad 190110716 , and darpa / afosr muri award f49620 - 02 - 1 - 0325 .
10 c. r. weisbin , j. blitch , d. lavery , e. krotkov , c. shoemaker , l. matthies , and g. rodriguez , `` miniature robots for space and military missions , '' , vol . 6 , no .
3 , pp . 918 , 1999 .
e. krotkov and j. blitch , `` the defense advanced research projects agency ( darpa ) tactical mobile robotics program , '' , vol .
18 , no . 7 , pp . 76976 , 1999 .
e. rybski , n. p. papanikolopoulos , s. a. stoeter , d. g. krantz , k. b. yesin , m. gini , r. voyles , d. f. hougen , b. nelson , and m. d. erickson , `` enlisting rangers and scouts for reconnaissance and surveillance , '' , vol . 7 , no .
1424 , 2000 . t. b. curtin , j. g. bellingham , j. catipovic , and d. webb , `` autonomous oceanographic sampling networks , '' , vol .
3 , pp . 8694 , 1993 .
r. m. turner and e. h. turner , `` organization and reorganization of autonomous oceanographic sampling networks , '' in _ ieee int .
conf . on robotics and automation _ , leuven , belgium , may 1998 , pp .
20607 .
e. eberbach and s. phoha , `` samon : communication , cooperation and learning of mobile autonomous robotic agents , '' in _
proceedings 11th international conf . on tools with artificial intelligence ( tai )
_ , chicago , il , nov . 1999 , pp
. 22936 .
a. okabe , b. boots , and k. sugihara , `` nearest neighbourhood operations with generalized voronoi diagrams : a review , '' , vol .
1 , pp . 4371 , 1994 .
z. drezner , ed .
, , springer series in operations research .
springer verlag , new york , ny , 1995 .
a. suzuki and a. okabe , `` using voronoi diagrams , '' in drezner @xcite , pp .
103118 . a.
okabe and a. suzuki , `` locational optimization problems solved through voronoi diagrams , '' , vol .
3 , pp . 44556 , 1997 .
a. okabe , b. boots , k. sugihara , and s. n. chiu , , wiley series in probability and statistics .
john wiley & sons , new york , ny , second edition , 2000 .
q. du , v. faber , and m. gunzburger , `` centroidal voronoi tessellations : applications and algorithms , '' , vol .
4 , pp . 637676 , 1999 .
s. p. lloyd , `` least squares quantization in pcm , '' , vol .
2 , pp . 129137 , 1982 , presented as bell laboratory technical memorandum at a 1957 institute for mathematical statistics meeting .
r. m. gray and d. l. neuhoff , `` quantization , '' , vol .
6 , pp . 23252383 , 1998 , commemorative issue 1948 - 1998 . h. yamaguchi and t. arai , `` distributed and autonomous control method for generating shape of multiple mobile robot group , '' in _
ieee / rsj int . conf . on intelligent robots & systems _ ,
munich , germany , sept .
1994 , pp .
800807 .
k. sugihara and i. suzuki , `` distributed algorithms for formation of geometric patterns with many mobile robots , '' , vol .
12739 , 1996 .
t. balch and r. arkin , `` behavior - based formation control for multirobot systems , '' , vol .
92639 , 1998 .
m. egerstedt and x. hu , `` formation constrained multi - agent control , '' , vol .
17 , no . 6 , pp . 94751 , 2001 .
j. p. desai , j. p. ostrowski , and v. kumar , `` modeling and control of formations of nonholonomic mobile robots , '' , vol .
17 , no . 6 , pp . 9058 , 2001 .
tabuada , g. pappas , and p. lima , `` feasible formations of multi - agent systems , '' , 2002 , submitted
. r. olfati - saber and r. m. murray , `` graph rigidity and distributed formation stabilization of multi - vehicle systems , '' in _ ieee conf .
on decision and control _ , las vegas , nv , 2002 , to appear .
c. tomlin , g. j. pappas , and s. s. sastry , `` conflict resolution for air traffic management : a study in multiagent hybrid systems , '' , vol .
4 , pp . 50921 , 1998 .
e. frazzoli , z. h. mao , j. h. oh , and e. feron , `` aircraft conflict resolution via semi - definite programming , '' , vol . 24 , no .
1 , pp . 7986 , 2001 .
h. choset , `` coverage for robotics - a survey of recent results , '' , vol .
31 , pp . 113126 , 2001 .
r. bachmayer and n. ehrich leonard , `` vehicle networks for gradient descent in a sampled environment , '' in _ ieee conf . on decision and control
2002 , to appear .
a. jadbabaie , j. lin , and a. s. morse , `` coordination of groups of mobile autonomous agents using nearest neighbor rules , '' , july 2002 , to appear .
e. klavins , `` communication complexity of multi - robot systems , '' in _ workshop on algorithmic foundations of robotics _ , nice , france , dec .
2002 , submitted .
j. corts , s. martnez , t. karatas , and f. bullo , `` coverage control for mobile sensing networks , '' in _ ieee int .
conf . on robotics and automation _ ,
arlington , va , may 2002 , pp . 13271332 .
j. corts , s. martnez , t. karatas , and f. bullo , `` coverage control for mobile sensing networks : variations on a theme , '' in _
mediterranean conference on control and automation _ ,
lisbon , portugal , july 2002 , electronic proceedings .
r. a. brooks , `` a robust layered control - system for a mobile robot , '' , vol .
1 , pp . 1423 , 1986 . c. w. reynolds , `` flocks , herds , and schools : a distributed behavioral model , '' , vol . 21 , no .
4 , pp . 2534 , 1987 .
r. c. arkin , , cambridge university press , new york , ny , 1998 .
m. s. fontan and m. j. mataric , `` territorial multi - robot task division , '' , vol .
5 , pp . 815822 , 1998 .
a. c. schultz and l. e. parker , eds .
, , washington , dc , june 2002 .
kluwer academic publishers , proceedings from the 2002 nrl workshop on multi - robot systems .
t. balch and l. e. parker , eds .
, , a. k. peters ltd . , 2002 .
l. e. parker , `` distributed algorithms for multi - robot observation of multiple moving targets , '' , vol .
3 , pp . 23155 , 2002 .
a. howard , maja j. mataric , and g. s. sukhatme , `` mobile sensor network deployment using potential fields : a distributed scalable solution to the area coverage problem , '' in _ proceedings of the 6th international conference on distributed autonomous robotic systems ( dars02 ) _ , fukuoka , japan , 2002 , pp .
299308 .
n. a. lynch , , morgan kaufmann publishers , san mateo , ca , 1997 .
g. tel , , cambridge university press , new york , ny , second edition , 2001
. j. n. tsitsiklis , d. p. bertsekas , and m. athans , `` distributed asynchronous deterministic and stochastic gradient optimization algorithms , '' , vol .
9 , pp . 80312 , 1986 .
d. p. bertsekas and j. n. tsitsiklis , , athena scientific , 1997 .
s. h. low and d. e. lapsey , `` optimization flow control i : basic algorithm and convergence , '' , vol . 7 , no . 6 , pp . 86174 , 1999 .
g. leibon and d. letscher , `` delaunay triangulations and voronoi diagrams for riemannian manifolds , '' in _ proceedings of the sixteenth annual symposium on computational geometry ( hong kong , 2000 ) _ , new york , 2000 , pp .
341349 , acm .
r. klein , , vol .
400 of _ lecture notes in computer science _ ,
springer verlag , new york , ny , 1989 .
a. suzuki and z. drezner , `` the p - center location problem in an area , '' , vol .
1/2 , pp . 6982 , 1996 .
j. corts and f. bullo , `` distributed lloyd flows for disk covering and sphere packing problems , '' 2002 , preprint . c. cattani and a. paoluzzi , `` boundary integration over linear polyhedra , '' , vol .
2 , pp . 1305 , 1990 .
j. gao , l. j. guibas , j. hershberger , li zhang , and an zhu , `` geometric spanner for routing in mobile networks , '' in _ acm international symposium on mobile ad - hoc networking & computing _ , long beach , ca , oct .
2001 , pp . 4555 .
x .- y . li and
wan , `` constructing minimum energy mobile wireless networks , '' , vol .
5 , no . 4 , 2001 .
s. meguerdichian , s. slijepcevic , v. karayan , and m. potkinjak , `` localized algorithms in wireless ad - hoc networks : location discovery and sensor exposure , '' in _ acm international symposium on mobile ad - hoc networking & computing _ , long beach , ca , oct . 2001 .
m. cao and c. hadjicostis , `` distributed algorithms for voronoi diagrams and application in ad - hoc networks , '' preprint , oct .
a. astolfi , `` exponential stabilization of a wheeled mobile robot via discontinuous control , '' , vol . 121 , no .
1 , pp . 1217 , 1999 . h. k. khalil , , prentice hall , englewood cliffs , nj , second edition , 1995 . d. g. luenberger , , addison - wesley , reading , massachusetts , second edition , 1984 .
in this section we collect some relevant facts on descent flows both in the continuous and in the discrete - time settings .
we do this following @xcite and @xcite , respectively .
we include proposition [ prop : discrete - lasalle - surprising ] as we are unable to locate it in the linear and nonlinear programming literature .
consider the differential equation @xmath231 , where @xmath232 is locally lipschitz and @xmath233 is an open connected set .
a set @xmath129 is said to be ( positively ) invariant with respect to @xmath234 if @xmath235 implies @xmath236 , for all @xmath237 ( resp .
@xmath238 ) . a descent function for @xmath234 on @xmath239 , @xmath240 , is a continuously differentiable function such that @xmath241 on @xmath239 .
we denote by @xmath242 the set of points in @xmath239 where @xmath243 and by @xmath129 be the largest invariant set contained in @xmath242 . finally , the distance from a point @xmath244 to a set @xmath129 is defined as @xmath245 .
[ lemma : continuous - lasalle ] let @xmath246 be a compact set that it is positively invariant with respect to @xmath234 .
let @xmath247 and @xmath248 be an accumulation point of @xmath249
. then @xmath250 and @xmath251 as @xmath252 .
let @xmath234 be a subset of @xmath1 .
an algorithm @xmath73 is a continuous mapping from @xmath234 to @xmath234 .
a set @xmath72 is said to be positively invariant with respect to @xmath73 if @xmath253 implies @xmath254 .
a point @xmath248 is said to be a fixed point of @xmath73 if @xmath255 .
we denote the set of fixed points of @xmath73 by @xmath256 . a descent function for @xmath73 on @xmath72 , @xmath257 ,
is any nonnegative real - valued continuous function satisfying @xmath258 for @xmath259 , where the inequality is strict if @xmath260 .
typically , @xmath261 is the objective function to be minimized , and @xmath73 reflects this goal by yielding a point that reduces ( or at least does not increase ) @xmath261 .
[ lemma : discrete - lasalle ] let @xmath262 be a compact set that it is positively invariant with respect to @xmath73 .
let @xmath263 and denote @xmath264 , @xmath265 .
let @xmath248 be an accumulation point of the sequence @xmath266 .
then @xmath267 , @xmath268 and @xmath269 as @xmath270 .
let @xmath248 be an accumulation point of @xmath271 and assume the whole sequence does not converge to it .
then , there exists an @xmath272 such that for all @xmath273 , there is a @xmath274 such that @xmath275
. let @xmath276 be the minimum of all the distances between the points in @xmath256 .
fix @xmath277 .
since @xmath73 is continuous and @xmath256 is finite , there exists @xmath278 such that @xmath279 , with @xmath280 , implies @xmath281 ( that is , for each @xmath282 , there exists such @xmath283 , and we take the minimum over @xmath256 ) .
now , since @xmath268 , there exists @xmath284 such that for all @xmath285 , @xmath286 . also , we know that there is a subsequence of @xmath287 which converges to @xmath248 , let us denote it by @xmath288 . for @xmath187
, there exists @xmath289 such that for all @xmath290 , we have @xmath291 .
let @xmath292 .
take @xmath66 such that @xmath293 then , @xmath294 now we are going to prove that @xmath295 . if @xmath296 , then this claim is straightforward , since @xmath297 .
if @xmath298 , suppose that @xmath299 . since @xmath300 , then @xmath301
therefore , there exists @xmath302 such that @xmath303 .
necessarily , @xmath304 .
now , by the triangle inequality , @xmath305 .
then , @xmath306 which contradicts
. therefore , @xmath307 .
this argument can be iterated to prove that for all @xmath308 , we have @xmath309 .
let us take now @xmath310 such that @xmath311 . since @xmath312 , we have @xmath313 , and therefore @xmath314 which is a contradiction .
therefore , @xmath287 converges to @xmath248 . | this paper presents control and coordination algorithms for groups of vehicles .
the focus is on autonomous vehicle networks performing distributed sensing tasks where each vehicle plays the role of a mobile tunable sensor .
the paper proposes gradient descent algorithms for a class of utility functions which encode optimal coverage and sensing policies .
the resulting closed - loop behavior is adaptive , distributed , asynchronous , and verifiably correct .
coverage control , distributed and asynchronous algorithms , centroidal voronoi partitions |
let @xmath1 be a group , @xmath2\}_{i\in i}$ ] a collection of conjugacy classes in @xmath1 , and @xmath3 some space .
any homomorphism @xmath4 induces a map , @xmath5 well - defined up to homotopy , which sends the circle @xmath6 to a loop in @xmath3 freely homotopic to the element @xmath7 .
[ d : geometric ] let @xmath8 be the free group of rank @xmath9 , and let @xmath10 be a @xmath11dimensional handlebody of genus @xmath9 .
we say a collection of conjugacy classes @xmath12,\ldots,[w_r]\}$ ] in @xmath13 is _ geometric _ if for some isomorphism @xmath14 the induced map @xmath15 is homotopic in @xmath10 to an embedding @xmath16 in this case we also say that the finite presentation @xmath17 is _
geometric_. let @xmath18 be an isomorphism as in definition [ d : geometric ] .
since any automorphism of a free group @xmath19 can be realized as a homeomorphism of the handlebody @xmath10 , the phrase `` for some isomorphism '' can be replaced by `` for any isomorphism '' .
( to see that any automorphism can be realized , first realize elementary nielsen transformations @xmath20 and @xmath21 and then use the fact that these transformations generate @xmath22 , ( * ? ? ?
* i.4 ) . )
a geometric presentation @xmath23 gives a ( not necessarily unique ) way to build a @xmath11manifold @xmath24 with @xmath25 , by attaching @xmath26handles to the boundary of the handlebody @xmath10 , along attaching curves @xmath27 for @xmath28 .
thus every geometric presentation is a presentation of the fundamental group of some @xmath11manifold .
moreover , the _ double _
@xmath29 @xmath30 is the fundamental group of a @xmath11manifold built by attaching thickened annuli to two copies of the handlebody @xmath10 of genus @xmath9 , along the attaching curves given by @xmath31 .
let @xmath12,\ldots,[w_r]\}$ ] be a collection of conjugacy classes in the free group @xmath32 . for @xmath33 a finite index subgroup , we define the _ lifts of @xmath34 _ to be the following set of @xmath35conjugacy classes : @xmath36\ \left|\ \begin{array}{c } g\in f,\ w\in{{\underline{w}}},\mbox { and } \\n(g)\in\z_{>0 } \mbox { minimal so that } gw^{n(g)}g^{-1}\in f ' \end{array } \right\}\right.\ ] ] the collection @xmath34 is _ virtually geometric _ if @xmath37 is geometric for some finite index @xmath33 . in this case
, we say the presentation @xmath38 is _ virtually geometric_. here is an equivalent topological formulation : realize @xmath34 as an embedded collection of circles @xmath39 in @xmath10 . if , for some finite cover @xmath40 , the inclusion @xmath41 is homotopic to an embedding into @xmath42 , then @xmath34 is virtually geometric .
[ r : vg3mfd ] if a presentation @xmath43 is virtually geometric , then there are positive integers @xmath44 so that @xmath45 is virtually a @xmath11manifold group . indeed , realizing @xmath46 as a collection of embedded loops @xmath39 in a genus @xmath9 handlebody @xmath10 ,
let @xmath47 be a _
finite cover in which the preimage @xmath48 is homotopic to an embedding in the boundary . for @xmath49 let @xmath50 be the loop in @xmath39 corresponding to @xmath51 .
each component of @xmath52 covers @xmath53 with the same degree ; let @xmath54 be this degree .
we construct a complex @xmath55 from @xmath56 by attaching @xmath54 disks to each component of @xmath52 , for each @xmath57 .
since @xmath58 is homotopic to an embedding into @xmath59 , the complex @xmath55 is homotopy equivalent to a @xmath11manifold .
moreover , the covering map @xmath60 extends to a covering map @xmath61 where @xmath62 is a complex with @xmath63 . ]
is a virtual @xmath11manifold group for some positive integer @xmath64 , but the author does not know an elementary argument . ]
( an aside : it is possible to show that every finitely presented group admits a presentation which is virtually geometric , but the virtually geometric presentation will nearly always be different from the original one .
briefly , one realizes the group as the fundamental group of a special polyhedron ; a finite branched cover of this polyhedron with positive even - degree branching over the center of every face will be thickenable .
it follows that a presentation coming from the special polyhedron is virtually geometric . for more on
_ special polyhedra _ and _ thickenability _ , see chapter 1 of @xcite . )
the double @xmath29 along a virtually geometric collection @xmath34 is also virtually the fundamental group of a @xmath11manifold , obtained by joining together a pair of handlebodies by thickened annuli .
a @xmath11manifold argument based on dehn s lemma ( * ? ? ?
* lemma 20 ) then shows that if @xmath29 is one - ended , it contains a closed surface subgroup .
gordon and wilton gave some examples of virtually geometric but non - geometric words in @xcite , and asked the following : ( * ? ? ?
* question 22)[q : main ] let @xmath13 be a free group of rank @xmath65 .
is @xmath66\}$ ] virtually geometric for every @xmath67 ? we give a negative answer to this question in section [ s :
answer ] , by showing that the word @xmath68 is not virtually geometric ( see corollary [ cor : k33 ] ) .
danny calegari ( personal communication ) pointed out a second motivation for studying question [ q : main ] , connected with _ stable commutator length _ @xcite .
if a homologically trivial collection @xmath34 is geometric , then @xmath69 is in @xmath70 , by ( * ? ? ?
* proposition 4.4 ) .
( this proposition is an application of gabai s result @xcite on the proportionality of thurston and gromov norms on the second homology of a @xmath11manifold . ) more generally , suppose @xmath34 is virtually geometric , so that @xmath37 is geometric for some @xmath35 with @xmath71<\infty$ ] .
a straightforward argument then shows that @xmath72}(\z+\frac{1}{2}).\ ] ] thus a positive answer to question [ q : main ] for words in @xmath73 $ ] would have given an alternate proof of the rationality of @xmath74 in free groups , which is the main theorem of @xcite .
in fact , the answer to question [ q : main ] is negative even in this more restricted setting ( see corollary [ cor : comm ] ) .
this section describes certain graphs which arise naturally from collections of words together with disk systems in handlebodies . the material in this section is discussed in more detail in berge @xcite . in the graphs we consider ,
multiple edges may connect a pair of vertices , but no edge joins a vertex to itself .
the _ valence _ of a vertex is the number of incident edges ( not the number of neighbors ) , and a graph will be said to be _ regular _ if every vertex has the same valence . a collection of conjugacy classes in a free group @xmath75
can also be thought of as a collection of _ cyclic words _ , i.e. , freely and cyclically reduced `` words '' @xmath76 whose letters are indexed mod @xmath77 where @xmath77 is the length of @xmath51 .
given a collection @xmath34 of cyclic words , one can form the _
whitehead graph _
@xmath78 as follows : 1 .
the @xmath79 vertices of @xmath78 are the generators @xmath80 of @xmath13 and their inverses .
2 . for each two - letter sequence @xmath81 in some @xmath82 ,
attach an edge from @xmath83 to @xmath84 .
( if some @xmath85 has length one , attach an edge from @xmath83 to @xmath86 . )
if @xmath13 is the free group @xmath87 , and @xmath88 , then the whitehead graph @xmath78 is isomorphic to the complete bipartite graph @xmath89 .
[ r : wggeo ] again , there is a topological interpretation : realize the collection @xmath34 as an embedded @xmath90submanifold @xmath64 in the handlebody @xmath10 .
there is a collection of disks @xmath91 in @xmath10 so that for each @xmath57 there is a loop in @xmath10 representing the generator @xmath92 which intersects @xmath93 exactly once , and has empty intersection with @xmath94 for all @xmath95 .
after homotoping @xmath64 to intersect @xmath96 minimally , split the handlebody along @xmath96 to give a ball with @xmath79 disks @xmath97 in the boundary , and @xmath98 a @xmath90submanifold with boundary in @xmath99
. the whitehead graph can then be obtained as the quotient of @xmath100 by crushing each disk to a point . notice that if @xmath64 was homotopic to an embedding into @xmath101 which still intersects @xmath96 minimally , then the whitehead graph is planar . as shown by an example of berge (
* section 12 ) , the assumption that @xmath64 still intersects @xmath96 minimally after being homotoped into @xmath101 is necessary , even for geometric @xmath34 ; there are examples of geometric collections of words with non - planar whitehead graph .
however , we will see in theorem [ t : planar ] that if @xmath96 is _ minimal _ in a certain sense , and @xmath34 is geometric , then the whitehead graph is planar . a _ complete system of disks _ for @xmath10 is a collection @xmath102 of properly embedded disks which is maximal with respect to the condition that @xmath103 is connected .
the system @xmath96 from remark [ r : wggeo ] is an example of a complete system
. different choices of disks determine different choices of free generators for @xmath104 .
( actually , there are multiple complete disk systems corresponding to any choice of free generators ; these disk systems are related to one another by homeomorphisms of @xmath10 which act trivially on @xmath104 . )
any collection of disjoint properly embedded disks in @xmath10 ( complete or not ) gives rise to a graph @xmath105 in exactly the same manner as described in remark [ r : wggeo ] .
so long as @xmath102 is a complete system , the graph @xmath106 always has @xmath79 vertices , but the number of edges depends on the particular complete system @xmath102 .
( the number of edges is the number of times @xmath102 intersects @xmath34 , after @xmath34 has been homotoped to make this intersection minimal . )
the complete system @xmath102 is said to be _ minimal _ ( with respect to @xmath34 ) if the number of edges is minimal over all complete systems of disks for @xmath10 .
as explained in @xcite , the following can be deduced from a result of zieschang ( * ? ? ?
* theorem 1 ) .
* section 11)[t : planar ] if @xmath102 is minimal , and @xmath34 is geometric , then @xmath106 is planar .
( in case the hypotheses of theorem [ t : planar ] are satisfied , berge ( * ? ? ?
* section 13 ) explains further how to obtain a _ heegaard diagram _ , and thus an explicit embedding of @xmath39 in @xmath101 , from @xmath106 . )
the next theorem is explained in ( * ? ? ?
* section 8) , and follows from the `` peak reduction '' lemma of higgins lyndon ( see @xcite or ( * ? ? ? *
i.4 ) , cf .
* section 4 ) for an account with a different flavor ) .
* section 8)[t : reducible ] if @xmath102 is not minimal , then @xmath106 contains a pair of vertices @xmath107 and @xmath108 with the same valence @xmath109 and a collection of @xmath110 edges @xmath111 which separate @xmath107 from @xmath108 .
the pair of vertices @xmath112 in theorem [ t : reducible ] is the pair of vertices coming from some single disk @xmath113 in @xmath102 .
one can obtain a new disk system @xmath114 for @xmath10 by deleting @xmath113 and adding a disk @xmath115 which intersects each of the edges @xmath116 exactly once , without intersecting any other edges .
the whitehead graph @xmath117 is obtained from @xmath106 by adding a new pair of vertices ( corresponding to @xmath115 ) , deleting @xmath118 , and `` splicing together '' the edges incident to @xmath118 .
( we will give a precise definition of _ splicing _ in the next section . )
an example of such an operation ( called a _ whitehead move _ ) is shown in figure [ f : whmove ] .
let @xmath119 and @xmath120 be graphs containing vertices @xmath121 , @xmath122 of the same valence .
let @xmath123 be a bijection from the edges incident to @xmath124 to the edges incident to @xmath125 .
we can form a new graph @xmath1 from this data by the procedure : 1 .
form the disjoint union of @xmath119 and @xmath120 .
2 . delete the vertices @xmath124 and @xmath125 ( and any adjacent edges ) .
3 . for each edge
@xmath126 connecting @xmath124 to some @xmath127 , the corresponding edge @xmath128 connects @xmath125 to some @xmath129 .
add an edge to @xmath1 connecting @xmath130 to @xmath131 .
if @xmath1 is obtained by such a procedure , we say that @xmath1 is obtained by _
splicing @xmath119 and @xmath120 at @xmath132_. ( there may be more than one such @xmath1 for a given @xmath133 . ) [ l : minor ] let @xmath134 be vertices for @xmath135 .
if @xmath136 is connected , and @xmath1 is obtained by splicing @xmath119 and @xmath120 at @xmath132 , then @xmath119 is a minor of @xmath1 .
collapse @xmath136 to a point to obtain a graph isomorphic to @xmath119 .
we say a graph is _ @xmath137edge - connected _ if no collection of @xmath138 or fewer edges disconnects the graph .
[ l : connected ] let @xmath139 , and suppose that @xmath1 is obtained by splicing @xmath119 and @xmath120 at @xmath132 . if @xmath119 and @xmath120 are regular , @xmath109valent , @xmath109edge - connected graphs , then so is @xmath1 . for @xmath135 ,
let @xmath140 be the graph obtained from @xmath141 by removing @xmath142 and all the edges adjacent to @xmath142 .
the graph @xmath140 will be thought of either as a subgraph of @xmath141 or as a subgraph of @xmath1 , depending on context . for @xmath135
we claim that @xmath140 is @xmath143edge - connected . indeed , suppose that @xmath144 is disconnected by @xmath145 or fewer edges into components @xmath146 for some @xmath147 . for some @xmath148 ,
the number of edges connecting @xmath124 to @xmath149 in @xmath119 is at most @xmath150 .
it follows that @xmath119 can be disconnected by at most @xmath151 edges , contradicting the @xmath109edge - connectedness of @xmath119 .
let @xmath152 be a set of @xmath153 or fewer edges in @xmath1 .
either @xmath144 or @xmath154 contains at most @xmath145 edges in @xmath152 .
switching labels if necessary , we suppose it is @xmath144 . since @xmath144 is @xmath143edge - connected , the subgraph @xmath155 is connected .
let @xmath156 be the ( not necessarily distinct ) vertices of @xmath119 connected to @xmath124 , and let @xmath157 be the set of @xmath109 edges in @xmath1 connecting @xmath144 to @xmath154 .
finally , let @xmath158 be @xmath159 notice that there is a quotient map @xmath160 which takes each @xmath161 to @xmath125 .
since @xmath120 is @xmath109edge - connected , @xmath162 does nt disconnect @xmath120 .
this implies that every vertex of @xmath163 is connected in @xmath164 to @xmath161 for some @xmath165 .
since the @xmath161 are all in @xmath144 , and @xmath166 is connected , @xmath167 is connected .
we now explain how splicing arises in building whitehead graphs .
let @xmath34 be a collection of words , as in section [ s : whitehead ] .
let @xmath102 be a complete system of disks for the handlebody @xmath10 , and let @xmath56 be some @xmath168fold cover of @xmath10 , with @xmath169 .
the preimage @xmath170 in @xmath56 has too many disks to be a complete disk system for @xmath56 , but we can still form the graph @xmath171 .
each component of the complement of @xmath170 looks exactly like the complement of @xmath102 in @xmath10 , so we obtain the following : [ l : disjoint ] the graph @xmath171 is isomorphic to a disjoint union of @xmath168 copies of @xmath106 . to get a complete system of disks for @xmath56
, we must delete some disks from @xmath170 .
each time we delete one of these disks , the number of complementary components goes down by one , and the whitehead graph changes by splicing two components together : [ l : splicecomplete ] let @xmath172 be disk systems in a handlebody @xmath10 so that @xmath173 is complete and @xmath174 is a single disk .
let @xmath34 be a collection of conjugacy classes in @xmath104 .
the graph @xmath175 is obtained from @xmath176 by splicing two components of @xmath177 together at a vertex .
( similarly , the whitehead move in figure [ f : whmove ] can be described as an `` unsplicing '' of the edges @xmath116 followed by splicing the resulting components together at @xmath112 . )
[ t : main ] let @xmath102 be a complete system of disks for the handlebody @xmath10 , and let @xmath34 be a collection of conjugacy classes in @xmath104 .
let @xmath178 .
if @xmath106 is a regular , @xmath109valent , @xmath109edge - connected non - planar graph , then @xmath34 is not virtually geometric .
let @xmath56 be a @xmath168fold cover of @xmath10 , corresponding to @xmath33 , where @xmath179 , and let @xmath170 be the preimage of @xmath102 in @xmath56 . by lemma [ l : disjoint ] , the whitehead graph @xmath171 is equal to @xmath168 disjoint copies of @xmath106 . if @xmath180 is a complete system of disks for @xmath56 , then lemma [ l : splicecomplete ] implies that @xmath181 is obtained from @xmath171 by successively splicing together distinct components in some way ( determined by the choice of complete system @xmath182 ) until only @xmath183 vertices remain
. exactly @xmath184 splices will be performed , and each splice reduces the number of components by one , by lemma [ l : connected ] , so @xmath181 will be connected .
moreover , induction on @xmath168 together with lemma [ l : connected ] imply that @xmath181 is @xmath109valent and @xmath109edge - connected . by theorem [ t :
reducible ] , the disk system @xmath96 is minimal . a @xmath109valent ,
@xmath109edge - connected graph can not be disconnected by removing a single vertex , so induction on @xmath168 and lemma [ l : minor ] imply that the graph @xmath106 is a minor of @xmath181 .
it follows that @xmath181 is non - planar .
theorem [ t : planar ] then implies that @xmath37 is not geometric in @xmath56 . since the finite cover @xmath56 was arbitrary , @xmath34 is not virtually geometric .
a concrete example is the word @xmath185 ; the whitehead graph @xmath186 is isomorphic to the complete bipartite graph @xmath89 .
this graph is @xmath11valent , @xmath11edge - connected , and non - planar . applying theorem [ t : main ] gives the following .
[ cor : k33 ] the word @xmath187 in @xmath188 is not virtually geometric . for an example lying in the commutator subgroup , we take @xmath189 .
( we use the common convention @xmath190 , @xmath191 , @xmath192 . ) the whitehead graph , shown in figure [ f : comm ] , is non - planar , @xmath193valent , and @xmath193edge - connected , so theorem [ t : main ] immediately gives the following . [ cor : comm ] the word @xmath194 in @xmath188 is not virtually geometric . the curious reader should be able to construct numerous similar examples , but it should be noted that either of the above examples gives non - virtually geometric words in @xmath19 for all @xmath65 , as the following proposition shows .
let @xmath195 be a non - virtually geometric collection of words in @xmath196 .
for all @xmath65 there is an embedding @xmath197 so that @xmath198 is not virtually geometric in @xmath19 . for @xmath199
we argue as follows : let @xmath200 be any map with image an index @xmath26 subgroup . indeed , the homomorphism @xmath201 is realized by a double cover of handlebodies @xmath202 , and any cover @xmath56 of @xmath203 either covers @xmath204 itself or has a double cover which covers @xmath204 .
the lifts of @xmath34 in this cover are contained in the lifts of @xmath205 .
thus if @xmath205 were virtually geometric , then @xmath34 would be virtually geometric . for @xmath206 , we let @xmath197 be any embedding of @xmath196 as a free factor of @xmath19 . to see why @xmath198 is non - virtually geometric , it is easiest to argue topologically .
let @xmath204 be the handlebody of genus @xmath11 ; we can obtain a handlebody @xmath207 of genus @xmath9 by attaching @xmath208 one - handles along @xmath209 disjoint disks in the boundary of @xmath204 .
realize @xmath34 as an embedded @xmath90manifold @xmath64 in @xmath210 .
suppose there is some finite cover @xmath211 in which the preimage of @xmath64 is homotopic to an embedded submanifold of the boundary .
there is a corresponding finite cover @xmath212 so @xmath213 embeds in @xmath56 .
there is also an infinite - sheeted cover @xmath214 corresponding to @xmath215 ; the space @xmath216 can be obtained from @xmath213 by attaching finitely many non - compact @xmath11manifolds with boundary @xmath217 to @xmath213 along disks .
each of the @xmath218 is homeomorphic to a ball with a cantor set removed from its boundary . @xmath219\ar[d]^{p '' } & \tilde{h}'\cup\bigcup_ia_i\\ \tilde{h}'\ar[ur]\ar[r]\ar[d]^{p ' } & \tilde{h}\ar[d]^p & \\ h_3 \ar[r ] & h_n & } } \]]notice that @xmath220 coincides with the union of the compact components of @xmath221 . if @xmath222 is homotopic to an embedding in the boundary of @xmath56 , then @xmath221 is homotopic to an embedding in the boundary of @xmath223 . since @xmath224 ,
this embedding is homotopic to an embedding in the boundary of @xmath213 , which covers @xmath204 .
this contradicts the assumption that @xmath34 is not virtually geometric in @xmath196 .
gordon and wilton s motivation for asking question [ q : main ] was the question of which doubles of free groups contain closed surface subgroups .
the double @xmath225 contains a surface subgroup , by an application of the main theorem of @xcite .
the double @xmath226 contains a closed surface subgroup by an explicit construction of sang - hyun kim @xcite .
the author thanks danny calegari for pointing out this problem , danny calegari , daryl cooper , and richard schwartz for useful conversations , and nathan dunfield for a python script which was useful for running computer experiments with john berge s program heegaard . thanks also to john berge for pointing out a bad typo in an earlier version of this note , and thanks to the anonymous referee for useful comments .
the author was visiting the caltech mathematics department while this work was done , and thanks caltech for their hospitality .
this work was partly supported by the national science foundation , grant dms-0804369 . | motivated by a question of gordon and wilton , we consider the question of which collections of words are `` virtually geometric '' .
in particular , we prove that some words ( e.g. @xmath0 ) are not virtually geometric . |
there has been much recent interest in the effects of galactic discs on the lensing cross - sections of galaxies ( maller , flores & primack 1997 ; wang & turner 1997 ; bartelmann & loeb 1998 ; keeton & kochanek 1998 ; koopmans , de bruyn & jackson 1998 ; mller & blain 1998 ) . a gravitationally dominant baryonic disc in the core of a dark matter halo , edge - on to the line of sight to a distant source ,
can significantly increase the mass surface density , and thus the lensing efficiency . the image geometry produced by an individual lens
can be modified significantly by the presence of an edge - on or almost edge - on disc ; however , the cross - section to multiple imaging is only increased by a factor of about 50per cent when averaged over all disc inclinations .
in contrast , the cross - section to magnifications greater than 10 is expected to be increased by a factor of a few after the same averaging process ( mller & blain 1998 ) .
hence , depending on the abundance and form of evolution of discs in dark matter haloes , the strong lensing optical depth at moderate and large redshifts could be increased very significantly .
bartelmann & loeb ( 1998 ) have also recently discussed this effect .
they note that it will probably not be possible to exploit any enhanced magnification in the optical waveband because the flux densities of the lensed images will be attenuated by dust associated with the disc . for the same reason , maller et al .
( 1997 ) did not consider the effect of disc lensing on the magnifications of multiple images because of possible differential extinction between the images .
dust extinction is important in the optical waveband ; however , a dusty lensing disc is optically thin in the radio and millimetre / submillimetre wavebands . indeed , because of @xmath0correction effects ( blain & longair 1993 ) , a lensing object is often expected to be fainter than the lensed images in the submillimetre waveband ( blain 1997a , 1998b ) .
hence , any enhanced magnification bias due to the effects of discs would be easy to detect in these wavebands .
existing radio lens surveys , such as the class survey ( browne et al .
1998 ) , appear to show an overabundance of lens candidates that can be identified as highly inclined discs .
the potential importance of submillimetre - wave lens surveys has been discussed for lensing by both clusters ( blain 1997a , 1998a ) and galaxies ( blain 1996 , 1997b , 1998b , c ) .
submillimetre - wave lensing by clusters has now been demonstrated to be a very useful phenomenon ( smail , ivison & blain 1997 ) .
the most striking feature of submillimetre - wave galaxy lens surveys is that the fraction of strongly lensed galaxies detected in a carefully selected sample can exceed 5 per cent ( blain 1996 , 1998b ) . in section 2
we discuss the statistical features of lensing by galaxies containing a disc , using techniques and models of lensing by disc galaxies developed by mller ( 1996 ) , maller et al .
( 1997 ) and mller & blain ( 1998 ) , and the statistical lensing formalism ( peacock 1982 ; pei 1995 ) applied in the submillimetre waveband by blain ( 1996 ) . in section3
we predict the effects of a potentially evolving population of discs ( lilly et al .
1998 ; mo , mao & white 1998 ; mao , mo & white 1998 ; marzke et al .
1998 ) on the lensing cross - sections . in section4
we estimate the surface density of strongly lensed galaxies and quasars in the light of recent submillimetre - wave observations by smail et al .
( 1997 ) using the new scuba camera at the james clerk maxwell telescope ( holland et al .
see smail et al .
( 1998 , 1999 , in preparation ) , ivison et al .
( 1998 ) , blain , ivison & smail ( 1998a ) , blain et al .
( 1998b , 1999a , b ) and frayer et al .
( 1998 ) for a discussion of the latest results . in section 5 the prospects for detecting such images in future surveys using forthcoming
ground - based and space - borne instruments are discussed .
the values of cosmological parameters have a significant effect on the properties of galaxy galaxy submillimetre - wave lensing ( blain 1998a ) . unless otherwise stated , we assume an einstein de sitter world model with hubble s constant @xmath1kms@xmath2mpc@xmath2 , in accordance with the models presented by maller et al .
( 1997 ) and mller & blain ( 1998 ) .
note that the abundance of submillimetre - wave lensed galaxies predicted in an einstein de sitter model is less than that expected in a model with a smaller density parameter ( blain 1998a ) .
millimetre / submillimetre - wave source counts of lensed images were predicted by blain ( 1996 ) , using formalism developed by peacock ( 1982 ) and pei ( 1995 ) to estimate the optical depth to strong lensing attributable to galaxies . in this formalism
the probability that a source at redshift @xmath3 is lensed by a magnification between @xmath4 and @xmath5 is @xmath6 .
this function is obtained by integrating the cross - sections of individual lenses to magnifications greater than @xmath4 , @xmath7 in the plane of the lens , both along the line of sight to @xmath3 and over the mass function of lenses @xmath8 , and then differentiating with respect to @xmath4 , @xmath9 here @xmath10 is the comoving radial distance element .
the number and flux density conservation conditions , @xmath11 must be satisfied . if the intrinsic luminosity function of galaxies @xmath12 is known , then the source count of unlensed galaxies @xmath13 @xmath14 is the luminosity of a source at a redshift @xmath3 that produces a flux density @xmath15 ; @xmath16 is the maximum redshift of the source population ; and @xmath17 is the angular diameter distance between redshifts @xmath18 and @xmath19 .
the count of lensed galaxies can be determined using equation ( 3 ) , if @xmath20 is replaced by an effective luminosity function , @xmath21 corrected to include the effects of lensing .
@xmath22 , the maximum magnification , is determined by the intrinsic size of the source and the geometry of the lens configuration . in general , larger values of @xmath22 are expected for smaller sources .
@xmath23 is expected for lensing by a singular isothermal sphere ( sis ) lens if the intrinsic size of the source is 1kpc ( peacock 1982 ) .
the cross - section to high - magnification lensing by an sis at redshift @xmath24 for a source at redshift @xmath3 , as measured in the plane of the lens , @xmath25 ^ 2 \left [ { { d(0,z')d(z',z ) } \over { d(0 , z ) } } \right]^2 \mu^{-2},\ ] ] ( schneider , ehlers & falco 1992 ) .
@xmath26 is the mass of the lens enclosed within a radius @xmath27 .
the distance ratio inside the second square bracket is referred to as @xmath28 in subsequent equations . in the case of lensing by siss ,
the magnification and redshift dependences of @xmath29 ( equation 1 ) can be separated at large magnifications , yielding the strong lensing probability @xmath30 that is applicable here .
the form of @xmath31 at all magnifications can be approximated by @xmath32 the first term describes the typically small magnification correction that must be applied to most sources in order to satisfy the number and flux density conservation conditions ( equation 2 ) .
the form of the function @xmath33 depends on both the values of cosmological parameters and the form of evolution of lensing galaxies ( blain 1996 , 1998a ) . if the distribution of lenses in mass @xmath34 is described by a press
schechter function ( press & schechter 1974 ) , @xmath35 \ , { \rm d}m,\ ] ] then equation ( 1 ) can be evaluated analytically .
@xmath36 includes all the details of the evolution of structure resulting from the growth of perturbations in an expanding universe ; @xmath37 if the primordial density fluctuations are scale - independent ; @xmath38 is the mean smoothed density of bound objects in the universe ; and @xmath34 is the mass of a region of the universe that has turned around from the hubble flow and is in the process of collapse at a particular epoch .
@xmath34 is related to the mass enclosed within a radius @xmath27 , @xmath26 ( equation 5 ) by the constant @xmath39 , that is , @xmath40 @xmath41 is the equivalently transformed version of @xmath42 .
this transformation allows the masses in equations ( 5 ) and ( 7 ) to be linked in the same expression .
for example , if the population of lensing galaxies consists entirely of siss , then by evaluating equation ( 1 ) , @xmath43 ^ 2 \gamma \left [ { 1 \over \gamma } + { 1 \over 2 } \right ] \times } \\ & & \>\>\>\>\>\>\>\>\>\>\>\>\>\>\>\>\ >
\int_0^z d_{\rm r}^2(z',z ) \,(1+z')^2 m_{\rm e}^*(z ' ) { { { \rm d}r } \over { { \rm d}z ' } } \ , { \rm d}z'.\end{aligned}\ ] ] if @xmath31 is given by equation ( 6 ) then @xmath44 in which @xmath45 .
this equation was used to derive the results for the probabilities of statistical lensing presented by blain ( 1996 , 1998a ) .
mller & blain ( 1998 ) recently presented the results of simulations of the high - magnification lensing cross - section for an exponential disc embedded within an sis halo .
when the disc was within about @xmath46 of being edge - on to the line of sight , this cross - section was found to be increased by a significant factor as compared with that for an sis halo alone .
an example of the relative magnification @xmath47 is plotted in fig.3(a ) of mller & blain ( 1998 ) , for a disc / halo combination well matched to that of the milky way .
we have now extended these simulations to include a wide range of halo and disc masses , and so to investigate the statistical properties of lensing by such objects .
the central surface mass density of the disc within the halo was kept constant ( mao et al . 1998 ) , while the scalelength of the disc was varied in such a way that the ratio of the mass of dark matter in the sis halo within a radius @xmath48kpc and the mass of the disc remained constant .
the results are now parametrized to generate an analytical description of the effects of lensing by a population of discs .
the relative cross - section to lensing by a magnification greater than @xmath4 due to a combination of a disc and an sis halo , as compared with an sis halo alone , was approximated by a function of the form @xmath49 where @xmath50.\ ] ] any systematic evolution of the scalelength / mass of discs within dark matter haloes can be incorporated by modifying the value of @xmath51 , defined as the mass of dark matter enclosed within 10kpc of the core of an sis halo , as a function of redshift .
0.75 mm ' '' '' 1.2 mm .the values of the best - fitting parameters derived from the extended results of mller & blain ( 1998 ) .
the associated best - fitting curves are shown in fig.1 .
the best - fitting values of the worst constrained parameters , @xmath52 and @xmath53 , are related by the equation @xmath54 .
[ cols= " < , < , < , < " , ] 1.2 mm ' '' '' 1.2 mm p1.1 cm p2.5 cm p1.2 cm p1.2 cm p1.2 cm a & no & 0.60 & 0 & 0 + b & yes & 0.60 & -1 & 1 + ' '' '' the optical depths to strong lensing @xmath55 and @xmath31 can now be calculated and compared . in fig.4 @xmath56
is shown for both models of disc evolution listed in table2 as a function of both magnification and redshift . the optical depth to strong lensing
is increased by incorporating the effect of lensing by discs by a factor between 2 and 3 at redshifts of about unity and magnifications of several tens .
the effect of including discs is more significant at greater magnifications .
the consequences for the counts of lensed galaxies and quasars in the submillimetre waveband are discussed below .
the selection functions in submillimetre - wave surveys are very broad and extend out to high redshifts .
this is because the flux density redshift relation for distant dusty galaxies is expected to be flat over the range of redshifts from about 0.5 to 10 , owing to large negative @xmath0corrections .
these arise when dust emission spectra , peaking in the rest - frame far - infrared waveband , are redshifted into the submillimetre waveband ( blain & longair 1993 ) .
a submillimetre - wave survey thus selects high - redshift galaxies preferentially , and because a flat flux density redshift relation naturally corresponds to a steep source count , magnification biases in such a survey are expected to be large ( blain 1996 ) . at that time , however , source counts of submillimetre - wave sources had not been determined .
now smail et al .
( 1997 ) have detected a population of distant dusty galaxies , including a dust - shrouded starburst / agn at a redshift @xmath57 ( ivison et al .
1998 ; frayer et al . 1998 ) . a fuller description of the survey ( smail et al .
1998 , 1999 , in preparation ; blain et al .
1999b ) and its consequences ( blain 1998c ; blain et al .
1998a , b , 1999a ) can be found elsewhere .
the results of other submillimetre - wave surveys are now becoming available ( barger et al . 1998 ; eales et al . 1998 ; holland et al .
1998 ; hughes et al . 1998 ) .
the surface density of galaxies with flux densities greater than 4mjy at a wavelength of 850@xmath4 m is @xmath58deg@xmath59 ( smail et al .
1997 ) . counts of lensed galaxies that are consistent with these observations , which do not include the effects of lensing by discs , were presented by blain ( 1998a , b , c ) .
the local population of _ iras _ galaxies ( saunders et al .
1990 ) appears to undergo pure luminosity evolution of the form @xmath60 where @xmath61 out to @xmath62 ( blain et al .
this model is consistent with counts at other wavelengths in the millimetre / submillimetre waveband ( wilner & wright 1997 ; kawara et al .
1998 ; lagache et al .
1998 ) , and with the intensity of diffuse background radiation in the submillimetre / far - infrared waveband ( puget et al . 1996
; guiderdoni et al . 1997 ; schlegel , finkbeiner & davis 1998 ; hauser et al . 1998 ; fixsen et al .
1998 ) .
we determine counts of lensed galaxies by using equation ( 3 ) with @xmath63 calculated using @xmath55 ( equation 15 ) , normalised as described above , in both models of disc evolution described in table2 .
the population of distant dusty galaxies assumed is described by the ` gaussian ' model of blain et al .
( 1999a ) , which accounts for all the available submillimetre - wave and far - infrared data currently available .
in order to illustrate the expected form of the counts of both lensed and unlensed galaxies at wavelengths of 1300 , 850 , 450 and 60@xmath4 m , the counts of galaxies expected in the absence of discs are compared in fig.5(a ) .
lensed counts are calculated , assuming that @xmath42 does and does not evolve .
if @xmath42 does not evolve , then the predicted count of lensed galaxies is greater . as a result ,
all previous predictions have been made assuming that @xmath42 evolves to produce conservative estimates of the counts of lenses . to demonstrate the effects of including lensing by discs , the relative size of the counts of lensed galaxies with discs as compared with the counts shown in fig.5(a )
are shown in fig.5(b ) .
the additional bias arising from disc lensing is typically a factor of 2 and 3 in models a and b respectively .
the only data currently available on the population of lensed dusty galaxies and quasars is provided by the identification of two luminous distant 60-@xmath4 m _ iras _ sources as lenses ( close et al .
1995 ; irwin et al . 1998 ; lewis et al .
1998 ) , and the identification of one known lensed quasar with a 60-@xmath4 m _ iras _ source ( kneib et al .
this indicates a 60-@xmath4 m count of lenses of at least @xmath64deg@xmath59 at a flux density of 0.2jy , fully consistent with the value of about @xmath65deg@xmath59 shown in fig.5(a ) .
the maximum magnification assumed in these calculations for a source at @xmath66 is @xmath67 , as assumed by blain ( 1996 ) .
a larger value of @xmath22 will lead to a more significant effect of lensing by discs , as discs are progressively more efficient at producing large magnifications as compared with sis haloes : see figs1(a ) and 4 .
this effect is emphasized in fig.6 , in which the additional magnification bias arising from lensing by discs is compared for different values of @xmath68 : 10 , 20 , 40 and 100 .
observations made using millimetre - wave interferometers ( solomon et al .
1997 ; mirabel et al . 1998 ; sakamoto et al .
1999 ) indicate that the most luminous far - infrared emitting regions of distant galaxies are rather small , with scales of less than several hundred parsecs .
as @xmath69 is expected for a 1-kpc source ( peacock 1982 ) , @xmath70 seems to be very reasonable .
larger magnifications may also be possible , at least for any very compact subsample of distant submillimetre - luminous galaxies and quasars , such as the ` extreme starbursts ' imaged by downes & solomon ( 1998 ) .
in this case , the numbers of lenses with flux densities of order 100mjy , and thus the magnification biases due to discs ( fig.5b ) , will be increased .
( a ) -5 mm ( b ) -4 mm the fraction of lensed galaxies detected in a carefully designed flux - limited submillimetre - wave survey is expected to be several per cent ( blain 1996 , 1998b ) , and it is reasonable to expect lensing by discs to increase this fraction by a factor of between 2 and 3 .
many forthcoming instruments will be useful for carrying out submillimetre - wave lens surveys .
foremost amongst them are large ground - based millimetre / submillimetre interferometer arrays ( mias ) , such as the mma ( brown 1996 ) : see also downes ( 1996 ) and ishiguro et al .
( 1992 ) . these instruments , with subarcsecond angular resolution , will be required in order to confirm submillimetre - selected lens candidates ( blain 1998b ) .
large single - antenna telescopes such as the us
mexican 50-m large millimeter telescope ( ltm / gtm ; schloerb 1997 ) and a 10-m telescope at the south pole ( stark et al . 1998 ) , are also being designed and built .
these telescopes will provide large and uniform focal planes , able to exploit large format bolometer array receivers ( glenn et al .
1998 ) , and would be very useful for wide - field millimetre / submillimetre - wave surveys ( blain 1998c ) .
planck surveyor _ ( bersanelli et al . 1996 ) and _ far - infrared and submillimetre telescope _
( _ first _ ; pilbratt 1997 ) space missions also have great potential .
\(a ) ( b ) ( c ) -3 mm -11.5 cm the relative speed of detecting lenses using the instruments discussed above is considered in more detail elsewhere ( blain 1998c ) .
there is freedom to choose a survey strategy for all but _ planck surveyor _ , and , if optimized , detection rates of order 0.03 - 0.3h@xmath2 are expected .
if the effect of lensing by discs is included , then these detection rates would be expected to increase by a factor of about 2 or 3 .
if the optical depth to submillimetre - wave lensing by discs is as large as we expect , then follow - up observations of detected lenses should reveal that a large fraction are lensed by edge - on discs .
submillimetre - wave point sources will be extracted from the all - sky _ planck surveyor _ survey ( blain 1998b ) , down to the confusion limit ( blain et al .
1998a , b ) .
more than a thousand lensed galaxies and quasars were predicted , without considering the effects of lensing discs .
this number should also be increased by a factor of 2 or 3 when lensing by discs is taken into account .
multiple lensed images formed by galaxies are separated on arcsecond scales . a large interferometer array , such as the mma ,
would thus be the ideal telescope with which to investigate lensing by distant galaxies in detail .
in fig.7 we show simulated images of a 10-arcsec wide field , contained within the primary beam of the mma , at an angular resolution of 0.1arcsec .
the contour levels correspond to three times the 850-@xmath4 m 1@xmath71 sensitivity limits in mma integrations lasting 1min , 1h and 1d : 220 , 28 and 5.8@xmath4jybeam@xmath2 respectively .
note that the primary beam area of the mma at 850@xmath4 m is about @xmath72deg@xmath59 , and so the brightest source in this area would be expected to have a flux density of 40@xmath4jy ( fig.5a ) . in fig.7(c ) a typical region of the background sky is shown .
the source distribution in this case is chosen so that one source will be magnified significantly by a disc galaxy at the centre of the figure .
all sources are gaussian and 0.2arcsec across at full width at half - maximum . an alignment between source and lens this close or closer
is expected in about 2per cent of cases . in fig.7(a )
the effects of lensing these sources by an edge - on disc galaxy , similar to the milky way , at a redshift of 0.3 is shown . the submillimetre - wave emission from the lensing galaxy
is calculated by transforming the observation of a nearby edge - on disc galaxy ( israel , van der werf & tilanus 1998 ) to @xmath73 , giving a total flux density of about 1mjy and a scalelength of 1arcsec . in fig.7(b ) the much less significant effects of lensing by an sis dark matter halo with the same mass at the same redshift are compared .
in addition to following up the sources detected in surveys using single - antenna telescopes ( blain 1998c ) , a deep targeted mia survey of known distant edge - on discs , provided for example by the sloan digital sky survey catalogue ( margon 1998 ) , would be a very powerful way of detecting lensed images of very faint distant submillimetre - wave galaxies .
integration times of about 24h per source would be required however .
1 . we have estimated the enhanced magnification bias expected for statistical lensing due to a population of baryonic discs in dark matter haloes .
the bias is expected to be enhanced by a factor of 2 or 3 as compared with a population of sis lenses .
this bias could be much larger if the sources are smaller than several hundred parsecs in size , and can thus be magnified by a factor in excess of 40 .
2 . in the submillimetre waveband
, lensed images will not suffer from the effects of dust absorption in the lensing galaxy .
hence images lensed by edge - on discs should be selected in an unbiased manner in submillimetre - wave surveys made using the mma , lmt / gtm , a 10-m telescope at the south pole , and the space - borne instruments _
planck surveyor _ and _ first_. 3 .
the subarcsecond angular resolution of the mma will be required in order to study lensed images in detail .
a targeted search of known edge - on discs could provide a fruitful technique for detecting and observing any very faint , distant dusty galaxies .
we thank jarle brinchmann , malcolm longair , priya natarajan , joel primack for helpful comments on the manuscript , the referee for his / her prompt reading of the manuscript and the rotary club of erice for their hospitality .
observations made by ian smail , rob ivison , jean - paul kneib and awb have assisted greatly in refining the conclusions of this paper .
blain a.w . ,
ivison r.j . , smail i. , kneib j .-
p . , 1998b , in colombi s. , mellier y. eds , wide - field surveys in cosmology .
proceedings xiv iap meeting , edition frontires , gif - sur - yvette , in press ( astro - ph/9806063 ) ishiguro m. , kawabe r. , nakai n. , morita k .-
okumura s.k .
, ohashi n. , 1992 , in ishiguro m. , welch w.j .
eds , astronomy with millimeter and submillimeter interferometry , pasp conference series , vol .
59 , p.405 stark a.a . ,
carlstrom j.e . , israel f.p .
, menten k.m . ,
peterson j.b . ,
phillips t.g .
, sironi g. , walker w.w . , 1998 , in phillips t.g .
ed . , advanced technology mmw , radio and terahertz telescopes .
spie vol . 3357 , spie , bellingham , p. 495
wang y. , turner e.l . , 1997 , mnras , 292 , 863 | the high - magnification gravitational lensing cross - section of a galaxy is enhanced significantly if a baryonic disc is embedded in its dark matter halo .
we investigate the effects of a population of such discs on the probability of detecting strongly lensed images of distant galaxies and quasars . the optical depth
to lensing is always more than doubled .
the effects are particularly significant at magnifications greater than about 50 , for which the optical depth typically increases by a factor greater than 5 .
if either the fraction of discs or the typical disc - to - halo mass ratio increases with redshift , then the optical depth is expected to be increased even further .
obscuration by dust in the lensing disc is expected to counteract the associated enhanced magnification bias for surveys in the optical waveband , but not in the radio and millimetre / submillimetre wavebands .
the presence of galactic discs should hence lead to a significant increase in the number of lensed galaxies and quasars detected in both radio- and millimetre / submillimetre - selected surveys .
an increase by a factor of about 3 is expected , enhancing the strong case for millimetre / submillimetre - wave lens surveys using the next generation of ground - based telescopes and the _ first _ and _ planck surveyor _ space missions .
galaxies : evolution galaxies : fundamental parameters cosmology : theory gravitational lensing
radio continuum : galaxies |
let @xmath0 be the complement of a reduced ( possibly reducible ) projective curve @xmath1 in the complex projective plane @xmath2 .
the space of characters of the fundamental group @xmath3 has an interesting stratification by subspaces , given by the cohomology of the rank one local system associated with the character : @xmath4 the closures of these jumping loci in @xmath5 were called in @xcite the characteristic varieties of @xmath6 .
more precisely , the characteristic varieties associated to @xmath6 were defined in @xcite as the zero sets of fitting ideals of the @xmath7$]-module which is the complexification @xmath8 of the abelianized commutator of the fundamental group @xmath9 ( cf . section [ prelim ] for more details ) .
in particular the characteristic varieties ( unlike the jumping sets of the cohomology dimension greater than one ) depend only on the fundamental group .
fox calculus provides an effective method for calculating the characteristic varieties if a presentation of the fundamental group by generators and relators is known . for each character @xmath10 the _ depth _ was defined in @xcite as @xmath11 so that the strata are the sets on which @xmath12 is constant . in @xcite
, we describe a geometric interpretation of the depth by relating it to the pencils on @xmath6 i.e. holomorphic maps @xmath13 having multiple fibers .
in fact the discussion in @xcite is in a more general context in which @xmath6 is a smooth quasi - projective variety .
the viewpoint of @xcite ( and @xcite ) is that such a pencil can be considered as a map in the category of orbifolds .
the orbifold structure of the curve @xmath14 is matched by the structure of multiple fibers of the pencil .
the main result of @xcite can be stated as follows : [ thm - main ] let @xmath15 be a pair consisting of a smooth quasi - projective manifold @xmath6 , whose smooth compactification @xmath16 satisfies @xmath17 , and let @xmath18 be a character of the fundamental group of @xmath6 .
assume that the depth of @xmath18 is positive .
then there is a ( possibly non - compact ) orbifold curve @xmath19 , a character @xmath20 of its orbifold fundamental group @xmath21 and @xmath22 strongly independent orbifold pencils @xmath23 such that @xmath24 .
moreover , 1 .
[ thm - main - part1 ] @xmath25 .
[ thm - main - part2 ] if @xmath19 is a global quotient orbifold of a one - dimensional algebraic group , then the inequality above is in fact an equality , @xmath26 .
vice versa , given an orbifold pencil @xmath27 and a character @xmath28 with a positive depth then @xmath29 has a positive depth at least as large as @xmath30 .
we refer to section [ prelim ] for all the required definitions , and in particular , the definition of strongly independent pencils . according to this result , given a character @xmath18 with a positive depth
, one automatically has an orbifold pencil with the depth @xmath12 being bounded below by the constant specified in theorem [ thm - main][thm - main - part1 ] by the geometric data .
one can compare this statement with the numerous previous results on existence of pencils on quasi - projective manifolds which mostly guarantee existence of ordinary pencils corresponding to the components of having a positive dimension ( cf .
@xcite and references therein ) .
for example , if @xmath18 has a positive depth and an infinite order , then it must belong to a component of characteristic variety of positive dimension since the isolated points of characteristic variety have a finite order by @xcite .
hence the results of @xcite can be applied to such a component to obtain a pencil @xmath27 and a character @xmath31 such that @xmath29 . here
@xmath19 is the complement in @xmath32 to a finite set containing say @xmath33 points . moreover , the number of independent pencils in the sense of section [ prelim ] is equal to one ( cf .
remark [ rem - non - orbifold ] ) and the depth of @xmath31 is equal to @xmath34 ) . since it was show in @xcite that @xmath35 the inequality in theorem [ thm - main][thm - main - part1 ] follows from the latter .
note that the orbifold pencils , which theorem [ thm - main ] claims , _ without the orbifold structure _ are just rational pencils and the connection with the jumping loci disappears .
the goal of this paper , is to illustrate in detail both parts [ thm - main - part1 ] and [ thm - main - part2 ] of theorem [ thm - main ] with examples in which orbifolds are unavoidable .
we start with a section reviewing mainly known results on the cohomology of local systems , characteristic varieties , orbifolds , and zariski pairs making possible to read the rest of the paper unless one is interested in the proofs of mentioned results .
then , firstly , in section [ sec - fermat ] , a family of curves is considered for which the characteristic variety containing isolated characters having torsion or arbitrary finite order and whose depth is 3 .
the calculations illustrate the use of fox calculus for finding explicit form description of the characteristic varieties . secondly , in the context of line arrangements , examples of ceva and arrangements
are considered in section [ sec - ceva ] .
their characteristic varieties have been studied in the literature via computer aided calculations based on fundamental group presentations and fox calculus . here
we present an alternative way to study such varieties independent of the fundamental group illustrating the geometric approach of theorem [ thm - main ] .
finally , in section [ sec - zariski - pair ] we discuss on a zariski pair of sextic curves , whose alexander polynomials coincide .
we determine this zariski pair by the existence of orbifold pencils .
the authors want to express thanks to the organizers of the _ intensive research period on configuration spaces : geometry , combinatorics and topology , _ at the _ centro di ricerca matematica ennio de giorgi _ , pisa , in may 2010 for their hospitality and excellent working atmosphere .
the second and third named authors want to thank the university of illinois and zaragoza respectively , where part of the work on the material of this paper was done during their visit in spring and summer 2011 .
[ sec - preliminaries ] in this section the necessary definitions used in theorem [ thm - main ] will be reviewed together with material on the characteristic varieties and zariski pairs with the aim to keep the discussion of the upcoming sections in a reasonably self - contained manner .
characteristic varieties appeared first in the literature in the context of algebraic curves in @xcite .
they can be defined as follows .
let @xmath36 be the decomposition of a reduced curve @xmath1 into irreducible components and let @xmath37 denote the degrees of the components @xmath38 .
let @xmath39 and @xmath40 .
then ( cf .
@xcite ) @xmath41 where @xmath42 is the homology class of a meridian of @xmath38 ( i.e. the boundary of small disk transversal to @xmath38 at a smooth point ) .
let @xmath43 be epimorphism of abelianization .
the kernel @xmath44 of @xmath45 , i.e. the commutator of @xmath46 , defines the universal abelian covering of @xmath6 , say @xmath47 , whose group of deck transformations is @xmath48 .
this group of deck transformations , since it acts on @xmath49 , also acts on @xmath50 .
on @xmath51 by conjugation ] this allows to endow @xmath52 ( as well as @xmath53 ) with a structure of @xmath54-module where @xmath55\approx { \mathbb c}[t_1^{\pm 1},\dots , t_r^{\pm 1}]/(t_1^{d_1}\cdot\ldots\cdot t_r^{d_r}-1).\ ] ] note that @xmath56 can be identified with the commutative affine algebraic group @xmath57 having @xmath58 tori @xmath59 as connected components .
indeed , the elements of @xmath60 can be viewed as the functions on the group of characters of @xmath46 . since @xmath46 is a finitely generated group , the module @xmath61 ( resp .
@xmath62 ) is a finitely generated @xmath54-module : the group @xmath51 is infinitely generated .
] in fact one can construct a presentation of @xmath61 ( resp .
@xmath62 ) with the number of @xmath60-generators at most @xmath63 ( resp .
@xmath22 ) , where @xmath22 is the number of generators of @xmath46 . if @xmath64 is not cyclic ( i.e. @xmath65 or @xmath66 and @xmath67 ) then @xmath60 is not a principal ideal domain . one way to approach the @xmath60-module structure of both @xmath61 and @xmath62 is to study their fitting ideals ( cf .
@xcite ) .
let us briefly recall the relevant definitions . let @xmath68 be a commutative noetherian ring with unity and @xmath69 a finitely generated @xmath68-module .
choose a finite free presentation for @xmath69 , say @xmath70 , where @xmath71 the homomorphism @xmath72 has an associated @xmath73 matrix @xmath74 with coefficients in @xmath68 such that @xmath75 ( the vectors below are represented as the column matrices ) .
[ def - fitting ] the @xmath76-th _ fitting ideal _
@xmath77 of @xmath69 is defined as the ideal generated by @xmath78 it will be denoted @xmath79 if no ambiguity seems likely to arise .
@xcite [ def - charvar ] with the above notations the _ @xmath76-th characteristic variety _ ( @xmath80 ) of @xmath81 can be defined as the zero - set of the ideal @xmath82 @xmath83 then @xmath84 is the set of characters in @xmath85 which do not belong to @xmath86 for @xmath87 .
if a character @xmath18 belongs to @xmath84 then @xmath76 is called the _ depth _ of @xmath18 and denoted by @xmath12 ( cf .
@xcite ) .
an alternative notation for @xmath88 is @xmath89 . essentially without loss of generality one can consider only the cases when the quotient by an ideal in the definition of the ring @xmath54 in is absent i.e. consider only the modules of the ring of laurent polynomials .
indeed , consider a line @xmath90 not contained in @xmath1 and in general position ( i.e. which does not contain singularities of @xmath1 and is transversal to it ) .
then @xmath91 is isomorphic to @xmath92 $ ] .
moreover , since we assume transversality @xmath93 , then the @xmath94-module @xmath95 does not depend on @xmath90 ( see for instance ( * ? ? ?
* proposition 1.16 ) ) . the characteristic variety @xmath96 determines @xmath89 ( cf .
@xcite ) . by abuse of language
it is called the _
@xmath76-th affine characteristic variety _ and denoted simply by @xmath97 .
one can also use the module @xmath62 to obtain the characteristic varieties of @xmath1 .
one has the following connection @xmath98 where @xmath99 denotes the trivial character
. the depth of a character appears in explicit formulas for the first betti number of cyclic and abelian unbranched and branched covering spaces ( cf .
@xcite ) [ charvargroups ] one can also define the @xmath76-th characteristic variety @xmath100 of any finitely generated group @xmath46 ( such that the abelianization @xmath101 or , more generally , for a surjection @xmath102 where @xmath103 is an abelian group ) as the @xmath76-th characteristic variety of the @xmath104$]-module @xmath105 obtained by considering the cw - complex @xmath106 associated with a presentation of @xmath46 and its universal abelian covering space @xmath107 ( respectively considering the covering space of @xmath106 associated with the kernel of the map to @xmath103 ) .
such invariant is independent of the finite presentation of @xmath46 ( resp .
depends only on @xmath102 ) .
this construction will be applied below to the orbifold fundamental groups of one dimensional orbifolds .
note that one has : * @xmath108 , * @xmath109 , for the affine case , and * @xmath110 , where @xmath111 is a @xmath58-th primitive root of unity for the curves in projective plane .
note also that in the case of a finitely presented group @xmath46 such that @xmath112 one has @xmath113 where as above @xmath104 $ ] and @xmath114 is a @xmath115-th primitive root of unity .
let @xmath6 be a smooth quasi - projective variety such that for its smooth compactification @xmath16 one has @xmath17 .
this of course includes the cases @xmath81 .
the structure of the closures of the strata @xmath84 is given by the following fundamental result .
[ arapurath ] the closure of each @xmath84 is a finite union of cosets of subgroups of @xmath116 .
moreover , for each irreducible component @xmath117 of @xmath84 having a positive dimension there is a pencil @xmath118 , where @xmath14 is a @xmath32 with deleted points , and a torsion character @xmath119 such that @xmath120 .
let @xmath121 be curve whose components form a subset of the set of components of @xmath1 .
there is a natural epimorphism @xmath122 induced by the inclusion .
this surjection induces a natural inclusion @xmath123 . with identification of the generators of @xmath60 with components of @xmath1 as above ,
this embedding is obtained by assigning @xmath124 to the coordinates corresponding to those irreducible components of @xmath1 which are not in @xmath125 ( cf .
@xcite ) .
the embedding @xmath123 induces the embedding @xmath126 ( cf .
@xcite ) ; any irreducible component of @xmath127 is the intersection of an irreducible component of @xmath128 with @xmath129 .
irreducible components of @xmath128 contained in @xmath129 for some curve @xmath130 are called _ coordinate components _ of @xmath128 .
if an irreducible coordinate component @xmath131 of @xmath127 is also an irreducible component of @xmath128 , then @xmath131 is called a _
non - essential coordinate component _ , otherwise it is called an _ essential coordinate component_. see @xcite for examples . a detailed discussion of more examples is done in sections [ sec - fermat ] , [ sec - ceva ] , and [ sec - zariski - pair ] . as shown in (
* lemma 1.4.3 ) ( see also ( * ? ? ?
* proposition 3.12 ) ) , essential coordinate components must be zero dimensional . in section [ libg - var ] the characteristic varieties of a finitely
presented group @xmath46 are defined as the zeroes of the fitting ideals of the module @xmath132 over @xmath64 .
this module is referred to in the literature as the _ alexander invariant _ of @xmath46 .
note , however , that this is not the module represented by the matrix of fox derivatives called the _ alexander module _ of @xmath46 .
our purpose in this section is to briefly describe the alexander invariant for fundamental groups of complements of plane curves and give a method to obtain a presentation of such a module from a presentation of @xmath46 . in order to do so ,
consider @xmath133 the fundamental group of the curve @xmath1 . without loss of generality
one might assume that 1 .
@xmath64 is a free group of rank @xmath134 generated by meridians @xmath135 , then one has the following any group @xmath46 as above satisfying @xmath136 admits a presentation @xmath137 where @xmath138 and @xmath139 satisfying : 1 .
@xmath140 , @xmath141 , and @xmath142 can be written in terms of @xmath143 and @xmath144 x_k^{-1}$ ] , where @xmath145 $ ] is the commutator of @xmath146 and @xmath147 .
a presentation satisfying @xmath148 is called a _
zariski presentation _ of @xmath46 .
from now on we will assume @xmath46 admits a zariski presentation as in . in order to describe elements of the module @xmath69
it is sometimes convenient to see @xmath149 $ ] as the ring of laurent polynomials in @xmath134 variables @xmath150 $ ] , where @xmath151 represents the action induced by @xmath42 on @xmath69 as a multiplicative action , that is , @xmath152 for any @xmath153 . 1
. one of course needs to convince oneself that action is independent , up to an element of @xmath154 , of the representative @xmath146 as long as @xmath140 .
this is an easy exercise .
we denote by `` @xmath155 '' equalities that are valid in @xmath69 .
[ exam - rel ] note that @xmath156\ { \smash{\mathop{=}\limits^{m}}}\ [ x , z]+t_x[y , z],\ ] ] where @xmath157 , @xmath158 , and @xmath159 are elements of @xmath46 and @xmath160 denotes @xmath161 in the multiplicative group .
this is a consequence of the following @xmath162=xyzy^{-1}x^{-1}z^{-1}=x(yzy^{-1}z^{-1})x^{-1 } xzx^{-1}z^{-1}\ { \smash{\mathop{=}\limits^{m}}}\ t_x[y , z]+[x , z].\ ] ] as a useful application of one can check that @xmath163\ { \smash{\mathop{=}\limits^{m}}}\ [ x , z]+(t_z-1)[y , x],\ ] ] where @xmath164 .
note that @xmath165 $ ] , @xmath166 and @xmath167 , @xmath168 are elements in @xmath44 , since @xmath169 .
therefore @xmath170 x_k^{-1}\ { \smash{\mathop{=}\limits^{m}}}\ t_k x_{i , j}\ ] ] ( see and @xmath148 ) .
moreover , [ prop - m1 ] for a group @xmath46 as above , the module @xmath69 is generated by @xmath171 and @xmath172 .
[ exam - rel2 ] the module @xmath69 is not freely generated by the set mentioned above , for instance , note that according to @xmath148 and any relation in @xmath46 , say @xmath173 ( as in ) can be written ( in @xmath69 ) in terms of @xmath174 and @xmath143 as @xmath175 . in other words
, @xmath176 is a relation in @xmath69 .
[ exam - rel3 ] even if @xmath46 were to be the free group @xmath177 , @xmath69 would not be freely generated by @xmath174 and @xmath143 .
in fact , @xmath178+(t_y-1)[z , x]+(t_z-1)[x , y]\ { \smash{\mathop{=}\limits^{m}}}\ 0\ ] ] for any @xmath157 , @xmath158 , @xmath159 in @xmath46 .
using example [ exam - rel ] repeatedly , one can check the following @xmath179 = \left\ { \array{l } { \smash{\mathop{=}\limits^{m}}}\ [ x , z]+t_x[y , z]\\ = [ y^{x^{-1}}x , z]\ { \smash{\mathop{=}\limits^{m}}}\ [ y^{x^{-1}},z]+ t_y [ x , z]\ { \smash{\mathop{=}\limits^{m}}}\ [ y , z]-(t_z-1)[x , y]+t_y [ x , z],\\ \endarray \right.\ ] ] where @xmath180 . the difference between both equalities results in @xmath181 .
such relations will be referred to as _ jacobian relations _ of @xmath69 .
a combination of examples [ exam - rel2 ] and [ exam - rel3 ] gives in fact a presentation of @xmath69 . [ prop - m2 ] the set of relations @xmath182, ... ,@xmath183 as described in example [ exam - rel2 ] and @xmath184 as described in example [ exam - rel3 ] is a complete system of relations for @xmath69 .
let @xmath185 be the free group in @xmath134 generators , for instance , the fundamental group of the complement to the union of @xmath186 concurrent lines . according to propositions [ prop - m1 ] and [ prop - m2 ]
, @xmath69 has a presentation matrix @xmath187 of size @xmath188 whose columns correspond to the generators @xmath189 $ ] and whose rows correspond to the coefficients of the jacobian relations @xmath190 , @xmath191 .
for instance , if @xmath192 @xmath193.\ ] ] such matrices have rank @xmath194 if @xmath195 for all @xmath196 , and hence the depth of a non - coordinate character is @xmath197 . on the other hand , for the trivial character @xmath99
, the matrix @xmath198 has rank 0 and hence @xmath99 has depth @xmath199 ( see definitions [ def - fitting ] and [ def - charvar ] for details on the connection between the rank of @xmath198 and the depth of a character ) . as a general reference for orbifolds and orbifold fundamental groups one can use @xcite , see also @xcite .
a brief description of what will be used here follows .
[ def - global ] an _ orbicurve _ is a complex orbifold of dimension equal to one .
an orbicurve @xmath19 is called a _ global quotient _ if there exists a finite group @xmath46 acting effectively on a riemann surface @xmath14 such that @xmath19 is the quotient of @xmath14 by @xmath46 with the orbifold structure given by the stabilizers of the @xmath46-action on @xmath14 .
we may think of @xmath19 as a riemann surface with a finite number of points @xmath200 labeled with positive integers @xmath201 ( for global quotients those are the orders of stabilizers of action of @xmath46 on @xmath14 ) .
a neighborhood of a point @xmath202 with @xmath203 is the quotient of a disk ( centered at @xmath204 ) by an action of the cyclic group of order @xmath205 ( a rotation ) . a small loop around @xmath204 is considered to be trivial in @xmath19 if its lifting in the above quotient map bounds a disk .
following this idea , orbifold fundamental groups can be defined as follows .
[ dfn - group - orb](cf .
@xcite ) consider an orbifold @xmath19 as above , then the _ orbifold fundamental group _ of @xmath19 is @xmath206 where @xmath207 is a meridian of @xmath208 and @xmath209 . according to remark [ charvargroups ] the definition [ def - charvar ]
can be applied to the case of finitely generated groups . in particular
one defines the @xmath76-th characteristic variety @xmath210 of an orbicurve @xmath19 as @xmath211 .
therefore also the concepts of a character @xmath18 on @xmath19 and its depth are well defined .
[ exam - orbicurve ] let us denote by @xmath212 an orbicurve for which the underlying riemann surface is @xmath32 with @xmath76 points removed and @xmath213 labeled points with labels @xmath214 . if @xmath215 ( resp .
@xmath216 ) we also use the notation @xmath217 ( resp .
@xmath218 ) for @xmath212 .
we suppress specification of actual points on @xmath32 .
note that @xmath219 ( here @xmath220 denotes a cyclic group of order @xmath221 with a generator @xmath222 ) .
note that a global quotient orbifold of @xmath223 by the cyclic action of order @xmath22 on @xmath32 that fixes two points , that is , @xmath224\mapsto [ \xi_nx : y]$ ] ( which fixes @xmath225 $ ] and @xmath226 $ ] ) is @xmath227 .
interesting examples of elliptic global quotients occur for @xmath228 , @xmath229 , and @xmath230 , which are global orbifolds of elliptic curves @xmath231 , @xmath232 , and @xmath233 respectively , see @xcite for a study of the relationship between these orbifolds ( @xmath234 ) and the depth of characters of fundamental groups of the complements to plane singular curves . a _ marking _ on an orbicurve @xmath19 ( resp .
a quasiprojective variety @xmath0 ) is a non - trivial character of its orbifold fundamental group ( resp .
its fundamental group ) of positive depth @xmath76 , that is , an element of @xmath235 ( resp .
@xmath236 ) which is in @xmath237 ( resp .
@xmath84 ) . a _ marked orbicurve _ is a pair @xmath238 , where @xmath19 is an orbicurve and @xmath20 is a marking on @xmath19 .
analogously , one defines a _ marked quasi - projective manifold _ as a pair @xmath15 consisting of a quasi - projective manifold @xmath6 and a marking on it .
a marked orbicurve @xmath238 is _ a global quotient _ if @xmath19 is a global quotient of @xmath14 , where @xmath14 is a branched cover of @xmath19 associated with the unbranched cover of @xmath239 corresponding to the kernel of @xmath240 . in other words ,
the covering space in definition [ def - global ] corresponds to the kernel of @xmath20 .
let @xmath6 be a quasi - projective variety , @xmath14 be a quasi - projective curve , and @xmath19 an orbicurve which is a global quotient of @xmath14 .
global quotient orbifold pencil _ is a map @xmath241 such that there exists @xmath242 where @xmath243 is a quasi - projective manifold endowed with an action of the group @xmath46 making the following diagram commute : @xmath244 the vertical arrows in are the quotients by the action of @xmath46 .
if , in addition , @xmath15 and @xmath238 are marked , then the global quotient orbifold pencil @xmath245 called _ marked _ if @xmath246 .
we will refer to the map of pairs @xmath247 as a _ marked global quotient orbifold pencil _ on @xmath15 with target @xmath238 .
[ def - indep ] global quotient orbifold pencils @xmath248 , @xmath249 are called _ independent _ if the induced maps @xmath250 define @xmath251$]-independent morphisms of modules @xmath252 that is , independent elements of the @xmath251$]-module @xmath253}(h_1(x_g,{\mathbb z}),h_1(c,{\mathbb z}))$ ] . in addition , if @xmath254 is surjective we say that the pencils @xmath255 are _ strongly independent_. note that if either @xmath256 or @xmath257 $ ] , then independence is equivalent to strong independence ( this is the case for remark [ rem - main - thm][rem - non - orbifold ] and theorem [ thm - main][thm - main - part2 ] ) .
the following are relevant improvements or additions to theorem [ arapurath ] : the isolated zero - dimensional characters of @xmath128 are torsion characters of @xmath258 . in ( * ? ? ?
* theorem 3.9 ) ( see also @xcite ) there is a description of one - dimensional components @xmath259 mentioned in theorem [ arapurath ] and most importantly , of the order of @xmath18 in terms of multiple fibers of the rational pencil @xmath260 . in @xcite
, an algebraic method is described to detect the irregularity of abelian covers of @xmath2 ramified along @xmath1 .
this method is very useful to compute _ non - coordinate components _ of @xmath128 independently of a presentation of the fundamental group of the complement @xmath6 of @xmath1 .
theorem [ thm - main ] ( see @xcite ) has ( * ? ? ?
* theorem 3.9 ) as a consequence , but uses the point of view of orbifold pencils . using this result
also the zero - dimensional components can be detected ( in particular essential coordinate components ) and in some cases characterized ( see section [ sec - ceva ] ) .
another improvement of theorem [ arapurath ] was given in @xcite were the point of view of orbifolds was first introduced as follows : [ thmprin ] let @xmath6 be a smooth quasi - projective variety .
let @xmath131 be an irreducible component of @xmath84 .
then one of the two following statements holds : 1 .
[ thmprin - orb ] there exists an orbicurve @xmath19 , a surjective orbifold morphism @xmath261 and an irreducible component @xmath117 of @xmath262 such that @xmath263 .
2 . [ thmprin - tors ]
@xmath131 is an isolated torsion point not of type [ thmprin - orb ] .
one has the following consequences from [ thm - main][thm - main - part2 ] that allows us to characterize certain elements of @xmath128 : [ rem - main - thm ] let @xmath15 be a marked complement of @xmath1 .
then possible targets for marked orbifold pencils are @xmath238 with @xmath264 ( see example [ exam - orbicurve ] ) .
assume that there are @xmath22 strongly independent marked orbifold pencils with such a fixed target @xmath238 .
then , 1 . [ rem - non - orbifold ] in case @xmath19 has no orbifold points , that is @xmath265 , the character @xmath18 belongs to a positive dimensional component @xmath131 of @xmath5 containing the trivial character . in this case ,
[ rem - torsion-2 ] in case @xmath18 is a character of order two , there is a unique marking on @xmath267 and @xmath12 is the maximal number of strongly independent orbifold pencils with target @xmath19 .
[ rem - elliptic ] in case @xmath18 has torsion 3,4 , or 6 , there is a unique marking on @xmath268 , @xmath269 , or @xmath270 respectively and @xmath12 is the maximal number of strongly independent orbifold pencils with target @xmath19 .
part [ rem - non - orbifold ] is a direct consequence of theorem [ arapurath ] and part [ rem - elliptic ] had already appeared in the context of alexander polynomials in @xcite . in section [ sec - ceva ]
we will describe in detail examples of corollary [ rem - main - thm][rem - torsion-2 ] for line arrangements
. we will give a very brief introduction to zariski pairs .
for more details we refer to @xcite and the bibliography therein . two plane algebraic curves @xmath1 and @xmath125 form a _
zariski pair _ if there are homeomorphic tubular neighborhoods of @xmath1 and @xmath125 , but the pairs @xmath271 and @xmath271 are not homeomorphic .
the first example of a zariski pair was given by zariski @xcite , who showed that the fundamental group of the complement to an irreducible sextic ( a curve of degree six ) with six cusps on a conic is isomorphic to @xmath272 whereas the fundamental group of any other sextic with six cusps is @xmath273 .
this paved the way for intensive research aimed to understand the connection between the topology of @xmath271 and the position of the singularities of @xmath1 ( whether algebraically , geometrically , combinatorially ... ) .
this research has been often in the direction of a search for finer invariants of @xmath271 .
characteristic varieties ( described above ) and the alexander polynomials ( i.e. the one variable version of the characteristic varieties ) , twisted polynomials @xcite , generalized alexander polynomials @xcite , dihedral covers of @xmath1 ( @xcite ) among many others are examples of such invariants .
if the alexander polynomials @xmath274 and @xmath275 coincide , then we say @xmath1 and @xmath125 form an _ alexander - equivalent zariski pair_. in section [ sec - zariski - pair ] we will use theorem [ thm - main ] to give an alternative proof that the curves in @xcite alexander - equivalent zariski pair , without computing the fundamental group .
consider the following family of plane curves : @xmath276 we will study the characteristic varieties of the quasi - projective manifolds @xmath277 , where @xmath278 , in light of the results given in the previous sections , in particular the essential torsion characters will be considered and their depth will be exhibited as the number of strictly independent orbifold pencils .
note that @xmath279 is nothing but the preimage by the kummer cover @xmath280{\smash{\mathop{\ \mapsto\ } \limits^{\kappa_n } } } [ x_0^n : x_1^n : x_2^n]$ ] of the following arrangement of three lines in general position given by the equation @xmath281 such a map ramifies along @xmath282 .
we will compute the fundamental group of @xmath283 as a quotient of the subgroup @xmath284 of @xmath285 associated with the kummer cover , where @xmath286 is a ceva arrangement .
more precisely , the quotient is obtained as a factor of @xmath284 by the normal subgroup generated by the meridians of the ramification locus @xmath287 in @xmath283 .
the fundamental group of the complement to the ceva arrangement @xmath288 is given by the following presentation of @xmath46 .
@xmath289=[e_3,e_5,e_1]=[e_3,e_4]=[e_5,e_2,e_4]=e_4e_3e_5e_2e_1e_0=1 \rangle\ ] ] where @xmath290 is a meridian of the component appearing in the @xmath291-th place in , @xmath292 $ ] denotes the commutator @xmath293 , and @xmath294 $ ] denotes the triple of commutators @xmath295 $ ] , @xmath296 $ ] , and @xmath297 $ ] leading to a triple of relations in .
in other to obtain one can use the _ non - generic _ zariski - van kampen method on figure [ fig - ceva ] ( see ( * ? ? ? * section 1.4 ) ) . the dotted line @xmath298 represents a generic line where the meridians @xmath299 are placed ( note that the last relation on is the relation in the fundamental group of @xmath300 ) .
the first two relations on appear when moving the generic line around @xmath301 .
the third and fourth relations come from moving the generic line around @xmath302 .
[ fig - ceva ] ( 0,0 ) ( 13,110)@xmath301 ( 43,110)@xmath298 ( 73,110)@xmath302 ( -5,60)@xmath303 ( -5,90)@xmath304 ( -5,20)@xmath305 the fundamental group of the complement to @xmath306 is equal to the kernel @xmath284 of the epimorphism @xmath307 since it is the fundamental group of the abelian cover with covering transformations @xmath308 .
therefore a presentation of the fundamental group of the complement to @xmath279 can be obtained by taking a factor of @xmath284 by the normal subgroup generated by @xmath309 , @xmath310 , and @xmath311 ( which are the meridians to the preimages of the lines @xmath312 , @xmath313 , and @xmath314 respectively ) . using the reidemeister - schreier method ( cf .
@xcite ) combined with the triviality of @xmath309 , @xmath310 , and @xmath311 one obtains the following presentation for @xmath315 : @xmath316}=1,\\ ( r6 ) & \prod_{k=0}^{n-1}e_{4,k , k}e_{3,k , k}e_{5,k , k}=1 \end{array } \ \rangle\ ] ] where @xmath317 and @xmath318 as a brief description of the reidemeister - schreier method , we recall that the generators of @xmath319 are obtained from a set - theoretical section of @xmath320 in ( in our case @xmath321 is given by @xmath322 ) as follows @xmath323 thus the set @xmath324 above forms a set of generators of @xmath319 .
finally a complete set of relations can be obtained by rewriting the relations of @xmath46 in ( and their conjugates by @xmath325 ) in terms of the generators of the subgroup @xmath319 . in order to illustrate the rewriting method we will proceed with the second relation of @xmath46 in .
@xmath326s(i , j)^{-1}=e_1^ie_2^j(e_3e_4e_3^{-1}e_4^{-1})e_2^{-j}e_1^{-i}=\\ ( e_1^ie_2^je_3e_2^{-j}e_1^{-i})\ ( e_1^ie_2^je_4e_2^{-j}e_1^{-i})\ ( e_1^ie_2^je_3^{-1}e_2^{-j}e_1^{-i})\ ( e_1^ie_2^je_4^{-1}e_2^{-j}e_1^{-i})= [ e_{3,i , j},e_{4,i , j } ] \endarray\ ] ] now we will discuss a presentation of @xmath327 as a module over @xmath328 , which will be referred to as @xmath329 . for details
we refer to section [ sec - alex - inv ] .
note that @xmath328 is isomorphic to @xmath330 and is generated by the cycles @xmath331 , @xmath332 , @xmath333 , ( @xmath317 ) where @xmath334 , @xmath335 , and @xmath336 satisfying @xmath337 is the morphism of abelianization ] .
let @xmath338 ( resp .
@xmath339 , @xmath340 ) be the generators of @xmath341 viewed as a multiplicative group corresponding to the additive generators @xmath331 ( resp .
@xmath332 , @xmath333 ) .
the characteristic varieties of @xmath319 are contained in @xmath342/(t_5^n\prod_j t_{3,j } \prod_i t_{4,i}-1).\ ] ] as generators of @xmath329 we select commutators of the generators of @xmath319 as given in . in order to do so , note that using relations @xmath343 in , a presentation of @xmath319 can be given in terms the @xmath344 generators @xmath345 , @xmath346 , and @xmath347 .
hence , by proposition [ prop - m1 ] , @xmath329 is generated by the @xmath348 commutators @xmath349,[e_{5},e_{4,i}],[e_{4,i},e_{3,j}],[e_{4,i_1},e_{4,i_2}],[e_{3,j_1},e_{3,j_2}]\}_{i_*,j_*\in { \mathbb z}_n},\ ] ] as a @xmath350$]-module . also , according to proposition [ prop - m2 ] , a complete set of relations of @xmath329
is given by rewriting the following relations @xmath351 & = & 0&\\ ( m2 ) & { [ } \prod_{i=0}^{n-1 } e_{5,i , j},e_{4,i } ] & = & 0&\\ ( m3 ) & { [ } e_{5,i , j+1},e_{3,i , j+1}e_{5,i , j+1}]e_{5,i , j+1}^{-1}&= & { [ } e_{5,i+1,j},e_{4,i+1,j}e_{5,i+1,j}]e_{5,i+1,j}^{-1}&\\ ( m4 ) & \prod_{i=0}^{n-1 } e_{4,i , i}e_{3,i , i}e_{5,i , i } & = & 0&\\ \endarray\ ] ] in terms of commutators and by the jacobian relations : @xmath352+(t_{4,i}-1)[e_{3,j},e_{5}]+(t_{5}-1)[e_{4,i},e_{3,j } ] & = & 0,\\ ( t_{3,j_1}-1)[e_{5},e_{3,j_2}]+(t_{3,j_2}-1)[e_{3,j_1},e_{5}]+(t_{5}-1)[e_{3,j_2},e_{3,j_1 } ] & = & 0,\\ ( t_{4,i_1}-1)[e_{5},e_{4,i_2}]+(t_{4,i_2}-1)[e_{4,i_1},e_{5}]+(t_{5}-1)[e_{4,i_2},e_{4,i_1 } ] & = & 0,\\ ... & \\ \endarray\ ] ] in order to rewrite relations @xmath353 one needs to use repeatedly . in what follows
, we will concentrate on the characters of @xmath354 contained in the coordinate axes @xmath355 .
computations for the general case can also be performed , but are more technical and tedious .
since we are assuming @xmath355 , and @xmath356 , relations in become @xmath357=[e_{3,j_2},e_{3,j_1}]=[e_{4,i_2},e_{4,i_1}]=0 $ ] and hence @xmath358 in become redundant .
a straightforward computation gives the following matrix where each line is a relation from written in terms of the commutators @xmath359,[e_5,e_{4,i}]\}_{i\in { \mathbb z}_n}$ ] .
@xmath360 } } \ ] ] more precisely , the first ( resp .
second ) block of @xmath198 corresponds to the @xmath22 relations given in @xmath361 ( resp .
@xmath362 ) of , @xmath363 , and @xmath364 .
the following @xmath22 blocks of @xmath198 ( between double horizontal lines ) correspond to the @xmath365 relations given in @xmath366 of .
note that the last row of each of these blocks is a consequence of the remaining @xmath367 rows .
the last block corresponds to the relation given in @xmath368 of . in order to illustrate
@xmath198 we will show how to rewrite the first relation for @xmath369 , that is , @xmath370\ { \smash{\mathop{=}\limits^{m}}}\ \phi_n [ e_5,e_{3,j}].\ ] ] using one has @xmath370\ { \smash{\mathop{=}\limits^{m}}}\ [ e_{5,0,j},e_{3,j}]+t[e_{5,1,j},e_{3,j}]+t^2[e_{5,0,j},e_{3,j}].\ ] ] therefore , it is enough to show that @xmath371=[e_5,e_{3,j}]$ ] . note that @xmath372 is a conjugate of @xmath373 ( using @xmath374 and @xmath374 ) , hence , by one obtains @xmath371=[e_5,e_{3,j}]$ ] ( since we are assuming @xmath375 ) .
also note that , performing row operations , one can obtain the following equivalent matrix @xmath376 } } \ ] ] finally , one can write the presentation matrix @xmath377 in terms of the basis @xmath378 - [ e_5,e_{3,i+1 } ] , [ e_5,e_{3,n-1 } ] , [ e_5,e_{4,i } ] - [ e_5,e_{4,i+1 } ] , [ e_5,e_{4,n-1 } ] \}_{i=0, ... ,n-2}\ ] ] resulting in @xmath379 } } \ ] ] one can use the units in the third block to eliminate columns , leaving the equivalent matrix @xmath380 } } \cong { \tiny { \left [ \begin{array}{c|ccc } 0 & -\phi_n & 0 & 0\\ 0 & 0 & -\phi_n & 0\\ \phi_n & 0 & 0 & 0\\ \hline 0 & 0 & 0 & \phi_n\\ \hline 0 & -1 & t^{n-2 } & 0\\ \end{array } \right ] . } } \ ] ] finally , a last combination of row operations using the units to eliminate columns results in @xmath381 } } \cong { \tiny { \left [ \begin{array}{c|cc } 0 & -\phi_n t^{n-2 } & 0\\ 0 & -\phi_n & 0\\ \phi_n & 0 & 0\\ \hline 0 & 0 & \phi_n \end{array } \right ] } } \cong { \tiny { \left [ \begin{array}{c|cc } 0 & \phi_n & 0\\ \phi_n & 0 & 0\\ \hline 0 & 0 & \phi_n \end{array } \right ] . } } \ ] ] hence the @xmath367 non - trivial torsion characters @xmath382 , @xmath249 belong to @xmath354 and have depth 3 , that is , @xmath383 . by theorem [ thm - main][thm - main - part1 ]
there are at most three strongly independent marked orbifold pencils from the marked variety @xmath384 .
our purpose is to explicitly show such three strongly independent pencils .
note that @xmath385),(\infty,[0:1]),(\infty,[1:1 ] ) } \\ & [ x : y : z ] & \mapsto & [ f_n : x_k^n ] , \endarray\ ] ] for @xmath386 are two natural orbifold pencils coming from the @xmath22-ordinary points of @xmath387 coming form the triple points of the ceva arrangement @xmath288 which are in @xmath388 .
consider the marked orbicurve @xmath389 , where @xmath390 , the first coordinate corresponds to the image of a meridian @xmath391 around @xmath225\in { \mathbb p}^1_{(n,[0:1]),(n,[1:0]),(\infty,[1:1])}$ ] and the second coordinate corresponds to the image of a meridian @xmath392 around @xmath226 $ ] ( note that @xmath393 ) . in order to obtain marked orbifold pencils with target @xmath389 one simply considers the following composition , where @xmath394 and @xmath395 are inclusions @xmath396),(\infty,[0:1]),(\infty,[1:1 ] ) } { \smash{\mathop{\ \hookrightarrow\ } \limits^{i } } } { \mathbb p}^1_{(n,[0:1]),(n,[1:0]),(\infty,[1:1])}.\ ] ] such pencils are clearly marked global quotient orbifold pencils from @xmath384 to @xmath389 , where @xmath389 is the marked quotient of @xmath397\}_{j\in { \mathbb z}_n}$ ] by the cyclic action @xmath224\mapsto [ \xi_nx : y]$ ] .
the resulting commutative diagrams are given by @xmath398 & \mapsto & [ w : x_k ] \\
{ \downarrow\rlap{$\vcenter{\hbox{$\scriptstyle\pi$}}$ } } & & \downarrow\\ { \mathcal x}_n & { \smash{\mathop{\ \rightarrow\ } \limits^{\psi_k } } } & { \mathbb c}_{n , n}\\ { [ } x_0:x_1:x_2 ] & \mapsto & [ f_n : x_k^n ] , \endarray\ ] ] @xmath399 , where @xmath400 is the smooth open surface given by @xmath401\in { \mathbb p}^3 \mid w^n = f_n\ } \setminus \{f_n\ell_1\ell_2=0\}$ ] .
note that there is a third quasitoric relation involving all components of @xmath279 , namely , @xmath402 and hence a global quotient marked orbifold map @xmath403),(n,[1:0]),(\infty,[1:1 ] ) } \\ & [ x : y : z ] & \mapsto & [ -f_nx_0^n : x_1^nx_2^n ] , \endarray\ ] ] which gives rise to the following diagram @xmath404 & \mapsto & [ -wx_0:x_1x_2 ] \\ { \downarrow\rlap{$\vcenter{\hbox{$\scriptstyle\pi_n$}}$ } } & & \downarrow\\ { \mathcal x}_n & { \smash{\mathop{\ \rightarrow\ } \limits^{\psi_k } } } & { \mathbb c}_{n , n}\\ { [ } x_0:x_1:x_2 ] & \mapsto & [ -f_nx_0^n : x_1^nx_2^n ] .
\endarray\ ] ] note that , when extending @xmath405 to a branched covering , the preimage of each line @xmath406 ( @xmath399 ) in @xmath279 ( @xmath407 ) decomposes into @xmath22 irreducible components @xmath408 and thus allows to consider @xmath409 ( @xmath399 , @xmath410 ) meridians around each component of @xmath411 . also consider a meridian @xmath412 around the preimage of @xmath387 .
the marked orbifold pencils @xmath413 , @xmath414 , and @xmath415 described above are strongly independent and hence they form a maximal set of strongly independent pencils
. consider @xmath416 $ ] , @xmath417 the three equivariant morphisms described above .
using the commutative diagrams and one can easily see that @xmath418 and @xmath419 and therefore @xmath420 are surjective @xmath421$]-module morphisms .
also note that @xmath422=\mu_n^j[\gamma_{k , i,0 } ] \in h_1(x_n;{\mathbb z})$ ] . consequently according to one has @xmath423 which implies that @xmath424 is surjective .
after the discussion of section [ sec - charvar ] , since the depth of @xmath425 is three , the set of strongly independent pencils is indeed maximal .
from theorem [ thm - main][thm - main - part2 ] , for any order two character @xmath18 of depth @xmath76 in the characteristic variety of the complement of a curve there exist @xmath76 independent pencils associated with @xmath18 whose target is a global quotient orbifold of type @xmath426 .
interesting examples for @xmath427 of this scenario are the arrangements @xmath428 , @xmath429 ( or _ erweiterte ceva _ cf .
* section 2.3.j , pg .
consider the following set of lines : @xmath430 the curve @xmath431 is a realization of the ceva arrangement @xmath432 ( a.k.a . braid arrangement or @xmath433-reflection arrangement ) .
note that this realization is different from the one used in section [ sec - fermat ] .
the curve @xmath434 is the arrangement @xmath435 ( a.k.a .
a realization of the non - fano plane ) .
the curve @xmath436 is the arrangement @xmath437 ( a.k.a .
a deleted @xmath433-arrangement ) . finally , @xmath438 is the arrangement @xmath439 . the characteristic varieties of such arrangements of lines are well known ( c.f @xcite ) .
such computations are done via a presentation of the fundamental group and using fox derivatives . in most cases
( except for the simplest ones ) the need of computer support is basically unavoidable . in (
* example 3.11 ) there is an alternative calculation of the positive dimensional components of depth 1 via pencils . here
we will give an interpretation via orbifold pencils of the characters of depth 2 , which will account for the appearance of these components of the characteristic varieties independently of computation of the fundamental group . note that @xmath440 and hence @xmath441 & \mapsto & [ \ell_1\ell_4:\ell_2\ell_5 ] \endarray\ ] ] is a pencil of conics such that @xmath442)=\ell_1\ell_4,f_c^{*}([1:0])=\ell_2\ell_5,f_c^{-1}([1:1])=\ell_3\ell_6)$ ] ( we will refer to it as the _ ceva pencil _ ) .
analogously @xmath443 and hence @xmath444 & \mapsto & [ \ell_1\ell_4\ell_7 ^ 2:\ell_2\ell_5\ell_8 ^ 2 ] \endarray\ ] ] is a pencil of quartics such that @xmath445)=\ell_1\ell_4\ell_7 ^ 2,f_{sc}^{*}([1:0])=\ell_2\ell_5\ell_8 ^ 2,f_{sc}^{*}([1:1])=\ell_3\ell_6\ell_9 ^ 2)$ ] ( we will refer to it as the _ pencil _ ) .
we include the structure of the characteristic varieties of these curves for the reader s convenience . as reference for such computations
see @xcite .
we will denote by @xmath448 the complement of the curve @xmath449 in @xmath2 , for @xmath450 . the characteristic variety @xmath452 consists of four non - essential coordinate components associated with the four triple points of @xmath451 ( see remark [ rem - main - thm][rem - non - orbifold ] ) and one essential component of dimension 2 and depth 1 given by the ceva pencil @xmath453,[1:0],[1:1]\}.\ ] ] the characteristic variety @xmath455 consists of six ( resp .
four ) non - essential coordinate components associated with the six triple points of @xmath454 ( resp .
four @xmath451-subarrangements ) of dimension 2 and depth 1 .
in addition , there is one extra character of order two , namely , @xmath456 of depth 2 .. similar notation will be used in the examples that follow . a second subscript ( when necessary )
will be used to index the characters considered . ] in order to check the value of the depth , one needs to find all marked orbifold pencils in @xmath457 of target @xmath458 where @xmath459 is the only possible non - trivial character of @xmath426 .
two such independent pencils are the following , @xmath460,[1:0],[1:1]\ } \to { \mathbb p}^1_{(2,[1:0]),(2,[1:1]),(\infty [ 0:1])}\ ] ] and @xmath461),(2,[1:1]),(\infty [ 0:1])}.\ ] ] this is the maximal number of independent pencils by theorem [ thm - main ] .
the characteristic variety @xmath463 consists of six ( resp .
five ) non - essential coordinate components associated with the six triple points of @xmath462 ( resp . four
@xmath451-subarrangements ) of dimension 2 and depth 1 .
in addition , there is one 3-dimensional non - essential coordinate component of depth 2 associated with its quadruple point ( see remark [ rem - main - thm][rem - non - orbifold ] ) .
consider the following pencil @xmath464),(\infty [ 0:1]),(\infty [ 1:0])}.\ ] ] computation of the induced map on the variety of characters shows that this map yields the only non - coordinate translated component of dimension 1 and depth 1 observed in the references above .
finally , there are two characters of order two , namely , @xmath465 of depth 2 . in order to check the value of the depth
, one needs to find two marked orbifold pencils on @xmath466 with target @xmath458 , where @xmath467),(2,[1:1]),(\infty [ 0:1])}\ ] ] and @xmath468 is the only non - trivial character of @xmath426 .
two such independent pencils can , for example , be given as follows @xmath469,[1:0],[1:1]\ } \to { \mathbb p}^1_{(2,[1:0]),(2,[1:1]),(\infty [ 0:1])}\ ] ] and @xmath470 ) } \setminus \{[1:0],[0:1]\ } \to { \mathbb p}^1_{(2,[1:0]),(2,[1:1]),(\infty [ 0:1])}.\ ] ] the characteristic variety @xmath472 consists of four ( resp . eleven ) non - essential coordinate components associated with the four triple points of @xmath471 ( resp .
eleven @xmath451-subarrangements ) , which have dimension 2 and depth 1 . in addition , there are three 3-dimensional non - essential coordinate components of depth 2 associated with the quadruple points of @xmath471 . consider the following pencil @xmath473,[0:1],[1:1]\}.\ ] ] computations of the induced map on the variety of characters show that this pencil yields the only non - coordinate translated component of dimension 2 and depth 1 observed in the references above .
finally , there are also three characters of order two @xmath474 of depth 2 . in order to check the value of the depth
, one needs to find two independent marked orbifold pencils on @xmath475 with target @xmath458 where @xmath476),(2,[1:0]),(\infty [ 1:1])}$ ] and @xmath468 is the only non - trivial character on @xmath426 .
two such independent pencils can be given , for example , as follows @xmath477,[1:0],[1:1]\ } \to { \mathbb p}^1_{(2,[0:1]),(2,[1:0]),(\infty [ 1:1])}\ ] ] and @xmath478,[1:0],[1:1]\ } \to { \mathbb p}^1_{(2,[0:1]),(2,[1:0]),(\infty [ 1:1])}.\ ] ] note that the depth @xmath479 characters in @xmath463 and @xmath472 lie in the intersection of positive dimensional components and this fact forces them to have depth greater than @xmath124 , see ( * ? ? ?
* proposition 5.9 ) . *
* depth conditions on the target : * first of all note that the condition on the target @xmath238 to have @xmath480 is essential in the discussion above , i.e. pencils with target satisfying @xmath481 may not contribute to the characteristic varieties .
for instance , the space @xmath482 also admits several global quotient pencils coming from the pencil , namely @xmath483),(2,[1:0]),(2,[1:1 ] ) } \to { \mathbb p}^1_{(2,[0:1]),(2,[1:0])}.\ ] ] however , the orbifold @xmath484 is a global quotient orbifold whose orbifold fundamental group is abelian , so no non - trivial characters belong to its characteristic variety . * * independence of pencils . *
here is an explicit argument for independence of pencils for one of the cases discussed in last section .
consider the pencils @xmath485 and @xmath486 described above as marked pencils from @xmath475 having @xmath458 as target .
the marking produces the following commutative diagrams : @xmath487 & \mapsto & [ \ell_1\ell_4:w ] \\ { \downarrow\rlap{$\vcenter{\hbox{$\scriptstyle\pi$}}$ } } & & { \downarrow\rlap{$\vcenter{\hbox{$\scriptstyle\tilde \pi$}}$}}\\ { \mathcal x}_9 & { \smash{\mathop{\ \rightarrow\ } \limits^{\psi_{9,2 } } } } & { \mathbb c}_{2,2}\\ { [ } x : y : z ] & \mapsto & [ \ell_1\ell_4:\ell_2\ell_5 ] , \endarray\ ] ] and @xmath488 & \mapsto & [ \ell_1\ell_4\ell_7:w\ell_8 ] \\ { \downarrow\rlap{$\vcenter{\hbox{$\scriptstyle\pi$}}$ } } & & { \downarrow\rlap{$\vcenter{\hbox{$\scriptstyle\tilde \pi$}}$}}\\ { \mathcal x}_9 & { \smash{\mathop{\ \rightarrow\ } \limits^{\psi_{9,3 } } } } & { \mathbb c}_{2,2}\\ { [ } x : y : z ] & \mapsto & [ \ell_1\ell_4\ell_7 ^ 2:\ell_2\ell_5\ell_8 ^ 2 ] , \endarray\ ] ] where @xmath489 is contained in @xmath490 \mid w^2=\ell_1\ell_4\ell_2\ell_5\}$ ] , @xmath491,[1:-1]\}$ ] and @xmath492 is given by @xmath493\mapsto [ u^2:v^2]$ ] .
+ consider @xmath494 , @xmath495 , @xmath399 the lifting of meridians around @xmath496 in @xmath489 . also denote by @xmath497 $ ] the ring of deck transformations of @xmath492 as before , where @xmath498 acts by multiplication by @xmath499 .
note that , as before @xmath500 and @xmath501 .
however , @xmath502 and @xmath503 . therefore @xmath485 and @xmath486 are independent pencils of @xmath475 with target @xmath458 .
consider the space @xmath504 of sextics with the following combinatorics : 1 .
@xmath19 is a union of a smooth conic @xmath505 and a quartic @xmath506 ; 2 .
@xmath507 where @xmath508 is a cusp of type @xmath509 and @xmath204 is a node of type @xmath510 ; 3 .
@xmath511 where @xmath508 is a @xmath512 on @xmath19 and @xmath68 is a @xmath513 on @xmath19 . in @xcite it is shown that @xmath504 has two connected components , say @xmath514 and @xmath515 .
the following are equations for curves in each connected component : @xmath516 for @xmath517 and @xmath518 for @xmath519 . the curves @xmath520 and @xmath521 form a zariski pair since their fundamental groups are not isomorphic .
this can not be detected by alexander polynomials since both are trivial . in @xcite the existence of an essential coordinate character of order two in the characteristic variety of @xmath521
was shown enough to distinguish both fundamental groups , since the characteristic variety of @xmath520 is trivial . by theorem [ thm - main][thm - main - part2 ]
this fact can also be obtained by looking at possible orbifold pencils .
note that there exists a conic @xmath522 passing through @xmath508 and @xmath68 such that @xmath523 , @xmath524 , and @xmath525 , @xmath525 .
consider @xmath526 the tangent line to @xmath527 at @xmath508 .
one has the following list of multiplicities of intersection : @xmath528 by @xcite , this implies that @xmath529 are members of a pencil of quartics . in other words , there is a marked orbifold pencil from @xmath530 marked with @xmath531 to @xmath532),(2,[1:0]),(\infty [ 1:1])}$ ] given by @xmath533\mapsto [ f_2^{(2)}\ell^2:q^2]$ ] whose target mark is the character @xmath468 .
e. artal , j. carmona , j.i .
cogolludo , and m. .
marco , _ invariants of combinatorial line arrangements and rybnikov s example _ , singularity theory and its applications , adv .
pure math .
43 , math .
japan , tokyo , 2006 , pp .
134 .
r. friedman and j. w. morgan , _ smooth four - manifolds and complex surfaces _ , ergebnisse der mathematik und ihrer grenzgebiete ( 3 ) [ results in mathematics and related areas ( 3 ) ] , vol .
27 , springer - verlag , berlin , 1994 .
suciu , _ fundamental groups of line arrangements : enumerative aspects _ , advances in algebraic geometry motivated by physics ( lowell , ma , 2000 ) , contemp . math . ,
276 , amer .
soc . , 2001 , pp .
4379 . ,
_ translated tori in the characteristic varieties of complex hyperplane arrangements _ , topology appl .
* 118 * ( 2002 ) , no . 1 - 2 , 209223 , arrangements in boston : a conference on hyperplane arrangements ( 1999 ) . | the present work is a user s guide to the results of @xcite , where a description of the space of characters of a quasi - projective variety was given in terms of global quotient orbifold pencils .
below we consider the case of plane curve complements and hyperplane arrangements .
in particular , an infinite family of curves exhibiting characters of any torsion and depth 3 will be discussed .
also , in the context of line arrangements , it will be shown how geometric tools , such as the existence of orbifold pencils , can replace the group theoretical computations via fundamental groups when studying characters of finite order , specially order two .
finally , we revisit an alexander - equivalent zariski pair considered in the literature and show how the existence of such pencils distinguishes both curves . |
the study of the in - medium properties of anti - kaon ( @xmath1 ) has attracted continuous attention @xcite .
@xmath1 has been theoretically expected to be bound by nuclear systems due to attractive strong interaction between @xmath1 and nucleon @xmath11 @xcite .
these @xmath1-nuclear bound systems assisted mainly by the strong interaction , which are so - called kaonic nuclei , have at least two important aspects ; the kaonic nuclei are strongly interacting exotic many - body systems , and they provide favorable sysmets for the studies of the @xmath1 properties at finite density .
however , in spite of experimental efforts to search for the @xmath1-nuclear bound systems @xcite , there are no clear evidences observed yet . in order to understand the @xmath1-nucleus systems ,
it is important to investigate theoretically the decay mechanism of the kaonic nuclei , or the absorption process of the @xmath1 into nuclei .
especially the decay ( absorption ) ratios of various modes and explicit numbers of the decay ( absorption ) widths are key quantities for the @xmath1-nucleus systems .
the decay of the kaonic nuclei in strong interactions can be categorized into two processes ; one is the mesonic process such as @xmath12 , and the other is the non - mesonic process such as @xmath13 , where @xmath14 denotes hyperon ( @xmath15 or @xmath16 ) .
the non - mesonic decays have advantage in experimental observations , since signals from kaonic nuclei are readily distinguished from backgrounds and no extra mesons do not have to be detected . therefore ,
systematic studies of the non - mesonic decay of the kaonic nuclei are desirable
. especially the ratios of the decay widths are interesting , since they will be insensitive to details of the production mechanism .
the absorption of the @xmath1 in nuclei may take place dominantly through the @xmath10 ( @xmath4 ) resonance , owning to the presence of the @xmath4 just below the @xmath17 threshold .
namely the @xmath4 can be a doorway of the @xmath1 absorptions in nuclei .
the @xmath4 doorway picture is more probable , in case that the @xmath4 is a quasi - bound state of @xmath17 @xcite , which has large @xmath17 components as almost real particles .
therefore the strong @xmath17 correlations are expected to be responsible for the @xmath4-induced decays of the @xmath1 in nuclei .
motivated by the @xmath4 doorway picture , we study the non - mesonic transition of @xmath18 ( @xmath19 and @xmath20 ) in uniform nuclear matter with a one - meson exchange approach .
now let us define the transition rate of the @xmath22 process by the transition probability divided by time @xmath23 as , _ yn _
d_2 |s - 1|^2 = _
d_2 |t_yn|^2 ( 2 ) ^4 ^4 ( p _ + p _ - p_n - p_y ) , [ eq : transition - rate ] with the @xmath24-matrix @xmath24 for the transition process given by the transition amplitudes @xmath25 as @xmath26 .
@xmath27 is volume of the system , and @xmath28 is the phase - space for the final state .
the transition rate depends on the center - of - mass energy @xmath29 , equivalently the initial nucleon momentum . summing up the @xmath22 transition rate in terms of the initial nucleon states
, we can estimate the non - mesonic decay width of the @xmath4 in nuclear medium as a function of the nucleon density .
the factor @xmath30 in the last form of eq .
( [ eq : transition - rate ] ) is responsible for the fact that only one @xmath11 exists in the initial state in the volume @xmath27 . [
cols="^,^,^,^ " , ] fixing the ratios of the @xmath4 coupling constants , @xmath31 , we plot the ratio of the non - mesonic decay widths of the in - medium @xmath1 to @xmath7 and @xmath32 as a function of the proton ( or neutron ) density in fig . [
fig : gvsrho](a ) .
this figure shows that the ratio of the non - mesonic decay widths @xmath33 is around @xmath34 almost independently of the nucleon density .
the density independence of the ratio of the decay widths is caused by sufficiently large phase - space in the final states .
we also obtain the absolute values of the non - mesonic decay widths of the @xmath4 in nuclear matter .
we use the couping constants of the @xmath4 to @xmath5 , @xmath6 and @xmath35 obtained by the chiral unitary approach , which gives the in - vacuum mesonic decay as @xmath36 with @xmath37 as observed in a @xmath38 initiated channel @xcite .
we show the non - mesonic decay widths in fig .
[ fig : gvsrho](b ) .
the linear dependence of the decay widths is caused by the large phase - space in the final states . at the normal nuclear density ( @xmath39 ) ,
the total non - mesonic decay width is @xmath40 , which is almost half of the mesonic decay width of @xmath4 ( @xmath41 ) .
we have investigated the non - mesonic transition @xmath42 in nuclear medium as one of the essential processes for the non - mesonic absorption of @xmath1 in nucleus . calculating the @xmath18 transition rates in the one - meson exchange processes
, we have found that the ratio of the @xmath43 transition rates to @xmath7 and @xmath44 strongly depends on the ratio of the @xmath4 couplings , @xmath45 . especially , larger @xmath4 couplings to @xmath17 lead to enhancement of the non - mesonic @xmath4 decay with @xmath7 emission .
furthermore , describing the @xmath4 properties by the chiral unitary approach , we have obtained the ratio of the non - mesonic decay widths to @xmath7 and @xmath46 as @xmath47 almost independently of the nucleon density .
we have also estimated that the total non - mesonic decay width is @xmath40 at the saturation density .
this study can be extend to the calculation of the absorption width of the @xmath48 in nuclear medium @xcite . c. j. batty , e. friedman and a. gal , phys .
* 287 * ( 1997 ) 385 ; t. waas and w. weise , nucl . phys . *
a625 * ( 1997 ) 287 ; a. ramos and e. oset , nucl .
a671 * ( 2000 ) 481 .
t. kishimoto , phys .
* 83 * ( 1999 ) 4701 ; e. friedman and a. gal , phys .
b * 459 * ( 1999 ) 43 ; y. akaishi and t. yamazaki , phys .
c * 65 * ( 2002 ) 044005 ; e. friedman and a. gal , phys .
* 452 * ( 2007 ) 89 .
t. kishimoto _ et al .
_ , prog .
* 149 * ( 2003 ) 264 ; m. agnello _ et al . _
[ finuda collaboration ] , phys .
* 94 * ( 2005 ) 212303 ; t. kishimoto _ et al .
_ , nucl .
a * 754 * ( 2005 ) 383 ; t. suzuki _ et al .
_ , nucl .
phys . a * 754 * ( 2005 ) 375 .
r. h. dalitz , t. c. wong and g. rajasekaran , phys .
* 153 * ( 1967 ) 1617 .
t. hyodo , d. jido and a. hosaka , phys .
c * 78 * ( 2008 ) 025203 .
t. sekihara , d. jido and y. kanada - enyo , phys .
c * 79 * ( 2009 ) 062201 .
e. oset and w. weise , nucl .
a * 319 * ( 1979 ) 477 .
d. jido , e. oset and j. e. palomar , nucl .
a * 694 * ( 2001 ) 525 .
n. kaiser , p. b. siegel and w. weise , nucl .
a * 594 * ( 1995 ) 325 ; e. oset and a. ramos , nucl .
a * 635 * ( 1998 ) 99 ; e. oset , a. ramos and c. bennhold , phys .
b * 527 * ( 2002 ) 99 ; d. jido , j. a. oller , e. oset , a. ramos and u. g. meissner , nucl .
a * 725 * ( 2003 ) 181 .
t. sekihara , t. hyodo and d. jido , phys . lett .
b * 669 * ( 2008 ) 133 .
o. braun _ et al .
_ , nucl . phys .
b * 129 * ( 1977 ) 1 .
d. jido , e. oset and t. sekihara , eur .
j. a * 42 * ( 2009 ) 257 . t. sekihara , j. yamagata - sekihara , d. jido and y. kanada - enyo , in preparation . | non - mesonic transition of @xmath0 is investigated as one of the essential processes for the non - mesonic absorption of @xmath1 in nuclei .
using one - meson exchange model in the calculation of the transition , we find that the non - mesonic transition ratio @xmath2 depends strongly on the ratio of the @xmath3 ( @xmath4 ) couplings to @xmath5 and @xmath6 .
especially a larger @xmath4-@xmath5 coupling leads to enhancement of the transition to @xmath7 . using the chiral unitary model for the description of the @xmath4
, we obtain @xmath8 which is almost independent of the nucleon density , and find the total non - mesonic decay width of the @xmath4 in uniform nuclear matter to be @xmath9 at the normal density . non - mesonic decay , kaonic nuclei ; @xmath10 doorway ; chiral unitary approach |
in a radioactive decay series , the end product is reached not only via the emission of @xmath3 and @xmath4 particles but also directly via heavy nuclei emission , the clusters , like carbon , oxygen , florine , neon , magnesium and silicon .
such a process is known as cluster decay , first proposed theoretically in 1980 by sndulescu , poenaru and greiner @xcite , before its experimental realization in 1984 by rose and jones @xcite . in this decay process a parent nucleus ( a , z ) , with mass number a and charge number z , breaks into two fragments , _ viz .
, _ the emitted cluster ( a@xmath5,z@xmath5 ) and the associated daughter ( a@xmath6,z@xmath6 ) , where a = a@xmath6+a@xmath5 , z = z@xmath6+z@xmath5 .
the light fragment ( a@xmath5,z@xmath5 ) is a cluster , heavier than the @xmath3-particle but lighter than the lightest fission fragment observed so far .
the detection of cluster - decay was hindered mainly due to large pile - up of @xmath3 particles , but with improved facilities these difficulties are removed and a large number of cluster - decays are observed from different radioactive nuclei . in the years that followed , the cluster decay process has been studied extensively using different theoretical models with different realistic nuclear interaction potentials . in general , two kinds of models are used for explaining the observed and/ or for predicting new decay modes . in one kind of these models ,
the @xmath3-particle as well as the heavy cluster(s ) were assumed to be pre - born in a parent nucleus , before they could penetrate the barrier with the available q - value .
these models are called the preformed cluster models @xcite . in such a model ,
the clusters of different sizes ( mass and/ or charge numbers ) are considered to be preformed in the parent nucleus , with different probabilities .
the gamow - like barrier penetration is also taken into account in these models . in the other kind of models , only the gamow s idea of barrier penetration
is used , without considering the cluster(s ) being or not being preformed in the parent nucleus . in other words ,
in this kind of models , called the unified fission models @xcite , the cluster radioactivity is considered as a simple barrier penetration phenomenon , in between the @xmath3-decay and the spontaneous fission . in this paper
, we attempt to give a model - independent , semiempirical formula for studying this above mentioned process of exotic cluster decay .
geiger and nuttal @xcite were the first who proposed a semiempirical law connecting the @xmath3-decay half - life and its q - value .
now , a large number of @xmath3-decays are observed from medium mass to superheavy nuclei , and several attempts have been made to give a universal formula @xcite .
these formulae vary among themselves mainly in the number of parameters .
one such scaling law , proposed recently by horoi _ et al _ @xcite , accounts for both the @xmath3 and cluster decays , and is given as , @xmath7+(a_2\mu^x+b_2 ) .
\label{eq:1}\ ] ] this scaling law involves six parameters @xmath8 , @xmath9 , @xmath10 , @xmath11 , @xmath12 and @xmath13 , which were obtained by fitting 119 @xmath3-decays and 11 cluster - decays from various even - even parents . in the following , we propose a new semiempirical formula based on only three parameters that are least squares fitted to the cluster - decay data alone .
so far , no attempt has been made to fit the @xmath3-decay data .
the evolution of our proposed formula is presented in section ii , followed by a brief description of the preformed cluster model ( pcm ) of gupta and collaborators @xcite in section iii .
the pcm is recently used by two of us and collaborators @xcite to calculate the cluster - decay half - lives of some deformed and superdeformed gd and nd parents . some of these results of ref .
@xcite are used here for comparisons with the calculations based on the proposed new semiempirical formula .
the results of our calculation are discussed in section iv , and the summary and conclusions are presented in section v.
we base the new formula for cluster decay half - lives on the following three simple experimental facts : + \(i ) it is known from experiments that the cluster decay half - lives increase with the size ( mass and/or charge ) of the clusters .
hence , the empirical formula should contain terms showing direct dependence on the mass number and charge number of the cluster .
\(ii ) the same cluster is emitted by different parents and hence the formula should contain dependence on the mass and charge asymmetries @xmath14 respectively .
\(iii ) since the @xmath3-decay and cluster - decay are physically similar processes , the q - dependence is taken to be the same as in geiger - nuttal law for @xmath3-decay , i.e. , @xmath15 . combining the above three results , we get @xmath16 where the constants @xmath17 , @xmath18 and @xmath19 are obtained by making a least squares fit of the available experimental half - lives for exotic cluster decays alone , with an rms deviation @xmath20=0.89(s ) , defined as @xmath21 ^ 2 $ ] , with @xmath22 denoting the function in eq .
( [ eq:3 ] ) , n the number of measurements and @xmath23 the experimentally observed values .
the @xmath24=@xmath25 , the variance , gives the standard deviation @xmath26 , with @xmath27 giving the arithmetic average of the expermentally measured quantities .
we refer to this formula as the _ az - formula _ ( _ azf _ ) .
note that the dependence on both @xmath28 and @xmath29 must be included in the formula ( [ eq:3 ] ) since these are separately measurable quantities , though the coupling between them is known to be weak @xcite , as is also evident when we consider the @xmath28 and @xmath29 dependence separately .
these special cases of az - formula , i.e. , @xmath30=0 or @xmath31=0 in ( [ eq:3 ] ) , respectively , are expected to give reasonably good results , though poorer than the azf results .
these truncated expressions are referred to as the _ af _ and _ zf _ formulae in the following , whose constants are also obtained directly by the least squares fit to data .
these are : @xmath32 and @xmath33 for af and @xmath34 and @xmath35 for zf , with rms deviations @xmath20=1.652 and 1.112 ( s ) , respectively .
apparently , the rms deviations for truncated expressions are larger , and hence the fits are poorer , compared to the total expression ( [ eq:3 ] )
. this will also be evident when we compare the results of these expressions with experimental data in section iv .
in the prefomed cluster model ( pcm ) of gupta and collaborators @xcite , the decay half - life is given in terms of three factors , the p@xmath36 , p and @xmath37 , as @xmath38 where @xmath39 is the preformation probability of the two fragments ( the cluster and daughter nuclei ) in their respective ground states , p the probability to tunnel the confining nuclear interaction barrier and @xmath37 as an assault frequency .
for calculating @xmath40 and p , gupta _ et al .
_ introduced the coupled motion in dynamical collective coordinates of mass asymmetry @xmath28 and relative separation r via a stationary schrdinger equation @xmath41 with the potential in it defined by the sum of experimental binding energies @xcite , the coulomb and the proximity @xcite potentials , as @xmath42 here the charges z@xmath43 are fixed by minimizing the potential ( without v@xmath44 ) in the charge asymmetry co - ordinate @xmath45 .
equation ( [ eq:5 ] ) is solved in the decoupled approximation , which gives @xmath46 @xmath47 only the ground state solution is relevant for the cluster decay to occur in the ground state of the daughter nucleus .
then , for @xmath28-motion , the properly normalized fractional preformation probability for , say , cluster a@xmath5 at a fixed r ( = r@xmath48=c@xmath49=c@xmath6+c@xmath5 , c@xmath43 being the sssmann central radii ) is @xmath50 for r - motion , the potential v(r ) is obtained from eq .
( [ eq:6 ] ) for fixed @xmath28 and , instead of solving the corresponding radial schrdinger equation , the wkb approximation is used for calculating the penetrability p. finally , the assault frequency @xmath37 in pcm is defined by considering that the total kinetic energy , shared between the two fragments , is the positive q - value : @xmath51 here @xmath52 is the spherical radius of the parent nucleus and @xmath53 , the nucleon mass .
figure 1 and table i give our calculated logarithms of decay half - lives by using the _ az - formula _ ( [ eq:3 ] ) for different clusters emitted from various radioactive parents , compared with the experimental data . in fig . 1
, we have also plotted the results of our calucations for _ af _ ( @xmath54 ) and _ zf _ ( @xmath55 ) versions of the _ az - formula_. it is evident that the azf fit to the data is better than the af and zf fits , as expected from our discussion above in section ii . in table
i , we have also added the experimental data on q - values and the results of another calculation using eq .
( [ eq:1 ] ) of horio et al . @xcite .
the comparison between the experiments and formulae ( [ eq:1 ] ) and ( [ eq:3 ] ) are also displayed in fig .
2 for the illustrative cases of @xmath56 and @xmath57 cluster decays ( respectively , the upper and lower panel ) . apparently , our semiempirical _
az - formula _ is much closer to experiments , as compared to the other formula due to horio et al .
@xcite .
finally , in order to compare the results of our semiempirical formula with the results of a model - dependent theoretical cluster decay calculation , we use a recent calculation @xcite of two of us ( mb and rkg ) and collaborators based on pcm for the emission of @xmath58 and @xmath56 clusters from @xmath59nd and @xmath60gd parents .
it may be mentioned here that these cluster - decay calculations are of interest only for nuclear structure information since their experimental observation may not be feasible in the near future @xcite .
3(a ) to ( d ) show the results of the pcm calculations , compared with those from the semiempirical _ az - formula _
( [ eq:3 ] ) and the scaling law ( [ eq:1 ] ) of horoi _ et al . _ @xcite .
we notice that , in general , the predictions of our semiempirical law lie higher than those of the pcm and horoi _
et al_. this is more so for nd parents than for gd parents where the predictions of the three calculations are nearly similar .
the interesting point is that the three calculations predict an exactly the same structure for @xmath61 vs. a , the parent mass number . for quantitative comparisons
, it may be reminded here that the constants a , b and c of our semiemprical formula are obtained by fitting the data from trans - actinide region , which may or may not be good for this trans - tin region .
in this paper , we have proposed a model - independent three parameter formula for calculating the half - lives of cluster decays of nuclei .
the evolution of the formula is based on three simple experimental observations about the characteristics of exotic cluster decays .
the inputs of the formula are simply the mass and charge numbers of the parent and cluster , along with the q - value of decay .
the predictions of the proposed formula are comparable to another model - independent scaling law as well as to the well known preformed cluster model .
the resulting good comparisons suggest that this formula could be used to make predictions for the guide of new experiments , similar to that of poenaru _ et al .
_ @xcite .
one of the authors ( m.b . ) acknowledges with thanks the partial financial support by the department of science and technology ( dst ) , vide grant no .
sr / ftp / psa-02/2002 . also , the support by dst under the fist programme vide letter no .
sr / fst / psi-005/2000 to the department of physics , m.s .
university , tirunelveli , india , is gratefully acknowledged .
1 the logarithms of decay half - lives for different clusters emitted from various radioactive parents , calculated by using _
az - formula _ ( azf ) and compared with experimental data . also , the results of calculations for af ( b=0 ) and zf ( a=0 ) truncations of azf are shown for comparisons .
the parents are labelled at the top x - axis and the corresponding clusters emitted by these parents are labelled at the bottom x - axis .
2 the experimental data on logarithms of decay half - lives for the emission of @xmath56 and @xmath62ne clusters from different radioactive parents , compared with the results of calculations using _ az - formula _ ( azf ) proposed here and the scaling law of ref .
@xcite .
fig.3 the logarithms of half - lives for the emission of @xmath63c clusters from different deformed and superdeformed nd and gd parents . our calculations , using _ az - formula _
( azf ) , are compared with the calculations based on pcm and the scaling law of horoi _ et al .
_ @xcite .
99 a. sndulescu , d.n .
poenaru and w. greiner , sovt .
* 11 * ( 1980 ) 528 .
h.j . rose and g.a .
jones , nature * 307 * ( 1984 ) 245 .
proceedings v@xmath64 international conference on nuclear reaction mechanisms _ , varenna , italy , 1988 , ed .
e. gadioli ( ricerca scientifica ed educazione permanente , milano , 1988 ) , p. 416 .
malik and r.k .
gupta , phys .
rev . * c 39 * ( 1989 ) 1992 .
s. kumar and r.k .
gupta , phys . rev .
* c 55 * ( 1997 ) 218 .
r. blendowske , t. fliessbach and h. walliser , nucl . phys . * a 464 * ( 1987 ) 75 .
r. blendowske and h. walliser , phys .
* 61 * ( 1988 ) 1930 .
y.j . shi and w.j .
swiatecki , phys .
* 54 * ( 1985 ) 300 .
pik - pichak , sovt .
. phys . * 44 * ( 1986 ) 923 .
g. shanmugam and b. kamalaharan , phys
* c 38 * ( 1988 ) 1377
. b. buck and a.c .
merchant , j. phys .
g : nucl . part .
* 15 * ( 1989 ) 615 .
a. sndulescu , r.k .
gupta , w. greiner , f. carstoiu , and m. horoi , int .
* e 1 * ( 1992 ) 374 .
g. royer , r.k .
gupta and v.yu .
denisov , nucl . phys . * a 632 * ( 1998 ) 275 . h. geiger and j.m .
nuttall , phil . mag . * 22 * ( 1911 ) 613 .
m. horoi , b.a .
brown and a. sndulescu , preprint : http:@xmath65arxiv.org/archive/nucl-th/9403008 a. brown , phys
* c 46 * ( 1992 ) 811 .
g. royer , j. phys .
g : nucl . part .
* 26 * ( 2000 ) 1149 .
gupta , s. dhaulta , r. kumar , m. balasubramaniam , g. mnzenberg and w. scheid , phys .
* c 68 * ( 2003 ) 034321 .
j. maruhn and w. greiner , phys .
* 32 * ( 1974 ) 548 .
gupta , w. scheid and w. greiner , phys .
* 35 * ( 1975 ) 353 .
gupta , in _
heavy elements and related new phenomena _ , edited w. greiner and r.k .
gupta , world sc .
singapore , 1999 , vol .
ii , p.730 .
g. audi and a.h .
wapstra , nucl . phys . * a 595 * ( 1995 ) 4 .
j. blocki , j. randrup , w.j .
swiatecki , and c.f .
tsang , ann .
( n.y ) * 105 * ( 1977 ) 427 .
r. bonetti , c. chiesa , a. guglielmetti , c. migliorino , p. monti , a.l .
pasinetti and h.l .
ravn , nucl .
phys . * a 576 * ( 1994 ) 21 .
price , j.d .
stevenson , s.w .
barwick and h.l .
ravn , phys .
* 54 * ( 1985 ) 297 .
e. hourani , l. rosier , g. berrier - ronsin , a. elayi , a.c .
mueller , g. rappenecker , g. rotbard , g. renou , a. libe and l. stab , phys . rev . * c 44 * ( 1991 ) 1424 .
r. bonetti , c. chiesa , a. guglielmetti , r. matheoud , c. migliorino , a.l .
pasinetti and h. l. ravn , nucl . phys . * a 562 * ( 1993 ) 32 .
e. hourani , m. hussonnois , l. stab , l. brillard , s. gales and j.p .
schapira , phys .
* b 160 * ( 1985 ) 375 .
r. bonetti , c. chiesa , a. guglielmetti , c. migiliorino , a. cesana and m. terrani , nucl .
phys . * a 556 * ( 1993 ) 115 .
p. qiangyan , y. weifan , y. shuanggui , l. zongwei , m. taotao , l. yixiao , k. dengming , q. jimin , l. zihua , z. mutian and w. shuhong , phys . rev . * c 62 * ( 2000 ) 044612 .
price , r. bonetti , a. guglielmetti , c. chiesa , r. matheoud , c. migiliorino and k.j .
moody , phys . rev . * c 46 * ( 1992 ) 1939 .
tretyakova , a. sndulescu , v.l .
mikheev , d. hasegan , i.a .
lebedev , yu.s .
zamyatnin , yu.s .
korotkin and b. f. myasoedov , jinr dubna rapid commun .
* 13 * ( 1985 ) 34 .
a. sndulescu , yu.s .
zamyatnin , i.a .
lebedev , b. f. myasoedov , s.p .
tretyakova and d. hasegan , jinr dubna rapid commun .
* 15 * ( 1984 ) 5 .
barwick , p.b .
price and j.d .
stevenson , phys .
rev . * c 31 * ( 1985 ) 1984 .
r. bonetti , c. chiesa , a. guglielmetti , c. migliorino , a. cesan , m. terrani and p. b. price , phys . rev . * c 44 * ( 1991 ) 888 .
moody , e.k .
hulet , s. wang and p. b. price , phys . rev .
* c 39 * ( 1989 ) 2445 .
tretyakova , yu.s .
zamyatnin , v.n .
kovantsev , yu .
s. korotkin , v. l. mikheev and g. a. timofeev , z. phys . * a 33 * ( 1989 ) 349 .
oglobin , n.z .
venikov , s.k .
lisin , s.v .
pirozhkov , v.a .
pchelin , yu.f .
rodionov , v.m .
semochkin , v.a .
shabrov , i.k .
shvetsov , v.m .
shubko , s.p .
tretyakova and v.l .
mikheev , phys . lett .
* b 235 * ( 1990 ) 35 .
tretyakova , v.l .
mikheev , v.a .
ponomarenko , a.n .
golovchenko , a.a .
oglobin and v.a .
shigin , jetp lett .
* 59 * ( 1994 ) 394 .
s. wang , d. snowden - ifft , p.b .
price , k. j. moody and e. k. hulet , phys .
rev . * c 39 * ( 1989 ) 1647 .
poenaru , d.schnabel , w. greiner , d. mazilu and r. gherghescu , at .
data nucl .
data tables * 48 * ( 1991 ) 231 . ccccccccparent & cluster & daughter & @xmath66 ( mev ) & @xmath67 ( s ) & @xmath68}$](s ) & @xmath69 ( s)&ref .
+ @xmath70 & @xmath56 & @xmath71 & 31.28 & 14.54 & 13.56 & 14.52 & @xcite + @xmath72 & @xmath56 & @xmath73 & 32.39 & 12.96 & 12.28 & 13.39 & @xcite + @xmath74 & @xmath56 & @xmath75 & 33.05 & 11.98 & 11.00 & 11.01 & @xcite + @xmath76 & @xmath56 & @xmath77 & 31.85 & 13.81 & 13.38 & 15.20 & @xcite + @xmath78 & @xmath56 & @xmath79 & 30.54 & 15.94 & 16.13 & 15.68 & @xcite + @xmath80 & @xmath56 & @xmath81 & 30.48 & 16.20 & 17.26 & 17.16 & @xcite + @xmath82 & @xmath56 & @xmath83 & 28.21 & 20.08 & 21.50 & 21.19 & @xcite + @xmath84 & @xmath85 & @xmath75 & 44.72 & 21.90 & 21.20 & 20.72 & @xcite + @xmath86 & @xmath87 & @xmath75 & 61.59 & 21.78 & 19.28 & 20.14 & @xcite + @xmath88 & @xmath89 & @xmath75 & 51.84 & 24.30 & 23.85 & 26.02 & @xcite + @xmath90 & @xmath57 & @xmath91 & 57.78 & 25.77 & 23.88 & 24.61 & @xcite + @xmath88 & @xmath57 & @xmath71 & 60.42 & 23.62 & 21.30 & 23.23 & @xcite + @xmath92 & @xmath57 & @xmath75 & 62.31 & 22.24 & 19.94 & 21.08 & @xcite + @xmath93 & @xmath57 & @xmath77 & 60.5 & 23.87 & 22.53 & 24.83 & @xcite + @xmath94 & @xmath57 & @xmath79 & 58.84 & 25.42 & 25.01 & 25.92 & @xcite + @xmath95 & @xmath57 & @xmath96 & 57.36 & 26.87 & 27.31 & 27.42 & @xcite + @xmath93 & @xmath97 & @xmath75 & 60.75 & 24.15 & 22.98 & 24.83 & @xcite + @xmath95 & @xmath97 & @xmath79 & 57.83 & 26.94 & 27.49 & 27.42 & @xcite + @xmath94 & @xmath98 & @xmath75 & 59.47 & 25.85 & 25.75 & 25.92 & @xcite + @xmath94 & @xmath99 & @xmath91 & 74.13 & 26.24 & 24.74 & 27.54 & @xcite + @xmath100 & @xmath99 & @xmath75 & 79.67 & 23.01 & 20.83 & 21.67 & @xcite + @xmath101 & @xmath99 & @xmath102 & 71.69 & 28.18 & 27.98 & 27.58 & @xcite + @xmath103 & @xmath99 & @xmath79 & 75.93 & 25.70 & 25.39 & 25.70 & @xcite + @xmath101 & @xmath104 & @xmath91 & 72.51 & 28.36 & 28.36 & 27.58 & @xcite + @xmath103 & @xmath104 & @xmath75 & 77.03 & 25.71 & 25.41 & 25.70 & @xcite + @xmath103 & @xmath105 & @xmath91 & 91.21 & 25.99 & 25.68 & 25.27 & @xcite + + | a new semiempirical formula , with only three parameters , is proposed for cluster decay half - lives .
the parameters of the formula are obtained by making a least squares fit to the available experimental data .
the calculated half - lives are compared with an earlier proposed model - independent scaling law . also , the calculated results of this formula are compared with the recent results of the preformed cluster model for @xmath0c and @xmath1c emissions from different deformed and superdeformed nd and gd parents .
the results are in good agreement with experiments as well as other models .
# 1@xmath2#1 |
measurements of the cosmic distance - redshift relation using supernovae @xcite , the cosmic microwave background @xcite , the hubble constant @xcite , and baryon acoustic oscillations ( bao ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have revealed that the expansion of the universe is accelerating ; either the energy density of the universe is dominated by an exotic `` dark energy '' , or general relativity requires modification .
the observed anisotropic clustering of galaxies can help distinguish between these possibilities by allowing simultaneous measurements of both the geometry of the universe and the growth rate of structure .
galaxy redshift surveys provide a powerful measurement of the growth rate through redshift - space distortions ( rsd ) @xcite .
although we expect the clustering of galaxies in real space to have no preferred direction , galaxy maps produced by estimating distances from redshifts obtained in spectroscopic surveys reveal an anisotropic galaxy distribution @xcite .
this anisotropy arises because the recession velocities of galaxies , from which distances are inferred , include components from both the hubble flow and from peculiar velocities driven by the clustering of matter ( see @xcite for a review ) .
despite the fact that galaxy light does not faithfully trace the mass , even on large scales , galaxies are expected to act nearly as test particles within the cosmological matter flow .
thus the motions of galaxies carry an imprint of the rate of growth of large - scale structure and allow us to both probe dark energy and test general relativity ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for recent studies ) .
the observed bao feature in the power - spectrum and correlation function of galaxies has been used to provide strong constraints on the geometry of the universe . while the bao method is expected to be highly robust to systematic uncertainties ( see , e.g. ,
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , it does not exploit the full information about the cosmological model available in the two - dimensional clustering of galaxies .
additional geometric information is available by comparing the amplitude of clustering along and perpendicular to the line - of - sight ( los ) ; this is known as the alcock - paczynski ( ap ) test @xcite .
rsd and ap tests rely on the measurements of anisotropy in the statistical properties of the galaxy distribution and are partially degenerate with each other , so that constraints on the growth of structure from rsd depend on the assumptions about the background geometry and vice versa @xcite .
however , given high precision clustering measurements over a wide range of scales , this degeneracy can be broken since rsd and ap have different scale - dependences .
recently , the wigglez survey @xcite has performed a joint rsd and ap analysis that constrains the expansion history in 4 bins across @xmath14 at the 10 - 15 per cent level @xcite .
using the sdss - ii lrg sample , @xcite perform a similar analysis to the present work .
they measure the angular diameter distance @xmath15 mpc and the hubble expansion rate @xmath16 km s@xmath4 mpc@xmath4 at @xmath17 after marginalizing over redshift space distortions and other parameter uncertainties . obtaining reliable cosmological constraints from the rsd and ap effects
demands precise modeling of the nonlinear evolution of both the matter density and velocity fields , as well as the ways in which the observed galaxies trace those fields .
the halo model of large - scale structure and variants thereof assume that galaxies form and evolve in the potential wells of dark matter halos , and provides a successful means of parametrizing the relation between the galaxy and halo density and velocity fields .
our modeling approach , based on @xcite , uses perturbation theory to account for the nonlinear redshift space clustering of halos in the quasilinear regime as a function of cosmological parameters , and then uses the halo model framework to motivate our choice of nuisance parameters describing the galaxy - halo relation .
we test these assumptions with a large volume of mock galaxy catalogs derived from @xmath18-body simulations .
the ongoing baryon oscillation spectroscopic survey ( boss ; * ? ? ?
* ) , which is part of sloan digital sky survey iii ( sdss - iii ) @xcite , is measuring spectroscopic redshifts of 1.5 million galaxies , approximately volume limited to @xmath19 ( in addition to spectra of 150,000 quasars and various ancillary observations ) .
the galaxies are selected from the multi - color sdss imaging to probe large - scale structure at intermediate redshift ; they trace a large cosmological volume while having high enough number density to ensure that shot - noise is not a dominant contributor to the clustering variance @xcite .
the resulting clustering measurements provide strong constraints on the parameters of standard cosmological models .
we use the cmass sample of boss galaxies that will be included in sdss data release 9 ( dr9 ) to constrain the growth of structure and geometry of the universe .
we apply rsd and ap tests to the data to measure the growth rate , the hubble expansion rate , and the comoving angular diameter distance at @xmath20 .
we improve on previous , similar analyses in a number of ways .
first , we use a model for the nonlinear anisotropic correlation function that is accurate to well below our statistical errors over the wide range of scales between @xmath21 and @xmath22 , which we validate using 68(@xmath23gpc)@xmath24 of @xmath18-body simulations populated with mock cmass galaxies .
next , rather than assuming a fixed underlying linear matter power spectrum , we use a prior on @xmath25 based on the wmap7 cosmic microwave background constraints @xcite and marginalize over the remaining uncertainties for all fits .
in addition to joint constraints on the geometric ( bao and ap ) and peculiar velocity ( rsd ) parameters , we present three `` null '' tests of the @xmath8cdm model . to begin , we simply ask whether any points in the @xmath8cdm parameter space allowed by wmap7 provide a good fit to the cmass clustering ; in this case , the only free parameters are those describing how galaxies trace matter . in the other two tests , we fit for the amplitude of peculiar velocities using wmap7 priors on geometric quantities , or we fit for the geometric parameters with wmap7 priors on rsd .
thus we can present the statistical precision with which our data measure either peculiar velocities or the ap effect in the @xmath8cdm model . finally , given the strong detection of the bao feature in the monopole of the correlation function
, we can break the degeneracy between @xmath26 and @xmath7 with our ap measurement .
we present the most constraining measurement of @xmath27 from galaxy clustering data to date using this technique , even after marginalizing over the amplitude of the rsd effect .
as our constraints exploit the full shape of the monopole and quadrupole correlation functions , they rely on further assumptions about the cosmological model : gaussian , adiabatic , power - law primordial perturbations , the standard number @xmath28 of massless neutrino species ( see discussion in * ? ? ?
* ) , and that dark matter is `` cold '' on the relatively large scales of interest . this paper is organized as follows . in sec .
[ sec : data ] we describe boss dr9 cmass data and in sec . [ sec : measurements ] we describe the measurements of two - dimensional clustering of galaxies used in this analysis .
[ sec : theory ] reviews the theory of the rsd and ap effects , and describes the theoretical model used to fit our measurements .
[ sec : analysis ] outlines the methods we use to analyse the data and sec .
[ sec : results ] presents the results of analysis .
we conclude by discussing the cosmological implications of our results in sec .
[ sec : discussion ] .
.,width=321 ] boss targets for spectroscopy luminous galaxies selected from the multi - color sdss imaging @xcite .
the target selection algorithms are summarized by @xcite and @xcite . for the galaxy sample
referred to as `` cmass '' , color - magnitude cuts are applied to select a roughly volume - limited sample of massive , luminous galaxies ; see @xcite for a detailed examination of the properties of cmass targets in the cosmos field . the majority of the galaxies are old stellar systems whose prominent 4000@xmath29 break in their spectral energy distributions makes them relatively easy to select in multi - color data .
most cmass galaxies are central galaxies residing in dark matter halos of @xmath30 , but a non - negligible fraction are satellites which live primarily in halos about 10 times more massive @xcite .
these galaxies are intrinsically very luminous and at the high mass end of the stellar mass function @xcite .
galaxies in the cmass sample are highly biased ( @xmath31 , * ? ? ?
in addition , they trace a large cosmological volume while having high enough number density to ensure that shot - noise is not a dominant contributor to the clustering variance , which makes them particularly powerful for probing statistical properties of large - scale structure . @xcite details the steps for generating the large - scale structure catalog and mask for dr9 , which includes the data taken by boss through july 2011 and covers 3275 @xmath32 of sky . in our analysis
we use galaxies from the boss cmass dr9 catalog in the redshift range of @xmath33 .
the sample includes a total of 264,283 galaxies , 207,246 in the north and 57,037 in the south galactic hemispheres .
figure [ fig : nz ] shows the redshift distribution and figure [ fig : mask ] shows the angular distribution of the galaxies in our sample .
to compute redshift space separations for each pair of galaxies given their angular coordinates and redshifts , we must adopt a cosmological model .
we use the same one as used to generate our mock catalogs , namely spatially - flat @xmath8cdm with @xmath34 .
this is the same cosmology as assumed in @xcite and in our companion papers @xcite .
our model accounts for this assumption and scales the theory prediction accordingly when testing a cosmological model with a different distance - redshift relation ; see sec .
[ apsec ] . using the galaxy catalog of @xcite we compute weighted `` data - data '' ( dd ) pair counts in bins of @xmath35 and @xmath36 @xmath37 where @xmath35 is the comoving pair separation in redshift space , @xmath36 is the cosine of the angle between the pair separation vector and the los and @xmath38 is the weight of @xmath39 galaxy in the catalog .
the double sum runs over all galaxies and @xmath40 if the pair separation between two galaxies falls into bin @xmath41 , and is zero otherwise .
three distinct effects contribute to the final weight @xmath42 of each galaxy .
these weights are described in more detail in @xcite and @xcite .
first , galaxies lacking a redshift due to fiber collisions or because their spectrum was not adequate to secure a redshift are accounted for by upweighting the nearest galaxy by weight @xmath43 , where @xmath44 is the number of near neighbors without a redshift .
second , we use the minimum variance weighting of @xcite , @xmath45 where @xmath46 is the expected number density of galaxies at the redshift of the galaxy and @xmath47 is the angularly averaged redshift - space correlation function integrated up to the separation of galaxies in the pair . for
every galaxy this weight will vary depending on which pair - counting bin it is assigned . for a constant radial selection function
this weighting scheme results in the minimal variance of the estimated correlation function ( for details see * ? ? ?
note that the analyses in @xcite and @xcite use scale - independent weights ; differences between the approaches are small in practice .
the third weight corrects for angular systematics , related to the angular variations in density of stars that make detection of galaxies harder in areas of sky closer to the galactic equator ( for details see * ? ? ?
the total weight is the product of these three weights .
we bin @xmath35 in 23 equal logarithmic bins between @xmath48 and @xmath49 with dlog@xmath50 , and 200 equally - spaced @xmath36 bins between 0 and 1 .
we compute `` data - random '' ( dr ) and `` random - random '' ( rr ) pair counts as in eqn . [ eq : dd ] , except that each random point is assigned only the @xmath51 weight and not the close - pair correction and systematic weights .
positions of objects in our random catalog are generated using the observational mask and redshifts are generated by picking a random redshift drawn from the measured redshifts of observed galaxies .
our random catalogs contain approximately 70 times more objects than the galaxy catalog .
following @xcite , the pair counts are combined to estimate the anisotropic correlation function as : @xmath52 figure [ fig : butterfly ] shows our measurement of @xmath53 in terms of los separation @xmath54 and transverse separation @xmath55 .
the central `` squashing '' is due to peculiar velocities . in the left panel ,
the bao ridge at @xmath56 @xmath23 mpc is evident . in the right panel , we show the clustering signal on smaller scales ; the `` finger - of - god '' effect is visible for small transverse separations but small on the scales we analyse .
the innermost contour in the right panel indicates the value of @xmath57 in the smallest separation bin included in our cosmological analysis . rather than work with the two - dimensional correlation function @xmath58
, we conduct our cosmological analysis on the first two even legendre polynomial moments , @xmath59 and @xmath60 , defined by @xmath61 or equivalently , @xmath62 here @xmath63 is the legendre polynomial of order @xmath64 . by symmetry
all odd-@xmath64 moments vanish and on large scales the measurements become increasingly noisy to larger @xmath64 .
the correlation functions @xmath65 and @xmath66 are estimated from @xmath53 using a riemann sum to approximate eqn .
[ eq : moments ] .
we include all galaxy pairs between 25 and @xmath67mpc in our analysis .
we also caution the reader that we have adopted logarithmically - spaced bins , while our companion papers @xcite analyze clustering in linearly - spaced bins of differing bin sizes .
our measurements of @xmath57 and @xmath68 , along with diagonal errors estimated from mock catalogs ( @xcite ; see sec .
[ sec : covmat ] ) are shown in figure [ fig : xilmeas ] .
the effective redshift of weighted pairs of galaxies in our sample is @xmath2 , with negligible scale dependence for the range of interest in this paper . for the purposes of constraining cosmological models
, we will interpret our measurements as being at @xmath2 .
the matrix describing the expected covariance of our measurements of @xmath69 in bins of redshift space separation depends in linear theory only on the underlying linear matter power spectrum , the bias of the galaxies , the shot - noise ( often assumed poisson ) and the geometry of the survey .
we use 600 mock galaxy catalogs , based on lagrangian perturbation theory ( lpt ) and described in detail in @xcite , to estimate the covariance matrix of our measurements .
we compute @xmath70 for each mock in exactly the same way as from the data ( sec .
[ sec : meas ] ) and estimate the covariance matrix as @xmath71 where @xmath72 is the monopole ( @xmath73 ) or quadrupole ( @xmath74 ) correlation function for pairs in the @xmath75 separation bin in the @xmath76 mock .
@xmath77 is the mean value over all 600 mocks .
the shape and amplitude of the average two - dimensional correlation function computed from the mocks is a good match to the measured correlation function of the cmass galaxies ; see @xcite and @xcite for more detailed comparisons .
the square roots of the diagonal elements of our covariance matrix are shown as the errorbars accompanying our measurements in fig .
[ fig : xilmeas ] .
we will examine the off - diagonal terms in the covariance matrix via the correlation matrix , or `` reduced covariance matrix '' , defined as @xmath78 where the division sign denotes a term by term division . in figure
[ fig : lptvslinear ] we compare selected slices of our mock covariance matrix ( points ) to a simplified prediction from linear theory ( solid lines ) that assumes a constant number density @xmath79 ( @xmath23 mpc)@xmath80 and neglects the effects of survey geometry ( see , e.g. , * ? ? ?
@xcite performed a detailed comparison of linear theory predictions with measurements from the las damas sdss - ii lrg mock catalogs @xcite , and showed that a modified version of the linear theory covariance with a few extra parameters provides a good description of the @xmath18-body based covariances for @xmath59 .
the same seems to be true here as well .
the mock catalogs show a deviation from the naive linear theory prediction for @xmath60 on small scales ; a direct consequence is that our errors on quantities dependent on the quadrupole are larger than a simple fisher analysis would indicate .
we verify that the same qualitative behavior is seen for the diagonal elements of the quadrupole covariance matrix in our smaller set of @xmath18-body simulations used to calibrate the model correlation function .
this comparison suggests that the lpt - based mocks are not underestimating the errors on @xmath68 , though more @xmath18-body simulations ( and an accounting of survey geometry ) would be required for a detailed check of the lpt - based mocks .
the lower panels of figure [ fig : lptvslinear ] compare the reduced covariance matrix to linear theory , where we have scaled the @xmath81 prediction from linear theory down by a constant , @xmath82 .
this comparison demonstrates that the scale dependences of the off - diagonal terms in the covariance matrix are described well by linear theory , but that the nonlinear evolution captured by the lpt mocks can be parametrized simply as an additional diagonal term . finally ,
while not shown here , the reduced covariances between @xmath57 and @xmath68 are small .
our analysis uses the lpt mock - based covariance matrix , which accurately accounts for both complexities of the survey geometry as well as nonlinear corrections to the growth of structure on the relatively large scales of interest here , and this allows us to accurately report uncertainties associated with both our measurements and parameter fits . and @xmath83 for separation bins of 33 and 103 @xmath23 mpc .
linear theory predictions for the reduced covariance matrices in the lower panels have been scaled by a constant factor to produce good agreement between linear theory and mock covariances for off - diagonal elements , demonstrating that the scale dependence of the off - diagonal terms matches the mock covariance matrix well , but that there is extra diagonal covariance in the mocks compared with linear theory .
elements of @xmath84 are small ( not shown).,width=321 ]
the effects of redshift space distortions in the linear regime are well - known ( * ? ? ?
* ; * ? ? ?
* see also @xcite for a comprehensive review ) .
we briefly summarize them here .
the redshift - space position , * s * , of a galaxy differs from its real - space position , * x * , due to its peculiar velocity , @xmath85 where @xmath86 is the change in the apparent los position of a galaxy due to the contribution of the los peculiar velocity @xmath87 to the galaxy s redshift .
since overdensities on large , linear scales grow in a converging velocity field ( @xmath88 ) , the effect of peculiar velocities induces a coherent distortion in the measured clustering of galaxies that allows us to measure the amplitude of the peculiar velocity field . in linear theory , and with some approximations , the anisotropic galaxy power spectrum becomes @xcite @xmath89 where @xmath90 is the linear galaxy bias , @xmath91 , @xmath92 is the logarithmic growth rate of matter fluctuations , and @xmath93 is the cosine of the angle between @xmath94 and the los . in linear theory ( eqn . [ eq : kaiser ] ) , only the @xmath96 and @xmath97 moments contribute to the power spectrum @xmath98 , and its fourier transform @xmath99 .
the two are simply related by @xmath100 order bessel functions : @xmath101 given a tight constraint on the underlying _ shape _ of the linear matter power spectrum , the two - dimensional clustering of galaxies constrains both @xmath102 and @xmath103 @xcite . in this work we measure and model only the monopole and quadrupole moments of the correlation function , @xmath104 .
these two moments are sufficient to constrain both @xmath102 and @xmath103 and encompass most of the available information on the peculiar velocity field for the highly biased galaxies of interest here @xcite , in addition to being more easily modelled than higher-@xmath64 moments ( but see also * ? ? ?
therefore , we collapse the anisotropic clustering information in @xmath105 into two one - dimensional functions @xmath104 before extracting cosmological information from the anisotropic galaxy clustering .
the kaiser formula describing the linear effect of redshift space distortions breaks down even on quite large scales .
an accurate model of the two - dimensional clustering of galaxies must account for nonlinear evolution in the real space matter density and velocity fields , nonlinear galaxy bias , and the nonlinear mapping between real and redshift space separations of pairs of galaxies .
the simplest picture of galaxy formation asserts that galaxies occupy dark matter halos , and so as a step towards understanding the clustering of galaxies , @xcite showed that a streaming model where the pairwise velocity probability distribution function is approximated as gaussian can be used to relate real space clustering and pairwise velocity statistics of halos to their clustering in redshift space .
we will demonstrate in sec .
[ halos2gals ] that the same model describes the clustering of galaxies : @xmath106 e^{-[r_\pi - y - \mu v_{12}(r)]^2/2\sigma_{12}^2(r,\mu ) }
\frac{dy}{\sqrt{2\pi\sigma^2_{12}(r,\mu ) } } \label{streamingeqn},\ ] ] where @xmath107 and @xmath108 are the redshift space transverse and los distances between two objects with respect to the observer , @xmath109 is the _ real _ space los pair separation , @xmath110 , @xmath111 is the real space galaxy correlation function , @xmath112 is the average infall velocity of galaxies separated by real - space distance @xmath113 , and @xmath114 is the rms dispersion of the pairwise velocity between two galaxies separated with transverse ( los ) real space separation @xmath115 ( @xmath109 ) .
@xmath116 , @xmath112 and @xmath114 are computed in the framework of lagrangian ( @xmath117 ) and standard perturbation theories ( @xmath118 , @xmath119 ) .
only two nuisance parameters are necessary to describe the clustering of a sample of halos or galaxies in this model : @xmath120 , the first - order lagrangian host halo bias in _ real _ space , and @xmath121 , an additive , isotropic velocity dispersion accounting for small - scale motions of halos and galaxies which will be described below .
further details of the model , its numerical implementation , and its accuracy can be found in @xcite and appendix [ cosmoxi2d ] .
@xcite examined the validity of eqn .
[ streamingeqn ] only for halo clustering rather than galaxies , thus the model must be extended and checked with a realistic sample of mock galaxies .
we use the machinery of the halo model ( see * ? ? ?
* for a review ) to describe the galaxy density field in terms of the density field of the host halos .
of particular importance for modeling redshift space distortions is the distinction between `` central '' and `` satellite '' galaxies @xcite . when modeling an approximately mass - limited galaxy sample , the first galaxy assigned to a host halo is considered central , and its position and velocity are that of the host halo center .
satellite galaxies orbit in the potential well of the host halos , and so are offset in both position and velocity from the halo center . in our particular implementation of the halo model , satellite galaxies
are randomly drawn from the dark matter particle members of the host halo in our simulation . as the virial velocities of massive halos can be large ( amounting to redshift space los separations of tens of @xmath23 mpc )
, intrahalo velocities can distort the redshift space correlation function . in the limit that these virial motions are uncorrelated with the quasilinear velocity field of interest
, they can be accounted for by additional convolution along the line of sight .
to assess the impact of virial motions on the observed galaxy clustering we use the mock catalogs described in @xcite , which closely match the small - scale clustering of cmass galaxies .
we compute @xmath104 from these mocks ( the average is shown as the error bars in figure [ fig : theoryfits ] ) and recompute @xmath122 after artificially setting the intrahalo velocities to 0 ( the dashed curves , with the one for @xmath57 covered by the solid line ) .
intrahalo velocities suppress the amplitude of @xmath68 on the smallest scales we attempt to model , reaching a 10 per cent correction at @xmath123mpc .
the reason for this suppression is that , on small scales , @xmath124 and non - negligible .
this causes a net transport of pairs to larger separations in redshift space .
note this is opposite to the effect of quasilinear peculiar velocities , which make the separation of a pair in redshift space on average smaller than in real space .
we include the effect of intrahalo velocities in our model by including an extra convolution with a gaussian of dispersion @xmath125 .
the solid curves in fig . [ fig : theoryfits ] are then the predictions for @xmath122 with the best fit @xmath126 or @xmath127 .
the model successfully describes the effect of intrahalo velocities on the monopole and quadrupole correlation functions . in appendix [ galaxydetails ]
we quantify the impact of our uncertainties in the halo occupation distribution ( hod ) of cmass galaxies and the possible breakdown of our assumption that the first galaxy assigned to each halo is `` central '' ( i.e. , has no intrahalo velocity dispersion ) on the value of our nuisance parameter @xmath121 . as a result of these considerations ,
we place a uniform prior on @xmath121 between 0 and 40 ( mpc)@xmath128 .
galaxy redshift surveys collect two angular coordinates and a redshift for each galaxy in the sample .
a fiducial cosmological model must be adopted to generate maps and measure clustering as a function of comoving separations .
this mapping depends on both the angular diameter distance and the inverse of the hubble parameter at the redshift of each galaxy pair . to a good approximation the inferred galaxy clustering in a different cosmological model
can be obtained from the fiducial one by a single rescaling of the transverse and parallel separations @xcite . rather than modify our observed galaxy clustering , we will account for the ap effect when we test different cosmological models by introducing two scale parameters , @xmath129 and @xmath130 , into the theoretical correlation functions we are fitting to : @xmath131 where @xmath6 denotes the physical angular diameter distance .
here @xmath132 is the expected two - dimensional correlation function _ if _ the measured galaxy correlation function were computed assuming the true redshift - distance relation in the cosmology being tested .
@xmath133 is the prediction for the measured correlation function , given that galaxy separations were computed using the fiducial cosmology model .
that is , @xmath129 and @xmath130 scale the ` true ' separations to the ones calculated using the fiducial cosmology .
the spherically averaged correlation function , @xmath57 , is sensitive to the parameter combination @xmath134 the quadrupole of the measured correlation function allows a measurement of a second combination @xcite @xmath135 thus breaking the degeneracy between @xmath136 and @xmath7 . to a good approximation , changing @xmath137 simply rescales the value of @xmath35 in the predicted correlation function , while @xmath138 primarily affects the quadrupole .
figure [ fig : theoryfits ] shows the effect of varying @xmath139 by @xmath140 per cent on @xmath68 at fixed @xmath137 with the dotted curves .
the scale - dependence of @xmath141 due to the ap effect will allow us to separate the effects of peculiar velocities and the ap effect .
given an underlying linear matter power spectrum shape @xmath142 , we consider the predicted galaxy clustering as a function of five parameters : @xmath143 .
since the normalization of @xmath144 ( denoted throughout as @xmath145 ) determines the amplitude of the second - order perturbation theory corrections in our model , in principle we should be able to separately determine @xmath102 , @xmath103 , and @xmath145 . in practice , the dependence is sufficiently small and degenerate with the nuisance parameters @xmath102 and @xmath121 that the degeneracy can not be broken ; see appendix [ sec : modelacc ] for details .
we have also assumed that any error in the fiducial cosmological model used to compute @xmath104 from the cmass galaxy catalog can be absorbed in a single scaling of distances , interpreted at the effective redshift of the survey .
if the assumed redshift dependence of @xmath146 and @xmath147 is grossly incorrect , we would expect a difference in the correlation functions split on redshift ; we see no evidence for this in our tests @xcite .
while the scale dependence of galaxy clustering itself can constrain the shape of the linear matter power spectrum , at present ( and certainly with the imminent public release of planck data ) the constraints enabled by cmb measurements are stronger .
the strong cmb constraints mean we can use the entire linear matter power spectrum as a standard ruler determined by observations of the cmb , rather than only the bao feature .
this approach relies on further cosmological model assumptions that are consistent with the current data , but from which moderate deviations are still allowed .
the temperature of the cmb has been measured exquisitely well @xcite , and determines the physical energy density in radiation , @xmath148 . in the minimal cosmological model allowed by current observations @xcite , namely a flat @xmath8cdm cosmology with nearly scale - invariant scalar , adiabatic , gaussian fluctuations along with the three standard ,
nearly massless neutrino species , only three additional parameters determine the shape of the underlying linear matter power spectrum , @xmath149 .
relative peak heights in the cmb determine the physical energy densities in cold and baryonic matter , @xmath150 , and the overall scale - dependence of the cmb power spectrum determines the spectral index @xmath151 of the nearly scale - invariant scalar primordial fluctuations .
constraints on these parameters do not rely on the distance to the cosmic microwave background , and thus are immune to the behavior of dark energy at lower redshifts than the last scattering surface .
moreover , small - scale cmb experiments @xcite now probe fluctuations on the same scales as galaxy clustering measurements , and find no compelling evidence for , e.g. , a running of the spectral index . allowing for running of the spectral index would degrade the cmb constraints on the linear matter power spectrum ( e.g. , * ? ? ?
* ) but we will not include this additional parameter obviously , our constraints should be interpreted in the context of our model assumptions .
one important extension of the minimal cosmological model is allowing neutrinos to have mass ; neutrino oscillation experiments suggest that @xmath152ev @xcite .
as the universe expands and cools , massive neutrinos become non - relativistic and modify the linear matter power spectrum inferred from the cmb , as well as alter the expansion history as compared with the massless neutrino case . in this work
we ignore this additional uncertainty in the shape of @xmath149 .
current upper bounds that combine several cosmological probes find @xmath153 ev ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , which is safely below the detectable level in the dr9 cmass sample @xcite .
as we explore the cosmological constraints available from our data set , we will consider a set of power - spectrum shapes parametrized by @xmath154 .
we will marginalize over parameters @xmath155 either by importance resampling @xcite the public monte carlo markov chains for a spatially flat @xmath8cdm model provided by the wmap collaboration , or by approximating the constraints on @xmath155 with a three - dimensional gaussian likelihood .
note that the three - dimensional constraints on linear matter power spectrum parameters are very nearly independent of cosmological model extensions that change the expansion rate only for @xmath156 ( i.e. , introducing @xmath157 or @xmath158 ) . in the present paper ,
we interpret the anisotropic clustering of cmass galaxies only in the context of the @xmath8cdm cosmology . by relaxing assumptions about the redshift - distance relation and/or the growth of structure in the @xmath8cdm model ,
we report the statistical precision with which cmass measurements constrain both the peculiar velocity field @xmath159 and geometric quantities @xmath160 and @xmath27 , without adopting a particular cosmological model that specifies how these quantities are related .
we consider the following four models : * * model 1 : wmap7+cmass flat @xmath8cdm .
* in the flat @xmath8cdm cosmology , @xmath159 , @xmath160 and @xmath27 are all determined once @xmath161 , @xmath145 , and @xmath162 are specified .
this provides a null test of our assumptions relating cmb fluctuations to predicted galaxy fluctuations . *
* model 2 wmap7+cmass @xmath8cdm geometry , free growth .
* the analysis for this model is the same as model 1 , except that we consider @xmath103 a free parameter in the cmass galaxy clustering fits . *
* model 3 wmap7+cmass @xmath8cdm growth , free geometry * the analysis for this model is the same as model 1 , except that we consider @xmath160 and @xmath27 as free parameters in the cmass galaxy clustering fits . * * model 4 wmap7+cmass , free growth , free geometry * : in this model , @xmath159 , @xmath160 and @xmath27 are all free parameters in the fit to the galaxy clustering data .
this allows us to determine how well our data break the degeneracy between the rsd and ap effects , and to present the constraints originating from the amplitude and scale - dependence of the galaxy quadrupole @xmath60 in the most model independent way possible .
this multivariate distribution can be used to constrain any model that does not alter the shape of the linear matter power spectrum at @xmath163 from that inferred from the cmb on the relatively large scales of interest here . in all four scenarios
we allow the value of @xmath121 to vary between @xmath164 and @xmath165 with a flat prior and marginalize over both @xmath121 and @xmath102 when deriving final results .
we detail our methods for sampling the multi - dimensional probability distribution functions of interest in models 1 through 4 in appendix [ sec : mcmcmethods ] .
this approach to parameter fitting allows our estimates of the growth of structure and geometry to be independent of many model assumptions and can be used to put constraints on more general models of gravity and dark energy .
however , they still rely on the standard model in three ways .
first , we assume that the processes in the early universe that were responsible for setting up the linear matter power spectrum at recombination do not change significantly , which is true for the most popular models of modified gravity and dark energy .
second , we assume that growth at the level of linear perturbation theory is scale independent between the cmb epoch and the effective redshift of our sample .
third , we use gr to compute the perturbation theory corrections to the galaxy clustering predictions .
the perturbation theory corrections are not large , and they are most important on small scales where @xmath121 also becomes important ; therefore we can not strongly constrain the amplitude of the higher - order corrections .
constraints on models with scale - dependent growth should be derived directly from the correlation function measurements and their covariance .
we follow the standard convention of denoting the amplitude of the matter power spectrum by @xmath166 , even though we restrict our analysis to scales @xmath167 @xmath23 mpc , so a different weighted integral over @xmath25 that is concentrated on larger scales would more accurately reflect our constraints on the growth rate of matter fluctuations . in particular , since @xmath25 is well - determined for @xmath168 in mpc@xmath4 ( e.g. * ? ? ?
* ) , and the bao scale provides a standard ruler with percent - level precision , our data constrain the amplitude of matter fluctuations on scales @xmath169mpc . in practice , the tight constraints on the shape of @xmath25 means that differences arising from how one specifies its amplitude are small when computing parameter constraints , as long as @xmath170 is well - determined in the model . in model 2 , we sample power spectra from the wmap @xmath8cdm chain , and take the traditional value of @xmath145 to relate the model parameter @xmath171 and the reported constraint @xmath103 . in model 4 , we do not specify a value of @xmath170 with each sampled power spectrum , so we normalize the power spectra by fixing @xmath172 , where @xmath173mpc . for power spectra drawn from wmap7 @xmath8cdm chains , @xmath174 ; the offset and variance between these parameters is negligible compared to our measurement errors on @xmath103 .
[ cols="<,<,<,<,<,<,<",options="header " , ] at fixed @xmath25 and geometric parameters , the fisher matrix analysis indicates a factor of 2 increase in the @xmath103 error , to @xmath175 , when @xmath121 is marginalized over without any prior , compared to when it is fixed at @xmath176 .
the marginalized error on @xmath121 in the former case is @xmath177 .
therefore , the hard prior @xmath178 substantially reduces this source of uncertainty .
table [ table : uncertainties ] indicates an increase of only 15 per cent in the error on @xmath103 , to 0.033 , when we marginalize over @xmath121 within our hard prior .
similarly , marginalizing over @xmath121 increases the error on @xmath139 from 0.019 to 0.022 .
therefore , further reduction in the uncertainty on @xmath121 with more detailed modeling of the small - scale clustering would only allow a slight reduction in the errors .
however , since our fits indicate a slight preference for @xmath179 compared to the fiducial @xmath127 , in future work we will revisit our choice of prior after a re - analysis of small - scale cmass clustering with a larger data set . of the three parameters defining the shape of the underlying linear matter @xmath25 in our analysis ,
the uncertainty in @xmath180 is the largest , particularly in the context of the constraining power of the cmass measurements . for this comparison
, we therefore hold @xmath181 and @xmath182 fixed , and examine how the central values and uncertainties in @xmath102 , @xmath103 or @xmath139 change as a function of @xmath183 , with all other parameters held fixed .
empirically , the uncertainties on these parameters do not depend much on @xmath183 .
therefore , we can estimate the impact on the uncertainty in @xmath183 through the dependence of the central value of parameter @xmath184 on @xmath183 , in units of the uncertainty on those quantities : @xmath185 where @xmath186 is the uncertainty at fixed @xmath183 . for gaussian probability distribution functions in @xmath187 and @xmath188 , the uncertainty on @xmath184 when marginalized over @xmath183
is increased by @xmath189 .
we find that @xmath190 for @xmath103 and @xmath191 for @xmath139 .
therefore , the current uncertainty in the underlying linear matter @xmath25 is negligible for the purpose of deriving constraints from the anisotropic clustering of cmass galaxies .
this justifies the use of a fixed power spectrum shape in the wigglez analyses of anisotropic clustering @xcite .
however , the uncertainty in @xmath25 does increase the uncertainty in @xmath102 , so applications such as galaxy - galaxy lensing / galaxy clustering combinations ( e.g. , * ? ? ?
* ) should marginalize over this additional uncertainty .
table [ table : uncertainties ] indicates that marginalizing over the uncertainty in the geometric quantities @xmath192 and @xmath27 [ or equivalently , @xmath193 and @xmath194 contributes negligibly to the uncertainty in @xmath103 for wmap7 uncertainties when a @xmath8cdm redshift - distance relation is assumed . in the @xmath8cdm model
, wmap7 constrains @xmath161 to @xmath195 per cent ; this translates into a relatively large uncertainty in the predicted growth rate of structure ( @xmath103 ) of 5.5 per cent .
marginalizing over this uncertainty increases our uncertainty on @xmath139 from 0.022 to 0.026 .
weak lensing ( e.g. , * ? ? ?
* ) and/or cluster abundances ( e.g. , * ? ? ?
* ; * ? ? ?
* ) reduce the @xmath8cdm uncertainty on a combination of @xmath145 and @xmath161 similar to the gr prediction for peculiar velocities , @xmath196 , and thus could potentially be used to reduce the uncertainty on @xmath103 in the @xmath8cdm growth scenario . in the present work , we do not explore other dataset combinations besides wmap7 and cmass . the wigglez survey has recently analyzed the anisotropic clustering of bright emission line galaxies over a broad redshift range @xcite .
their growth rate constraints assumed a fixed underlying linear matter power spectrum and redshift - distance relation , for which our error is 0.033 . with a factor
@xmath198 fewer galaxies , they achieve comparable precision and good agreement with our central value in their @xmath199 bin : @xmath200 . in their analysis
they include modes up to @xmath201 and marginalize over a lorentzian model to account for small scale nonlinearities .
we find that the quoted wigglez error is in good agreement with our fisher matrix prediction if we assume @xmath202 , @xmath203 , @xmath204 , a gaussian damping @xmath205 and @xmath206mpc , i.e. , a factor of two smaller scale than we have adopted for our analysis .
exploration using the fisher matrix suggests that the difference in number densities between cmass and wigglez has a negligible impact on the uncertainties , while the lower bias of their sample implies a 15 per cent ( 10 per cent ) improvement at @xmath207 mpc . by far
the dominant difference arises because they are fitting out to @xmath208 ; their error would increase by a factor of 2 if they adopted our minimum scale .
another issue is that our mock covariance matrix , which accounts for both nonlinear growth of structure and our complex survey geometry , yield errors 40 per cent larger than our naive fisher matrix analysis would predict , with the difference between the linear and mock covariance matrices increasing on small scales ( for @xmath68 ) .
while it is not clear how this difference scales with galaxy bias or depends on survey geometry , the wigglez use of linear theory covariance matrices down to much smaller scales could cause their uncertainty to be underestimated .
another recent wigglez analysis @xcite simultaneously fits @xmath209 and @xmath210 at @xmath211 for a fixed underlying @xmath25 , and makes use of modes with @xmath212 , which can be compared with the last two lines of table [ table : uncertainties ] and model 4 in table [ tab : bf4b4 ] . as in the previous discussion ,
our best fit values for these parameters are consistent , and using clustering information on small scales ( i.e. , larger @xmath213 ) permits a factor of @xmath214 tighter constraint on @xmath139 at a fixed number of galaxy spectra compared with our clustering analysis of cmass galaxies
. however , the higher bias of cmass galaxies permits relatively tighter constraints on @xmath139 than @xmath103 as compared with wigglez .
the wigglez and boss cmass surveys target very different galaxy types , which will have different nonlinear properties and modeling uncertainties . without a detailed study it is impossible to understand the robustness of various assumptions for the nonlinear distortions ,
particularly when multiple sources of nonlinearity contribute .
these can potentially cancel when examining clustering , meaning that goodness - of - fit must be used carefully when assessing robustness and potential biases in a model fit .
in this work we have adopted a conservative approach based on the best model available that has been compared with a large volume of @xmath18-body mock galaxy catalogs , and only fit over those scales where we are confident that the signal is dominated by the quasilinear velocity field of interest , and where the impact of small - scale random motions can be simply modeled and marginalized over .
we have analyzed the anisotropic clustering of boss dr9 cmass galaxies with an accurate analytic model for the monopole and quadrupole correlation functions , which we have validated using a large volume of @xmath18-body based mock galaxy catalogs .
the combination of the bao standard ruler and the alcock - paczynski effect allows us to separately constrain the comoving angular diameter distance and hubble expansion rate at the effective redshift of our sample , @xmath2 , while redshift space distortions allow us to constrain the amplitude of the peculiar velocity field , a direct measurement of the growth _ rate _ of structure , @xmath215 .
table [ tab : bf4b4 ] summarizes our constraints under several assumptions .
figure [ fig : twodhist ] shows the degeneracy between the alcock - paczynski parameter @xmath139 and the growth rate of structure in our measurements ; this explains why our constraints improve considerably if we make further model assumptions about the geometry or growth . in the most general case where all three parameters vary independently , we find @xmath216 , @xmath217 mpc and @xmath218 km s@xmath4 mpc@xmath4 . to illustrate the cosmological constraining power of our measurements , we summarize our results as three distinct tests of a minimal @xmath8cdm cosmology ( see section [ cosmomodel ] ) that link the observed cmb anisotropies at @xmath219 and cmass galaxy fluctuations at @xmath220 , independent of the expansion history and growth of structure in the universe at redshifts below our sample ( @xmath221 ) .
we use a flat @xmath8cdm model to illustrate our constraining power on the behavior of the universe at @xmath11 , and explore more general cosmological models in our companion paper , @xcite .
+ _ is the shape of the power spectrum of matter density fluctuations inferred from the cmb consistent with the one inferred from galaxy fluctuations after a factor of @xmath222 amplification ? _ + if the dominant component of the energy density is `` cold '' , then in the linear regime perturbation growth in general relativity is scale - independent .
modulo our corrections for non - linear @xmath8cdm evolution and galaxy biasing ( @xmath223 per cent on the scales we analyse ) , the shapes of the linear matter power spectrum inferred from the cmb and galaxy clustering are consistent : our best fit @xmath8cdm model gives @xmath224 for 44 degrees of freedom .
we quantify this statement further using our fit to the location of the broad turnover in @xmath25 , which indicates the horizon size at matter - radiation equality and thus the physical matter density , assuming the radiation density is known . combining our constraint ( @xmath225 ) with those in the literature for the sdss - ii lrg sample and wigglez yields @xmath226 , where the error neglects the expected small but non - zero covariance between the galaxy samples , and also fixes @xmath227 and @xmath151 to best fit cmb values .
the cmb constraint is only slightly more precise than the combined galaxy measurement ( @xmath228 ) , and the two are in excellent agreement . with this test passed , in the rest of our analysis
we make use of the full matter power spectrum ( rather than just the bao feature ) as a standard ruler in galaxy clustering measurements .
see @xcite for an exploration of other cosmological models using the cmass monopole correlation function shape .
in particular , the good agreement between constraints on the linear matter power spectrum from the cmb and cmass galaxy clustering limit the allowed contribution from species such as massive neutrinos which induce scale - dependent growth .
+ _ do our constraints on the geometry of the universe require dark energy at @xmath229 ? _ + distance constraints from both the cmb and the bao feature are determined relative to the sound horizon at @xmath230 , the redshift of decoupling , and at @xmath231 , the redshift when baryons were released from the compton drag of the photons and @xmath232 from @xcite and @xcite , respectively . ] . within the constraints of our cosmological model assumptions ( see section [ cosmomodel ] ) , @xmath233 and @xmath234 depend only on @xmath235 and @xmath227 .
we project the observed _ difference _ in comoving angular diameter distance @xmath236 , @xmath237 onto the parameter @xmath238 in a flat @xmath8cdm cosmology and marginalize over the other parameters .
the result is shown by the dashed line in figure [ fig : ol ] : @xmath239 .
we quote the maximum likelihood and 68 per cent confidence region around it unless otherwise noted .
we have purposely chosen a variable that is independent of the expansion history at @xmath240 .
similarly , the cmb constraint on @xmath241 with our measurement of @xmath242 requires @xmath243 between @xmath2 and @xmath244 ( dotted curve in figure [ fig : ol ] ) .
+ _ does the observed growth rate of structure at @xmath245 require dark energy ? _ + because wmap places such a tight constraint on the amplitude of curvature perturbations deep in the matter - dominated epoch at @xmath246 mpc@xmath4 ( 1.8 per cent ) , we can translate our measurement of @xmath103 into a constraint on @xmath247 , where @xmath248 is the usual growth function that depends only on @xmath161 or @xmath238 in a flat @xmath8cdm cosmological model : @xmath249 ^ 3}.\ ] ] to do so , we require the additional cosmological model assumptions listed in section [ cosmomodel ] .
we marginalize over the wmap7 uncertainties in the parameters that convert curvature perturbations to the integrated amplitude of matter perturbations on scales of @xmath250 mpc , @xmath251 , namely @xmath241 , @xmath227 , and @xmath151 .
section [ sigma8def ] details the relation between @xmath251 and @xmath145 , which can be regarded as equal for our purposes .
this comparison between the fluctuation amplitude at @xmath244 and @xmath2 requires @xmath252 within @xmath253 $ ] ( central 68 per cent confidence ) , with a maximum likelihood at 0.76 .
the distribution is shown as the dot - dashed curve in figure [ fig : ol ] .
+ in combination with the cmb , both our geometric and growth rate constraints require a value of @xmath238 at @xmath254 that is consistent with the concordance model expectation . combining all three constraints with the wmap7 distance and normalization priors , and marginalizing over @xmath255 and @xmath241
, we find @xmath256 ( solid curve in figure [ fig : ol ] ) .
+ our analysis required further assumptions compared with a bao - only analysis @xcite that constrains @xmath257 .
however , we were able to perform tests of the scale - independence and growth rate of cosmic structure between recombination and @xmath258 , as well as to break the degeneracy between @xmath136 and @xmath7 using the alcock - paczynski test .
so far , our measurements do not unveil any deviations from the minimal @xmath8cdm model we have examined . under the assumption of a @xmath8cdm cosmology extending to @xmath259
, the bao feature adds the most constraining power to wmap7 on @xmath8cdm parameters ; we find @xmath260 , @xmath261 , and @xmath262 km s@xmath4 mpc@xmath4 in this model .
we anticipate statistical improvements on these results with the completed boss galaxy dataset covering a footprint three times larger , as well as developments in the theoretical modeling that will allow tighter cosmological constraints .
funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the u.s .
department of energy office of science .
the sdss - iii web site is http://www.sdss3.org/. sdss - iii is managed by the astrophysical research consortium for the participating institutions of the sdss - iii collaboration including the university of arizona , the brazilian participation group , brookhaven national laboratory , university of cambridge , carnegie mellon university , university of florida , the french participation group , the german participation group , harvard university , the instituto de astrofisica de canarias , the michigan state / notre dame / jina participation group , johns hopkins university , lawrence berkeley national laboratory , max planck institute for astrophysics , max planck institute for extraterrestrial physics , new mexico state university , new york university , ohio state university , pennsylvania state university , university of portsmouth , princeton university , the spanish participation group , university of tokyo , university of utah , vanderbilt university , university of virginia , university of washington , and yale university .
the simulations used in this paper were analysed at the national energy research scientific computing center , the shared research computing services pilot of the university of california and the laboratory research computing project at lawrence berkeley national laboratory .
bar gratefully acknowledges support provided by nasa through hubble fellowship grant 51280 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 .
mw is supported by the nsf and nasa .
ls and wp thank the european research council for support .
wp also acknowledges support from the uk science and technology facilities council .
mecs was supported by the nsf under award no .
to compute overdensities of the galaxy field , defined as @xmath263 we must know the unperturbed density field @xmath264 . while the angular selection function of the survey is usually well known
, the radial distribution is not easy to model accurately .
usually , the unperturbed radial distribution of galaxies is modeled from the distribution of observed redshifts by either shuffling them or splining with a smooth curve .
different ways of constructing a random catalog will result in different estimates of correlation function ( see , e.g. , * ? ? ? * ; * ? ? ? * ) . to estimate the magnitude of this effect on the measurements of the moments of correlation function we take 600 mock catalogs of cmass sample and apply simplified version of our analysis to the measurements produced using different ways of reconstructing @xmath265 .
we reconstruct @xmath265 first by shuffling `` observed '' redshifts in mock catalogs and then by splining that distribution with 10 , 20 and 30 node cubic spline fits .
we find that the ways of reconstructing @xmath265 for random catalogs do not affect the rsd measurements significantly . in our current analysis we are using the `` shuffled '' catalogs since they introduce the least bias in the measurements of @xmath266 @xcite .
we are unable to obtain redshifts for @xmath267 5 per cent of the galaxies due to fiber collisions no two fibers on any given observation can be placed closer than @xmath268 . at @xmath269
this @xmath268 exclusion corresponds to @xmath270mpc .
redshifts for some of the collided galaxies can be reclaimed in regions where plates overlap , but the remaining exclusion must be accounted for .
we account for fiber - collided galaxies by assigning its weight to its nearest neighbor on the sky .
tests on mock catalogs presented in @xcite indicate that the nearest neighbor correction adopted in this work is accurate to better than 1 per cent for both @xmath271 and @xmath272 at the scales used in our analysis .
redshift failures are discussed in detail in @xcite ; we also correct for them with a nearest - neighbor upweighting scheme .
redshift measurement errors smooth the apparent galaxy density field , in the same fashion as described with our nuisance parameter @xmath121 .
the median redshift error for our sample is 42 km s@xmath4 , which translates into an additive contribution to @xmath273 mpc@xmath128 .
equations [ eq : kaiser ] and [ streamingeqn ] assume that a `` plane - parallel '' approximation , which states that the sample is far enough from the observer so that all line - of - sights are parallel to each other , is accurate enough
. this approximation will fail at some scale for wide surveys ( see e.g. * ? ? ? * ) .
we approximate the scale dependent magnitude of wide - angle effects for our sample in a similar manner to @xcite .
the effective redshift of our galaxy sample is @xmath2 , which in the best - fit wmap7 cosmology corresponds to a comoving distance of approximately @xmath274 .
the largest scale that we consider in this analysis is @xmath22 , which corresponds to an opening angle of about 3 degrees .
we estimate the wide - angle corrections to be at most 10 per cent of the statistical errors on the largest scales and do not try to correct for this effect in current analysis .
@xcite demonstrated that eqn .
[ streamingeqn ] provides an accurate description of @xmath275 when real space clustering and velocity statistics inside the integrand are measured directly from @xmath18-body simulations .
moreover , in the regime where @xmath276 ( appropriate for cmass galaxies ) , the real space clustering and velocity statistics can be computed analytically with sufficient precision to predict @xmath275 . in detail , this `` sweet spot '' in the precision of the model arises because the functions @xmath112 and @xmath277 entering eqn .
[ streamingeqn ] were evaluated as a function of cosmological parameters in standard perturbation theory accounting for only the linear bias of the tracer . at the redshift of interest , the second order bias @xmath278 crosses zero near @xmath279 , and the calculation of these functions neglecting @xmath278 is sufficient for our purposes .
we therefore caution against the use of our code for tracers with bias substantially different from 2 .
note that @xmath280 is evaluated in lpt @xcite and includes the contribution to the real space clustering from second - order lagrangian bias , to the first order bias , @xmath281 , through the peak background split ; these parameters are not varied independently . ] .
+ standard perturbation theory is known to have inaccuracies in describing the bao feature , and we found that on bao scales errors in our perturbation theory calculation of pairwise velocity statistics caused inaccuracies in the prediction for @xmath275 .
however , on scales @xmath282 mpc , the redshift space version of lpt @xcite is very accurate ( see figure 2 of * ? ? ?
we therefore interpolate between our evaluation of eqn .
[ streamingeqn ] at smaller scales and lpt on large scales , with the transition fixed at 100 mpc .
finally , we model the effect of galaxy intrahalo velocities ( traditional `` fingers - of - god '' ) by convolving our model @xmath283 with an additional gaussian velocity dispersion of variance @xmath121 .
using the gaussian form allows @xmath121 to be included directly in the gaussian in eqn .
[ streamingeqn ] for faster evaluation of our model as we explore cosmological parameter space .
we have developed a code to numerically evaluate eqn .
[ streamingeqn ] as well as all the relevant perturbation theory integrals as a function of an input linear matter power spectrum and nuisance parameters @xmath90 and @xmath121 . the alcock - paczynski effect is easily incorporated using eqn .
[ eq : apeqn ] before computing legendre polynomial moments from @xmath275 .
the internal units of * cosmoxi2d * are mpc , in which the underlying linear matter power spectrum is most tightly constrained .
the code is publicly available .
because we have such a large volume of simulations , we can use the difference between the theoretical model at the known cosmological parameters of the @xmath18-body simulation and the measured correlation function from the mock galaxy catalogs to quantify our theoretical systematic error ; we find @xmath284 at the best fit value of @xmath121 .
we also compare the scale dependence of the model error with the 5 parameters we are fitting , and find @xmath285 shifts compared to the _ unmarginalized _ uncertainties on all parameters ( i.e. , the uncertainty on each parameter if all the others were known perfectly ) .
therefore we conclude that our systematic error is negligible in the context of this analysis .
another concern is that the model becomes inaccurate rapidly on scales smaller than our minimum fitting scale , @xmath286 .
when alcock - paczynski parameters vary , scales smaller than @xmath286 contribute to the model .
however , we have verified that for variations of @xmath287 per cent in @xmath137 and/or @xmath288 per cent in @xmath139 ( i.e. , much larger than the final uncertainties ) , the theoretical error induces @xmath289 shifts in all parameters compared with the unmarginalized uncertainties . finally , we point out that unlike in linear theory , our model depends on @xmath90 , @xmath171 , and @xmath145 separately , rather than the only the combinations @xmath102 and @xmath103 .
however , at the particular @xmath102 and @xmath103 values of our sample , we find that changes in the predicted @xmath122 with @xmath145 can be absorbed by changes in the value of @xmath102 .
quantitatively , for a @xmath140 per cent change in away from our fiducial @xmath290 and at fixed @xmath25 , ap parameters , and @xmath291 mpc@xmath128 , @xmath102 shifts by @xmath292 with no measurable shift in the central value of @xmath103 .
as shown in figure [ fig : theoryfits ] , our model with @xmath293 fits the @xmath104 of our mock galaxy catalogs , based on the best - fitting hod in @xcite . in this section
we quantify how uncertainties in both the theoretical modeling and data analysis cause uncertainty in the expected value of @xmath121 for the cmass sample .
we address several aspects of this problem separately .
* _ one - halo vs two - halo contributions to @xmath122 and fiber collision corrections _ : the formalism of the halo model distinguishes between `` one - halo '' and `` two - halo '' pairs depending on whether the two galaxies occupy the same or different halos . in figure [ fig:1hvs2h ]
we show the total change in @xmath122 due to satellite galaxy intrahalo velocities ( ihv ) by the dashed ( dotted ) curves in our mock catalogs .
the dash - dot and solid curves show the contributions to this change in @xmath122 from one - halo pairs , which are localized to along the los with @xmath294 .
because the one - halo pairs contribute to such a small @xmath295 , they can be neglected on the scales used in our cosmological parameter fits , @xmath296 .
this fact is important to establish since our method for fiber collision corrections will correctly recover the distribution of pair separations for pairs of galaxies with separations larger than the fiber collision scale , but suppresses the contribution of pairs of galaxies at the fiber collision scale ( @xmath297 ) . * _ uncertainty in the hod at fixed cosmology _ : uncertainties in the hod parameters will introduce an uncertainty in @xmath121 .
while @xmath121 is roughly proportional to the satellite fraction , it also depends on the distribution of host halo mass increasing @xmath298 and @xmath299 increases host halo mass at fixed satellite fraction , which increases @xmath121 .
we use mcmc chains from @xcite to estimate our uncertainty on @xmath121 at fixed cosmology to be 6 mpc@xmath300 . *
_ breakdown of the `` central '' galaxy assumption _ : the analysis of @xcite suggests that the brightest galaxy in a halo is not always the `` central '' one .
we test the impact of relaxing our assumption that the velocity of mock central galaxies have no intrahalo velocities by assigning them the intrahalo velocity of a random dark matter particle halo member in our simulations some fraction @xmath184 of the time .
when we have more than one galaxy in a halo , we assume that the chance of not including the `` central '' galaxy in our sample is lower , @xmath301 . for @xmath302 , we find @xmath121 at our fiducial hod is increased by 9 mpc@xmath128 . * _ variations in the halo mass function with cosmological parameters _ :
a broad range of observations shows good agreement between the concordance @xmath8cdm halo mass function and the multiplicity of galaxy groups and clusters ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , so this uncertainty is subdominant : even if our fiducial hod masses were scaled by a factor of 2 , @xmath303 would change by 11 mpc@xmath128 .
given the above considerations , we adopted a generous hard prior on @xmath121 between 0 and 40 mpc@xmath128 .
we adopt a hybrid mcmc / importance resampling approach to explore the boss likelihood surface with various priors imposed from the wmap7 likelihood in models 1 - 4 .
this approach is necessary in our case because our model evaluation is slow , and we must marginalize over @xmath102 and @xmath121 at each point in parameter space that we consider .
this section describes our methods in each case .
wmap7 mcmc chains are publicly available .
these chains provide a fair sample of the wmap7 likelihood surface .
importance resampling @xcite allows us to compute how constraints on the model parameters change given an additional constraint by evaluating the new likelihood at a subsample of the original mcmc chain , multiplying the original weight of each element by the new likelihood , and then recomputing confidence intervals .
for this model we marginalize only over two parameters , so it is feasible to directly compute a marginalized cmass likelihood : @xmath304 we then use importance resampling of the wmap7 chain parameters .
[ [ model-2-wmap7cmass - lambdacdm - geometry - free - growth - model-3-wmap7cmass - lambdacdm - growth - free - geometry ] ] model 2 wmap7+cmass @xmath8cdm geometry , free growth ; model 3 wmap7+cmass @xmath8cdm growth , free geometry ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in models 2 and 3 , for each linear matter power spectrum , we must vary three or four extra parameters describing the galaxy clustering observations .
we therefore explore the likelihood surface at each point in the wmap7 chain by mcmc , and thereby sample the cmass likelihood distribution @xmath305 . in model 2 , @xmath306 , and in model 3 ,
mcmc chains at a fixed @xmath308 sample parameter space proportional to @xmath309 , but in order to compute the marginalized likelihood of @xmath103 in model 2 or @xmath310 in model 3 , we must determine the relative likelihood of the mcmc chains evaluated at different @xmath311 .
since @xmath312 where @xmath18 is an overall normalization , we can combine the @xmath313 computed at any point with the probability density estimated by our mcmc to determine the relative normalization of @xmath314 . in practice , we first normalize each mcmc distribution so that @xmath315 , find @xmath316 with the minimum value of @xmath313 in each chain , and integrate @xmath317 in a small , fixed size region of parameter space around @xmath318 , which we call @xmath319 . , we restrict @xmath320 to be sufficiently far from the hard prior boundary of @xmath121 , so that the volume element @xmath321 does not intersect the hard boundary . the difference between @xmath322 and the global @xmath313 minimum is small , since @xmath121 is poorly constrained by our measurements . ] the relative weight of each point in the wmap - only mcmc chain is then determined by the cmass - only likelihood value and @xmath323 : @xmath324 we find that our constraints are the same if we neglect this volume weighting factor @xmath325 , indicating that the effective volume of parameter space allowed by the cmass measurements does not strongly depend on the underlying cosmological parameters @xmath326 when exploring the region of this parameter space allowed by wmap7 .
in this case , we use wmap data to provide a prior on the shape of the linear matter power spectrum , which is well approximated by a multivariate gaussian in the parameters @xmath327 , @xmath227 , @xmath151 .
wmap7 constraints on these parameters primarily come from ratios of peak heights and the overall shape , rather than the locations of the peaks ( which are sensitive to the angular diameter distance to the last scattering surface ) .
therefore , marginalized likelihood for these parameters is nearly independent of the adopted model for the low redshift expansion history ( i.e. , whether @xmath158 , or @xmath157 are freely varied ) .
thus we are able to make use of the cmb information on the underlying linear matter power spectrum that is independent of the model for the low redshift expansion history , and does not contain information on the distance to the last scattering surface . by using the linear matter power spectrum as a `` standard ruler '' , we are able to infer information about the geometric parameters @xmath193 and @xmath139 . in this case
we run an mcmc chain with the following 8 parameters , adding an additional multivariate gaussian likelihood representing the cmb prior on the three linear matter power spectrum parameters : @xmath328 . | we analyze the anisotropic clustering of massive galaxies from the sloan digital sky survey iii baryon oscillation spectroscopic survey ( boss ) data release 9 ( dr9 ) sample , which consists of 264283 galaxies in the redshift range @xmath0 spanning 3275 square degrees .
both peculiar velocities and errors in the assumed redshift - distance relation ( `` alcock - paczynski effect '' ) generate correlations between clustering amplitude and orientation with respect to the line - of - sight . together with the sharp baryon acoustic oscillation ( bao )
standard ruler , our measurements of the broadband shape of the monopole and quadrupole correlation functions simultaneously constrain the comoving angular diameter distance ( @xmath1 mpc ) to @xmath2 , the hubble expansion rate at @xmath2 ( @xmath3 km s@xmath4 mpc@xmath4 ) , and the growth rate of structure at that same redshift ( @xmath5 ) .
our analysis provides the best current direct determination of both @xmath6 and @xmath7 in galaxy clustering data using this technique .
if we further assume a @xmath8cdm expansion history , our growth constraint tightens to @xmath9 . in combination with the cosmic microwave background , our measurements of @xmath6 , @xmath7 , and @xmath10 all separately require dark energy at @xmath11 , and when combined imply @xmath12 , independent of the universe s evolution at @xmath13 . in our companion paper @xcite
, we explore further cosmological implications of these observations .
-1.25 cm [ firstpage ] cosmology : large - scale structure of universe , cosmological parameters , galaxies : haloes , statistics |
the observed neutrino oscillation@xcite implies strong lepton flavor violation(lfv ) . in the standard model(sm )
, lepton number is exactly preserved in contradiction with the neutrino oscillation .
some shortcomings of sm , for instance , the emergence of quadratic divergences in the higgs sector , imply that the sm need to be extended . the minimal supersymmetric standard model ( mssm)@xcite is one of the most promising candidates in all extensions of the sm .
the most general superpotential in supersymmetry theory contains bilinear and trilinear terms which do not conserve either the baryon number ( b ) or the lepton number ( l ) .
the simultaneous presence of both lepton number violating and baryon number violating could lead to very rapid proton decay . in order to turn off proton rapid decay ,
supersymmetric models introduce a discrete r - parity symmetry@xcite implying a conserved quantum number @xmath5 , where b , l and s are the baryon number , the lepton number and the spin of particles respectively .
however , such a stringent symmetry appears short of theoretical basis .
especially we know that a stable proton can survive by imposing only one of l- and b- conservation
. moreover , non - zero r - violating couplings might provide small neutrino masses , which would explain the phenomena of neutrino oscillation experiments .
thus , there is strong theoretical and phenomenological motivation to introduce partial r - parity violations into the most general representations of superpotential , which can be written as @xcite @xmath6 where @xmath7 = 1,2,3 are generation indices ; @xmath8 = 1,2 are su(2 ) isospin indices and @xmath9 are su(3 ) color indices .
@xmath10 are dimensionless r - violating yukawa couplings behaving as @xmath11 , @xmath12 . in above superpotential , the last bilinear terms mix the lepton superfield and the higgs one , which might generate masses of neutrinos and consequently introduce compatible description of neutrino oscillation in a natural way@xcite .
all the other trilinear terms only violate either l- or b - symmetry respectively , and the terms that may produce both l- and b - violation simultaneously are forbidden in superpotential so that a stable proton is ensured . experimental detection on the signals of r - parity violating interactions would be the most outstanding evidence of new physics beyond the sm .
if nature is supersymmetric and neutrino oscillations are really induced by r - violating couplings included in a general superpotential , then the prediction of single sneutrino production modes at colliders is straightforward .
therefore , probing single sneutrino production on a high energy collider is specially attractive in experimental searches and relevant phenomenological studies about r - violating interactions .
such exploration ought to be promising if carried out on fine high energy lepton colliders .
there were some experimental bounds on lfv di - lepton cross - section measurement from lep experiments@xcite , which could attribute to single sneutrino productions and would set mass constraint on these neutrino super - partners .
we have discussed the possibility of probing off - shell sneutrino effect induced by @xmath13 terms on a 500gev linear collider @xcite , and concluded that with current constraints on @xmath14 parameters one can extend sneutrino search up to @xmath15tev region which is far beyond the constraints from lep .
nevertheless , despite the high performance of linear colliders and clean environment in their detectors , none of tesla , jlc or even ilc projects would be in commission in this decade .
on the other hand , the tevatron hadron collider is currently in its upgrade run - ii physics and the lhc will be put into running in three years , and consequently more careful studies on direct susy searches including the potential detection of single sneutrino production on hadron colliders should be addressed in detail both experimentally and phenomenologically . the first two terms @xmath16 and @xmath17 in the superpotential eq.([sup ] ) may lead to r - violating single sneutrino production and sequential lfv decay on hadron colliders .
many papers studied these single sneutrino production modes both on- and off - mass - shell of sneutrino resonance via @xmath17 interactions and decay in r - parity conservation mode as @xmath18 @xcite .
charginos and subsequential neutrolinos could decay in r - conserving way as well , e.g. @xmath19 and @xmath20 ; then sleptons from neutrolino would decay in r - violating modes as @xmath21 or @xmath22 . with these r - violating interaction vertexes involved in sneutrino and chargino / neutrolino decay chains , single sneutrino will subsequently decay to all the sm stable particles at last .
namely , these single sneutrino production and subsequential @xmath23 decay processes are doubly r - violated , and unless light neutrolinos would live long enough to escape from detectors there would be more than three leptons in final states .
these characteristic triple - lepton signals are obviously good for tagging new physics .
but there are problems in susy parameter determination from these multiple leptons channels : firstly , the tri - lepton production can also be induced by r - conserving supersymmetric models , for instance , @xmath24 associated production may have the subsequential decay to three leptons and two stable @xmath25 as the lightest supersymmetric particle ( lsp ) which performances as missing energy in detector .
so , these tri - lepton channels might not be able to distinguish supersymmetric r - violating interaction signals from those of r - conserving ones .
secondly , since there are too many parameters involving in @xmath26 and chargino and neutrolino cascade decay , it is complicated to determine any of susy parameters decisively , and one has to adopt a very compactly simplified version of parameters such as msugra . on the contrary
, the lfv di - lepton processes of sneutrino rare decay @xmath27 on hadron collider would provide an ideal detection channel for r - violating interactions and a potential means of abstracting some susy parameters experimentally . as we will show in following sections , with some reasonable assumption on parameter space and appropriate treatment on experimental observables ,
one is not only able to select lfv di - lepton signals from numerous sm physics background , but also able to decouple the contribution of sneutrino part in signals from that of other susy parts , so that an unique sneutrino mass parameter could be abstracted explicitly .
leaving alone the lhc in construction , tevatron run - ii upgrade is now in progress .
it is expected to deliver an proton - antiproton collision integrated luminosity as @xmath28(@xmath29 ) per experiment at the energy of 1.96 tev , which will be quantitatively one order higher than the luminosity acquired in run - i with center of mass energy 1.8 tev .
therefore , before lhc , the two experiments d and cdf on the tevatron are the only available facilities to test r - violating interaction and search for sneutrino . in this paper , we will discuss the possibility of probing r - violating lfv di - lepton signals and abstracting sneutrino information from experimental observation at the tevatron run - ii .
we investigate two di - lepton signal processes in which final states can be identified precisely even in hadron colliding environment , namely @xmath30 and its charge - conjugated process @xmath31 . by integrating the first two lepton number violation terms in eq.([sup ] )
, one can obtain the lagrangian relevant to present discussion as @xmath32 where superscript @xmath33 refers to charge conjugation . in this work ,
we simply take the r - violating parameters @xmath14 and @xmath34 to be real to avoid further complication . in order to get optimal efficiency of signal detection , we choose not to distinguish the absolute sign of lepton charge , so that di - track events of @xmath35 final states would be treated equally as those of @xmath36 ones .
this summation treatment of @xmath3 events denotes the summation of two final states @xmath35 and @xmath36 events . ]
is based on the following consideration : firstly , it can double the statistic of signal events , which will improve the significance and confidence on collected signal samples .
secondly , the charge determination of tracks with high transverse momentum(@xmath37 ) relies on many realistic detector - relevant factors , thus indiscrimination between @xmath38 and @xmath39 will remove systematic uncertainty on charge measurement . to reflect the summation measurement in theoretical calculation , we phenomenologically introduce an unique parton - level @xmath40 , to denote polar angles of both outgoing @xmath41 and @xmath42 with respect to the incoming parton from proton beam , and adopt a consistent momentum notation to all signal processes as @xmath43 where @xmath44 are the first generation partons , @xmath45 stands for the four - momenta of partons from proton beam whose direction of vector @xmath46 is along @xmath47 axis , and @xmath48 stands for those from anti - proton beam .
we always use the four - momentum @xmath49 to denote outgoing @xmath41 and @xmath42 in signal processes , so that the open angle between @xmath50 and @xmath51 is precisely @xmath52 defined above .
correspondingly , @xmath53 represents the four - momenta of final @xmath54 particles . neglecting electron and muon masses ,
we introduce mandelstam invariant variables to describe kinematics of all the four signal subprocesses as @xmath55 and feynman diagrams of @xmath3 signals at parton level are depicted in figure 1 accordingly .
we define the amplitude of @xmath56 subprocess as @xmath57 , where the superscript minus sign in parenthesis indicates the processes with negative charged electrons in final states , and in the following we will use the superscript @xmath58 to denote processes with positive charged outgoing positrons .
the amplitude @xmath59 can be represented by @xmath60- and @xmath61-channel parts @xmath62 with @xmath63 @xmath64 and the notations of propagator factors @xmath65 are defined as follows @xmath66 where a sparticle width @xmath67 is introduced into the @xmath68-channel propagators to suppress resonance singularity .
according to the feynman diagrams in figure 1 , the amplitude of @xmath69 subprocess has only @xmath70-channel part and can be denoted as @xmath71 with @xmath72 in above equation , @xmath73=@xmath74 are left- and right - hand project operators .
@xmath75 is generation index of supersymmetric particles , @xmath76 denotes the two mass eigenstates of each scalar quark flavor . @xmath77 and @xmath78 are the @xmath79 matrices used to diagonalize various up and down - type squark mass matrices , respectively . the amplitudes of the other type of signal subprocess @xmath80 , represented by eq.([proc2 ] ) , can be written out straightforward from above expressions of eq.([proc1 ] ) as @xmath81 the amplitudes of initial first generation sea quark collision subprocesses , represented by eq.([proc3],[proc4 ] ) , can be obtained from those of valence incoming quark subprocesses of eq.([proc1],[proc2 ] ) under following replacement @xmath82 by now all the signal amplitudes at parton level are given , where those of subprocess ( [ proc1 ] ) are written out explicitly , while those of the other three ones are obtained by appropriate initial and final state exchange respectively .
in this paper , we mainly focus on probing sneutrino resonance effect via potential lfv @xmath3 signals at the tevatron run - ii .
since the @xmath61 or @xmath70-channel squark internal exchange diagrams contribute into signal measurement and arose ambiguity in exploring sneutrino information , which makes sense via the @xmath68-channel exchange , we name all the squark contribution as pollution of the signal .
fortunately , because of the r - violating scalar - pseudoscalar(s - p ) yukawa couplings of sfermion to dirac fermions as shown in eq.([lag ] ) , the interference terms among @xmath60-channel sneutrino exchange diagrams and @xmath61 and @xmath70-channel squark interchange ones get vanished in all subprocesses .
we call the vanishing of the interference terms between amplitudes of @xmath83 and those of @xmath84 , as a decouple feature between the two parts of sneutrino and squark contributions . in this way
, we might be able to abstract sneutrino information from @xmath85 measurement . in order to reduce the number of parameters in sneutrino and squark sections
, we simply take that the mass spectrum of all flavor scalar quarks is highly degenerated , and reasonably assume the mass splitting among three heavy sneutrinos is trivial so that an unique mass scale parameter could be introduced in the sneutrino section .
that means @xmath86
@xmath87 to reflect the @xmath3 summation and the decouple feature between sneutrino and squark contributions in the calculation of the signal production cross - section at hadron - level , we define a set of parton - level differential cross - section components as @xmath88 where @xmath89\cdot { \nonumber}\\ & & \sum_{i , i^{'}=1}^{3}\sum_{k , k^{'}=1}^{2}(~\lambda^{'}_{1i1 } \lambda^{'}_{2i1}|r_{k1}^{\tilde u_i}|^2\cdot \lambda^{'}_{1i^{'}1}\lambda^{'}_{2i^{'}1}|r_{k^{'}1}^{\tilde u_{i^{'}}}|^2~ ) \\ { \nonumber}\\ \label{mtuuu }
m}^{\hat t + \hat u}_{\bar u u}|^2 = |{\cal m}^{(-)\hat u}_{u\bar{u}}|^2+|{\cal m}^{(+)\hat t}_{u\bar u}|^2 { \nonumber}\\ & = & [ ~\hat u^2\cdot { \cal p}(\hat u , m_{\tilde q})^2+\hat t^2\cdot { \cal p}(\hat t , m_{\tilde q})^2~]\cdot { \nonumber}\\ & & \sum_{i , i^{'}=1}^3\sum_{k , k{'}=1}^2 ( ~\lambda^{'}_{21i}\lambda^{'}_{11i}|r_{k2}^{\tilde d_i}|^2 \cdot \lambda^{'}_{21i^{'}}\lambda^{'}_{11i^ { ' } } |r_{k^{'}2}^{\tilde d_{i^{'}}}|^2~)\end{aligned}\ ] ] the above expressions manifest that with the summation treatment on @xmath3 signal processes , the @xmath60-channel contribution is doubled , while @xmath61 and @xmath70-channel contributions always come together and will cancel their individual forward - backward distribution asymmetry on polar angle @xmath52 with each other at parton level . due to the decouple feature between sneutrino and squark sections , we are safe to divide the calculation of @xmath3 signal total cross - section at the tevatron into two separate parts named as sneutrino and squark cross - section contributions , respectively as @xmath90 = \sigma^{(\tilde{\nu } ) } + \sigma^{(\tilde{q})}\ ] ] and the differential cross - sections are @xmath91 and @xmath92 we adopt the pythia @xcite built - in parton distribution function ( p.d.f ) of the proton and antiproton , cteq5l , and implement above hadron - level calculation by convoluting eq.([equation 2 ] ) and eq.([equation 3 ] ) with 1.96 tev @xmath93 collision energy . here we should clarify that the above defined parton - level @xmath94 , which is the polar angle of outgoing @xmath95 with respect to the incoming parton from proton beam , would be boosted to an experiment observable angle @xmath96 between @xmath95 and proton beam . in this way , we get an event generator of lfv @xmath3 production at the tevatron based on the package pythia . now the question is , apart from the theoretical calculation , how we can find some appropriate observables in which the overlap of sneutrino and squark contribution parts is minimized , so that one can suppress squark pollution to a trivial level and abstract sneutrino information clearly from experiment .
we take the r - parity violating parameters @xmath14 and @xmath34 under the experimental constraints presented in ref.@xcite .
all the scale factors in the constraints which are susy particle mass dependent are simply ignored , so that the r - violating yukawa couplings could be naturally kept at values less than @xmath97(@xmath98 ) .
thereby , there is only mass scale parameter @xmath99 in sneutrino section as free parameter which should be determined by experiment .
the degenerate sneutrino mass parameter has to be constrained by the latest concrete lep collider data .
for example , opal experiment has set an upper limit on @xmath100 as @xmath101 with @xmath102 at 95% cl @xcite . to get a consistent value with this observation
, one can arrive at@xcite @xmath103 a unique sneutrino decay width @xmath67 is set to be @xmath104 , which results in a very short @xmath105 lifetime . the degenerate scalar quark mass
is taken as @xmath106 except otherwise specified , and @xmath79 @xmath107 and @xmath108 mixing matrices are set to be unit for simplification .
according to the decouple feature of sneutrino and squark section at parton level , we plot two independent contributions to the @xmath3 signal production cross - section on @xmath109 collider at @xmath110 , as the functions of corresponding mass parameter @xmath111 and @xmath112 respectively in figure 2 . due to the @xmath60-channel resonance enhancement ,
the sneutrino contribution is dominant over the squark one in most region of parameter space .
however , on account of the less density of valence quark p.d.f in higher energy region , the resonance enhancement would decrease with the increment of sneutrino mass , and the cross section drops to a few femto barn in the vicinity of @xmath113gev .
the calculation shows as well that the cross - section at hadron level contributed by scalar quark exchange @xmath61+@xmath70-channel is 5@xmath114 when @xmath115gev , and becomes negligible when @xmath116gev . a set of mc distributions of @xmath3 signal observables at the tevatron , i.e. polar angle @xmath117 , pseudo - rapidity @xmath118 and transverse momentum @xmath37 of outgoing electron , and invariant mass of @xmath3 system are plotted in figure 3 , with @xmath119gev and @xmath115gev where the pollution from scalar quark section is maximized .
the sneutrino contribution is characterized by the di - lepton invariant mass peak around @xmath120 , and the lepton transverse energy tendency as @xmath121 , while the two leptons from squark section are much softer .
therefore , with a proper cut on @xmath3 invariant mass , we ought to get rid of squark pollution and derive a purified sneutrino section .
it should be clarified here we adopt the eq .
( [ equation 2 ] ) and eq.([equation 3 ] ) to generate the above two individual virtual mc event samples contributed by squark and sneutrino respectively , but would generate the true samples by using the differential cross - section contributed by both squark and sneutrino sections in following event selection strategy studies . for the background of the signal
, we consider the following processes which have significant @xmath3 candidates in their final states @xmath122 where @xmath123 refers to decay products of @xmath124,@xmath125,@xmath126 and @xmath127 other than @xmath3 .
as discussed in ref@xcite , @xmath128 process can be removed by calorimeter - based energy isolation cut on both @xmath3 candidates , for these leptons from b - quark semi - leptonic decay are very soft and always associated with charmed meson as @xmath129 . however , for @xmath130 events where top quark decay to @xmath131 exclusively , there are possible @xmath3 final states from @xmath132 subsequential decay , together with b - quarks fragmentized into two jets which can not be tagged easily .
different from @xmath133 collider case , these @xmath130 production events with @xmath3 in final states accompanied by two b - quarks jets might not be distinguished from those signal events , in which besides @xmath3 there are two jets produced by remnant parton collision or multiple interactions in high luminosity hadron collider environment .
therefore , there are three kinds of processes which should be taken into account as physical background , and relevant cross - sections at the tevatron run - ii are @xmath134 \times 2\cdot br[w\rightarrow e\nu_e ] \cdot br[w\rightarrow \mu\nu_{\mu } ] { \nonumber}\\ & = & 8.45~pb \times 2\cdot 11\% \cdot 11\% = 204.5~fb\\ { \nonumber}\\ \sigma_{tt } & = & \sigma[p\bar p\rightarrow t\bar t ] \times 2\cdot br[w\rightarrow e\nu_e ] \cdot br[w\rightarrow \mu\nu_{\mu } ] { \nonumber}\\ & = & 8.32 pb \times 2\cdot 11\% \cdot 11\% = 201.3~fb\\ { \nonumber}\\ \sigma_{\tau\tau } & = & \sigma[p\bar p\rightarrow \tau^+ \tau^- ] \times
2\cdot br[\tau \rightarrow e\nu_e\nu_{\tau } ] \cdot br[\tau \rightarrow \mu\nu_{\mu}\nu_{\tau } ] { \nonumber}\\ & = & 229.67~pb \times 2\cdot 17\% \cdot 17\% = 13.27~pb\end{aligned}\ ] ] some mc distributions of @xmath130 and @xmath135 background are depicted in figure 4 and figure 5 respectively .
distributions of @xmath132 are very similar to those of @xmath130 and thus are not plotted here .
these distributions show that in @xmath130 and @xmath132 events there are quite large missing transverse energy and non - collinearity of @xmath3 in azimuthal plane , while @xmath3 final states in @xmath135 events are very soft and derive small invariant mass .
these features of background would be useful to develop an efficient event selection strategy to suppress background and improve the significance of signal .
taking the advantage of the collinearity and the resonance effect of energetic @xmath3 signal contributed by sneutrino section , we suggest following off - line event selection strategy at run - ii tevatron to reduce sm physical background as well as squark pollution 1 .
* cut1 * geometry acceptance : heavily boosted at the tevatron , the pseudo - rapidity distribution of signals does not show distinct difference from those of background , namely all @xmath3 candidates would tend to forward - backward distributed .
accordingly , we merely adopt the real detector acceptance on both signal and background to avoid signal loss . for demonstration
, we take the real geometry of upgraded ddetector@xcite in simulation , where the pseudo - rapidity coverage of muon+tracker system is as latexmath:[\[\begin{aligned } and the coverage of the central calorimeter which provides better energy resolution than the two endcap ones is as latexmath:[\[\begin{aligned } 2 .
* cut2 * transverse momentum constraint : in order to reject tremendous soft qcd jets which might produce counterfeit electron and muon simultaneously in high luminosity hadron collision environment , we use @xmath37 cut on both electron and muon candidates as : @xmath137 3 . * cut3 * invariant mass constraint : in order to remove the scalar quark pollution from total signals
, we make an invariant mass cut as @xmath138 which is just greater than half of 250gev , the current lower bound on sneutrino mass given in eq.([msnlimt ] ) .
we deliberately choose such a loose cut to leave room for resolution and uncertainty in real track transverse momentum and em object energy scale measurement .
in this way , @xmath135 events will be suppressed heavily .
additionally , this invariant mass constraint would also be efficient on some instrumental background , e.g. single @xmath139 production which decays via @xmath140 where one of its decay produced muons occasionally deposits nearly all its energy in calorimeter and thus provides a fake electron candidate .
however , a good - sized part of @xmath132 and @xmath130 events could survive after this cut yet .
cut4 * collinearity constraint : assuming high spatial resolution on x - y plane vertical to the beam , we set a cut on @xmath3 @xmath141-difference to select back - to - back events as : @xmath142 where we allow a @xmath143 redundancy , considering that the remnant parton collisions recoiled in signal processes might carry some significant amount of transverse momenta in total .
cut5 * missing transverse energy constraint : in lfv signals there is not any missing @xmath144 .
however , in consideration of detector resolution we cut on a non - zero value as @xmath145 four sets of mc samples for @xmath146,@xmath147,@xmath148 background and lfv signal productions with @xmath119 and @xmath115gev are generated by pythia at the tevatron run - ii @xmath109 collider with @xmath149 , scaled with different assumed luminosity .
each sample is simulated with the five - step event selection , and the numbers of events passing individual cut are listed in table [ tablesb],respectively .
.event selection efficiency on background and signal .
the first 6 rows are numbers of events before and after individual cut .
the values of event selection efficiency on different samples are given in the 7th row . [ cols="<,^,^,^,^ " , ] despite different assumed luminosity in generating the mc samples , we define a selection efficiency variable on both background and signal as @xmath150 it s demonstrated that with the five - step strategy we are able to reduce the background by 2 - 3 orders , i.e. @xmath151 , @xmath152 and @xmath153 , while keeping the selection efficiency on the r - violating @xmath3 signal as high as @xmath154 .
with certain integrated luminosity l , the number of background events b and that of @xmath3 signal s dominated by the contribution of sneutrinos with @xmath111=500@xmath155 after selection are given by @xmath156 where @xmath157 is the sum of @xmath146 , @xmath147 and @xmath130 contribution after cuts , and @xmath158 is the cross - section of r - violating lfv @xmath3 signal .
the significance of signal over background is defined as @xmath159 in the following discussion , we optimistically assume some 10@xmath29 integrated luminosity would be accumulated at both the d and cdf experiment eventually , which is not an unreachable goal for the tevatron run - ii luminosity upgrading is in progress remarkably . accordingly , the significance of signal from 500 @xmath155 sneutrino contribution is about 3 .
the significance @xmath160 as functions of sneutrino mass @xmath111 at the tevatron is plotted in figure 6 , with integrated luminosity being 5@xmath29 and 10@xmath29 respectively , where l=5@xmath29 is taken as conservative estimate for reference .
one can see clearly that lfv signals contributed by sneutrinos lighter than 430@xmath155 would derive significance of signal over the sm background greater than 5 for l=10@xmath29 , and thus would be detected explicitly .
the signal cross - section dominated by the contribution of 430@xmath155 sneutrinos would hold as large as 4.2@xmath114 after selection , and there will be some 40 signal events on record together with 70 background events in total @xmath1 data .
the transverse energy distributions of selected electron candidates are plotted in figure 7 . despite trivial remnant of scalar quark pollution , the outstanding peak in higher region of selected electron @xmath161 distribution
denotes the half value of sneutrino mass scale , and reveals the r - violating single sneutrino production distinctively .
figure 6 and 7 show that with an optimized 10@xmath29 luminosity recorded in total , lfv signals of degenerated sneutrinos with masses smaller than 400@xmath155 would be detected at the tevatron at more than 5 significance , and an apparent precipitate structure shown in transverse energy distributions of selected electron candidates would reveal the whereabout of sneutrino mass scale .
the ratio of signal to background can be as large as 0.6 , which is acceptable for a stable data analysis .
consequently , if the sneutrino mass scale really lies in the region of @xmath162 , the d experiment at the run - ii tevatron will discover these supersymmetric scalar partners of dirac neutrinos via lfv @xmath3 signal processes ; while if no experimental evidence is observed , one can exclude sneutrino up to @xmath163 at 95% cl where significance is smaller than 2 .
we adopt the criterion that the r - violating lfv signal is discovered at @xmath164 and excluded when @xmath165 .
accordingly , figure 8 shows the discovery and exclusion regions of sneutrino as functions of integrated luminosity at the tevatron .
the solid and dotted curves correspond to the significance of the signal over background(sb ) being @xmath166 and @xmath167 , and thus the sneutrino mass parameter regions above or under the hatched area can be excluded or discovered , respectively .
we have studied the lepton flavor violation processes @xmath168 at the tevatron run - ii with @xmath169 colliding energy , in the framework of the r - parity violating mssm .
there are two types of r - violating interactions involved in the lfv signal processes , namely the couplings of sneutrino - leptons(quarks ) vertices and those of squark - quark - lepton ones . fortunately , due to the r - violating s - p yukawa couplings of sfermion to dirac leptons and quarks , the contribution of sneutrino section is decoupled with that of squark part . accordingly , one is able to cut on an appropriate observable such as invariant mass @xmath170 to remove most of scalar quark pollution , and derive sneutrino mass parameter from purified data sample which is dominated by sneutrino contribution .
making use of the sneutrino resonance effect and @xmath3 collinearity in transverse plane vertical to beam pipe , we develop a set of event selection strategies . under these strategies ,
the sm background can be under control , so that it is possible to probe the r - parity violating interactions and to abstract the sneutrino information from experimental observation .
we conclude that with an assumption of 10@xmath29 integrated luminosity , the experiments at the tevatron run - ii machine would have the potential to discover sneutrino in the region of @xmath171 by detecting lfv @xmath3 signals , or extend the mass scale constraint up to @xmath172 at 95% cl within the mssm framework with r - parity violation .
s25 y. fukuda et al . , super - kamiokande collaboration , phys .
* b433 * ( 1998)9 ; phys . lett .
* b436 * ( 1998)33 ; phys .
* 81 * ( 1998)1562 ; m. apollonio et al .
, chooz collaboration , phys .
* b420 * ( 1998)297 .
r. barbieri et al . , hep - ph/9810232 ; b. allanach et al . , hep - ph/9906224 ; f. deliot et al , phys . lett.*b475 * ( 2000)184 ; g. moreau et al .
* b604 * ( 2001)3 ; s. bar - shalom et al .
phys . rev . *
d64 * ( 2001 ) 095008 .
the d detector is described in detail elsewhere as : t.lecompte and h.t.diehl , " the cdf and d upgrades for run ii , ann .
nucl . part .
sci . * 50 * , 71 ( 2000 ) ; v. abazov , _ et al .
_ , in preparation for submission to nucl .
methods phys .
a. * fig.7 * distributions of electron candidates after selection .
the numbers are normalized with integrated luminosity @xmath1 .
the hatched histogram is for sm background , and the unhatched one is background folded with 430gev sneutrino signal . * fig.8 * discovery(sb=5 , solid line ) and exclusion(sb=2 , dot line ) limits on sneutrino mass parameter as a function of the tevatron integrated luminosity
. the region between these two lines are those which can not be excluded or discovered explicitly . | we investigated the lepton flavor violation processes @xmath0 induced by r - parity violating interactions at the tevatron hadron collider .
the theoretical calculation and monte carlo simulation demonstrate that with a set of suitable cuts on experimental observables , one might be capable to reduce the standard model physical background to a controllable level so that the signals of r - parity violating interactions could be detected distinctively .
furthermore , clear sneutrino information could be abstracted from the purified event sample where other susy scalar quark pollution is heavy suppressed .
we conclude that with a reasonable assumption of @xmath1 integrated luminosity , the experiments at the tevatron machine would have potential to discover sneutrino in the region of @xmath2 via lepton flavor violation @xmath3 production channels , or extend the mass scale constraint up to @xmath4 at 95% cl . *
pacs : 11.30.fs , 11.30.pb , 12.60.jv , 14.80.ly * |
formaldehyde , h@xmath6co , is a poisonous molecule in the aldehyde group . on earth
it plays a part in troposphere chemistry dynamics as the main source of oh via photo - dissociation and is formed from photo - oxidation in the atmosphere or through the incomplete burning of biomass @xcite .
traces of formaldehyde have tentatively been detected in the martian atmosphere @xcite where it is believed to be derived from the oxidation of methene ( c@xmath7h@xmath8 ) @xcite .
formaldehyde was the first polyatomic molecule to be detected in the interseller medium ( ism ) @xcite and is extremely abundant @xcite .
this has made it useful in investigating the isotope composition of carbon in the galaxy @xcite .
the proposed mechanism of production is via the successive hydrogenation of co @xcite on icy grain mantles : @xmath9 further hydrogenation produces methanol through an intermediate methyl radical h+h@xmath7co@xmath10 ch@xmath11o@xmath10 h+ ch@xmath11o @xmath10 ch@xmath11oh .
common reactions include that with ammonia which produces amines @xcite and polymerisation with other h@xmath7co molecules . as a result
, formaldehyde is believed to be the major precursor for the formation of complex organic molecules in the ism that include interstellar glycolaldehyde @xcite and amino acids @xcite .
formaldehyde s astrophysical relevance does not end in the ism .
recently , it has been detected in comets @xcite , such as 103p / hartley 2 @xcite , c/2007 n3@xcite and hale - bopp @xcite , where it is thought to originate from the degredation of polyoxymethylene @xcite .
it is also present in protoplanetary discs around low mass young stars ( taurus - auruga class i / ii ) @xcite as circumstellar ice with an abundance ratio of @xmath12 compared to the more ubiquitous water - ice .
because of h@xmath7co s role as a precursor to complex organic molecules , it is considered a possible biomarker .
the rna world hypothesis suggests an early earth with a co@xmath7 , h@xmath7o and n@xmath7 rich atmosphere @xcite . illuminating this mix with ultraviolet ( uv )
radiation should lead to a large amount of formaldehyde being fixed in the atmosphere before being deposited into the prebiotic oceans @xcite .
alternatively , the source of prebiotic chemical compounds may be derived without need of illuminating uv radiation via glancing icy body impacts @xcite .
such impacts would produce shock - compression conditions that lead to formation of hcn molecules .
these hcn molecules can be hydrolyzed to form formaldehyde and from there produce amino acids .
thus a planet rich in formaldehyde may indicate one undergoing the stages of pre - life .
finally , formaldehyde masers @xcite are a reliable and proven tracer for high - density environments such as star - forming regions in galaxies due to its ubiquity and large number of long wavelength transitions @xcite .
currently , there are 19 extragalactic sources @xcite of these masers including iras 18566 + 0408 , which is notable for detection of the first h@xmath7co maser flare @xcite .
formaldehyde masers ( and maser flares ) have mostly been observed via the @xmath13 and @xmath14 @xmath15-doublet transitions at 6.1 cm and 2.2 cm respectively
. the wide - range of interactions in atmospheric , terrestrial , astrophysical and astrobiological phenomena makes formaldehyde a relevant molecule in the chemistry of exoplanets and their atmospheres .
therefore a complete , high - resolution , line list for h@xmath7co should provide an important aid for characterisation and modelling of formaldehyde .
these considerations led us to study formaldehyde as part of the exomol project @xcite , which aims to produce comprehensive molecular line lists for studies of the atmospheres of exoplanets and cool stars .
high - resolution , room - temperature formaldehyde spectra have been well - studied in the laboratory @xcite ; the early work was reviewed by @xcite . currently , the major source of publicly available spectroscopic data on h@xmath7co is the hitran database @xcite which has recently been updated to include long - wavelength data from the cdms database @xcite .
the spectral regions covered in the database are 0 100 @xmath4 , 1600 1800 @xmath4 @xcite and the 2500 3100 @xmath4 @xcite at up to 10@xmath16 cm / molecule sensitivity for @xmath17=296 k. however , this accounts for only 40 000 transitions extending up to @xmath18 and covers only four of the six fundamental vibrational bands as well as the ground state rotational spectrum .
this deficiency arises from an apparent lack of absolute intensities in the 100 1600 @xmath4 range .
additional observed transitions are available @xcite and include line positions @xcite , and intensities @xcite of some of the fundamental bands and hot bands @xcite .
the incompleteness and low rotational excitations available in hitran limits the applicability of this data for temperatures above 300 k. the theoretical spectra presented in this paper aims to provide a more complete and accurate picture of the spectra of formaldehyde up to 10 000 @xmath4 and for temperatures up to 1500 k. our line list should therefore be useful for modelling higher temperature environments as well as studies on non - lte transitions such as those observed in masers . theoretically , electric dipole transition intensities of h@xmath7co were studied by @xcite and @xcite ; see also the review by @xcite .
@xcite used an _ ab initio _ mp2/6 - 311 g * * dms to simulate the photoacoustic spectrum of high c - h stretching overtones of h@xmath7co .
@xcite generated an _ ab initio _
couple - cluster ccsd(t)/aug - cc - pvtz dipole moment surface ( dms ) for h@xmath7co ; they used an effective charges representation to compute ( relative ) rovibrational line intensities for h@xmath7co reproducing the hitran data @xcite with reasonable agreement .
@xcite computed an _ ab initio _
dms using the qcisd/6 - 31111g(d , p ) level of theory and presented it as an expansion . despite these works
there is no comprehensive line list for formaldehyde available in the literature .
the goal of this work is to bridge this gap .
we use the variational program trove @xcite in conjunction with an initial potential energy surface ( pes ) obtained ` spectroscopically ' by @xcite and a new _ ab initio _
dipole moment surface ( dms ) for formaldehyde and generate an extensive line list for h@xmath0c@xmath1o applicable for the temperatures up to @xmath2 k. in the following h@xmath6co and formaldehyde will refer to the main isotopologue h@xmath0c@xmath1o .
h@xmath7co is a prolate asymmetric top molecule that belongs to the c@xmath19 molecular symmetry group @xcite .
the group has four irreducible representations @xmath20 , @xmath21 , @xmath22 and @xmath23 .
once the h atom nuclear spin is taken into account the ` para ' @xmath3 representations are singly degenerate and the ortho @xmath24 representations are triply degenerate . as h@xmath7co has four atoms , it has six vibrational modes ; table [ tab : vibcoord ] shows the vibrational modes and their corresponding symmetries , band centers and descriptions .
coriolis interactions occur strongly between the @xmath25 and @xmath26 modes , and weakly between the @xmath27 and @xmath25 modes @xcite which couples their energy levels and wavefunctions .
this manifests itself in the @xmath27 , @xmath25 and @xmath26 mode interaction as overlapping bands which make these three bands difficult to distinguish from each other .
c*3l mode & band centers & symmetry & description + @xmath28 & 2782.46 & @xmath20 & symmetric c - h stretching + @xmath29 & 1746.01 & @xmath20 & c - o stretching + @xmath27 & 1500.18 & @xmath20 & symmetric o - c - h bending + @xmath25 & 1167.26 & @xmath22 & out - of - plane bending + @xmath30 & 2843.33 & @xmath23 & asymmetric c - h stretching + @xmath26 & 1249.10 & @xmath23 & asymmetric o - c - h bending + full details of their pes calculation are given by @xcite , so only a brief summary is presented here .
the initial pes was computed _
ab initio _ using the ccsd(t)/aug - cc - pvqz theory .
variational calculations with this surface give a root - mean - square ( rms ) error of 5.1 @xmath4 for the fundamental band centers .
@xcite refined the _ ab initio _
pes using a @xmath31 formulation where @xmath32 , @xmath33 and @xmath34 are the original _ ab initio _ ,
_ correction _ and _ refined _ pes respectively .
the eigenfunctions of the original _ ab initio _
hamiltonian , @xmath35 , are used as the basis functions for the extended hamiltonian @xmath36 where @xmath37 is typically small and almost diagonal correction .
@xmath33 is expanded in taylor series and expansion coefficients are obtained in a variational least - squares fit to a high - resolution spectroscopic data using trove .
the rms error against experimental energy levels with @xmath38 of this semi - empirical pes , called h2co-2011 , is 0.04 @xmath4 .
it should be noted however that the excellent accuracy of the refined pes caused serious problems with the absolute intensities of some bands .
the intensities based on the initial , less accurate _ ab initio _ pes , the intensities of the @xmath39 , @xmath40 and @xmath41 bands were observed to agree much better with the experiment .
figure [ fig : comp_pes](a ) highlights this effect , where an order of magnitude difference in absolute intensity and cross - section was observed .
initially blamed on the dms , it was soon discovered that the original _ ab initio _
pes did not have this problem . in order to address this issue
we have repeated the refinement process making it less aggressive with careful observation of the transition moments . in the present work
we have also increased the size of the basis set ( see discussion below ) .
we used the same fitting set of experimental energies as well as the same functional form for pes .
the rms deviation of the experimental term values used in the pes fit ( @xmath42 ) against the computed energies in ayty is 0.18 @xmath4 ( 0.006 @xmath4 for pure rotational term values ) .
the potential parameters as well as the associated fortran 95 program are given as supplementary material .
the resulting line intensities have returned to quality of the _ ab initio _ levels as can seen in figure [ fig : comp_pes](b ) .
= 15.0 cm the trove program suite @xcite is employed to compute our formaldehyde line list as well to perform the least - squares fit of the _ ab initio _
pes discussed above .
trove is designed to compute variational ro - vibrational energy levels , associated eigenfunctions and transition intensities for molecules of arbitrary structure .
variational methods are often limited in their efficiency due to the need to diagonalize increasingly large hamiltonian matrices as the complexity of the molecule increases .
however with the improved computational power and parallelism of modern cpus , it is now feasible to solve ro - vibrational schrdinger equations for polyatomic molecules .
trove has been successfully used to produce high - accuracy line lists and spectra for tetratomic molecules such as hsoh @xcite , nh@xmath11 @xcite , so@xmath11 @xcite , ph@xmath11 @xcite , and sbh@xmath11 @xcite , as well as recently the pentatomic molecule ch@xmath43 @xcite . trove approximates the kinetic energy operator ( keo ) by a truncated taylor series expansion in generalized coordinates .
previous calculations for h@xmath7o@xmath7 by @xcite show that computations using series representation of the keo for non - linear molecules , such as h@xmath7co , converge as well as exact keo - based programs such as wavr4 @xcite and , particularly for excited rotational states , at cost of significantly less computational time . in this work
we use a kinetic expansion order of @xmath44 for optimal convergence while providing reasonable computation times and similarly , we use a potential energy expansion order of @xmath45 . in trove , the primitive vibrational basis set is represented by a symmetrized product of six one - dimensional vibrational functions @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath50 , and @xmath51 , where @xmath52 denotes the associated local mode vibrational quanta , @xmath53 are linearized versions @xcite of the coordinates @xmath54 , and @xmath55 , respectively , and @xmath56 is the dihedral angle between the och@xmath57 and och@xmath6 planes .
the functions @xmath58 are obtained by solving the corresponding 1d schrdinger equation @xcite for the vibrational motion associated with the corresponding coordinate @xmath59 @xmath60 @xmath61 , with the other coordinates held fixed at their equilibrium values , where the numerov - cooley method @xcite is used .
the direct product of the 1d basis functions is contracted using the polyad condition : @xmath62 which in terms of the normal mode quantum numbers @xmath63 reads @xmath64 this polyad rule is based on the approximate relationship between the h@xmath7co fundamental frequencies ( see table [ tab : vibcoord ] ) : @xmath65 the vibrational basis set is further optimized by solving four reduced eigen - problems variationally for @xmath66 , @xmath67 , @xmath68 , and @xmath69 to produce four sets of wavefunctions @xmath70 , @xmath71 , @xmath72 , and @xmath73 , respectively . at the step
2 the pure vibrational ( @xmath74 ) problem is solved variationally using the basis set constructed as a symmetrized direct product of @xmath75 ( @xmath76 ) and @xmath77 ( @xmath78 or @xmath79 ) contracted through the polyad number condition ( [ eq : polyad ] ) and symmetrized according to the c@xmath19(m ) molecular symmetry group . in this work
the basis set is truncated at @xmath80 as the relative simplicity of the molecule means that this gives well - converged results . the maximum polyad number @xmath81 restricts the number of combinations of @xmath82 and @xmath83 for which @xmath84 .
the resulting eigenfunctions @xmath85 obtained for each c@xmath19(m ) symmetry @xmath86 and @xmath87 together with the symmetrized rigid rotor wavefunctions @xmath88 form our @xmath74 basis set representation @xcite , where the ro - vibrational basis functions are given as a direct product of @xmath85 and @xmath88 . here @xmath89 is the rotational parity defined by @xcite , and @xmath15 is the projection of the angular momentum on the body - fixed axis @xmath90 .
the latter is defined according with the eckart conditions @xcite and is oriented approximately along the co bond . in c@xmath19(m ) symmetry , @xmath15 and @xmath89 correlate with the customary @xmath91 and @xmath92 rotational quantum numbers as @xmath93 the vibrational part of the @xmath74 basis set is truncated using the energy threshold of @xmath94 cm^-1^and thus consists of 2310 , 1531 , 1688 , and 2112 functions for the @xmath95 and @xmath87 symmetries , respectively .
the resulting ro - vibrational hamiltonian matrix in the @xmath74 representation exhibits a block diagonal structure where each of four blocks represents an irreducible representation @xmath95 or @xmath87 and can be diagonalised independently .
each of these blocks displays a band - diagonal structure whose bandwidth and length is determined by the @xmath74 basis set size and the level of rotational excitation respectively . in generating our line list
we employed an upper eigenvalue limit of 18 000 @xmath4 as the intensity of transitions involving higher energy states are too weak to be important .
the @xmath74 matrix blocks produced by trove were on average dimensions @xmath96 @xmath97 @xmath96 .
the rule of thumb for the average size of a block for @xmath98 is @xmath99 .
the largest @xmath100 computed was @xmath5 which required the diagonalisation of matrices in the order of @xmath101 for eigenvalues and eigenvectors .
the linear algebra libraries lapack @xcite and scalapack @xcite were employed to solve for the eigenvalues and eigenvectors .
intensity computation requires high quality electric dms .
we use an _ ab initio _
dms computed at the ccsd(t)/aug - cc - pvqz level of theory in the frozen - core approximation using cfour @xcite .
three symmetry - adapted projections of the dipole moment cartesian components , @xmath102 , @xmath103 , and @xmath104 , are given in the analytical representations with each component expanded in taylor series ( 185 parameters in total ) in terms of internal coordinates around the equilibrium configuration using the form developed by @xcite to represent the dipole moment of h@xmath7cs .
these parameters reproduce the _ ab initio_dipole moment values of the @xmath102 , @xmath103 , and @xmath105 components with rms errors of 0.0002 debyes for each component .
the equilibrium value of our dipole moment is 2.3778 d ( at @xmath106 = 1.2033742 , @xmath107 = 1.10377 , @xmath108 = 121.844@xmath109 ) , which can be compared to the experimental value of the ground vibrational state dipole moment of @xmath110=2.3321(5 ) d measured by @xcite .
the eigenvectors , obtained by diagonalization , are used in conjunction with the dms to compute the required linestrengths ( and from that the einstein-@xmath3 coefficients and absolute intensities ) of transitions that satisfy the rotational selection rules @xmath111 and the symmetry selection rules @xmath112 the einstein-@xmath3 coefficient for a particular transition from the _ initial _ state @xmath113 to the _ final _ state @xmath114 is given by : @xmath115 where @xmath116 is the rotational quantum number for the initial state , @xmath117 is planck s constant , @xmath118 is the transition frequency ( @xmath119 ) , @xmath120 and @xmath121 represent the eigenfunctions of the final and initial states respectively , @xmath122 is the electronically averaged component of the dipole moment along the space - fixed axis @xmath123 ( see also @xcite ) . from this
the absolute absorption intensity is determined by : @xmath124 , \label{eq : intens}\ ] ] where @xmath125 is the boltzmann constant , @xmath17 the absolute temperature and @xmath126 is the nuclear spin statistical weight factor .
@xmath127 , the partition function , is given by : @xmath128 where @xmath129 is the degeneracy of a particular state @xmath113 with energy @xmath130 .
for h@xmath7co , @xmath129 is @xmath131 with @xmath132 for @xmath20 and @xmath21 symmetries and @xmath133 for @xmath22 and @xmath23 symmetries .
the transitions were computed using the energy limits @xmath134 8000 and @xmath134 18000 @xmath4 for the lower and upper states , respectively .
although diagonalisation of the hamiltonian matrices is very demanding on computer resources , it is the calculation of the einstein-@xmath3 coefficients which dominates the actual computer time due the sheer number of these and the large size of the eigenvectors .
graphics processing units ( gpu ) were therefore employed to accelerate computation of the intensities . to do this required the development of a new algorithm to allow these fast but memory poor processors to be used efficiently . a paper discussing this
will be published elsewhere @xcite .
the line list produced , which we call ayty , contains around 10 billion transitions with wavenumbers up to 10 000 @xmath4 .
the transitions are sorted in increasing transition frequency and then converted into the exomol format @xcite .
an extract of the state file and transition file can be seen in tables [ tab : levels ] and [ tab : trans ] .
spectra at arbitrary temperatures can be computed using the einstein-@xmath3 coefficients from the transition files .
the theoretical error as estimated by the fitting rms deviation of 0.18 @xmath4 .
this means our transition frequencies and energy levels should be reliable to about 0.2 @xmath135 with low - lying levels , particularly the pure rotational ones , being much more accurate than this and levels for vibrational states for which there are no available laboratory data much less so .
.extract from the h@xmath6co state file .
the full table is available at http://cdsarc.u - strasbg.fr / cgi - bin / vizier ?-
source = j / mnras/. [ cols="<,>,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] [ tab : missingbandv6v4v3 ]
we have computed the frequency and einstein-@xmath3 coefficients of almost 10 billion transitions of formaldehyde , which cover wavelengths longer that 1 @xmath110 m and includes all rotational excitations up to @xmath5 , making the line list applicable for temperatures up to 1500 k. the ayty line list gives a room - temperature spectrum in excellent agreement with available experimental data .
we have highlighted those regions missing from the hitran database with the hope that they will be investigated further experimentally .
the new line list may be accessed via www.exomol.com or http://cdsarc.u - strasbg.fr / viz - bin / qcat?j / mnras/. the cross - sections of h@xmath7co can be also generated at www.exomol.com as described by @xcite .
this work was supported by the erc under the advanced investigator project 267219 and made use of the dirac@darwin , dirac@cosmos hpc cluster and emerald cfi cluster .
dirac is the uk hpc facility for particle physics , astrophysics and cosmology and is supported by stfc and bis .
the authors would like to acknowledge the work presented here made use of the emerald high performance computing facility provided via the centre for innovation ( cfi ) .
the cfi is formed from the universities of bristol , oxford , southampton and ucl in partnership with stfc rutherford appleton laboratory .
we thank clara sousa - silva and duncan a. little for help during the writing of this paper , afa would also like to thank dr .
faris n. al - refaie , lamya ali , sarfraz ahmed aziz , and rory and annie gleeson for their support .
tennyson j. , hill c. , yurchenko s. n. , 2013 , in aip conference proceedings , vol .
1545 , 6@xmath136 international conference on atomic and molecular data and their applications icamdata-2012 , aip , new york , pp . | a computed line list for formaldehyde , h@xmath0c@xmath1o , applicable to temperatures up to @xmath2 k is presented .
an empirical potential energy and _ ab initio _
dipole moment surfaces are used as the input to nuclear motion program trove .
the resulting line list , referred to as _
ayty _ , contains 10.3 million rotational - vibrational states and around 10 billion transition frequencies .
each transition includes associated einstein-@xmath3 coefficients and absolute transition intensities , for wavenumbers below 10 000 @xmath4 and rotational excitations up to @xmath5 .
room - temperature spectra are compared with laboratory measurements and data currently available in the hitran database .
these spectra show excellent agreement with experimental spectra and highlight the gaps and limitations of the hitran data .
the full line list is available from the cds database as well as at www.exomol.com .
[ firstpage ] |
the observation of a significant triangular flow parameter @xmath0 of charged hadrons is considered one of the most interesting results from recent heavy ion experiments at the relativistic heavy ion collider ( rhic ) and at the large hadron collider ( lhc ) energies @xcite .
this large value is attributed to the fluctuations in the initial qcd matter density distribution @xcite . as thermal emission of photons
is known to be sensitive to the initial temperature of the system , the study of photon @xmath0 can provide valuable information about initial state fluctuation and and its evolution complementary to hadronic observables .
hydrodynamic model calculation with event - by - event ( e - by - e ) fluctuating initial conditions ( ic ) explain the data for hadronic spectra and elliptic flow well upto a large @xmath2 even for most central collisions @xcite .
it has also been shown that fluctuations in the ic enhance the production of thermal photons significantly for @xmath6 2 gev/@xmath4 compared to a smooth initial - state - averaged profile and a better approximation of the direct photon data is obtained in that @xmath2 region @xcite . in a recent study
we have demonstrated that e - by - e fluctuating ic result in much larger elliptic flow of thermal photons than a smooth ic for @xmath7 2 gev/@xmath4 both at rhic and at the lhc energies @xcite .
however , we see that the effect of ic fluctuations in not sufficient to explain the direct photon @xmath1 data and our elliptic flow results still underestimate the data by a large margin @xcite . since a nonzero @xmath0 arises only if there are fluctuations in the initial matter density distribution , photon @xmath0 can provide information about the fluctuating initial state more directly than the elliptic flow parameter @xcite .
we calculate thermal photon @xmath0 for 040% central collisions of pb nuclei at lhc using the ( 2 + 1 ) dimensional e - by - e hydrodynamic framework developed in @xcite . this ideal hydrodynamic model with fluctuating ic
is in good agreement with charged particle spectra and elliptic flow @xcite as well as spectra and elliptic flow of thermal photons at rhic and at the lhc energies @xcite .
we use state of the art photon rates ( complete leading order plasma rates from @xcite and hadronic photons from @xcite ) integrated over the hydrodynamical evolution to calculate the spectra and the anisotropic flow parameters . the initial formation time @xmath8 of the plasma is assumed to be 0.14 fm / c and the fluctuation size parameter @xmath9 as 0.4 fm ( see @xcite and references therein for details ) .
spectra from smooth ( sic ) and fluctuating ( fic ) initial conditions along with alice direct photon data @xcite for 040% central collision of pb nuclei at lhc .
@xmath2 spectra calculated using leading order ( lo ) plasma rates @xcite , next - to - leading order ( nlo ) plasma rates @xcite and nlo pqcd photons are also plotted for comparison [ see @xcite].,width=491 ] figure [ spec_lhc ] shows thermal photon spectra from e - by - e hydrodynamics along with alice direct photon data @xcite for 040% central collisions of pb nuclei at lhc @xcite .
similar to rhic , the @xmath2 spectrum from fluctuating initial conditions ( fic ) is found to be much harder than the result obtained from smooth initial conditions ( sic ) in the region @xmath6 2 gev/@xmath4 .
thermal photons from fluctuating ic along with next - to - leading order ( nlo ) pqcd photons ( which starts dominating the direct photon spectrum for @xmath10 3.5 gev/@xmath4 ) explain the alice data well for @xmath11 gev/@xmath4 .
however , for @xmath12 2 gev/@xmath4 our result underestimates the data which can only be explained if the contribution from the hadronic phase is increased by an order of magnitude .
we see that the inclusion of the nlo plasma rates @xcite to our calculation enhance the production of thermal photons by a factor of about 1020% in the region @xmath12 2 gev/@xmath4 . for @xmath7 2 gev/@xmath4 however , the difference between the complete lo and nlo results is not significant as shown in figure [ spec_lhc ] .
in addition , e - by - e calculation of anisotropic flow parameter using the nlo plasma rate is numerically expensive process .
hence , in practice we calculate thermal photon @xmath0 using the complete lo plasma rates @xcite which is an excellent approximation .
the triangular flow parameter @xmath0 of thermal photons for 040% central collisions of lead nuclei at 2.76a tev at lhc is shown in figure [ v3 ] .
@xmath0 is calculated with respect to the participant plane ( pp ) by taking average over a sufficiently large number of events using the relation @xmath13 here the participant plane angle @xmath14 is defined as @xmath15 @xmath16 is the energy density , @xmath17 , and @xmath18 is the azimuthal angle in the above equation . the initial triangular eccentricity of the matter density is calculated using @xmath19 \varepsilon \left ( x , y,\tau_{0}\right ) } { \int \mathrm{d } x \mathrm{d } y \ ; r^{3 } \varepsilon \left ( x , y,\tau _ { 0}\right ) } \ , .\ ] ] with respect to participant plane ( pp ) and reaction plane ( rp ) for 2.76a tev pb+pb collisions at lhc and for 040% centrality bin along with photon @xmath1 results @xcite . ]
@xmath0(pp ) and the reaction plane ( rp ) @xmath0 are shown in figure [ v3 ] along with the elliptic flow parameter @xmath1 of thermal photons for comparison .
we see a large value of @xmath0(pp ) for @xmath12 6 gev/@xmath4 compared with the elliptic flow results .
in addition , @xmath0(pp ) shows similar qualitative nature to the photon @xmath1 where it is small at low @xmath2 , then rises with @xmath2 , peaks around 23 gev/@xmath4 , and then drops for larger values of @xmath2 .
it is well known that the average collision geometry plays a significant role in determining @xmath1 and thus one finds large elliptic flow even with smooth initial conditions and with respect to the reaction plane angle .
however , @xmath0 ( rp ) from individual events is found to be both positive and negative as @xmath5 is uncorrelated with the rp , and thus after averaging over a large number of events we find @xmath0(rp ) close to zero . in case of hadrons ,
the magnitude of the initial triangular eccentricity @xmath5 is the main determining factor for the size of the final state @xmath0 .
however this is not true for photons . from a study of individual events
we see that events with larger early time values of transverse flow velocity ( @xmath20 ) correlate best with @xmath0 for thermal photons and that conversely the magnitude of @xmath0 does not strongly correlate with the value of @xmath5 .
we conclude that fast @xmath20 builtup driven by fluctuations dominates over the initial geometry effects in determining the @xmath0 of thermal photons ( see @xcite for details ) .
the photon @xmath0 results are expected to be sensitive to the initial parameters of the model calculation and we study this dependence by changing @xmath9 and @xmath8 from their default values
. a larger @xmath9 results in a smaller triangular flow whereas a larger @xmath8 gives larger @xmath0 especially in the region @xmath6 2 gev/@xmath4 .
the initial density distribution becomes smoother for larger @xmath9 and we get smaller @xmath0 . on the other hand for larger @xmath8 , a large fraction of the high @xmath2 photons with smaller flow velocities are not included in the calculation and we get a much larger @xmath0 at higher @xmath2 ( at the expense of significantly reducing the total yield ) .
we see a positive and substantial value of the triangular flow parameter @xmath0 of thermal photons at lhc from e - by - e ideal hydrodynamic model .
@xmath0(pp ) is found to be about half of the value of @xmath1(pp ) at @xmath2= 1 gev/@xmath4 .
however , for @xmath21 gev/@xmath4 the difference between these two anisotropy parameters is less than 25% .
we know that for hadronic @xmath0 the initial triangular geometry of the overlapping zone dominates over the local fluctuations , however this is not true for photon @xmath0 .
we conclude from a study of individual events that unlike in the case of hadron @xmath0 where the magnitue of @xmath5 is a good predictor for the observed value , photon @xmath0 is driven by a different dynamics having to do with fast buildup of transverse flow due to fluctuations .
rc gratefully acknowledge he financial support by the dr .
k. s. krishnan research associateship from variable energy cyclotron centre , department of atomic energy , government of india .
tr is supported by the academy researcher program of the academy of finland , project no .
we acknowledge the computer facility of the csc computer centre , espoo . | we calculate the triangular flow parameter @xmath0 of thermal photons for 040% central collisions of pb nuclei at lhc using an event - by - event hydrodynamic model with fluctuating initial conditions .
thermal photon @xmath0 with respect to the the participant plane angle is found to be positive and significant compared to the elliptic flow parameter @xmath1 of thermal photons .
in addition , photon @xmath0 as a function of @xmath2 shows similar qualitative nature to photon @xmath1 in the region @xmath3 gev/@xmath4 .
we argue that while @xmath0 originates from @xmath5 deformations of the initial state density distribution , fast buildup of radial flow due to fluctuations is the main driving mechanism for the observed large value . triangular flow , event - by - event hydrodynamics , thermal photons |
direct detection is a crucial next step in our understanding of extrasolar planets .
although the radial velocity technique continues to generate valuable discoveries ( marcy & butler 2000 ) , the composition , origins , and essential natures of exoplanets will likely remain hidden unless we can study them directly . to date , we know of over 100 exoplanets , all with semimajor axes @xmath1 6 au . using the radial velocity technique , detection of exoplanets with orbital radii similar to those of saturn , uranus , and
neptune would require timelines on the order of decades . however , the existence of these outer planets , theories of planetary dynamics ( e.g. rasio & ford , 1996 ) , and the shaping of dusty debris disks ( holland et al 1998 ; greaves et al . 1998 ; liou & zook 1999 , ozernoy et al .
2000 ) all lend credence to the idea of planets with semimajor axes of tens to hundreds of au .
the primary requisites of a workable strategy for detection of extrasolar planets are high - contrast imaging and subarcsecond resolution .
these capabilities are available via the use of ground - based adaptive optics ( ao ) systems or direct imaging from space .
a choice of ao in turn mandates the use of large telescopes , since the time required to achieve a given s / n ratio varies as the inverse fourth power of mirror diameter ( olivier et al . 1997 ) .
yet even with keck ao , planets with semi - major axes of order tens of au can be spatially resolved only around stars within @xmath2 60 pc of earth .
current ao sensitivities also rule out the detection of solar - age planets in reflected light , however , young ( @xmath1 60 myr ) , warm planets may be seen via their near - infrared emission .
selection of young targets is effected chiefly through the study of galactic space motions ( u , v , w ) in conjunction with excess emissions of stars .
these may appear as x - rays , h@xmath3 and ca ii h & k emission , or far - infrared excesses ( e.g. spangler et al . 2001 ) .
lithium absorption may also be an important youth indicator .
techniques for identifying young stars are further described in zuckerman et al .
( 2001 ) .
once targets have been identified and observed , registration of long - exposure images and a process called unsharp - masking increase our sensitivity to point sources . unsharp - masking is a process in which a smoothed version of an image is subtracted from the original so as to improve the contrast between the background noise and small - scale structure .
as can be seen in the next section , unsharp - masking improves our sensitivity to point sources by as much as two magnitudes within 2 of a target star .
our study utilizes the ao system on keck ii ( wizinowich et al .
2000 ) . over the past two years
, we have observed @xmath2 100 nearby , young stars using kcam , nirspec s slit - view camera ( scam ) , and most recently nirc2 , the first camera optimized for keck
we estimate that more than half of our target stars are @xmath1 100 myr old .
figure 1 shows sensitivity calculations for representative targets from the scam and nirc2 datasets .
as expected , sensitivity improves both for nirc2 data and for images which have been unsharp - masked .
llllllll + + star & d & age & strehl & band & 2 m@xmath4 & 5 m@xmath4 & 10 m@xmath0 + & ( pc ) & ( myr ) & & & r ( au ) & r ( au ) & r ( au ) + + + gj803 & 10 & 12 & .15 & h & 9 & 5 & 3 + hd207129 & 16 & 30 & .10 & h & 32 & 16 & 10 + hip 560 & 39 & 12 & .40 & k&78 & 35 & 23 + hip82587 & 30 & 30 & .20 & h & n / d & 37 & 25 + hd159911 & @xmath260 & 30 & .40 & k & n / d & 54 & 30 + + using the age of gj 803 , we derived orbital separations at which this study would detect planets of 2 , 5 , and 10 jupiter masses . by scaling with respect to star age and data quality ( measured by strehl ratio ) , we may extend this result to other target stars ( table 1 ) .
our sensitivity to planets of a given mass varies with such factors as age , distance , and magnitude of the host star , as well as conditions under which the data were taken .
however , for the best stars in our sample , of which gj 803 is an example , this survey is able to detect 1 m@xmath0 planets with orbital radii of only 19 au .
burrows a. , marley , m. , hubbard , w. b. , lunine , j. i. , guillot , t. , saumon , d. , freedman , r. , sudarsky , d. , sharp , c. 1997 , 491 , 856 greaves , j. s. , holland , w. s. , moriarty - schieven , g. , jenness , t. , dent , w. r. f. , zuckerman , b. , mccarthy , c. , webb , r. a. , butner , h. m. , gear , w. k. , walker , h. j. 1998 letters , 506 , 133 holland , w. s. , greaves , j. s. , zuckerman , b. , webb , r. a. , mccarthy , c. , coulson , i. m. , walther , d. m. , dent , w. r. f. , gear , w. k. , robson , i. 1998 nature , 392 , 788 liou , j. , zook , h.a .
1999 , , 118 , 580 mclean , i.s .
1998 , spie , 3354 , 566 marcy , g. w. , butler , r. p. 2000 , pasp , 112 , 137 olivier , s. s. , max , c. e. , brase , j. m. , gavel , d. t. , macintosh , b. , carrano , c. 1998 , in asp conference series 134 : brown dwarfs and extrasolar planets , eds .
rebolo , r. , martin , e.l . , zapatero osorio , m.r .
, ( san francisco : asp ) , 262 .
ozernoy l.m .
, gorkavyi n.n . , mather j.c .
, taidakova t.a .
2000 , letters , 537 , 147 spangler , c. , sargent , a. i. , silverstone , m. d. , becklin , e. e. , zuckerman , b. 2001 , , 555 , 932 rasio , f. a. , ford , e. b. 1996 , science , 274 , 954 wizinowich , p. , et al .
2000 , proc .
spie , 4007 , 64 .
zuckerman , b. , song , i. , bessell , m. s. , webb , r. a. 2001 , 562 , 87 | adaptive optics , large primary mirrors , and careful selection of target stars are the keys to ground - based imaging of extrasolar planets .
our near - ir survey is capable of identifying exoplanets of 1 - 10 m@xmath0 within 100 au of young ( t@xmath160 myr ) , nearby ( d@xmath160 pc ) stars .
for very young and proximate targets such as gj 803 ( 12 myr , 10 pc ) we are able to detect exoplanets with parameters approaching those of planets our solar system ( 1 m@xmath0 at 19 au , 2 m@xmath0 at 9 au ; 5 m@xmath0 at 5 au ) .
we have thus far imaged over 100 stars with keck ao . here
we report on our progress , discuss specific observing strategies , and present detailed sensitivity limits .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
the study of bilayer graphene ( blg ) is a focused area in the condensed - matter physics because of the potential application of blg to new electronic devices.@xcite one of the fundamental subjects is to explore the physics of the ground state of electrons in blg .
a number of experiments @xcite performed on high quality suspended blg samples have provided the evidence that the ground state is gapped at the charge neutrality point ( cnp ) . in particular , a recent experiment by velasco _ et al . _
@xcite has observed that ( i ) the ground state is insulating in the absence of external electric and magnetic fields , with a gap @xmath2 2 mev that can be closed by a perpendicular electric field of either polarization , ( ii ) the gap grows with increasing magnetic field @xmath0 as @xmath3 with @xmath4 1 mev and @xmath5 5.5 mevt@xmath6 , and ( iii ) the state is particle - hole asymmetric . on the other hand ,
theories have predicted various gapped states , such as a ferroelectric - layer asymmetric state @xcite or quantum valley hall state ( qvh),@xcite a layer - polarized antiferromagnetic state ( af),@xcite a quantum anomalous hall state ( qah),@xcite a quantum spin hall state ( qsh ) , @xcite and a superconducting state in coexistence with antiferromagnetism ( saf ) .
@xcite the ferroelectric - layer asymmetric and qah and qsh states all have been ruled out by the experiment.@xcite the saf state is excluded because the real system is an insulator . the af state can not reproduce the gap behavior with varying the magnetic field .
recently , the loop - current state has been studied by numerical diagonalization of an effective mean - field hamiltonian for a finite size lattice @xcite and by analytically solving a two - band continuum model ( 2bcm ) .
@xcite whether the model of this state agrees with the experimental observations on the electronic properties of blg remains a question . in this work , using the four - band continuum model ( 4bcm ) for electrons with finite - range repulsive interactions in blg , we study the ordered - current state ( ocs ) at the cnp with a rigorous formalism and compare the results with the experimental observations .
the importance of using the 4bcm to describe quantitatively the many - body properties of the electron liquid in the blg has been stressed by the existing works.@xcite we here investigate the gap behavior of the ocs with varying the magnetic field @xmath0 , and the particle - hole asymmetry spectra at finite @xmath0 , and the phase transitions in the electron system in the presence of the electric and magnetic fields .
we will show that the puzzles ( i)-(iii ) of the experimental observations can be resolved by the present model of the ocs . with the same 4bcm
, we also study the af state and show that the af state is not able to reproduce the experimental result for the gap as a function of the magnetic field .
the lattice structure of a blg is shown in fig .
1 . the unit cell of blg contains four atoms denoted as a@xmath7 and b@xmath7 on top layer , and a@xmath8 and b@xmath8 on bottom layer with interlayer distance @xmath9 .
the lattice constant defined as the distance between the nearest - neighbor ( nn ) atoms of a sublattice is @xmath10 .
the energies of intralayer nn [ between a@xmath7 ( a@xmath8 ) and b@xmath7 ( b@xmath8 ) ] and interlayer nn ( between b@xmath7 and a@xmath8 ) electron hopping are @xmath11 2.8 ev and @xmath12 0.39 ev , respectively . the first brillouin zone and the two valleys @xmath13 and @xmath14 in the momentum space
are depicted in fig .
2 . for the carrier concentration close to the cnp , we need to consider only the states with momenta close to the dirac points @xmath15 and @xmath16 .
we here define the operator @xmath17 , where @xmath18 or @xmath14 , @xmath19 creates a spin-@xmath20 electron of momentum @xmath21 in valley @xmath22 of @xmath23 sublattice , and @xmath21 is measured from the dirac point @xmath13 ( @xmath14 ) and confined to a circle @xmath24 in @xmath13 ( @xmath14 ) valley . with the operator @xmath25 , the hamiltonian describing the noninteracting electrons is given by @xmath26 with @xmath27 where @xmath28 , @xmath29 ( -1 ) for @xmath21 in the valley @xmath13 ( @xmath14 ) , and @xmath30 .
we hereafter use the units of @xmath31 = 1 and @xmath32 = 1 .
the interaction hamiltonian is @xmath33 where @xmath34 is the number deviation of electrons with spin @xmath20 from its average occupation at site @xmath35 of sublattice @xmath23 ( hereafter denoted as @xmath36 for short ) , @xmath37 , @xmath38 is the on - site interaction , and @xmath39 is the interaction between electrons at sites @xmath36 and @xmath40 . within the mean - field approximation ( mfa ) , since the interaction @xmath39 appears in the exchange self - energy , it can be considered as a finite - range interaction by taking into account the screening effect due to the electronic charge fluctuations.@xcite the total hamiltonian @xmath41 satisfies the particle - hole symmetry .
in the ordered - current state for which there is no antiferromagnetism , the effective interaction under the mfa is given by @xmath42 where the self - energy @xmath43 is defined by @xmath44\exp(i\vec k\cdot\vec d ) \label { mfa2}\end{aligned}\ ] ] and @xmath45 is the vector from the position @xmath36 to @xmath40 .
first , we consider the diagonal self - energy and denote @xmath46 by @xmath47 for brevity . now
, the function @xmath48 can be written as @xmath49 where the @xmath21 summation runs over the first brillouin zone , and @xmath50 is the total number of unit cells on single layer graphene .
note that the function @xmath51 in the integrand in eq .
( [ mfa3 ] ) is sizable only in areas close to the two dirac points .
figure 3 shows the typical behaviors of the two functions @xmath52 defined in the two valleys @xmath13 and @xmath53 , respectively .
the result in fig .
3 is obtained by the self - consistent solution to the ocs without external fields .
the functions are nonvanishing only within @xmath54 with @xmath32 as the lattice constant .
then , the @xmath21 integration in eq .
( [ mfa3 ] ) can be confined to two valleys .
since the range of the exchange interaction @xmath47 is finite due to the electronic charge - fluctuation screening , the phase @xmath55 in the factor @xmath56 can be safely approximated as @xmath57 .
therefore , we can write the formulas for @xmath58 and @xmath59 as @xmath60 where the @xmath21 summation is confined to a single valley , and the quantities @xmath61 and @xmath62 are defined by @xmath63 the quantity @xmath64 can be written as @xmath65 with @xmath66 as the average electron doping concentration on sublattice @xmath23 . for the doping concentration close to the cnp
, we need to consider only the low energy quasiparticles with momenta close to the dirac points . then by expanding the self - energy with respect to the momentum @xmath21 in the two valleys and taking only the leading terms ,
we get @xmath67 with @xmath68 physically , the imaginary part @xmath59 is proportional to a bond current .
all the bond currents in the lattice constitute to the current loops .
the existence of the bond currents breaks the time - reversal symmetry . in fig .
4 , we draw out some of the bond currents on the same sublattice connected to a given site @xmath35 . clearly , the total current density at site @xmath35 is zero .
next , we consider the quantity @xmath69 with @xmath70 .
for example , consider the case for @xmath23 = a@xmath7 and @xmath71 = b@xmath7 on the top layer .
suppose the quantity is not vanishing . as shown in fig .
5 , the bond currents all with a fixed bond length @xmath72 result in three kinds of hexagon current loops with positive , negative , and zero fluxes [ supposing the flux is positive ( negative ) for counterclockwise ( clockwise ) current loop ] , respectively . from the particle conservation law ,
the current along the boundary between the positive and the negative flux hexagons is two times of that along the boundary between the zero and the positive / negative flux ones .
the hexagon current loops imply not only the breaking of time - reversal symmetry but also the breaking of translational invariance ( homogeneity ) .
the breaking of translational invariance to a low symmetry state requires the relevant interaction strong enough .
note that there is no a common periodicity for the two kinds of hexagon current loops with different side length @xmath73 in the lattice .
the coexistence of the different hexagon current loops corresponds to completely an inhomogeneous system and can not be realized for the electrons with finite - range interactions .
the most favorable case is the smallest hexagon loops may exist when the interaction between the nn a@xmath7 and b@xmath7 atoms is strong enough .
the argument applies to all @xmath69 with @xmath70 . for
weak to medium interactions , we here assume all the currents between the sites of different sublattices are negligible small . on the other hand , the off - diagonal averages @xmath74 with @xmath70 can be pure real quantities without breaking homogeneity of the system .
the real quantities describe the electron hopping and renormalize the noninteracting hamiltonian .
we here assume that such renormalization has already been included in @xmath75 , we therefore do not take into account these hopping processes more again .
( in the presence of external electric or magnetic field , even if the renormalization depends on the field , we will neglect the field effect . )
we suppose @xmath76 and @xmath77 that means the breaking of the layer inversion symmetry . for the homogeneous system at the cnp
, we have @xmath78 . as a result ,
the effective mfa hamiltonian @xmath79 is obtained by adding the diagonal matrix diag(@xmath80 ) to @xmath81 : @xmath82 note that the matrices @xmath79 and @xmath83 are related by @xmath84 , where @xmath85 is a @xmath86 matrix @xmath87 if @xmath88 is an eigenfunction of @xmath79 with eigenvalue @xmath89 ( with @xmath90 = 1 , 2 , 3 , 4 ) , then @xmath91 is an eigenfunction of @xmath83 with the same eigenvalue .
therefore , the whole energy spectra can be obtained from the eigenstates only in a single valley . under the mfa and with the wave functions
@xmath88 s , the order parameters @xmath92 and @xmath93 are determined by @xmath94 where @xmath95 is the fermi distribution function , @xmath96 is the @xmath97th component of the eigenfunction @xmath88 , and @xmath98 is the total area of one layer . from the 2bcm,@xcite
we know that the valence and conduction bands are connected to the electronic motions in the a@xmath7 and b@xmath8 sublattices . therefore , the energy gap between the valence and conduction bands is determined by @xmath99 .
to reproduce the experimental data @xmath100 = 1 mev at the cnp , @xmath101 needs to be @xmath102 ev . supposing the effective interaction @xmath103 \end{aligned}\ ] ] ( decaying as @xmath104 , a typical behavior in the two - dimensional electron liquid @xcite ) with @xmath105 as the screening constant of high frequency limit of blg , we obtain the desired value @xmath106 with @xmath107 . another coupling constant is obtained as @xmath108 .
table i summaries all the parameters for the 4bcm .
.[tab]parameters for the 4bcm . [ cols="^,^,^,^,^,^,^",options="header " , ] in the presence of the magnetic field @xmath0 applied perpendicularly to the sample plane , we take the landau gauge for the vector potential , @xmath109 . with this gauge ,
the @xmath110 component momentum @xmath111 is a good quantum number . replacing the variable @xmath112 and the operator @xmath113 with the raising and lowering operators @xmath114 and @xmath32 , @xmath115 and @xmath116 , we can rewrite the effective hamiltonian in real space . at the @xmath13 valley ,
the hamiltonian is @xmath117 here @xmath0 is in the unit of @xmath118 t .
the @xmath13-valley eigenfunction @xmath119 is expressed as @xmath120 for @xmath121 2 , where @xmath122 is the @xmath1th level wave function of a harmonic oscillator of frequency @xmath123 and mass 1/2 centered at @xmath124 , and the superscript @xmath125 means the transpose of the vector .
the vector @xmath126 and the eigenenergy @xmath127 are determined by @xmath128 with @xmath129 the vector @xmath130 is normalized to unity . for each @xmath131 ,
the four energy levels appear at the valence , conduction , and other two bands about @xmath132 far from the zero energy , respectively .
for @xmath1 = 1 , there are only three states with @xmath133 and the other three components and eigenvalues are determined by the upper left 3@xmath1343 matrix of @xmath135 . for @xmath1 = 0 , we have only one state @xmath136 and @xmath137 . at the @xmath14 valley ,
the hamiltonian is @xmath138 since the hamiltonian has the symmetry @xmath139 , the eigenfunction @xmath140 is therefore given as @xmath141 with the same eigen value @xmath127 . in the presence of the magnetic field @xmath0 , the formulas determining the order parameters @xmath142 are different from eqs .
( [ pd1 ] ) and ( [ pd2 ] ) .
the @xmath21 summations in eqs .
( [ pd1 ] ) and ( [ pd2 ] ) are now replaced with the summations over @xmath111 and the landau index @xmath1 .
correspondingly , the wavefunction @xmath143 is replaced with @xmath144 with @xmath145 as the length of the blg in @xmath112 direction . by denoting the length in @xmath110 direction as @xmath146
, we have @xmath147 .
the @xmath111 summation is performed as @xmath148 where @xmath149-integral has been carried out using the normalization condition for the wave functions of the harmonic oscillator .
the equations for determining the order parameters are obtained as @xmath150 the solution to the landau levels at @xmath0 = 1 t is shown in fig .
6(a ) . only the levels in the conduction and valence bands are depicted . for @xmath1 = 1
, there is a level @xmath151 slightly above @xmath152 in the valence band .
there is no state in the conduction band for @xmath1 = 0 and 1 . only when @xmath131 , the level @xmath153 in the conduction band appears .
the energy gap is @xmath154 clearly , the particle - hole symmetry is no longer valid at finite @xmath0 , in agreement with the experiment.@xcite figure 6(b ) shows @xmath155 of the ocs as function of @xmath0 .
the af calculation of the same 4bcm ( see sec .
iv ) and experimental results for @xmath155 are also plotted for comparison . here , the only fitting parameter is @xmath101 for reproducing @xmath156 at @xmath0 = 0 .
the theoretical result for @xmath155 of the ocs as a function of @xmath0 is in surprisingly good agreement with the experiment.@xcite the above solution to the order parameters is only in the branch of @xmath157 . at weak magnetic field @xmath158
, there is another branch of @xmath159 . in this case , the two levels of @xmath1 = 0 and 1 appear in the conduction band but not in the valence band , and the energy gap is given by @xmath160 . in fig .
7 , we show the hysteresis curves for the ocs order parameters @xmath161 and @xmath162 and the gap @xmath163 . for @xmath164 t , there are two branches for @xmath155 . in the lower gap branch ,
the gap decreases with increasing @xmath165 .
this behavior of @xmath155 is in qualitative agreement with the experimental observation by weitz _
et al._@xcite that indicates two peaks in the electric conductance appearing at @xmath166 t ( where the real gap reaches the minimum ) , respectively .
in the af state , the magnetization at site @xmath167 is defined as @xmath168 where in the second equality we have used the facts that @xmath169 and the total number of up - spin electrons coincides with that of down - spin electrons .
the magnetizations in an unit cell are given by @xmath170 .
the order parameters are defined as @xmath171 . under the mfa ,
the interaction hamiltonian reads @xmath172 by supposing @xmath74 is real , the second term in right hand side of eq .
( [ mf1 ] ) then describes the electron hopping and is a renormalization of the noninteracting hamiltonian . as aforementioned ,
we suppose such a renormalization has already been included in the noninteracting hamiltonian ; we will not take into account this exchange effect again . with the mfa
, we obtain an effective hamiltonian as @xmath173 where @xmath20 = + 1 ( -1 ) for spin - up ( down ) electrons .
note the matrices @xmath174 and @xmath175 are related by @xmath176 if @xmath88 is an eigenfunction of @xmath177 with eigenvalue @xmath178 ( @xmath90 = 1,2,3,4 ) , then @xmath179 is an eigenfunction of @xmath180 with the same eigenvalue .
therefore , we need to find out only the eigenstates of up - spin electrons . using the property of the wave functions ,
we can obtain the equations for determining the order parameters . for @xmath92 , for example , we get @xmath181 here , the @xmath21 summation in the first line runs over the first brillouin zone , while it runs over a single valley in the second line ( because both valleys give the same contribution ) .
similarly , we obtain for @xmath93 , @xmath182 equations ( [ d1 ] ) and ( [ d2 ] ) for determining the af order parameters happen to be the same as eqs .
( [ pd1 ] ) and ( [ pd2 ] ) for the ocs order parameters by setting @xmath183 .
since the valence and conduction bands are connected to the electronic motions in the a@xmath7 and b@xmath8 sublattices , the energy gap between the valence and conduction bands is determined by @xmath99 . to reproduce the experimental data @xmath100 = 1 mev at the cnp , @xmath38 needs to be @xmath184 ev .
this value of u is larger than 9.3 ev of the recent _ ab initio _
calculation,@xcite which means the af state of @xmath185 ev can not reproduce the experimental data @xmath186 .
we now consider the behavior of the order parameters in the presence of the magnetic field @xmath0 applied perpendicularly to the blg plane . since the system under the magnetic field is not homogeneous , the hamiltonian can not be written in momentum space .
for low energy electrons , however , their overall momenta are close to the dirac points @xmath13 and @xmath14 .
we here formulate the problem by a different way . from the beginning , we write the electron operator @xmath187 as @xmath188 where @xmath189 is a fermion operator in valley @xmath190 separated from the fast phase factor @xmath191 $ ] and annihilates electrons of valley @xmath190 and spin @xmath20 at site @xmath167 of @xmath23 sublattice .
the operator @xmath189 weakly depends on coordinate @xmath167 . for later use
, we here define the operator @xmath192 where @xmath18 or @xmath14 is the valley index . in the presence of @xmath0 , as did in sce .
iii , we take the landau gauge for the vector potential @xmath193 and use the raising and lowering operators @xmath114 and @xmath32 .
we get the effective hamiltonian for af state as @xmath194 with @xmath195 for electrons at @xmath13 valley , and @xmath196 for electrons at @xmath14 valley .
the hamiltonian satisfies the transformation @xmath197 . as mentioned above
, we need to find out the eigenstates of up - spin electrons , @xmath198 for @xmath90 = 1,2,3,4 , and @xmath1 = 0 , 1 , @xmath199 . for each index @xmath1 ,
the four energy levels ( if they exist ) appear at the valence , conduction , and other two bands about @xmath132 far from the zero energy , respectively . at @xmath13 valley ,
the eigenfunction is given by @xmath200 for @xmath121 2 .
the vector @xmath201 and the eigenenergy are determined by @xmath202 with @xmath129 the vector @xmath130 is normalized to unity . for @xmath1 = 1
, there are only three states with @xmath203 and the other three components and eigenvalues are determined by the upper left 3@xmath1343 matrix of @xmath135 . for @xmath1 = 0 , we have only one state @xmath136 and @xmath137 .
note that this energy level is close to a level of @xmath204 . on the other hand , at valley @xmath14
, the eigenfunction is given by @xmath205 for @xmath131 .
the eigen equation reads @xmath206 with @xmath207 for @xmath1 = 1 , we have three states with @xmath208 and the other three components and the eigenvalues are determined by the lower right 3@xmath1343 matrix of @xmath209 . for @xmath1 = 0 ,
we have only @xmath210 and @xmath211 ( close to a level of @xmath212 ) . the order parameter @xmath92 is determined by @xmath213 where the first line is the definition ; the second line represents the averages in terms of the wave functions with @xmath214 as the @xmath97th component of @xmath215 , @xmath216 has been used for spin down electrons , and a factor @xmath217 , the area of the unit cell of one layer graphene , comes from the fact that @xmath218 is the probability density of electrons around site @xmath167 and the multiplication with @xmath217 gives rise to the probability of electrons in the cell at site @xmath167 ; in the last line , the @xmath111 summation is carried out according to eq .
( [ kysum ] ) .
analogously , the order parameter @xmath93 is determined by @xmath219 at the cnp and zero temperature , the order parameters @xmath92 and @xmath93 are self - consistently determined by eqs .
( [ egnk])-([ad2 ] ) . in fig
. 8 , we show the results for @xmath92 and @xmath93 at zero temperature as functions of the magnetic field @xmath0 and compare them with the experimental data for @xmath220 . clearly ,
even though @xmath92 and @xmath93 grow with increasing @xmath0 , their dependence of @xmath0 is not strong enough to match the experimental result.@xcite therefore we can not expect the af state as the candidate for the ground state of electrons in blg .
the landau levels @xmath221 of the af state at @xmath222 t are shown in fig .
9 . in different from the ocs ,
the distributions of the levels in the two valleys are now different .
especially , in the @xmath13 valley , there are no levels of @xmath1 = 0 and 1 in the conduction band ( for positive @xmath142 ) , while they appear in the conduction band but disappear in the valence band in the @xmath14 valley .
the energy gap is therefore the indirect gap @xmath223 .
as known , there is a momentum cutoff @xmath224 for the 4bcm .
the corresponding cutoff for the landau levels is given by @xmath225 . at small @xmath0
, @xmath226 is very large . for accelerating the numerical computation ,
we have used the super - high efficiency algorithm for sum of series.@xcite according to the algorithm , one needs to compute only a number of selected landau levels .
when an external electric field @xmath227 is applied perpendicularly to the blg plane , there is an effective potential difference @xmath228 between the two layers .
the hamiltonian @xmath79 for the ocs now is obtained by adding the diagonal matrix diag@xmath229 to @xmath81 . here
the terms @xmath230 appear because of the electric polarization by @xmath227 .
note that @xmath64 has the same sign of @xmath231 , and thereby @xmath232 , which means that the order parameters are even functions of @xmath231 .
the model shows that if @xmath227 closes the energy gap , then @xmath233 does it either . for the sake of illustration , we here consider the case of @xmath234 and @xmath235 .
the results for other cases can be deduced by the symmetry of the hamiltonian . for @xmath234 and @xmath235
, we still have two cases : @xmath236 and @xmath237 . here
, we consider the case of @xmath236 .
the discussion can be extended to the case of @xmath237 . at @xmath238 ,
the effective gap parameter is @xmath239 .
the positive voltage @xmath231 pushes this level from the valence band toward to the conduction band .
the critical potential @xmath240 closing the effective gap is obtained as @xmath241 mev .
the critical field of the experimental data @xcite is @xmath242 mv@xmath6 , which corresponds to @xmath243 mev ( using @xmath244 ) .
since the system satisfies the particle - hole symmetry at @xmath238 , we can take the chemical potential as zero for the system at the cnp .
then , the level @xmath245 is occupied if it is negative , otherwise it is empty .
therefore , with increasing @xmath231 from 0 , the system undergoes a phase transition at @xmath246 from the state with the level @xmath245 occupied to the state with the level empty .
thus , to search the critical @xmath240 where the gap closes at finite @xmath0 , we study the phase transition . at finite @xmath0 ,
a state at @xmath247 can be obtained by continuously changing the parameters @xmath0 and @xmath231 from the state at @xmath248 . if @xmath249 , then the level @xmath245 is empty .
note that @xmath245 is the only landau level of @xmath1 = 0 at finite @xmath0 and there is another level of @xmath1 = 1 close to it similarly as the case of @xmath250 .
so the two levels of @xmath1 = 0 and 1 in the @xmath13 valley keep empty on the path from @xmath248 to @xmath247 . on the other hand ,
if one starts from an initial state with @xmath251 , then the two levels of @xmath1 = 0 and 1 keep filled .
( we denote the filling number as @xmath252 and 1 for the two levels empty and filled , respectively . )
we thus have two states at @xmath247 . by comparing their energies , the ground state at @xmath247
is uniquely determined . at the critical potential @xmath253 ,
the two states have the same ground - state energy .
the ground - state energy per unit cell , @xmath254 , is given by @xmath255 \label{gse}\end{aligned}\ ] ] where @xmath256 is the self - energy matrix given by @xmath257 .
the formula ( [ gse ] ) can be derived according to many - particle theory.@xcite note that the energy levels of the ocs at finite @xmath231 are not degenerate for interchanging the indices of the two valleys .
especially , the landau levels @xmath258 and @xmath259 are given by @xmath260 and @xmath261 , respectively .
for positive @xmath231 and @xmath92 , the level @xmath259 is always occupied . in fig .
10 , we exhibit the result for @xmath253 as function of @xmath0 and compare it with the experimental data.@xcite the experimental data are obtained by converting the critical electric field @xmath227 to @xmath240 according to @xmath262 with the dielectric constant @xmath263 ( solid points ) and 3 ( diamonds ) . as seen from fig .
10 , the behavior of @xmath253 by the theoretical calculation is in fairly good agreement with the experiment @xcite with @xmath263 in the converting from @xmath227 to @xmath253 .
as already seen , there is another solution of @xmath237 in the range @xmath264
t. we show in the insert in fig .
10 the phase boundary for this case .
we see that the state of @xmath237 in @xmath265 t is unstable with respect to a small @xmath227 .
the range for the stable state of @xmath237 is reduced to @xmath266 t , with @xmath267 t close toward to the experimental data @xcite @xmath268 t.
with the mfa to the 4bcm , we have studied the ocs and the af state of the electrons with finite - range repulsive interactions in blg at the cnp .
we have shown that the result of af state is not in agreement with the experimental observation on the energy gap behavior that grows with increasing the magnetic field @xmath0 . however , for the ocs with only one coupling constant @xmath101 fitting the experimental gap at @xmath0 = 0 , the obtained energy gap at finite @xmath0 is in surprisingly good agreement with experimental data.@xcite the results for the phase transition in the system in the presence of external electric and magnetic fields , and the particle - hole asymmetry spectra in the presence of @xmath0 are in qualitative agreements with the experimental observations.@xcite there is also the intermediate experimental support @xcite to the prediction for the hysteresis energy gap behavior with varying @xmath0 .
these facts show that the ocs is a possible ground state of electrons in blg .
the model explored here can be useful for understanding the physics of the electrons in blg that is expected as a new generation of semiconductor .
this work was supported by the national basic research 973 program of china under grants no .
2011cb932702 and no .
2012cb932302 , nsfc under grant no . 10834011 , and the robert a. welch foundation under grant no . | based on the four - band continuum model , we study the ordered - current state ( ocs ) for electrons in bilayer graphene at the charge neutrality point .
the present work resolves the puzzles that ( a ) the energy gap increases significantly with increasing the magnetic field @xmath0 , ( b ) the energy gap can be closed by the external electric field of either polarization , and ( c ) the particle - hole spectrum is asymmetric in the presence of @xmath0 , all these as observed by the experiment .
we also present the prediction of the hysteresis energy gap behavior with varying @xmath0 , which explains the existing experimental observation on the electric conductance at weak @xmath0 .
the large energy gap of the ocs is shown to originate from the disappearance of landau levels of @xmath1 = 0 and 1 states in conduction / valence band . by comparing with the existing models and the experiments ,
we conclude that the ocs is a possible ground state of electrons in bilayer graphene . |
smooth manifolds are some of the most important objects in mathematics .
they contain a wealth of geometric information , such as tangent spaces , tangent bundles , differential forms , de rham cohomology , etc .
, and this information can be put to great use in proving theorems and making calculations . however
, the category of smooth manifolds and smooth maps is not closed under many useful constructions , such as subspaces , quotients , function spaces , etc . on the other hand ,
various convenient categories of topological spaces are closed under these constructions , but the geometric information is missing .
can we have the best of both worlds ?
since the 1970 s , the category of manifolds has been enlarged in several different ways to a well - behaved category as described above , and these approaches are nicely summarized and compared in @xcite . in this paper , we work with diffeological spaces , which were introduced by j. souriau in his 1980 paper @xcite , and in particular we study the natural topology that any diffeological space has .
a diffeological space is a set @xmath3 along with a specified set of maps @xmath4 for each open set @xmath5 in @xmath6 and each @xmath7 , satisfying a presheaf condition , a sheaf condition , and a non - triviality condition ( see definition [ de : diffeological - space ] ) .
given a diffeological space @xmath3 , the @xmath0-topology on @xmath3 is the largest topology making all of the specified maps @xmath4 continuous . in this paper , we make the first detailed study of the @xmath0-topology .
our results include theorems giving properties and characterizations of the @xmath0-topology as well as many examples which show the behaviour that can occur and which rule out some natural conjectures .
our interest in these topics comes from several directions .
first , it is known @xcite that the topological spaces which arise as the @xmath0-topology of a diffeological space are precisely the @xmath1-generated spaces , which were introduced by jeff smith as a possible convenient category for homotopy theory and were studied by @xcite .
some of our results help to further understand which spaces are @xmath1-generated , and we include illustrative examples .
second , for any diffeological spaces @xmath3 and @xmath8 , the set @xmath9 of smooth maps from @xmath3 to @xmath8 is itself a diffeological space in a natural way and thus can be endowed with the @xmath0-topology . since the topology arises completely canonically , it is instructive to compare it with other topologies that arise in geometry and analysis when @xmath3 and @xmath8 are taken to be smooth manifolds .
a large part of this paper is devoted to this comparison , and again we give both theorems and illustrative examples . finally , this paper arose from work on the homotopy theory of diffeological spaces , and can be viewed as the topological groundwork for this project .
it is for this reason that we need to focus on an approach that produces a well - behaved category , rather than working with a theory of infinite - dimensional manifolds , such as the one throroughly developed in the book @xcite .
we will , however , make use of results from @xcite , as many of the underlying ideas are related .
here is an outline of the paper , with a summary of the main results : in section [ se : background ] , we review some basics of diffeological spaces . for example , we recall that the category of diffeological spaces is complete , cocomplete and cartesian closed , and that it contains the category of smooth manifolds as a full subcategory .
moreover , like smooth manifolds , every diffeological space is formed by gluing together open subsets of @xmath6 , with the difference that @xmath10 can vary and that the gluings are not necessarily via diffeomorphisms . in section
[ se : d - topology ] , we study the @xmath0-topology of a diffeological space , which was introduced by iglesias - zemmour in @xcite . we show that the @xmath0-topology is determined by the smooth curves ( theorem [ th : curves ] ) , while diffeologies are not ( example [ diffeology ] ) .
we recall a result of @xcite which says that the topological spaces arising as the @xmath0-topology of a diffeological space are exactly the @xmath1-generated spaces ( proposition [ delta ] ) .
we give a necessary condition and a sufficient condition for a space to be @xmath1-generated ( propositions [ la ] and [ countable ] ) and show that neither is necessary and sufficient ( proposition [ pr : fc ] and example [ lpc ] ) .
we can associate two topologies to a subset of a diffeological space .
we discuss some conditions under which the two topologies coincide ( lemmas [ l1 ] and [ l2 ] , proposition [ pr : topologies - agree ] , and corollary [ d - open ] ) .
section [ se : function - spaces ] contains our most substantial results .
we compare the @xmath0-topology on function spaces between smooth manifolds with other well - known topologies .
the results are ( 1 ) the @xmath0-topology is almost always strictly finer than the compact - open topology ( proposition [ pr : d - vs - compact - open ] and example [ compare ] ) ; ( 2 ) the @xmath0-topology is always finer than the weak topology ( proposition [ weakvsd ] ) and always coarser than the strong topology ( theorem [ th : d - vs - strong ] ) ; ( 3 ) we give a full characterization of the @xmath0-topology as the smallest @xmath1-generated topology containing the weak topology ( theorem [ conj : d - top ] ) ; ( 4 ) as a consequence , we show that the weak topology is equal to the @xmath0-topology if and only if the weak topology is locally path - connected ( corollary [ iff : weak = d ] ) ; ( 5 ) in particular , when the codomain is @xmath6 or the domain is compact , the @xmath0-topology coincides with the weak topology ( corollary [ cor : n = r ] and corollary [ m - cpt ] ) , but not always ( example [ weak - neq - d ] ) .
all smooth manifolds in this paper are assumed to be hausdorff , second countable and without boundary .
we would like to thank andrew stacey and chris schommer - pries for very helpful conversations , and jeremy brazas for the idea behind example [ lpc ] .
here is some background on diffeological spaces .
a good reference for the material in this section is the book @xcite . [
de : diffeological - space ] a * diffeological space * is a set @xmath3 together with a specified set @xmath11 of maps @xmath12 ( called * plots * ) for each open set @xmath5 in @xmath6 and for each @xmath7 , such that for all open subsets @xmath13 and @xmath14 : 1 .
( covering ) every constant map @xmath12 is a plot ; 2 .
( smooth compatibility ) if @xmath12 is a plot and @xmath15 is smooth , then the composition @xmath16 is also a plot ; 3 .
( sheaf condition ) if @xmath17 is an open cover and @xmath12 is a set map such that each restriction @xmath18 is a plot , then @xmath12 is a plot .
we usually use the underlying set @xmath3 to denote the diffeological space @xmath19 .
let @xmath3 and @xmath8 be two diffeological spaces , and let @xmath20 be a set map .
we call @xmath21 * smooth * if for every plot @xmath22 of @xmath3 , the composition @xmath23 is a plot of @xmath8 .
the collection of all diffeological spaces with smooth maps forms a category , and we will denote it by @xmath24 . given two diffeological spaces @xmath3 and
@xmath8 , we write @xmath25 for the set of all smooth maps from @xmath3 to @xmath8 .
an isomorphism in @xmath24 will be called a * diffeomorphism*. there is a fully faithful functor from the category of smooth manifolds to @xmath24 .
every smooth manifold @xmath26 is canonically a diffeological space with the same underlying set and plots taken to be all smooth maps @xmath27 in the usual sense .
we call this the * standard diffeology * on @xmath26 . by using charts , it is easy to see that smooth maps in the usual sense between smooth manifolds coincide with smooth maps between them with the standard diffeology . from now on , unless we say otherwise , every smooth manifold considered as a diffeological space is equipped with the standard diffeology .
it is clear that given a fixed open subset @xmath5 of @xmath6 , the set of all plots from @xmath5 to a fixed diffeological space @xmath3 is equal to @xmath28 . given a set @xmath3 , let @xmath29 be the set of all diffeologies on @xmath3 ordered by inclusion . then @xmath29 is a complete lattice .
this follows from the fact that @xmath29 is closed under arbitrary ( small ) intersection .
the largest element in @xmath29 is called the * indiscrete diffeology * on @xmath3 , which consists of all set maps @xmath12 , and the smallest element in @xmath29 is called the * discrete diffeology * on @xmath3 , which consists of all locally constant maps @xmath12 . the smallest diffeology @xmath30 on @xmath3 containing a set of maps @xmath31 is called the diffeology * generated * by @xmath32 .
the diffeology @xmath30 consists of all maps @xmath33 such that there exists an open cover @xmath34 of @xmath35 such that @xmath21 restricted to each @xmath36 factors through some element @xmath18 in @xmath32 via a smooth map @xmath37 .
the standard diffeology on a smooth manifold is generated by any smooth atlas on the manifold .
for every diffeological space @xmath3 , @xmath11 is generated by @xmath38 .
generalizing the previous paragraph , let @xmath39 be a set of functions from some diffeological spaces to a fixed set @xmath3 .
then there exists a smallest diffeology on @xmath3 making all @xmath40 smooth , and we call it the * final diffeology * defined by @xmath32 . for a diffeological space @xmath3 with an equivalence relation @xmath41
, the final diffeology defined by the quotient map @xmath42 is called the * quotient diffeology*. similarly , let @xmath43 be a set of functions from a fixed set @xmath8 to some diffeological spaces .
then there exists a largest diffeology on @xmath8 making all @xmath44 smooth , and we call it the * initial diffeology * defined by @xmath45 . for a diffeological space @xmath3 and a subset @xmath32 of @xmath3 , the initial diffeology defined by the inclusion map @xmath46
is called the * sub - diffeology*. more generally , we have the following result : the category @xmath24 is both complete and cocomplete
. limits and colimits are constructed as follows .
the forgetful functor @xmath47 , where @xmath48 denotes the category of sets , must preserve both limits and colimits since it has both left and right adjoints , given by the discrete and indiscrete diffeologies .
the diffeology on the ( co)limit is the initial ( final ) diffeology defined by the natural maps .
more precisely , let @xmath49 be a functor from a small category @xmath50 .
then @xmath51 is a plot if and only if the composition @xmath52 is a plot of @xmath53 for each @xmath54 . and @xmath55 is a plot if and only if locally it factors as @xmath56 for some @xmath54 , with the first map a plot of @xmath53 .
[ th : cartesian - closed ] the category @xmath24 is cartesian closed . given two diffeological spaces @xmath3 and @xmath8 , the set of maps @xmath57 is smooth@xmath58 forms a diffeology on @xmath25 .
we call it the * functional diffeology * on @xmath25 , and we always equip hom - sets with the functional diffeology . furthermore , for each diffeological space @xmath8 , @xmath59 is an adjoint pair . a smooth manifold of dimension @xmath10 is formed by gluing together some open subsets of @xmath6 via diffeomorphisms . a diffeological space is also formed by gluing together open subsets of @xmath6 ( with the standard diffeology ) via smooth maps , possibly for all @xmath7 .
to be precise , let s introduce the following concept : let @xmath60 be the category with objects all open subsets of @xmath6 for all @xmath7 and morphisms the smooth maps between them .
given a diffeological space @xmath3 , we define @xmath61 to be the category with objects all plots of @xmath3 and morphisms the commutative triangles @xmath62_p \ar[rr]^f & & v \ar[dl]^q \\ & x , } \ ] ] with @xmath63 plots of @xmath3 and @xmath21 a smooth map .
we call @xmath61 the * category of plots of @xmath3*. it is equiped with a forgetful functor @xmath64 sending a plot @xmath12 to @xmath5 regarded as a diffeological space and sending the morphism displayed above to @xmath21
. we can use @xmath65 to show that any @xmath3 can be built out of euclidean spaces : [ colim ] the colimit of the functor @xmath64 is @xmath3 .
clearly there is a natural cocone @xmath66 sending the above commutative triangle to itself . for each diffeological space @xmath8 and cocone @xmath67
, we define a set map @xmath68 by sending @xmath69 to @xmath70 , where by abuse of notation the second @xmath71 denotes the plot @xmath72 with image @xmath69 .
note that @xmath73 induces a ( unique ) cocone map since @xmath74 for each plot @xmath22 and each @xmath75 , which also implies the smoothness of @xmath73 .
the result is essentially the same as ( * ? ? ?
* exercise 33 ) . given a diffeological space @xmath3 , the category @xmath61 can be used to define geometric structures on @xmath3 .
see @xcite for a discussion of differential forms and the de rham cohomology of a diffeological space , and see @xcite for tangent spaces and tangent bundles .
a diffeological space is a set with some extra structure .
we can associate to every diffeological space the following interesting topology : given a diffeological space @xmath3 , the final topology induced by its plots , where each domain is equipped with the standard topology , is called the * @xmath0-topology * on @xmath3 . in more detail , if @xmath76 is a diffeological space , then a subset @xmath32 of @xmath3 is open in the @xmath0-topology of @xmath3 if and only if @xmath77 is open for each @xmath78 .
we call such subsets * @xmath0-open*. if @xmath29 is generated by a subset @xmath79 , then @xmath32 is @xmath0-open if and only if @xmath77 is open for each @xmath80 .
a smooth map @xmath81 is continuous when @xmath3 and @xmath82 are equiped with the @xmath0-topology , and so this defines a functor @xmath83 to the category of topological spaces .
\(1 ) the @xmath0-topology on a smooth manifold with the standard diffeology coincides with the usual topology on the smooth manifold .
\(2 ) the @xmath0-topology on a discrete diffeological space is discrete , and the @xmath0-topology on an indiscrete diffeological space is indiscrete .
every topological space @xmath8 has a natural diffeology , called the * continuous diffeology * , whose plots @xmath84 are the continuous maps . a continuous map @xmath85 is smooth when @xmath8 and @xmath86 are equiped with the continous diffeology , and so this defines a functor @xmath87 .
[ dc ] the functors @xmath88 are adjoint , and we have @xmath89 and @xmath90 . the adjointness is ( * ? ? ?
* proposition 3.1 ) , and the rest is easy .
[ la ] for each diffeological space , the @xmath0-topology is locally path - connected .
however , not every locally path - connected space comes from a diffeological space ; see example [ lpc ] .
we say that a sequence @xmath91 in @xmath6 * converges fast * to @xmath71 in @xmath6 if for each @xmath92 the sequence @xmath93 is bounded .
note that every convergent sequence has a subsequence which converges fast .
[ lem : scl ] let @xmath91 be a sequence which converges fast to @xmath71 in @xmath6 . then there is a smooth curve @xmath94 such that @xmath95 for @xmath96 , @xmath97 for @xmath98 , @xmath99 for each @xmath100 , and @xmath101 maps @xmath102 $ ] to the line segment joining @xmath103 and @xmath91 .
[ th : curves ] the @xmath0-topology on a diffeological space @xmath3 is determined by @xmath104 , in the sense that a subset @xmath32 of @xmath3 is @xmath0-open if and only if @xmath77 is open for every @xmath105 .
( @xmath106 ) this follows from the definition of the @xmath0-topology .
( @xmath107 ) suppose that @xmath77 is open for every @xmath105 .
consider a plot @xmath108 , and let @xmath109 .
suppose that @xmath110 converges fast to @xmath71 . by the special curve lemma , there is a smooth curve @xmath111 such that @xmath99 for each @xmath112 and @xmath113 .
since @xmath114 is open , @xmath91 is in @xmath115 for @xmath112 sufficiently large .
so @xmath115 is open in @xmath5 .
[ diffeology ] let @xmath116 with the standard diffeology , and let @xmath8 be the set @xmath117 with the diffeology generated by @xmath118 .
then @xmath119 is homeomorphic to @xmath120 since @xmath121 , but @xmath3 and @xmath8 are not diffeomorphic since the identity map @xmath122 does not locally factor through curves . in other words ,
the @xmath0-topology is determined by smooth curves , but the diffeology is not . in this example
, @xmath8 has the smallest diffeology such that @xmath123 consists of the usual smooth curves .
in contrast , by boman s theorem ( * ? ? ?
* corollary 3.14 ) , @xmath3 has the _ largest _ diffeology such that @xmath123 consists of the usual smooth curves . that is , @xmath22 is a plot if and only if for every smooth function @xmath124 , the composite @xmath125 is in @xmath104 . a topological space @xmath3 is called @xmath1-generated
if the following condition holds : @xmath126 is open if and only if @xmath127 is open in @xmath128 , the standard @xmath10-simplex in @xmath129 , for each continuous map @xmath130 and each @xmath7 .
it is not hard to show that being @xmath1-generated is the same as being @xmath131-generated or @xmath132$]-generated , i.e. that one can determine the open sets of a @xmath1-generated space using just the continuous maps @xmath133 or @xmath132 { \to}x$ ] .
this follows from the existence of a surjective continuous map @xmath134 that exhibits @xmath128 as a quotient of @xmath131 .
note the similarity to theorem [ th : curves ] .
more on @xmath1-generated topological spaces can be found in @xcite .
[ delta ] the spaces in the image of the functor @xmath0 are exactly the @xmath1-generated topological spaces .
since the argument is easy , we include a proof .
let @xmath3 be a diffeological space , and consider @xmath135 .
suppose @xmath127 is open in @xmath131 for all continuous @xmath136 .
then @xmath127 is open in @xmath131 for all smooth @xmath137 .
thus @xmath32 is open in @xmath119 , and so @xmath119 is @xmath1-generated .
now suppose that @xmath8 is @xmath1-generated . by adjointness ,
the identity map @xmath138 is continuous .
we claim that it is a homeomorphism , and so @xmath8 is in the image of @xmath0 . indeed , suppose @xmath139 is open .
that is , @xmath127 is open in @xmath131 for all smooth @xmath140 .
that is , @xmath127 is open in @xmath131 for all continuous @xmath141 .
then , since @xmath8 is @xmath1-generated , @xmath32 is open in @xmath8 .
because of this , it will be helpful to better understand which topological spaces are @xmath1-generated .
[ countable ] every locally path - connected first countable topological space is @xmath1-generated .
let @xmath142 be a locally path - connected first countable topological space .
then for each @xmath69 , there exists a neighborhood basis @xmath143 of @xmath71 , such that 1 .
each @xmath144 is path - connected ; and 2 .
@xmath145 .
this is because , for a neighborhood basis @xmath146 of @xmath71 , we can define @xmath147 to be the path - component of @xmath148 containing @xmath71 , and @xmath144 to be the path - component of @xmath149 containing @xmath71 for @xmath150 . since @xmath3 is locally path - connected , each @xmath144 is open .
now let @xmath151 be the final topology on @xmath3 for all continuous maps @xmath152 for all @xmath7 .
clearly @xmath153 .
suppose @xmath32 is not in @xmath154 .
this means that there exists @xmath155 such that for each @xmath156 which is a neighborhood of @xmath71 , there exists @xmath157 .
let @xmath143 be a neighbourhood basis for @xmath71 with the above two properties , and write @xmath158 accordingly .
define @xmath159 { \to}x$ ] by letting @xmath160}$ ] be a continuous path connecting @xmath161 to @xmath162 in @xmath144 , and @xmath163 .
it is easy to see that @xmath21 is continuous for @xmath142 , but @xmath127 is not open in @xmath164 $ ] .
so @xmath32 is not in @xmath151 .
recall from propositions [ la ] and [ delta ] that every @xmath1-generated space is locally path - connected . however ,
not every @xmath1-generated space is first countable : [ pr : fc ] let @xmath3 be a set with the complement - finite topology .
we write @xmath165 for its cardinality
. then 1 .
@xmath3 is @xmath1-generated if @xmath166 or @xmath167 ; 2 .
@xmath3 is not @xmath1-generated if @xmath168 .
note that @xmath3 is not first countable when @xmath167 .
this provides a counterexample to the converse of proposition [ countable ] .
\(1 ) if @xmath3 is a finite set , then the complement finite topology is the discrete topology .
hence @xmath3 is @xmath1-generated .
assume @xmath167 , and let @xmath45 be a non - closed subset of @xmath3 , that is , @xmath169 and @xmath170 .
we must construct a continuous map @xmath171 such that @xmath172 is not closed in @xmath131 .
note that in this case , every injection @xmath133 is continuous .
take an injection @xmath173 .
we can extend this to an injection @xmath171 with @xmath174 .
this map is what we are looking for .
\(2 ) if @xmath168 , then every continuous map @xmath132 { \to}x$ ] is constant . otherwise , since every point in @xmath3 is closed , @xmath132 $ ] would be a disjoint union of at least @xmath175 and at most countably many non - empty closed subsets , which contradicts a theorem of sierpiski ( see , e.g. , ( * ? ?
* a.10.6 ) or the slick argument posted by gowers @xcite ) .
since @xmath3 is not discrete , it is not @xmath1-generated .
assume the continuum hypothesis .
then the above proposition says that a set @xmath3 with the complement finite topology is @xmath1-generated if and only if @xmath3 is not an infinite countable set .
here is an example showing that not every locally path - connected topological space is the @xmath0-topology of a diffeological space : [ lpc ] as a set , let @xmath3 be the disjoint union of copies of the closed unit interval indexed by the set @xmath50 of countable ordinals .
we write elements in @xmath3 as @xmath176 with @xmath177 $ ] and @xmath178 .
let @xmath8 be the quotient set @xmath179 , where the only non - trivial relations are @xmath180 for all @xmath181 . since we will only work with @xmath8
, we denote the elements of @xmath8 in the same way as those of @xmath3 .
the topology on @xmath8 is generated by the following basis : 1 .
the open interval @xmath182 for each @xmath183 and @xmath178 ; 2 .
the set @xmath184 ) \cup ( \cup_{c < a } ( x_c,1_c])$ ] for each @xmath178 and @xmath185 .
one can show that @xmath8 is locally path - connected ( but not first countable ) .
however , @xmath8 is not @xmath1-generated .
indeed , let @xmath186 $ ] .
then @xmath32 is not open in @xmath8 . for every continuous map @xmath130
, we claim that @xmath127 is open in @xmath128 .
otherwise , there exists @xmath187 such that no open neighborhood of @xmath188 is contained in @xmath127 .
since the intervals @xmath189 are open , we must have @xmath190 , the common point .
choose a sequence @xmath191 converging to @xmath188 such that each @xmath192 is not in @xmath127 .
then @xmath193 for some countable ordinals @xmath194 .
let @xmath195 be a countable ordinal larger than each @xmath194 . then @xmath196 is an open set containing @xmath197 but none of the @xmath198 , so @xmath198 is not convergent to @xmath190 , which conflicts the continuity of @xmath21 .
let @xmath3 be a diffeological space , and let @xmath8 be a quotient set of @xmath3 .
then we can give @xmath8 two topologies : 1 .
the @xmath0-topology of the quotient diffeology on @xmath8 ; 2 .
the quotient topology of the @xmath0-topology on @xmath3 .
since @xmath199 is a left adjoint , these two topologies are the same .
similarly , let @xmath3 be a diffeological space , and let @xmath32 be a subset of @xmath3 .
then we can give @xmath32 two topologies : 1 .
@xmath200 the @xmath0-topology of the sub - diffeology on @xmath32 ; 2 .
@xmath201 the sub - topology of the @xmath0-topology on @xmath3
. however , these two topologies are not always the same . in general
, we can only conclude that @xmath202 .
[ 2q ] ( 1 ) let @xmath32 be a subset of @xmath131 . then @xmath203 is discrete if and only if @xmath32 is totally disconnected under the sub - topology of @xmath131 . in particular , if @xmath204 , then @xmath205 is the discrete topology , which is strictly finer than the sub - topology @xmath206 .
\(2 ) let @xmath207 be a continuous and nowhere differentiable function , and let @xmath208 be its graph , equipped with the sub - diffeology of @xmath117 .
then @xmath203 is the discrete topology , which is strictly finer than the sub - topology of @xmath117 .
here is the proof .
let @xmath209 defined by @xmath210 be a smooth map whose image is in @xmath32 .
assume that @xmath211 for some @xmath212 . then by the inverse function theorem
, @xmath213 is a local diffeomorphism around @xmath214 . since @xmath215 , we have @xmath216 , which implies that @xmath217 around @xmath218 , contradicting nowhere differentiability of @xmath21
. therefore , any plot of the form @xmath219 is constant . by theorem
[ th : curves ] , @xmath203 is discrete . on the other hand ,
the sub - topology @xmath220 is homeomorphic to the usual topology on @xmath131 .
we are interested in conditions under which @xmath221(a ) .
[ l1 ] let @xmath32 be a convex subset of @xmath6 .
then @xmath222 .
following the idea of the proof of ( * ? ? ?
* lemma 24.6(3 ) ) , let @xmath223 be closed in the @xmath203-topology , and let @xmath224 be the closure of @xmath45 in @xmath32 for the @xmath220-topology . note that the @xmath220-topology is the same as the sub - topology of @xmath6 . hence ,
for any @xmath225 , we can find a sequence @xmath226 in @xmath45 which converges fast to @xmath195 . since @xmath32 is convex
, the special curve lemma ( lemma [ lem : scl ] ) says that there is a smooth curve @xmath227 such that @xmath228 and @xmath229 for each @xmath230 .
therefore , @xmath231 by the definition of the @xmath0-topology .
[ l2 ] if @xmath32 is a @xmath0-open subset of a diffeological space @xmath3 , then @xmath222 .
let @xmath45 be in @xmath203 . to show that @xmath45 is in @xmath220 , it suffices to show that @xmath45 is @xmath0-open in @xmath3 .
let @xmath22 be an arbitrary plot of @xmath3 .
since @xmath32 is @xmath0-open in @xmath3 , @xmath77 is an open subset of @xmath5 .
hence the composition of @xmath232 is also a plot for @xmath3 , which factors through the inclusion map @xmath233 .
since @xmath234 , @xmath235 is open in @xmath77 , which implies that @xmath236 is open in @xmath5 .
thus @xmath45 is @xmath0-open in @xmath3 , as required .
@xmath237 is @xmath0-open in @xmath238 .
hence , @xmath239 . also see corollary [ d - open ] for another example .
note that lemma [ l2 ] is not true if we change @xmath0-open to @xmath0-closed : let @xmath240 .
then @xmath32 is @xmath0-closed in @xmath131 .
it is easy to check that @xmath203 is discrete , and is strictly finer than @xmath220 .
[ pr : topologies - agree ] let @xmath3 be a diffeological space and let @xmath32 be a subset of @xmath3 . if there exists a @xmath0-open neighborhood @xmath241 of @xmath32 in @xmath3 together with a smooth retraction @xmath242 , then @xmath243 .
( here both @xmath241 and @xmath32 are equipped with the sub - diffeologies from @xmath3 . )
let @xmath234 .
then @xmath244 is @xmath0-open in @xmath3 .
therefore , @xmath245 . given a smooth manifold @xmath26 of dimension @xmath246 , by the strong whitney embedding theorem
, there is a smooth embedding @xmath247 . if we view @xmath26 as a subset of @xmath248 , then @xmath249 since there is an open tubular neighborhood @xmath5 of @xmath26 in @xmath248 together with a smooth retraction @xmath27 .
let @xmath26 and @xmath250 be smooth manifolds . recall that the set @xmath251 of smooth maps from @xmath26 to @xmath250 has a functional diffeology described just after theorem [ th : cartesian - closed ] .
in this section , we consider the topological space obtained by taking the @xmath0-topology associated to this diffeology , and we compare it to other well - known topologies on this set : the compact - open topology , the weak topology , and the strong topology
. here is a review of these three topologies and their relationship . the books @xcite are good references for the weak and strong topologies .
the compact - open topology on @xmath251 has a subbasis which consists of the sets @xmath252 , where @xmath253 is a non - empty compact subset of @xmath26 and @xmath254 is an open subset of @xmath250 .
( this makes sense for any diffeological spaces @xmath26 and @xmath250 , where @xmath253 is then required to be compact in @xmath255 and @xmath254 to be open in @xmath256 . )
we now describe a subbasis for the weak topology on @xmath251 . for @xmath257 , @xmath258 a chart of @xmath26 , @xmath259 a chart of @xmath250 , @xmath260 compact , @xmath261 with @xmath262 , and @xmath263
, we define the set @xmath264 to be @xmath265 for each @xmath266 and each multi - index @xmath267 with @xmath268 .
these sets form a subbasis for the weak topology . here
@xmath269 is a multi - index in @xmath270 with @xmath271 , @xmath272 , and @xmath273 is the differential operator @xmath274 .
a subbasis for the strong topology on @xmath251 is similar , but it allows constraints using multiple charts .
more precisely , if @xmath275 is a family of subbasic sets for the weak topology such that the collection @xmath276 is locally finite , then the intersection of this family is a subbasic set for the strong topology .
in fact , one can show that these intersections form a base for the strong topology .
each of these is at least as fine as the previous one , i.e. , @xmath277 the first inclusion is proved in lemma [ le : wcco ] , and the second is clear .
the compact - open topology and the weak topology coincide if and only if @xmath26 or @xmath250 is zero - dimensional ( see example [ compare ] ) .
moreover , the weak topology and the strong topology coincide if the domain @xmath26 is compact and are different if @xmath26 is non - compact and @xmath250 has positive dimension ( see @xcite ) .
now we start our comparison of the @xmath0-topology with these topologies .
the following lemma is needed for the subsequent proposition .
[ le : product ] let @xmath3 and @xmath8 be two diffeological spaces such that @xmath119 is locally compact hausdorff .
then the natural bijection @xmath278 is a homeomorphism .
note that when @xmath3 is a smooth manifold , @xmath119 is locally compact hausdorff .
first observe that the natural bijection @xmath279 is a homeomorphism for @xmath5 and @xmath35 open subsets of euclidean spaces , since in this case the @xmath0-topology is the usual topology .
the functors @xmath280 , @xmath281 for any diffeological space @xmath282 and @xmath283 for any locally compact hausdorff space @xmath254 all preserve colimits since they are left adjoints .
thus the claim follows from proposition [ colim ] , using that @xmath119 is locally compact hausdorff , as is each @xmath284 for @xmath5 an open subset of euclidean space . for general @xmath3 and @xmath8
, one can show using a similar argument that the @xmath0-topology on @xmath285 corresponds under the bijection above to the smallest @xmath1-generated topology containing the product topology on @xmath286 .
[ pr : d - vs - compact - open ] for diffeological spaces @xmath3 and @xmath8 , the @xmath0-topology on @xmath25 contains the compact - open topology . this result is a stepping stone to proving the stronger statement that the @xmath0-topology contains the weak topology .
recall that the compact - open topology has a subbasis which consists of the sets @xmath287 , where @xmath253 is a non - empty compact subset of @xmath119 and @xmath254 is an open subset of @xmath120 .
we will show that each @xmath288 is @xmath0-open .
let @xmath289 be a plot of @xmath25 .
since the corresponding map @xmath290 is smooth , @xmath291 is open in @xmath292 .
so for each @xmath293 , @xmath294 is in the open set @xmath291 .
note that the natural map @xmath295 is a homeomorphism by lemma [ le : product ] . by the compactness of @xmath253 and the definition of the product topology , @xmath296 for some open neighborhood @xmath35 of @xmath188 in @xmath5 , which implies that @xmath297 is open in @xmath5 .
thus @xmath288 is open in the @xmath0-topology .
we will see in example [ compare ] that the @xmath0-topology is almost always strictly finer than the compact - open topology .
the next lemma will be used to show that the @xmath0-topology contains the weak topology for function spaces between smooth manifolds .
[ differential ] let @xmath5 be an open subset in @xmath6 .
then @xmath298 is smooth .
let @xmath299 be a plot with @xmath300 .
this means that the associated map @xmath301 defined by @xmath302 is smooth .
write @xmath303 for the vector @xmath304 with @xmath305 the origin of @xmath306
. then @xmath307 is smooth .
since @xmath308 , @xmath309 is a plot , which implies the smoothness of @xmath273 .
note that the smoothness of @xmath273 does not imply its continuity in general .
it is an easy exercise that for @xmath310 and @xmath311 , @xmath273 is not continuous in the compact - open topology but is continuous in both the weak and strong topologies .
now we can compare the @xmath0-topology with the weak topology for function spaces between smooth manifolds : [ weakvsd ] let @xmath26 and @xmath250 be smooth manifolds .
then the @xmath0-topology on @xmath251 contains the weak topology .
recall that the weak topology on @xmath251 has the sets @xmath264 , described at the beginning of section [ se : function - spaces ] , as a subbasis .
let @xmath312 be a plot , that is , @xmath313 given by @xmath314 is smooth .
if @xmath315 , then by proposition [ pr : d - vs - compact - open ] , lemma [ differential ] and the facts that @xmath316 and @xmath317 are diffeomorphisms , only finitely many differentials are considered , @xmath253 is compact and @xmath35 is open , it is not hard to see that there exists an open neighborhood @xmath318 of @xmath319 in @xmath254 such that @xmath320 .
therefore , @xmath264 is @xmath0-open .
since the weak topology is almost always strictly finer than the compact - open topology , so is the @xmath0-topology : [ compare ] the @xmath0-topology on @xmath321 is strictly finer than the compact - open topology . to prove this ,
consider @xmath322,1)$ ] , where @xmath323 is the zero function .
this is open in the weak topology and thus is open in the @xmath0-topology .
we claim that no open neighborhood of @xmath323 in the compact - open topology of @xmath321 is contained in @xmath5 .
otherwise , we may assume @xmath324 for some @xmath263 and some compact @xmath253 , since if @xmath325 , then @xmath326 for each @xmath267 and @xmath327 .
then clearly @xmath328 defined by @xmath329 is in @xmath330 for any @xmath253 .
but @xmath21 is not in @xmath5 since @xmath331 . using a similar argument , with bump functions
, one can show that when @xmath26 and @xmath250 are smooth manifolds of dimension at least 1 , then the weak topology is strictly finer than the compact - open topology .
thus the @xmath0-topology is strictly finer than the compact - open topology in this situation .
in general , the weak topology is different from the @xmath0-topology on @xmath251 : [ weak - neq - d ] ( 1 ) let @xmath332 and @xmath333 be equipped with the discrete diffeologies .
let @xmath334 be the constant function sending everything to @xmath335 , and let @xmath336 be defined by @xmath337 . note that @xmath338 converges to @xmath21 in the weak topology for the following reason .
since each element in the subbasis of the weak topology depends only on the values of the function and its derivatives on a compact subset of @xmath332 , any of them containing @xmath21 must contain all @xmath338 for @xmath10 large enough .
on the other hand , we claim that for each @xmath10 there is no continuous path @xmath339
\to c^{\infty}(\n , \{0 , 1\})$ ] with @xmath340 and @xmath341 , where the codomain is given the weak topology .
since the weak topology contains the compact - open topology , such an @xmath65 gives rise to a continuous function @xmath132 \times \n \to \{0 , 1\}$ ] , i.e. a homotopy from @xmath342 to @xmath343 .
since these maps are clearly not homotopic , no such @xmath65 exists . thus the weak topology is not locally path - connected .
it follows from proposition [ la ] that the weak topology is different from the @xmath0-topology on @xmath344 .
the above argument in fact shows that every continuous path in @xmath345 with respect to a topology containing the compact - open topology is constant . in particular
, this holds for the @xmath0-topology , and since the @xmath0-topology is @xmath1-generated , it must be discrete .
\(2 ) let @xmath3 be a countable disjoint union of copies of @xmath346 , i.e. @xmath347 with each @xmath348 .
then the weak topology on @xmath349 is not locally path - connected , by a similar argument , with @xmath350 defined by @xmath351 , and @xmath352 defined by @xmath353 \(3 ) the weak topology on @xmath354 is not locally path - connected , by a similar argument , with @xmath355 defined by @xmath356 and @xmath357 defined by @xmath358 where @xmath359 is a strictly increasing smooth function with @xmath360 for @xmath361 and @xmath362 for all @xmath363 .
the above examples all show that the weak topology is not locally path connected and in particular that it is not @xmath1-generated .
the @xmath0-topology is a @xmath1-generated topology which contains the weak topology , and the following theorem says that , given this , it is as close to the weak topology as possible .
[ conj : d - top ] let @xmath26 and @xmath250 be smooth manifolds
. then the @xmath0-topology on @xmath2 is the smallest @xmath1-generated topology containing the weak topology .
first note that by proposition [ weakvsd ] , the @xmath0-topology contains the weak topology , and by proposition [ delta ] , the @xmath0-topology is @xmath1-generated .
so we must prove that the @xmath0-topology on @xmath251 is contained in every @xmath1-generated topology containing the weak topology .
so let @xmath154 be a @xmath1-generated topology containing the weak topology and assume that @xmath364 not open in @xmath154 .
since @xmath154 is @xmath1-generated , there is a @xmath154-continuous map @xmath365 such that @xmath77 is not open in @xmath131 .
since @xmath154 contains the weak topology , @xmath366 is weak continuous . by composing with a translation in @xmath131
, we can assume that @xmath335 is a non - interior point of @xmath77 .
thus we can find a sequence @xmath367 of real numbers converging to @xmath335 so that @xmath368 for each @xmath369 . by theorem [ th : weak - scl ]
, there is a smooth curve @xmath370 such that @xmath371 for each @xmath303 and @xmath372 .
this shows that @xmath32 is not open in the @xmath0-topology .
since every @xmath1-generated space is locally path connected ( see propositions [ la ] and [ delta ] ) , the previous result is in fact a special case of the next result . [ th : lpc ] let @xmath26 and @xmath250 be smooth manifolds . then the @xmath0-topology on @xmath2 is the smallest locally path connected topology containing the weak topology .
suppose @xmath154 is a locally path - connected topology that contains the weak topology , @xmath32 is not @xmath154-open , and @xmath373 is not @xmath154-interior to @xmath32 .
since the weak topology on @xmath374 is first countable , there is a countable weak neighborhood basis @xmath375 of @xmath21 .
contained in each @xmath376 there is a path - connected @xmath154-neighborhood @xmath377 of @xmath21 . for each @xmath369 , choose an @xmath378 and a @xmath154-continuous ( and therefore weak continuous ) path from @xmath21 to @xmath379 lying entirely in @xmath380 .
we can concatenate these paths to produce a weak continuous path @xmath366 such that @xmath381 and @xmath382 . by theorem
[ th : weak - scl ] , there is a smooth curve @xmath370 such that @xmath383 and @xmath384 .
then @xmath115 contains @xmath335 but not @xmath385 for any @xmath303 , so @xmath32 is not open in the @xmath0-topology . as a corollary
, we have the following necessary and sufficient condition for the weak topology to be equal to the @xmath0-topology : [ iff : weak = d ] let @xmath26 and @xmath250 be smooth manifolds .
then the weak topology on @xmath251 coincides with the @xmath0-topology if and only if the weak topology is locally path - connected .
this follows from theorem [ th : lpc ] ( or from theorem [ conj : d - top ] , using that the weak topology is second countable @xcite ) .
this allows us to give a situation in which the @xmath0-topology and the weak topology coincide .
( see also corollary [ m - cpt ] . )
[ cor : n = r ] let @xmath26 be a smooth manifold .
then the weak topology on @xmath386 coincides with the @xmath0-topology . by lemma [ le : convex ]
, the weak topology on @xmath387 has a basis of convex sets .
a linear path is smooth and hence weak continuous , so it follows that this topology is locally path connected .
our next goal is to show that the @xmath0-topology is contained in the strong topology .
we first need some preliminary results . [
le : locally - convex ] let @xmath26 be a smooth manifold and let @xmath250 be an open subset of @xmath388 .
then the @xmath0-topology on @xmath374 is contained in any topology that contains the weak topology and has a basis of convex sets .
here we say that a subset of @xmath374 is * convex * if it is convex when regarded as a subset of the real vector space @xmath389 .
a convex set is nt necessarily path connected , since linear paths may not be continuous .
thus theorem [ th : lpc ] does nt apply directly .
however , in the proof of theorem [ th : lpc ] , all that is used is that the subsets @xmath377 are path connected in the _ weak _ topology .
since linear paths are smooth , they are weak continuous , and so the proof goes through .
[ le : strong - open - subspace ] let @xmath26 be a smooth manifold and let @xmath250 be an open subset of @xmath388 . then @xmath374 is an open subspace of @xmath389 when both are equipped with the strong topology .
we first prove that the strong topology on @xmath2 is the subspace topology of the strong topology on @xmath390 . since the inclusion map @xmath391 induces a continuous map in the strong topologies ( ( * ? ? ?
* exercise 10(b ) , page 65 ) ) , the intersection of a strong open set in @xmath390 with @xmath2 is open in @xmath2 . on the other hand , the data for each weak
subbasic set @xmath32 in @xmath2 defines a weak subbasic set in @xmath390 whose intersection with @xmath2 is @xmath32 . since the strong subbasic sets are certain intersections of the weak subbasic sets , our claim follows .
now we show that @xmath374 is an open subset of @xmath389 , following the argument in lemma [ le : wcco ] . for @xmath392 , choose charts for @xmath26 and @xmath250 and compact sets @xmath393 as described in lemma [ le : charts](b ) .
then @xmath394 where each @xmath395 is understood to be a subbasic set for @xmath389 .
so @xmath374 is open in the strong topology .
[ th : d - vs - strong ] let @xmath26 and @xmath250 be smooth manifolds .
then the @xmath0-topology on @xmath374 is contained in the strong topology .
choose an embedding @xmath396 , and let @xmath5 be an open tubular neighborhood of @xmath250 in @xmath388 , so that the inclusion @xmath397 has a smooth retract @xmath398 . since @xmath267 and @xmath369 induce continuous maps in both the strong topology ( ( * ? ? ?
* exercise 10 , page 65 ) ) and the @xmath0-topology ( an easy argument ) , @xmath251 is a subspace of @xmath399 when both are equipped with either of these topologies .
so if these topologies agree on @xmath400 , then they agree on @xmath374 .
thus it suffices to prove the result when @xmath250 is open in @xmath388 .
assume that this is the case .
we first prove that the strong topology on @xmath389 has a basis of convex sets .
if @xmath401 is a basic open set of the strong topology , as described at the beginning of section [ se : function - spaces ] , and if @xmath402 , then by the proof of lemma [ le : convex ] , @xmath403 which shows that @xmath32 is covered by convex strong open sets . by lemma
[ le : strong - open - subspace ] , @xmath374 is open in @xmath389 , so it too has a basis of convex sets .
thus , by lemma [ le : locally - convex ] , the @xmath0-topology on @xmath374 is contained in the strong topology .
m - cpt ] let @xmath26 and @xmath250 be smooth manifolds with @xmath26 compact .
then the @xmath0-topology on @xmath251 coincides with the weak topology .
the @xmath0-topology is trapped between the weak topology ( proposition [ weakvsd ] ) and the strong topology ( theorem [ th : d - vs - strong ] ) , and these coincide when @xmath26 is compact . here
is one application of our results : [ d - open ] let @xmath26 be a smooth compact manifold , and let @xmath404 be the set of all diffeomorphisms from @xmath26 to itself with the sub - diffeology of @xmath405 .
then @xmath404 is @xmath0-open in @xmath405 .
hence , @xmath406 in the sense of section [ sse : subset ] . as mentioned in corollary [ m - cpt ] , when @xmath26 is compact , the weak , strong and @xmath0-topologies on @xmath407 all coincide .
the first claim is then the restatement of ( * ? ? ?
* theorem 2.1.7 ) , and the second part follows from lemma [ l2 ] .
similarly , many results in ( * ? ? ? * chapter 2 ) can be translated into those for the @xmath0-topology .
when @xmath26 is non - compact and @xmath250 has positive dimension , the weak topology is different from the strong topology @xcite .
since the weak topology and the @xmath0-topology coincide for @xmath408 , it follows that the @xmath0-topology and the strong topology are different for @xmath408 when @xmath26 is non - compact .
we can make this explicit in the next example .
[ cinfty ] it is not hard to show that the strong topology on @xmath321 has a basis @xmath409 , where @xmath410 for each @xmath411 . on the other hand , since the @xmath0-topology agrees with the weak topology on @xmath321 , it has a basis @xmath412 , where @xmath413 for each@xmath414\}$ ] .
it follows that the strong topology is strictly finer than the @xmath0-topology on @xmath321 . on the other hand
, it can be the case that the @xmath0-topology is different from the weak topology but agrees with the strong topology .
for example , this happens in case ( 1 ) of example [ weak - neq - d ] , where it is easy to see that the strong topology is also discrete .
the book @xcite also studies function spaces between smooth manifolds , but uses a different smooth structure on the function space to ensure that the resulting object has the desired local models . by (
* lemma 42.5 ) , the smooth structure has fewer smooth curves than the diffeology studied here , and as a result the natural topology discussed in ( * ? ? ?
* remark 42.2 ) is larger than the @xmath0-topology .
in fact , according to that remark , it is larger than the strong topology ( which is called the @xmath415-topology in @xcite ) .
in this appendix , our goal is to prove a theorem about the weak topology on function spaces which is analogous to the special curve lemma ( lemma [ lem : scl ] ) .
this is theorem [ th : weak - scl ] . before proving the theorem
, we collect together and prove some basic results about the weak topology on function spaces , and state the following lemma .
a. there exists a locally finite countable atlas @xmath416 of @xmath26 and a compact set @xmath417 , for each @xmath267 , such that @xmath418 , where @xmath419 denotes the interior of @xmath420 . b. for any smooth map @xmath421 , there exist @xmath422 as in ( a ) and a countable atlas @xmath423 of @xmath250 such that @xmath424 for each @xmath267 .
consider @xmath426 where @xmath427 is compact and @xmath428 is open .
let @xmath429 .
choose charts for @xmath26 and @xmath250 and compact sets @xmath420 as described in lemma [ le : charts](b ) .
choose @xmath303 so that @xmath430
. then @xmath431 so @xmath288 is open in the weak topology .
[ le : convex ] let @xmath26 be a smooth manifold . the sets @xmath432 , where@xmath257 , @xmath433 , @xmath258 is a chart of @xmath26 , @xmath434 is compact and @xmath263 , form a subbasis for the weak topology on @xmath389 .
in particular , the weak topology on @xmath389 has a basis of convex sets .
consider a subbasic set @xmath435 containing a function @xmath436 .
first observe that @xmath437 for some @xmath438 , since these sets are determined by comparing finitely many norms on a compact set .
one can then show that @xmath439 for some @xmath440 , using bounds on the derivatives of @xmath317 on @xmath441 .
finally , we claim that @xmath442 for some @xmath443 . to see this , cover @xmath441 by finitely many open balls @xmath444 such that @xmath445 for each @xmath446 , and
let @xmath443 be the minimum of the radii and @xmath440 .
then if @xmath447 and @xmath448 , we have @xmath449 for some @xmath446 and @xmath450 , so @xmath451 .
for @xmath250 open in @xmath388 , we will implicitly use that the inclusion map induces a continuous map @xmath452 in the weak topologies , which follows from the fact that the weak topology is functorial in the second variable ( see ( * ? ? ? * exercise 10(a ) , page 64 ) ) .
( in fact , the weak topology and the subspace topology on @xmath2 agree , but we wo nt need this . ) although @xmath2 need not be an open subset of @xmath390 , it has the following weaker property .
[ le : app ] let @xmath26 be a smooth manifold and let @xmath250 be an open subset of @xmath388 . if@xmath453 and @xmath253 is a compact subset of @xmath26 , then there is a convex basic weak @xmath390-neighborhood of @xmath21 whose elements map @xmath253 into @xmath250 .
the set @xmath454 is open in the compact - open topology on @xmath455 and so is open in the weak topology by lemma [ le : wcco ] . by lemma [ le : convex ] , the weak topology on @xmath389 has a basis of convex sets .
thus any @xmath421 has such a convex basic set as a weak neighbourhood .
[ th : weak - scl ] let @xmath26 and @xmath250 be smooth manifolds .
suppose @xmath456 is weak continuous and @xmath367 is a sequence of real numbers converging to zero .
then there is a subsequence @xmath457 and a smooth curve @xmath458 such that @xmath459 for each @xmath303 and @xmath372 .
we first reduce to the case where @xmath250 is open in @xmath388 . as in theorem
[ th : d - vs - strong ] , choose an embedding @xmath396 , and let @xmath5 be an open tubular neighborhood of @xmath250 in @xmath388 , so that the inclusion @xmath397 has a smooth retract @xmath398 . by ( * ? ? ?
* exercise 10(a ) , page 64 ) , the map @xmath460 sending @xmath363 to @xmath461 is weak continuous , so if the theorem holds for @xmath462 , then there is a smooth curve @xmath463 such that @xmath464 for each @xmath303 and @xmath465 . then the map sending @xmath363 to @xmath466 is smooth , @xmath467 for each @xmath303 , and @xmath468 ,
so we are done .
thus we may assume that @xmath250 is open in @xmath388 .
if @xmath367 is eventually constant we may take @xmath469 to be a constant function , so suppose it is not .
choose charts @xmath470 for @xmath26 and compact sets @xmath471 as described in lemma [ le : charts](a ) .
let @xmath472 . for @xmath473 , the sets , @xmath474 are weak @xmath390-neighborhoods of @xmath21 so we may choose a strictly monotone subsequence @xmath457 such that @xmath475 for each @xmath303 .
set @xmath476 .
now compose @xmath366 with a continuous function taking @xmath385 to @xmath457 for each @xmath303 to obtain a weak continuous function @xmath477 that satisfies @xmath478 for @xmath473 and @xmath479 .
fix @xmath480 . by lemma [ le : app ] , for each @xmath481 $ ] , there is a convex neighborhood of @xmath482 whose elements map @xmath483 into @xmath250 . by compactness , there is a @xmath484 such that any subinterval of @xmath164 $ ] of length at most @xmath485 is mapped by @xmath477 into one of these neighborhoods .
thus , for each @xmath363 , any convex combination of elements in @xmath486 \cap [ 0,1])$ ] maps @xmath483 into @xmath250 .
let @xmath487 be the strictly decreasing sequence obtained by ordering the set @xmath488 .
note that @xmath489 and @xmath490 for @xmath473 . fix a non - decreasing @xmath491)$ ] such that @xmath492 in a neighborhood of @xmath493 $ ] and @xmath494 in a neighborhood of @xmath495 .
let @xmath496 } |\mu^{(\ell')}(t)|$ ] .
define @xmath497 by @xmath498 for @xmath499 , @xmath500 for @xmath501 , and @xmath502 for @xmath503 , @xmath473 .
note that for each @xmath504 $ ] , @xmath505 is a convex combination of elements of @xmath506 \cap [ 0,1])$ ] .
clearly , @xmath507 is constant on @xmath493 $ ] and constant on @xmath495 .
the choice of @xmath508 ensures that it is constant in a neighborhood of @xmath509 for each @xmath303 and smooth on @xmath510 .
thus , @xmath507 is smooth on @xmath511 . to see that it is also smooth at @xmath512 , fix a positive integer @xmath513 , set @xmath514 , @xmath515 for each @xmath303 , and @xmath516 . it will suffice to show that all partial derivatives of @xmath517 exist and equal zero on @xmath518 .
certainly @xmath519 there and if @xmath520 is any composition of partial differentiation operators such that @xmath521 vanishes on @xmath522 then the partial derivative of @xmath521 with respect to any of @xmath523 also vanishes there . to complete the induction , it is enough to show that the partial derivative of @xmath521 with respect to @xmath363 also vanishes on @xmath522 .
where @xmath517 is @xmath524 , the order of mixed partials is unimportant so @xmath525 off @xmath522 for some @xmath526 and some multi - index @xmath267 .
choose @xmath50 so that @xmath527
. then @xmath528 is a tail of the sequence @xmath529 .
so if @xmath530 and @xmath531 , then @xmath532 and , for @xmath533 , @xmath534 if @xmath535 , then @xmath536 and @xmath537 so @xmath538 on @xmath539 . thus , for any @xmath540 , @xmath541 where we have used that @xmath542 in the last inequality . since @xmath303 can be arbitrarily large , this inequality holds for all sufficiently small @xmath363 , so the partial derivative of @xmath521 with respect to @xmath363 ( from the right ) exists and equals zero .
the partial derivative from the left is trivially zero .
this completes the induction and the proof that @xmath507 is smooth .
let @xmath545 be a smooth partition of unity on @xmath26 with @xmath546 supported in @xmath547 and define @xmath469 by @xmath548 .
then @xmath549 is a smooth curve such that @xmath550 for each @xmath303 , and of course @xmath551 .
it remains to show that @xmath552 takes values in @xmath250 for each @xmath553 .
let @xmath554 .
there are finitely many @xmath480 such that @xmath555 ; among them , choose @xmath556 so that @xmath557 is as large as possible .
then , for any @xmath363 and any @xmath480 such that @xmath555 , @xmath505 is a convex combination of elements of @xmath558 \cap [ 0,1])$ ] .
thus , @xmath559 is also a convex combination of elements of @xmath560 \cap [ 0,1])$ ] and therefore maps @xmath561 to @xmath250 .
but @xmath562 so @xmath563 . hence , @xmath564 , that is , @xmath565 .
we conclude that @xmath566 .
this completes the proof . | diffeological spaces are generalizations of smooth manifolds which include singular spaces and function spaces . for each diffeological space ,
p. iglesias - zemmour introduced a natural topology called the @xmath0-topology .
however , the @xmath0-topology has not yet been studied seriously in the existing literature . in this paper
, we develop the basic theory of the @xmath0-topology for diffeological spaces .
we explain that the topological spaces that arise as the @xmath0-topology of a diffeological space are exactly the @xmath1-generated spaces and give results and examples which help to determine when a space is @xmath1-generated .
our most substantial results show how the @xmath0-topology on the function space @xmath2 between smooth manifolds compares to other well - known topologies . |
a gas of excitons is a canonical example for an interacting many - body system @xcite .
excitons are bound states of an electron and a hole , which can be considered as a system with no effective electric charge .
this justifies the assumption of a short - range interaction .
the latter is caused by the pauli principle because of the fermionic constituents . at low temperatures
the excitons can form a superfluid as a collective many - body state . moreover , in the presence of a periodic potential a mott - insulating ( mi ) state is possible where the excitons form a state which is commensurate to the periodic structure of the potential minima .
both states , the superfluid and the mi phase , and the phase transition between them have been observed in an ultracold gase of bosonic rubidium atoms @xcite .
although it is more difficult to control the parameters of an exciton gas with present technologi es , the similarity of the exciton gas and a bosonic gas of real atoms indicates that the formation of these two states and the phase transition between them should be accessible under proper experimental conditions .
superfluidity in a two - dimensional ( 2d ) system of electron - hole pairs was predicted on the basis of cooper pairing within a bcs mean - field approach @xcite in ref .
the bcs phase of electron - hole cooper pairs in a dense electron - hole system and a dilute gas of indirect excitons , formed as bound states of electron - hole pairs , were also analyzed in coupled quantum wells ( cqw ) @xcite .
this was followed by a number of detailed theoretical @xcite as well as experimental studies @xcite . besides the superfluid phase
, a wigner supersolid state caused by dipolar repulsion in electron - hole bilayers , was described @xcite .
the recent theoretical and experimental achievements in the studies of the superfluid dipolar exciton phases in cqws were reviewed in ref .
@xcite , and various experimental studies of excitonic phases in cqws were described in ref .
@xcite .
the formation of a superfluid with indirect excitons was recently proposed in ref .
@xcite using mos@xmath0 layers separated by a hexagonal boron nitride insulating barrier and surrounded by hexagonal boron nitride cladding layers and in ref .
@xcite for dipolar excitions in two parallel transition metal dichalcogenides layers .
a perpendicular electric field modifies the band structure in a way that it becomes advantageous for optically excited electrons and holes to reside in the opposite mos@xmath0 monolayers to form indirect excitons . as an extension of these proposals , we consider excitons in a somewhat modified design , using a structured gate to create an additional periodic potential , to study the linear response of the excitonic superfluid to an electromagnetic wave at low and high concentrations
this provide a new controllable scale through the length of the periodicity .
besides the question about how to create collective states in an exciton gas it is also important to characterize and analyze these states experimentally .
although the system is assumed to be translational invariant , there are characteristic spatial correlations of quantum fluctuations . here
we suggest to measure the linear response to an external electromagnetic field .
this response depends on the wavevector @xmath1 of the external field , which provides information related to the spatial properties of the state . using the static structure factor ( ssf )
, the response is related to the spatial correlation function of the exciton density @xcite . in this paper we will study the ssf inside the superfluid phase and how it changes when we get close to the phase boundaries in the dilute regime and in the high density regime ( i.e. , near the transition to an mi phase ) .
the paper is organized in the following way . in sec .
[ sect : model ] , starting from an interacting gas of electrons and holes , we derive an effective model for excitons formed by bound electron - hole states .
assuming coulomb interaction , we neglect the possibility for a dissociation of electrons and holes .
this model is treated in mean - field approximation and the green s function of quasiparticles are calculated in subsect .
[ sect : mfa_quasi ] . in sec . [ sect : thermal ] we study the corresponding ssf at non - zero temperature and compare it with the result of zero temperature . the results are discussed for different regimes in sec . [
sect : discussion ] .
conclusions follow in sec . v.
the hamiltonian of a the system of electrons and holes , either in single layer or spatially separated in a double layer , can be written in momentum representation as h=__=e , h(_,-_)c^_c _ + _ , _ 1,_2 u_c^_-_1,hc_-_2,hc^__1,ec__2,e , [ hamilton ] where @xmath2 ( @xmath3 ) is the creation ( annihilation ) operator for electrons , and @xmath4 ( @xmath5 ) are the corresponding operators for holes . while @xmath6 is the chemical potential of elctrons
, @xmath7 is the chemical potential of holes , assuming the concentrations of electrons and holes are equal in order to have a neutral electron - hole plasma .
this is justified since the electrons and holes are created always pairwise by an external laser source . the electron and hole single - particle energy spectra @xmath8 are defined in a tight - binding approximation , reflecting an external periodic field which is applied to the system .
the electron - hole attraction due to coulomb interaction is given in momentum space as @xmath9 , which is finite for @xmath10 as a result of screening .
its specific form is different whether we have direct or indirect excitions . in eq .
( [ hamilton ] ) the spins of electrons and holes are neglected , because we are not interested in magnetization effects . the fermionic hamiltonian could be treated within the bcs mean - field approximation to obtain the order parameter for cooper pairs @xcite for many physical quantities , such as the superfluid concentration , this approximation is sufficient .
correlations of excitons , though , would also require to include quantum fluctuations generated by quasiparticles .
there are several options for including quantum excitonic effects .
a very direct way is to consider an effective hamiltonian @xmath11 which describes the dynamics of excitons on the lattice , based on tunneling of excitons between neighboring lattice sites .
when we assume that the effect of the coulomb interaction is strong , for instance , in case of a small distance between electrons in one and holes in another layer , the excitons will not dissociate into electrons and holes .
then we can consider the excitons as stable quantum particles , where the shortest distance of the model is the lattice spacing , because the excitons tunnel between the neighboring minima of the periodic potential .
the fact that excitons are formed by fermions leads to a local repulsive interaction .
similar to the famous hubbard model , the competition of nearest neighbor tunneling on the lattice and the on - site repulsion can lead to the rich physics of `` strong correlations '' with different quantum states , including superfluidity , mott and topological states .
this requires the tuning of the tunneling rate and the concentration of particles .
both parameters are tunable in our experimental proposal through the variation of an external electric field .
first , the idea is to derive from the fermionic hamiltonian ( [ hamilton ] ) an effective hamiltonian for bound electron - hole pairs ( excitons ) whose creation operators in configuration space can be written as a_^=_,k_-,-c_,h^c_,e^d^2r d^2r , [ exc_op1 ] where the kernel @xmath12 decays exponentially on a characteristic length @xmath13 away from @xmath14 .
this length represents the effective size of the exciton .
now we assume that the lattice spacing is much larger than @xmath13 , which implies that the exciton creation and annihilation operators are in a good approximation local : a_^=c_,h^c_,e^ , a_=c_,ec_,h .
[ exc_op2 ] these operators resemble hard - core bosons , since they satisfy the bosonic commutation relations with the additional condition @xmath15 due to the pauli principle of the fermionic constituents .
second , we assume that the excitons can not dissociate into fermions . therefore , the effective hamiltonian is expressed by the operators @xmath16 and @xmath17 only .
as a first order approximation we can write h_eff=_,(j_,+_,)a_^a_ , [ eff_ham00 ] where the summation is with respect to the minima of the periodic potential , using a tight - binding approximation . in eq .
( [ eff_ham00 ] ) the hopping rate is j_,= describes the tunneling of excitons between neighboring potential minima , while @xmath18 is the chemical potential of excitons that controls their concentration .
this hamiltonian describes essentially the physics of the exciton gas in a periodic potential with on - site interaction .
its extension to other types of interaction , including long - range dipole - dipole interaction , is obvious but not important for the subsequent discussion .
hamiltonian ( [ eff_ham00 ] ) can have a normal and a superfluid state for the exciton gas .
in particular , the mi state with concentration @xmath19 is an eigenstate of @xmath20 .
considering that the system is at temperatures well below the kosterlitz - thouless transition temperature , one can use a mean - field approximation and follow the idea of the bogoliubov approach .
then the superfluid phase is distinguished from the normal phase by a spatially uniform order parameter @xmath21 , which vanishes outside the superfluid phase .
then the exciton operators in the superfluid phase read a_=+ _ , a_^=^*+_^ , [ operators ] where @xmath22 is the creation operator for bosonic quasiparticles .
if @xmath23 is the number of bosons in the superfluid phase and @xmath24 is the number of lattice sites , @xmath25 is the concentration of superfluid excitons per unit cell of the lattice .
@xmath23 is obtained from the order parameter as @xmath26 .
a mean - field approximation for the hamiltonian ( [ eff_ham00 ] ) gives at zero temperature @xcite @xmath27 moreover , the total excitonic concentration @xmath28 per unit cell of the lattice is obtained from the free energy per site f=-(tr e^-h_b ) as @xmath29 therefore , concentrations are measured in units of the lattice constant and appear in our calculation as dimensionless quantities . in fig .
[ fig : densities ] the total and the superfluid concentration are plotted in the superfluid phase , shown in fig .
[ fig : phased ] , as a function of the chemical potential . in the dilute regime , that corresponds to low total concentration , almost all excitons are in the superfluid phase , whereas a reduction of the superfluid concentration with increasing total concentration is caused by the formation of a kind of supersolid due to the interplay of the interaction and the underlying lattice structure .
the rather simple results in eqs .
( [ n0 ] ) and ( [ ntot ] ) can be used to calculate other physical quantities of the excitonic gas .
for instance , the effect of quantum fluctuations in momentum representation is described by the quasiparticle green s function of @xmath30 @xcite : _ , _ - , = , [ gf1 ] where @xmath31 is the dispersion in the periodic potential , normalized with @xmath32 , and @xmath33 is the frequency of the external electromagnetic field .
this result is similar to the two - particle green s function obtained for the exciton gas in ref .
the response of a many - body quantum system to a weak external electromagnetic field of frequency @xmath33 and wavevector @xmath1 is given by fermi s golden rule , where the latter is characterized by the structure factor @xmath34 @xcite .
this quantity is defined as the fourier transform of the truncated density - density correlation function : s(,)=_,e^itdt d^2p d^2k , where the wavevector @xmath1 is measured in units of the inverse lattice constant @xmath35 . using the truncated correlation function ( i.e. , substracting the product of concentrations by the second term ) eliminates a dirac delta function at @xmath36 . with the operators in eq .
( [ operators ] ) , @xmath37 and the quasiparticle green s function ( [ gf1 ] ) we obtain s(,)=_,_- , .
integration over all frequencies @xmath33 gives us the ssf as s()s(,)d= [ static_sf2 ] with the lattice tight - binding dispersion @xmath31 .
in the case of a square lattice with lattice constant @xmath38 we get @xmath39 . replacing @xmath40 by the dispersion of a free bose gas @xmath41 and assuming that @xmath42
, we obtain from eq .
( [ static_sf2 ] ) the well - known result of feynman for the low concentration regime @xmath43 @xcite : s()~ .
the ssf at zero temperature is characterized by a vanishing behavior for a vanishing wavevector @xmath1 .
the question is now whether or not this behavior is also visible in a realistic experiment where we have thermal fluctuations .
we will demonstrate subsequently that this is not the case , since even very small thermal fluctuations alter the behavior of the ssf substantially at small wavevectors .
the calculation for including thermal fluctuations requires a summation with respect to matsubara frequencies @xmath44 , where @xmath45 @xcite , rather than a frequency integration .
this summation can be expressed again by an integral in the complex plane @xcite , which gives us for the superfluid phase @xcite @xmath46 and with the bare dispersion @xmath47 for the square lattice .
the ssf is very sensitive to thermal fluctuation at small @xmath1 , as illustrated in fig .
[ fig : struct ] for @xmath48 : it vanishes for @xmath49 and @xmath36 , whereas it increases strongly for @xmath50 if @xmath51 .
the reason for this behavior is the @xmath52 term , which diverges for @xmath53 .
this factor is 1 for strictly zero temperature but diverges for any finite @xmath54 when @xmath51 .
let us distinguish three different regimes for the ssf : @xmath55 ( low concentration regime ) , @xmath56 ( maximum superfluid concentration ) and @xmath57 ( vanishing superfluid at high concentration ) .
for the behavior of the ssf it is crucial that the limit @xmath58 followed by the limit @xmath59 leads to a different result for the structure factor than the limit @xmath59 followed by the limit @xmath58 .
therefore , we have to be careful with the regime @xmath60 at low temperature , because the ssf is very sensitive to small changes of parameters .
this can be seen in figs .
[ fig : struct ] , [ fig : struct2 ] , where either a small change in temperature or a small change in the superfluid concentration has a strong effect on the ssf . _ i. low concentration regime _ in the dilute regime with @xmath55 we have @xmath61 , such that we obtain from eq .
( [ structurf2 ] ) s()~ ~ , [ sf_dilute1 ] which is in agreement with the well - known result for the weakly interacting bose gas and with the ssf of the ideal ( noninteracting ) bose gas @xcite . for low concentration , when @xmath62 one can obtain from eq .
( [ sf_dilute1 ] ) s()~ .
[ sf_dilute ] interaction between the excitons can have two effects : it can create a superfluid with concentration @xmath63 and , together with the periodic potential , it can destroy the superfluid and create a mi state .
thus , besides the weakly interacting regime with @xmath64 , there is an intermediate regime with @xmath65 and maximum superfluid concentration @xmath66 , and a strongly interacting regime of a dense exciton gas with @xmath67 and @xmath68 . since @xmath69 is small at low and at high concentration ( cf . fig .
[ fig : densities ] ) , the ssf differs only by the prefactor @xmath70 in these two regimes , which is 1 in the dilute regime and very small in the dense regime . _
maximum superfluid concentration _ in this case we have @xmath56 , which implies according to eqs .
( [ n0 ] ) and ( [ ntot ] ) @xmath71 and @xmath72 ( half - filled lattice ) ; i.e. , @xmath73 .
moreover , we get for the quasiparticle dispersion @xmath74 and , therefore , s()~ ~ ~ for j1 .
[ asympt2 ] _ iii .
high concentration regime _
@xmath57 implies for the concentrations @xmath75 and @xmath67 ; i.e. , @xmath76 .
thus , in the dense regime , i.e. close to the mi phase when @xmath77 , the ssf vanishes with @xmath69 as s()~n_0 ~ for j_1 , [ asympt3 ] in contrast to eq .
( [ sf_dilute ] ) for the dilute regime .
the dependence of the ssf on @xmath78 in the dilute regime is depicted in fig .
[ fig : struct ] . at fixed temperature
the ssf is plotted in fig .
[ fig : struct2 ] for different values of @xmath69 .
both figures reflect the asymptotic behavior for small @xmath1 obtained in eqs .
( [ sf_dilute ] ) , ( [ asympt2 ] ) and ( [ asympt3 ] ) . in the experiment with a dilute exciton gas
we should always see a broad maximum at small @xmath1 which decays according eq .
( [ sf_dilute ] ) like @xmath79 .
this maximum disappears at higher exciton concentrations , and the ssf is relatively flat with a very weak increase , as illustrated in fig .
[ fig : struct2 ] . we have not calculated the ssf in the mi phase because the latter is not accessible with the present experimental techniques .
such a calculation would require a different approach to obtain the corresponding quasiparticle green s function .
but this is a straightforward task when we use the concept developed , for instance , in ref . @xcite .
we predict a method to control the state of an exciton system by introducing a spatially periodic potential through a profiled external gate or by a periodic superlattice structure .
the control can be performed also by change of the chemical potential of the system , driven by laser pumping .
the transition of superfluid ( bec ) state to localized state takes place .
this phase transition is controlled by the parameters of the external potential .
the transition can be revealed by study of exciton flow induced by a gradient of the exciton concentration , originated from a photon pumping spot . besides , the transition can be observed by elastic light scattering .
the elastic light scattering cross section is proportional to the static structure factor .
we demonstrated that the static structure factor drastically changes at the superfluid to localized states transition .
the ssf , which can be measured by non - resonant scattering of electromagnetic waves , is a useful quantity to characterize a superfluid state in an excitonic system .
since the latter was proposed some time ago @xcite and observed in a recent experiment @xcite , a detailed analysis , for instance , based on x - ray raman scattering ( xrs ) @xcite could reveal more details of the nature of this state . in particular , there is a pronounced maximum of the ssf at small wavevectors in the case of a dilute exciton gas and an almost flat ssf at higher exciton concentrations .
this high sensitivity can be employed for accurate measurements of the superfluid properties at small wavevector @xmath1 . in particular
, it would be possible to observe the reduction of the superfluid concentration and the crossover to the mi phase under an increasing total exciton concentration . in a proper experimental set - up the latter
could be controlled by an external gate .
d. w. snoke , in quantum gases : finite temperature and non - equilibrium dynamics ( vol .
1 , cold atoms series ) , n. p. proukakis , s. a. gardiner , m. j. davis , and m. h. szymanska , eds . , imperial college press , london , 2013 .
d. s. chemla , j. b. stark , and w. h. knox , in `` ultrafast phenomena viii '' , eds .
martin et al . , springer , 21 ( 1993 ) ; c. w. lai , j. zoch , a. c. gossard , and d. s. chemla , science * 303 * , 503 ( 2004 ) .
l. v. butov , a. zrenner , g. abstreiter , g. bohm and g. weimann , phys .
lett . * 73 * , 304 ( 1994 ) ; l. v. butov , c. w. lai , a. l. ivanov , a. c. gossard , and d. s. chemla , nature * 417 * , 47 ( 2002 ) ; l. v. butov , a. c. gossard , and d. s. chemla , nature * 418 * , 751 ( 2002 ) . | we consider excitons in a two - dimensional periodic potential and study the linear response of the excitonic superfluid to an electromagnetic wave at low and high densities .
it turns out that the static structure factor for small wavevectors is very sensitive to a change of density and temperature .
it is a consequence of the fact that thermal fluctuations play a crucial role at small wavevectors , since exchanging the order of the two limits , zero temperature and vanishing wavevector , leads to different results for the structure factor .
this effect could be used for high accuracy measurements in the superfluid exciton phase , which might be realized by a gated electron - hole gas .
the transition of the exciton system from the superfluid state to a non - superfluid state and its manifestation by light scattering are discussed . |
soft x - ray transients ( sxt ) are binaries in which a black hole or a neutron star primary accretes matter from a main sequence or giant secondary ( van paradijs & mcclintock 1995 , tanaka & shibazaki 1996 ) .
several sources containing a black hole show outbursts every few decades , which last a few months . during the long - lasting quiescent state
the systems are very dim ( mcclintock & remillard 1990 , mcclintock et al .
the x - ray luminosity observed during this state from binaries containing a black hole is systematically lower than that from systems with a neutron star .
but the luminosity from the neutron star systems is not as high as expected for accretion of all matter at the neutron star s surface .
for a recent compilation of asca observations of sxts in quiescence see asai et al .
it was suggested that in the neutron star systems part of the matter is thrown out by a magnetic propeller ( illarionov & sunyaev 1975 ) and therefore the luminosity is reduced ( asai et al . 1998 , menou et al .
1999b ) . for the x - ray
nova a0620 - 00 in quiescence the comparison of observed optical and x - ray luminosity showed that the observed x - ray flux is much less than expected if one assumes the same accretion flow everywhere in the disk ( see mcclintock et al . 1995 and references therein ) .
this observation led to the suggestion that most of the thermal and kinetic energy is carried into the black hole by an advection - dominated accretion flow ( adaf ) and is not radiated away .
the adaf model includes the physics of the hot accretion flow ( narayan et al .
1996 , 1997 , for a recent review see narayan et al .
1999 ) and is successful in describing the observed spectra of black hole sxts in quiescence .
it provides strong evidence for the black hole nature of the accreting primaries .
the configuration of the black hole sxt is the following .
matter from the roche - lobe filling secondary star flows over to an accretion disk . in quiescence
the disk is cool in the outer part . at a certain transition radius , more accurately a transition region ,
the accretion flow is transformed into a hot coronal flow and farther in no underlying cool disk exists anymore ( esin et al . 1997 ) .
this coronal flow continues inward and , at very high temperatures , becomes an advection - dominated accretion flow . for a schematic drawing see fig
. 1 .
our present investigation studies evaporation as the source of the matter for the adaf .
we have shown earlier ( meyer & meyer - hofmeister 1994 , liu et al 1995 , liu et al .
1997 ) that evaporation of matter from a cool disk into an overlying hot corona is important for the evolution of accretion disks in binaries during quiescence .
the same process works in dwarf nova disks and in disks of x - ray novae ( mineshige et al .
1998 ) . in dwarf
nova disks the inner part of the disk around the white dwarf primary can be evaporated , depending on the amount of mass flow in the cool disk . in black hole
sxts the horizon lies at much smaller radii and the coronal flow continues into a probably more spherical inner area filled with very hot gas . if the standard thin accretion disk would extend to such small radii with the same mass flow rate as in the outer regions the matter would become ionized and the disk instability would not allow the whole disk to remain in a quiescent state .
lasota et al .
( 1996 ) pointed out that the inner part of the thin disk has to be truncated . in sect .
2 we discuss the physics of the evaporation process applied to a black hole source .
we point out that the interaction of evaporation and mass accretion in the cool disk are similar for dwarf novae and transients .
we compute the disk evolution using the well observed system a0620 - 00 as an example ( sect .
3 ) . in sect .
4 we compare our results with the findings from the spectral fit based on the adaf model . in sect .
5 we summarize the general features .
we analyse the disks around black holes .
the situation for disks around neutron stars is more complex .
irradiation by the hot neutron star surface has to be taken into account .
but it is expected that for both , black hole and neutron star sources the accretion flow near the compact star occurs as an adaf ( menou et al .
1999b ) .
the evaporation model decribes how above the cool inert disk a hot self - sustained coronal layer is built up , fed by matter from the disk underneath ( meyer & meyer - hofmeister 1994 , meyer 1999 ) .
the vertical structure of the corona is established by the balance between heat generation and radiation losses at the coronal - chromospheric transition layer .
based on a one - zone model liu et al .
( 1995 ) computed the evaporation rate @xmath0 ( evaporation into the corona above and below the cool disk , both sides ) for different white dwarf masses up to @xmath1 .
the rate varies with the distance to the primary star @xmath2 and its mass @xmath3 as @xmath4.\ ] ] due to wind loss the coronal mass flow rate at the inner edge of the cool disk towards the compact object , @xmath5 , is about 80% of @xmath0 .
we checked the applicability of this result to larger compact object masses with respect to the following physical conditions : the electron - proton collision frequency is high enough that equilibrium is established within the dynamical timescale , the mean free path of transporting electrons is small compared to the temperature scale length in the conductive boundary layer .
the ratio of advective to conductive energy flux remains the same .
all these three ratios remain nearly invariant for large changes in central mass . an important feature of the accretion disk in quiescence is the existence of a transition between cool disk and hot coronal flow . to evaluate the transition radius @xmath6 one needs to know the mass flow rate @xmath7 in the disk .
the assumption of a quasi - stationary disk allows such an estimate . to determine how this location changes during the long period of quiescence one has to compute the evolution of the disk .
we describe here briefly earlier results to show the similarity in the evolution of the inner disk in different binaries .
to follow such a disk evolution it is necessary to include conservation of mass and angular momentum in the interacting cool disk and corona ( liu et al .
the angular momentum released farther inside by the accreting matter flows outward in the corona and even forces a small part of coronal matter to flow outward .
this matter condenses later in the outer cool disk ( compare fig .
( = location of the transition between cool disk and coronal flow ) .
@xmath8 depends on the evaporation rate @xmath9 and on the mass flow rate @xmath7 in the cool disk , @xmath3 primary mass .
the upper panel shows @xmath9 [ eq . 1 ] and @xmath8 for 2 different rates @xmath7 : for lower @xmath10 the transition occurs farther out . for explanation of the critical mass flow rates @xmath11 and @xmath12 see fig . 3 where the corresponding values @xmath13 and @xmath14 are given .
the lower panel shows @xmath9 for 2 different primary masses and the resulting change of @xmath8 : for smaller @xmath3 the transition occurs farther in .
( cgs units),width=332 ] liu et al .
( 1997 ) studied the evolution of the disk around a @xmath15 primary star in an ordinary dwarf nova system with a relatively high mass overflow rate of @xmath16= @xmath17 from the secondary .
disk evaporation creates a hole in the inner disk and the transition remains near @xmath18 cm while matter piles up in the outer disk until the onset of instability occurs near @xmath19 cm . without evaporation
this would have happened near to the white dwarf surface . for a comparison of accretion disk instabilities in dwarf novae and
transient x - ray sources see cannizzo ( 1998 ) .
wz sge stars , a subgroup of dwarf nova ( ritter & kolb 1998 ) , have also very long outburst recurrence times similar to black hole sxts considered here ( kuulkers 1999 ) .
this makes wz sge an interesting candidate to study the effect of evaporation . for the parameters we had chosen according to observations ( liu et al .
1998 ) we found that during the first years of quiescence the mass flow rate in the inner disk is lower than the evaporation rate , @xmath20 , and a hole is formed . with the increasing accumulation of matter
@xmath21 rises and the transition shifts towards the white dwarf until after about 9 years the hole is closed .
evaporation goes on all the time , even after the inner edge of the disk has reached the white dwarf surface , until the outburst is triggered ( fig . 1 , meyer - hofmeister et al . 1998 ) .
generally the inner edge of the cool disk will be established where the evaporation rate equals the mass flow rate in the disk .
this balance determines the transition radius if the mass of the primary and the mass flow rate in the disk are known .
the fact that the disks are quasi - stationary during long quiescence intervals allows a first estimate for @xmath6 .
we show in a schematic drawing how @xmath6 changes if @xmath21 changes ( fig .
2a ) and how a different primary mass influences the evaporation rate @xmath9 and therefore the location of the transition .
the system a0620 - 00 ( nova mon 1917,1975 ) , discovered by _
ariel _ in 1975 ( elvis et al . 1975 ) , is one of the best studied x - ray novae .
for a recent detailed compilation of the observations for a0620 - 00 see kuulkers ( 1998 ) .
the very first theoretical investigations of the accretion disk ( huang & wheeler 1989 and mineshige & wheeler 1989 ) aimed to clarify whether the outburst behaviour of a0620 - 00 could be understood as an accretion disk instability . in later work the emphasis was put mostly on modelling the observed exponential decay of the outburst luminosity .
recently cannizzo ( 1998 ) reviewed these investigations and included an example of outburst modelling based on disk evolution with an assumed law for loss of matter from the cool inner disk , but without the flow of angular momentum in the combined disk + coronal accretion .
our investigation concentrates on the evolution of the accretion disk during quiescence using a computer code which includes the physics of interaction of disk and corona ( see sect .
. observations for this state are more difficult , but have been carried out in the optical , uv , euv and x - ray bands . for a description of the observations and deduction of system parameters
see narayan et al .
( 1996 , 1997 ) .
the orbital period is 7.8hr ( mcclintock et al .
the mass of the black hole follows from the mass function if the inclination is known .
narayan at al .
( 1996,1997 ) considered different values for the inclination i and the corresponding primary mass @xmath3 : i=@xmath22 and @xmath3=@xmath23 , i=@xmath24 and @xmath3=@xmath25 and in the later paper also i=@xmath26 and @xmath3=@xmath27 , as recently suggested by barret et al .
( 1996 ) . for our modelling of the disk evolution
we take bh masses of @xmath28 and @xmath29 .
the viscosity parameter ( shakura & sunyaev 1973 ) assumed was @xmath30= 0.05 , a standard value for the cool disk in dwarf nova outburst modelling .
we assume that only a small amount of matter was left over in the disk after the last outburst ( the value of the critical surface density @xmath31 , compare fig .
3 , corresponds to @xmath32=0.3 .
note that the viscosity in the hot state used in the adaf model is not relevant for the modelling of the quiescent phase ) .
we start with an outer disk radius of 9.210@xmath33 cm ( geometry connected with the orbital period , secondary mass taken as 0.4@xmath34 ) . during the evolution
the disk grows to 10@xmath35 cm .
the same code was used as for wz sge ( solution of the diffusion equation , changes of outer radius due to conservation of mass and angular momentum , evaporation in the inner disk into the coronal flow , changes of the inner edge of the disk , meyer - hofmeister et al .
( 1998 ) ) .
the lightcurve of a620 - 00 in b magnitude for the outburst in 1917 , reconstructed by eachus et al .
( 1976 ) , has a shape similar to the lightcurve of the 1975 outburst ( lloyd et al .
1977 ) , indicating that the two outbursts were similar .
though it is not certain that no outburst was missed it is generally assumed that no outburst ocurred in between .
basic constraints on the quiescent disk evolution are then the two facts : ( 1 ) the outburst recurrence time of 58 years and ( 2 ) the amount of matter stored in the disk and released in the outburst . the amount of energy in the 1975 outburst was estimated as 3 - 4@xmath36erg ( mcclintock et al .
1983 , white et al .
this estimate refered to a neutron star primary and to isotropic radiation , a distance of 0.9 kpc was assumed .
if we use 1 kpc for the distance as narayan et al .
( 1997 ) and assume a black hole and the disk seen under an inclination of @xmath37 we obtain for the matter accreted in the outburst @xmath38= 6 - 8 @xmath39 g ( for the inclination see also the newer analysis of barret et al .
1996 ) . and @xmath14 critical surface densities .
for the surface densities between these values two states , hot and cold , are possible .
when accumulation of matter reaches @xmath40 an outburst sets in , width=332 ] the total amount of matter is closely related to the rate of mass overflow from the companion star .
we study the situation for rates in the range @xmath41 ( mcclintock et al .
1995 ) to @xmath42 , the latter corresponding to the results of narayan et al .
narayan et al .
( 1996 , 1997 ) studied the adaf in a0620 - 00 in two investigations .
the first results from the fit to the spectra were mass flow rates in the disk ( assuming that the same rate flows through the outer disk and in the adaf ) of only 7@xmath43 to 2@xmath44 for a primary mass of 4.4@xmath34 .
but in the second investigation , based on very elaborate calculation techniques and testing different heating of electrons , the resulting mass flow rates in the adaf for most of the models lie in the range 1 to @xmath45 .
most of the models were computed for a primary mass of 6.1 @xmath34 , but the resulting mass flow rate was about the same for 4.4 @xmath34 as taken in the earlier investigation .
it was found that nearly all the optical flux comes from the adaf , the outer disk being quite negligible . from our point of view
this means that the mass overflow rate is higher by the additional amount of matter accumulated in the outer disk for the outburst and by the amount lost in the wind when the coronal flow is formed .
our earlier investigations of the evaporation process gave about 20% wind loss ( liu et al .
the mass overflow rate therefore might be as high as @xmath42 . in an earlier investigation
( meyer - hofmeister & meyer 1999 ) first results for an evolution based on a low mass overflow rate were presented . , mass flow rates 30 years after outburst ( compare fig .
3 ) : @xmath7 mass flow in the standard cool disk , @xmath46 hot coronal flow and total flow .
the negative values for the coronal flow indicate a flow outward ( for conservation of angular momentum during disk evolution).,width=332 ] the aim of our computation is to find out whether one can model the disk evolution so that it complies with the constraints on recurrence time and amount of matter stored in the disk @xmath47and also agrees with mass flow rate in the adaf and transition radius derived from the spectral fit by narayan et al .
our standard case represents such a consistent disk evolution .
we take the following parameters : primary mass @xmath3=4 @xmath34 , mass overflow rate 1.9@xmath48 . in fig .
3 we show the evolution of the thin cool accretion disk with a hot corona above . the onset of the outburst occurs after 59 years . during quiescence the surface density in the disk increases steadily everywhere as the distributions after 15 , 30 and 59 years show .
the surface density in the outer disk determines the total amount of matter stored at the onset of the outburst @xmath47 , in our example 6.9 @xmath39 g .
the transition radius adjusts so that the evaporation rate is the same as the mass flow rate in the inner cool disk ( if the transition would occur only in a very narrow radial interval , the evaporation rate and the mass flow rate there would be the same , but in our computations the evaporation of matter into the hot corona is spread out over a transition area ) . in our standard case
we get for the location of the inner edge of the cool disk @xmath49 cm . in fig .
4 we show the mass flow in the disk after 30 years of quiescence for the
standard case , a situation comparable to the present state of a0620 .
6@xmath50g / s , half of the overflow rate , flows inward towards the black hole . in the transition region
a fraction of the matter flows outward in the corona and condenses again in the cool disk .
such an outward flow is necessary for the conservation of angular momentum in the co - existing cool disk and coronal disk . in the transition region matter
is also lost via a wind from the corona .
this division of mass flows is about the same over the total computed quiescence evolution .
we also studied the disk evolution for other parameters .
the smaller the rate @xmath51 the more time is needed to accumulate enough matter to reach the critical surface density @xmath14 .
the balance of mass flow in the disk and evaporation determines the transition radius .
a lower rate @xmath51 results in a more extended hole ( compare fig .
2 , upper panel ) and the critical surface density can then only be reached farther out in the disk .
this means the surface density is higher everywhere in the disk and therefore more matter piles up in the outer disk .
the quiescence lasts longer .
we found that for an overflow rate of 1.8@xmath48 , only a little lower than in our standard case , no outburst occurs anymore and the disk is stationary .
such a system would not easily be recognized . on the other hand a slightly higher rate
triggers the outburst early ( see table 1 ) . for a slightly higher viscosity parameter @xmath52=0.06 the accumulated amount of matter
@xmath47 is less and the recurrence time is shorter than for the standard value 0.05 .
( small variation of the parameters within the allowed range would lead to exact agreement with the observed recurrence time ) .
.parameters of models of a0620 [ cols="^,^,^,^,^,^ , > " , ] @xmath53 models 1 and 3 of narayan et al .
( 1997 ) + @xmath54 models 1b and 3b ( c.f . table 1 ) , values 20 to 30 years after outburst in table 2 we summarize the comparison of the results from our disk evolution computations and the results from the spectral fit of the adaf model to the observations .
the transition radius @xmath55 used for the adaf fit is estimated from observations on the basis of the largest velocity @xmath56 seen in the @xmath57 line , 2100@xmath58 ( marsh et al .
1994 , orosz et al .
for the inclinations @xmath37 and @xmath59 and the related black hole masses the values @xmath6 follow , the smaller value for the smaller inclination and higher primary mass .
the adaf fit yields a mass flow rate @xmath60 .
this result does not depend sensitively on @xmath6 .
our computations are carried out for two values of viscosity , @xmath52=0.05 , the standard value from dwarf nova outburst modelling and 0.06 .
we take 4 and 6@xmath34 for the primary mass .
we consider only cases of disk evolution which comply with the constraints ( 1 ) recurrence time of 59 years of a0620 and ( 2 ) amount of matter accumulated for the outburst in agreement with the outburst energy ( only a rough value ) . for 6@xmath34
the second constraint is less well met .
the flux measurements used for the adaf fit were obtained many years apart as pointed out by narayan et al .
( 1996 ) . for our models 1b and 3b ( compare table 1 )
we therefore give the transition radius and the coronal mass flow rate ( as supply for the adaf ) for the evolutionary times of 20 and 30 years after the outburst .
this documents the change of values during evolution . comparing the rates @xmath61 and @xmath46 the values are closer for the 6@xmath34 primary .
but we have mentioned already that in this case somewhat more matter is accumulated than estimated from the outburst energy . keeping in mind
that here the results are brought together from very different physical processes , ( a ) the radiation from the innermost very hot sphere and ( b ) the evolution of the cool disk with evaporation of matter into a coronal layer , the derived rates @xmath62 and @xmath46 are remarkably similar .
the computation of the disk evolution for a set of parameters allows us to learn about the general features of the transition of the cool disk to a coronal disk . as pointed out in sect .
2 the transition radius is determined by the balance of mass flow in the disk and evaporation . this situation is the same in disks around white dwarfs and around black holes . in dwarf
nova systems for a high mass flow rate , the transition radius might move in to the white dwarf surface or even disappear .
then no hole is created in the disk but evaporation still goes on continuously . in disks around black holes , as for example in a0620 , for the low accretion rates in quiescence a transition at a distance of about @xmath63 cm from the black hole results .
the transition radius can be estimated without following the disk evolution if one assumes a mass flow rate in the disk .
but the evaluation of radii together with the disk evolution includes the constraints on the evolution and is therefore more firm . taking the transition radius and the rate @xmath61 and calculating the evaporation rate one can compare these results .
this was done in the investigations of meyer ( 1999 ) and also by liu et al .
( 1999 ) . in the latter work
this comparison was also performed for advection - dominated accretion flows onto massive black holes in galactic nuclei .
[ [ the - relation - between - the - mass - overflow - rate - and - the - mass - flow - rate - in - the - corona ] ] the relation between the mass overflow rate and the mass flow rate in the corona ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ for the adaf fit it was assumed that the mass flow rate in the outer cool disk and in the inner very hot spherical region would be the same . since narayan et al . ( 1997 ) found that most of the optical light also originates in the adaf the mass flow rate in the outer disk is not important for the spectrum .
but it is of interest to find out what the mass overflow rate from the companion star is ; in our best example ( model 1b ) it is @xmath64 .
this raises interesting questions about the nature of the magnetic braking mechanism in this binary ( for a discussion see king et al .
1996 and menou et al .
1999a ) . from our computations we find that , rather independent of the time passed after the last outburst , about 1/3 of the matter flowing over from the secondary is stored in the outer disk for the next outburst .
the other 2/3 is evaporated into the corona . from this amount
20 % is lost in a wind and the remaining 80% flows inward in the corona and provides the matter for the adaf in the very hot inner region close to the black hole .
this means that about half of the matter flowing over from the companion star accretes onto the black hole in quiescence . in fig .
6 we show this division of the flow of matter in the cool standard disk and the corona . our results for disk evolution show ( compare table 1 ) that small changes in the mass overflow rate from the companion star affect the outburst recurrence time essentially .
for rates only slightly less than the rate which provides the onset of instability after 59 years ( example a0620 - 00 ) the critical surface density can not be reached , the system is globally stable .
probably many such systems exist , faint persistent soft x - ray binaries .
asai k.,dotani t.,hoshi r. et al . , 1998 ,
pasj 50 , 611 barret d. , mcclintock j.e . , grindlay j.e . , 1996 ,
apj 473 , 963 cannizzo j.k . , 1998 ,
apj 494 , 366 eachus l.j .
, wright e.l .
liller w. , 1976 , apj 203 , l17 elvis m. , page c.g .
, pounds k.a .
et al , 1975 , nature , 257 , 656 esin a.a . ,
mcclintock j.e , narayan r. , 1997 , apj 489 , 865 huang m. , wheeler j.c . , 1989 ,
apj 343 , 229 illarionov a.f . ,
sunyaev r.a . , 1975 ,
a&a 39 , 185 king a.r . , kolb u. , burderi l. , 1996 , apj 464 , l127 kuulkers e. , 1998 , new astronomy reviews 42 , 1 kuulkers e. , 1999 , in : proc . of disk instablities in close binary systems , eds .
s. mineshige and j.c .
wheeler , universal academic press , kyoto , p.169 lasota j .-
p . , narayan r. , yi i. , 1996 , a&a 314 , 813 liu b.f . , meyer f. , meyer - hofmeister e. , 1997 , a&a 328 , 247 liu b.f .
, yuan w. , meyer f. , meyer - hofmeister e. , xie g. , 1999 , submitted to apj liu f.k .
, meyer f. , meyer - hofmeister e. , 1995 , a&a 300 , 823 marsh t.r . , robinson e.l . ,
wood j.h . , 1994 , mnras , 266 , 137 mcclintock j.e . ,
petro , l.d .
, remillard r.a .
, ricker g.r .
, 1983 , apj 226 , l27 mcclintock j.e . ,
remillard r.a . , 1990 ,
baas , 21 , 1206 mcclintock j.e . , horne k. , remillard r.a . , 1995 ,
apj 442 , 358 menou k. , narayan r. , lasota j .-
, 1999a , apj 513 , 811 menou k. , esin a. , narayan r. et al .
, 1999b , apj ( in press ) meyer f. , 1999 , in : proc . of disk instablities in close binary systems , eds
. s. mineshige and j.c .
wheeler , universal academic press , kyoto , p.209 meyer f. , meyer - hofmeister e. 1994 , a&a 288 , 175 meyer - hofmeister e. , meyer , 1999 , in : proc . of disk instablities in close binary systems , eds .
s. mineshige and j.c .
wheeler , universal academic press , kyoto , p.217 meyer - hofmeister e. , meyer f. , liu b.f .
, 1998 , a&a 339 mineshige s. , wheeler j.c . , 1989 , apj 343 , 241 mineshige s. , liu b.f . , meyer f. , meyer - hofmeister e. 1998 , pasj 50 , l5 narayan r. , mcclintock j.e.,yi i. 1996 , apj 457 , 82 narayan r. , barret d. , mcclintock j.e . , 1997 , 482 , 448 narayan r. , mahadevan r. , quartaert e. , 1999 , in : the theory of lack hole accretion discs , eds . m.a
. abramowics et al . , cambridge university press orosz j.a . , bailyn c.d . ,
remillard r.a . , 1994 , apj , 436 , 848 ritter h. , kolb u. , 1998 , a&as 129 , 83 shakura n.i .
, sunyaev r.a . , 1973 ,
a&a 24 , 337 , 1995 , apj 446,741 tanaka y. , shibazaki n. , 1996 , ara&a , 34 , 607 van paradijs j. , mcclintock j.e . , 1995 , in : x - ray binaries , ed.w .. h.g .
lewin , j. van paradijs , e.p.j .
van heuvel , cambridge univ . press ,
white n.e .
, kaluziewski j.l . ,
swank j.h . , 1984 , in : high energy transients in astrophysics , aip conference proc .
woosley , aip new york , p.31 | using the black hole transient x - ray source a0620 - 00 as an example we study the physical interplay of three theoretical constituents for modelling these transient sources : ( 1 ) the advection - dominated accretion flow ( adaf ) onto the central black hole , ( 2 ) the evaporation of matter from the cool outer disk forming a coronal flow and ( 3 ) standard disk evolution leading to outburst cycles by accretion disk instability ( dwarf nova mechanism ) .
we investigate the evolution of accretion disks during quiescence including the evaporation of gas in the inner part of the disk . about 20% of the matter
is lost in a wind from the corona .
the mass flow rate obtained from our model for the coronal flow towards the black hole , is about half of the matter flowing over from the companion star .
it agrees with the rate independently derived from the adaf spectral fits by narayan et al .
( 1997 ) .
about one third of the matter accumulates in the outer cool disk .
the computed disk evolution is consistent with the observational data from the outburst in 1975 .
the evolution of the accretion disk until the instability is reached shows that the critical surface density can not be reached for rates only slightly less than the rate derived here for a0620 - 00 .
systems with such accretion rates would be globally stable , suggesting that many such faint permanently quiescent black hole x - ray binaries exist . |
galaxy formation remains one of the more vexing problems in extragalactic astronomy .
as shown in recent papers such as @xcite , @xcite , @xcite , and @xcite , galaxy simulations that aim to reproduce the observed properties of galaxies and their dark matter halos over a wide range of masses have become increasingly complex and comprehensive .
simulations are probing , among other topics , the seeding and growth of black holes , the chemical enrichment of galaxies , the formation of bulgeless disk galaxies , and the ability of stellar feedback to modify the inner structure of dark matter halos .
the simulations often tackle more than one of these issues simultaneously . to evaluate how realistic the modeled formation processes and resulting simulated galaxies are , having a well - studied and understood sample of real , observed galaxies with which to compare
is necessary .
galaxy observations provide important constraints on the initial input parameters to the simulations and also define the final galactic structure and composition that should be achieved . in this letter ,
we compile data for a sample of dwarf spheroidal and early and late - type disk galaxies and characterize their inner mass structure .
we investigate the amount of total enclosed dynamical mass inside of 500 pc across galaxy type and total baryonic mass .
we also directly compare recent galaxy simulations to the data to determine how well they reproduce observed galaxy structure .
we collect data for a sample of 50 dispersion- and 61 rotation - supported systems .
the dispersion - supported systems are local group dwarf spheroidal ( dsph ) galaxies with measured half - light radii and stellar velocity dispersions .
the data are taken from @xcite , @xcite , and @xcite .
these systems have total luminosities ranging from @xmath1 10@xmath3@xmath410@xmath5 @xmath6 .
the rotation - supported systems are disk galaxies of all morphological types with published rotation curves from @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
these include bright , early - type spirals and low mass dwarf irregulars , and both high and low surface brightness galaxies .
the sample of disk galaxies covers a large range in mass , with @xmath7 ranging from @xmath1 20@xmath4300 km s@xmath8 .
the unifying property of the sample is the availability of dynamical data at small radii . for each dsph and disk galaxy ,
we compile @xmath9 , @xmath10 , @xmath11 , @xmath12 , and @xmath13(@xmath14500 pc ) . for the disk galaxy sample
, we also compile the disk type using the hyperleda database @xcite .
@xmath9 is the total stellar mass of a galaxy .
we assume a stellar mass - to - light ratio @xmath15 = 1 when computing @xmath9 for the dsphs @xcite .
for the disk galaxies , we calculate the total stellar mass using the baryonic tully - fisher relation ( btfr ) of @xcite . for @xmath10 , 1.33@xmath16
is taken as the total gas mass of a galaxy .
@xmath11 , the total baryonic mass of a galaxy , is the sum of @xmath9 and @xmath10 . the effective radius , @xmath12 , is the half - light radius in @xmath17 for the dsph sample , and 1.7 times the disk scale length in @xmath18 or @xmath19 for the disk galaxy sample .
we measure the total dynamical mass ( stars+gas+dark matter ) @xmath13(@xmath14500 pc ) inside of 500 pc .
our primary motivation for choosing 500 pc is so that we can evaluate dsph and disk galaxies at the same scale . as discussed in detail in section [ results ] , this radius is a compromise for both types of galaxies : the half - light radius for dsphs is often less than 500 pc , but resolution and non - circular motions can affect rotation curves at smaller radii . to calculate @xmath13(@xmath14500 pc ) for the dispersion - supported dsph galaxies , we use equation 10 of @xcite @xmath20},\ ] ] and evaluate it at r = 500 pc .
as discussed in the @xcite paper , this equation provides enclosed mass estimates for dispersion - supported galaxies that are comparable to the masses obtained from a full jeans / mcmc analysis if it is applied near " the half - light radius .
the half - light radii of the dsphs in our sample range from tens of pc to @xmath1 1 kpc , with the majority being around 250 pc .
we discuss the applicability / implementation of equation 1 to our dsph sample below .
for eight classical " dsphs in our sample , @xmath13(@xmath14500 pc ) is provided by m.g . walker ( 2013 , private communication ) .
these values are calculated using a complete jeans / mcmc analysis .
we use these data to determine a correction factor , @xmath21 , that we apply to the @xmath13(@xmath14500 pc ) values calculated for the dsph sample using equation 1 .
this correction factor is the ratio of the fully modeled mass to the mass estimated by equation 1 as a function of the difference between 500 pc and the half - light radius : @xmath22 we discuss in section [ m500 ] the typical size of this correction and the uncertainty of @xmath13(@xmath14500 pc ) values calculated for the dsphs using equation 1 .
we lack detailed mass models for some of the assembled disk galaxies , so we assume spherical symmetry : @xmath23 with @xmath24 pc .
this is a tolerable assumption for very late types , which tend to be dominated by their dark matter halos , and for very early types in which the bulge component dominates the inner mass budget . for intermediate types ,
the stellar disk can dominate the mass budget at small radii , but deviations from the spherical approximation are at most tens of percent in mass , a negligible amount over the scales considered .
in figure [ plot1 ] , we plot @xmath13(@xmath14500 pc ) as a function of @xmath9 , @xmath10 , @xmath11 , and @xmath12 for the combined dsph and disk galaxy samples .
we find that @xmath13(@xmath14500 pc ) increases very slowly over many orders of magnitude in the other parameters , and that there is a smooth transition between dispersion - supported systems and rotation - supported systems .
specifically , over @xmath25 orders of magnitude in baryonic mass and @xmath26 in length , the total dynamical mass inside of 500 pc increases only 1.5 orders of magnitude for all but the most massive early - type ( s0-sa ) disk galaxies .
@xmath13(@xmath14500 pc ) increases from a few @xmath2710@xmath28 m@xmath29 for the lowest luminosity dsph galaxies to @xmath30 for sb - sc type disk galaxies in the sample ( see also figure [ plot2 ] ) .
the dynamical mass enclosed by 500 pc is remarkably similar for vastly different galaxies .
figure [ plot1 ] shows that galaxies with different kinds of internal motions ( pressure support vs. rotational support ) , different primary baryonic mass components ( stellar mass vs. gas mass ) , different environments ( satellite galaxies vs. isolated field galaxies ) , different physical sizes , etc .
, are generally rather similar .
they are not identical as we do find a gradual trend for the inner dynamical mass to increase with total stellar mass : @xmath31 .
there is some scatter about this relation , but the majority of galaxies fall within half a dex of this line . the weak correlation defined by the majority of the assembled sample
is violated by the most massive early - type ( s0-sa ) disks .
these have a significantly higher central concentration of mass ( fig . [ plot2 ] ) .
this sets in around @xmath32 with a fairly narrow transition region where galaxies of this mass either lie on or above the relation .
indeed , the data give the visual impression of a gap between the most concentrated early - type galaxies and the rest of the sample .
it is unclear whether this gap is real , or a chance consequence of the sample size .
the sudden change in enclosed central mass of the earliest type galaxies in our sample is real .
the large central masses of early - type galaxies are probably dominated by stars .
this is consistent with the findings of @xcite , from whom most of the early - type data originate . in detailed mass models of early - type systems , it is common to see an inner peak in the rotation curve due to the bulge and a subsequent peak farther out .
the latter might be attributed to either disk or dark matter , but one dark matter halo can not produce both peaks , requiring the inner peak to be dominated by the stellar bulge . the apparent dichotomy in inner mass is suggestive of distinct bulge formation mechanisms between early- and late - type spirals .
the mechanism in early - type spirals is evidently very efficient at packing lots of stars into a small radius .
perhaps these are classical @xmath33 bulges built by mergers , or cold flows are more efficient at delivering baryons to small radii in these massive galaxies @xcite .
contrawise , the less dense central regions of late - type galaxies might grow by secular evolution of the disk .
our results are robust against errors and changes in our assumptions .
uncertainties in @xmath15 for the dsphs or the calibration of the btfr for the disks will change @xmath9 by a factor of a few at most .
this translates to small shifts left or right in figure [ plot1 ] that are insignificant compared to the entire range of @xmath9 that is plotted , particularly given the shallow slope of @xmath13(@xmath14500 pc ) with total stellar mass .
uncertainties in @xmath10 and @xmath12 will produce similarly negligible horizontal shifts in figure [ plot1 ] .
the largest uncertainty is in the measurement of @xmath13(@xmath14500 pc ) .
as discussed in section [ data ] , equation 1 is a good substitute for full modeling when it is used to determine the enclosed mass of a dispersion - supported system at a radius close to the half - light radius . comparing the fully - modeled masses of dsphs provided by walker to the estimates obtained using equation 1
, we find that equation 1 gives values that are typically 1.3 times larger than those determined from the full modeling .
this would be a small vertical shift in figure [ plot1 ] .
our correction factor , @xmath21 , compensates for this difference when @xmath13(@xmath14500 pc ) is calculated for the remainder of the dsph sample .
those dsphs with half - light radii beyond the range of the walker sample ( @xmath1 180@xmath4700 pc ) are least certain .
these are typically the lowest luminosity ( lowest @xmath9 ) systems .
even if we exclude these galaxies , the remaining dsph sample continues the trend set by the disk galaxies , as highlighted by the gray band in the figures .
the uncertainty on @xmath13(@xmath14500 pc ) for the disk galaxies is dependent on how well the rotation velocity at 500 pc can be measured .
500 pc is at the lower limit of the spatial resolution of most rotation curves and non - circular motions become more important at small radii . in addition
, it becomes more difficult to constrain @xmath34(500 pc ) if the galaxy has a bulge since the rotation curve will rise very quickly .
we are therefore forced to exclude galaxies whose rotation curves lack the rising portion at small radius .
these are typically very massive galaxies with @xmath7 in excess of 300 km s@xmath8 .
these systems would populate the very upper right corner of figure [ plot1 ] .
much like the constant mass within 300 pc in the milky way dsphs that was emphasized by @xcite , the total dynamical mass inside of 500 pc appears to be nearly constant in galaxies spanning an enormous range of total baryonic mass , type , and size .
it is only in the earliest type disk galaxies with total stellar masses greater than @xmath1 10@xmath35m@xmath29 that a significant amount of mass builds up in the inner 500 pc . in the next section
, we explore whether recent galaxy simulations are successful in matching the inner regions of observed galaxies . in figure [ plot3 ] , we bin the observed galaxy data and plot the median stellar mass and median @xmath13(@xmath14500 pc ) for each bin .
the dsph galaxies are binned by stellar mass such that each bin contains 10 galaxies .
the disk galaxies are binned in stellar mass by type and contain between 9 and 15 galaxies per bin .
around a few times 10@xmath36 m@xmath29 in total stellar mass , a gradual upturn in the total dynamical mass inside of 500 pc begins to appear .
this upturn is roughly where the baryons begin to dominate the inner mass @xcite .
we also plot in figure [ plot3 ] the total dynamical mass inside of 500 pc for recent galaxy simulations .
these simulations employ feedback prescriptions to modify the inner mass distribution of the galaxy - halo system , prevent the formation of massive bulges , and produce flat rotation curves , exponential surface brightness profiles and dark matter cores .
specifically , we include the simulations of @xcite , the simulated galaxies presented in @xcite , and the simulations of @xcite .
we have chosen these simulations as a comparison sample because @xmath34(500 pc ) is measurable in the published rotation curves ( or @xmath13(@xmath14500 pc ) is tabulated ) and @xmath9 is also provided .
not all recent simulations provide the necessary information to perform our test ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
though @xmath13(@xmath14500 pc ) does not distinguish between luminous and dark mass , the binned galaxy observations do provide a robust measure of the _ total _ inner mass .
@xmath13(@xmath14500 pc ) can therefore be used to gauge how realistic and/or successful the simulations are in reproducing observed total mass distributions .
if they are successful in this regard , one can then ask if they are also able to match the ratios and radial distributions of the different mass components . with this in mind
, we now examine how the simulations compare to the observations in figure [ plot3 ] .
the total mass inside of 500 pc in the simulations with @xmath9 @xmath1410@xmath36 m@xmath29 are in the same range of the central masses of observed galaxies of similar total stellar mass , though typically on the high side .
the feedback processes that are acting in these simulations are keeping the total inner dynamical masses consistent with observations .
these simulations also seem to be making progress in matching the mass profile ( cuspy vs. cored ) of the dark matter halo ( see figure 1 in governato et al .
2012 and figure 6 in brook et al .
2012b ) . the simulations with @xmath9 @xmath37 10@xmath36 m@xmath29
show a similar upturn in @xmath13(@xmath14500 pc ) as the observed galaxies . while the @xcite simulations in this range have significantly _ more _ total mass in their central 500 pc than the observed late - type sd and sc disk galaxies with comparable total stellar masses , the two most massive @xcite simulations are in much better agreement with the observations .
these are their sg3 and sg4 simulations ; most of the merger activity and star formation takes place prior to @xmath38 = 2 in sg3 ( the less massive galaxy of the two ) , while sg4 builds up its mass at a much later time after @xmath38 = 1 .
the @xcite simulation , identified as an sbc / d galaxy , has significantly more mass in its center than observed galaxies of similar type .
this is not an uncommon result .
simulated disk galaxies often have very steep rotation curves , even when star formation , feedback and outflows are modeled ( see , for example , * ? ? ?
* ; * ? ? ?
the decomposition of the @xcite rotation curve into stellar and dark matter mass components shows that it is not just a build - up of baryonic mass , but also a substantial increase in dark matter mass that is responsible for the large total mass at the center : the simulated galaxy contains @xmath1 3 @xmath27 10@xmath39 m@xmath29 of dark matter and 2.1 @xmath27 10@xmath36 m@xmath29 in stars in the inner 500 pc .
it is likely that adiabatic contraction has played a role in creating such a dense center .
the two @xcite simulations with low inner dynamical masses fall in the range of observed sb and sc galaxies .
these simulations utilize high early stellar feedback and high diffusion to regulate / modify star formation .
the five other @xcite simulations have much higher inner dynamical masses , more similar to the bin of early - type s0-sa galaxies . in general , these simulated galaxies have slightly lower total stellar masses than the observed early - type galaxies . alternatively ,
if the morphology of the simulated galaxies is closer to sb - sc , the inner dynamical masses of the simulations are too high , much like the @xcite simulation .
the @xcite simulations cover a narrow range in mass , so it is unclear how their feedback prescription would fare for low mass galaxies .
the baryonic feedback models that are being developed to address observational constraints are making headway towards matching observed galaxies .
there is a tendency for the total inner dynamical masses of the simulations to be larger than those of observed galaxies of similar total stellar mass , with the details depending on the particulars of each simulation .
so far , we have only been able to check the dynamical mass at a particular radius for those simulations that provide the necessary information .
a much stronger test could be made if the radial distribution of baryonic surface density and gravitational force were provided for simulated galaxies @xcite .
we have shown that for a sample of dsph and early- and late - type disk galaxies , the total ( baryons+dark matter ) dynamical mass inside of 500 pc is nearly independent of the overall mass of the galaxy .
there is a weak scaling , @xmath40 . over @xmath1 9
decades in baryonic mass and @xmath1 3 in length , the central mass increases only 1.5 dex .
the earliest type ( s0-sa ) disks are an exception to this weak scaling .
their central masses increase 1.5 dex rather suddenly for @xmath32 .
we speculate that this might be the consequence of a distinct formation mechanism for their dense stellar bulges .
we have also compared the results of recent galaxy simulations to the data to investigate how well the simulations reproduce observed galaxy structure .
the simulations display a slow increase in total inner dynamical mass with increasing total stellar mass that is similar to the trend seen in observed galaxies , though the degree of detailed agreement varies with individual models .
considerably stronger tests could be made if more information were provided by the models .
we would like to thank matt walker for sharing his data .
we also thank the referee for comments that helped to improve the galaxy sample and sample of comparison simulations .
the work of s.s.m . is supported in part by nasa adap grant nnx11af89 g .
we acknowledge the usage of the hyperleda database ( http://leda.univ-lyon1.fr ) .
agertz , o. , kravtsov , a. v. , leitner , s. n. , & gnedin , n. y. 2013 , , 770 , 25 begum , a. & chengalur , j.n .
2004 , , 424 , 509 brook , c.b . ,
stinson , g. , gibson , b.k . ,
roskar , r. , wadsley , j. , & quinn , t. 2012a , , 419 , 771 brook , c.b . , stinson , g. , gibson , b.k .
, wadsley , j. & quinn , t. 2012b , , 424 , 1275 collins , m.l.m . , chapman , s.c . , rich , r.m .
, et al . 2013 , , 768 , 172 conroy , c. & van dokkum , p.g .
2012 , , 760 , 71 de blok , w.j.g . & bosma , a. 2002 , , 385 , 816 de blok , w.j.g . ,
mcgaugh , s.s . , & rubin , v.c .
2001 , , 122 , 2396 de blok , w.j.g . , walter , f. , brinks , e. , et al .
2008 , , 136 , 2648 de zeeuw , p.t . , bureau , m. , emsellem , e. , et al .
2002 , , 329 , 513 dekel , a. , birnboim , y. , engel , g. , et al .
2009 , , 457 , 451 di cintio , a. , brook , c.b . ,
maccio , a.v . ,
2013 , , submitted ( arxiv:1306.0898 ) governato , f. , zolotov , a. , pontzen , a. , et al .
2012 , , 422 , 1231 guedes , j. , callegari , s. , madau , p. , & mayer , l. 2011 , , 742 , 76 hopkins , p. f. , kere , d. , murray , n. , quataert , e. , & hernquist , l. 2012 , , 427 , 968 kuzio de naray , r. , mcgaugh , s.s .
, & de blok , w.j.g .
2008 , , 676 , 920 martin , n.f . ,
de jong , j.t.a . , &
rix , h.w .
2008 , , 684 , 1075 mcconnachie , a.w .
2012 , , 144 , 4 mcgaugh , s. s. 2004 , , 609 , 652 mcgaugh , s.s .
2012 , , 143 , 40 mcgaugh , s.s .
, de blok , w.j.g . , schombert , j.m .
, kuzio de naray , r. , & kim , j.h . 2007 , , 659 , 149 noordermeer , e. , van der hulst , j.m . , sancisi , r. , swaters , r.s .
, & van albada , t.s .
2007 , , 376 , 1513 oh , s .- h . ,
de blok , w.j.g . , brinks , e. , walter , f. , & kennicutt , r.c .
2011 , , 141 , 193 paturel g. , petit c. , prugniel p. , theureau g. , rousseau j. , brouty m. , dubois p. , cambrsy l. , 2003 , a&a , 412 , 45 piontek , f. & steinmetz , m. 2011 , , 410 , 2625 portinari , l. , sommer - larsen , j. , & tantalo , r. 2004 , , 347 , 691 sanders , r.h .
1996 , , 473 , 117 sofue , y. , tutui , y. , honma , m. , tomita , a. , takamiya , t. , koda , j. , & takeda , y. 1999 , , 523 , 136 stinson , g.s . , brook , c. , maccio , a.v .
, wadsley , j. , quinn , t.r . , & couchman , h.m.p .
2013 , , 428 , 129 strigari , l.e . ,
bullock , j.s . ,
kaplinghat , m. , simon , j.d . ,
geha , m. , willman , b. , & walker , m.g .
2008 , , 454 , 1096 trachternach , c. , de blok , w.j.g .
, mcgaugh , s.s .
, van der hulst , j.m . , & dettmar , r.j .
2009 , , 505 , 577 verheijen , m.a.w . , & sancisi , r. 2001 , , 370 , 765 vogelsberger , m. , genel , s. , sijacki , d. , torrey , p. , springel , v. , & hernquist , l. 2013 , , submitted ( arxiv:1305.2913 ) walker , m.g . ,
mateo , m. , olszewski , e.w .
, et al . 2007 , , 667 , l53 walker , m.g . , mateo , m. , olszewski , e.w . ,
penarrubia , j. , evans , n.w . , & gilmore , g. 2009 , , 704 , 1274 walker , m.g .
, mcgaugh , s.s . ,
mateo , m. , olszewski , e.w .
, & kuzio de naray , r. 2010 , , 717 , l87 | we compile a sample of dispersion - supported and rotation - supported galaxies for which the dynamical mass enclosed at 500 pc can be measured .
we find that this dynamical quantity increases only slowly with stellar mass ( @xmath0 ) over @xmath1 9 decades in baryonic mass and @xmath1 3 in length scale , with a sudden upturn at the highest masses ( @xmath2 ) .
this upturn occurs for the earliest type ( s0 and sa ) disk galaxies , and is consistent with the `` extra '' mass within 500 pc being predominantly stellar .
this sudden change may be indicative of different bulge formation mechanisms between early and late type disks .
we also compare the data to the results of recent galaxy formation simulations .
these broadly track the observed trend , with some tendency to put too much mass in the inner 500 pc , depending on the details of each simulation . |
over the past forty years , black - hole physics has become quite a mature field @xcite . driven by a desire to understand their fundamental properties and potential importance to astrophysics ,
the study of black holes has produced a number of outstanding results .
for example , hawking has proven that black holes never bifurcate and that the areas of black - hole event horizons never decrease to the future @xcite .
further , through the efforts of a number of researchers , the black hole `` no - hair '' theorems @xcite have established that stationary electrovacuum black holes are far simpler objects than were once imagined @xcite . in all of these studies ,
the interiors of black holes have been entirely ignored , as well they should as these regions neither affect nor are observable from the asymptotic region of the spacetime ( by their very definition ) .
so , while a great deal has been learned about black - hole exteriors , comparatively little is known about black - hole interiors .
are there any features that black - hole interiors share ? there are at least two results in this direction . first , penrose has shown that if a spacetime contains a future trapped surface then , provided the cosmic censorship hypothesis holds @xcite , it must lie within the black - hole region and that such a spacetime must be singular in the sense that it can not be future null geodesically complete @xcite . while a truly amazing and important theorem in the history of black - hole physics , it tells us very little about the extent to which the spacetime is singular . are just a few , most , or all of the geodesics in the interior region incomplete ?
how long does it take for such a singularity to develop ?
second , a number of people using a variety of methods ( see @xcite and the references therein ) , have examined whether the cauchy horizons that occur in the interiors of the kerr - newman black holes will be present in more `` realistic '' spacetimes . in particular , poisson and israel have argued that these horizons are destroyed ( generically ) due to an unbounded `` inflation '' of the mass function @xmath4 on what was the cauchy horizon inside the black hole @xcite . although arguments have been given that indicate that these results are quite general , a proof remains to be found @xcite . here ,
in the effort to explore the features of black - hole interiors , we examine a single issue : how long can the journey of an observer in a black hole last ?
must it be finite ?
if so , can an upper bound on the length of such a journey be given in terms of some characteristic of the black hole ?
it turns out that there are two forms of this problem that are best considered separately .
first , we restrict ourselves with observers that enter a black hole from the asymptotic region . that is , we are restricting our attention to the portion of a black - hole region lying to the future of past null infinity @xmath5 .
second , we drop this restriction and look at the entire black - hole region .
however , we shall see that this generalization does come at a small price .
a restriction on the cauchy surface topology is necessary ( while none appears necessary for the first case ) along with a weaker bound on the lifetimes of observers within the black - hole region .
consider first the case of an observer entering a black hole from an asymptotic region @xcite @xmath6 also known as the domain of outer communications @xcite . in the maximally extended schwarzschild spacetime , this is region ii depicted in fig .
[ schwarz1 ] . a partial answer to our problem can obtained from the work of friedman , schleich , and witt which can be stated as : in an asymptotically flat spacetime satisfying the null - convergence condition @xcite with distinct null infinities @xmath7 , the sets @xmath8 are empty for @xmath9 @xcite .
therefore , fixing one null infinity @xmath10 it is impossible for an observer to leave the asymptotic region @xmath11 by entering a black hole and then live indefinitely by escaping to another asymptotic region .
the maximally extended schwarzschild spacetime depicted in fig .
[ schwarz1 ] exemplifies this impossibility .
although this result is very encouraging , it does not eliminate other possibilities for an observer existing indefinitely within a black hole .
based on the general considerations presented in sec .
[ sec : formulating ] and its successful proof in the spherically symmetric case ( theorem 1 below ) , we conjecture that this result can indeed be strengthened . _ conjecture 1 . _
fix a globally hyperbolic asymptotically flat spacetime satisfying the dominant - energy and non - negative - pressures conditions @xcite with a non - empty black - hole region .
fix a connected component @xmath12 of the event horizon and denote the supremum of its area as measured over the set of all of its spatial cross - sections by @xmath13 .
then , there exists a constant @xmath14 ( independent of which spacetime and event horizon is chosen ) such that the lifetime of any observer within the black hole is bounded above by @xmath15 provided that the black hole is entered by crossing @xmath12 in the future of @xmath16 .
recall that the irreducible mass associated with a connected component of black - hole s event horizon at a given `` time '' is obtained by calculating its surface area @xmath1 on a spatial cross - section and setting @xcite @xmath17 the quantity is `` irreducible '' as it is non - decreasing to future , which follows from the fact that the area @xmath1 of the event horizon is non - decreasing to the future @xcite .
so , with this definition , conjecture 1 bounds the lifetime of observers within a black hole by a multiple of the supremum of the irreducible mass calculated on the black - hole s event horizon ( provided the black hole is entered from the asymptotic region ) . to explore the features of conjecture 1 ,
consider a black - hole spacetime associated with two distant stars that each collapses forming black holes .
should these two black holes never collide with one another , the event horizon will then have two connected components and conjecture 1 asserts that the lifetime of an observer that enters one of the black holes ( to the future of @xmath16 ) is bounded by a quantity computable from the geometry of the event horizon of the black hole entered .
however , should the two black holes collide and coalesce into a single black hole , the event horizon will then have a single connected component . in this case , the bound given by conjecture 1 is insensitive as to whether the black hole was entered before or after coalescence and to which black hole was entered .
so , the bound is the same for observers entering a small black hole and a large black hole as long as the two eventually coalesce .
this is probably a gross overestimate in the case of an observer entering the smaller black hole before coalescence and probably can be refined . next , suppose we slightly modify conjecture 1 by having it place a bound on the lifetime of observers in the portion of the black hole lying to the future of @xmath16 .
fix any two points @xmath18 and @xmath19 in this region with @xmath20 and fix any causal curve @xmath21 from @xmath16 to @xmath18 . then , as @xmath21 must intersect the event horizon somewhere , conjecture 1 gives an upper bound on the lengths of all causal curves from @xmath18 to @xmath19 ( that is independent of @xmath18 of @xmath19 ) , thereby bounding the lifetimes of all observers in this region .
however , suppose that the black hole is sufficiently complicated that its boundary ( part if not all of the total event horizon ) is not connected .
for example , we could imagine a spacetime containing a wormhole with its mouths sufficiently separated and moving in such a way that the event horizons enclosing each mouth never coalesce .
then , if @xmath18 can be reached from @xmath16 by two causal curves that intersect different horizons ( i.e. , thread different mouths ) , we would have a choice of which component to use in establishing an upper bound to the lifetimes of observers therein .
can we use the smaller of the two , or must we always use the larger ? or should we use a bound based on total area of the event horizons associated with our region ?
we leave this problem open for investigation . in conjecture 1 ,
we have limited ourselves to observers entering a black hole from an asymptotic region .
might the bound given by conjecture 1 hold for all observers within a given black hole ?
clearly not , if we do not put some restriction on the cauchy surface topology as the maximally extended schwarzschild spacetime provides a simple counterexample . here
, the problem is that there as an entire asymptotic region within the black hole .
( see fig .
[ schwarz1 ] . ) even if we restrict our consideration to spacetimes with cauchy surfaces having the topology of a compact orientable connected three - manifold @xmath22 minus a point ( so that there are no other asymptotic regions to escape to ) then , as is demonstrated in sec .
[ sec : topology ] , if we do not further restrict @xmath22 , asymptotically flat black - hole spacetimes can be constructed with `` compact internal regions '' and yet admitting infinitely long timelike curves in the black - hole region .
furthermore , as is demonstrated in sec .
[ sec : formulating ] , even if we restrict ourselves to spacetimes with @xmath0 cauchy surfaces , there is no @xmath14 for which the bound given by conjecture 1 will hold . however , here the only problem appears to be that the bound is not quite appropriate .
the idea is that the bound in conjecture 1 is constructed from the geometry of the event horizon and therefore can be calculated without a knowledge of the interior of the black hole . for more general observers ,
a knowledge of the interior geometry ( at least on a cauchy surface ) is necessary in obtaining a bound on the lifetimes of observers therein .
based on these considerations and its successful proof in the spherically symmetric case ( theorem 2 below ) , we are led to the following conjecture .
_ conjecture 2 .
_ fix a globally hyperbolic asymptotically flat spacetime satisfying the dominant - energy and non - negative - pressures conditions @xcite . then , provided the cauchy surfaces are diffeomorphic to @xmath0 , there is a number @xmath14 ( independent of which spacetime is chosen ) and a quantity @xmath23 computable from the initial data on a cauchy surface @xmath24 , such that the lifetimes of all observers within the black - hole region is bounded above by @xmath25 . in this conjecture , @xmath23 is intended to represent the least upper bound to a quasilocal mass function ( associating a number with each given spacelike two - sphere in a spacetime ) over the set of all two - spheres in @xmath24 . as we have not specified
what quasilocal mass function is to be used ( see however ref .
@xcite and the references therein ) , this portion of the conjecture is imprecise and we leave the task of making it precise as part of the problem .
however , when we specialize to the spherically symmetric spacetimes , we expect this mass to be simply @xmath26 , where @xmath4 is the mass function constructible in any spherically symmetric spacetime as given by eq .
( [ defm ] ) in sec .
[ sec : elementaries ] . for simplicity ,
we have limited the cauchy surface topology in conjecture 2 to @xmath0 . in sec .
[ sec : topology ] , we generalize this conjecture by broadening the class of spatial topologies .
furthermore , a theorem is presented which states ( roughly ) that the lengths of causal curves in the portion of the black hole not reachable from the asymptotic region are bounded provided the spacetime admits a maximal cauchy surface .
although this is far from a proof of conjecture 2 , it is encouraging .
as the calculation of @xmath13 in conjecture 1 requires a knowledge of the geometry of spacetime in the far future ( and therefore can be quite difficult to calculate in practice ) one might worry that @xmath13 could be arbitrarily large ( or even infinite ) thereby making the bound given by conjecture 1 rather weak .
however , from an analysis of the spherically symmetric case ( see theorem a1 in appendix a ) , we expect the three masses @xmath27 , @xmath2 , and @xmath23 to be related according to the following inequalities @xmath28 whether the second inequality actually holds can not be investigated until @xmath23 is given a precise definition . however
, since the total area @xmath29 of the black hole at a given `` time '' is the sum of the areas associated with each connected component , the first inequality always holds provided that @xmath30 .
( this inequality is similar to , though distinct from , the gibbons - penrose isoperimetric inequality for initial data sets in general relativity which has its inspiration from this inequality @xcite . ) that such an inequality should hold can be argued as follows .
consider this inequality in an asymptotically flat electrovacuum spacetime whose event horizon is connected ( i.e. , so there is only one black hole at late times ) .
as the area of the event horizon is non - decreasing to the future , this inequality is most difficult to satisfy at late times .
however , it is expected that the black hole will eventually settle - down and approach a stationary state which , by the `` no - hair '' theorems , must be one of the kerr - newman spacetimes where it is known that the inequality does indeed hold .
( as far as the author is aware , whether this inequality always holds is unresolved .
theorem a1 in appendix a proves it does hold for the spherically symmetric spacetime satisfying the dominant - energy condition . )
therefore , should the first inequality in eq .
( [ mass_ineq ] ) indeed hold , then the bound given by conjecture 1 will always be finite . furthermore , independent of this inequality , an alternate version of conjecture 1 can be obtained by replacing the bound by @xmath31 .
should the irreducible mass always be bounded by the adm mass ( as expected ) , then conjecture 1 would imply this alternate version .
conjectures 1 and 2 are similar to the conjecture that closed universes possessing @xmath32 or @xmath33 cauchy surfaces ( and obeying the dominant - energy and non - negative - pressures conditions ) must have finite lifetimes @xcite , which is a slight variant of the closed - universe recollapse conjecture @xcite .
in fact , the similarity is more than superficial .
the idea is that the requirement that the cauchy surfaces be compact in the closed - universe finite lifetime conjecture can probably be replaced by the more general requirement that the geometry on cauchy surfaces be bounded in an appropriate sense . in the spherically symmetric case
we can make this precise .
if we replace the requirement that the cauchy surfaces be compact with bounds on @xmath34 and @xmath4 , we can again establish a finite lifetime for the spacetime .
( actually , if this is done , the final bound on the lifetime may be slightly different from that stated in ref . @xcite . ) with this in mind , consider the collapse of a star leading to the formation of a black hole as depicted in fig . [ collapse ] .
the black - hole region @xmath35 of this spacetime is globally hyperbolic with its boundary @xmath36 ( the event horizon of the black hole ) acting as a cauchy surface for this region .
although @xmath37 is non - compact it does enjoy the property that the geometry thereon is essentially bounded .
in particular , the area of the black hole ( as measured on the event horizon ) is bounded by its future asymptotic value .
therefore , it would seem that proofs of these two conjectures would be very similar and that one could be obtained from the other with only slight modifications being required . indeed , a proof of conjectures 1 and 2 for the spherically symmetric case can be obtained using the techniques developed in the course of the proof that spherically symmetric closed universes have finite lifetimes . in fact , we prove somewhat more .
_ theorem 1 .
_ fix a globally hyperbolic asymptotically flat spherically symmetric spacetime satisfying the dominant - energy and non - negative - pressures conditions @xcite and possessing a cauchy surface that is geodesically complete .
then , the lifetime of an observer in the black hole is bounded above by @xmath38 provided either : ( a ) the black hole is entered from the asymptotic region of the spacetime ; or ( b ) the spacetime possesses a spherically symmetric cauchy surface with @xmath39 everywhere outgoing ; or ( c ) the observer lies to the future of the component of a cauchy surface connected to @xmath40 on which @xmath39 is outgoing . note that in the spherically symmetric case , on a sphere of symmetry of area @xmath41 , @xmath27 is simply @xmath42 . as shown by theorem a1 in appendix
a , provided the dominant - energy condition holds , for spheres on the event horizon @xmath43 ( i.e. , @xmath42 on the event horizon is always bounded above by @xmath2 ) .
therefore , the bound in theorem 1 can always be replaced by @xmath44 thereby given a bound that is calculable from the initial data on a cauchy surface .
further , provided that the event horizon is complete in the sense given by theorem a2 in appendix a , then @xmath45 and @xmath4 have the same future asymptotic limits on the event horizon .
therefore , in this case , the bound in theorem 1 is equivalent to @xmath46 times the future asymptotic value of the mass @xmath4 on the event horizon .
each part of theorem 1 has its own strengths and weaknesses .
part ( a ) of this theorem is best used when we know that the observer is entering the black hole from the asymptotic region . however , suppose we are given the location of an observer on a ( spherically symmetric ) initial data surface @xmath24 ( a cauchy surface ) and that he enters the black hole .
do we need to find the rest of the spacetime from the initial data on @xmath24 to find out whether the observer was indeed in the asymptotic region ?
if @xmath39 is outgoing on all of @xmath24 , then by ( b ) the length of his ( and indeed all observers ) journey in the black hole is bounded by a multiple of the future asymptotic irreducible mass of the black hole . however , suppose that @xmath39 is not outgoing on all of @xmath24 ( i.e. , @xmath24 contains a future or past outer trapped surface ) .
then , provided our observer begins his journey on the portion of @xmath24 connected to @xmath40 on which @xmath39 is outgoing , then , by ( c ) , his lifetime within the black hole is still bounded by a multiple of the future asymptotic irreducible mass of the black hole . in theorem 1 , we have demanded that the spacetime admit a cauchy surface that is geodesically complete .
this is automatically satisfied by asymptotically flat cauchy surfaces with @xmath0 topology , but not may not be satisfied by those with @xmath47 topology .
we make this requirement as it guarantees that @xmath4 will be everywhere non - negative , which is needed for the method of proof used herein ( see lemma 1 ) .
it is likely , though not for certain , that the proof can be improved upon , allowing us to lift the requirement that a cauchy surface be complete .
the next theorem gives an upper bound on the lifetimes of observers within the black - hole region of a spherically symmetric spacetime that makes no requirements on how the black hole was entered nor on how @xmath39 behaves on a cauchy surface .
note that the price for this is that a restriction on the cauchy topology is necessary and that the bound need not be as tight as that given by theorem 1 .
_ theorem 2 .
_ fix an asymptotically flat spherically symmetric spacetime that possesses @xmath0 cauchy surfaces and that satisfies the dominant - energy and non - negative - pressures conditions @xcite . then , for any cauchy surface @xmath24 , the lifetime of an observer in the black - hole region @xmath35 with event horizon @xmath48 is bounded above by @xmath49 where @xmath50 is the subset of @xmath24 on which @xmath39 is past - directed timelike , past - directed null , or zero .
to understand the features of the bound given by theorem 2 , consider the black hole formed by the collapse of a spherical star from a state that is benign in the sense that on a spherically symmetric cauchy surface @xmath24 there are no future or past outer trapped surfaces ( so in particular @xmath39 is outgoing everywhere thereon ) .
then , @xmath50 is empty , the maximum of @xmath4 on @xmath51 occurs on its boundary ( which lies on @xmath37 ) where it is no greater than @xmath27 .
so , in this case , theorem 2 reproduces the bound given by part ( b ) of theorem 1 .
if , however , the spacetime contains no such cauchy surface ( see for example the black - hole spacetime given by fig .
[ db2 ] ) , then @xmath52 can be larger than @xmath53 . as discussed in sec .
[ sec : formulating ] , this larger bound is necessary .
furthermore , as @xmath54 is a subset of @xmath55 , this quantity is constructed from the geometry on the portion of @xmath24 that is not in the asymptotic region .
lastly , we note that the rather complicated bound given by theorem 2 can always be replaced by the simpler but cruder bound @xmath56 as @xmath57 is a subset of @xmath24 and @xmath58 .
( note that this theorem suggests a strengthening of conjecture 2 . )
the bound given by theorem 1 has the feature that it increases as matter flows through the event horizon into the black hole thereby making it larger .
therefore , it might seem that we could help an observer that has entered a black hole live longer by gathering up all the material available to us outside the black hole and throw it in after him .
does increasing the black - hole s mass in this way really help the observer ? in sec .
[ sec : do ] , we formulate and study a precise version of this problem .
we find that while generally the lifetime of the observer can be increased , there are circumstances in which our effectiveness is limited and others in which there is nothing that can be done to help the unlucky traveler .
the remainder of this work is organized as follows . in sec .
[ sec : formulating ] , we go through the steps leading to the formulation of conjectures 1 and 2 . in sec . [ sec : topology ] , we generalize the class of cauchy surface topologies for which we believe conjecture 2 holds and prove a theorem which bounds the lengths of causal curves lying `` deep within '' a black hole in a spacetime possessing a maximal cauchy surface . in sec .
[ sec : spherical ] , we provide the proofs of theorems 1 and 2 . in sec .
[ sec : do ] , we investigate the problem of whether we can increase the lifetime of an observer within a black hole .
lastly , in sec . [ sec : discussion ]
, we make a few final remarks and discuss the hopes for extending the results presented .
our conventions are those of ref .
in particular , metrics are such that timelike vectors have negative norms and the riemann and ricci tensors are defined by @xmath59 } \omega_c = r_{abc}{}^d \omega_d$ ] and @xmath60 , respectively .
all metrics are taken to be @xmath61 . for two sets @xmath1 and @xmath35
, @xmath62 denotes the elements in @xmath1 that are not in @xmath35 ( i.e. , the set difference ) , @xmath63 denotes the elements that are not in @xmath1 ( i.e. , its complement ) , and @xmath64 denotes the boundary of @xmath1 .
lastly , our units are such that @xmath65 .
an observer beginning a journey from the asymptotic region of a maximally extended schwarzschild spacetime of positive mass @xmath66 ( region i in fig . [ schwarz1 ] ) into the black hole ( regions ii and iv ) must enter region ii .
as is well known , this journey within the black hole must end in a time no greater than @xmath67 as the observer will be crushed as he approaches the singularity within .
does a similar result hold more generally ? consider the following conjecture
. _ conjecture 3 ( false ) .
_ there is a constant @xmath14 such that the lifetime of any observer in a black hole of mass @xmath66 is no greater than @xmath68 .
we begin by testing this conjecture against the stationary ( electrovacuum ) black - hole spacetimes . for a kerr - newman black hole of mass @xmath66 , angular momentum @xmath69 , and electric charge @xmath70 , with @xmath71
, we immediately realize that some care is needed in addressing this problem . here
, an observer can travel into the black hole and cross the cauchy horizon associated with a cauchy surface @xmath24 that extends to spacelike infinity at both ends .
( see fig .
28(i ) of ref .
@xcite or fig .
12.4 of ref .
@xcite . ) after crossing this horizon , he can travel to another asymptotically flat region and thereby live an indefinitely long time .
however , that such a journey can be achieved is perhaps not surprising once we realize that the observer , having gone outside of the domain of dependence of @xmath24 , is in a region that is unpredictable ( from the initial data on @xmath24 ) .
it has been conjectured ( the cosmic censorship conjecture @xcite ) and investigations have indicated that such cauchy horizons do not occur for `` realistic '' spacetimes @xcite .
so , while a few spacetimes will have such horizons beyond which the journey of an observer could continue ( such as kerr - newman with @xmath72 ) this will usually not be the case as the spacetime will become sufficiently singular that the journey will be inextendible .
( just how singular it will become has become a matter of debate @xcite . )
it is because of such counterexamples that we have restricted ourselves to globally hyperbolic spacetimes in conjectures 1 and 2 .
should either conjecture hold , we then have an upper bound on the time that an observer can remain within the domain of dependence of @xmath24 once inside the black hole . whether the spacetime can be extended ( so that the journey of the observer can continue ) or otherwise ( so that the journey is at an end ) is then a matter for further investigation . restricting ourselves to the globally hyperbolic portion of the kerr - newman spacetimes ( with cauchy surfaces extending to spacelike infinity in both asymptotic regions ) we find that conjecture 3 holds with @xmath66 being the irreducible mass associated with the event horizon ( which is constant thereon ) and @xmath73 .
( our reasons for identifying @xmath66 with @xmath27 and not @xmath2 or some other quantity will be given below . ) therefore , the kerr - newman spacetimes obey conjecture 1 provided @xmath73 .
furthermore , these spacetimes show that @xmath74 is the least value of @xmath14 for which the conjecture can hold as it can be shown that the ratio of the maximum lifetime of an observer within the portion of the black hole lying to the future of @xmath75 and @xmath27 , while always less than @xmath74 , approaches @xmath74 in a sequence of kerr - newman black holes where @xmath76 ( i.e. , the ratio approaches @xmath74 in the limit that the black hole becomes extreme ) . in conjecture 3 ,
no conditions are stated on the material content of the spacetime .
without such a condition we can easily counterexample the conjecture as follows . from an extended schwarzschild spacetime of positive mass , construct a new spacetime by multiplying the schwarzschild metric by a conformal factor that is unity outside the black - hole region ( i.e. on regions i and iii in fig . [ schwarz1 ] ) .
then , as the two spacetimes share the same causal structure and agree outside the black - hole region of schwarzschild , the share the same black - hole region .
by choosing the conformal factor appropriately , we can quite easily ensure the existence of arbitrarily long timelike curves in the black - hole region ( in particular , region ii ) .
we conjecture that spacetimes with `` ordinary '' matter do not exhibit this type of behavior .
just what restrictions on the matter content is appropriate is an issue for investigation , however , in conjectures 1 and 2 we have imposed the dominant - energy and non - negative - pressures conditions @xcite .
as is demonstrated by theorems 1 and 2 , together these conditions are sufficient for establishing a theorem in the spherically symmetric case .
it is likely that these conditions can be weakened as the kerr - newman spacetimes satisfy the conclusions of conjecture 1 ( with @xmath77 ) without satisfying the non - negative - pressures condition ( in the @xmath78 case ) .
further , in conjecture 3 there is no restriction on how the observer enters the black hole .
this is important as the black - hole region associated with the asymptotic region i of schwarzschild is all of regions ii and iv in fig . [ schwarz1 ] .
in particular the other asymptotic end is inside the black hole ( with our definition of black hole ) .
so , clearly there are infinitely long lived observers therein .
we need some type of condition to rule out this and similar counterexamples .
is it enough to restrict ourselves to observers that begin their journey from outside the black hole ?
consider an observer that begins his journey in region iii .
while initially outside the black hole , he can go inside and reach the other asymptotic region and thereby live forever .
however , an observer that begins in the asymptotic portion , region i , must enter region ii upon entering the black hole wherein his lifetime must end in a finite time .
so , one way to take care of this problem is to restrict ourselves to observers who enter the black hole from `` sufficiently far away '' .
for example , we can restrict ourselves to observers that begin their journey somewhere in the asymptotic region @xmath79 ( region i in fig .
[ schwarz1 ] ) as we have done in conjecture 1 . or ,
somewhat more weakly , we can restrict ourselves to observers that lie to the future of @xmath75 .
a second possibility is to restrict ourselves to spacetimes with @xmath0 cauchy surfaces as we have done in conjecture 2 .
this way , any possibility of escaping to another asymptotic region is eliminated _
ab initio_. ( more general topologies are considered in sec . [ sec : topology ] . )
lastly , in conjecture 3 , we have not specified what is meant by the mass @xmath66 .
should we use the irreducible mass ( as we have done in conjecture 1 ) , the adm mass of the spacetime , a quasilocal mass constructed on the event horizon of the black hole , a quantity constructed from the initial data ( as we have done in conjecture 2 ) , or something else ? to investigate this issue , consider the black - hole spacetimes associated with the gravitational collapse of uniform density `` dust ball '' such as that depicted in fig .
( these are the spacetimes considered by oppenheimer and snyder in their pioneering work on gravitational collapse @xcite . )
these dust - ball spacetimes are easily constructed by `` gluing '' a portion of a dust - filled robertson - walker spacetime to schwarzschild @xcite and we refer the reader to appendix b for a brief review of their construction and the establishment of notation for the following .
note that for these spacetimes , the mass function @xmath4 coincides with the adm mass in the vacuum region and since all of the matter eventually goes into the black hole , @xmath27 and @xmath4 both have @xmath2 as future asymptotic limits on the event horizon .
so , while these spacetimes can not distinguish which of these three quantities should be used in conjecture 3 , they can test whether one ( and hence all ) will fail to work . to see why all three are inappropriate for conjecture 2 , fix a dust - ball spacetime constructed from a @xmath80 robertson - walker spacetime ( so the dust ball is gravitational bound ) that is described by the constants @xmath22 and @xmath81 .
( see fig .
[ db2 ] . ) here , the black hole is the union of regions @xmath82 , ii , @xmath83 , @xmath84 , and iv . for observers within the black hole , not necessarily entering from region i , it can be argued that the maximum lifetime that can be spent within the black hole is attained by the curve corresponding to the center of the dust ball .
therefore , @xmath85 now consider a sequence of such spacetimes in which @xmath86 . in this limit , the coefficient of @xmath2 in eq .
( [ coeff ] ) diverges .
therefore , this sequence of spacetimes shows that , provided that we always identify @xmath66 as the adm mass of the spacetime ( or the future asymptotic values of @xmath27 or @xmath4 on the event horizon ) , there is no constant @xmath14 for which the above conjecture will hold .
this is not surprising once we realize that in such a sequence ( with fixed @xmath22 ) the maximum lifetime of observers inside the black hole , given by eq .
( [ life_c ] ) , is fixed , while the mass of the black hole , as measured using the adm mass given by eq . ( [ dust - ball - adm - mass ] )
is becoming arbitrarily small in this limit .
this shows , in particular , that an upper bound on the lifetime of observers inside of a black hole can not , in general , be determined by examining the spacetime outside of the black hole . with @xmath87 close to @xmath88 , the size of the spheres of symmetry inside the black hole and the amount of matter
therein can be much greater than would be guessed by examining the mass of the black hole from the outside .
indeed , the spacetime within the black hole acts very much like a closed universe ( in particular like a @xmath80 robertson - walker spacetime ) except that it is connected to an asymptotically flat spacetime by a small `` throat '' .
so , conjecture 2 would fail if we identified @xmath23 with @xmath2 or the future asymptotic values of @xmath27 or @xmath4 on the event horizon .
is there some other quantity that we can construct that is sensitive to the geometry inside of the black hole ?
note that for these spacetimes with @xmath80 and @xmath89 , the mass @xmath4 is not maximal on the outer boundary of the dust ball .
it is maximal on the surface @xmath90 .
fixing a spherically symmetric cauchy surface @xmath24 in our spacetime and using eq .
( [ m_robertson - walker ] ) , we see that the supremum of @xmath4 over @xmath24 is simply @xmath91 . therefore , by eq .
( [ life_c ] ) , for these spacetimes we have @xmath92 so , conjecture 2 does hold for these spacetimes provided that we take @xmath23 to be the supremum of @xmath4 over a cauchy surface and @xmath93 . analyzing the remaining dust - ball spacetimes
, we again find that conjecture 2 holds provided that @xmath94 and @xmath95 .
( note that this lower bound on the allowed values of @xmath14 is not least .
finding the least upper bound to the lengths of timelike curves in the black holes of the dust - ball spacetimes with @xmath96 or @xmath80 and @xmath97 does not appear to be an easy task .
for instance , is the geodesic at the center of the dust ball longest ? with some work it can probably be shown that the conjecture holds for all of these spacetimes with @xmath98 . ) from these observations and in view of theorem 2 , it would seem that the proper identification for @xmath23 conjecture 2 would be something that generalizes @xmath26 .
the most obvious candidate is to generalize @xmath4 by some quasilocal mass function and then find its supremum over the set of all two - spheres on a cauchy surface @xmath24 . however , whether this is truly appropriate we shall leave as a matter for future investigation ( i.e. , finding such a quantity for conjecture 2 and proving it , or somehow show that there is no such notion for which conjecture 2 will hold ) . analyzing the dust - ball spacetimes ( @xmath80 , @xmath99 , and @xmath96 ) we find that conjecture 1 always holds with @xmath100 .
( again , this lower bound on @xmath14 is not least and it is likely that the conjecture holds for all @xmath93 . )
so , from these observations , it would seem that @xmath2 or the future asymptotic values of @xmath27 or @xmath4 could be appropriate for the mass in conjecture 1 .
it is in the proof of conjecture 1 in the spherically symmetric case ( i.e. , theorem 1 ) that we discover that @xmath27 is the most natural candidate , which is why this is the quantity chosen for conjecture 1 .
in this section we investigate the problem of relaxing the restriction to @xmath0 cauchy surfaces in conjecture 2 .
as we are interested in asymptotically flat spacetimes , we begin by restricting ourselves to manifolds where all of the `` interesting topology '' is bounded away from infinity .
we make this idea precise by demanding that the cauchy surfaces have the topology of a connected , orientable , three - manifold without boundary @xmath22 less a point which we label @xmath40 .
the idea here is that deleted point @xmath40 `` represents '' spatial infinity .
so , outside of a sufficiently large sphere ( i.e. , inside of a two - sphere sufficiently near @xmath40 ) the space has the topology @xmath101 . further , to exclude the possibility of escaping to other asymptotic regions , as in the maximally extended schwarzschild spacetime , we demand that @xmath22 be compact .
this way all of the non - compactness of a cauchy surface @xmath24 is due to the single asymptotic region .
we summarize these requirements by demanding that @xmath102 , where @xmath22 is a connected , orientable , closed ( compact without boundary ) three - manifold . as a simple example
, @xmath103 is of this form with @xmath104 . restricting ourselves to spacetimes with cauchy surfaces
diffeomorphic to @xmath105 , does conjecture 2 hold for all closed three - manifolds @xmath22 ?
remarkably , the answer is no .
we can construct a spacetime meeting the conditions of conjecture 2 with @xmath106 ( the three - torus ) and yet possessing infinitely long timelike curves within the black hole as follows @xcite . in constructing the @xmath107 dust - ball spacetimes
, we took the region lying inside of a spherically symmetric timelike three - surface in a @xmath107 robertson - walker spacetime and attached it to schwarzschild as we wanted the model to have @xmath0 cauchy surfaces .
taking the region lying outside of this timelike three - surface and attaching it to schwarzschild , we obtain a globally hyperbolic asymptotically flat spherically symmetric spacetime with @xmath101 cauchy surfaces and infinitely long lived observers within the black hole just like in schwarzschild .
the spatial topology still has the form @xmath105 , although @xmath22 here being @xmath0 is not compact .
however , the spatial homogeneity of the robertson - walker spacetimes allows an alternative construction that gives us black - hole spacetimes with @xmath22 closed . from a @xmath107 robertson - walker spacetime
we construct a new spacetime by simply identifying the the spacetime under a discrete group of isometries so that the spatial topology is now @xmath108 .
now choosing a ( locally spherically symmetric ) two - sphere in this spacetime and generating a timelike three - surface by carrying the sphere along the flow of the dust and then attaching the exterior of this three - surface appropriately to a portion of maximally extended schwarzschild , we again have a globally hyperbolic asymptotically flat spacetime , but now the cauchy surfaces are diffeomorphic to @xmath108 minus a point .
this spacetime admits infinitely long timelike curves in the black - hole region and therefore , independent of what is chosen for @xmath23 ( as long as it finite ) there is no finite @xmath14 for which conjecture 2 would hold for cauchy surfaces diffeomorphic to @xmath109 .
( note that we can perform a similar construction on the @xmath110 robertson - walker spacetimes . )
so , as conjecture 2 does not hold for all cauchy surfaces of the form @xmath105 with @xmath22 closed , are there any restrictions on the allowed topologies of @xmath22 that will give a generalized conjecture 2 a chance to hold ?
a hint is provided by the realization that in the study of the closed - universe recollapse conjecture , a restriction on the cauchy surface topology is necessary as otherwise there may be a topological obstruction to the existence of a maximal cauchy surface .
furthermore , in the case where a maximal cauchy surface exists , there is a finite upper bound to the lengths of all causal curves in the spacetime ( provided certain energy and genericity conditions are met ) @xcite .
does a similar result hold for the asymptotically flat spacetimes ? that is ,
if the spacetime does admit a maximal cauchy surface , then must there be an upper bound to the lengths of timelike curves within the black - hole region ? whether this is the case is unknown , however ,
adapting the ideas of bartnik @xcite , we can bound the lengths of timelike curves in the portion of the black hole that is not reachable from the asymptotic region .
consider the portion of the black hole that is so deep within that it is `` hidden '' in the sense that it can not be reached from @xmath79 , i.e. , @xmath111 this region is globally hyperbolic ( provided the entire spacetime is globally hyperbolic ) and is simply the intersection of the black - hole and white - hole regions associated with an asymptotic end . in schwarzschild , this is all of region iv in fig .
[ schwarz1 ] . in the dust - ball spacetime of fig .
[ db1 ] , this set is empty . in the dust - ball spacetime of fig .
[ db2 ] , this is the union of regions @xmath82 , @xmath83 , @xmath84 , and iv .
note that this is precisely the region of the black hole that is not covered by conjecture 1 .
furthermore , for the spacetimes with cauchy surfaces diffeomorphic to a closed three - manifold minus a point , these regions are very much like closed universes in that for any cauchy surface @xmath24 , the domain of dependence of the compact set @xmath112 contains the hidden black - hole region .
therefore , the evolution of these regions is determined by data on a compact set .
this property provides another connection between the conjectures that observers in closed - universes and black - hole regions must have finite lifetimes and using it we now show that if the spacetime admits a maximal cauchy surface , then the lengths of all causal curves in this hidden portion of the black hole are bounded above . _
theorem 3 .
_ fix an asymptotically flat spacetime satisfying the timelike - convergence condition @xcite and possessing a maximal cauchy surface @xmath113 diffeomorphic to @xmath105 where @xmath22 is a connected , orientable , closed three - manifold .
then , there exists an upper bound on the lengths of all causal curves in the hidden portion of the black hole ( i.e. , the region given by eq .
( [ hidden ] ) ) .
_ denoting the hidden portion of the black hole by @xmath114 , our task is to show that @xmath115 is finite . since @xmath116
, we need only show that @xmath117 and @xmath118 are finite .
[ recall that @xmath119 is the distance function defined as the least upper bound to the length of all continuous causal curves connecting @xmath18 to @xmath19 if @xmath20 , and zero otherwise , and @xmath120 is defined for subsets @xmath121 and @xmath122 of @xmath66 as the least upper bound of @xmath119 over all @xmath123 and @xmath124 @xcite . ]
we bound @xmath118 as follows .
fix any point @xmath125 and let @xmath126 be a longest timelike curve from @xmath113 to @xmath18 . as is well known , @xmath126 intersects @xmath113
orthogonally , is geodetic , and has no points between @xmath113 and @xmath18 conjugate to @xmath113 .
( see , for example , sec .
9.3 of ref .
@xcite . ) however , this last requirement is difficult to satisfy for long curves as a congruence of timelike geodesics meeting @xmath113 orthogonally has vanishing expansion on @xmath113 since @xmath113 has zero trace extrinsic curvature @xmath127 by virtue of its being a maximal hypersurface .
were @xmath127 everywhere bounded away from zero by a negative constant @xmath128 on @xmath113 , or merely on @xmath129 , then the length of @xmath126 ( which equals @xmath130 ) could be no larger than @xmath131 .
( see theorem 9.5.1 of ref .
@xcite . )
however , @xmath127 is merely zero on @xmath113 . to take care of this problem
, we construct a cauchy surface @xmath132 to the future of @xmath113 having negative trace extrinsic curvature bounded away from zero on @xmath133 by `` pushing '' or `` evolving '' @xmath113 to the future .
we construct @xmath132 by first constructing a family of cauchy surfaces @xmath134 by evolving @xmath113 with zero shift ( so the displacement is purely normal ) and non - constant lapse @xmath135 .
the trace of the extrinsic curvature of the surface @xmath134 then evolves according to @xmath136 where @xmath137 is the future - directed unit normal to the surfaces , @xmath138 is the extrinsic curvature of the surfaces , and @xmath139 is the derivative operator associated with the induced metric on each surface .
note that with @xmath140 , the right - hand side of eq .
( [ dkdt ] ) is always non - negative as @xmath141 is manifestly non - negative and @xmath142 is non - negative by the timelike - convergence condition .
were the sum of these quantities positive everywhere on @xmath113 , then by simply evolving @xmath113 with @xmath140 , we would have a cauchy surface with @xmath143 for any @xmath144 .
however , this need not be the case , so we need to be a bit more subtle in our choice of @xmath135 .
let @xmath145 be any compact subset of @xmath113 ( with smooth boundary ) whose interior contains the compact set @xmath146 . then , @xmath147 .
there is a great deal of freedom in our construction of @xmath134 and hence @xmath132 , but for definiteness , we shall consider any family of hypersurfaces @xmath134 that arises from @xmath113 and a lapse that obeys the property that , on @xmath113 , @xmath135 is the unique solution of the elliptic partial differential equation @xmath148 for some scalar field @xmath149 that is positive in the interior of @xmath145 and zero on @xmath150 and that satisfies the boundary condition that @xmath135 be zero on @xmath150 . by hopf s maximum principle @xcite , @xmath135 is positive everywhere in the interior of @xmath145 , so by eq .
( [ dkdt ] ) , @xmath151 on @xmath145 and therefore @xmath152 on @xmath153 for some positive constant @xmath154 .
so , for sufficiently small @xmath155 , on @xmath156 the trace of the extrinsic curvature @xmath127 is bounded away from zero by the negative constant @xmath157 .
we shall take @xmath132 to be any such surface .
so , as @xmath158 is a subset of of @xmath159 and @xmath160 thereon , by the argument above , @xmath161 and therefore @xmath162 . noting that @xmath163 and the fact that @xmath164 is finite ( as @xmath113 and @xmath132 coincide outside of compact sets ) we have established the existence of an upper bound on @xmath118 . by a time - reversed argument , a similar bound can be established on @xmath117 .
therefore , we have established the existence of a finite upper bound on @xmath115 as was to be shown .
@xmath165 so , should the spacetime admit a maximal cauchy surface , then there will not be arbitrarily long timelike curves contained within the hidden portion of the black hole .
while sufficient conditions have been given for the existence of such surfaces @xcite , not all asymptotically flat spacetimes admit maximal cauchy surfaces .
indeed , the cauchy surface topology may provide an obstruction to the existence of such a surface .
this follows from the simple fact that scalar constraint equation of general relativity demands that the ricci scalar curvature associated with the induced metric on such a surface must be non - negative provided the spacetime satisfies the non - negative - energy condition @xcite while most three - manifolds of the form @xmath105 do not admit riemannian metrics with non - negative scalar curvature @xcite . for instance , our @xmath166 black - hole spacetime above does not admit a maximal cauchy surface because of this obstruction .
what three - manifolds @xmath167 ( with @xmath22 as before ) admit asymptotically flat metrics of non - negative scalar curvature , i.e. , those that do not provide a topological obstruct the existence of a maximal cauchy surface ?
witt has shown that any such manifold @xmath22 also admits a riemannian metric with positive scalar curvature @xcite .
it then follows from the work of gromov and lawson that @xmath22 must be @xmath32 , @xmath33 , or a manifold that can be constructed from one of these by making certain identifications and connected summations @xcite .
therefore , we generalize the cauchy surface topology in conjecture 2 to include each of these manifolds less a point . while theorem 3 does support the plausibility of conjecture 2 ( in that some bound should exist )
, it would be nice to strengthen its results beyond the hidden portion of the black hole .
however , with the current method of proof , it does not appear that such a generalization is possible .
the problem is that @xmath168 need not be contained in a compact subset of @xmath113 and therefore the construction of a cauchy surface @xmath132 that is contracting at a rate bounded away from zero on subset that acts as a cauchy surface for the region of the black hole lying to the future of @xmath132 will fail .
the dust - ball spacetimes in figs .
[ db1 ] and [ db2 ] provides simple examples of this behavior .
however , in these spacetimes , for a timelike curve in the black - hole region , the closure of @xmath169 is a compact subset of @xmath113 .
so , here we could show that there are no infinite length causal curve ( although we would not have an upper bound independent of the curve ) .
can this weaker result be shown to hold ?
there is no ( known ) reason why @xmath170 must meet a cauchy surface within a compact subset thereof as happens when @xmath126 `` runs into '' a cauchy horizon ( as seen in kerr - newman ) .
lastly , we note that the lengths of causal curves in the region of the black hole @xmath35 lying to the past of @xmath113 are bounded as @xmath171 is the compact set @xmath172 .
so , a timelike curve in a black - hole region of a spacetime satisfying the timelike - convergence condition and admitting a maximal cauchy surface can not have infinite length to the past .
our strategy in proving theorems 1 and 2 is quite simple .
the idea is to first establish a number of properties of @xmath34 and @xmath4 on the event horizon .
next , these properties are used to establish an upper bound for @xmath34 in the black - hole region ( or just in the part lying to the future of @xmath75 ) .
once this is done , the theorem follows quite easily with the aid of the following lemma established earlier in the analysis of the closed - universe recollapse conjecture ( and whose proof is essentially the subject of sec .
iii of ref .
@xcite ) .
_ lemma 1 .
_ in a globally hyperbolic spherically symmetric spacetime satisfying the non - negative - pressures condition @xcite and on which @xmath4 is everywhere non - negative and @xmath34 is everywhere bounded above by some constant @xmath173 , the lengths of all timelike curves are bounded above by @xmath174 . in sec .
[ sec : elementaries ] , the elementary features of the spherically symmetric spacetime are reviewed . in sec .
[ sec : r_and_m ] , some properties of @xmath34 and @xmath4 associated with asymptotic flatness are established . in sec .
[ sec : bounding_r_b ] , bounds for @xmath34 on the black - hole region are established in a variety of circumstances .
lastly , in sec
. [ sec : complete ] , the proofs of theorem 1 and 2 are completed . in this section , the elementary features of the spherically symmetric spacetimes needed here are briefly reviewed . for a more complete presentation ,
see ref .
@xcite .
a spacetime @xmath175 is said to be spherically symmetric if it admits a group @xmath176 of isometries , acting effectively on @xmath66 , each of whose orbits is either a two - sphere or a point .
denote the orbit of a point @xmath18 by @xmath177 .
the value of the non - negative scalar field @xmath34 at each @xmath178 is defined so that @xmath179 is the area of @xmath177 .
so , in particular , @xmath180 if @xmath181 , while @xmath182 if @xmath177 is a two - sphere . furthermore , we shall say that @xmath183 is a sphere of symmetry if @xmath184 for some @xmath178 and @xmath183 is a two - sphere . where @xmath185 , we decompose the metric @xmath186 into the sum @xmath187 , where @xmath188 is the projection operator onto the tangent space of each sphere of symmetry and @xmath189 is the projection operator onto the tangent space of each two - surface perpendicular to the spheres of symmetry . using the fact that there exists a preferred `` unit - metric '' @xmath190 on each sphere of symmetry , we have @xmath191 ( where @xmath192 and @xmath193 ) .
this gives us the final decomposition of @xmath186 as @xmath194 for the spherically symmetric spacetimes , the mass @xmath4 is defined by @xmath195 defining @xmath196 to be either of the two antisymmetric tensor fields such that @xmath197b}$ ] , and denoting the `` radial part '' of the einstein tensor @xmath198 by @xmath199 ( i.e. , @xmath200 ) we have @xmath201 from which it follows that @xmath202 as the space of vectors perpendicular to the spheres of symmetry is two - dimensional and the metric thereon is lorentzian , we shall label a radial vector as being either outgoing , ingoing , future - directed , or past - directed according to which quadrant the vector lies with the convention that @xmath39 is outgoing near @xmath203 .
radial null vectors will have two such labels ( e.g. , outgoing and future - directed ) while the zero vector will have all four . ( in other words , the set of vectors under a particular label is closed . )
note that these designations are well defined globally for any spherically symmetric spacetime that is simply connected as are the spherically symmetric spacetimes with @xmath0 or @xmath101 cauchy surfaces .
our first lemma establishes a few basic properties of @xmath34 .
_ lemma 2 . _
fix a globally hyperbolic asymptotically flat spherically symmetric spacetime satisfying the null - convergence condition @xcite . then : ( a ) on @xmath204 , @xmath39 can not be future - directed or ingoing ; ( b ) on @xmath205 , @xmath39 can not be past - directed or ingoing ; ( c ) on @xmath79 , @xmath39 must be outgoing spacelike ; ( d ) on @xmath206 , @xmath39 must be either past - directed or outgoing and @xmath34 is non - decreasing to the future . _ proof . _ to prove ( a ) , fix any sphere of symmetry @xmath207 .
if @xmath39 where future - directed or ingoing on @xmath183 , then @xmath183 would be a future outer trapped surface in @xmath204 which is impossible by proposition 9.2.8 of ref .
@xcite . part ( b ) follows from ( a ) by simply reversing the time - orientation of the spacetime .
part ( c ) follows from ( a ) and ( b ) and noting that outgoing spacelike is the only possibility left . that @xmath39 must be either past - directed or outgoing in ( d ) follows from part ( a ) and continuity
from which it follows that @xmath34 is non - decreasing to the future .
( alternatively , part ( d ) follows from hawking s black - hole area theorem . )
@xmath165 our next lemma establishes properties of @xmath4 in the asymptotic region and on the event horizon . _
_ fix a globally hyperbolic asymptotically flat spherically symmetric spacetime satisfying the dominant - energy condition @xcite . then , on @xmath79 , @xmath208 .
further , on the portion of the event horizon lying the future of @xmath75 [ i.e. , @xmath209 , @xmath4 is non - decreasing to the future and has a future limit which is no greater than @xmath2 . _ proof .
_ as @xmath39 is outgoing spacelike on @xmath79 by lemma 2 , @xmath210 is outgoing thereon by eq .
( [ d2 m ] ) .
fix any point @xmath211 , and a spherically symmetric cauchy surface @xmath24 containing @xmath18 .
as @xmath4 is non - decreasing outwards on @xmath212 and since @xmath213 as we approach @xmath40 in @xmath24 , @xmath208 at @xmath18 and hence on all of @xmath214 . by lemma 2
, @xmath39 must be outgoing ( and yet not past - directed ) on @xmath215 , and hence @xmath210 must be outgoing thereon .
therefore , @xmath4 must be non - decreasing to the future thereon . by continuity , @xmath4 is bounded by @xmath2 on @xmath216 and so @xmath4 must have a limit no greater than @xmath2 to the future . @xmath165 it is worth noting that the upper bound of @xmath2 on @xmath4 on the event horizon need not hold everywhere on the event horizon .
a simple example is provided by the dust - ball spacetime in fig .
on the portion of the event horizon lying to the future of some cauchy surface , @xmath4 is bounded as follows . _ lemma 4 . _ fix a globally hyperbolic asymptotically flat spherically symmetric spacetime satisfying the non - negative - pressures condition @xcite . fix a cauchy surface @xmath24 therein and denote by @xmath50 the subset of @xmath24 on which @xmath39 is past - directed timelike , past - directed null , or zero . on the portion of the black - hole s event horizon @xmath37 in the future domain of dependence of @xmath24 ,
i.e. , @xmath217 , @xmath4 is bounded above by the expression @xmath218 _ proof . _ by lemma 2 , @xmath39 is either outgoing or past - directed on @xmath37 . where @xmath39 is outgoing , then @xmath219 , where the first inequality follows directly from the definition of @xmath4 .
where @xmath39 is past - directed , by lemma c1 of appendix c , @xmath220 .
combining these results , eq .
( [ lemma_4 ] ) follows .
lastly , that @xmath4 is everywhere non - negative for the spacetimes under consideration is given by the following lemma .
_ lemma 5 .
_ on a spherically symmetric spacetime satisfying the dominant - energy condition @xcite and admitting a complete spherically symmetric cauchy surface @xmath24 , @xmath4 is everywhere non - negative .
_ we first show that @xmath4 is non - negative on @xmath24 and then show that the non - negativity of @xmath4 everywhere else follows .
suppose that at some point @xmath221 that @xmath222 . then , by eq .
( [ defm ] ) , @xmath39 is necessarily spacelike at @xmath18 .
denote the subset of @xmath24 where @xmath185 by @xmath223 ( which is just @xmath24 less one , two , or possibly no points ) and let @xmath224 the spherically symmetric unit - vector field on @xmath223 such that @xmath225 at @xmath18 .
consider the maximal integral curve @xmath226 of @xmath224 emanating from @xmath18 .
then , as @xmath227 where @xmath39 is spacelike with @xmath225 , @xmath4 is negative and @xmath39 is spacelike all along @xmath226
. therefore , along @xmath226 , the change in @xmath34 with respect to the distance @xmath228 measured along @xmath226 from @xmath18 is bounded by @xmath229 . integrating this and using the fact that @xmath185 on @xmath223 , the length of @xmath226 is bounded above by @xmath230 .
the curve @xmath226 is geodetic ( in @xmath24 ) and , furthermore , it can not be extended to a greater length in @xmath24 as @xmath4 is zero where @xmath34 is zero .
therefore , with @xmath24 complete , @xmath4 must be non - negative everywhere on @xmath24 .
having established that @xmath4 is non - negative on @xmath24 , we now show that @xmath4 is non - negative everywhere as follows .
suppose that @xmath231 at some point @xmath232 .
then , @xmath39 is necessarily spacelike at @xmath18 .
so , consider the past - directed radial null geodesic @xmath21 with future endpoint @xmath18 such that @xmath233 at @xmath18 , where @xmath234 is a past - directed tangent vector to @xmath21 . then , either : ( i ) @xmath39 remains spacelike with @xmath185 on @xmath21 in which case it must intersect @xmath24 at some point @xmath19 .
so , as @xmath235 , we have @xmath236 ; or ( ii ) @xmath237 remains spacelike on @xmath21 but it intersects the `` center of symmetry '' where @xmath238 at some point @xmath19 , so that @xmath239 ; or ( iii ) at some point @xmath19 on @xmath21 , @xmath39 fails to remain spacelike in which case it must be null or zero so that @xmath240 . in all three cases ,
@xmath241 is non - negative , so @xmath4 is non - negative on @xmath242 . that @xmath4 is non - negative on @xmath243 follows by a similar argument .
@xmath165 here we establish bounds on @xmath34 in the black - hole region under a variety of circumstances .
_ lemma 6 .
_ fix a globally hyperbolic asymptotically flat spherically symmetric spacetime satisfying the null - convergence condition @xcite and possessing a black hole @xmath35 with event horizon @xmath36 . fix a spherically symmetric cauchy surface @xmath24 for this spacetime and denote by @xmath244 the component of @xmath24 connected to @xmath40
on which @xmath39 is outgoing .
then , @xmath245 on : ( a ) that portion of the black hole lying in the future of the past null - infinity [ i.e. , on @xmath246 ; ( b ) on the portion of the black hole lying in the future of @xmath244 [ i.e. , on @xmath247 ; ( c ) on all of the black - hole region provided that @xmath39 is outgoing on all of @xmath24 . _ proof . _ to prove ( a ) , fix any point @xmath248 .
let @xmath21 be an outgoing null geodesic generator of @xmath249 .
then @xmath21 has a future endpoint on @xmath177 and no past endpoint .
further , it can be shown that @xmath21 is contained in @xmath250 and enters @xmath214 and hence must intersect @xmath37 . therefore ,
as @xmath39 is not past - directed or ingoing on @xmath250 by lemma 2 , @xmath34 is decreasing to the future on @xmath21 and hence @xmath251 . to prove ( b ) , fix any point @xmath252 .
if @xmath230 is zero , then eq .
( [ lemma_6 ] ) is immediate .
otherwise , consider the ( unique ) radial ingoing and future - directed null geodesic @xmath21 with past endpoint @xmath19 on @xmath244 and future endpoint @xmath18 .
denote by @xmath253 and @xmath254 the parameter and associated tangent vector , respectively , in an affine parameterization of @xmath21 .
then , @xmath255 at @xmath19 as @xmath39 is outgoing at @xmath19 .
further , along @xmath21 , @xmath256 remains negative as can be seen as follows .
consider , @xmath257 the second equality follows from the fact that @xmath21 is geodetic .
the inequality follows from eq .
( [ defm ] ) , the fact that @xmath254 is null , that @xmath258 ( sign depending on one s choice of @xmath259 ) , and the fact that @xmath260 satisfies the null - convergence condition .
[ note that eq .
( [ raychaudhuri ] ) is the raychaudhuri equation for a congruence of radial null geodesics where their expansion @xmath261 is given by @xmath262 @xcite . ] therefore , @xmath34 is non - increasing to the future along @xmath21 .
should @xmath21 intersect the event horizon , then @xmath230 is bounded above by @xmath263 which in turn is bounded above by @xmath264 .
otherwise , @xmath21 has its past endpoint on @xmath265 , so as @xmath39 is outgoing on @xmath244 , @xmath34 on @xmath265 is bounded above by @xmath34 on its outer boundary , which gives @xmath266 . finally , to prove ( c )
, we note that @xmath267 and so by part ( b ) , eq .
( [ lemma_6 ] ) holds on @xmath268 . to establish this bound on @xmath269 , fix any point @xmath18 therein .
the proof proceeds identically to the above except now @xmath21 is a past - directed ingoing radial null geodesic with past endpoint @xmath18 and future endpoint @xmath19 on @xmath24 .
[ note that @xmath270 since @xmath271 and @xmath20 . ]
this time @xmath34 is non - increasing to the past along @xmath21 giving us the bounds @xmath272 .
@xmath165 _ lemma 7 .
_ fix an asymptotically flat spherically symmetric spacetime that possesses @xmath0 cauchy surfaces and that satisfies the dominant - energy and non - negative - pressures condition @xcite .
then for any spherically symmetric cauchy surface @xmath24 therein , on the black - hole region @xmath35 with event horizon @xmath48 , we have @xmath273 where @xmath50 is the subset of @xmath24 on which @xmath39 is past - directed timelike , past - directed null , or zero .
_ we first establish eq .
( [ bound_r_3 ] ) for the portion of the black hole lying on or to the future of @xmath24 . for this region ,
the surface @xmath274 defined as the union of @xmath51 together with @xmath217 , acts as a cauchy surface for @xmath275 as any past - directed causal curve with future endpoint @xmath276 and no past endpoint must intersect @xmath274 . to see this , for any such point @xmath18 , consider any past - directed causal curve @xmath253 with future endpoint @xmath18 and without past endpoint .
the curve @xmath253 must intersect @xmath24 at some point @xmath19 as @xmath232 . if @xmath270 , then @xmath277 . if @xmath278 , then there must a point on @xmath253 between @xmath18 and @xmath19 that lies on @xmath37 and thus @xmath253 intersects @xmath217 .
therefore , appealing to lemma 3 of ref .
@xcite ( or more correctly just repeating the argument used therein as our region is not quite a spacetime in that it has boundaries and further noting that the dominant - energy requirement was superfluous ) , on @xmath35 we have @xmath279 however , we have @xmath280 equation ( [ b2 ] ) follows from lemma c2 in appendix c by taking @xmath281 and noting that the boundary of @xmath22 is a subset of @xmath37 .
equation ( [ b3 ] ) follows from lemma 4 . combining these results , eq .
( [ bound_r_3 ] ) follows .
( note : it is in the establishment of eq .
( [ b2 ] ) that our restriction to @xmath0 cauchy surfaces is used . had @xmath101 cauchy surfaces been allowed , @xmath22 would not have been compact and @xmath34 may be unbounded on @xmath51 as it is in the maximally extended schwarzschild spacetime . )
we now establish eq .
( [ bound_r_3 ] ) for the portion of the black hole lying on or to the past of @xmath24 . for this region ,
the surface @xmath51 acts as a cauchy surface for @xmath282 in the sense that any future - directed causal curve with past endpoint @xmath283 and no future endpoint must intersect @xmath51 .
( any such curve @xmath253 must intersect @xmath24 at some point @xmath19 as @xmath284 .
however , @xmath270 as @xmath271 and hence @xmath253 intersects @xmath51 . )
the proof now follows as before with @xmath285 .
@xmath165 to complete the proofs of theorem 1 and 2 , we simply observe that theorem 1 follows from lemmas 1 , 5 , and 6 , while theorem 2 follows from lemmas 1 , 5 , and 7 .
a massive spherical shell surrounds a spherical vacuum ( schwarzschild ) black hole so that @xmath286 inside of the shell and @xmath287 outside of the shell .
an observer begins his journey on a future marginally outer trapped sphere of symmetry @xmath183 ( which will be on the event horizon if the shell of matter does not fall into the black hole ) . if the shell does not collapse becoming trapped within the black - hole region , then theorem 1 [ part ( c ) ] guarantees that the lifetime of our traveler is bounded above by @xmath288 , which is on the order of the true maximum lifetime @xmath289
. however , suppose that the shell does fall inward creating a black hole of final mass @xmath290 .
in this case , theorem 1 bounds the lifetime of the observer by @xmath291 , which is much larger than the vacuum maximum of @xmath289 .
can the observer truly live this long ? more generally , does the fact that the bound in theorem 1 is a multiple of the future asymptotic size of the black hole indicate that one can increase the lifetime of an observer therein by throwing additional matter into the hole to make it larger ?
further , as it appears that the only natural upper bound for the size of the black hole @xmath27 is the mass of the spacetime @xmath2 ( lemma a1 ) , which corresponds to throwing all the available matter into the black hole ( not that this is always possible ) , might one be able to increase the lifetime of our traveler within the black hole to be on the order of @xmath2 ? that this is not possible in our scenario is a consequence of the following theorem .
_ theorem 4 . _ fix a globally hyperbolic spherically symmetric spacetime satisfying the dominant - energy and non - negative - pressures conditions @xcite and possessing a spherically symmetric cauchy surface that is geodesically complete .
then , for any sphere of symmetry @xmath183 that is future trapped , the lengths of all causal curves in @xmath292 are bounded above by @xmath293 . _ proof . _
this theorem follows from lemma 1 and the fact that @xmath294 on @xmath292 , which we now establish .
fix any point @xmath295 with @xmath182 and fix any maximal length timelike curve @xmath126 from @xmath183 to @xmath18 .
the curve @xmath126 is radial and geodetic with @xmath34 positive thereon .
it follows from the non - negative - pressures condition that @xmath296 on @xmath126 ( where @xmath155 is the proper time along @xmath126 ) , and since @xmath4 is non - negative by lemma 5 , @xmath297 on @xmath126 .
however , as @xmath39 is future - directed timelike , future - directed null , or zero on @xmath183 , @xmath298 must be nonpositive on @xmath299 .
therefore , @xmath300 on all of @xmath126 .
thus , @xmath301 . that @xmath302 on the boundary of @xmath303 follows by continuity of @xmath34 ( or by a slight modification of the above argument for the null case ) .
@xmath165 this theorem bounds the length of the journey of the observer in the scenario above by @xmath304 , which shows that the additional matter can not increase his lifetime to be on the order of @xmath290 .
in fact , it turns out that in this case ( and those like it ) the additional matter never increases the lifetime of an observer within .
( see theorem 6 below . ) to address the more general problem of whether we can increase the lifetime of an observer within a black hole , some care is needed as the problem is subtle for at least three reasons .
first , while it may be possible to increase the amount of time an observer can possibly exist within a black hole , it may just be that none of this increase is available to the observer that is _ already _ within the black - hole region .
that is , while we expect that observers can exist longer in larger black holes ( as is certainly true for vacuum schwarzschild black holes ) , and it is true that by throwing in matter we can increase the amount of time an observer has spent within a black hole ( by making the black hole larger , which is related to our next point ) , it may be that an observer still can not live any longer than he would have otherwise .
second , the very notion of a black hole is a nonlocal concept requiring a knowledge of the entire future history for its determination .
for example , while the spacetime corresponding to a static spherically symmetric shell of matter with a flat interior has no black - hole region , if we choose to `` turn off '' the pressure that keeps the shell from collapsing inwards , the new spacetime will possess a nonempty black - hole region .
so , while under one evolution an observer may begin his journey outside of the black - hole region , in another he may already be within this region .
third , the manipulation of matter is a delicate issue in general relativity .
we can not move matter around as we like , e.g. , pulling two stars together or pushing them apart .
the pushing and pulling involve stresses that must be taken into account as demanded by einstein s theory of gravity . to take care of these subtleties
, we formulate our problem as an initial - value problem as follows .
recall that an initial data set @xmath305 for general relativity consists of a triple @xmath306 @xcite . here
@xmath24 is a three - manifold ( physically the spacetime at an `` instant '' ) , @xmath307 is a riemannian metric on @xmath24 ( physically the metric induced from the spacetime metric @xmath186 ) , and @xmath138 is a symmetric tensor on @xmath24 ( physically the extrinsic curvature of @xmath24 in @xmath66 ) .
on @xmath24 , @xmath307 and @xmath138 are not freely specifiable . instead
, the initial - value constraint equations of general relativity impose a relation between these quantities and the energy density and mass - current density of the matter fields .
( see eqs .
( 10.2.41 ) and ( 10.2.42 ) of ref . @xcite . )
fix an initial data set @xmath305 and a point @xmath18 in the three - manifold @xmath24 associated with @xmath305 .
our knowledge of @xmath305 corresponds to a knowledge of the state of the universe at an `` instant of time '' , while our knowledge of @xmath18 corresponds to our knowledge of the position of our observer at that time . if we had a detailed description of the matter fields ( in particular their initial data satisfying the constraint equations and the equations governing their evolution ) then with einstein s equation we could evolve this initial data to produce a unique spacetime ( up to diffeomorphisms ) .
however , we wish investigate the behavior of gravitational fields without this detailed knowledge of the matter fields , one way around this problem is to restrict ourselves to vacuum spacetimes . instead
, we choose to proceed as follows . to restrict ourselves to spacetimes that are `` physical '' in the sense that their matter content could be considered in some sense `` ordinary '' , we shall restrict ourselves to spacetimes whose einstein tensor satisfies a fixed set of inequalities ( energy conditions ) . for definiteness , here we shall demand that the spacetime satisfy the dominant - energy and non - negative - pressures conditions @xcite .
we shall call a spacetime @xmath175 satisfying these energy conditions a _ possible evolution _ of an initial data set @xmath305 if there is a cauchy surface @xmath308 in @xmath66 such that the induced data on @xmath308 coincides with that of @xmath305 .
note that without the energy conditions , there will be many possible evolutions of any given initial data set .
in fact , we can `` evolve ''
any initial data set @xmath309 to any other initial data set @xmath310 in the sense that there will be a single spacetime that is a possible evolution of both @xmath309 and @xmath310 .
( in the terminology of geroch @xcite , without restrictions , we can _ build _ any initial data set from any other initial data set . )
it is worth noting that when we restrict ourselves to spacetimes satisfying the dominant - energy condition , if the initial data is vacuum in the sense that the associated energy and mass - current densities ( which can be calculated using the initial - value constraint equations of general relativity ) , then a possible evolution of @xmath305 is everywhere vacuum and therefore is unique ( up to diffeomorphisms ) .
more generally , however , there will be many possible evolutions of a given initial data set . from a possible evolution of @xmath305 , we can calculate the least upper bound to the lengths of causal curves having past endpoint @xmath18 . by looking at this least upper bound over the set of all possible evolutions , we can find which possible evolution gives our observer the longest possible lifetime . if @xmath18 is not within the black hole for _ every _ possible evolution of @xmath305 , then there is no finite upper bound on the lifetime of our observer as there will be a possible evolution for which there is a causal curve with past endpoint @xmath18 and future endpoint on @xmath311 ( with such a curve having infinite length ) .
in other words , in this case we can choose our possible evolution so that the observer in fact was never within a black hole .
however , if @xmath18 is in the subset @xmath312 of @xmath24 consisting of all the points of @xmath24 that lie within the black - hole region of each possible evolution , then an observer that begins a journey at @xmath313 is _ guaranteed _ to lie within the black - hole region independent of which possible evolution of @xmath305 we choose .
note that @xmath312 is a closed subset of @xmath24 ( being the intersection of a collection of closed sets ) , and that when there is only a single possible evolution ( e.g. , a vacuum spacetime ) , then @xmath314 .
of course , @xmath312 may be empty . with this definition , a question relevant to our original problem is the following . _
question 1 .
_ given an initial data set @xmath305 , what is the least upper bound on the lengths of causal curves in @xmath315 over the set of all possible evolutions of @xmath305 ?
physically , this upper bound corresponds to the greatest lifetime an observer , beginning his journey at some point in @xmath312 , can hope to achieve by our ( allowed ) manipulations of the matter in the spacetime .
unfortunately , this question appears to be as difficult to answer as proving conjectures 1 and 2 .
even the task of computing @xmath312 from @xmath305 appears difficult as finding all possible evolutions and their associated black - hole regions is nontrivial . are there any sufficient conditions for knowing whether a given point @xmath316 is in @xmath312
? it is well known that future trapped surfaces must lie within the black - hole region of an asymptotically predictable spacetime satisfying the null - convergence condition .
further , the total trapped region @xmath317 associated with an initial data surface ( or initial data set ) shares this property @xcite .
therefore , @xmath318 . in other words , if an observer lies in the total trapped region , it is assured that there is no possible evolution that will allow him to escape to infinity .
so , restricting ourselves to observers that begin their journey from some point in @xmath317 , we are led to the following question . _
question 2 .
_ given an initial data set @xmath305 , what is the least upper bound on the lengths of causal curves in @xmath319 over the set of all possible evolutions of @xmath305 ?
this question also appears quite difficult to answer .
however , restricting ourselves to spherically symmetric initial data sets and spherically symmetric possible evolutions ( spherically symmetric spacetimes ) , a partial answer to this last question is given by the following theorem .
_ theorem 5 . _ fix a spherically symmetric initial data set @xmath305 with three - manifold @xmath103 . denoting the total trapped region in @xmath24 by @xmath317 ,
the lengths of all causal curves in @xmath319 are bounded above by @xmath320 for any possible evolution satisfying the dominant - energy and non - negative - pressures conditions @xcite .
_ this theorem follows from lemma 1 and the fact that @xmath321 on @xmath319 , which we now establish . as in the proof of lemma 7
, we begin by noting that the boundary of @xmath319 acts much like a cauchy surface for @xmath319 . on @xmath322 , denote a null generator of this null boundary by @xmath21 , affinely parameterized by @xmath253 .
then , by the null - convergence condition ( which follows from both the dominant - energy and non - negative - pressures conditions ) , we have @xmath323 .
therefore , as @xmath324 on the apparent horizon @xmath325 , @xmath326 all along @xmath21 to the future of @xmath327 , and so @xmath328 thereon as well .
further , by lemma c2 , @xmath329 , where we have used the fact that @xmath330 . repeating the argument of lemma 3 in ref .
@xcite [ and noting that @xmath39 can not be past - directed timelike on @xmath331 , we have @xmath332 on @xmath319 . @xmath165 therefore , we have a bound on the lifetime of observers beginning their journey in the total trapped region of a cauchy surface in terms of the initial data on that region .
in particular , this bound is insensitive to the final asymptotic size of the black hole ( unlike the bounds given by theorems 1 and 2 ) .
so , can we increase the lifetime of an observer in a black hole by manipulating the matter ?
in general , yes , since not all possible evolutions allow for equally long causal curves . can we increase it to be on the order of @xmath2 ? while this remains unanswered , for certain observers in the spherically symmetric case , theorems 4 and 5
show that this is not always possible . are there any initial data sets where adding matter can not lengthen the lifetime of an observer in the black hole at all ?
we state , without proof , the following theorem giving a class of initial data ( a vacuum schwarzschild black hole of mass @xmath66 surrounded by matter ) showing that the lifetime of observers in the schwarzschild region can not be increased beyond the vacuum maximum @xmath67 .
_ theorem 6 . _ fix a spherically symmetric initial data set @xmath305 such that : ( 1 ) the data on its total trapped region @xmath317 is isometric to the data induced on a cauchy surface in maximally extended schwarzschild spacetime of positive mass @xmath66 with the region corresponding to @xmath317 lying in region ii of fig . [ schwarz1 ] ( i.e. , @xmath39 is future - directed timelike everywhere inside of @xmath317 ) ; ( 2 ) outside of @xmath317 , @xmath39 is outgoing spacelike . restricting ourselves to spacetimes satisfying the dominant - energy and non - negative - pressures conditions @xcite , the lifetime of an observer to the future of some point @xmath333 under a spherically symmetric possible evolution of @xmath305 is no greater than the maximum lifetime under a spherically symmetric possible evolution with @xmath319 being vacuum , which is no greater than @xmath67 .
more weakly , the least upper bound to the lengths of causal curves in @xmath319 over the set of spherically symmetric possible evolutions is @xmath67 .
as the proofs of conjectures 1 and 2 in the spherically symmetric case involve a detailed use of the fields @xmath34 and @xmath4 , concepts for which we do not have an adequate generalization ( at least none that are obviously useful for our purposes ) , it would seem that their proof in the general case will be a difficult task .
however , we do have a few hints as to how one might proceed .
one possible step towards a proof of conjecture 1 is provided by generalizing the lemma bounding the size of the spheres of symmetry @xmath34 in the black hole region of a spherically symmetric spacetime ( lemma 6 ) along the following lines .
consider a `` cut '' @xmath127 of past null infinity @xmath16 and consider the null boundary @xmath334 .
the null generators of this boundary are past complete and have positive expansion thereon ( to the past ) .
therefore , the area of a cross - section of @xmath335 will be decreasing to the future and , therefore , the areas of cross - sections of @xmath335 in the black - hole region will be bounded by the area associated with a cross - section of the event horizon , which is bounded by the future asymptotic area of the black hole s event horizon .
( this is simply a mimicry of the proof used in lemma 6 . )
even with such a result , this is still far from showing that there are no long causal curves in the portion of the black hole region lying to the future of @xmath5 .
perhaps a good place to start in the proof of conjecture 2 would be to attempt a proof of a generalized version of theorem 5 , i.e. , drop the spherically symmetry requirement and replace @xmath336 by some more general quantity @xmath337 constructed on @xmath317 .
although such a theorem would not place an upper bound on the lifetimes of all observers in a black hole region , it would bound those lying to the future of a total trapped region ( which is a subset of the total black hole region ) .
further , such a theorem would be a very nice strengthening of penrose s 1965 theorem that shows that @xmath338 can not be future null complete @xcite .
we end with a few remarks and questions for consideration .
( 1 ) do conjectures 1 and 2 in fact hold with @xmath77 ( the smallest possible value for which they can hold as shown by our analysis of the kerr and dust - ball spacetimes ) ? ( 2 ) can weaker versions of conjectures 1 and 2 be proven ?
for instance , a theorem that merely asserts that the lifetimes of observers in the black - hole region will be bounded , or merely that the lifetimes must be finite ( no global upper bound ) , would be of interest .
( 3 ) if conjecture 1 or 2 is false , does a counterexample to either conjecture exist in the literature , or are the spacetimes providing counterexamples peculiar in some way so that they have been missed ? ( 4 ) can the energy conditions in theorems 1 and 2 be weakened ?
in particular , it would be nice if the non - negative - pressures condition could be replaced by a weaker condition stating that the pressure is not `` too negative '' compared to the energy density .
( 5 ) as has been mentioned , theorems 1 and 4 are slightly awkward in that they require the existence of a geodesically complete spherically symmetric cauchy surface so that @xmath4 will be non - negative everywhere as needed by lemma 1 .
can the requirement that @xmath4 be non - negative in lemma 1 be dropped , thereby allowing for improved versions of theorems 1 and 4 ?
i would like to thank robert geroch for initially directing my attention to the possibility of proving a theorem such as theorems 1 and 2 and for the discussions with both him and robert wald regarding this work .
_ theorem a1 .
_ for a globally hyperbolic asymptotically flat spherically symmetric spacetime satisfying the dominant - energy condition @xcite and with @xmath339 future complete , the area @xmath1 of the event horizon is always bounded above by @xmath340 equivalently , on the event horizon : @xmath43 ; and @xmath341 .
_ the basic idea behind the proof is to construct a `` retarded - time function '' @xmath342 such that @xmath343 as we approach the event horizon of the black hole provided that @xmath344 is complete and then show that this is not this case if @xmath34 is anywhere larger than @xmath345 on the event horizon . essentially , the proof uses the outgoing eddington - finkelstein coordinates ( @xmath34 and @xmath342 , together with two angular coordinates @xcite ) .
however , we choose not to actually `` work in these coordinates '' as it is easier to just treat @xmath34 and @xmath342 as fields and work with their properties .
we begin by constructing a `` retarded - time function '' @xmath342 on @xmath204 as follows . through each point
@xmath18 in the spacetime with @xmath182 , there exists two spherically symmetric null surfaces containing @xmath18 . in the region outside of the black hole ,
i.e. , @xmath204 , we shall `` label '' the outgoing surfaces ( by which is meant those whose tangent vectors can be taken to be outgoing and future - directed ) by defining the scalar field @xmath342 to be constant on these null surfaces and requiring that @xmath346 as we approach @xmath339 .
this choice uniquely defines @xmath342 everywhere on @xmath347 ( up to the addition of an overall constant to @xmath342 ) and has the advantage that the completeness of @xmath339 is then equivalent to the property of @xmath342 assuming all values on @xmath204 .
set @xmath348 .
then @xmath234 is radial , outgoing , future - directed , null , tangent to surfaces of constant @xmath342 , and geodetic ( as is any null vector field arising from a gradient ) .
further , on @xmath347 ( where @xmath185 ) , define @xmath254 to be the radial outgoing past - directed null vector field such that @xmath349 . then , setting @xmath350 , we have @xmath351 equation ( [ lu ] ) follows from the definition of @xmath234 and our normalization requirement for @xmath254 .
equation ( [ lr ] ) follows from the fact that @xmath352 so that @xmath353 .
[ note that @xmath354 is positive for all @xmath355 , as otherwise the sphere of symmetry containing @xmath18 would be an outer trapped surface .
such surfaces can not occur in this region provided the null - convergence condition holds , as it does in our case .
indeed , by the null - convergence condition , @xmath356 , and since @xmath357 is unity at @xmath339 , we have @xmath358 everywhere on @xmath347 . ]
now suppose that for some sphere of symmetry @xmath183 on the event horizon @xmath359 .
setting @xmath360 , then @xmath361 everywhere on @xmath362 .
( see fig .
[ theorema1 ] . ) to see this , fix any point @xmath363 and let @xmath253 denote the spherically symmetric ingoing null surface containing @xmath18 .
similarly , let @xmath226 be the spherically symmetric ingoing null surface containing @xmath183 . since @xmath34 and @xmath342 are continuous , for any number @xmath364 we can find a constant @xmath365 such that @xmath366 for all @xmath367 with @xmath368 . let @xmath21 be the maximal integral curve of @xmath234 containing @xmath228 .
then , setting @xmath369 , it follows from the fact that @xmath39 is outgoing or past - directed ( yet not parallel to @xmath234 ) on @xmath370 that @xmath371 .
further , choosing @xmath228 so that @xmath372 ( i.e. , so that @xmath19 lies to the future of @xmath18 ) we have @xmath373 . putting this all together , we have , @xmath374 for all @xmath364 and so @xmath375 on @xmath362 . ( that @xmath376 on @xmath362 follows from the fact that @xmath39 is outgoing spacelike on the open set @xmath362 . )
further , on @xmath362 , we have @xmath377 to see this , consider how @xmath378 changes along an integral curve of @xmath234 parameterized by @xmath34 , @xmath379 the first equality follows from the definition of @xmath357 , the second from the fact that @xmath380 , and the third follows from eq .
( [ ddr ] ) and the dominant - energy condition .
therefore , dividing eq .
( [ last ] ) by eq .
( [ lr ] ) and noting that @xmath381 is positive on @xmath362 ( as it is a subset of @xmath79 ) , we have @xmath382 using the facts that @xmath383 on @xmath362 , we find @xmath384 integrating this and using the fact that @xmath385 as @xmath386 along the integral curve of @xmath234 [ by eq .
( [ lr ] ) and the fact that @xmath208 on @xmath79 ] , gives eq .
( [ lowerbound_lr ] ) .
we can now establish an upper bound on @xmath342 as follows .
let @xmath253 be a maximal integral curve of @xmath254 in @xmath362 .
parameterizing this curve by @xmath342 , we have @xmath388 setting @xmath389 and integrating eq .
( [ dr*du ] ) along the curve @xmath253 from a point @xmath18 to a point @xmath19 lying to the future of @xmath18 , we have @xmath390 since @xmath391 on @xmath362 , @xmath392 is bounded from below by @xmath393 on @xmath362 , which is finite as @xmath394 . therefore , eq
. ( [ bound_u ] ) gives an upper bound on allowed values of @xmath342 contradicting the completeness of @xmath344 .
so , on the event horizon , @xmath34 can be no greater than @xmath395 .
@xmath165 _ theorem a2 .
_ fix a globally hyperbolic asymptotically flat spherically symmetric spacetime satisfying the dominant - energy condition @xcite with @xmath339 future complete and the event horizon complete in the sense that a future incomplete timelike or null geodesic in @xmath79 can be assigned an endpoint on the event horizon ( in other words , the event horizon is `` all there '' ) and the null geodesic generators of the event horizon are future complete .
then , @xmath42 and @xmath4 have the same limit to the future on the event horizon .
( in other words , the irreducible mass @xmath27 and mass @xmath4 have the same future asymptotic limits to the future on the event horizon . ) _ proof . _
denote by @xmath253 and @xmath234 the parameter and associated tangent vector , respectively , in an affine parameterization of a null geodesic generator @xmath21 or the event horizon .
further , let @xmath254 denote the radial null vector parallel transported along @xmath21 with @xmath396 thereon . denote the future asymptotic value of @xmath4 on the event horizon by @xmath397 .
if @xmath398 on all of @xmath21 , then since @xmath399 on the portion of @xmath21 lying to the future of @xmath75 , @xmath34 must have @xmath400 as a future limit .
so , suppose that at some point on @xmath21 that @xmath401 . since @xmath34 is non - decreasing to the future on @xmath21 , this inequality will continue to hold .
then , repeating the argument that led to eq .
( [ bounding_lr ] ) , this equation again holds with our @xmath254 on the portion of @xmath21 lying to the future of @xmath75 . as @xmath402 on this portion
, we have @xmath403 integrating this we have @xmath404 for some positive constant @xmath22 . noting that @xmath352 on @xmath21 , so that @xmath405 , we have @xmath406 integrating this , we see that @xmath386 as @xmath407
however , this contradicts the fact that @xmath34 is bounded by @xmath345 on the event horizon by theorem a1 .
therefore , the event horizon can not be complete without @xmath34 and @xmath408 having the same future limits
. @xmath165
in some detail , the construction of a dust - ball spacetime proceeds as follows .
the robertson - walker spacetimes are spherically symmetric and can be coordinated ( everywhere excepting where @xmath238 ) so that the metric is given by @xmath409 where @xmath410 is a unit - metric on the two - sphere , @xmath411 and @xmath412 tells us which family we are considering .
( the sign of @xmath412 gives the sign of the intrinsic ricci scalar curvature of the surfaces of homogeneity . )
surfaces of constant @xmath413 are the surfaces of homogeneity and the integral curves of the fluid flow lie in surfaces of constant @xmath414 and are perpendicular to the spheres of symmetry .
note that the allowed range for @xmath414 in expressing the metric in this form is @xmath415 in the @xmath96 and @xmath99 cases while @xmath416 in the @xmath80 case . from eq .
( [ robertson - walker_metric ] ) , the size @xmath34 of a sphere of symmetry is given by @xmath417 with dust as a source , @xmath418 is given by @xmath419 for some constant @xmath22 .
using eq .
( [ r_robertson - walker ] ) and eq .
( [ defm ] ) , the mass @xmath4 associated with each sphere of symmetry is calculated to be @xmath420 notice that @xmath4 is constant along the integral curves of the fluid flow , as it must be as dust has zero pressure ( see eq .
( [ d2 m ] ) ) . to construct a dust - ball spacetime
, we fix a dust - filled robertson - walker spacetime as described above ( i.e. , fix some @xmath412 and @xmath22 ) and choose a number @xmath87 ( within the allowed range ) .
we then take the region in the robertson - walker spacetime with @xmath421 and attach its boundary @xmath317 ( being the timelike three - surface where @xmath422 ) to an appropriate similar three - surface @xmath423 of an appropriate extended schwarzschild spacetime . in order that there not be any `` surface layers '' where the spacetimes are glued together , it is necessary that @xmath4 be a continuous function on the newly created spacetime . therefore , the mass of the appropriate schwarzschild spacetime is simply @xmath4 evaluated on @xmath317 . as the adm mass @xmath2 of the spacetime so constructed is the mass of the schwarzschild portion , we therefore have @xmath424 we shall omit the remainder of the details of this matching except to note that , in the @xmath80 case , when : @xmath425 , @xmath423 lies in regions i , ii , and iii and thereby `` covers up '' all of region iv in fig .
[ schwarz1 ] ; otherwise @xmath423 lies in regions ii , iii , and iv and thereby can not be reached from the asymptotic region ( region i ) without entering the black hole . in figs .
[ db1 ] and [ db2 ] we have sketched two such dust - ball spacetimes . in fig .
[ db1 ] , @xmath426 , while in fig .
[ db2 ] , @xmath81 . in each
, we have divided the spacetime into a number of regions according to the following scheme . on i , @xmath427 , @xmath428 , and @xmath82 , @xmath39 is outgoing and : i lies to the future of @xmath75 and the past of @xmath339 ( i.e , i is the asymptotic region ) ; @xmath427 is in the future of @xmath75 but not in the past of @xmath339 ; @xmath428 lies past of @xmath339 but not in the future of @xmath429 ; @xmath82 lies neither in the future of @xmath75 nor in the past of @xmath339 . on ii and @xmath83
, @xmath39 is future - directed and ii lies to the future of @xmath75 while @xmath83 does not . on iii and @xmath84
, @xmath39 is past - directed and iii lies to the past of @xmath339 while @xmath84 does not .
on iv , @xmath39 is ingoing . for the dust - ball spacetimes , regions i , ii , and
iii always exist while : @xmath427 and @xmath428 exist only for @xmath430 ; @xmath82 exists only for @xmath431 ; @xmath83 , @xmath84 , and iv exist only for @xmath432 .
note that the outer surface of the dust ball crosses the event horizon of the black hole where the surface @xmath422 intersects the surface @xmath433 , i.e. , at @xmath434 .
( see figs .
[ rwdust ] , [ db1 ] , and [ db2 ] . ) therefore , on the event horizon in the dust - filled portion of the spacetime we have @xmath435 .
in particular , in the @xmath80 case , we see that the black hole is eternal , in the sense that every cauchy surface will intersect the black - hole region , if @xmath436 .
_ lemma c1 . _ fix globally hyperbolic spherically symmetric spacetime satisfying the non - negative - pressures condition @xcite .
fix any cauchy surface surface @xmath24 therein and let @xmath50 denote the subset of @xmath24 on which @xmath39 is past - directed timelike , past - directed null , or zero . then , for a point @xmath232 at which @xmath39 is past - directed timelike , past - directed null , or zero , we have @xmath438 _ proof . _
the first inequality is immediate from the definition of @xmath4 while the proof of the second is obtained by repeating the argument given in the latter portion of the proof of lemma 3 in ref .
@xcite and noting that the theorem s requirement that the dominant - energy condition hold can be dropped .
( that is , as the null - convergence condition follows from the non - negative - pressures condition and it can be shown that the existence of a timelike vector @xmath439 for which the last term in eq .
( 2.16 ) in ref .
@xcite is non - negative follows from the null - convergence condition , the requirement that the dominant - energy condition hold was superfluous . )
@xmath165 _ lemma c2 .
_ fix a globally hyperbolic spherically symmetric spacetime and a spherically symmetric cauchy surface @xmath24 therein . for a spherically symmetric compact subset @xmath22 of @xmath24 , define @xmath440 to be those components of @xmath441 for which @xmath39 is spacelike and @xmath442 for vectors @xmath443 on @xmath441 ( tangent to @xmath24 ) and pointing out of @xmath22 .
then , we have @xmath444 _ proof . _
let @xmath18 be any point where @xmath34 reaches its maximum value on @xmath22 . if @xmath39 is non - spacelike at @xmath18 , then @xmath445 , where the first inequality follows from the definition of @xmath4 given by eq .
( [ defm ] ) .
otherwise , if @xmath39 is spacelike at @xmath18 , then @xmath18 must be on @xmath446 for otherwise if @xmath18 is in the interior of @xmath22 or on any other part of the boundary , there exist vectors @xmath443 ( either in @xmath22 or pointing into @xmath22 ) such that @xmath442 thereby violating the maximality of @xmath34 at @xmath18 on @xmath22 .
therefore , in this case @xmath447 . combining these two cases , eq .
( [ bound_r_on_c ] ) follows .
@xmath165 k. s. thorne , _ black holes and time warps : einstein s outrageous legacy _ ( w. w. norton , new york , 1994 ) .
w. israel , in _ three hundred years of gravitation _ , edited by s. w. hawking and w. israel ( cambridge university press , cambridge , england , 1989 ) , chap .
s. w. hawking and g. f. r. ellis , _ the large - scale structure of space - time _
( cambridge university press , cambridge , england , 1973 ) .
r. m. wald , _ general relativity _
( university of chicago press , chicago , 1984 ) .
p. t. chruciel , in _ differential geometry and mathematical physics _ , edited by j. beem and k. l. dugal ( american mathematical society , providence , 1994 ) .
b. carter , in _ black holes _ , edited by c. dewitt and b. s. dewitt ( gordon and breach science publishers , new york , 1973 ) , pp .
r. penrose , rev .
nuovo cimento , * 1 * , 252 ( 1969 ) .
r. penrose , in _ theoretical principles in astrophysics and relativity _ , edited by n. r. lebovitz , w. h. reid , and p. o. vandervoort ( university of chicago press , 1978 ) , chap
r. penrose , in _ general relativity : an einstein centenary survey _ , edited by s.
w. hawking and w. israel ( cambridge university press , cambridge , england , 1979 ) , chap . 12 .
r. p. geroch and g. t. horowitz , in _ general relativity : an einstein centenary survey _ , edited by s. w. hawking and w. israel ( cambridge university press , cambridge , england , 1979 ) .
chap . 5 .
r. penrose , phys .
* 14 * , 57 ( 1965 ) .
e. poisson and w. israel , phys . rev .
d * 41 * , 1796 ( 1990 ) . c. barrabs , w. israel , and e. poisson , class . quantum grav .
* 7 * , l273 ( 1990 ) .
a. bonanno , s. droz , w. israel , and s. m. morsink , phys .
d * 50 * , 7372 ( 1994 ) .
a. bonanno , s. droz , w. israel , and s. m. morsink , university of alberta report no .
alberta - thy-35 - 94 ( gr - qc/9411050 ) .
a null infinity associated with an asymptotically flat spacetime is denoted here by @xmath203 , which is the disjoint union of a past part @xmath16 ( past null - infinity ) and a future part @xmath311 ( future null - infinity ) . spatial infinity
is denoted by @xmath40 .
( see ref .
@xcite for a discussion of these notions . ) j. l. friedman , k. schleich , and d. m. witt , phys .
* 71 * , 1486 ( 1993 ) . a spacetime with ricci tensor @xmath448 and einstein tensor @xmath449
is said to satisfy : ( 1 ) the null - convergence condition ( also known as the null energy condition ) if @xmath450 ( equivalently @xmath451 ) for all null @xmath234 ; ( 2 ) the timelike - convergence condition ( also known as the strong energy condition ) if @xmath452 for all timelike @xmath439 ; ( 3 ) the non - negative - energy condition ( also known as the weak energy condition ) if @xmath453 for all timelike @xmath439 ; ( 4 ) the dominant - energy condition if @xmath454 for all future - directed timelike @xmath439 and @xmath455 ; ( 5 ) the non - negative - pressures condition if @xmath456 for all spacelike @xmath457 . s. a. hayward , phys . rev .
d * 49 * , 831 ( 1994 ) .
g. w. gibbons , commun .
* 27 * , 87 ( 1972 ) .
r. penrose , ann .
ny acad .
sci . * 224 * , 125 ( 1973 ) .
p. s. jang and r. w. wald , j. math .
* 18 * , 41 ( 1977 ) .
p. s. jang , phys .
d * 20 * , 834 ( 1979 ) .
g. w. gibbons , in _ global riemannian geometry _ , edited by t. j. willmore and n. j. hitchin ( ellis horwood , chichester , england , 1984 ) , pp .
g. a. burnett , phys . rev .
d * 43 * , 1143 ( 1991 ) .
g. a. burnett , phys .
d * 48 * , 5688 ( 1993 ) .
g. a. burnett , phys .
d * 50 * , 6158 ( 1994 ) .
g. a. burnett , phys .
d * 51 * , 1621 ( 1995 ) . j. d. barrow , g. j. galloway , and f. j. tipler , mon . not .
r.astron .
soc . * 223 * , 835 ( 1985 ) .
j. d. barrow and f. j. tipler , mon . not .
r. astron .
216 * , 395 ( 1985 ) .
j. d. barrow and f. j. tipler , _ the anthropic cosmological principle _
( oxford university press , oxford , england , 1988 ) . j. e. marsden and f. j. tipler , phys . rep . * 66 * , 109 ( 1980 ) .
j. r. oppenheimer and h. snyder , phys .
56 * , 455 ( 1939 ) . c. w. misner , k. s. thorne , and j. a. wheeler , _ gravitation
_ ( w. h. freeman and company , san francisco , 1973 ) .
d. r. brill , in _ proceedings of the third marcel grossmann meeting on the recent developments of general relativity _ , edited by h. ning ( north holland , new york , 1983 ) .
r. bartnik , commun . math .
117 * , 615 ( 1988 ) .
r. bartnik , in _ proceedings of the fifth marcel grossmann meeting on general relativity _ , edited by d. g. blair and m. j. buckingham ( world scientific , singapore , 1989 ) .
m. h. protter and h. f. weinberger , _ maximum principles in differential equations _
( springer - verlag , new york , 1984 ) .
r. bartnik , commun .
phys . * 94 * , 155 ( 1984 ) .
r. bartnik , p. t. chruciel , and n. murchadha , commun .
math.phys . * 130 * , 95 ( 1990 ) .
d. m. witt , phys .
* 57 * , 1386 ( 1986 ) .
d. m. witt , ( unpublished ) .
m. gromov and h. b. lawson , jr .
hautes etudes sci.publ . math .
* 58 * , 83 ( 1983 ) .
r. p. geroch , j. math .
* 23 * , 611 ( 1982 ) . | we study the problem of how long a journey within a black hole can last . based on our observations ,
we make two conjectures .
first , for observers that have entered a black hole from an asymptotic region , we conjecture that the length of their journey within is bounded by a multiple of the future asymptotic `` size '' of the black hole , provided the spacetime is globally hyperbolic and satisfies the dominant - energy and non - negative - pressures conditions .
second , for spacetimes with @xmath0 cauchy surfaces ( or an appropriate generalization thereof ) and satisfying the dominant energy and non - negative - pressures conditions , we conjecture that the length of a journey anywhere within a black hole is again bounded , although here the bound requires a knowledge of the initial data for the gravitational field on a cauchy surface .
we prove these conjectures in the spherically symmetric case .
we also prove that there is an upper bound on the lifetimes of observers lying `` deep within '' a black hole , provided the spacetime satisfies the timelike - convergence condition and possesses a maximal cauchy surface .
further , we investigate whether one can increase the lifetime of an observer that has entered a black hole , e.g. , by throwing additional matter into the hole .
lastly , in an appendix , we prove that the surface area @xmath1 of the event horizon of a black hole in a spherically symmetric spacetime with adm mass @xmath2 is always bounded by @xmath3 , provided that future null infinity is complete and the spacetime is globally hyperbolic and satisfies the dominant - energy condition . |
recent observations of cosmic ray positron / electron fluxes by the pamela satellite @xcite and atic balloon experiments @xcite have shown excess of events compared with expected astrophysical backgrounds .
ppb - bets @xcite , which is another balloon experiment , also reported cosmic ray electron flux consistent with atic results .
although more recent results of fermi satellite @xcite and hess telescope @xcite on the electron flux did not confirm such a sharp excess observed by atic , the positron excess observed by pamela still remains viable .
while these anomalous positron / electron fluxes may be explained by some nearby high - energy astrophysical sources @xcite , they can also be interpreted as evidence of dark matter annihilation / decay @xcite .
if the recent fermi data is fitted by dark matter annihilation / decay , a heavy dark matter with mass of order a few tev is favored @xcite . in the case of annihilation ,
the required annihilation cross section for reproducing the pamela anomaly is around @xmath0 , which is orders of magnitude larger than the standard value , @xmath1 . in order to fill this gap
, one needs to invoke nonthermal dark matter production scenario @xcite or sommerfeld enhancement mechanism @xcite giving velocity - dependent annihilation rate . in the case of decay
, the required decay rate is about @xmath2 . in order to confirm that the positron / electron excesses are caused by dark matter annihilation / decay ,
cross checks by other observations are necessary .
gamma - rays @xcite , anti - protons @xcite , anti - deuterons @xcite and radio emission associated with the same annihilation / decay processes @xcite may be used for this purpose . at the present stage , however ,
no clear excess was found on those observations , and they give constraints on annihilating / decaying dark matter models .
big - bang nucleosynthesis @xcite and the reionization @xcite also give stringent constraints on annihilating dark matter models .
it was also pointed out that dark matter annihilation / decay which explains the pamela anomaly may predict potentially large neutrino flux as well @xcite , which are constrained from observations of super - kamiokande ( sk ) toward the direction of the galactic center @xcite .
dark matter annihilation ( decay ) yields high - energy neutrinos at the galactic center where the number density is enhanced , and they reach to the earth without energy loss . during the propagation , the neutrino oscillation effects mix three flavors ( @xmath3 ) . among them , some fraction of muon neutrinos are converted into high - energy muons due to the interaction with matter inside the earth , and if the interaction point is within around 1 km from the detector , produced muons penetrate the detector leaving characteristic signatures like the cherenkov light from a muon itself and/or from electromagnetic shower produced by it @xcite . since the direction of a primary neutrino is conserved even after it is converted into a muon , sk limit on the muon flux from the galactic center direction provides useful constraint on dark matter annihilation / decay models @xcite . in this paper we further study the neutrino flux from the galactic center generated by dark matter annihilation / decay and how they are detected at the neutrino detector . since both conversion rate of neutrino into muon and muon energy loss processes in the earth matter
depend on its energy , the signal of the high - energy neutrino has energy dependence . in particular , we distinguish _ showering _ and _ non - showering _ muon events at sk , following the criterion given in ref .
as is expected , higher energy muons are likely to lose energy by radiative losses , developing electromagnetic shower inside the detector .
interestingly , the probability for causing shower events becomes significant around the muon energy @xmath4 , which is a natural scale of dark matter mass .
therefore such a separation , showering / non - showering events , has rich information on the initial neutrino energy spectrum and we show that it is a possible method to discriminate dark matter properties . this analysis can be applied to future neutrino detectors , such as km3net , for searching for dark matter , or even the present on - going experiments , such as sk , may be able to improve constraints on dark matter properties . in sec .
[ sec : muon ] we provide formulae to evaluate the upward showering and non - showering muon fluxes from the primary neutrino flux generated by dark matter annihilation / decay at the galactic center .
applications to some dark matter annihilation / decay models is performed in sec .
[ sec : dmmodel ] .
[ sec : conc ] is devoted to conclusions .
in this section we develop a formalism to evaluate the upward muon flux starting from the initial neutrino energy spectrum produced by dark matter annihilation / decay .
we assume that dm particle , denoted by @xmath5 , with mass @xmath6 annihilates ( decays ) into some final state @xmath7 with annihilation cross section ( decay rate ) @xmath8 ( @xmath9 ) .
the primary neutrino flux from the galactic center is given by @xmath10 for the case of annihilation , and @xmath11 for the case of decay , with @xmath12 where @xmath13 , and @xmath14 denotes the probability that the @xmath15 at the production is observed as @xmath16 at the earth due to the effect of neutrino oscillation and given as @xmath17 , and @xmath18 here , @xmath19 kpc and @xmath20 are the distance of the solar system from the galactic center and the local dark matter density near the solar system , respectively .
@xmath7 collectively denotes the primary annihilation / decay modes ( e.g. , @xmath21 , etc . ) , and @xmath22 represents the @xmath15 spectrum arising from the final state @xmath7 which is simulated by pythia package @xcite .
the dependence on the dark matter halo density profile is contained in the remaining factor @xmath23 , defined by @xmath24 where @xmath25 is the distance from us along the direction @xmath26 , which is the cone - half angle from the galactic center within the range @xmath27 , and @xmath28 is the solid angle over which the neutrino flux is averaged .
dark matter halo profile is characterized as follows : @xmath29^{(\beta-\gamma)/\alpha}},\ ] ] where @xmath30 is the distance from galactic center .
here we consider two typical density profiles : navarro - frenk - white ( nfw ) profile @xcite ( @xmath31kpc ) and cored isothermal profile ( @xmath32kpc ) .
typical values of @xmath33 for these profiles are summarized in table [ table : j ] .
as is expected , dark matter halo with large core @xcite predicts smaller neutrino flux , but the dependence on the halo profile is not so large compared with gamma - ray flux from the galactic center @xcite , as long as the angular resolution is larger than @xmath34 .
thus following discussion does not much depend on the halo profile and it is possible that neutrino signals are detectable while the gamma - ray constraints are satisfied .
.typical values of @xmath35-factor for nfw and isothermal profiles for @xmath36 and @xmath37 .
[ cols="<,<,<,<,<,<",options="header " , ] high - energy neutrinos produced by the dark matter annihilation / decay at the galactic center come in the earth and some of them interact with matter inside the earth , generating high - energy muons . since muons have electric charges , they lose their energy during the propagation inside the earth .
typical propagation distance of high - energy muons with energy @xmath38 tev is around 1 km , as will be shown .
therefore , muons appearing only within 1 km around the detector can reach the detector , leaving their characteristic signals .
the search strategy is to find upward muons in order to avoid cosmic - ray muon background .
taking these processes into account , showering muon flux from the galactic center within cone - half angle @xmath39 is given by @xmath40 where @xmath41 is the parent differential neutrino flux evaluated in the previous subsection , @xmath42 is the parent neutrino energy , @xmath43 denotes the probability that a neutrino with energy @xmath42 is converted into a muon with observable energy range , and @xmath44 is fraction of the showering muon in the total observed muon events with parent neutrino energy @xmath42 .
the last ingredient , @xmath44 , is read off from fig . 5 of ref . @xcite . by taking the ratio between black and black plus red lines in that figure
, we obtain a following fitting formula , @xmath45 where @xmath46 and @xmath47 , and @xmath48 .
( note that this differs from the quantity @xmath49 used in ref .
@xcite . ) the result is shown in fig .
[ fig : showprob ] .
note that this formula is valid for @xmath50 .
if we want to know total muon events , including non - showering events , we have only to replace @xmath44 with 1 : @xmath51 as a function of initial neutrino energy @xmath42 given in ref . @xcite . ] in order to evaluate @xmath43 , the muon - nucleon scattering cross section and the muon energy loss in the rock surrounding the sk must be known .
it is given by @xmath52 where @xmath53 is the number density of proton ( neutron ) in the rock and @xmath54 is the scattering cross section of a neutrino with proton ( neutron ) to create a muon with energy @xmath55 , given by @xcite and @xmath56 .
thus primary neutrino flux in eq .
( [ shower ] ) should only count neutrinos , not anti - neutrinos ( assuming that primary flux of neutrinos and anti - neutrinos are equal ) . ] @xmath57 the mass density of the standard rock is @xmath58 @xcite , and hence @xmath59 where @xmath60 is the avogadro s number .
@xmath61 is the distance which muons with energy @xmath55 can travel until their energies decrease to the threshold energy @xmath62 in the rock .
it is obtained by solving the energy loss equation , @xmath63 where @xmath64 represents the distance in units of @xmath65 .
the first term @xmath66 is from ionization loss , and it only logarithmically depends on @xmath55 .
the second term denotes other radiative energy loss effects due to bremmstrahlung and pair creation , which may cause electromagnetic shower inside the detector .
the precise formulae for those energy losses are given in ref .
@xcite . in fig .
[ fig : r_mu ] we show the propagation distance of muons in the rock . for reference
, we also have shown the result when only the ionization loss is taken into account . for convenience , we provide fitting formula for @xmath67 in the rock , which is valid for @xmath68 , as @xmath69},\ ] ] where @xmath70 and @xmath71 , and @xmath72 .
notice that both the interaction cross section of neutrinos with nucleon and propagation distance of muons are roughly proportional to @xmath73 . therefore heavier dark matter
has a tendency to provide more muon signals .
this situation is quite different from other indirect signals , such as gamma - rays or positrons . for fixed annihilation cross section or decay rate , smaller fluxes of gamma - ray or positron
will be predicted for heavier dark matter , since the number density of dark matter is inversely proportional to the dark matter mass . in the case of neutrino - induced muon signals
, however , higher energy neutrinos more likely produce muon signals , so that it compensates the suppression from smaller number density of dark matter for heavier mass @xcite . before
going into predictions of dark matter models , we evaluate the background muon events coming from atmospheric neutrinos .
since we are interested in the upward muon events , we do not need to worry about background from atmospheric muons .
detailed calculations for the atmospheric neutrino spectrum are performed in refs .
@xcite . using the atmospheric neutrino flux given in ref .
@xcite , the background upward muon events are evaluated as @xmath74 and @xmath75 from the cone half angle @xmath76 ( corresponding to @xmath77 and @xmath78 , respectively ) .
if dark matter prediction is larger than or comparable to those values , dark matter detection by neutrino - induced muon signals is promising .
now let us evaluate the upward showering / non - showering muon flux following the procedure given in the previous section .
we consider dark matter mainly annihilating ( decaying ) into @xmath79 , and @xmath80 .
following results are not changed so much between @xmath81 and @xmath82 , and hence we call these models as `` leptonic annihilation ( decay ) '' .
similarly , annihilating ( decaying ) into @xmath83 and @xmath84 lead to almost the same results , and we call those cases as `` hadronic annihilation ( decay ) '' . taking account of neutrino
oscillation effects , annihilating ( decaying ) into @xmath85 and @xmath86 yield completely the same muon flux , but the case of @xmath87 predicts slightly smaller flux . in the following we call the former case as `` neutrino annihilation ( decay ) '' , since the characteristic behaviors are essentially the same .
resulting total and showering muon fluxes are shown in figs .
[ fig : nu]-[fig : had ] , for the above mentioned three cases : neutrino , leptonic , and hadronic annihilation ( decay ) . in these figures , the vertical axis is @xmath88 for the case of annihilation ( top panels ) , and @xmath89 for the case of decay ( bottom panels ) .
contours of @xmath90 and @xmath91 are shown by solid and dashed lines , respectively .
the shaded region is excluded at @xmath92 c.l . by the observation of sk toward the direction of galactic center with cone half angle
@xmath93 @xcite . if other values of cone angle are chosen , the constrained region becomes widened or narrowed . and
@xmath91 are shown by solid and dashed lines , respectively .
the horizontal axis is dark matter mass and the vertical axis is @xmath88 for the case of annihilation ( top ) , and @xmath89 for the case of decay ( bottom ) .
here dark matter is assumed to annihilate ( decay ) into @xmath85 .
results in the case of @xmath86 are the same . in the case of @xmath87 ,
both fluxes are slightly reduced by an amount of factor 2 .
the shaded region is excluded at @xmath92 c.l . by the observation of sk toward the direction of galactic center with cone half angle @xmath93.,title="fig : " ] and @xmath91 are shown by solid and dashed lines , respectively .
the horizontal axis is dark matter mass and the vertical axis is @xmath88 for the case of annihilation ( top ) , and @xmath89 for the case of decay ( bottom ) .
here dark matter is assumed to annihilate ( decay ) into @xmath85 .
results in the case of @xmath86 are the same . in the case of @xmath87 ,
both fluxes are slightly reduced by an amount of factor 2 .
the shaded region is excluded at @xmath92 c.l . by the observation of sk toward the direction of galactic center with cone half angle @xmath93.,title="fig : " ] , but for annihilating ( decaying ) into @xmath81 . in the case of @xmath94 ,
muon fluxes are slightly smaller by an amount of @xmath95 .
, title="fig : " ] , but for annihilating ( decaying ) into @xmath81 .
in the case of @xmath94 , muon fluxes are slightly smaller by an amount of @xmath95 .
, title="fig : " ] as is easily seen from these figures , in the case of annihilation , the total muon flux remains roughly constant even if dark matter mass is increased for fixed annihilation cross section .
this is because the annihilation rate is proportional to @xmath96 reflecting the fact that the rate is proportional to the square of dark matter number density , but the muon production cross section inside the earth and the muon propagation distance are both roughly proportional to @xmath6 , which cancels the @xmath6 dependence .
however , the muon propagation distance deviates from the simple scaling @xmath97 when the radiative energy loss becomes important at @xmath98 gev ( see fig . [
fig : r_mu ] ) .
thus for heavy dark matter with @xmath99gev , the predicted total muon flux shows mass dependence . on the other hand ,
the showering muon is caused by high - energy component of the primary neutrino flux , and hence its mass dependence is quite different from the total muon flux , although the flux itself is suppressed .
however , the showering muon flux prefers heavier dark matter mass , and atmospheric background flux is significantly reduced for higher energy region . thus we may have a chance to detect muon signals and observe separately non - showering and showering muons .
mass scaling in the case of decaying dark matter can also be recognized in a similar way . since the decay rate per unit volume is only proportional to the dark matter number density ,
the total muon flux is proportional to @xmath6 for fixed decay rate .
this feature can be read off from the figures . , but for annihilating ( decaying ) into @xmath83 .
results in the case of @xmath84 are almost the same.,title="fig : " ] , but for annihilating ( decaying ) into @xmath83 .
results in the case of @xmath84 are almost the same.,title="fig : " ] and @xmath84 . ]
[ fig : ratio ] shows the ratio of the shower / total muon flux for various final states : @xmath100 and @xmath84 .
this ratio is not sensitive to the neutrino flavors : @xmath101 and @xmath86 yield a common ratio .
it is seen that the shower / total muon ratio is quite sensitive to the dark matter mass irrespective to the dark matter annihilation / decay modes . since this ratio expected from the atmospheric neutrino background
is about 0.12 , the dark matter signals predict clearly large shower / total muon ratio . thus searching for muons signals with distinguishing showering and non - showering muons from particular direction may be useful for discriminating the signal from the background and constraining / extracting properties of dark matter .
in this paper , we have studied how dark - matter originated neutrinos can be detected and how can we extract useful information on the dark matter properties , motivated by recent observations of anomalous cosmic - ray positron / electron fluxes . for this purpose ,
we have focused on the separation between showering and non - showering upward muon fluxes inside the neutrino detector . since muons with energy larger than about 500 gev lose their energy dominantly by radiative processes , such as bremmstrahlung and pair creation processes , searching for electromagnetic shower is a useful way to discriminate signals of high - energy neutrinos from atmospheric neutrino background . moreover , since the energy dependence of the total muon flux and showering muon flux are different from each other , measuring the ratio between them gives us information on the dark matter mass .
a good point of muon search compared with gamma - ray and positron / electron flux is that the muon flux is enhanced for heavier dark matter due to the enhancement of the muon production and propagation distance inside the earth .
thus our method has a particular benefit for rather heavy dark matter case , as is favored from the recent observations of the cosmic ray electron flux by the fermi satellite .
the previous bound on the neutrino flux from the galactic center given in ref .
@xcite comes from data of sk - i .
since the superkamiokande accumulates more data now , the bound is expected to be improved soon .
finally we mention possible applications to the on - going or future neutrino detectors .
the icecube detector can not see the upward muon signals from the galactic center , since it is located at the south pole .
a planned extension , called icecube deepcore @xcite , may have a potential to search galactic center by looking at downward muons if atmospheric muons can be removed to the level of expected signals .
although the sensitivity of the deepcore on the muon flux from the galactic center is difficult to estimate at the present stage , other targets such as dwarf galaxies may be used as a source of high - energy neutrinos .
future megaton scale water tank detector , hyper - kamiokande , and kilometer size detector , such as km3net , is expected to significantly improve current sensitivities of sk and may be ideal experiments to search for the dark matter signatures by looking at the galactic center .
detailed estimates on the sensitivities in these experiments are beyond the scope of this paper , and we leave this issue for future work . j.h
. appreciates dr .
kaneyuki and dr .
ibarra for useful discussion .
k.n . would like to thank the japan society for the promotion of science for financial support .
this work is supported by grant - in - aid for scientific research from the ministry of education , science , sports , and culture ( mext ) , japan , no .
20244037 , no .
2054252 ( j.h . ) , and also by world premier international research center initiative , mext , japan .
d. hooper , p. blasi and p. d. serpico , jcap * 0901 * , 025 ( 2009 ) [ arxiv:0810.1527 [ astro - ph ] ] ; h. yuksel , m. d. kistler and t. stanev , arxiv:0810.2784 [ astro - ph ] ; k. ioka , arxiv:0812.4851 [ astro - ph ] ; n. j. shaviv , e. nakar and t. piran , arxiv:0902.0376 [ astro-ph.he ] ; y. fujita , k. kohri , r. yamazaki and k. ioka , arxiv:0903.5298 [ astro-ph.he ] .
l. bergstrom , t. bringmann and j. edsjo , phys .
d * 78 * , 103520 ( 2008 ) [ arxiv:0808.3725 [ astro - ph ] ] ; m. cirelli and a. strumia , arxiv:0808.3867 [ astro - ph ] ; v. barger , w. y. keung , d. marfatia and g. shaughnessy , phys .
b * 672 * , 141 ( 2009 ) [ arxiv:0809.0162 [ hep - ph ] ] ; c. r. chen , f. takahashi and t. t. yanagida , phys . lett .
b * 671 * , 71 ( 2009 ) [ arxiv:0809.0792 [ hep - ph ] ] ; i. cholis , l. goodenough , d. hooper , m. simet and n. weiner , arxiv:0809.1683 [ hep - ph ] ; m. cirelli , m. kadastik , m. raidal and a. strumia , nucl . phys .
b * 813 * , 1 ( 2009 ) [ arxiv:0809.2409 [ hep - ph ] ] .
n. arkani - hamed , d. p. finkbeiner , t. r. slatyer and n. weiner , phys .
d * 79 * , 015014 ( 2009 ) [ arxiv:0810.0713 [ hep - ph ] ] ; m. pospelov and a. ritz , phys .
b * 671 * , 391 ( 2009 ) [ arxiv:0810.1502 [ hep - ph ] ] ; c. r. chen and f. takahashi , jcap * 0902 * , 004 ( 2009 ) [ arxiv:0810.4110 [ hep - ph ] ] ; i. cholis , d. p. finkbeiner , l. goodenough and n. weiner , arxiv:0810.5344 [ astro - ph ] ; y. nomura and j. thaler , arxiv:0810.5397 [ hep - ph ] ; r. harnik and g. d. kribs , arxiv:0810.5557 [ hep - ph ] ; d. feldman , z. liu and p. nath , phys . rev .
d * 79 * , 063509 ( 2009 ) [ arxiv:0810.5762 [ hep - ph ] ] ; p.
f. yin , q. yuan , j. liu , j. zhang , x. j. bi and s. h. zhu , phys .
d * 79 * , 023512 ( 2009 ) [ arxiv:0811.0176 [ hep - ph ] ] ; k. ishiwata , s. matsumoto and t. moroi , arxiv:0811.0250 [ hep - ph ] ; y. bai and z. han , arxiv:0811.0387 [ hep - ph ] ; p. j. fox and e. poppitz , arxiv:0811.0399 [ hep - ph ] ; k. hamaguchi , e. nakamura , s. shirai and t. t. yanagida , arxiv:0811.0737 [ hep - ph ] ; a. ibarra and d. tran , jcap * 0902 * , 021 ( 2009 ) [ arxiv:0811.1555 [ hep - ph ] ] ; i. cholis , g. dobler , d. p. finkbeiner , l. goodenough and n. weiner , arxiv:0811.3641 [ astro - ph ] .
t. moroi and l. randall , nucl .
b * 570 * , 455 ( 2000 ) [ arxiv : hep - ph/9906527 ] ; m. fujii and k. hamaguchi , phys .
b * 525 * , 143 ( 2002 ) [ arxiv : hep - ph/0110072 ] ; phys .
d * 66 * , 083501 ( 2002 ) [ arxiv : hep - ph/0205044 ] ; m. nagai and k. nakayama , phys .
d * 76 * , 123501 ( 2007 ) [ arxiv:0709.3918 [ hep - ph ] ] ; b. s. acharya , p. kumar , k. bobkov , g. kane , j. shao and s. watson , jhep * 0806 * , 064 ( 2008 ) [ arxiv:0804.0863 [ hep - ph ] ] .
s. profumo and p. ullio , jcap * 0407 * , 006 ( 2004 ) [ arxiv : hep - ph/0406018 ] ; m. nagai and k. nakayama , phys . rev .
d * 78 * , 063540 ( 2008 ) [ arxiv:0807.1634 [ hep - ph ] ] ; p. grajek , g. kane , d. j. phalen , a. pierce and s. watson , arxiv:0807.1508 [ hep - ph ] .
j. hisano , s. matsumoto and m. m. nojiri , phys . rev
. lett . * 92 * , 031303 ( 2004 ) [ arxiv : hep - ph/0307216 ] ; j. hisano , s. matsumoto , m. m. nojiri and o. saito , phys .
d * 71 * , 063528 ( 2005 ) [ arxiv : hep - ph/0412403 ] .
g. bertone , m. cirelli , a. strumia and m. taoso , jcap * 0903 * , 009 ( 2009 ) [ arxiv:0811.3744 [ astro - ph ] ] ; e. nardi , f. sannino and a. strumia , jcap * 0901 * , 043 ( 2009 ) [ arxiv:0811.4153 [ hep - ph ] ] ; m. kawasaki , k. kohri and k. nakayama , arxiv:0904.3626 [ astro-ph.co ] ; m. cirelli and p. panci , arxiv:0904.3830 [ astro-ph.co ] .
f. donato , d. maurin , p. brun , t. delahaye and p. salati , phys . rev . lett . * 102 * , 071301 ( 2009 ) [ arxiv:0810.5292 [ astro - ph ] ] ; p. grajek , g. kane , d. phalen , a. pierce and s. watson , arxiv:0812.4555 [ hep - ph ] . d. t. cumberbatch , j. zuntz , h. k. k. eriksen and j. silk , arxiv:0902.0039 [ astro-ph.ga ] ; j. zhang , x. j. bi , j. liu , s. m. liu , p. f. yin , q. yuan and s. h. zhu , arxiv:0812.0522 [ astro - ph ] ; k. ishiwata , s. matsumoto and t. moroi , phys . rev .
d * 79 * , 043527 ( 2009 ) [ arxiv:0811.4492 [ astro - ph ] ] ; l. bergstrom , g. bertone , t. bringmann , j. edsjo and m. taoso , arxiv:0812.3895 [ astro - ph ] .
j. hisano , m. kawasaki , k. kohri and k. nakayama , phys .
d * 79 * , 063514 ( 2009 ) [ arxiv:0810.1892 [ hep - ph ] ] ; j. hisano , m. kawasaki , k. kohri , t. moroi and k. nakayama , phys .
d * 79 * , 083522 ( 2009 ) [ arxiv:0901.3582 [ hep - ph ] ] .
a. v. belikov and d. hooper , arxiv:0904.1210 [ hep - ph ] ; s. galli , f. iocco , g. bertone and a. melchiorri , arxiv:0905.0003 [ astro-ph.co ] . j. hisano , m. kawasaki , k. kohri and k. nakayama , phys . rev .
d * 79 * , 043516 ( 2009 ) [ arxiv:0812.0219 [ hep - ph ] ] . | we study upward muon flux at neutrino detectors such as super - kamiokande resulting from high - energy neutrinos produced by the dark matter annihilation / decay at the galactic center .
in particular , we distinguish showering and non - showering muons as their energy loss processes inside the detector , and show that this information is useful for discriminating dark matter models . |
the flourish of cloud computing technology nowadays is largely due to its outstanding features , on - demand self - service , resource pooling , rapid elasticity , etc.[1 ] all users in cloud environment share a public pool of configurable and virtualized computing resources , such as cpus , disks or network . users can easily scale - up or scale - down their cloud resources according to their real time demands .
for example , before the landing of nasa s curiosity rover , it engineers are allowed to deploy as many servers running on the aws ( amazon web services ) cloud as they need . then , when they are done , they may shut down additional servers to avoid paying for those resources [ 2 ] .
besides , the centralization of servers makes cloud computing technology more environment friendly and energy saving .
compared to setting up their own data center , individuals and enterprises are now becoming more favorable to deploy their businesses on cloud [ 3 ] [ 4 ] .
cloud computing has leveraged users from hardware requirements , while reducing overall client side requirements and complexity [ 5 ] . as the fast growth of cloud computing , security issues
are considered as the obstacles on the highway , which largely hinder the big enterprises wills of porting their business from traditional data center to cloud .
apparently , the security of cloud computing seems to be improved due to the centralization of data and increased security - focused resources [ 6 ] .
the fact , however , is that the security of cloud computing now is considered still in infancy [ 7 ] , especially the network security which faces many new challenges . generally , to protect an enterprise network against cyber - attack , we traditionally adopt network security devices such as firewalls , dmz hosts or intrusion detection systems ( ids ) [ 8 ] .
these traditional network defense strategies , however , can not be applied to cloud computing environment adaptively due to not only the attacks can rise internally but also the dynamic and elastic features of cloud computing [ 9 ] . to settle such problems , new methods are proposed continually in the past years , such as distributed cloud intrusion detection model proposed by irfan gul and m. hussain[10 ] , integrating an ids into cloud computing environment proposed by claudio mazzariello , roberto bifulco and roberto canonico[11 ] or control the inter - communication among virtual machines method proposed by hanqian wu , yi ding[12 ] , etc .
these novel methods , however , merely try to reinforce cloud computing s internal network via porting traditional network defense means .
such methods are not only unsystematic but also impossible to implement when the scale of physical hosts reaches at least half million [ 13 ] . besides
, once the cyber - attack causes some vms overloaded which in turn causes physical hosts which they reside overloaded , all the services on this overloaded physical hosts will be affected or even be corrupted . moreover , due to the logical coupling between vms in a common virtual sub - network , for example , the coupling relationship between load - balancers and servers or between servers and databases , disasters will spread dramatically and then quickly collapse a large part of cloud network . in this paper , we first propose a new two layers model to describe the cloud s complex internal network with the full consideration of the interactions between physical and virtual networks .
based on this model and complex network theory , a novel solution is introduced to systematically and globally settle such a problem that the traditional network defense strategies are no longer suitable for cloud . this solution can make the whole network in cloud computing environment more robust to resist the malicious cyber attacks and to maintain the infrastructures as operatively as possible , even before collapsing .
different from the traditional networks , network in cloud computing environment can be divided into two layers : the virtual layer and the physical layer .
the physical layer contains chunks of physical network facilities , such as switchers , routers , severs or other common network devices .
the virtual layer , however , is built on the physical layer and is implemented via various virtualization technologies , such as container technologies , virtual machine technologies or software define network technologies [ 14 ] .
all these virtual resources , such as vm instances , distributed databases or distributed storage , run on physical hosts and are inter - connected via virtual networks which also run on some physical hosts [ 15 ] .
fig 1 shows the relationship between virtual and physical layers . due to
the sharing of physical resource pool , crash of one vm instance can cause other vm instances on the same physical host to collapse .
furthermore , because of the logical coupling of different components in a sub - virtual network , such collapsing may spread along different paths on both physical layer and virtual layer .
this process is the avalanche effect in cloud computing s cyber attack .
2 shows the avalanche process when only one vm instance in the network is attacked by malicious hacker . according to the complex network theory ,
such avalanche effect can ruin a network rapidly , even the scale of a network is really large [ 16 ] .
in the real cloud environment , vm instances and other virtual components compose the virtual sub - network which represents a full functional application , such as a web application or scientific computing platform .
the whole virtual layer consists of various virtual sub - networks that have different scales . in this article , we use scale - free network to model such a virtual sub - network , due to the fact that many kinds of practical computer networks , including the internet and local area networks , are all scale - free [ 17 ] . here
, we suppose that the distribution of virtual sub - virtual networks with different scale obey the power law distribution , i.e. the larger the virtual sub - network is , the litter it appears [ 18 ] . after this
, we can deploy these virtual components onto physical machines randomly .
the modeling process can be divided into two steps : 1 ) generate various scale - free networks whose scale distribution obey the power law distribution .
each vertex in this network represents a virtual component , such as vm instance on which runs different services .
2 ) create the two - layers model by adding physical vertices into the network and randomly add edges between these physical vertices and the vertices in virtual layer .
to apply complex network methods to the model , we need to simplify the two layers model into single layer . ignoring the specific functions of facilities in the two layers , all virtual or physical components in cloud can be treated as a vertex in network , and inter - connections between vertices
can be abstracted as edges . due to that virtual network
is built on the physical network , edges between vertices on physical layer can be omitted .
so , this two - layer network can be further simplified to a single layer network as we can see from fig .
this simplification reduces the complexity of analysis and makes it possible to use the mature complex network theory and tools .
then , we come to analyze the robustness of this system .
we usually use the size of giant component after initially removing a fraction @xmath0 of nodes to measure the robustness of a network .
first , we consider the situation in which no immune nodes are set up to guarantee the function of the whole network .
bond percolation process can be a great tool to model the dynamic process in the system .
edges are occupied only when the end nodes of the edges are not initially removed and both the end nodes are not infected ( node @xmath1 are infected with probability @xmath2 , we will discuss it later ) .
we define @xmath3 as the probability that node @xmath1 belongs to a small clusters of exactly @xmath4 nodes .
since the network is sparse enough , we can assume that the network topology is locally tree - like .
this means that in the limit of large network size an arbitrarily large neighborhood around any nodes takes the form of a tree , then the calculation using message - passing algorithms can give a good approximation of the clusters . assuming that the networks to be locally tree - like , according to brian karrer and m. e. j. newman s recent theory [ 22 ] , @xmath3 can be write as : @xmath5 \delta \big(s-1 , \sum_{j \in n_{j } } s_{j}\big)\ ] ] where @xmath6 is the kronecker delta which is defined as follows : @xmath7 we can now introduce a probability generating function @xmath8 , whose value is given by [ 22 ] : @xmath9 \delta(s-1,\sum_{j \in n_{i}}s_{j})\\ & = z\prod_{j \in n_{i}}\sum_{s_{j}=0}^{\infty}\pi_{i \gets j}(s_{j})z^{s_j } \end{split}\ ] ] we can simplify the equation as [ 22 ] : @xmath10 where @xmath11 . to calculate @xmath12 , we note that @xmath13 is zero if the edge between @xmath1 and @xmath14 is unoccupied ( with probability @xmath15 ) and nonzero otherwise ( @xmath16 ) , which means that @xmath17 in which : @xmath18 where @xmath19 stands for the fraction of nodes initially removed . and for @xmath20 : @xmath21 \delta(s-1 , \sum_{k \in n_{j\backslash i}}s_k ) \end{split}\ ] ] where the @xmath22 denotes that the set of neighbors of @xmath14 without @xmath1 . substituting this equation into the definition of @xmath23 above
, we then find that : @xmath24 then the expected fraction @xmath25 of the network occupied by the entire percolating cluster is given by the average over all nodes : @xmath26=1-\frac{1}{n}\sum_{i=1}^n\prod_{j\in n_i}h_{i\gets j}(1)\ ] ] setting @xmath27 in equation ( 7 ) we have : @xmath28 we can calculate the size of the remaining greatest connected component of the networks , i.e. the percolating cluster by solving this equation .
some nodes in this network may have the immune ability against malicious attacks due to that they are well protected by some virtual network security equipments which are deployed by professional network administrators .
usually , large corporations have enough money and awareness to employ professional security counselors and managers to protect their it facilities ( physical or virtual ) from cyber - attacks . according to this common sense , in our model
, vertices in a large virtual sub - network will have a high probability to avoid crash when they are attacked by hackers .
the immunity probability @xmath29 of a specific node v can be calculated as : @xmath30 where @xmath25 is the number of vertices in a virtual sub - network which node v belongs to and @xmath31 is the number of vertices in the whole virtual layer .
coefficient @xmath32 stands for that even a virtual sub - network is well protected , it is also possible to be ruined inevitably by some cases . to enhance the robustness of cloud computing network
, we may place some key vertices behind virtual network security components , such virtual firewall , virtual ids etc . [
21 ] in our virtual layer model , we do nt take account of virtual network security components due to that they are transparent to the application users and ca nt be attacked directly .
virtual network security components can be deployed rapidly and conveniently without much more consumptions .
the key vertices which are selected to protect are that have the highest degrees in the network . usually , vertex has high degree somehow means that they are important or even crucial . .
in this figure , vm2 is protected by a virtual firewall . in our model
, we can place such key nodes behind the security components.,width=302 ] to simulate the crashing process , initially , we randomly remove some vertices from the network to simulate that some vms are ruined . then all the vertices which are the neighbors of crashed nodes are affected . due to
that each node in the network has its own immune coefficient which we have mentioned before , the neighboring nodes may survive and avoid crashing during the process .
these new crashed nodes in turn affected their own neighbors .
this process will continue until the system reaches a stable state that no more vertices are affected . in each spreading step , we use the number of nodes in the largest connected cluster to represents the current state of network . based on the model
we have discussed before , we now consider the situation with immune nodes which are totally immune to the infections and will never collapse with some protection . in this paper
, we select the nodes with greatest degrees as the immune nodes . as the introduce of immune nodes into the system
, there will be some changes for @xmath16 .
@xmath33(1-p_{j\in b})\bigg)\ ] ] where @xmath34 stands for the probability that @xmath14 is in the selected group of immune nodes .
if the immune nodes are randomly selected , @xmath35 for all @xmath14 and @xmath19 is the fraction of protected nodes .
it is obvious that @xmath36 , thus the expectation of size of giant component will be greater . in this paper , we selected the nodes with greatest degrees as immune nodes . thus @xmath34 will be a function of its degree and @xmath19 . and
if @xmath14 is in the group with greatest degrees , the effect of protecting this node will be greater .therefore the network will be more robust .
from what we have discussed in the above sections , we can conclude that as long as the immune nodes are added into the network , the probability of existing larger cluster is improved as well . to verify the robustness improvement after applying our novel method to cloud computing s network
, we have simulated this avalanche process with different ratio of initial immune nodes and initial attacked nodes .
we use 5000 physical hosts with 10 vm instances running on each of them to simulate the attack process . here
, we assume that the largest scale of virtual sub - network contains at most 500 vm instances . the results in fig .
5 show that the number of key nodes that have the ability to resist the cyber - attack will finally affect the robustness of the whole network , and the initial number of attacked nodes also affects the network s robustness . as the ratio of initial attacked nodes increase , the number of survived nodes in the largest connected cluster decrease accordingly . also , with different ratio of protected nodes , the robustness ( measured by the number of nodes in the largest connected cluster ) of network varies significantly .
the more the nodes are protected , the higher the robustness is .
5 demonstrates that if we only select 5% ( 2500 vms , 250 physical hosts ) key nodes to give the ability to resist the cyber - attack , the ratio of final survived nodes to the total nodes can increased over 40% or even 70% .
also , if we protect 20% key nodes , this ratio will stably over 60% and in some optimistic cases it will over 90% ( 0.5% nodes initially be attacked ) . in practice , benefited from the elastic and dynamic features of cloud computing , nodes can by rapidly protected by virtual network security devices on demand .
besides , the sdn technology has the ability to detect the real time topology and to re - calculate the degree of all nodes in network rapidly .
so that , when we detect the change of network , no matter physical or virtual , we can re - select the key nodes ( nodes have the highest degree ) and protect those new key nodes by the virtual network components to obtain the immunity in a short period .
in summary , we have introduced a novel method based on complex network theory that can significantly improve the robustness of cloud computing s network to defense malicious attacks with low costs .
our approach shows that with a reasonable protection of some key nodes in the network , significant gains can be achieved for the robustness while the network s functional topology keep unchanged .
this result reveals the fact that instead of deploying security equipment on each rack , protecting the key nodes with virtual network security components is more efficient , economic and energy - saving .
the applications of our results are imminent on one hand to guide the improvement of the existing cloud computing networks but also serve on the other hand to design future cloud infrastructures with improved robustness .
mell , peter , and tim grance .
`` the nist definition of cloud computing . '' national institute of standards and technology 53.6 ( 2009 ) : 50 .
cloud insights .
( 2014 , aug ) .
case study : big data cloud computing helps nasa rover curiosity land on mars . [ online ] .
available : https://www.cloudinsights.com/case-study-nasa-jpl-666995202.html rightscale .
2015 state of the cloud report .
[ online ] .
available : http://assets.rightscale.com/uploads/pdfs/rightscale-2015-state-of-the-cloud-report.pdf amazon .
amazon.com announces first quarter sales up 15% to $ 22.72 billion .
[ online ] .
available : http://phx.corporate-ir.net/phoenix.zhtml?c=97664&p=irol-newsarticle&id=2039598 zissis , dimitrios , and dimitrios lekkas .
`` addressing cloud computing security issues . ''
future generation computer systems 28.3 ( 2012 ) : 583 - 592 .
so , kuyoro .
`` cloud computing security issues and challenges . ''
international journal of computer networks 3.5 ( 2011 ) .
carlin , sean , and kevin curran .
`` cloud computing security . ''
( 2011 ) . sans institute infosec reading room .
building a secure internet data center network .
[ online ] .
available : http://www.sans.org/reading-room/whitepapers/modeling/building-secure-internet-data-center-network-infrastructure-73 shin , seungwon , and guofei gu .
`` cloudwatcher : network security monitoring using openflow in dynamic cloud networks ( or : how to provide security monitoring as a service in clouds ? ) . '' network protocols ( icnp ) , 2012 20th ieee international conference on .
ieee , 2012 .
gul , irfan , and m. hussain .
`` distributed cloud intrusion detection model . ''
international journal of advanced science and technology 34 ( 2011 ) : 71 - 82 .
mazzariello , claudio , roberto bifulco , and roberto canonico .
`` integrating a network ids into an open source cloud computing environment . ''
information assurance and security ( ias ) , 2010 sixth international conference on .
ieee , 2010 .
wu , hanqian , et al .
`` network security for virtual machine in cloud computing . ''
computer sciences and convergence information technology ( iccit ) , 2010 5th international conference on .
ieee , 2010 .
amazon data center size .
13/amazon- data- center- size/. malhotra , lakshay , devyani agarwal , and arunima jaiswal .
`` virtualization in cloud computing . ''
j inform tech softw eng 4.136 ( 2014 ) : 2 .
openstack , openstack installation guide for red hat enterprise linux 7 , centos 7 , and fedora 21 .
[ online ] .
available : http://docs.openstack.org/kilo/install-guide/install/yum/content/ch_overview.html .
barrat , alain , marc barthelemy , and alessandro vespignani .
dynamical processes on complex networks .
cambridge university press , 2008 .
mathematics and the internet : a source of enormous confusion and great potential .
defense technical information center , 2009 .
power law .
[ online ] .
available : https://en.wikipedia.org/wiki/power_law .
kivel , mikko , et al .
`` multilayer networks . ''
journal of complex networks 2.3 ( 2014 ) : 203 - 271 .
openstack , firewall - as - a - service ( fwaas ) overview .
[ online ] .
available : http://docs.openstack.org/admin-guide-cloud/content/install_neutron-fwaas-agent.html openstack , openstack networking ( `` neutron '' ) .
[ online ] .
available : https://wiki.openstack.org/wiki/neutron#why_neutron.3f karrer , brian , m. e. j. newman , and lenka zdeborov .
`` percolation on sparse networks . '' physical review letters 113.20 ( 2014 ) : 208702 . | as a novel technology , cloud computing attracts more and more people including technology enthusiasts and malicious users .
different from the classical network architecture , cloud environment has many its own features which make the traditional defense mechanism invalid . to make the network more robust against a malicious attack , we introduce a new method to mitigate this risk efficiently and systematically . in this paper , we first propose a coupled networks model which adequately considers the interactions between physical layer and virtual layer in a practical cloud computing environment . based on this new model and our systematical method , we show that with the addition of protection of some specific nodes in the network structure , the robustness of cloud computing s network can be significantly improved whereas their functionality remains unchanged .
our results demonstrate that our new method can effectively settle the hard problems which cloud computing now is facing without much cost . |
in 1970s , the discovery of black hole entropy , the four laws of classical black hole mechanics and hawking radiation revealed a mysterious but charming connection between gravitation and thermodynamics . in 1995 , jacobson derived the einstein equations as equilibrium equation of state by assuming proportionality of entropy and horizon area , and the clausius relation .
since then , a large number of papers have sprung up to generalize the approach to other gravity theories @xcite@xcite@xcite .
last year , verlinde @xcite argued the gravity is not only an emergent force , but particularly an entropy force .
his work is motivated by the jacobson s work as well as the gauge / gravity correspondence .
recently , li and his collaborators gave a new scenario of entropy force and holographic thermodynamics @xcite .
first , a surface stress tensor is introduced on a holographic screen while brown - york quasi - local energy @xcite is adopted to replace verlinde energy .
this tensor would give both the surface energy and pressure .
then , a closed screen used by verlinde is extended to an open or closed one .
finally , a time - like screen is used rather than a null screen .
the thermodynamics is self - consistent and back to the jacobson s form while the time - like screen is approaching the null screen . as an important part , chemical potential on holographic screen
is discussed in detail , and specific examples are calculated in li s paper .
for example , a screen sweeping vacuum outside a black hole horizon gives a form of chemical potential according to thermodynamic relations . in this paper
, we propose a different interpretation of chemical potential on holographic screen .
we ponder that such holographic entropy is the entropy bound , independent of concrete matter , since the entropy is given within an ideal system .
there are two reasons in favor of our point .
one is the verlinde temperature is general , having nothing to do with the composition of matter .
it only requires the information of metric .
the other is birkhoff s theorem ensure the irrelevance between exterior metric and interior matter composition . in section iii ,
we discuss two ideal cases , each of which will lead to the entropy bound respectively . for the first case
, we analyze the system with sparse but incompressible liquid like star of uniform density and then give a holographic derivation of bekenstein entropy bound . for the second case , we consider a spherically symmetric system , and compute its holographic entropy on a screen at infinity .
since the temperature we use is not related to specific matter system , the entropy we obtain is not the usual statistical one ( like the black hole entropy ) . at infinity
, we also can not recognize the form of system , so the holographic entropy which we calculated is the entropy bound .
in this paper , we choose the natural units(ie .
in this paper , we need the form of the surface stress tensor and other indispensable quantities , thus we have to review @xcitemiao:2011er , and lay the foundation of our calculation in section iii .
introducing some basic definitions is beneficial .
we set a 2 + 1 dimensional timelike hypersurface @xmath1 embedded in the 3 + 1 dimensional spacetime @xmath2 .
@xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath7 and @xmath8 ( here @xmath9,@xmath10 run from 0 to 3 , and @xmath11,@xmath12 run from 0 to 2 ) denote the coordinates , metric and covariant derivatives of @xmath2 and @xmath1 respectively .
the hypersurface @xmath1 can be described by a scalar function @xmath13 analogous to jacobson s procedure @xcite , we suppose an energy flux @xmath14 passing through an open but time - like patch on holographic screen @xmath15 @xmath16where @xmath17 is the stress tensor in the bulk @xmath2 , @xmath18 is the killing vector , and @xmath19 is the unit vector normal to @xmath1 . obviously , a vector normal to @xmath1 is proportional to @xmath20 , and can be normalized as @xmath21 from the viewpoint of the holographic principle , every physical process in the bulk is corresponding to a process on the holographic screen . when an energy flux @xmath14 passes through the screen @xmath1 from the left side to the right side , the energy will change on the right side , and thus a corresponding change on the screen due to the holographic principle .
the energy density @xmath22 and energy flux @xmath12 on @xmath1 take the form @xmath23where @xmath24 is the energy - momentum tensor on @xmath1 , @xmath25 and @xmath26 are the unit vectors normal to the screen s boundary @xmath27 ( we will give the expressions of @xmath25 and @xmath28 below eq.([mi ] ) ) , and @xmath29 is the killing vector on @xmath1 .
naturally , @xmath24 should only rest on the geometry of the boundary , thus we suppose ( in miao:2011er , a more general and appropriate ansatz has been given , which , however , do not influence our discussion ) @xmath30where @xmath31 is a constant to be determined , and @xmath32 is the extrinsic curvature on @xmath1 defined as @xmath33here @xmath34 is the projection operator satisfying @xmath35 .
it should be stressed that eq.(t ) is the key ansatz in this scheme to derive the einstein equations .
the total energy variation on the screen then becomes @xmath36notice that there are two contributions to @xmath14 , namely the variation of the energy density @xmath37 and the energy flux @xmath12 flowing into the screen .
@xmath38 is the determination of the reduced metric on @xmath39 , with @xmath8 the covariant derivative on @xmath1 and @xmath40 a unit vector in @xmath1 and normal to @xmath39 .
let us choose a suitable function @xmath41 on @xmath1 to denote the boundary @xmath39 , then we have @xmath42when @xmath43 is along the direction of @xmath44(@xmath45 ) it becomes @xmath46 , and when it is along the direction of @xmath47(@xmath48 ) it becomes @xmath49 . from eq.([t ] ) and the gauss - codazzi equation @xmath50 , one can rewrite the eq.([ee2 ] ) as@xmath51 in the following derivations , we will only focus on the quasi - static process for simplicity .
recalling that we have use eq.([fsigma ] ) to denote the holographic screen @xmath1 . in the quasi - static limit
, @xmath52 is independent of time , so @xmath53 .
note that @xmath54 in the quasi - static limit , thus we have @xmath55 .
then the killing vector @xmath29 on @xmath1 can be induced from the killing vector @xmath56 in the bulk @xmath57 , @xmath58thus , in the quasi - static limit , we can rewrite eq.([e1 ] ) as @xmath59from the eq.([e11 ] ) and eq.([e22 ] ) , it follows @xmath60note that @xmath61 , thus we can add a cosmological constant term @xmath62 to the left of the above equation . defining the newton s constant as @xmath63 , we then get the einstein equations , @xmath64substitute @xmath65 into eq.([t ] ) , we obtain , @xmath66which is just the quasi - local stress tensor defined in @xcite . as emphasized in @xcite , @xmath24 characterizes the entire system , including contributions from both the gravitational field and the matter fields .
in addition , generally , @xmath24 need to be subtracted in order to derive the correct energy . to find the relationships between the energy in the bulk and on the holographic screen
, we require a closed screen . using eq.([e1 ] ) and eq.([ee2 ] ) , together with the conservation of energy in the bulk@xmath67we get@xmath68where the fact that a closed screen has no time - like boundary has been used .
this shows the relationship between the bulk energy and the energy on the holographic screen , which also make the logic in section iii complete .
obviously , they differ from each other by a time independent constant at most .
this constant is associated with the subtraction scheme of the energy on the screen , and generally it depends on the position of the screen . due to the constant independent of time , the subtraction scheme does not affect our derivation of the einstein equations .
in this section , we follow the procedure in sec .
vi of @xcite , and delve into the first law of thermodynamics on the holographic screen .
let us consider a 2 + 1 dimensional holographic screen @xmath1 with a constant time slice @xmath57 . for simplicity , we will deal with the case of a static , spherically symmetric spacetime@xmath69where @xmath70 and @xmath38 are functions of @xmath71 only , and @xmath1 is a hypersurface described by @xmath72 mass is distributed around the origin point with spherical symmetry .
we use the quasi - local energy in @xcite to investigate the first law of thermodynamics on screen . as mentioned in sec .
ii , an appropriate subtraction scheme for the stress tensor on @xmath1 is necessary to derive the correct energy . for the asymptotically flat space ,
brown and york propose to deduct the contributions of energy from the flat background spacetime , which gives rise to the same result at infinity as the adm energy for asymptotically flat space .
thus , we concentrate on spherically symmetric asymptotically flat spacetime in this section . more details of the subtraction scheme available from sec .
vi of @xcite , we merely use the results without proof .
we choose newton constant @xmath73 for simplicity . following the subtraction scheme used in @xcite ,
the proper surface energy density on screen becomes @xmath74while the proper surface momentum density @xmath75 vanishes due to spherical symmetry .
and the pressure @xmath76 , trace of the spatial stress , is @xmath77@xmath25 is the unit vector normal to boundary @xmath39 , and @xmath78is the killing vector on @xmath1.@xmath79 is the induced metric on two - boundary @xmath57 . both @xmath80 and @xmath76 are measured by observer at infinity and vanish in flat spacetime .
thus , the quasi - local energy on a time slice @xmath57 is@xmath81introduce the verlinde temperature on the screen as@xmath82 then , we assume that the full form of the first law of thermodynamics on screen is @xmath83where @xmath84 is the entropy of screen , @xmath85 is the chemical potential and @xmath86 is the number of freedom . here we suppose @xmath87 as in verlinde s paper .
we will discuss an ideal model in which mass density @xmath37 is a sufficiently small constant independent of @xmath71 .
it suggests incompressible liquid and infinite sound speed like stars of uniform density .
thus , the holographic entropy we calculate is rather the entropy bound of ordinary matter .
it will give some intuitions about specific physical system .
in gr , with a constant @xmath37 , there are exact solutions of metric @xcite : @xmath88@xmath89where @xmath90 is the radius of matter distribution and @xmath2 is the total mass .
we initially suspend the chemical potential and assume the first law of thermodynamics on the holographic screen @xmath91using brown and york s energy definition , relative quantities are as follows : @xmath92 @xmath93 @xmath94 thus we can get @xmath95when @xmath96 , we have @xmath97 .
actually , without matter , no normal entropy appears .
thus , we introduce a holographic chemical potential , and then the 1st law is back to the form of eq.([law1 ] ) .
then , we immediately have @xmath98this is proportional to @xmath99 and @xmath100 , which denotes the influnce of particles passing through the screen .
finally , we get the entropy bound @xmath101 rather than the entropy of concrete matter . as the second model
, we place a screen sweeping vacuum at infinity in the spherically symmetric space . according to birkhoff s theorem ,
no matter what the composition of matter is and how it distributes ( up to spherical symmetry ) and evolves , the geometry of outer space - time is described by schwarzschild vacuum solution , such as black hole , star system .
this is very similar to the case with constant bulk density .
again , the specific mathematical calculation of entropy is just entropy bound of some special physical system . still without the chemical potential like eq.([no che ] ) , those quantities have different forms:@xmath102 @xmath103 @xmath104 the entropy is approximately@xmath105when @xmath106 , the mass @xmath2 can be treat as a small quantity , then we expand @xmath84 as@xmath107this is the bekenstein bound .
the entropy @xmath108 while the adm mass @xmath109 , therefor the holographic chemical potential vanishes .
the reason why we lay the screen at infinity is attributed to the non - locality of gravity .
it requires an asymptotic flat spacetime to define the total energy . to include the whole physics , especially the gravity , we choose an isolated system which have infinit boundary to satisfy the asymptotic flat condition . only in this way
can we obtain the entropy bound of the system .
to some extent , this is coincident with two simple calculations in @xcite ( page 105 and 112 ) which show that a large @xmath71 is needed to reach the entropy bound for photons .
those examples correspond concrete material system and are included in our general calculation let us analyze the mathematical structure of the first law of holographic thermodynamics . in this paper , we just care about the physics of isolated systems .
consequently we consider the adm mass @xmath2 and constant density @xmath110 are parameters rather than variables . with the only variable @xmath71 ,
the holographic thermodynamic equation is totally differential .
on verlinde and li s projects , gravity is treated as entropy force which is not related to elementary particles , while the chemical potential is related to the variation of the number of particles in thermodynamics . in the first model ,
the screen moving outward in the star leads to the increase of particles , therefore the holographic chemical potential appears in the screen , whereas the chemical potential in the second model does not arise owing to the screen moving in the vacuum .
since the specific matter constitution is not considered , the holographic chemical potential can not be deduced from the first principle in this paper .
we can only judge the origin of the chemical potential from simple discussion .
since the gravity is not fundamental and have no relevant elementary particles , we can relate it to phonon .
the degrees of freedom are equal to the area in our convention .
while the screen expanding in the vacuum , the degrees of freedom also rise without particles , thus it is contributed by the particles and gravity respectively .
the effect of gravity in the holographic screen is the increasing number of modes of simple harmonic oscillation .
a larger screen area implies the possibility of excitation of higher energy levels .
we know the phonon have no chemical potential , therefore the screen moving in vacuum will not lead a chemical potential as the second case .
having calculated these two special examples , we find that entropy bound is a natural result of the first law of thermodynamics on the holographic screen .
however , more research on holographic chemical potential need to be done in future .
we thank miao li and xiao - dong li for valuable discussion and suggestion .
10 c. eling , r. guedens , and t. jacobson , phys .
* 96 , * 121301 ( 2006 ) .
e. elizalde and p. j. silva , phys .
d * 78 , * 061501 ( 2008 ) .
r. brustein and m. hadad , phys .
* 103 , * 101301 ( 2009 ) [ erratum - ibid .
* 105 * ( 2010 ) 239902 ] .
e. p. verlinde , arxiv:1001.0785 [ hep - th ] .
w. gu , m. li , and r. x. miao , arxiv:1011.3419 [ hep - th ] .
j. d. brown and j. w. york , phys .
d * 47 * 1407 ( 1993 ) . | we introduce a new interpretation of chemical potential and show that holographic entropy is entropy bound , which is supported by two ideal cases discussed in detail .
one is sparse but incompressible liquid like a star of uniform density and the other is a screen at infinity in spherically symmetric sapcetime .
our work is based on the new scenario of entropy force and holographic thermodynamics , and the brown - york quasi - local energy . |
while studying the transformation properties of @xmath6 under @xmath7 , richard dedekind , in the 1880 s @xcite , naturally arrived at what we today call the _ dedekind sum _
@xmath8 where @xmath3 and @xmath4 are positive integers , and @xmath9 is the _ sawtooth function _ defined by @xmath10 here @xmath11 denotes the _ fractional part _ of @xmath12 . the dedekind sum and its generalizations have since intrigued mathematicians from various areas such as analytic @xcite and algebraic number theory @xcite , topology @xcite , algebraic @xcite and combinatorial geometry @xcite , and algorithmic complexity @xcite .
almost a century after the appearance of dedekind sums , leonard carlitz introduced the following polynomial generalization , which we will call a _ dedekind carlitz polynomial _ : @xmath13 here @xmath1 and @xmath2 are indeterminates and @xmath3 and @xmath4 are positive integers .
undoubtedly the most important basic property for any dedekind - like sum is _
reciprocity_. for the dedekind
carlitz polynomials , this takes on the following form @xcite .
[ carlitzreciprocity ] if @xmath1 and @xmath2 are indeterminates , and @xmath3 and @xmath4 are relatively prime positive integers , then @xmath14 carlitz s reciprocity theorem generalizes that of dedekind @xcite , which states that for relatively prime positive integers @xmath3 and @xmath4 , @xmath15 dedekind reciprocity follows from theorem [ carlitzreciprocity ] by applying the operators @xmath16 twice and @xmath17 once to carlitz s reciprocity identity and converting the greatest - integer functions into fractional parts .
our motivation for this paper stems from the appearance of dedekind sums in lattice - point enumerators for rational polyhedra .
the first such instance was discovered by mordell @xcite ( the case @xmath18 in the following theorem ) and vastly generalized by pommersheim @xcite , who laid the foundation for the appearance of dedekind - like sums in _ ehrhart polynomials _
@xmath19 ; here @xmath20 is a lattice @xmath21-polytope ( i.e. , @xmath20 has integral vertices ) and @xmath22 denotes a positive integer .
[ thm.l ] let @xmath23 be the convex hull of @xmath24 , @xmath25 , @xmath26 , and @xmath27 , where @xmath28 and @xmath29 are pairwise relatively prime positive integers .
then the ehrhart polynomial of @xmath23 is @xmath30 there are natural geometric interpretations for three of the four coefficients appearing in the ehrhart polynomial of the mordell pommersheim tetrahedron @xmath23 : the leading coefficient is the volume of @xmath23 , the second leading coefficient equals half of the sum of the areas of the faces of @xmath23 , and the constant term is the euler characteristic of @xmath23 .
( in fact , similar interpretations hold for any @xmath21-dimensional lattice polytope @xcite . )
however , aside from toric varieties attached to @xmath23 @xcite , a geometric reason for the appearance of dedekind sums in the linear term of the ehrhart polynomial in theorem [ thm.l ] has so far eluded mathematicians . in this paper
we attempt to shed some light on their appearance .
we begin by studying dedekind
carlitz sums through the generating function attached to the lattice points of a rational polyhedron @xmath31 , namely , the _ integer - point transform _ @xmath32 where @xmath33 .
our goals in this paper are as follows : * we show that dedekind
carlitz polynomials appear naturally in integer - point transforms of rational cones ( section [ carlitzappearssection ] ) .
* we give novel _ geometric _ proofs of theorem [ carlitzreciprocity ] , some of its generalizations , and some new reciprocity theorems ( sections [ carlitzrecsection ] and [ variationssection ] ) .
* we show that our geometric setup immediately implies that ( higher - dimensional ) dedekind carlitz polynomials can be computed in polynomial time ( section [ complexitysection ] ) . *
we realize the equivalence of dedekind carlitz polynomials and the integer - point transform of a two - dimensional analogue of the mordell pommersheim tetrahedron ( section [ trianglesection ] ) . *
we give an intrinsic geometric reason why dedekind sums appear in theorem [ thm.l ] by applying brion s decomposition theorem @xcite to the mordell pommersheim tetrahedron ( section [ mordellsection ] ) .
while we believe that this paper constitutes the first fundamental study of carlitz dedekind sum through a geometric setup , a part of this setup can implicitly be found in papers by solomon @xcite and chapman @xcite on generalized dedekind sums , as well as in @xcite .
we start by decomposing the first quadrant @xmath34 into two cones , namely @xmath35 thus @xmath36 is closed , @xmath37 is half - open , and @xmath34 is the disjoint union of @xmath36 and @xmath37 . let s compute their integer - point transforms . by a simple tiling argument ( see , for example , ( * ? ? ?
* chapter 3 ) ) , @xmath38 where @xmath39 is the _ fundamental parallelogram _ of the cone @xmath36 .
analogously , we can write @xmath40 where @xmath41 note that we need to include different sides of the half - open parallelograms @xmath42 and @xmath43 . .
] now we list the integer points in the half - open parallelepiped @xmath42 .
we may assume that @xmath3 and @xmath4 are relatively prime .
then , since @xmath42 has height 1 , @xmath44 whence @xmath45 we obtain the integer - point transform for @xmath46 in the same way , carefully adjusting our sums for the half - open parallelepiped @xmath43 : @xmath47
the reciprocity theorem for dedekind
carlitz sums now follows almost instantly from our geometric setup .
we have constructed two cones @xmath36 and @xmath37 such that @xmath48 as a disjoint union . in the language of integer - point transforms ,
this means @xmath49 we just computed the rational generating functions on the right - hand side , and the integer - point transform for the first quadrant is simple : @xmath50 .
thus becomes @xmath51 which yields the identity of theorem [ carlitzreciprocity ] after clearing denominators .
our new , geometric proof of carlitz s reciprocity theorem has a natural generalization to higher dimensions , which yields the reciprocity identity for the _ higher - dimensional dedekind
carlitz polynomials _
@xmath52 where @xmath53 are indeterminates and @xmath54 are positive integers .
the reciprocity theorem for these polynomials is due to berndt and dieter @xcite , and we give a novel proof using essentially the same geometric picture as in our previous proof .
[ carlitz reciprocity n - dim ] if @xmath54 are pairwise relatively prime positive integers , then @xmath55 analogous to our proof of theorem [ carlitzreciprocity ] , we construct a single ray in @xmath56-dimensional space , and then we decompose the non - negative orthant into @xmath56 cones as follows .
let @xmath57 , denote the @xmath58 unit vector by @xmath59 , and define @xmath60 the fundamental parallelepiped of @xmath61 is @xmath62 thus a point in @xmath63 will look like @xmath64 and a slice of this parallelepiped at @xmath65 ( where @xmath66 ) will contain @xmath67 as the only integer point .
note also that the integer point @xmath68 is in @xmath63 . combining this information as in section [ carlitzappearssection ]
yields the integer - point transform @xmath69 where @xmath70 and @xmath71 .
since @xmath72 is a disjoint union of the nonnegative @xmath56-dimensional orthant , @xmath73 theorem [ carlitz reciprocity n - dim ] follows upon clearing denominators in this identity . to virtually every theorem in this paper , there exist translate companions , i.e. , we can shift the cones involved in our proofs by a fixed vector .
this gives rise to shifts in the greatest - integer functions , and the resulting carlitz sums are polynomial analogues of _ dedekind rademacher sums _
for the sake of clarity of exposition , we only give integer - vertex versions of our cones , but the reader should keep in mind that arbitrary vertices do , in principal , not cause any additional problems .
dedekind s reciprocity law together with the identity @xmath74 , if @xmath75 , allows us to compute the classical dedekind sum in a euclidean - algorithm style and thus very efficiently , namely in linear time .
( computational complexity is measured in terms of the input length , e.g. , in this case @xmath76 . )
we do not know how to apply a similar reasoning to the dedekind
carlitz polynomials via carlitz s reciprocity law theorem [ carlitzreciprocity ] ; however , the following central theorem of barvinok @xcite allows us to deduce that ( higher - dimensional ) dedekind carlitz polynomials can be computed efficiently . in fixed dimension ,
the integer - point transform @xmath77 of a rational polyhedron @xmath20 can be computed as a sum of rational functions in @xmath78 in time polynomial in the input size of @xmath20 .
note that barvinok s theorem says , in particular , that the rational functions whose sum represents @xmath79 are _ short _ , i.e. , the set of data needed to output this sum of rational functions is of size that is polynomial in the input size of @xmath20 .
the application of barvinok s theorem to any of the cones appearing in the proof of theorem [ carlitz reciprocity n - dim ] immediately yields the following novel complexity result .
[ carcomplexity ] for fixed @xmath56 , the higher - dimensional dedekind
carlitz polynomial + @xmath80 can be computed in time polynomial in the size of @xmath54 .
in particular , this result says that there is a more economical way to write the long " polynomial @xmath80 as a short sum of rational functions .
theorem [ carcomplexity ] implies that any dedekind - like sum that can be efficiently derived from dedekind carlitz polynomials ( e.g. , by applying differential operators ) can also be computed efficiently .
the geometry we have used so far is very simple : it is based on one ray in space .
naturally , this can be extended in numerous ways .
to illustrate one of them , consider two rays through the points @xmath81 and @xmath82 in the first quadrant of the plane , where @xmath83 are positive integers .
this construction decomposes the first quadrant into three rational cones as shown in figure [ tworays ] .
the integer - point transform of each of the two exterior cones , @xmath36 and @xmath84 , is easily computed in the same manner as in section [ carlitzappearssection ] .
however , the cone in the middle , @xmath37 , is bounded on either side by a non - unit vector , and therefore the interior points of the fundamental parallelogram of this cone are not trivial to list .
we will use unimodular transformations to compute @xmath85 .
this construction leads to a new reciprocity identity ( theorem [ carlitz abcd ] below ) .
we start by computing the integer - point transform for a general rational cone in @xmath86 , i.e. , @xmath37 in figure [ tworays ] .
[ general2dimcone ] suppose @xmath83 are positive integers such that @xmath87 and @xmath88 , and let @xmath89 such that @xmath90
. then the cone @xmath91 has the integer - point transform @xmath92 to compute the generating function @xmath93 , note that the linear transformation given by @xmath94 maps @xmath95 to the cone @xmath96 generated by @xmath97 and @xmath98 , whose integer - point transform we know how to compute from section [ carlitzappearssection ] . since @xmath99
we obtain @xmath100 [ carlitz abcd ] suppose @xmath83 are positive integers such that @xmath87 and @xmath88 , and let @xmath89 such that @xmath101 . then @xmath102 again , we prove the theorem geometrically .
let @xmath103 so that @xmath37 is closed and @xmath36 and @xmath84 are half - open , and @xmath104 is a disjoint union of the first quadrant . with the methods introduced in section [ carlitzappearssection ] , the integer - point transforms of @xmath36 and @xmath84
are @xmath105 and the integer - point transform of @xmath37 was computed in lemma [ general2dimcone ] . by our construction
, @xmath106 from which theorem [ carlitz abcd ] follows upon clearing denominators .
theorem [ carlitz abcd ] is the polynomial analogue of the following result due to pommersheim ( * ? ? ?
* theorem 7 ) : [ pomm3term ] suppose @xmath83 are positive integers such that @xmath88 , and let @xmath89 such that @xmath101 .
then @xmath107 this , in turn , generalizes rademacher s three - term reciprocity theorem @xcite : [ radcor ] if @xmath108 are pairwise coprime positive integers then @xmath109 where @xmath110 , @xmath111 , and @xmath112 .
the case @xmath113 of berndt
dieter s theorem [ carlitz reciprocity n - dim ] is the polynomial analogue of this three - term law .
clearly , dedekind s reciprocity theorem is implied by corollary [ radcor ] , and girstmair showed @xcite that implies theorem [ pomm3term ] , so that , in fact , dedekind s , rademacher s , and pommersheim s reciprocity theorems are equivalent .
such an equivalence can clearly not hold on the polynomial level , but one could ask whether theorems [ carlitzreciprocity ] and [ carlitz abcd ] are equivalent . theorem [ carlitz abcd ] could be generalized in several ways , e.g. , to higher dimensions or to more than three cones in dimension 2 ( this yields a carlitz polynomial analogue of ( * ? ? ? * theorem 8) ) , but we digress .
theorem [ carlitz abcd ] simplifies when @xmath114 : then the third dedekind
carlitz polynomial disappears .
geometrically , this stems from the fact that the cone @xmath37 is _ unimodular _ , i.e. , its fundamental parallelogram contains only the origin . if @xmath83 are positive integers such that @xmath114 , then @xmath115 this is the polynomial analogue of the following result due to rademacher @xcite .
[ rademacher abcd ] if @xmath83 are positive integers such that @xmath116 , then @xmath117
we have shown that dedekind
carlitz polynomials appear as natural ingredients of integer - point transforms of cones . in this light , it should come as no surprise that the following conic decomposition theorem of brion @xcite will prove useful .
we remind the reader that a _
rational polytope _ is the convex hull of finitely many rational points in @xmath118 . if @xmath119 is a vertex of the polytope @xmath20 , then the _ vertex cone _
@xmath120 is the smallest cone with apex @xmath119 that contains @xmath20 .
[ brionthm ] suppose @xmath20 is a rational convex polytope
. then we have the following identity of rational functions : @xmath121 where the sum is over all vertices of @xmath122 .
brion s theorem allows us to give a novel expression for the dedekind
carlitz polynomial as the integer - point transform of a certain triangle .
[ thm.brion2d ] let @xmath3 and @xmath4 be relatively prime positive integers and @xmath1 and @xmath2 be indeterminates . if @xmath123 is the triangle with vertices @xmath124 , @xmath125 , and @xmath126 , then the dedekind
carlitz polynomial @xmath127 and the integer - point transform of @xmath123 are related in the following manner : @xmath128 the triangle @xmath123 comes with the vertex cones @xmath129 once more we apply the methods of section [ carlitzappearssection ] and obtain the integer - point transforms @xmath130 now brion s theorem [ brionthm ] gives @xmath131 or , after some manipulations , @xmath132 the two dedekind carlitz polynomials in this expression are actually related by reciprocity : after a simple change of variables , carlitz s theorem [ carlitzreciprocity ] implies @xmath133 and substituting this into finishes the proof .
we remark that we could also deal with triangles with rational vertices .
the setup is exactly the same , however , the formulas become quite a bit messier .
in this final section we derive a generating - function equivalent to theorem [ thm.l ] .
it is natural to extend the application of brion s theorem in the last section to polytopes in higher dimensions , which we will do with the tetrahedron @xmath23 .
since our ultimate goal is a geometric proof of theorem [ thm.l ] , we will apply brion s theorem to the dilate @xmath134 for any positive integral @xmath22 . as we will see
, the following higher - dimensional extensions of dedekind carlitz polynomials will appear naturally in the integer - point transforms of the vertex cones of @xmath23 . for positive integers @xmath135 , and indeterminates @xmath136
, we define the _ dedekind rademacher carlitz ( drc ) sum _ @xmath137 the drc sums are the main ingredients for the integer - point transform of @xmath23 .
the computation of the integer - point transforms of the vertex cones should be routine by now ; we will derive only one of them in detail .
the vertex cone @xmath36 has the fundamental parallelepiped @xmath142 there exists exactly one integer point in @xmath143 at each integer @xmath144 along the @xmath145-axis and @xmath146 along the @xmath147-axis . in other words , for @xmath148 , @xmath149 is written as @xmath150 and hence @xmath151 and @xmath152 . thus @xmath153 and @xmath154 } { \left ( u-1 \right ) \left ( u^a - v^b \right ) \left ( u^a - w^c \right ) } \ , .
\end{aligned}\ ] ] similarly , we derive @xmath155 } { \left ( v-1 \right ) \left ( u^a - v^b \right ) \left ( v^b - w^c \right ) } \ , , \\ \sigma_{{{\mathcal k}}_3 } ( u , v , w ) & = \frac { w^{(t+2)c } \left [ ( w-1 ) + { { \rm drc}}\left ( w^{-1 } , u , v ; c , a , b \right ) \right ] } { \left ( w-1 \right ) \left ( u^a - w^c \right ) \left ( v^b - w^c \right ) } \ , , \end{aligned}\ ] ] and @xmath156 substituting these rational functions into and clearing denominators gives the theorem .
theorem [ thm.brion3d ] gives @xmath157 with numerator @xmath158 and denominator @xmath159 .
we obtain the ehrhart polynomial @xmath160 by taking the limit of this rational function as @xmath161 .
we need to use lhospital s rule and take partial derivatives with respect to @xmath1 once , @xmath2 twice , and @xmath162 three times in both numerator and denominator before substituting @xmath163 .
this is easily done for the denominator @xmath164 with the result @xmath165 .
the numerator @xmath166 of @xmath167 is not handled as easily . after differentiating and setting @xmath163
it becomes @xmath168 plus numerous terms that do not depend on @xmath22 .
fortunately , because the constant term of @xmath169 is @xmath170 ( see , for example , ( * ? ? ?
* corollary 3.15 ) ) , we are not concerned with these extra terms . as before , we replace all greatest - integer functions with fractional - part functions to modify to @xmath171 now we use three elementary identities : @xmath172 they allow us to simplify to @xmath173 we divide by the denominator @xmath165 and add the constant term 1 to arrive at the desired formula for @xmath174 .
matthias beck , _ geometric proofs of polynomial reciprocity laws of carlitz , berndt , and dieter _ , diophantine analysis and related fields 2006 , sem .
35 , keio univ . ,
yokohama , 2006 , pp . | we study higher - dimensional analogs of the _ dedekind carlitz polynomials _
@xmath0 where @xmath1 and @xmath2 are indeterminates and @xmath3 and @xmath4 are positive integers .
carlitz proved that these polynomials satisfy the _ reciprocity law _
@xmath5 from which one easily deduces many classical reciprocity theorems for the dedekind sum and its generalizations .
we illustrate that dedekind
carlitz polynomials appear naturally in generating functions of rational cones and use this fact to give geometric proofs of the carlitz reciprocity law and various extensions of it .
our approach gives rise to new reciprocity theorems and computational complexity results for dedekind
carlitz polynomials , a characterization of dedekind carlitz polynomials in terms of generating functions of lattice points in triangles , and a multivariate generalization of the mordell pommersheim theorem on the appearance of dedekind sums in ehrhart polynomials of 3-dimensional lattice polytopes . |
quantum mechanics is one in a large class of theories which are consistent with relativity in the sense that they do not allow signals to be sent faster than the speed of light .
many of these theories exhibit strong non - local correlations between distant particles that can not be explained by the properties of the individual particles alone .
surprisingly , quantum mechanics is not the most non - local among them , which raises the question about the physical principle that singles out quantum mechanics and sets the limit on the possible strength of correlations in nature . here
we experimentally address this fundamental question by testing the principle of information causality in the classical , quantum and post - quantum regime .
while the no - signaling principle limits the speed with which distant parties can communicate , information causality states that the accessible information can not be more than the information content of a communicated message , no matter what other shared resources are used .
both classical and quantum mechanics satisfy this principle , while it is violated by most post - quantum theories @xcite .
we experimentally emulate correlations of various strengths from classical to almost maximally non - local and demonstrate a violation of the principle of information causality in the case where the simulated correlations are beyond the quantum regime .
apparent super - quantum correlations are , in our approach , a consequence of the non - unitary evolution of quantum states when subjected to polarization - dependent loss with post - selection @xcite . for moderate loss , we find that initially entangled states can result in super - quantum correlations , while unentangled states still appear classical . for higher loss on the other hand
we observe super - quantum correlations even for classical input states .
no - signaling resources can formally be treated as pairs of black boxes shared between arbitrarily separated alice and bob @xcite , see fig .
[ fig : icmotivation]a ) .
each box has a single input and output and the correlation between them is only restricted by the no - signaling principle .
this means that the local outcome only depends on the local input , such that alice can not learn anything about bob s input from only her output .
a typical quantum example of such a resource is a pair of entangled particles , shared between alice and bob , where inputs correspond to measurement settings and outputs to measurement outcomes . since the work of john bell and numerous subsequent confirming experiments it is now widely accepted that these particles exhibit non - local correlations , which have no classical explanation . under the no - signaling constraint alone , however , there are even stronger non - local correlations than quantum entanglement @xcite .
the maximum that is compatible with relativity is achieved by the so - called popescu - rohrlich ( pr)-box @xcite , characterized by perfect correlations of the form @xmath0 , between alice s and bob s inputs @xmath1 and @xmath2 and outputs @xmath3 and @xmath4 , respectively .
here @xmath5 denotes addition modulo @xmath6 , equivalent to the logical xor , where @xmath7 when @xmath8 and @xmath9 otherwise .
a convenient operational way of quantifying non - locality is the clauser - horne - shimony - holt ( chsh ) inequality @xcite .
this experimentally testable reformulation of bell s inequality is satisfied by any correlation that can be described by a local hidden variable model .
such models are a description of correlations that can arise in classical systems , but can not describe non - local correlations obtained from e.g. entangled quantum states . written in terms of correlations of the form @xmath10
the inequality takes the form @xmath11 here @xmath12 denotes the probability for obtaining outputs @xmath13 , which satisfy @xmath14 given the inputs @xmath1 for alice and @xmath2 for bob . while this inequality is satisfied by any classical correlations , it can be violated in the quantum case .
this violation , however , is bounded to a value of @xmath15 , known as tsirelson s bound @xcite .
note , that inequality is presented here in a slightly different form than conventionally @xcite , where the classical bound is @xmath6 and tsirelson s bound is @xmath16 .
they are , however , linearly related and the difference is a simple rescaling of @xmath17 . despite being a simple consequence of the mathematical formalism of quantum mechanics ,
it is unclear what the physical motivation is for this seemingly sub - optimal limit on the strength of quantum correlations .
in fact even the algebraic maximum @xmath18 can be achieved ( by the pr - box ) without violating the no - signaling principle .
this principle is physically motivated by the fact that , according to special relativity , faster - than - light information transfer would allow information to be sent backwards in time and thus violate causality . nevertheless
, it does not explain why super - quantum correlations such as the pr - box are incompatible with quantum mechanics and seem not to exist in nature . a possible explanation is offered by the principle of information causality a generalization of no - signaling which states that there can not be more information available than was transmitted @xcite .
this can be understood on the basis of the following elementary information - theoretic protocol : bob tries to gain information from a set of data that is only known to alice .
the parties are allowed to use an arbitrary amount of shared no - signaling resources , but may not communicate more than @xmath19 classical bits . in this case , the information causality principle states that the amount of information accessible to bob should be limited to @xmath19 classical bits @xcite . and
@xmath4 for alice and bob , when they input @xmath1 and @xmath2 , respectively . in the case of a pr - box
the outputs of the left ( l ) and right ( r ) box would be perfectly correlated according to @xmath0 . the inputs and outputs depicted here
correspond to the simplest instance of the information causality protocol . *
b ) * example of the multilevel information causality protocol for @xmath20 .
alice has a list of @xmath21 bits @xmath22 and bob tries to guess the bit @xmath23 ( shown in bold , red ) using @xmath24 pairs of shared black boxes on @xmath25 levels ( corresponding boxes labeled @xmath26 , @xmath27 ) .
bob s inputs @xmath28 and choice of boxes are determined by the binary decomposition @xmath29 . from his outputs @xmath30 and alice s @xmath9-bit message @xmath31 bob computes a final guess @xmath32 for alice s bit @xmath33 .
note that bob only needs to use one box on each level and ignores the outputs of all the other boxes .
hence , his input to these boxes can be arbitrary and in the experiment we chose to use the same input for all boxes on one level . ] in the simplest instance , alice has a set of two bits @xmath34 and bob wants to guess one of them , which we denote @xmath33 @xcite , see fig .
[ fig : icmotivation]a ) . alice and bob then input @xmath35 and @xmath2 into their respective black box and obtain outputs @xmath3 and @xmath4 . from this output alice computes an @xmath36-bit message @xmath37 and sends it to bob , who calculates his guess for alice s bit as @xmath38 . in the case of a shared pr - box , bob can guess either one of alice s bits perfectly , since in that case @xmath0 and thus @xmath39 . in the more general case considered here , alice has a dataset @xmath40 of @xmath41 bits and bob wants to guess the bit with index @xmath29 . as discussed in ref .
@xcite , alice and bob can achieve this task by using a nested version of the protocol outlined above , with @xmath42 black boxes on @xmath43 levels and @xmath9 bit of classical communication .
the protocol is illustrated in fig .
[ fig : icmotivation]b ) for the case @xmath20 . from every output alice computes a temporary message @xmath44 , where @xmath45 denotes the level and @xmath46 the number of the box on that level . since she is only allowed @xmath9 bit of communication , she uses these temporary messages as the inputs for the boxes on the next - lower level and only sends the final message to bob . depending on @xmath47 bob then decodes either @xmath48 or @xmath49 and then moves on to the next - higher level until he reaches the bit of interest .
bob s success can then be quantified by @xmath50 where @xmath51 is the shannon mutual information between the @xmath45th bit of alice s list and bob s guess for it @xcite .
this quantity can further be bounded as @xmath52 where @xmath53 is the binary entropy of the success probability @xmath54 for guessing the @xmath45th bit .
experimentally , we generate apparent super - quantum correlations based on the effect of polarization - dependent loss in a post - selected bell - test experiment @xcite , see fig .
[ fig : setup]a ) .
we use photon pairs created from a continuous - wave pumped spontaneous parametric down - conversion source in a polarization sagnac design @xcite , as illustrated in fig .
[ fig : setup]b ) . using this approach
we obtain photon pairs with very high efficiency and in a continuously tunable fashion that enables us to produce any bipartite quantum state @xcite .
in particular , we used the maximally entangled state @xmath55 as the initial state , where @xmath56 represent horizontal and vertical polarization , respectively . for comparison
, we also considered the corresponding fully decohered and thus separable state @xmath57 .
this state was produced as a mixture of the two pure state components @xmath58 and @xmath59 by probabilistically mixing the respective coincidence counts .
mm long periodically poled ktiopo@xmath60 ( ppktp ) crystal inside a polarization sagnac interferometer using a grating stabilized continuous wave pump laser ( l ) at a wavelength of @xmath61 nm . by controlling the phase and polarization of this laser and adjusting the additional half - wave - plate in bob s arm of the source , hwp3 , any two - qubit states can be produced .
* c ) * polarization - dependent loss is introduced to the system in a controllable way using an interferometer based on calcite beam displacers ( bd ) , which split the horizontal and vertical polarization components into two spatial modes .
the two hwps in the interferometer are set to rotate the polarization by @xmath62 , which ensures equal path - length of the two spatial modes upon recombination at the second set of bd .
the degree of loss for each polarization is then proportional to the offset of the corresponding hwp from this setting .
finally , a series of quarter - wave plates ( qwp ) , hwp and polarizing beam splitter ( pbs ) is used to perform the bell measurements . note : additional polarizers may be introduced before the interferometer to produce high quality separable states . ] the initial state is then subjected to polarization - dependent loss , introduced to the system by means of a jamin - lebedev polarization - interferometer , which allows individual control of the degree of loss for each polarization mode for both alice and bob , see fig .
[ fig : setup]c ) . in the symmetric case considered here
the loss was parametrized by a single parameter @xmath63 , where @xmath64 corresponds to the loss - free scenario and @xmath65 means complete loss of one polarization . with this setup we simulated correlations of increasing strength , ranging from classical to quantum and close to maximal
non - signaling as discussed in detail in the methods section . using these correlations we investigated the information causality protocol on up to four levels ( corresponding to a @xmath66-bit data - set for alice ) with @xmath9-bit of communication .
crucially , we implemented the protocol in fig .
[ fig : icmotivation]b ) on a shot - by - shot basis , rather than estimating the performance from coincidence probabilities . for this we used an ait - ttm8000 time - tagging module with a temporal resolution of @xmath67 ps to register the single photon counts for all possible outcomes . from this data , using passive feed - forward , i.e. at the processing stage , we were able to reconstruct over @xmath68 individual trials of the protocol for each of the @xmath69 settings of uniformly increasing @xmath63 .
at a correlation strength of @xmath70 , the information available to bob is at least @xmath71 bits , despite only receiving @xmath9 bit from alice . for four nesting levels of the protocol
we establish lower bounds as high as @xmath72 bit , which violates the information causality inequality @xmath73 by almost @xmath74 standard deviations .
similarly for weaker correlations , bob has more information available than contained in alice s message for all nesting levels as soon as the correlation strength surpasses @xmath75 .
the fact that this value is significantly higher than tsirelson s bound of @xmath76 emphasizes that the quantity @xmath77 only recovers this bound in the asymptotic limit @xmath78 .
iin the following we therefore consider an alternative figure of merit , motivated by identifying the protocol in fig .
[ fig : icmotivation]b ) as a special case of a so - called random access code @xcite . using similar ideas as in ref .
@xcite , the efficiency of this task can be bounded by @xmath79 which thus also encompasses the principle of information causality @xcite .
this bound , however , can indeed be saturated by quantum states for any size of alice s dataset , as illustrated in fig .
[ fig : resultsic : rac ] .
note that our data violates the bound before the correlations surpass tsirelson s bound .
this is a result of a slight anisotropy in the simulated correlations due to experimental imperfections and a resulting bias for certain data - sets .
it is not present when considering isotropic correlations , see fig .
[ fig : resultsic : rac]b ) .
crucially , this highlights the dependence of both figures of merit and to the specific random choice of alice s data - set .
( blue circles ) , @xmath20 ( red squares ) , @xmath80 ( yellow diamonds ) and @xmath81 ( green triangles ) levels in the protocol , where at each level a random dataset @xmath82 was used .
error - bars represent the standard deviation of 5 individual runs of every protocol .
the lines correspond to theoretical expectations for the given correlation strength . *
b ) * a zoom into the region where our data violates tsirelson s bound ( indicated by the grey , vertical line ) . our data violates the bound of @xmath83 already before the correlation strength surpasses tsirelson s bound , which is a result of a finite sample size and the particular choice of random dataset , see sec [ sec : supp1 ] . in the right panel , the same plot for isotropic correlations obtained from using the protocol of ref .
@xcite shows very good agreement with the theoretical predictions . ]
in particular , the separable state used in the simulation produces entanglement - like correlations for one measurement choice of alice and uncorrelated outputs for the other , see fig .
[ fig : correlations ] . hence , depending on the choice of data - set the figures of merit @xmath84 and @xmath77 might resemble the behavior expected for an entangled state , for a completely mixed state or , for higher nesting level , anything in - between . only when averaging over all possible datasets , @xmath82 , for a given level or employing the `` depolarization '' protocol introduced in ref . @xcite to make the correlations isotropic without changing the chsh value , can the quantities and be used as reliable figures of merit , see fig .
[ fig : correlations ] and [ fig : resultsdepolsep - i ] .
note , however , that anisotropic super - quantum correlations ( averaged over all datasets ) do not necessary violate tsirelson s bound . in this case
the principle of information causality can not be probed using the depolarization approach , since it would result in isotropic correlations and information causality would not be violated .
in contrast to the full set of no - signaling correlations , and the set of classical correlations , which both have the form of a well - characterized polytope , much less is known about the quantum set @xcite .
understanding the set of quantum correlations theoretically and characterizing it experimentally should thus be a primary aim from a practical as well as a fundamental perspective .
information causality , which has been proposed as a physical principle to reconstruct the set of quantum correlations , has already proven successful in recovering the famous tsirelson bound .
this limit of quantum correlations , however , is only one extremal point on the continuous boundary and there exist correlations below it , which nevertheless do not admit a quantum description @xcite .
information causality also rules out such correlations for some 2-dimensional slices of the full ( 8-dimensional ) no - signaling polytope , while it does not for other slices @xcite .
this shortcoming , nevertheless is not definite and might just be a result of a suboptimal protocol in fig .
[ fig : icmotivation]b ) .
a violation of information causality would in particular imply that the tested theory does not admit a suitable measure of one of the most elementary information theoretic quantities : entropy @xcite .
such a measure is assumed to be consistent with the classical limit and such that the entropy change @xmath85 of a composite system @xmath86 satisfies @xmath87 under local evolution of the subsystems @xmath88 and @xmath89 .
hence , a failure of these requirements could be interpreted as allowing for the generation of non - local correlations via local transformations .
similar consequences might also arise from the violation of alternatives to information causality , which are more or less successful in recovering part of the quantum boundary .
examples include the principles of local orthogonality @xcite , the requirement that the theory has a suitable classical limit @xcite or that certain communication @xcite or computational tasks @xcite are non - trivial .
our method of simulating super - quantum correlations could be adapted to explore some of these alternative principles as well .
of particular interest , however , would be a test of information causality in the multipartite case , since most of the above principles are formulated in the bipartite setting , which is bound to fail in recovering the full quantum boundary due to the existence of multipartite super - quantum correlations , which obey every bipartite principle @xcite . while there are studies of information causality for higher - dimensional systems , which strengthen its position as a physical principle that determines quantum correlations @xcite , a suitable generalization to the multipartite case is still an open problem .
as highlighted by our experiment , special focus has to be put on anisotropic regions of the no - signaling polytope .
specifically we find that the introduced figures of merit are not valid in a single instance of the protocol and have to be averaged over all possible datasets or estimated from the depolarized , isotropic data .
this subtle , but very important detail is clearly highlighted by our experimental results , where we show how even a small amount of imbalance can result in a violation of the principle by quantum states for a specific choice of parameters , while obeying the principle on average .
examining the results of a chsh - inequality test make it clear where our data crosses the boundary of the quantum set . in our investigation
we focused on the scenario of a fixed maximally entangled state @xmath90 in situations with different amounts of loss , as shown in fig .
[ fig : results ] .
we further considered the state @xmath91 , which resembles the state @xmath90 after full decoherence as might happen during propagation between alice and bob .
this allows for an intuitive comparison between the entangled and unentangled case .
the tested inequality has the form of a chsh - inequality with measurements in the yz - plane of the bloch sphere . for the lossless case @xmath92 , alice s and bob s measurements
can be viewed as the application of appropriate basis - rotations ( around the x - axis ) followed by projective measurements in the @xmath56-basis .
these rotations can also be seen as phase - gates in the diagonal polarization basis @xmath93 . in the case where polarization dependent loss is present ,
these phase - gates become non - unitary .
they act as the identity on the state @xmath94 and impose a phase on the non - orthogonal state @xmath95 , where @xmath96 . the precise relation between @xmath63 and the degree of loss
is discussed in ref .
@xcite . as non - unitary operations
can only be performed non - deterministically , postselection on success is required , which results in the observation of apparent super - quantum correlations .
finally , we use the first step of the depolarization procedure in ref .
@xcite to symmetrize the simulated correlations , while preserving their possible anisotropy .
curiously , we note that moderate polarization - dependent loss can lead to super - quantum correlations for entangled states without invalidating the chsh inequality for separable states , as suggested in ref . @xcite .
this observation even holds when optimizing the separable state for maximal chsh - value , for each degree of loss @xcite .
note , however , that these results were obtained using the same measurements for both separable and entangled states , whereas arbitrary hidden variable theories would allow arbitrary measurements . for both the entangled state @xmath97 ( blue circles ) and the separable state @xmath98 ( red squares ) , together with the theoretical predictions ( blue and red lines , respectively ) for these states , versus the amount of polarization - dependent loss as parametrized by @xmath63 .
the gray dashed line represents the theoretical expectation for the optimal separable state for a given amount of loss . in the experiment
we observe a violation of tsirelson s bound for @xmath99 .
interestingly , we identify a region ( @xmath100 ) where the quantum bound of the inequality is violated , while the classical bound still holds for all separable states . with the chosen , fixed , separable state @xmath98 we observe a first violation at @xmath101 .
errors from a monte - carlo sampling of the poissonian counting statistics are not visible on the scale of this plot . ]
figure [ fig : results ] illustrates the obtained values of the chsh parameter @xmath17 and compares them to the ideal case , which , for the initially entangled state , is described by @xmath102 here @xmath103 is a function of @xmath63 , which can be analytically approximated by @xmath104 , as discussed in ref .
@xcite .
we experimentally violate tsirelson s bound by more than @xmath105 standard deviations , @xmath106 at a loss parameter of @xmath107 . at this point ,
the achieved value for the unentangled state , @xmath108 , is indeed well below the classical bound of @xmath109 and even the optimal unentangled state does not violate the inequality until @xmath110 . in the region
@xmath111 it is therefore possible to exploit super - quantum correlations from entangled states while unentangled states still appear classical . with increasing loss
, both states eventually violate tsirelson s bound and approach the numerical maximum of @xmath18 , with experimental values of @xmath112 and @xmath113 for the entangled and separable state , respectively .
the increasing deviation from the theoretical predictions in fig .
[ fig : results ] is a result of the decreasing signal - to - noise ratio in the single - photon detectors for high - loss settings .
related experiments have observed apparent violations of tsirelson s bound as a consequence of explicit violations of the detection loophole @xcite or the fair - sampling assumption @xcite .
the latter is in fact typically violated if the quantum system of interest has more ( possibly ) correlated degrees of freedom than those tested in the bell - inequality @xcite .
violation of tsirelson s bound has also been considered as an intermediate step in deriving three - qubit inequalities @xcite .
we thank tim ralph and cyril branciard for helpful discussions .
we also thank the team from the austrian institute of technology for kindly providing the time - tagging modules for this experiment .
we acknowledge financial support from the arc centres of excellence for engineered quantum systems ( ce110001013 ) and quantum computation and communication technology ( ce110001027 ) .
is funded by an australian research council future fellowship ( ft100100761 ) , and a.f by an australian research council discovery early career researcher award ( de130100240 ) .
a.g.w . acknowledges support from a uq vice - chancellor s senior research fellowship .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . & .
_ _ * * , ( ) . , , &
_ _ * * , ( ) . .
_ _ * * , ( ) .
_ et al . _
_ _ * * , ( ) .
( ) . , & .
_ _ * * , ( ) . , , & .
_ _ * * , ( ) .
et al . _
_ _ * * , ( ) . & .
_ _ * * , ( ) . , &
_ _ * * , ( ) . , , , & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) .
_ et al .
_ _ _ * * , ( ) . & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . &
. _ _ * * , ( ) . , , & .
_ _ * * , ( ) .
_ _ * * , ( ) .
_ _ * * , ( ) . , & _ _ * * , ( ) . , , & .
_ _ * * , ( ) . , , , & _ _ * * , ( ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . * supplemental material * here we discuss in more detail the subtleties of the information causality protocol for individual choices of the data - set for alice . in particular we will discuss the effect of anisotropic correlations of decohered entangled states and present results for the case where
these are transformed to isotropic correlations by incorporating the protocol proposed in ref .
@xcite into our experiment .
as discussed previously , the experiment was performed for both entangled and separable initial states . while the entangled state allows for generation of super - quantum correlations in the range @xmath114 , the separable state covers a larger range of @xmath115 .
this range in particular includes a large part of the set of classical ( @xmath116 ) and quantum correlations ( @xmath117 ) . here
it is important to reiterate the form of the used state . since it has been chosen as the fully decohered version of @xmath90 it still retains correlations of the same strength in one basis , while correlations in any orthogonal basis are lost a classically correlated state . .
* shown are the theoretically expected measurement probabilities for both the * a ) * entangled and * b ) * separable state for a lossless bell - test experiment . note that the correlations are the same for the two states when alice chooses to measure @xmath9 , but there are no correlations for the separable state when she measures @xmath118 . ] an important consequence of this feature , as discussed in the main text , is that the success probability in the information causality protocol then depends on the specific choice of data - set for alice .
as an example consider the simplest instance , where alice has two bits @xmath119 and @xmath120 .
alice uses @xmath121 as her input and bob uses @xmath122
. the probability of success in this scenario will be the same as for the entangled state whenever @xmath123 ( that is for the data - sets @xmath124 and @xmath125 ) and the same as for random guessing ( @xmath126 ) in the other two cases , where @xmath127 .
although ref .
@xcite discussed the related case where the probability of success may depend on bob s choice , the feature revealed here has important practical implications , since the calculation of the figures of merit always requires to consider all of bob s possible choices , while only one data - set for alice has to be considered .
depending on the choice of data - set , any degree of efficiency can be achieved with classical states .
there are several ways to circumvent these problems .
clearly , averaging over all possible data - sets will recover the performance expected from the chsh - value corresponding to the respective state .
however , as the possible choices for alice s data - set grow doubly - exponentially , this approach is typically unfeasible , in particular when testing the protocol on a shot - by - shot basis .
here we employed the `` depolarization '' protocol introduced by ref .
@xcite , which takes any set of correlations to an isotropic one without changing the chsh - value , using 3 bits of shared randomness .
the drawback of this method , however , is that it precludes a test of information causality in the anisotropic regime of super - quantum correlations below tsirelson s bound , see ref .
figure [ fig : resultsdepolsep - i ] shows the experimental lower bounds on the mutual information measure @xmath77 for the isotropic correlations obtained from the depolarization protocol applied to the initially separable state .
figure [ fig : resultsdepolsep - rac ] shows the random access code efficiency @xmath84 for the same data . in both cases we observe excellent agreement with the theoretical predictions .
( blue circles ) , @xmath20 ( red squares ) , @xmath80 ( yellow diamonds ) and @xmath81 ( green triangles ) levels in the protocol , where at each level a random dataset @xmath82 was used .
error - bars represent the standard deviation of 5 individual runs of every protocol .
the lines correspond to theoretical expectations for the given correlation strength . ]
( blue circles ) , @xmath20 ( red squares ) , @xmath80 ( yellow diamonds ) and @xmath81 ( green triangles ) levels in the protocol , where at each level a random dataset @xmath82 was used .
error - bars represent the standard deviation of 5 individual runs of every protocol .
the lines correspond to theoretical expectations for the given correlation strength . ] | quantum correlations can be stronger than anything achieved by classical systems , yet they are not reaching the limit imposed by relativity .
the principle of information causality offers a possible explanation for why the world is quantum and why there appear to be no even stronger correlations . generalizing the no - signaling condition it suggests that the amount of accessible information must not be larger than the amount of transmitted information .
here we study this principle experimentally in the classical , quantum and post - quantum regimes .
we simulate correlations that are stronger than allowed by quantum mechanics by exploiting the effect of polarization - dependent loss in a photonic bell - test experiment .
our method also applies to other fundamental principles and our results highlight the special importance of anisotropic regions of the no - signalling polytope in the study of fundamental principles . |
in the universe , the galaxies forming stars with the highest rates and/or exhibiting the most extreme nuclear radiation turn out to be the most obscured ones .
their internal activity being partly dust - enshrouded , it remains underestimated in visible observing bands and shows up in the infrared regime .
such so - called luminous infrared galaxies ( lirgs ) seem to be more prevalent in relatively dense environments like groups where tidal galaxy interactions play a role in enhancing their activity . on the other hand ,
the densest regions like clusters are usually considered as being particularly unfavorable for the global activity of their host galaxies .
the combined action of multiple high - speed collisions , of the global cluster tidal field , and of the ram - pressure exerted by the rich intra - cluster medium contribute to strip galaxies of their disk and halo gas reservoirs . without this fuel ,
star formation is quenched and eventually stops while the galaxy center is no longer fed and any nuclear activity ends .
we present here data obtained with the isocam camera on board the infrared space observatory ( iso ) that reveals the presence of confirmed cluster members with high infrared luminosities and inferred star formation rates of several tens of solar masses per year .
these lirgs were discovered in the x - ray luminous cluster j1888.16cl , situated at a redshift of 0.56 .
the observations performed in the lw3 band centered at 15 @xmath1 m covered a strip of @xmath2 square arcminutes .
follow - up ground based observations were carried out using eso facilities in chile .
they consist of optical b , v , r wfi/2.2 m images , near - infrared j , h , ks sofi / ntt images and optical fors1/vlt multi - slit spectra . this collection of data allowed us to assess the cluster membership of the iso sources and study their spectral properties .
a total of 44 sources were detected at 15 @xmath1 m and the vast majority were unambiguously identified on optical / near - infrared images , thanks to the fair spatial resolution of isocam ( 4.6 " at 15 @xmath1 m ; see fig . 1 ) .
among the 27 objects for which spectra could be obtained , six have a redshift consistent with cluster membership while two other sources with slightly higher redshifts may lie in infalling groups .
the photometric redshift of all remaining mid - infrared emitters are consistent with a foreground or background location .
the most distant source at a redshift of 1.7 is an atypical hyperluminous broad absorption line quasar ( ( * ? ? ?
* duc et al . , 2002 ) ) .
the iso sources in the cluster are uniformly distributed along the 6 1.3 mpc@xmath3 strip mapped by isocam and have a velocity dispersion of @xmath4 , which is marginally comparable to that measured for all cluster members ( @xmath5 ) .
the confirmed cluster members detected by isocam have 15 @xmath1 m fluxes ranging between 200 and 750 @xmath1jy .
we estimated their total infrared luminosity using the mid - infrared to far - infrared correlation that has been well established for nearby galaxies observed with the iras satellite ( ( * ? ? ? * chary & elbaz 2001 ) ) .
all sources have an inferred ir luminosity above @xmath6 , and hence may be considered as genuine luminous infrared galaxies ( see fig . 2 ) .
either nuclear activity or star formation could heat the dust responsible for the infrared emission .
recently , a high number of galaxies shining in the hard x - rays ( ( * ? ? ?
* martini et al . , 2002 ) ;
* johnson et al . , 2003 ) ) and in the radio centimetric regime ( ( * ? ? ?
* morrison & owen 2003 ) ; ( * ? ? ?
* best , 2004 ) ) was found in chandra images and very large array maps of distant clusters , respectively .
they are suspected to host an agn .
however , studies based on infrared data obtained with iso and x - ray data collected by xmm - newton and chandra indicate that , statistically , infrared - selected luminous galaxies are predominantly powered by star formation , at least in the field environment ( ( * ? ? ?
* fadda et al . , 2002 ) ) .
all j1888 iso sources have weak emission lines in their optical spectrum ; none of the latter are broad enough to be classified as type 1 agns .
they typically exhibit an [ oii]3727 emission line with a moderate equivalent width ( less than 40 ) and strong early balmer hydrogen absorption lines ( the equivalent width of h@xmath7 is above 4 ) .
this spectroscopic signature is best explained by spectrophotometric models including a selective dust extinction depending on the stellar age and a high rate of current star formation ( ( * ? ? ?
* poggianti et al . , 2001 ) ) .
such an optical signature gives further support to the hypothesis that the infrared emission detected in the cluster results from a dust enshrouded starburst . from the inferred total infrared luminosities ,
one may deduce individual star formation rates ( sfr ) ranging between 20 and 120 ( a maximum of 270 including the two ulirgs for which the cluster membership is not assessed )
the discovery of extremely active galaxies in x ray luminous clusters was a priori unexpected given all the processes that eventually inhibit star formation in such an environment .
optical surveys of local clusters indicate a lack of galaxies with blue colors or exhibiting emission lines in their spectrum and hence showing signs of star formation activity .
such galaxies are however more numerous in more distant and younger clusters ( ( * ? ? ?
* dressler et al . , 1999 ) ) .
another evolutionary effect is the increased number of galaxies with spectra characterized by strong balmer absorption lines and no emission lines .
this signature is consistent with that expected for galaxies having suffered a recent but now faded intense star formation episode ( ( * ? ? ?
* poggianti et al . , 1999 ) ) .
this interpretation however was challenged ( ( * ? ? ?
* balogh et al . , 1999 ) ) since prompt truncated star formation may have the same effect in some cases .
the presence of post - starburst galaxies implies that starbursts may occur in clusters ; they would have remained unnoticed in optical surveys because of their dust cocoon . what triggered the enhanced infrared activity in the iso j1888 galaxies ?
the answer will very much depend on when and where the lirg event was switched on .
one simple hypothesis is that the mid - infrared emitters have recently been accreted by the cluster and that the luminous infrared phase actually started in the coeval field where lirgs are not uncommon . under this assumption , the infall rate that would account for the number of detected galaxies is constrained by the duration of the infrared luminous phase which can not exceed 100 myr . during that period
at least six field lirgs should have been accreted .
the proportion of lirgs in the field is still poorly known but could be of the order of 510 % in the redshift interval 0.5 - 0.6 ( ( * ? ? ?
* flores et al . , 1999 ) ) .
the total inferred infall rate ( including the ir - quiet galaxies ) would then be of about 100 massive galaxies in 100 myr , which seems unrealistic .
numerical simulations and x ray observations show however that accretion onto clusters from the field is not a spherically symmetric process , but occurs along filaments or via mergers with other groups and clusters .
one therefore can not exclude the possibility that the lirgs observed in j1888.16cl belonged to such a recently accreted structure . where tidal interactions could be an efficient mechanism in triggering the star formation .
the alternative is that the luminous infrared phase and associated starburst event occurred within the cluster .
there the large velocity dispersion does not favor tidal galaxy - galaxy collisions and mergers .
ram - pressure may create some shocks and instabilities in the gaseous component while it is stripped ; the changing global cluster tidal field may destabilize the galaxy disks , when they lie in infalling groups ( ( * ? ? ?
* bekki , 1999 ) ) or when a cluster - cluster merger is involved ( ( * ? ? ?
* miller & owen , 2003 ) ) .
such mechanisms can account for a mild star formation episode but are probably not efficient enough to concentrate large quantities of gas in the nuclear regions , as required to trigger a starburst or an agn activity with an intensity comparable to the local lirgs .
although the effective triggering mechanism remains unclear perhaps tidal interactions in sub - structures , our observations clearly indicate that a significant activity may occur in clusters , even at moderate redshifts and hence a long time after their collapse and the formation of the bulk of their stellar populations .
this result is corroborated with other studies based on isocam mid infrared data ( ricardo - perez et al .
, in these proceedings ) , optical data ( see the review by a. dressler ) , radio continuum maps and perhaps millimetric scuba surveys ( webb et al .
, these proceedings ; ( * ? ? ?
* best , 2002 ) ) .
the mips camera on board the infrared satellite sptizer will also very soon provide ample information on the dust enshrouded activity in clusters .
this secondary peak of activity may correspond to a phase in the evolution of clusters when accretion of galaxies , especially through the merging of groups , was conspicuous . | analysing mid infrared isocam images of the cluster of galaxies j1888.16cl , we identified among its members several particularly active galaxies with total infrared luminosities well above @xmath0 . if powered by dust enshrouded starbursts , as suggested by their optical spectra , these luminous infrared galaxies would exhibit star formation rates surprisingly high in the cluster environment .
the triggering mechanism is unclear but could be tidal collisions within sub - structures or infalling groups . |
in an x - ray image obtained with rosat , the vela supernova remnant ( snr ) is roughly circular with a diameter @xmath10 @xcite . at a distance of 250 pc @xcite
this corresponds to about 35 pc .
six bow shaped `` knots '' of emission lie beyond the nominal circumference of the remnant .
the shape and location of these knots led aschenbach et al . to suggest that they were due to ejecta `` bullets '' that had overtaken the blast wave .
of the six features , the closest to the remnant and also the brightest in x - ray is `` knot d '' .
the optical nebula rcw 37 @xcite , lies along the outer edge of knot d. this association with bright optical filaments also distinguishes knot d from the other five knots .
although it is grouped with the ejecta bullets , there are two other explanations for the origin of knot d. one scenario , suggested by @xcite is that knot d is a shock that has broken out from the vela snr due to inhomogeneities in the ambient medium .
they analyzed _
chandra_-acis spectra of knot d and found that the oxygen abundance was solar and that the neon abundance was modestly enhanced , and they found no evidence for abundance differences at different locations in the knot .
these results as well as the morphology of the x - ray emission led @xcite to favor a shock break - out origin for knot d. the other suggestion @xcite is that knot d is associated with rxj0852.0 - 4622 , a remnant that in projection is within the boundaries of the vela snr .
they analyzed optical echelle data of rcw 37 and found a velocity split near the bright optical edge in both [ s ii ] and [ o iii ] emission , and an almost complete velocity ellipse across the nebula in the latter .
they inferred that the geometry of the optical nebula is either an incomplete funnel or a wavy sheet .
based on the kinematic structure of emission as well as the optical and x - ray morphologies of the nebula they suggested that an outflow from rxj0852.0 - 4622 impacted the pre - existing wall of the vela snr to produce knot d. as they point out , for their model to work both remnants have to be at about the same distance @xcite .
regardless of its origin , it is known that knot d is a source of shock excited emission .
@xcite presented far - ultraviolet observations of the region obtained with the _ voyager 2 _ ultraviolet spectrometer .
strong emission from c iii @xmath6977 and o vi @xmath21032,1038 was detected in the spectrum .
the authors noted the sheet - like morphology of the filaments and the spatial coincidence of the x - ray and optical emission and argued that these were characteristics of a shock - cloud interaction ( such as had been seen in the cygnus loop snr ) and that this interaction produced the ultraviolet lines as well as the optical emission .
they suggested that the large _ voyager _ fov included shocks with a range of velocities , and that @xmath11 km s@xmath5 shocks were responsible for the c iii emission and 160300 km s@xmath5 shocks produced the o vi emission . in this paper
we take a detailed look at the shock - cloud interaction and determine the shock wave paramaters .
our results throw some light on the nature of knot d , but do not unambiguously determine its origin .
we use ultraviolet spectra obtained with the _ hopkins ultraviolet telescope _ ( hut ) and with the _ far ultraviolet spectroscopic explorer _ ( fuse ) .
these data have much higher spectral resolution than the _ voyager _ spectrum and the aperture fovs are over a thousand times smaller
. we also present optical longslit spectra of the filaments .
we confirm that the optical emission is shock excited . by comparing the measured far - ultraviolet emission line strengths with shock model predictions , we find the properties of the shocks around knot d. the kinematic distribution of the emission around the bow shock is clearly revealed in the high resolution fuse data .
the observations are described in 2 , and the results presented in 3 .
then , in 4 we analyze the results and present our interpretation .
the last section , 5 summarizes the central results of this work and presents some of the issues yet to be answered .
images of the region of interest are shown in figure [ f_image ] .
the top panel is a three color image of the southern half of knot d. narrowband h@xmath3 and [ o iii ] @xmath65007 images ( first presented by @xcite ) are shown in red and green , respectively .
the blue is the _ chandra _ x - ray image , presented by @xcite ( and kindly provided to us by the lead author ) .
overlaid on the image are the locations of the fuse lwrs and mdrs apertures .
the bottom panel is a two color image of a region around the bright filament .
only the optical emission is shown ( h@xmath3 in red , [ o iii ] in green ) overlaid with the hut and fuse lwrs aperture positions .
the hut observations were made during the astro-2 space shuttle mission on 1995 march 15 .
( the basic hut design is described by @xcite , and the improvements to the instrument for astro-2 by @xcite . )
the wavelength coverage of hut was 820
1840 with a resolution of @xmath12 3 .
the 56@xmath1310 aperture was centered at @xmath14 , @xmath15 , and placed at a position angle of 18 which aligned the aperture with the optical filament . after approximately 1000 s of exposure time at this position , the aperture was offset to a location @xmath1230 east , perpendicular to the aperture s long dimension .
we refer to these positions as p1 and p2 , respectively ( figure [ f_image]b ) .
the data were obtained in a time - tag mode which allowed the separation of orbital day and night photons .
since the daytime photons are contaminated by airglow emission , only the nighttime data are used here .
the effective exposure times are 932 s for p1 and 1162 s for p2 .
the hut data were processed using an iraf package originally developed to process hut data from the astro-1 mission , and described by @xcite .
the flux calibration applied to these data is based on in - orbit observations of white dwarf stars fitted with theoretical models .
other corrections , such as pulse persistence , dark count and ly@xmath3 scattering have been characterized and included here @xcite .
the fuse observation ( program i d b1080201 ) was obtained on 2001 april 2 as part of a guest investigator project to study the vela - puppis region .
eleven exposures with a total integration time of 14,133 s were obtained with the low resolution ( lwrs ) 30@xmath1330 aperture centered at @xmath14 , @xmath15 .
this position is the same as the center of the hut p1 observation , and lies on the bright optical filament ( figure [ f_image]b ) .
data were obtained simultaneously through the medium resolution ( mdrs ) 20@xmath134 aperture .
the mdrs aperture was located 208 away , 24 west of north from the lwrs aperture on an x - ray bright region away from the bright optical emission ( figure [ f_image]a ) .
the wavelength range covered by fuse is 905 1187 . in this paper
we present data from segments sic2a ( @xmath12905 1000 ) , lif1a ( @xmath121000 1100 ) and lif2a ( @xmath121100 1187 ) , the segments with largest effective area in their wavelength ranges @xcite .
the data in other segments were used for comparison to distinguish between real features and possible detector artifacts .
the raw data from all the exposures were combined and the pipeline calfuse version 1.8.7 was used to produce calibrated spectra .
a shift was applied to the lif1a and sic2a flux vectors to line up the geocoronal emission ( ly@xmath16 and ly@xmath17 respectively for the two channels ) at the appropriate wavelength in the heliocentric frame of reference .
this was done for both lwrs and mdrs spectra .
the source is an extended emission object and fills the apertures .
for such a target , the effective spectral resolution is a combination of the instrumental resolution ( @xmath120.05 ) , the slit width , and a further degradation caused by detector astigmatism @xcite .
the width of the airglow lines in a spectrum is a measure of the effective spectral resolution . in our data ,
the effective spectral resolutions are @xmath120.34 for the lwrs data and @xmath120.09 for the mdrs data .
( these numbers correspond to about 100 km s@xmath5 and 30 km s@xmath5 , respectively , at 1000 . )
the longslit optical spectra presented here were obtained in 1989 february at las campanas observatory .
the observations used the du pont 2.5 m telescope , the modular spectrograph and an @xmath18 pixel ti ccd .
a 600 line mm@xmath5 5000 blaze grating was used with an 85 mm camera , providing spectral coverage , 4750
7150 at 3.0 per pixel .
the slit width was 2 and the effective spectral resolution was 8 .
the spatial scale along the slit was 08 pixel@xmath5 , with a usable slit length of 7 .
the slit was placed along the east - west direction , passing through the optical filaments @xmath19 below the slit center of the hut position p1 .
the data were reduced using standard iraf procedures , including bias subtraction , flat fielding , image rectification and extraction of one dimensional spectra .
the background was determined from a portion of the slit lying outside the remnant and subtracted from the spectra .
the flux calibration was done using observations of standard stars from the list of @xcite .
fiducial stars along the slit , and the known spatial scale were used to extract regions of the long slit data corresponding to the intersection of the slit with the hut apertures . the optical images ( figure [ f_image]b ) show that the emission is fairly uniform within the apertures , so the optical spectra should be representative of the overall emission within the hut apertures at p1 and p2 .
the optical narrowband images show a striated emission structure .
there are bands of green ( strong [ o iii ] ) , yellow ( strong [ o iii ] and h@xmath3 ) and red ( strong h@xmath3 ) running parallel to the shock front ( figure [ f_image]b ) , indicating a systematic variation of the h@xmath3 to [ o iii ] flux ratio .
the h@xmath3 emission is at its brightest in the region between about 50 and 70 behind the leading [ o iii ] edge .
the [ o iii ] emission reaches its peak at about the same location as the h@xmath3 ( @xmath20 behind the edge ) . in the green ( [ o iii ] dominated ) band of emission , the [ o iii ] flux is about 40% of its peak value , while the h@xmath3 flux remains below 20% of its peak value .
the brightest optical filaments lie approximately parallel to the southeast edge of the x - ray emitting region ( figure [ f_image]a ) and there is a gap of about 1 between the x - ray edge and the filaments .
observations of other shock - cloud interactions in the cygnus loop @xcite and in vela @xcite have shown that optical filaments are typically coincident with the edges of bright x - ray features , and not separated as in knot d. so though the association of x - ray and optical regions is expected in snr shocks , the existence of a gap between the two is surprising .
the optical spectra of the regions of bright and faint h@xmath3 ( p1 and p2 , respectively ) are shown in figure [ f_opt ] .
we measured the integrated line fluxes using the iraf task , ` splot ` , and calculated the surface brightnesses .
we derive the color excess from the observed ratio between between the h@xmath3 and h@xmath16 fluxes .
the intrinsic ratio between the two lines for shock excited emission is 3.0 @xcite .
the observed ratio at p1 , i@xmath21/i@xmath22 , implies that @xmath23 , a value consistent with those obtained for other regions in vela @xcite . at position
p2 , we only have an upper limit for the h@xmath16 flux and can not estimate the color excess .
therefore we assume that the p1 value applies to the whole region .
we used the extinction curve suggested by @xcite , and total to visual selective extinction @xmath24 to calculate correction factors for all the lines .
the surface brightnesses of the optical lines , corrected for interstellar extinction are presented in table [ tbl_opt ] . the brightness ratio , i@xmath25}}$]/i@xmath21 is greater than 1 at both positions p1 and p2 .
the high ratios imply that the gas is collisionally excited and confirm that the optical filaments are due to shocks .
as expected from the narrowband images , the optical spectra at the two positions differ significantly in the strength of [ o iii ] emission relative to emission from the lower ionization lines .
the [ o iii ] @xmath24959,5007 flux is about 3 times weaker at position p2 than at p1 , whereas all the other lines are about an order of magnitude weaker at p2 .
some lines , including [ o i ] @xmath26300,6364 are not detected at the latter position .
this difference suggests that at p2 , the shock is `` incomplete '' - it has not yet swept up enough material for the post - shock gas to have recombined and produced strong h@xmath3 , [ s ii ] and other low ionization lines .
the hut spectra of the two positions in knot d contain numerous emission lines expected from snr shocks ( figure [ f_hut ] ) . as expected from the _ voyager _ uvs spectrum @xcite , c iii @xmath6977 and o vi @xmath21032,1038 are the strongest lines below 1200 .
other lines such as n iii @xmath6991 , which were lost in the low resolution _ voyager _
spectrum , are detected in these hut data .
longward of 1200 , the strongest lines are c iv @xmath61549 , o iii ] @xmath61663 and n iii ] @xmath61751 .
many ( but not all ) of the detected lines are marked in figure [ f_hut ] .
one surprise is that n v @xmath61240 is relatively weak : in spectra of radiative shocks in the cygnus loop the line is comparable in strength to the c iv line @xcite .
the integrated line fluxes in the spectra were measured using the iraf tasks ` splot ` and the more sophisticated ` specfit ` @xcite , the latter for blended lines .
the pair of lines , si iv @xmath61402 and o iv ] @xmath61403 could not be deblended automatically .
therefore , it was assumed that the si iv @xmath61402 flux was exactly half the si iv @xmath61394 flux ( valid for optically thin emission ) .
this amount was subtracted from the measured flux of the 1403 feature and added to the 1394 line .
the surface brightnesses of the lines , corrected for interstellar extinction , are listed in table [ tbl_hut ] along with the correction factors .
it should be noted that the interstellar extinction curve at wavelengths below 1000 is poorly known and is the main source of uncertainties in the intrinsic brightness of the shortest wavelength lines .
the line strengths at p1 are about twice those at p2 , but apart from that overall factor the spectra at the two positions are similar .
the dramatic differences in the relative intensities of optical lines are not present for the ultraviolet lines .
this implies that the shock conditions are similar at the two positions .
for instance i@xmath26/i@xmath27}1664}$ ] , a ratio that is very sensitive to the maximum temperature reached behind the shock , is @xmath28 at p1 and @xmath29 at p2 .
thus the far - ultraviolet spectra support the idea that the difference in the optical emission between p1 and p2 is due to different levels of shock completeness , rather than differences in shock velocity .
the main difference between the hut p1 and p2 spectra is the ratio between the two o vi doublet lines : i@xmath30/i@xmath31 at p1 and @xmath32 at p2
. the ratio approaches 0.5 when the line optical depth tends to zero , and approaches 1.0 when the line is optically thick .
the measured ratios show that the line fluxes are affected by resonance line scattering and that the effect is stronger at p1 than at p2
. the ratio between the o vi line fluxes can be affected by absorption of the 1038 line by molecular hydrogen ( the lyman band transitions , r(1)@xmath33 and p(1)@xmath33 , at 1037.15 and 1038.16 , respectively ) and by c ii * ( 1037.02 ) .
the ratio can also be affected by self - absorption of the lines by o vi in the ism .
@xcite find the column of low velocity h@xmath34 in the j=1 state @xmath35 , 10@xmath36 , and 10@xmath37 towards three stars behind vela .
even the highest of these would absorb only about 9% of the flux from a 150 km s@xmath5 width line centered at zero velocity .
the more typical lower values for the column would have negligible effect on the o vi @xmath61038 flux .
the c ii * line is stronger , but is at @xmath38 km s@xmath5 relative to the o vi @xmath61038 rest wavelength .
the velocity profile of the o vi lines in the fuse lwrs spectrum ( see below ) shows that c ii * absorption would not reduce the o vi flux significantly . the o vi column out to the distance of vela ( 250 pc ) is @xmath39 @xmath40 @xcite , which would have negligible affect on the line fluxes .
also , knot d lies beyond the remnant boundary , and it is unlikely that other shocked clouds associated with vela lie in front of it .
therefore , the ratios measured in the hut spectra should not be modified significantly by foreground absorption .
the fuse lwrs spectrum is shown in figure [ f_fuse ] .
the strongest snr lines in the fuse bandpass are c iii @xmath6977 and o vi @xmath21032,1038 .
these lines are allowed to go off - scale in the plot in order to show a number of the weaker lines present in the spectrum .
the weak lines are well separated from each other and are easily identified .
furthermore , snr lines can be distinguished from airglow lines by comparing the total spectrum with a spectrum screened to include only orbital night data .
many of these weaker lines have been detected in snr spectra for the first time by fuse .
for example , this paper reports the first such detection of s iii @xmath61077 .
note that the region around he ii@xmath41[n ii ] @xmath61085 is not included in the plots .
this range lies at the edge of only one detector segment @xcite , and the telescope effective area is very small in this region .
the total line fluxes in the lwrs spectrum were obtained by trapezoidal integration of the flux vector with suitable background subtraction .
the strong lines contain several thousand counts , so the random errors in their measured fluxes are low ( @xmath12 1% ) .
the errors are about 6% for some of the weaker lines , such as s iii @xmath61015 and s iv @xmath61063 . for the weakest line fluxes , particularly those detected in the sic 2a segment ( e.g. s vi @xmath2933 , 944 ) , uncertainties in the background placement contribute to the total error , which is about 20% .
the uncertainty in the absolute flux calibration is about 10% @xcite , which dominates the statistical errors in the measured fluxes of the strong lines and is comparable to those of the weak lines .
the surface brightnesses of the lines , corrected for interstellar extinction , are presented in table [ tbl_lwrs ] .
the fluxes of the strong lines ( c iii and o vi ) plotted on a velocity scale are shown in figure [ f_lw ] .
( the velocity is relative to the local standard of rest , in which frame the sun is moving away from vela at about 12.5 km s@xmath5 . )
the top panel is an overlay of the two o vi lines .
the lines are centered at @xmath42 km s@xmath5 and the fwhm of each line is @xmath43 km s@xmath5 . if the emission fills the slit uniformly , then the observed line profile is the intrinsic profile convolved with the line spread function , which , to very good approximation , is a 106 km s@xmath5 wide `` tophat '' function . under this assumption
the o vi lines have intrinsic fwhms of @xmath44 km s@xmath5 .
the surface brightnesses of the o vi in the fuse lwrs spectrum lies between the hut p1 and p2 values ( tables [ tbl_hut ] , [ tbl_lwrs ] ) .
however , the ratio i@xmath30/i@xmath45 in the fuse lwrs spectrum , which is lower than the value at either hut position .
thus , when detected over the more spatially extended lwrs aperture , resonance scattering has a smaller effect on the o vi emission than when detected over the hut aperture at p1 or p2 .
the feature on the blue wing of o vi @xmath61038 is identified as c ii @xmath61037.02 , a line also detected in a fuse spectrum of an x - ray bright knot in the center of vela @xcite .
the transition leading to the 1037.02 line has a lower state 63 cm@xmath5 above the ground state .
the companion line , c ii @xmath61036.34 is a strong ground state transition @xcite and is not detected because it is absorbed by the interstellar gas between us and the snr .
the bottom panel of figure [ f_lw ] shows an overlay of the c iii and the o vi @xmath61032 lines .
the c iii line is significantly broader than the o vi line , and it has a central reversal . by fitting the wings of the c iii profile with gaussians
we find that the observed peaks are centered at @xmath47 km s@xmath5 and @xmath48 km s@xmath5 , and that the components have fwhms @xmath4 km s@xmath5 .
c iii is a strong resonance line , and while the observed profile is consistent with there being two kinematic components , we expect optical depth effects to influence the line profile .
some idea about the intrinsic velocity distribution of low ionization emission is gleaned from the [ s ii ] @xmath66716 position - velocity data presented by @xcite . in an echelle spectrum with the slit cutting across the optical filament about 2 south of the fuse lwrs position , @xcite detected bright [ s ii ] emission around zero - velocity .
in addition to this central component , they found fainter red - shifted and blue - shifted peaks ( see their figure 4 ) .
the intrinsic velocity structure of the c iii emission is likely to be as complex , in which case the reversal in the observed profile is due to zero - velocity emission being self - absorbed by both filament material and the ism along the sightline .
c iii @xmath6977 and the o vi doublet are detected in the fuse mdrs spectrum , obtained at a position several arcminutes behind the bright optical filaments .
the line fluxes are plotted against lsr velocity in figure [ f_md ] .
the two o vi lines track each other closely ( figure [ f_md ] top panel ) .
each o vi line profile has two peaks , centered at @xmath49 km s@xmath5 and @xmath48 km s@xmath5 .
the width of the gap between the two components is about 90 km s@xmath5 , which is much greater than the typical velocity width @xmath50 km s@xmath5 for absorption by interstellar o vi @xcite .
we conclude that we are seeing two distinct emitting components ( rather than a single , self - absorbed component ) along the line of sight at this location .
the kinematic structure of the c iii emission follows that of the o vi emission ( figure [ f_md ] bottom panel ) .
the c iii line has two peaks , centered at the same velocities as the o vi lines .
the separation between the two c iii velocity components is larger in the the mdrs spectrum than in the lwrs spectrum .
because of its bow shape , and because of the lack of complex optical morphology at the location of the mdrs aperture ( figure [ f_image ] ) , we conclude that the two emission components are due to shocks on the front and back sides of knot d. the blue - shifted o vi lines have fwhms of about 70 km s@xmath5 , and are broader than their red - shifted counterparts , which are about 50 km s@xmath5 wide ( fwhm ) .
both c iii components have fwhms of about 50 km s@xmath5 .
the surface brightnesses , corrected for interstellar extinction , of the blue and red shifted components of the three lines are presented in table [ tbl_mdrs ] .
the blue shifted component of c iii @xmath6977 is about 1.3 times as bright as the red shifted component .
in contrast , the total o vi flux of the blue shifted component is just half that of the red shifted component .
i@xmath51/i@xmath52 is about 0.9 for the red shifted component and about 0.4 for the blue shifted component .
the ratios between the doublet lines , i@xmath30/i@xmath53 are 0.54 for the blue shifted component and 0.56 for the red shifted component .
these ratios are closer to the optically thin limit compared with the values obtained for the hut and fuse lwrs spectra , which is consistent with the less edge - on viewing geometry expected at this position .
to facilitate interpretation of the ultraviolet spectra , we compare the data with shock model calculations .
we focus on the bright filament covered by the hut p1 and fuse lwrs apertures , and generate a list of line strengths relative to i@xmath27 } } = 100 $ ] .
the two apertures cover different regions on the sky , and the o vi surface brightness is lower in the fuse lwrs spectrum .
we use the o vi to o iii ] line ratio measured in the hut p1 spectrum .
the line strengths of other high ionization lines below 1200 , s vi , ne vi ] and ne v ] , are scaled relative to i@xmath51 based on the fuse lwrs measurement , and then rescaled relative to i@xmath27}}$ ] based on the hut p1 o vi to o iii ] line ratio .
the fuse lwrs surface brightnesses are used for all other lines in the fuse bandpass .
these data are shown in column 3 of table [ tbl_mods ] .
shock models were calculated using an updated version of the code presented by @xcite .
the updates are discussed by @xcite , and some atomic rates for si iii , si iv and s iv lines have been updated based on the chianti database @xcite .
the model follows the emission and cooling behind a steady shock moving into a constant density medium .
the main inputs are the shock velocity , the pre - shock density and the elemental abundances .
models are presented for a range of shock velocities . in each case the pre - shock hydrogen number density , @xmath54 is 1 @xmath1 .
( the ultraviolet line fluxes scale linearly with pre - shock density so model line ratios do not depend on @xmath54 . )
elemental abundances , h : he : c : n : o : ne : mg : si : s : ar : ca : fe : ni = 12.00 : 11.00 : 8.55 : 7.97 : 8.79 : 8.07 : 7.58 : 7.55 : 7.21 : 6.60 : 6.36 : 7.51 : 6.25 on a logarithmic scale , are used in the models . except for the oxygen abundance ( discussed below ) these are solar abundances based on @xcite and tabulated by @xcite . in all the models the pre - shock magnetic field is 1@xmath55 g
, the temperature of the pre - shock gas 10,000 k , and the gas is fully ionized .
these parameters do not affect the ultraviolet line strengths significantly .
for each model , the calculation is followed until the post - shock gas reaches about 1000 k , by which point the recombination zone is complete .
the ratio of n iii ] @xmath61750 to o iii ] @xmath61664 is relatively insensitive to the shock velocity . for the range of shock velocities between about 80 km s@xmath5 and 120 km s@xmath5 , the ratio is a measure of the relative abundances of nitrogen and oxygen .
there is some amount of uncertainty in the value for solar oxygen abundance .
@xcite , from which we have taken the other abundances , lists [ o ] = 8.87 .
recently , it has been suggested that the actual value is lower than normally assumed : @xcite gives [ o ] = 8.73 , and @xcite give [ o ] = 8.69 .
we ran a set of 100 km s@xmath5 shock models varying the oxygen abundance between 8.69 and 8.87 ( and keeping other abundances fixed ) .
the observed n iii ] to o iii ] line ratio is obtained for oxygen abundance , [ o ] = 8.79 , and we used this value in the models presented in columns 49 of table [ tbl_mods ] .
the optical depths of the strongest resonance lines observed are of order unity in the direction normal to the shock front . when viewed close to edge - on ( as expected at and near p1 ) , the optical depths are higher . a significant fraction of the line photons emitted by the shocked gas are scattered out of the line of sight , reducing the observed fluxes . while comparing spectra with shock model predictions , it is necessary to take into consideration this decrease in fluxes due to resonance line scattering . in the case of the o vi lines ,
the ratio between the intensities of the two lines of the doublet is a measure of the line optical depth .
if we assume that the optical depth along the line of sight is much higher than the transverse optical depth then , as described by @xcite , we can calculate the intensity correction factor . for an observed ratio i@xmath30/i@xmath31 ,
the intensity correction factor is @xmath56 ( for the sum of both lines ) .
the carbon lines , c iii @xmath6977 , c ii @xmath61335 and c iv @xmath61549 are all significantly affected by resonance scattering .
the flux ratio between c iii @xmath6977 and c iii @xmath61176 provides an estimate of the effects of resonance scattering on the 977 line .
the models predict a ratio of about 80 , while the observed ratio is about 20 , indicating a correction factor of about 4 .
this conclusion is somewhat compromised because of uncertainties in the model predictions of the 1176 line .
we expect that n v @xmath61240 flux is less affected by resonance scattering than any of these other lines because nitrogen is less abundant .
specifically , we may assume that the correction factor for the 1240 line ( actually an unresolved doublet ) is not more than the correction factor for the o vi doublet derived above .
the ultraviolet spectrum provides several diagnostic line ratios that can be used to estimate the shock velocities .
the ratios n iv ] @xmath61490 to n iii ] @xmath61750 and o iv ] @xmath61403 to o iii ] @xmath61604 are particularly useful : the lines are intercombination lines and so are not subject to optical depth effects , and the ratios do not depend on abundances . at the hut p1 position , i@xmath57}}$]/i@xmath58}}~\sim0.32 $ ] and i@xmath59}}$]/i@xmath27}}~\sim0.44 $ ] .
a shock velocity @xmath60 km s@xmath5 is required to reproduce these observed values .
it is clear from table [ tbl_mods ] that such a low velocity shock can not produce the observed flux of higher ionization lines such as n v @xmath61240 and the o vi doublet .
if we assume that both n v and o vi are equally affected by resonance scattering then i@xmath51/i@xmath61 . to produce this ratio requires a shock velocity of about 170 km s@xmath5 .
if n v is less affected by resonance scattering , then the ratio is higher and so is the required shock velocity .
the ratio between ne vi ] @xmath61006 and ne v ] @xmath61146 is less accurate because the lines are weak , but these lines are not affected by resonance scattering , and the observed ratio implies shock velocities in excess of 180 km s@xmath5 .
finally , we note that the ratio of optical lines i@xmath62}}$]/i@xmath25}}~\sim3.1 $ ] ( table [ tbl_opt ] ) , which requires shock velocities of at least 80 km s@xmath5 .
( the forbidden line strengths are predicted by shock models but are not presented in table [ tbl_mods ] . ) from these comparisons we find that the lower ionization lines , including o iii ] , o iv ] , n iii ] and n iv ] , come predominantly from shocks with velocities @xmath63 km s@xmath5 and the higher ionization lines such as o vi , n v , s vi , ne v ] and ne vi ] come from faster shocks , with velocities about 180 km s@xmath5 .
the latter velocity falls within the range given by @xcite for o vi producing shocks , but our value for the velocity of the slower shock is less than their estimate for the c iii producing shocks ( see 1 ) .
intermediate velocity shocks can not contribute a significant fraction of the emission since that would affect all the line ratios in the wrong way .
for example , if there were a substantial contribution from say a 140 km s@xmath5 shock , then i@xmath57}}$]/i@xmath58}}$ ] would be higher than observed while i@xmath51/i@xmath64 would be lower than observed . the lack of correlation between the c iii and o vi velocity distributions in the fuse lwrs spectrum ( figure [ f_lw ] ) provides corroborating evidence that the two lines arise in different shocks . since the ultraviolet line ratios are similar in both hut spectra ( 3.2 )
, the arguments and conclusions presented above for p1 also hold for p2 .
the observed o iii ] @xmath61664 surface brightness , @xmath65 erg s@xmath5 @xmath40 arcsec@xmath66 ( table [ tbl_hut ] ) . to proceed with our analysis
, we make the simplifying assumption that a 100 km s@xmath5 shock is responsible for all the o iii ] emission .
a 100 km s@xmath5 shock model using a pre - shock hydrogen number density @xmath67 @xmath1 predicts an o iii ] line intensity @xmath68/@xmath69 erg s@xmath5 @xmath40 sr@xmath5 ( table [ tbl_mods ] ) . this is equivalent to @xmath70 erg s@xmath5 @xmath40 arcsec@xmath66 .
the model prediction scales with the pre - shock density .
also the model predicts the intensity emerging perpendicular to the shock front . since the shock front is viewed close to edge - on , the effective area of the shock observed is larger than the aperture area .
the @xmath71 hut aperture is placed parallel to the shock front .
if the path length through the o iii ] emitting gas is @xmath72 ( ) , then the ratio of the shock area to aperture area is @xmath73 .
thus , @xmath74 , where @xmath75 is the pre - shock density for the 100 km s@xmath5 shock component . substituting the values for observed and model surface brightnesses , we obtain : @xmath76 . below we find that the dynamic pressure of the shocks observed at the fuse mdrs position , @xmath77 dyne @xmath40 . assuming that this is the pressure driving the shocks at p1 , for @xmath78 km s@xmath5 and [ he ] = 11.0 , we obtain @xmath79 @xmath1 .
this , in turn , implies that the path length through the emitting gas is about 164 , which is about 0.2 pc at the distance of vela .
the length of the optical filament in the plane of the sky is about 430 , which is significantly higher than the derived path length through the o iii ] emitting gas .
the difference is probably due to the curvature of the shock front into the plane of the sky , but could also be because the cloud that the shock is running into is approximately cylindrical with its long direction oriented in the plane of the sky .
as we showed above , the high ionization lines are produced by faster shocks .
for a 180 km s@xmath5 shock , the isobaric condition yields a pre - shock density @xmath80 @xmath1 . at the fuse mdrs position , the observed o vi to c iii ratios are 0.4 for the blue - shifted component and 0.9 for the red - shifted component ( table [ tbl_mdrs ] ) .
( note : no correction for resonance scattering was made in this case . )
we ran a set of shock models with velocities in the range 140160 km s@xmath5 spaced by 5 km s@xmath5 . by interpolating from the predicted line ratios we found that the observed values for the blue - shifted and red - shifted components are satisfied for shock velocities of @xmath81 km s@xmath5 and @xmath82 km s@xmath5 , respectively .
we note for shock velocities from 140 to 160 km s@xmath5 that @xmath83 increases rapidly while i@xmath52 stays fairly constant ( table [ tbl_mods ] ) . if our line of sight intersects with a shock front once , then the following holds @xcite : @xmath84 , where @xmath85 is the intensity of the shock viewed face - on , @xmath86 is the observed radial velocity ( absolute value ) and @xmath87 is the shock velocity . we apply this to the c iii emission .
the observed surface brightnesses are @xmath88 and @xmath89 erg s@xmath5 @xmath40 arcsec@xmath66 for the blue- and red - shifted components ( table [ tbl_mdrs ] ) .
the central velocities of the components are @xmath49 km s@xmath5 and @xmath48 km s@xmath5 . using these values in the equation above yields @xmath90 erg s@xmath5 @xmath40 arcsec@xmath66 .
the face - on c iii intensity predicted by a 150 km s@xmath5 shock model for pre - shock density 1.0 @xmath1 , @xmath91 erg s@xmath5 @xmath40 arcsec@xmath66 .
the line intensity scales with pre - shock density , so the pre - shock density required to produce the observed c iii brightness is @xmath92 @xmath1 . using the values for shock velocity and pre - shock density derived above , we find that the pressure in the x - ray knot driving the shock ( @xmath93 ) is about @xmath94 dyne @xmath40 ( @xmath95 @xmath1 k ) .
the dynamical pressure of the shock derived above is about twice the value derived from the hut spectrum of a face - on shock near the center of the remnant @xcite , and about seven times the value in the x - ray region associated with the face - on shock @xcite .
it is also about 25 times the characteristic pressure of the x - ray emitting gas derived by @xcite , corrected for the revised distance of 250 pc to vela .
the high pressure is consistent with knot d originating in a bullet near the center of the remnant in that the bullets have to have ram pressures substantially larger than the average to punch out through the shell .
the enhanced pressure would not be consistent with the knot being a blister on the surface , such as formed in the model suggested by @xcite for expanding snr shells .
it is interesting that @xcite derived a pressure slightly higher than our value in the region around the line of sight towards hd 72089 , which is also near some arcuate optical filaments .
they also found several high velocity components in the absorption spectrum of the star .
the enhanced pressure suggests that the high velocity emission and optical filaments at the position observed by @xcite may be related to a bullet similar to knot d but traveling towards us .
we have used steady flow shock models to explain the observed emission lines . at the mdrs location , these lines arise in a shell within the surface of an evolving bow shock . @xcite
approximate a herbig - haro bow shock by a series of plane parallel , steady flow oblique shocks ; an approximation that they find valid if the cooling time is short compared to dynamical times .
knot d is about 30 ( @xmath96 cm ) beyond the main blast wave . assuming that it is moving at @xmath97 km s@xmath5 @xcite , it has taken about 4000 years for it to get to its current position since overtaking the blast wave .
the cooling time for a 150 km s@xmath5 shock running into 5 @xmath1 material is about 300 years .
the front of the bow shock is running into a denser part of the cloud .
this encounter results in a slower shock that gives rise to the optical filaments as also the low ionization ultraviolet lines .
the use of steady flow models has another limitation : shocks of 180 km s@xmath5 are subject to thermal instabilities @xcite .
the structure and spectrum of such a shock fluctuate in time @xcite , but the spectrum averaged over time is not greatly affected @xcite .
the spectrograph apertures include emission from sufficiently complex filaments that the fluctuations are probably smoothed out .
we have analyzed far - ultraviolet and optical observations of knot d in the vela snr , and shown that the emission is due to shocks driven into a medium with typical interstellar cloud densities and normal elemental abundances .
we have also found that the dynamical pressure of the shocks is higher than the characteristic value for the remnant .
we have estimated that the path length through the optical filaments is shorter than their extent in the plane of the sky implying a curved shock front .
the velocity profiles of the lines revealed by high resolution fuse mdrs is consistent with emission from a bow shock .
the over - pressure in knot d supports the idea that it is a bow shock around a bullet or cloud that originated near the center of the remnant and has punched through the shell . however , there is no evidence for enhanced abundances .
if the driver of the bow shock is a shrapnel of ejecta @xcite , then it remains undetected .
an alternative candidate for the driver is an accelerated cloud that originated near the center .
the existence of such clouds was suggested by @xcite to explain some of the high velocity emission observed in snrs .
knot d , in spite of its unique properties , may be one of many such features including the ones observed in the plane of the sky around vela , as well as others moving radially towards or away from us . in any case
, our study provides an estimate of the properties of knot d that can be usefully compared with results derived from x - ray data .
the properties of knot d can also be compared with those of other regions to find out the extent of their similarities .
the suggestion that knot d is associated with rxj0852.0 - 4622 @xcite has been made mainly on morphological grounds .
the connection also depends on the smaller remnant being at about the same distance as the vela snr .
the distance and also the age of rxj0852.0 - 4622 are highly uncertain @xcite and provide very few constraints on the properties of putative blow - out regions .
we have found that the interstellar extinction towards the optical filaments is typical for the vela snr .
furthermore our ultraviolet data do not show any peculiarities ( e.g. in the velocity distribution ) that suggest a connection between the knot and the newly discovered remnant .
therefore , we believe that knot d is associated with the vela snr .
it will require additional studies to prove or disprove conclusively the association between rxj0852.0 - 4622 and knot d. one major issue that remains unanswered is the gap of about 1between the optical filaments and the edge of the x - ray knot ( figure [ f_image ] ) .
this is not seen in typical shock - cloud interactions ( see references in 3.1 ) where the x - ray edge and optical emission are co - incident . in those cases , the relationship of the two components is well explained by a model in which the blast wave hits the cloud and a reverse shock is propagated back into the remnant @xcite . in knot d , the gap must be a consequence of the dynamics of the bow shock .
the gap region could contain gas that is hot but with a temperature not high enough to make x - rays .
the region may then be a source of bright o vi emission .
it could also be a transient region of low pressure as predicted by models of thermal instability @xcite .
a third possibility is that it is a region of cool gas supported by magnetic pressure .
each of these possibilities has interesting implications for the structure and evolution of astrophysical bow shocks , but the existing data do not allow us to distinguish among them . however , it is worth noting that the spatial extent of the gap may be related to the cooling length of the shocked gas . at the distance of vela ,
1 is about @xmath98 cm , which is roughly the cooling length scale of a 240 km s@xmath5 shock in a medium with density 3 @xmath1 .
these conditions may have obtained in the course of the evolution of the bow shock .
detailed models and direct observations are needed for further elucidation .
we thank paul plucinsky for useful discussions , and for providing us with the chandra image presented in figure 1 .
we also thank matt redman for giving us access to his [ s ii ] position - velocity data .
h2ools , a software package written by steve mccandliss was used in estimating the effects of molecular hydrogen absorption .
the referee s comments helped improve the flow and focus of the paper .
this work has been supported by nasa grant nag5 - 10248 and nasa contract nas5 - 32985 , both to the johns hopkins university .
h@xmath16 & 4861 & 9.1 & @xmath991.0 & 1.40 + [ o iii ] & 4959 & 25.9 & 9.2 & 1.38 + [ fe iii ] & 4987 & 2.4 & & 1.38 + [ o iii ] & 5007 & 80.7 & 28.7 & 1.38 + [ fe ii]+fe[iii ] & 5270 & 0.9 & & 1.35 + [ n ii ] & 5754 & 1.2 & & 1.30 + he i & 5876 & 0.7 & & 1.29 + [ o i ] & 6300 & 2.6 & & 1.26
+ [ o i ] & 6364 & 0.9 & & 1.26 + [ n ii ] & 6548 & 8.4 & 0.7 & 1.24 + h@xmath3 & 6563 & 27.5 & 2.7 & 1.24 + [ n ii ] & 6583 & 26.5 & 2.3 & 1.24 + [ s ii ] & 6716 & 19.7 & 2.5 & 1.23 + [ s ii ] & 6731 & 14.6 & 1.5 & 1.23 + s vi & 933 & 3.6 & 3.1 & 5.38 + s vi & 944 & 2.9 & 2.5 & 5.12 + c iii & 977 & 42.4 & 19.7 & 4.47 + n iii & 991 & 15.6 & 7.8 & 4.25 + o vi & 1032 & 22.5 & 10.0 & 3.73 + o vi & 1038 & 17.0 & 6.1 & 3.66 + s iv & 1064 & 1.8 & 0.7 & 3.42 + s iv & 1074 & 2.6 & 1.6 & 3.34 + he ii+[n ii ] & 1085 & 4.5 & 1.6 & 3.25 + [ ne v ] & 1146 & 1.3 & & 2.88 +
c iii & 1176 & 2.6 & 1.4 & 2.74 + n v & 1240 & 5.2 & 3.4 & 2.52 + c ii & 1335 & 4.5 & 2.2 & 2.30 + o v & 1371 & @xmath990.7 & @xmath990.6 & 2.24 + si iv & 1393 & 9.3 & 4.7 & 2.21 + o iv ] & 1403 & 13.7 & 9.1 & 2.20 +
n iv]+[n iv ] & 1485 & 3.9 & 3.0 & 2.11 + c iv & 1549 & 26.4 & 15.2 & 2.07 + [ ne iv ] & 1602 & 0.8 & 1.2 & 2.04 + he ii & 1640 & 11.4 & 3.7 & 2.03 + o iii ] & 1664 & 31.1 & 13.8 & 2.03 + n iii ] & 1750 & 12.8 & 5.5 & 2.02 + s vi & 933 & 1.6 + s vi & 944 & 1.0 + c iii & 977 & 40.6 + n iii & 991 & 11.1 + ne vi & 1006 & 0.8 + s iii & 1015 & 0.7 + s iii & 1021 & 0.9 + o vi & 1032 & 17.6 + o vi & 1038 & 10.3 + s iv & 1063 & 2.2 + s iv & 1074 & 1.9 + s iii & 1077 & 0.6 + si iii & 1110 & 0.3 + si iii & 1113 & 0.4 + si iv & 1123 & 0.3 + si iv & 1128 & 0.6 + ne v & 1137 & 0.2 + ne v & 1146 & 0.5 + c iii & 1176 & 1.8 + s vi & 937 & 12 & 0 & 0 & 2 & 11 & 29 & 30 + c iii & 977 & 131 & 1073 & 1154 & 1070 & 528 & 351 & 418 + n iii & 991 & 36 & 86 & 88 & 93 & 101 & 67 & 67 + ne vi ] & 1006 & 4 & 0 & 0 & 0 & 0 & 2 & 15 + s iii & 1015 & 5 & 5 & 4 & 5 & 4 & 3 & 3 + o vi & 1034 & 127 & 0 & 0 & 0 & 12 & 447 & 2221 + s iv & 1070 & 13 & 5 & 7 & 9 & 12 & 7 & 8 + s iii & 1077 & 2 & 2 & 2 & 2 & 2 & 1 & 1 + si iii & 1112 & 2 & 14 & 10 & 4 & 3 & 3 & 3 + si iv & 1125 & 3 & 3 & 4 & 2 & 1 & 1 & 1 + ne v ] & 1146 & 3 & 0 & 0 & 1 & 4 & 20 & 45 + c iii & 1176 & 6 & 12 & 14 & 13 & 7 & 4 & 5 + n v & 1240 & 17 & 0 & 1 & 14 & 81 & 229 & 184 + c ii & 1335 & 14 & 263 & 243 & 210 & 96 & 80 & 89 + o v & 1371 & @xmath992 & 0 & 0 & 1 & 4 & 20 & 32 + si iv & 1396 & 30 & 129 & 113 & 47 & 23 & 22 & 26 + o iv ] & 1403 & 44 & 9 & 49 & 85 & 131 & 183 & 191 + n iv ] & 1490 & 13 & 6 & 20 & 30 & 46 & 34 & 30 + c iv & 1549 & 85 & 83 & 398 & 711 & 638 & 317 & 360 + he ii & 1640 & 37 & 4 & 19 & 44 & 38 & 33 & 36 + o iii ] & 1664 & 100 & 100 & 100 & 100 & 100 & 100 & 100 + n iii ] & 1750 & 41 & 42 & 41 & 42 & 39 & 27 & 27 + + i(o iii ] ) & 1664 & & 0.33 & 0.45 & 0.61 & 1.21 & 1.53 & 1.56 + | we present spectra of optical filaments associated with the x - ray knot d in the vela supernova remnant .
it has been suggested that knot d is formed by a bullet of supernova ejecta , that it is a break - out of the shock front of the vela snr , and also that it is an outflow from the recently discovered remnant rxj0852.0 - 4622 .
we find that knot d is a bow shock propagating into an interstellar cloud with normal abundances and typical cloud densities ( n@xmath0 4 11 @xmath1 ) .
optical longslit spectra show that the [ s ii ] @xmath26716,6731 to h@xmath3 line ratio is greater than unity , proving that the optical filaments are shock excited .
the analysis of far - ultraviolet spectra obtained with the hopkins ultraviolet telescope and with the far ultraviolet spectroscopic explorer ( fuse ) lwrs aperture show that slower shocks ( @xmath4 km s@xmath5 ) produce most of the low ionization lines such as o iii ] @xmath61662 , while faster shocks ( @xmath7 km s@xmath5 ) produce the o vi @xmath21032,1038 and other high ionization lines .
c iii and o vi lines are also detected in the fuse mdrs aperture , which was located on an x - ray bright region away from the optical filaments .
the lines have two velocity components consistent with @xmath8 km s@xmath5 shocks on the near and far sides of the knot .
the driving pressure in the x - ray knot , @xmath9 @xmath1 k , is derived from the shock properties .
this is over an order of magnitude larger than the characteristic x - ray pressure in the vela snr .
the velocity distribution of the emission and the overpressure support the idea that knot d is a bow shock around a bullet or cloud that originated near the center of the vela remnant . |
quantum entanglement is nonlocal and excess - classical correlation between separate parties , which is a most important character of quantum mechanics .
lots of interest has focused on the nature of entanglement and the structure of entangled states .
besides , entanglement is also the most central and indispensable resource in quantum information processing such as quantum computation , quantum teleportation @xcite , quantum dense coding @xcite , and quantum key distribution @xcite . in a word ,
quantum entanglement is not only importance in theory but also in practical applications .
one the other hand , it is well known that decoherence @xcite is a vital factor that should not be neglected in quantum information processing .
the coupling between any quantum system and its environment is inevitable .
thus the entanglement will evidently be reduced and even disappear because of this system - environment coupling , e.g. in a large scale quantum computer or during the course of entanglement distribution via noisy channels .
the stability of entanglement depends on the initial entangled system its entanglement structure and its size
. we may look on entanglement as a bond between different qubits , just like the chemical bond between different atoms .
people studied the behavior of chemical bonds in different environment to understand how chemical bonds are formed . thus investigating
the entanglement dynamics of different types of entangled states in the decoherence environment may help to gain some insight into the properties of the decoherence and the entanglement , which will provide useful hints for maintaining entanglement .
and what kind of entanglement bond is the most stable under different decoherence models is an interesting problem . in this paper
, we investigate the evolution of pairwise entanglement for two - qubit entangled states in the decoherence model which is described by general pauli channels .
we use the concurrence of wootters@xcite , which is related straightforwardly to the entanglement of formation ( eof ) , as the measure of entanglement for two - qubit entangled states .
the most interesting problem is that given some general decoherence model , what kind of entangled states can maintain entanglement best .
we find that with the same initial entanglement , the lifetime of entanglement in some specific class of entangled states is the longest . for a special decoherence model ,
that is depolarizing channels , all pure states together with some mixed states , which we call as decoherence path states ( dps ) is the most entanglement - stable .
we present the analytic dynamics of two - qubit entanglement for these special entangled states .
furthermore , we also study the stability of the nearest neighbor entanglement in the ground state of some spin models such as heisenberg and xxy model .
coincidentally , the conclusion is that in some noise models , the nearest neighbor entanglement in the ground state is also the most stable , though it is not maximized @xcite .
the paper is organized as follows . in sec .
ii we introduce the entanglement measure of two - qubit entanglement and the decoherence model , which can be viewed as a completely positive map . in sec .
iii we investigate the dynamics of two - qubit entanglement under the influence of decoherence and try to find the special entangled states of which the pairwise entanglement is the most stable . in sec .
vi we examined the evolution of entanglement for some specific and maybe important mixed states , e.g. the ground states of spin models and the maximally entangled mixed states etc . in sec .
iv conclusions and discussions , together with some interesting open questions are presented .
there have been a number of measures for two - qubit entanglement , such as the entanglement of formation @xcite , negativity @xcite and relative entropy of entanglement @xcite etc . in this paper
, we adopt the well - established measure of entanglement concurrence as the measure of two - qubit entanglement .
consider a general two - qubit state , the density matrix is @xmath0 .
then its time - reversed matrix is defined as @xmath1 the concurrence of @xmath0 is given by @xcite @xmath2 where @xmath3s are the eigenvalues of @xmath4 in decreasing order .
the corresponding entanglement of formation can be evaluated as @xmath5 where @xmath6 is the shannon s entropy function . however , it is not a very simple task to calculate the concurrence of a two - qubit system in an analytic way .
here we adopt a new method of calculating the entanglement of formation and thus the concurrence , which is based on lorentz singular - value decomposition @xcite . for an arbitrary @xmath7 state @xmath0 ,
there exists a @xmath8 matrix with elements @xmath9 . in the real @xmath10-picture
, the density matrix @xmath0 can be written as @xmath11 where @xmath12 are the pauli matrices . _ lemma 1 :
_ the @xmath8 matrix @xmath10 can be decomposed as @xmath13 , with @xmath14 , @xmath15 finite proper orthochronous lorentz transformations given by @xmath16 , @xmath17 , where @xmath18{cccc}% 1 & \cdot & \cdot & 1\\ \cdot & 1 & 1 & \cdot\\ \cdot & i & -i & \cdot\\ 1 & \cdot & \cdot & -1 \end{array } \right ) $ ] .
the normal form @xmath19 is either of real diagonal form @xmath20 $ ] with @xmath21 , or of the form @xmath22{cccc}% a & \cdot & \cdot & b\\ \cdot & d & \cdot & \cdot\\ \cdot & \cdot & d & \cdot\\ c & \cdot & \cdot & b+c - a \end{array } \right)\ ] ] with @xmath23 , @xmath24 , @xmath25 , @xmath26 real . and
the lorentz singular values of the second normal form are given by @xmath27=[\sqrt{(a - b)(a - c ) } ,
\sqrt{(a - b)(a - c ) } , d , -d]$ ] .
_ lemma 2 : _ given a state @xmath0 and the corresponding matrix @xmath10 , the concurrence of @xmath0 is @xmath28 depending on the lorentz singular values of @xmath10 . and @xmath29 , where @xmath14 , @xmath15 are proper orthochronous lorentz transformations . based on the above two useful lemmas
, we can see that the concurrence of a given density matrix @xmath0 is only determined by the lorentz singular values of the corresponding @xmath10-matrix , which are the only invariants of a state under determinant 1 slocc operations @xcite . in the following ,
we adopt this method for calculating the residual entanglement of an initial entangled state in the decoherence environment .
it is shown that the influence of the environment on the pairwise entanglement is reflected by changing the lorentz singular values . before proceeding to the details ,
we first introduce the decoherence model generated by pauli operators . for a general decoherence model
, it can be denoted as a completely positive map with an operator - sum representation .
the effect of the general pauli channels on a qubit @xmath0 is described as follows @xmath30 where @xmath31 , @xmath32 , and @xmath33 are pauli operators .
this decoherence model includes some representative noise channels . when @xmath34 it is just the depolarizing channel , which describes the decoherence process related to the couplings of quantum system to the thermal reservoir in the large temperature limit . and
when @xmath35 the noise model is the dephasing channel , without energy exchange between the system and the environment , and only lose phase information
. these kinds of decoherence models are common in several physical systems . in the rest of this paper
, we will investigate how the decoherence will influence the pairwise entanglement in details .
we fist assume that each qubit is independently coupled to the environment .
the environment is characterized by the noisy channels generated by pauli operators as shown in eq ( 6 ) .
the initial state @xmath0 , associated with this state @xmath13 , is an entangled two - qubit state .
then after some time , @xmath0 will be transformed to another state @xmath36 with much less entanglement , due to the action of the noisy channels , that is @xmath37 . according to eq ( 6 ) ,
it can be obtained easily that @xmath38 where @xmath39 and @xmath40 .
now we can transformed this map into the @xmath10-picture .
denote the @xmath10-matrix associated with the state @xmath36 as @xmath41 , then @xmath42 where @xmath43 and @xmath44 are lorentz transformations given by @xmath45 and @xmath46 . for simplification
, we can introduce @xmath47 and @xmath48 .
therefore , the state evolution under decoherence is simply characterized by @xmath49 in the real @xmath10-picture . for the decoherence model we discussed here , @xmath50 , @xmath51 . after simple calculation
, it can be seen that @xmath52}$ ] , where @xmath53 , @xmath54 and @xmath55 .
it is obvious that the action of the noisy channels on the entanglement can be viewed as shrinking the lorentz singular values by the above three coefficients .
if the initial entangled states are set as pure states , then according to the schmidt decomposition theorem , an arbitrary two - qubit pure state @xmath56 can be expressed as @xmath57 , where here @xmath58 and @xmath59 are non - negative real numbers satisfying @xmath60 .
that is there always exist local unitary operations @xmath61 and @xmath62 , which satisfy @xmath63 , where @xmath64 . here
@xmath65 , @xmath66 are the @xmath67 , @xmath68 eigenstates of the pauli @xmath69 matrix . in the @xmath10-picture
, @xmath70 corresponds to the matrix : @xmath71{cccc}% 1 & \cdot & \cdot & \lambda_{2}^{2}-\lambda_{1}^{2}\\ \cdot & 2\lambda_{1}\lambda_{2 } & \cdot & \cdot\\ \cdot & \cdot & 2\lambda_{1}\lambda_{2 } & \cdot\\ \lambda_{1}^{2}-\lambda_{2}^{2 } & \cdot & \cdot & -1 \end{array } \right)\ ] ] and the local unitary operations on @xmath72 correspond to left and right multiplication of @xmath73 with orthogonal matrices , therefore in the @xmath10-picture , an arbitrary pure state @xmath74 corresponds to the matrix : @xmath75 where @xmath76{cccc}% 1 & \cdot & & \\ \cdot & o_{1 } & & \\ & & & \end{array } ) = t(u\otimes u^{\ast})t^{\dagger}$ ] and @xmath77{cccc}% 1 & \cdot & & \\ \cdot & o_{2 } & & \\ & & & \end{array } ) = t(v\otimes v^{\ast})t^{\dagger}$ ] , with @xmath78 and @xmath79 are real @xmath80 orthogonal matrices with determinant @xmath81 . in the following
we will investigate how the pairwise entanglement changes under decoherence in the real @xmath10-picture and try to find what kind of entangled states , with the same initial entanglement , can maintain entanglement best .
we starting by considering the initial entangled state is in the schmidt decomposition form @xmath72 .
as discussed above , due to the coupling between the system and environment , @xmath82 is transformed into another mixed states @xmath83 . in the @xmath10-picture
, this action can be expressed as @xmath84 , that is : @xmath85{cccc}% 1 & \cdot & \cdot & ( \lambda_{2}^{2}-\lambda_{1}^{2})q_{3}\\ \cdot & 2\lambda_{1}\lambda_{2}q_{1}^{2 } & \cdot & \cdot\\ \cdot & \cdot & 2\lambda_{1}\lambda_{2}q_{2}^{2 } & \cdot\\ ( \lambda_{1}^{2}-\lambda_{2}^{2})q_{3 } & \cdot & \cdot & -q_{3}^{2}% \end{array } \right)\ ] ] therefore the concurrence of @xmath86 can be obtained easily according to lemma 2 .
@xmath87 where @xmath88 is the initial entanglement .
actually , this result can also be obtained by the conventional way of calculating the concurrence .
note that @xmath89{cccc}% a & \cdot & \cdot & b\\ \cdot & d & c & \cdot\\ \cdot & c & e & \cdot\\ b & \cdot & \cdot & a \end{array } \right)\ ] ] where @xmath90 , @xmath91 , @xmath92 , @xmath93 and @xmath94 . for a state having a density matrix of the above form
, the concurrence is given by @xmath95 , where @xmath96 and @xmath97 .
we note @xmath98 is always larger than @xmath99 , thus @xmath100 which agrees with the above result . for an arbitrary pure state @xmath101
, the corresponding @xmath10 matrix is shown in eq .
( 10 ) . in the decoherence environment , @xmath101 is transformed to @xmath36 . in the real @xmath10-picture
, the evolution of the state is described by @xmath102 we then multiply @xmath103 and @xmath104 to @xmath10 from left and right respectively , and get another @xmath10-matrix @xmath105 , where @xmath106 and @xmath107 .
this corresponds to local unitary operations on the state @xmath36 in the @xmath0-picture , thus the concurrence of @xmath108 is identical to the one of @xmath41 .
as we have discussed above , the concurrence of a given @xmath10 matrix is only determined by the lorentz singular values , and the action of the decoherence on the entanglement can be viewed as shrinking the lorentz singular values . if we look on @xmath109 as the components of a vector , then the influence of decoherence is just shrinking this vector according to the coefficients @xmath110 , @xmath111 and @xmath112 .
this can be reflected by the residual entanglement shown in eq .
the action of the noisy channels is characterized by the three shrinking coefficients .
in addition , we note that @xmath113 .
therefore the effects of @xmath114 and @xmath104 on @xmath14 and @xmath15 respectively is changing the shrinking directions . then if we order the coefficients @xmath115 such that @xmath116 .
based on the above analysis , it can easily be seen that the maximum residual entanglement is @xmath117 where @xmath118 is the initial entanglement .
the above maximum residual entanglement can be achieved by appropriate local unitary operations .
the corresponding initial pure states are @xmath119\}\ ] ] this set of pure states present those special states that are the most entanglement - stable .
the minimum residual entanglement and the corresponding initial pure states can be also derived in a similar way .
given a specific decoherence model , the relation between @xmath110 , @xmath111 , and @xmath112 is known , then the pure states in @xmath120 can be written explicitly .
for example , in the dephasing channels model , the coefficients are @xmath121 and @xmath122 .
therefore both @xmath123 and @xmath124 belong to the set @xmath120 .
the other states contained in @xmath120 can also be obtained easily according eq .
( 16 ) .
although all the pure states with the same entanglement is equivalent under local unitary operations , it can be seen from the above results that under decoherence the behavior of different pure states are not all the same .
this reflect the properties of the decoherence model and its influence on the entanglement .
our results suggest that if the decoherence model is fixed , there exists a special class of pure states , with the same initial entanglement , which are more favorable for maintaining entanglement .
this gives some useful hints for maintaining and distributing entanglement .
for example , if we want to distribute an entangled pure state between two separate parties through noisy channels , it will be helpful to apply some local unitary operations beforehand to transform the entangled pure state into the form of the states in the above set @xmath120 .
for an initial general mixed entangled state @xmath0 with the initial entanglement @xmath118 , to drive an analytic evolution equation for its entanglement maybe intractable .
however , we derive a upper bound of the residual entanglement in the following .
this upper bound is the corresponding residual entanglement @xmath125 for the states in the set @xmath120 with the same initial entanglement @xmath118 .
it has be shown in @xcite , there exists an optimal decomposition @xmath126 , such that @xmath127 for each @xmath128 , with @xmath129 . then under decoherence @xmath0 is transformed into another state @xmath130 .
according to the convexity of the concurrence , we know that @xmath131 .
since @xmath132 , it is obvious that @xmath133 .
this suggests that the state in the set @xmath120 is the most entanglement - stable of all the states with the same initial entanglement , no matter pure or general mixed states .
it is well known that , an arbitrary pure state can be transformed into the states , with the same entanglement , of the set @xmath120 by local unitary operations .
thus in the sense discussed in this paper , the pairwise entanglement in pure states are more favorable for maintaining entanglement compared to the generic mixed states .
this conclusion is valid for any decoherence model which can be verified from the above discussions .
up till now , we have investigate the entanglement dynamics of two - qubit states assuming that each qubit is independently coupled to the environment .
in the following , we consider the situation that only one qubit is under decoherence and the initial state is a pure entangled state .
we are interested in whether the lifetime of entanglement also depends on the initial entanglement in this situation .
as discussed above , the entanglement dynamics for pure states with the same initial entanglement are also dependent on their forms .
to be comparable , we set the initial pure state in the schmidt decomposition form @xmath134 . if only the first qubit is under decoherence , then the two - qubit system becomes @xmath135 .
in the @xmath10-picture , the matrix corresponding to @xmath36 is @xmath136 , that is @xmath137{cccc}% 1 & \cdot & \cdot & \lambda_{2}^{2}-\lambda_{1}^{2}\\ \cdot & 2\lambda_{1}\lambda_{2}q_{1 } & \cdot & \cdot\\ \cdot & \cdot & 2\lambda_{1}\lambda_{2}q_{2 } & \cdot\\ ( \lambda_{1}^{2}-\lambda_{2}^{2})q_{3 } & \cdot & \cdot & -q_{3}% \end{array } \right)\ ] ] therefore the concurrence of @xmath36 is @xmath138 this result shows that the lifetime of the entanglement is independent on the initial entanglement . no matter how much the initial two - qubit state is entangled , it becomes separable in a constant time .
this somewhat interesting phenomena reflect that entanglement is some kind of nonlocal property . in the above
, we discuss the dynamics of entanglement under the noisy channels generated by pauli matrices . when the parameters satisfy @xmath139 and @xmath140 , the decoherence model in eq .
( 6 ) is the depolarizing channels .
the depolarizing channels describe the system - environment coupling in the large temperature limit @xmath141 .
it can be realized by random von neumann measurements .
again we assume that each qubit is independently coupled to the environment .
the shrinking coefficients are @xmath142 .
taking into account of the strength of the system - environment coupling and the interaction time , we can write @xmath143 . thus the corresponding matrices in the @xmath10-picture are @xmath144 $ ] . given an arbitrary initial entangled pure state @xmath101 , the corresponding schmidt decomposition normal form is @xmath123 associated with the @xmath10-matrix @xmath73 . then after time @xmath145 , the residual entanglement is dependent on the @xmath146 .
note that @xmath147 and @xmath148 here
. therefore the concurrence at time @xmath145 is @xmath149 from the above evolution function of the pairwise entanglement , we can find that the residual entanglement at time @xmath145 , only depends on its initial entanglement @xmath118 .
thus for all two - qubit pure states coupled with the same depolarizing environment , the stability of the entanglement is only governed by their initial entanglement , although these pure states could be in different forms .
in other words , all pure states are the most entanglement - stable , need not to apply local unitary operations beforehand .
recalling the above analysis in the @xmath10-picture , the reason for this interesting result is that the shrinking of the lorentz singular values of the associated @xmath10-matrix , which is introduced by the depolarizing channels , is isotropic . besides , it is also obviously that even for two generic mixed state , if there are @xmath150 equivalent then the dynamics of two - qubit entanglement are also equal . furthermore , there are some special mixed entangled states which have the same entanglement dynamics as pure states , that is also the most entanglement - stable .
we will discuss the situation of mixed states in the next section . in fig .
1 we present a visual example by plotting the dynamics of residual entanglement for pure states and some other generic mixed states with the same initial entanglement .
residual entanglement @xmath151 as function of time @xmath145 .
pure state ( solid curve ) ; some other generic mixed states ( dashed and dotted curve ) . ]
_ definition : _ given a decoherence model characterized by a completely positive map @xmath152 , decoherence path states ( dps ) are those transient states @xmath0 obtained from the pure states , that is @xmath153 @xmath154 which satisfies @xmath155 .
we consider the depolarizing channels , and the initial entangled state @xmath0 is a decoherence path state with the initial entanglement @xmath156 .
thus there exists @xmath154 with the entanglement @xmath157 and some time @xmath158 that satisfy @xmath159 and @xmath160 .
then after some time @xmath145 , the decoherence path state @xmath0 evolves to another state @xmath161 with the entanglement @xmath162 .
this can be simplified to @xmath163 .
it is obvious that this is the same with the dynamics of entanglement for pure states with the same initial entanglement @xmath118 , as shown in eq .
in fact , the familiar werner states belong to the decoherence path states .
therefore in the depolarizing channels not only all pure states but also some special mixed states have the same entanglement dynamics .
in other words , the pairwise entanglement of the decoherence path states are also the most stable .
we plot the dynamics of entanglement in decoherence path states with different initial entanglement , as depicted in fig .
2 . from fig.2 ( a )
it can be seen that the entanglement in the two - qubit system decreases with time due to its interaction with the decoherence environment .
there exists some time @xmath164 when its entanglement @xmath165 for @xmath166 .
thus @xmath164 is the critical time when the system entanglement disappears , i.e. the two - qubit system becomes separable .
we can easily calculate the critical time : @xmath167}{2\kappa}%\ ] ] the relation between @xmath164 and the initial entanglement @xmath118 is depicted in fig .
certainly , the lifetime of entanglement is longer if the initial entanglement is larger . for the singlet state ,
the lifetime of entanglement is @xmath168 as function of the initial entanglement @xmath118 and time @xmath145 .
( b ) critical time @xmath169 as function of the initial entanglement @xmath118 . ]
however , we can not exclude the possibility that there exist non - dps states with the same entanglement dynamics as dps states . in the following ,
we can see a concrete example .
but as a special class of mixed states , the decoherence path states of a given decoherence model are expected to exhibit some other interesting properties . in this section we will investigate the stability of the nearest neighbor entanglement of the ground states of spin models .
the hamiltonian of the translationally invariant xxz spin chain with periodic boundary condition is @xmath170\ ] ] when @xmath171 the above hamiltonian represents the heisenberg antiferromagnetic model .
the ground state @xmath172@xcite is translationally invariant and the total z component of spin is zero .
thus the reduced density matrix of the @xmath173 and @xmath174 site is : @xmath175{cccc}% u & 0 & 0 & 0\\ 0 & x & z & 0\\ 0 & z^{\ast } & y & 0\\ 0 & 0 & 0 & v \end{array } \right)\ ] ] when @xmath176 and @xmath177 it corresponds to the ground state of the heisenberg antiferromagnetic model .
the density matrices of the maximally entangled mixed states @xcite are also in the above form . in the general pauli channels of ( 6 ) , the state @xmath178 evolve to @xmath179{cccc}% a & 0 & 0 & e\\ 0 & b & f & 0\\ 0 & f^{\ast } & c & 0\\ e^{\ast } & 0 & 0 & d \end{array } \right)\ ] ] where @xmath180 , @xmath181 , @xmath182 , @xmath183 , @xmath184 and @xmath185 with @xmath186 , @xmath187 , @xmath188 and @xmath189 .
we note that @xmath190 because @xmath178 is positive .
thus @xmath191 .
therefore the concurrence of @xmath192 is given by @xmath193 .
for the special heisenberg antiferromagnetic model and the depolarizing channels , the residual entanglement is @xmath194 , where @xmath195 .
note that @xmath196 , @xmath197 and @xmath198 .
thus @xmath199 , which is the same as the pure states .
therefore in the depolarizing channels , the nearest neighbor entanglement of the ground state of the heisenberg antiferromagnetic model is the most stable , though it is not maximized @xcite .
this result study the entanglement of the ground states of spin models from a new point of view .
in addition , we can verify that @xmath178 does not belong to the decoherence path states .
this suggest that in this certain decoherence model , several mixed states other than decoherence path states are also the most entanglement - stable .
in conclusion , we have investigate the entanglement dynamics of a two - qubit system under a general decoherence model , that is pauli channels . given a decoherence model , we find the special class of pure states that are the most entanglement - stable and present the analytic entanglement dynamics of these states .
since any pure states with the same entanglement are @xmath150 equivalent , we show that pure states are more favorable for maintaining entanglement than general mixed states . therefore in the situation of maintaining or distributing entanglement , it is helpful to using pure entangled states and to apply appropriate local unitary operations beforehand to transform the entangled states to the most entanglement - stable form .
particularly , we investigate a certain decoherence model i.e. the depolarizing channels . in this case , a special class of mixed states that is decoherence path states are as entanglement - stable as pure states .
the familiar werner states are indeed dps states . in addition , we investigate the entanglement dynamics of some specific class of mixed states , such as the ground states of xxz , heisenberg antiferromagnetic spin model and the maximally entangled mixed states . it is shown that in the depolarizing channels , the nearest neighbor entanglement of the ground state of heisenberg antiferromagnetic spin model is coincidentally the most entanglement - stable .
another interesting result is that if only one qubit is coupled with the decoherence environment then the life time of entanglement is independent on the initial entanglement .
this just reflects that entanglement is some kind of nonlocal property . for the decoherence model considered in this paper ,
we find the most entanglement - stable form of entangled states .
the extension of this study to more general decoherence models is very meaningful , which is also related to the important work in @xcite .
furthermore , we also introduce a special class mixed states i.e. decoherence path states and find that they are also the most entanglement - stable states in the depolarizing channels .
however , they are expected to exhibit more interesting properties in general decoherence models .
the authors thank dr .
xiang - fa zhou for valuable discussions .
this work was funded by national fundamental research program ( 2001cb309300 ) , the innovation funds from chinese academy of sciences , z .- w .
zhou acknowledges funds from national natural science foundation of china ( grant no .
10204020 ) . | consider the dynamics of a two - qubit entangled system in the decoherence environment , we investigate the stability of pairwise entanglement under decoherence . we find that for different decoherence models , there exist some special class of entangled states of which the pairwise entanglement is the most stable .
the lifetime of the entanglement in these states is larger than other states with the same initial entanglement .
in addition , we also investigate the dynamics of pairwise entanglement in the ground state of spin models such as heisenberg and xxy models . |
research on type ia supernovae ( sn ia ) has been guided for many years by the general agreement that the progenitors are carbon - oxygen ( c / o ) white dwarf stars ( wd ) as first predicted by @xcite .
this hypothesis is based on observational and theoretical evidence that thermonuclear burning of a c / o wd can produce light curves and spectra that are very similar to those observed for sn ia . beyond that general statement
, we lack many of the details concerning the composition of the progenitors and the physics of the explosions . in order to increase the effectiveness of sn ia as distance indicators
we must improve our ability to predict the intrinsic brightness of individual events .
observations that provide new information about the chemical structure deepen our understanding of the explosion physics and may contribute constraints to theoretical models . here
we identify , measure , and compare high - velocity ( @xmath0 20,000 km s@xmath1 ) and photospheric - velocity ( pvf ) absorption features from lines of , , , , and in a sequence of optical spectra of sn ia 2009ig .
the spectra of sn 2009ig at phases earlier than @xmath13 d were previously published by @xcite .
the phase of each observation is expressed in days with respect to the time of @xmath4-band maximum light , @xmath4-max ( sep 6.0 ut = jd 2,455,080.5 ; see 2.1 ) .
high - velocity ( hvf ) absorption features from lines are frequently identified in spectra of sn ia obtained a week or more before @xmath4-max , but reliable identifications of hvf from other elements are rare .
the presence of hvf from other lines is inferred from unusual pvf line profiles in very early - time spectra .
hvf components of blended @xmath76355 features are reported by @xcite , @xcite , @xcite , @xcite , @xcite , and @xcite .
evidence for hvf in features is discussed by @xcite , @xcite , @xcite , @xcite , and @xcite .
the two earliest spectra of sn 2009ig ( @xmath14 d , @xmath15 d ) clearly resolve high - velocity @xmath76355 as a separate absorption feature produced in a detached line forming region .
all features identified in the earliest spectra of sn 2009ig are hvf .
the first unambiguous detection of pvf occurs at @xmath10 d , and we measure both hvf and pvf velocities from @xmath10 d until hvf are no longer detectable at about @xmath11 d. all absorption features in spectra obtained after @xmath16 d are pvf .
the early start and dense coverage of our sample allows us to map a complete transition from hvf to pvf in sn 2009ig . the strongest absorption features in early - time spectra of sn ia are the h&k lines and the near - infrared triplet ( ir3 ) .
most of the pre - maximum spectra in our sn 2009ig sample include both of these lines , presenting an opportunity to study the development of hvf and pvf in . @xcite and @xcite also reported simultaneous observations of hvf and pvf from both lines .
hvf from are usually observed as very broad features that extend 15,00025,000 km s@xmath1 beyond typical photospheric velocities .
they are most often detected in spectra obtained earlier than @xmath16 d , but it is not unusual for hvf to persist as late as maximum light .
detections of high - velocity in early spectra of sn ia are reported and discussed by @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and others . @xcite compare and discuss hvf in sn ia observed near @xmath4-max .
@xcite assert that all spectra of sn ia observed more than one week before maximum will exhibit high - velocity .
@xcite use the same pre - maximum data of sn 2009ig as we do to identify and discuss differences in line profiles of h&k , ir3 , and @xmath76355 .
they reveal the two - component nature of early @xmath76355 features and propose that hvf may be ubiquitous in sn ia . here
we focus on the location and composition of hvf line forming regions .
in particular , we examine the characteristics of hvf from lines that are not .
we directly compare the development of @xmath76355 features in the spectra of sn 2009ig to sequences of early - time spectra of five other sn ia .
velocity measurements show that the hvf and pvf line forming regions remain separated by about 8,000 km s@xmath1 in sn 2009ig .
the expansion of sn ia is assumed to be homologous , so that measured velocities are proportional to distances from the center of the sn .
consequently , a line forming region that produces high - velocity absorption features is physically outside of a region that produces lower velocity features .
homologous expansion ensures that they will remain in the same relative positions .
the formation of detached hvf requires a region with a localized enhancement of abundance or density
. the hvf layer must also be farther from the center of the sn than the photospheric layer .
we are a long way from a consensus for how hvf line forming regions come into existence .
a few of the proposed explanations for hvf are uneven density profiles in the ejecta @xcite , a clumpy structure or a thick torus @xcite , and interaction between the sn ejecta and surrounding material @xcite .
evidence for polarization of the hvf in spectra of sn ia was first presented by @xcite and discussed by @xcite .
if the hvf regions are not spherically symmetric , then viewing angles will influence the observed velocities .
in [ obs ] we describe the observations of sn 2009ig and the reduction of the data .
a discussion of the methods used for identification and measurement of absorption features is presented in [ i d ] .
the characteristics of hvf from specific lines are presented in
a discussion of pvf and their measurements appears in [ ps ] .
analysis of the composition and location of hvf line forming regions is found in [ lfr ] . in [ comp ] we compare the results for sn 2009ig to early - time spectra of other sn ia .
possible clues to the origins of hvf regions in sn ia are discussed in [ sources ] .
section [ conc ] presents a summary and conclusions .
sn 2009ig in ngc 1015 ( redshift @xmath17 ; ned ) was discovered @xcite by the lick observatory supernova search ( loss ; * ? ? ?
* ) with the 0.76 m katzman automatic imaging telescope ( kait ) at a magnitude of 17.5 on 2009 aug .
20.5 ( ut dates are used throughout this paper ) .
the last known nondetection was also by loss / kait on aug .
16 to a limiting magnitude of 18.7 . on aug .
21.1 the first spectrum was obtained with the asiago 1.82 m telescope by @xcite .
it revealed characteristics of a sn ia very soon after the explosion with a high expansion velocity for @xmath76355 ( 24,500 km s@xmath1 ) and no detection of @xmath75971 or @xmath75641 . because sn 2009ig was nearby and
the detection was very early , we were presented with a unique opportunity to conduct extensive observations of a young sn ia .
photometric data were obtained in the @xmath18 bands with the 0.8 m tsinghua - naoc telescope ( tnt ) , located at xinglong station of the national astronomical observatory of china , 180 km from beijing .
details of the tnt , detectors , and observing conditions are described by @xcite .
all photometric data were reduced using standard iraf packages .
the light curves were fit to mlcs2k2 templates @xcite .
the light curves and the best - fit templates are displayed in figure [ lc ] .
we adopted @xmath19 mag for the milky way extinction @xcite and a time of maximum @xmath4-band brightness of @xmath20 mjd ( 2009 sep .
mlcs2k2 finds @xmath21 for the best - fit light - curve parameter , @xmath22 mag for the distance modulus ( assuming h@xmath23 km s@xmath1 mpc@xmath1 ) , and @xmath24 mag for the host - galaxy extinction . from the mlcs2k2 results we find @xmath25 and @xmath26 mag on 2009 sep .
the @xmath27 parameter corresponds to a decline - rate value of @xmath28 mag .
these values suggest that sn 2009ig is slightly more luminous and has a slower decline rate than a `` normal '' sn ia , but the @xmath29 and @xmath30 values are not exceptional .
these parameters are consistent with those of @xcite , who used kait and _ swift _ photometry of sn 2009ig to derive values of @xmath31 mag and @xmath24 mag .
small differences in the photometric parameters have no effect on the spectroscopic results presented in this paper .
light curves of sn 2009ig obtained at the tsinghua - naoc telescope ( tnt ) . the black solid line is the mlcs2k2 fit and the dotted lines are the @xmath32 boundaries .
template fitting gives @xmath33 mag on 2009 sep .
6.0 with @xmath6 mag .
[ lc],scaledwidth=52.0% ] our complete sample includes 32 optical spectra of sn 2009ig obtained from @xmath14 d to + 12.5 d with respect to the time of @xmath4-max .
spectra obtained before sep . 9 ( + 3 d ) were previously published by @xcite . in most cases where the phase is discussed in the text
, we round the value to whole days for simplicity .
the _ swift _ @xmath34-grism spectra are included in this list of optical " spectra since we only display and analyze the portion of these spectra in the range 32004400 .
figure [ fullstack ] displays spectra obtained on 25 of the 28 nights between @xmath2 d and + 13 d. the maximum wavelength range of 32009000 reaches both the h&k lines ( @xmath73945 ) and the ir3 ( @xmath78579 ) .
full wavelength coverage is achieved on every night from @xmath2 d to @xmath16 d. for some of the dates , observations of sn 2009ig that cover slightly different wavelengths occurred at two different telescopes within a few hours of each other .
we combine these data to create a single continuous spectrum by trimming the data from one source at a specific wavelength and continuing from that wavelength with data from the other source without overlap .
the transition points are selected to optimize the signal - to - noise ratio for the combined spectrum .
spectral features in sn ia change rapidly at early times , but a daily cadence is sufficient to trace the development of individual features .
the use of a single spectrum per day makes it easier to measure and display the data .
a complete list of the observational details can be found in table [ obstbl ] , including the wavelengths at which the spectra were joined . for the ground - based spectroscopy
, the slit was generally oriented along the parallactic angle to minimize differential slit losses caused by atmospheric dispersion @xcite .
sn 2009ig was observed using the shane 3 m telescope at lick observatory with the kast spectrograph @xcite on aug .
22 , 24 , 25 , 27 , and 28 .
an additional spectrum was obtained on aug .
22 at the 10 m keck i telescope with the low resolution imaging spectrometer ( lris ; * ? ? ?
* ) equipped with an atmospheric dispersion corrector .
the lick / kast spectra used a 600/4310 grism on the blue side and a 300/7500 grating on the red side with a @xmath35 wide slit .
the keck / lris spectrum was obtained with a 400/3400 grism on the blue side and a 600/7500 grating on the red side using a @xmath36 wide slit .
the 9.2 m hobby - eberly telescope ( het ; * ? ? ?
* ) with the marcario low - resolution spectrograph ( lrs ; * ? ? ?
* ) was used to observe sn 2009ig on aug .
23 ( @xmath9 d ) .
additional het observations were made on aug .
29 , 30 , and 31 , as well as on sep . 2 , 3 , 4 , 5 , 9 , 10 , and 14 .
the het / lrs spectra have an effective wavelength range of 44009200 .
reduction of the het / lrs data was also conducted with standard iraf procedures .
low - dispersion spectra of sn 2009ig were obtained on aug .
26 , 27 , 28 , 29 , and 30 using the 6.5 m mmt with the blue channel spectrograph @xcite , covering the wavelength range 32008200 .
ccd processing and spectrum extraction for these observations were completed with iraf and the data were extracted with the optimal algorithm of @xcite . after the wavelength calibration was derived from low - order polynomial fits to calibration - lamp spectra , additional small adjustments were applied by cross - correlating a template sky spectrum to the night - sky lines that were extracted along with the sn .
idl routines were used to flux calibrate the data and to remove telluric lines @xcite .
the _ swift _ satellite observed sn 2009ig with the uv grism mode of the ultraviolet / optical telescope ( uvot ; * ? ? ?
* ) onboard the spacecraft @xcite .
_ swift _ data were obtained on aug .
23 , 25 , and 27 , and on sep . 1 , 3 , and 6 .
we do not use all of the _ swift _ data for our analysis , but the complete sequence of _ swift _ spectra and photometric measurements was published by @xcite .
_ swift _ data were extracted and reduced using the `` uvotimgrism '' package of ftools .
the effective wavelength coverage is 17004900 , but in this analysis we use only a small portion of the _ swift _ spectra covering the wavelength region 32004400 that contains the h&k feature .
4560 ( left ) and @xmath7 6355 ( right ) in sn 2009ig are displayed in velocity space ( blueshifted velocities are expressed with positive values ) .
rest wavelengths are indicated on the top abscissa .
the spectra were obtained between @xmath2 d and @xmath4-max , and the phase of each each spectrum is marked near the red end .
the line profiles for @xmath76355 in the earliest spectra clearly require separate hvf and pvf components .
the phases from @xmath10 d to @xmath37 d reveal contributions from both hvf and pvf .
narrow d features are visible in the @xmath7 6355 profiles . hvf @xmath74560 is blended with pvf @xmath74481 beginning about @xmath12 d. the velocities measured for each feature are indicated by a blue hashmark for hvf and a red hashmark for pvf .
the dotted line at 21,500 km s@xmath1 and the dashed line at 13,000 km s@xmath1 are the same in all figures .
[ pvsi],scaledwidth=50.0% ] 5453 ( left ) and @xmath75641 ( right ) are marked on a sequence of spectra of sn 2009ig obtained between @xmath2 d and + 0 d. rest wavelengths are indicated on the top abscissa .
the phase of each spectrum is marked near the red end .
hvf are strongest before @xmath12 d and pvf are stronger after @xmath37 d. the measured velocities are indicated by a blue mark for hvf and a red mark for pvf .
the pvf in @xmath75453 are compromised by blending after @xmath16 d and we do not measure the velocities ( see [ ps ] ) . the dotted line at 21,500 km s@xmath1 and the dashed line at 13,000 km s@xmath1 are the same in all figures .
[ pvs],scaledwidth=50.0% ] 3945 ( left ) and ir3 @xmath78579 ( right ) in sn 2009ig are displayed in velocity space ( blueshifted velocities are expressed with positive values ) .
rest wavelengths are indicated on the top abscissa .
the spectra were obtained between @xmath2 d and @xmath38 d and the phase of each each spectrum is marked near the red end .
hvf velocities for are higher than for other lines in the early spectra ( compare offsets from the dotted line ) , but they eventually decline to the mean hvf velocity for all lines ( dotted line ) .
hvf are detected for a few days longer than hvf for other lines , and the pvf appear later . comparing the features by phase
suggests that the hvf component of h&k @xmath73945 is exaggerated by a strengthening contribution from pvf @xmath73858 beginning about @xmath12 d. the measured velocities are indicated by a blue mark for hvf and a red mark for pvf .
the dotted line at 21,500 km s@xmath1 and the dashed line at 13,000 km s@xmath1 are the same in all figures .
[ pvca],scaledwidth=50.0% ] 5018 ( left ) and @xmath75169 ( right ) are marked on a sequence of spectra of sn 2009ig obtained between @xmath2 d and + 0 d. rest wavelengths are indicated on the top abscissa .
the phase of each spectrum is marked at the red end .
the dashed line in the velocity space of @xmath75018 and the dotted line in velocity space for hvf @xmath75169 are in the same physical location .
the feature at that position is produced by hvf only in the first two spectra , by a blend of pvf and hvf from about @xmath39 d to @xmath11 d , and by pvf only after @xmath11 d. comparing offsets from the dotted lines shows that hvf of are deeper , wider , and at higher velocities than hvf of ( fig .
[ pvsi ] ) and ( fig.[pvs ] ) .
the measured velocities are indicated by a blue mark for hvf and a red mark for pvf . the dotted line at 21,500 km s@xmath1 and the dashed line at 13,000 km s@xmath1 are the same in all figures .
[ pvfe],scaledwidth=50.0% ]
eight lines have been identified that produce both hvf and pvf absorptions in spectra of sn 2009ig .
the features associated with these lines are displayed by phase and velocity space in four figures : @xmath74560 and @xmath7 6355 ( figure [ pvsi ] ) , @xmath7@xmath7 5453 , 5641 ( figure [ pvs ] ) , h&k @xmath73945 and ir3 @xmath78579 ( figure [ pvca ] ) , and @xmath7@xmath7 5018 , 5169 ( figure [ pvfe ] ) .
the velocity measurements at each phase are indicated by a blue hashmark for hvf and a red hashmark for pvf . to facilitate comparison of the components between features ,
all of the figures also have a dotted line plotted at 21,500 km s@xmath1 and a dashed line at 13,000 km s@xmath1 .
these velocities were selected to be representative of the hvf at @xmath12 d and the pvf near @xmath4-max . a detailed discussion of individual features is found in [ hv ] ( hvf ) and [ ps ] ( pvf ) . the hvf and pvf of @xmath76355 are easily identified and relatively free from blending .
these characteristics make them useful templates that guide efforts to identify absorption features associated with other lines .
we compare the velocities , the phases of initial and final detections for hvf and pvf , the relative strengths of the hvf and pvf components at a given phase , and the velocity separation between the hvf and pvf .
the presence of simultaneous hvf and pvf from several lines increases the number of blended features in the early - time spectra of sn 2009ig .
velocities and line profiles of individual features are measured directly from the flux - calibrated spectra rather than a flattened continuum because the shapes and slopes of the local continua vary widely in different regions of each spectrum . in the velocity figures ,
the flux is scaled to make it easier to see details of the absorption features .
a representative flux ratio is maintained between the two panels in each figure in order to compare the relative strengths of features from similar lines , but it is not possible to maintain a consistent scale between the figures . the velocity range for the and plots 48,000 to @xmath408,000 km s@xmath1 . for and
velocity plots the range is 36,000 to @xmath408,000 km s@xmath1 . throughout this paper
we reference blueshifted velocities using positive values
. 6355 feature in the mmt spectrum of sn 2009ig obtained on aug .
27 ( @xmath41 d ) .
rest wavelengths are indicated on the top abscissa .
the pvf component is dominant at this phase , but there is clear evidence for a contribution from the hvf component .
the solid blue line shows that a gaussian profile with a minimum at 20,700 km s@xmath1 is a good fit to the data .
the red dotted lines form gaussians with minima that have been shifted to 19,700 and 21,700 km s@xmath1 .
the red lines do not fit the data as effectively as the blue line .
the photospheric component is fit with a gaussian ( solid green line ) having a minimum at 14,400 km s@xmath1 .
[ gfit],scaledwidth=52.0% ] at many phases one of the components may lack a distinct minimum but still makes an obvious contribution to the overall line profile . for example
, the right panel of figure [ pvsi ] shows that the pvf component from @xmath76355 at @xmath10 d produces only a small indentation on the red side of the larger absorption feature dominated by hv .
as time passes the strength of the hvf signal diminishes and the pvf component becomes stronger . at @xmath41
d the pvf absorption is dominant and the hvf component forms a small but distinct indentation on the blue side of the pvf absorption feature .
we use this spectrum to illustrate a method for measuring velocities from absorption components that form incomplete features .
figure [ gfit ] shows the @xmath76355 feature from the august 27 spectrum ( @xmath41 d ) .
the photospheric component is simply fit by a gaussian ( solid green line ) having a minimum at 14,400 km s@xmath1 .
the data in the region of the indentation created by the hvf component are fit with a gaussian profile having a minimum at 20,700 km s@xmath1 ( solid blue line ) .
the dotted red lines show alternative gaussians with minima that have been shifted by @xmath42 1,000 km s@xmath1 from the minimum of the blue line .
neither of the line profiles plotted in red fit the data as well as the blue line .
the optimal position for the gaussian is determined by changing the minimum in steps of 100 km s@xmath1 and inspecting the results by eye .
it is nearly always easy to determine when the gaussian profile deviates from the actual data by more than a few hundred km s@xmath1 .
uncertainties due to this measurement technique are 1001,000 km s@xmath1 depending on the quality of the data and how well the minima are defined .
the measurements in tables [ hvtbl ] ( hvf ) and [ pstbl ] ( pvf ) were obtained by this method .
figure [ pvsi ] illustrates that for @xmath76355 , uncertainties will be higher for hvf measured after @xmath41 d and for pvf measured before @xmath43
d. the synow code @xcite is another tool for line identification .
synow creates model spectra that can compare relative line strengths and positions by wavelength .
synow makes it possible to explore the contributions of individual ions to a complete spectrum , and it is particularly helpful for distinguishing between the relative contributions of different lines in a blended feature .
in most sn ia , produces the only detectable hvf . in sn 2009ig ,
eight lines from five ions produce distinct hvf in the early spectra : @xmath74560 and @xmath76355 ( figure [ pvsi ] ) , @xmath7@xmath75453 , 5641 ( figure [ pvs ] ) , and @xmath7@xmath75018 , 5169 ( figure [ pvfe ] ) .
there are also strong suggestions of hvf from several other lines that are associated with blended absorption features for which we are unable to confirm the contributions of individual lines .
hvf are clearly identified from all lines in the @xmath2 d spectrum of sn 2009ig , and hvf remain strong through @xmath10 d. the strength of the hvf absorptions weaken with time , and in most lines hvf are no longer detectable after @xmath11 d. beginning at about @xmath12 d , many of the hvf are affected by blends with pvf that are increasing in strength . the highest velocities for features from each line are measured in the earliest spectrum at @xmath2 d. for , @xmath44 32,600 km s@xmath1 , followed by at 26,600 km s@xmath1 , at 25,700 km s@xmath1 , at 23,900 km s@xmath1 , and at 22,800 km s@xmath1 .
this range of velocities for the hvf is apparently caused by differences in the effective opacities of the lines at that phase .
by @xmath10 d , the hvf velocities for all lines except are grouped closely together with a mean of 22,600 km s@xmath1 .
hvf velocities for all lines converge near 21,500 km s@xmath1 by @xmath11 d. the hvf of @xmath76355 provides a useful guide for the behavior of hvf in general because it is essentially unblended .
the hvf component dominates the profile from @xmath2 d to @xmath10 d at velocities that are about 1,000 km s@xmath1 lower than and about 2,000 km s@xmath1 lower than .
both hvf and pvf contribute to the @xmath76355 line profile between @xmath10 d and @xmath45 d , and hvf are no longer detected from after @xmath11 d. narrow d features are visible in the spectra at about 23,500 km s@xmath1 , which is zero velocity for na d , and at 24,200 km s@xmath1 , which is the velocity of the host galaxy , ngc 1015 .
has many other lines that often produce detectable features in spectra of sn ia .
possible evidence exists for hvf components from @xmath7@xmath73858 , 4130 , 5051 , and 5972 , but the features are compromised by blending .
the data do not include sufficiently short wavelengths to cover the hvf of @xmath74560 at @xmath9 d , but this feature is strong at @xmath2 d and @xmath10 d. contributions from pvf @xmath74481 and pvf @xmath74407 begin to affect hvf @xmath74560 by @xmath43 d. we measure hvf until @xmath46 d , when the absorption minimum shifts abruptly to the red , indicating that the hvf contribution has ended .
synow models suggest that the pvf component is stronger than pvf from @xmath12 d until about @xmath47 d. absorption features in the spectra can possibly be associated with several other doubly ionized lines .
possible hvf are found for @xmath75740 , @xmath74264 , and @xmath7@xmath74407 , 5129 , but as for many other tentative associations , blending prevents definitive identifications .
figure [ pvs ] shows that hvf for both @xmath75453 and @xmath75641 are stronger at @xmath10 d than they are at @xmath2 d. hvf velocities for are between the velocities of hvf and hvf .
the figure also shows that the @xmath75641 hvf and the @xmath75453 pvf form a blended feature .
we identify the red component of this blend as hvf @xmath75641 .
the hvf is strongest from @xmath2 d to @xmath10 d , contributions from both hvf and pvf are evident from @xmath12 d to @xmath16 d , and weak absorptions remain near the locations of both hvf as late as @xmath47 d. we stop measuring hvf @xmath75453 after @xmath11 d and hvf @xmath75641 after @xmath16 d. there is possible evidence in the spectra for hvf from @xmath75032 and @xmath75429 .
hvf are not detected from carbon , oxygen , or magnesium , so there is no evidence for unburned fuel or carbon burning products in the hvf region of sn 2009ig .
if the @xmath74481 line produced a hvf at @xmath2 d , it would have a velocity of about 24,000 km s@xmath1 , and it would appear at about 29,000 km s@xmath1 in the velocity space of @xmath74560 ( left panel of figure [ pvsi ] ) . at @xmath12 d , the hvf velocity would be about 21,500 km s@xmath1 , corresponding to 26,500 km s@xmath1 in the figure . the spectra are steeply sloped toward the red in this region , which makes it more difficult to discern small absorptions , but careful inspection reveals no evidence of absorption features that can be associated with hvf from .
the velocity plots and tables show that hvf of and are wider and deeper , with absorption minima at higher velocities than the hvf of other lines .
@xcite observed that hvf of both lines are stronger and have higher velocities than hvf of any other lines in the spectra of sn 2009ig .
the @xmath75169 hvf produces a very strong absorption from @xmath2 d to @xmath43 d that is nearly as broad and deep as the hvf .
the feature is subsequently affected by pvf @xmath75018 , pvf @xmath75051 , and pvf @xmath75032 beginning about @xmath12 d. we stop measuring hvf @xmath75169 after the pvf blend becomes dominant at @xmath16 d. hvf @xmath75018 is the strongest part of a blend with the hvf components of @xmath75032 and @xmath75051 from @xmath2 d to about @xmath12 d. this identification is confirmed by the close agreement of the measured velocities of hvf @xmath75169 and hvf @xmath75018 .
this feature loses a distinct minimum after @xmath45 d and we stop measuring hvf @xmath75018 at that point .
possible evidence exists for hvf components from @xmath7@xmath74924 , 5129 .
figure [ pvca ] displays the h&k blend @xmath73935 ( left ) and the ir3 @xmath78579 ( right ) . at @xmath2 d ,
the absorption minima for both blends are near 32,000 km s@xmath1 and the features have blue wings that extend beyond 40,000 km s@xmath1 .
simultaneous observations of h&k and ir3 are of particular interest as they provide a rare opportunity to compare the evolution of these features .
as shown in figure [ pvca ] and table [ hvtbl ] , the measured velocities and rates of change are very similar for hvf from both of the strong blends .
although hvf velocities are about 7,000 km s@xmath1 higher than hvf and nearly 10,000 km s@xmath1 higher than hvf at @xmath2 d , they have a significantly faster decline rate , and the hvf velocities for all lines converge near 21,500 km s@xmath1 at @xmath11 d. hvf from the ir3 are unblended , and they are weak but detectable until @xmath4-max .
that is about 5 days after the last hvf detection from other lines .
the hvf component of h&k @xmath73945 is blended with pvf @xmath73858 .
side by side comparison of the h&k and ir3 line profiles suggests that the hvf from h&k are broadened by pvf beginning about @xmath12 d and the pvf influence gets stronger with time .
we find that @xmath48 d is the final phase at which the absorption minimum of the ca / si blend is consistent with the hvf of ir3 , so we stop measuring h&k hvf after that phase .
@xcite demonstrated that there are conditions under which the absorption feature attributed to hvf h&k can be produced by a combination of and a particular density profile for .
@xcite used polarization data in a @xmath11 d spectrum from sn 2004dt to suggest that provides the dominant contribution to this blend at @xmath11 d. @xcite compare the average strength of the hvf component in spectra from sn ia at @xmath4-max .
they find that slowly declining sn ia produce either high photospheric velocities or strong hvf close to @xmath4-max , but not both .
sn 2009ig fits this pattern with @xmath49 13,400 km s@xmath1 , while the ir3 series in figure [ pvca ] shows that hvf is very weak at @xmath4-max .
we use velocity offsets from pvf and their relative strengths by phase to confirm the identifications of hvf in spectra of sn 2009ig .
pvf are identified for each of the same lines that produce hvf detections : @xmath74560 and @xmath76355 ( figure [ pvsi ] ) , @xmath7@xmath75453 , 5641 ( figure [ pvs ] ) , and @xmath7@xmath75018 , 5169 ( figure [ pvfe ] ) .
pvf velocity measurements are in table [ pstbl ] .
there are weak suggestions of pvf for some lines in the @xmath2 d spectrum of sn 2009ig , but the first unambiguous detections of pvf are at @xmath10 d. @xcite note that @xmath10 d is the phase at which the rate of change slows down for @xmath50 colors .
pvf become progressively stronger with time , and the pvf of most lines are stronger than the hvf by about @xmath12 d. at @xmath11 d the pvf components dominate all blended absorption features with little or no influence from the hvf .
pvf velocities in sn 2009ig decline steeply at first and then more gradually .
they become essentially constant after @xmath46 d. this behavior is consistent with the typical pattern for pvf velocities in sn ia @xcite . the pvf velocities of sn
2009ig are about 2,000 km s@xmath1 greater than velocities in a
normal " sn ia , but they are not exceptional
. we measure @xmath8 13,400 km s@xmath1 , which places sn 2009ig in the _ hv _ group defined by @xcite .
@xcite show that @xmath51 13,500 km s@xmath1 for @xmath52% of sn ia having @xmath53 mag .
the change in @xmath76355 velocity between + 0 d and + 10 d is @xmath54 km s@xmath1 per day .
that puts sn 2009ig in the _ low velocity gradient _ group defined by @xcite .
pvf @xmath73858 is blended with hvf h&k .
the contribution is significant as early as @xmath12 d , but it is not clear when the pvf become strong enough to define the minimum in this feature .
we do not take velocity measurements from this line .
pvf @xmath74130 is the probable source of the absorption visible as a small notch near 4,000 in the p cygni emission peak from h&k .
hvf @xmath74264 is a likely contributor to this feature .
pvf of @xmath7@xmath75051 , 5972 are possibly identified .
most of the pvf are strongly affected by blending with other pvf .
pvf @xmath75018 ( left panel figure [ pvfe ] ) is blended with pvf @xmath75051 and pvf @xmath75032 .
the and lines may pull the observed minimum of this feature to the red since the pvf velocities for @xmath75018 before @xmath4-max are 1,0001,500 km s@xmath1 lower than pvf @xmath75169 .
the strong absorption feature near 4950 is formed by pvf @xmath75169 with contributions from pvf @xmath75266 and pvf @xmath75129 .
the blend of pvf @xmath75018 , @xmath75051 , and @xmath75032 is observed near 4800 , and this feature continues to grow stronger after @xmath4-max .
the p cygni emission component of this feature pushes the minimum for pvf @xmath75169 to velocities that are about 1,000 km s@xmath1 lower than pvf @xmath75018 and other pvf lines after @xmath4-max .
pvf @xmath75453 is blended with hvf @xmath75641 from @xmath10 d to @xmath41 d. we stop measuring pvf @xmath75453 after @xmath16 d when the velocity abruptly drops from 12,400 km s@xmath1 at @xmath16 d to 10,400 km s@xmath1 at @xmath47 d. that is not consistent with the other pvf lines at this phase including the other line at @xmath75641 .
the displacement of the pvf @xmath75453 feature suggests that it is also moved to the red by p cygni emission from the strong blend of fe - group lines observed at 4950 .
pvf from lines of doubly ionized ions are weak at @xmath2 d and @xmath9 d , become significantly stronger at @xmath10 d , and disappear at about @xmath47 d. figure [ pvsi ] shows this pattern for pvf @xmath74560 .
other possible but unconfirmed identifications of pvf from twice - ionized lines show similar behavior : @xmath75740 , @xmath74264 , and @xmath7@xmath74407 , 5129 .
the hvf components of lines also appear to be stronger at @xmath10 d than at @xmath2 d. pvf @xmath74481 is a major contributor to the very strong absorption found near 4220 in a blend with pvf @xmath74407 and hvf @xmath74560 .
the left panel of figure [ pvsi ] shows the minimum of this feature near the dotted line at 21,500 km s@xmath1 until @xmath47 d. a pvf with a velocity of 14,000 km s@xmath1 would appear at about 19,000 km s@xmath1 on the velocity scale of @xmath74560 .
synow models show a better fit to the data when the pvf component is stronger than from @xmath12 d until about @xmath47 d. the synow result is consistent with the abrupt shift of this feature to the red at @xmath47 d that aligns the minimum with characteristic pvf velocities for @xmath74481 .
the evidence for pvf from unburned c and o in sn 2009ig is inconclusive .
@xmath74743 makes a reasonable fit to absorption features for both hvf near 4370 and pvf near 4480 ( figure [ pvsi ] , wavelength axis on top ) .
if , however , @xmath74743 is the source of these features , then @xmath76580 should also be easily detected .
the pvf component of @xmath76580 would appear near 3000 km s@xmath1 in the velocity space of @xmath76355 .
@xcite cite work with synow models @xcite that finds @xmath76580 to be present in the early - time spectra of sn 2009ig .
is purported to cause a flattening of the red wing in the line profile of @xmath76355 from @xmath2 d to @xmath10 d. it may be that including improves the synow fit to the line profile , but we do not find any absorption features at this location ; the tiny notch at 6200 is telluric .
@xmath77773 develops a broad pvf line beginning about @xmath47 d , but it is much weaker than pvf found in many sn ia .
the features near 7400 in the spectra from @xmath47 d to + 9 d correspond to pvf of @xmath77773 at 11,00012,000 km s@xmath1 , consistent with other pvf velocities those phases .
5169 ( blue ) , and @xmath76355 ( green ) are plotted together in velocity space at phases @xmath2 d , @xmath10 d , @xmath12 d , @xmath11 d , and @xmath55 d ( top to bottom ) . both hvf of occupy the same physical space at @xmath2 d. the hvf has a similar profile with an offset of about 7,000 km s@xmath1 .
the hvf of @xmath75018 forms the notch near 35,000 km s@xmath1 . at @xmath2 d
, the hvf is confined to the lower velocity half of the region . at @xmath10 d ,
hvf remain strong from all lines and weak pvf are detected from and .
weakening hvf of and are still detected at @xmath12 d , and hvf remain strong . at this phase hvf @xmath75169 is affected by blending with pvf @xmath75018 , and hvf h&k is blended with @xmath73858 . at @xmath11 d , the only evidence for hvf is very weak , and at @xmath55 d , all absorptions are pvf .
the hvf of ( @xmath2 d ) are asymmetric with steeper blue sides while pvf of siii ( @xmath11 d,@xmath55 d ) are nearly symmetric.,scaledwidth=45.0% ] [ area4 ] d and @xmath37 d are plotted with spectra from synow models .
the data are shown in red .
model spectra with only pvf are dotted blue and model spectra with both pvf and hvf are solid black .
the addition of hvf lines clearly improves the agreement between the model and the data.,scaledwidth=52.0% ] [ syn ]
figure [ area4 ] can be used to trace the location of the line forming regions for hvf of h&k , ir3 , @xmath75169 , and @xmath76355 at phases @xmath2 d , @xmath10 d , @xmath12 d , @xmath11 d , and @xmath55 d ( top to bottom ) .
velocity serves as a proxy for radial distance , with the center of the sn on the right at 0 km s@xmath1 and the radius increasing with velocity from right to left .
the spectra are normalized to 1.0 at 6500 and the features are normalized to a flat continuum .
the flux scale is the same in each panel . at @xmath2 d ,
all features have the same flux calibration . at later phases ,
all other features retain the same flux calibration , but h&k hvf is scaled down to fit in the figure . at @xmath2 d ,
both hvf ( purple and red ) occupy the same velocity space , and a blue wing is present in both features that extends from 40,000 to @xmath5655,000 km s@xmath1 .
the minima for the hvf are at velocities greater than 30,000 km s@xmath1 ( table [ hvtbl ] and figure [ vplot ] ) .
the hvf velocity from @xmath75169 is near 26,000 km s@xmath1 .
the line profile for at this phase is similar to that of h&k in both width and depth .
a velocity difference of about 5,000 km s@xmath1 is evident between the line forming region ( purple line ) and line forming region ( blue line ) on the red sides of the features .
the blue sides , however , are closely aligned .
this discrepancy is caused by the hvf of @xmath75018 that creates a notch observed near 35,000 km s@xmath1 in the velocity space of @xmath75169 ( blue line ) . between the red edge of this notch and the continuum ,
the side of the @xmath75169 hvf is shifted by about 7,000 km s@xmath1 to the blue .
the hvf at @xmath2 d has a minimum at about 23,000 km s@xmath1 , and the line forming region is confined to a velocity space that is almost entirely in the lower velocity half of the hvf .
the line profile at @xmath2 d ( green ) is clearly asymmetric , with a steeper blue side and an abrupt transition at the blue edge from the absorption to the continuum .
the blue limit of the hvf has nearly the same velocity as the absorption minimum for .
in contrast , the pvf of ( @xmath11 d,@xmath55 d ) are nearly symmetric . at @xmath10
d , hvf are strong from all lines , with velocities for about 29,000 km s@xmath1 , for near 23,000 km s@xmath1 , and for near 21,000 km s@xmath1 .
the hvf is again formed in a smaller region than for other lines .
the blue limit of the hvf is close to the same velocity as it was at @xmath2 d. pvf from and can be detected near 16,000 km s@xmath1 at this phase , with the pvf significantly distorting the line profile . at @xmath12 d ,
the hvf of is noticeably weaker , but it is clearly present as it forms the distortion in the blue side of the pvf . using the methods described in [ i d ] , the velocity of hvf from is measured to be near 21,000 km s@xmath1 .
the hvf velocity of is near 22,000 km s@xmath1 and the hvf remains strong at about 27,000 km s@xmath1 . at this phase
, the hvf of @xmath75169 has a significant contribution from pvf @xmath75018 that narrows the observed profile .
the h&k hvf also receives a strong contribution from pvf @xmath73858 . at @xmath11
d , the most obvious evidence for hvf is a very weak absorption from the ir3 hvf .
the blue line absorption near 21,000 km s@xmath1 is primarily due to the pvf of @xmath75018 .
the h&k hvf absorption ( purple ) near 23,000 km s@xmath1 is a blend with pvf @xmath73858 .
the magnitude of the contribution from can be estimated by comparing the hvf of h&k to the hvf of ir3 .
the broad ca / si profile does not provide a hvf detection . at this phase ,
has a pvf at 14,000 km s@xmath1 , and the line profile is nearly symmetric in velocity space with no evidence of a hvf .
figure [ area4 ] shows that the absorption minima for hvf are always observed above 20,000 km s@xmath1 .
there is no evidence at any phase for hvf with lower velocities .
by @xmath55 d , all absorptions are easily explained by associations with pvf at velocities near 13,000 km s@xmath1 . the blue wing of the ir3 pvf at this phase is slightly flattened by the remnant of the hvf .
figure [ vplot ] plots hvf and pvf velocity measurements by phase for @xmath76355 , @xmath74560 , @xmath7@xmath75453 , 5641 , @xmath7@xmath73945 , 8579 , and @xmath7@xmath75018 , 5169 .
all lines contribute both hvf and pvf measurements to the figure .
the plotted data are found in tables [ hvtbl ] and [ pstbl ] .
this figure is used to visually compare the relative velocities of both hvf and pvf from different lines at the same phases .
variations in the decline rates are easy to discern .
hvf and pvf from the same lines or the same ions can also be compared .
hvf velocities converge to a common value near 21,500 km s@xmath1 and pvf velocities are grouped together at all phases .
a significant gap in velocity space is evident between the hvf and pvf regions for all lines at all phases .
the separation is about 8,000 km s@xmath1 , except which displays a greater separation before @xmath16 d and @xmath76355 for which the separation at all phases is about 7,000 km s@xmath1 .
no features are found at velocities that would place them between these layers .
figure [ vplot ] shows that from @xmath2 d to @xmath12 d , the hvf velocities of both lines are intermediate between the higher hvf velocities and the lower hvf velocities for other lines . by @xmath41 d ,
hvf velocities are essentially the same as the velocities of other hvf that are not .
the patterns of velocity by phase can be compared the behavior of the hvf line profiles by phase that are displayed in figure [ area4 ] . and both have their maximum velocities in the earliest spectrum at @xmath2 d. at that phase , the hvf line profile is wide , deep and more like the the hvf of then the hvf of .
the velocity difference between the hvf of and is 3,600 km s@xmath1 . at @xmath12 d ,
the hvf line profiles are less distinct due to blending with pvf .
the line profile is narrower and weaker and the velocities of and are separated by 1,600 km s@xmath1 .
the hvf of at this phase remain broad and at higher velocities . at @xmath11 d ,
the hvf of and are very weak and their measured velocities are within 400 km s@xmath1 . the velocity dispersion of the hvf components of , , , and is confined within a narrow velocity region of less than 2000 km s@xmath1 from @xmath10 d until hvf are no longer detected . by @xmath41 d ,
the mean hvf velocity ( excluding ) is about 21,500 km s@xmath1 , and velocities reach this value a few days later .
after the hvf velocities reach @xmath56 21,500 km s@xmath1 , the velocities remain constant within the range of scatter introduced by measurement uncertainties .
the mean pvf velocity is observed to be about 13,000 km s@xmath1 at @xmath46 d , and it declines less than 1,000 km s@xmath1 through the end of our observations at @xmath3 d. ( most of the decline is caused by @xmath75169 ; see [ ps ] . )
the relative velocities of the hvf and pvf regions imply that , with homologous expansion , the radius of the hvf line forming region is @xmath57 times greater than the radius of the pvf line forming region .
figure [ syn ] shows synow model spectra plotted with data from sn 2009ig obtained at @xmath43 d and @xmath37 d. these phases are chosen because they include hvf and pvf , and the relative strengths of the components change in the time interval .
the data are plotted in the figure with red lines .
synow spectra from models with only pvf velocities are plotted in blue with dotted lines , and model spectra with both hvf and pvf are solid black . at both phases ,
the addition of hvf noticeably improves the models in the regions of the and features .
more subtle improvements are evident around the and features .
the synow models using only pvf velocities produce a ir3 feature that is weak and shifted to longer wavelengths because the lines at @xmath43 d and @xmath37 d are dominated by the hvf component .
the relative weakness of the pvf ir3 feature in the model is a consequence of synow setting the optical depth for each ion to represent the strongest line ( in this case h&k ) and then calculating the strength of other lines from that ion assuming local thermodynamic equilibrium ( lte ) .
the synow parameters for the models are given in table [ syntbl ] .
in these simple models , the same hvf and pvf velocities are assigned to all ions except for high - velocity and .
figure [ syn ] can be compared to figure 4 of @xcite , which uses high - velocity to improve the synow fit to a @xmath10 d spectrum from sn 1994d .
additional hvf contributions from and can be added to further improve the agreement between synow model spectra and the data , but hvf from these lines are not clearly identified in sn 2009ig . and are not included in the velocity tables or plots .
figure [ psi6 ] shows the development of @xmath76355 absorption features by phase in six sn ia for which there are high - cadence spectroscopic observations that begin at least 12 days before @xmath4-max .
were obtained through the _ wiserep _ data archive @xcite .
the original published sources of the spectra are sn 2003du @xcite , sn 2005cf @xcite , sn 2011fe @xcite ( _ snfactory _ ) , and sn 2012fr @xcite . the sn 2012cg data are previously unpublished fast spectra from the cfa archive . ]
we note that none of these sn ia are rapid decliners ; their @xmath30 values ( in sn order of discovery ) are 1.02 , 1.05 , 0.90 , 1.11 , 0.86 , and 0.80 .
the spectra obtained before @xmath12 d provide strong evidence for detached hvf regions from both and in five of the six sn . with one exception , the earliest spectra require separate hvf and pvf components to fit the line profiles .
only sn 2011fe at @xmath58 d has a broad absorption feature that can be fit with a single gaussian .
the line profiles for every spectrum in figure [ psi6 ] can be correlated with a profile from one of the phases in the spectral sequence of sn 2009ig . in some cases ,
the phase must be offset slightly to establish the best match .
for example , the sequence of line profiles from @xmath12 d to @xmath16 d in sn 2005cf is nearly identical to the @xmath10 d to @xmath45 d profiles in sn 2009ig .
the features of sn 2003du from @xmath10 d to @xmath37 d are similar to the features of sn 2009ig at the same phases , but the hvf diminish more rapidly in sn 2003du .
the most extreme phase offset relative to sn 2009ig is found in sn 2011fe , for which the sequence from @xmath58 d to @xmath12 d corresponds to sn 2009ig from @xmath10 d to @xmath11 d. although the first spectrum of sn 2011fe was obtained earlier than the first spectrum of sn 2009ig with respect to @xmath4-max , the spectrum of sn 2011fe corresponds to a phase in the sequence that is after the detached hvf are observed .
there is only a it is possible that even earlier observations of sn 2011fe could have revealed hvf .
previous authors have studied the eccentricities of the @xmath76355 line profile in sn ia without directly detecting hvf .
the presence of a hvf component from @xmath76355 is inferred from unusual line profiles that are wide , square , or triangular .
none of the spectra available before sn 2009ig display separate and distinct high - velocity features such as found at @xmath2 and @xmath9 d in sn 2009ig and from @xmath2 d to @xmath43 d in sn 2012fr .
@xcite use synow to model early - time spectra of eight sn ia .
they show that fitting the @xmath76355 line profile requires a detached " region at 20,00022,000 km s@xmath1 for seven of the eight sn ia .
the other sn in their sample requires a detached region at 24,500 km s@xmath1 .
@xcite compare the early spectra of six sn ia and report that each of the @xmath76355 line profiles are well fit by a double - gaussian function with separate central wavelengths .
they show that two separate absorption features can be combined to produce flat - bottomed and triangular line profiles .
figure [ psi6 ] shows that for different sn ia the characteristic velocities of the hvf and pvf regions can vary by up to 3,000 km s@xmath1 , but in all cases a gap of at least 5,000 km s@xmath1 is maintained between the hvf and pvf .
@xcite , @xcite , @xcite , and @xcite describe the separation between hvf and pvf components to be about 8,000 km s@xmath1 , the same value as the velocity gap in sn 2009ig .
the correlation of line profiles from pre - maximum spectra of other sn ia to the spectral sequence of sn 2009ig confirms a suggestion made by @xcite that the `` peculiar '' profiles of @xmath76355 in early - time spectra of sn ia can be explained as part of a common evolutionary sequence that includes separate hvf and pvf components .
@xcite also propose that the strength of the hvf components from and lines is correlated in sn ia .
thus , if one of these lines produces strong hvf , then the other will also have a strong hvf component .
the spectra of sn 2009ig confirm this relationship , and they also demonstrate that the presence of strong hvf from and may be an indicator for hvf of , , and .
the agreement between the sn 2009ig results and these well - studied sn ia demonstrates that the behavior of hvf line forming regions in spectra of sn 2009ig is not unique , and in fact is quite common .
the papers discussed in this section identify fifteen sn ia with spectra obtained at @xmath12 d or earlier .
all fifteen sn display high - velocity at velocities between 17,000 and 22,000 km s@xmath1 , and all but sn 2011fe show evidence for separate hvf and pvf absorption components .
the observations of detached hvf in sn 2009ig suggest the presence of a region at 20,00023,000 km s@xmath1 in the velocity space of the sn with a localized enhancement of element abundance or density , or both .
the proposed line forming region would be the source of hvf for , , and from @xmath10 d to @xmath16 d. this simple model is complicated by the fact that hvf are initially detected at a range of even higher velocities . in addition , the hvf velocities of , , and decline at different rates .
this suggests that the lines may sample different density structures .
although the velocity ranges overlap , the hvf of each line are not even detected in the same region .
figure [ area4 ] shows that at @xmath2 d , @xmath76355 has the narrowest detached hvf with a velocity range from 10,000 to 30,000 km s@xmath1 . at the same phase ,
the hvf of and hvf extend from 10,000 to 50,000 km s@xmath1 , but they are offset from each other .
absorptions are exceptionally broad in velocity space , with absorption minima that are several thousand km s@xmath1 higher than the hvf components of other lines . due to the very low excitation potentials of the lines ( @xmath59 ev ) , they are capable of producing detectable absorptions at much lower abundances than most other lines found in sn ia
. lines also have low excitation potentials ( @xmath60 ev ) . in sn 2009ig ,
the hvf from are nearly as wide and deep as hvf of , with velocities between and .
the simplest explanation for both the hvf ca and fe observations is that primordial abundances of these elements are sufficient to produce the observed absorptions .
the line profiles can be produced by light from the sn passing through a line forming region that extends from about 10,000 to 50,000 km s@xmath1 in velocity space .
the observed minima for these features would move to lower velocities with time , as expansion of the sn reduces opacity at high velocities more rapidly than it does at lower velocities .
as the column depth is reduced , the observed minima become nearly constant at velocities near to the inner edge of the line forming region .
if the hvf of and are formed in regions of unburned material from the progenitor , it is still a challenge to explain the simultaneous presence of detached hvf for @xmath76355 ( 8.1 ev ) and ( 13.7 ev ) . an explanation is also required for the fact that the hvf velocities for all lines , as measured at the absorption minima , eventually converge near 21,500 km s@xmath1 . since figure [ area4 ] shows that the line forming materials are distributed through a wide range of velocities , there must be some motivation for the line opacities of , , , , and to have approximately equal values at the same phases and in a relatively narrow velocity range .
attempts have been made to explain hvf without invoking a separate line forming region .
the flat - bottomed and triangular shapes observed in the @xmath76355 line profile at early times receive particular attention .
none of the physical models used to generate unusual line profiles survive comparisons with early observations of multiple sn ia .
hvf from have also been attributed to opacity effects from recombination , but the time scale of recombination is a day or two and hvf from are observed from 8 to 15 days in individual sn ia .
explosion models for sn ia make predictions for structures that range from spherical to highly asymmetric .
the resulting chemical distributions may be simply stratified by atomic mass , violently mixed between regions of burned material , or contain bubbles of material moving between burning regions .
model results can be tuned to produce specific layering compositions and velocity distributions that may mimic hvf of specific atoms at specific locations .
we are not aware of any explosion models that predict the relative uniformity across a range of absorbing materials that is observed for hvf in sn 2009ig .
one way to increase the local density in high - velocity regions is by interaction between the expanding ejecta and nearby circumstellar material ( csm ; * ? ? ?
the shock front sweeps up csm and a shell is locked in " after the shock moves on
. the shell will continue in homologous expansion with the rest of the sn ejecta .
this model satisfies the constraint that the hvf line forming region exist in a narrow , unchanging velocity window and explains the asymmetric line profiles of the @xmath7 6355 hvf .
three of the recent models have been able to produce variable density structures : symmetric ignitions that generate strong mixing @xcite , gravitationally confined detonations with burning regions on the surface of the wd @xcite , and double - detonation models that produce a high - velocity burning region with a density gap between it and the center of the sn @xcite .
we present a time series of spectra of sn 2009ig obtained between @xmath14 d and @xmath61 d with respect to the time of @xmath4-max .
the two earliest observations in our sample ( @xmath14 d and @xmath15 d ) are the first spectra of a sn ia to resolve the high - velocity component of @xmath76355 as a distinctly separate absorption feature .
hvf are identified in the spectra from , , , , and , and we show that the line profiles require a detached hvf line forming region .
simultaneous detections are made for hvf and pvf of the same lines from @xmath10 d to @xmath16 d. velocity measurements demonstrate that the hvf and pvf line forming regions are separated by @xmath62 km s@xmath1 as long as hvf are detected . using a mean velocity of 21,500 km s@xmath1 for the hvf layer and 13,000 km s@xmath1 for the pvf layer
, we show that @xmath63 .
the initial hvf velocities and the velocity decline rates are different for each line . the fact that the hvf velocities eventually converge suggests the presence of a line forming region in sn 20019ig at velocities between 20,000 and 23,000 km s@xmath1 .
a comparison of the hvf from , , and by phase in velocity space shows that hvf from different lines are formed in overlapping but not identical regions .
hvf of and have strong features that range from 10,000 to 50,000 km s@xmath1 in the earliest spectra , but the locations of their absorption minima differ by @xmath56 7,000 km s@xmath1 . at the same phases , hvf
are confined to a much narrower region that extends from 10,000 to only 30,000 km s@xmath1 .
we explore the physical conditions that can produce detached hvf , and find that it is challenging for any model to reproduce the observations of sn 2009ig .
the spectra of sn 2009ig form a complete map of the transition from the earliest observations of detached hvf to features with contributions from both hvf and pvf , and finally to phases when all of the absorptions are pvf .
hvf of both and are frequent characteristics of sn ia observed before @xmath12 d , and that allows us to compare our results to observations of several other sn ia .
all of the line profiles found in early - time spectra of sn ia can easily be correlated with individual profiles that are part of the development sequence defined by sn 2009ig .
we interpret this as evidence for a common evolutionary sequence in sn ia that requires both hvf and pvf line forming regions .
the authors wish to thank the chairmen of the tacs from the university of texas and penn state university for providing director s discretionary time for the het / lrs observations .
thanks nick suntzeff and rob robinson for insightful comments .
also thanks mark phillips for providing perspective , as well as michael childress for helpful discussions and for sharing an advance copy of his paper on sn 2012fr .
we thank mark sullivan , isobel hook , peter nugent , andy howell , and bill vacca for sharing their own data on line profiles in young sn ia .
the authors make frequent use of david bishop s excellent webpage listing recent supernovae and valuable references associated with them : www.rochesterastronomy.org / snimages/. j.v .
is supported by hungarian otka grants k-76816 and nn-107637 , nsf grant ast-0707769 , and texas advanced research project arp-009 .
the research of j.c.w . in supported in part by nsf grant ast-1109801 .
the cfa supernova program is supported by nsf grant ast-1211196 to the harvard college observatory .
x. wang is supported by the natural science foundation of china ( nsfc 11178003 , 11073013 ) , the china-973 program 2009cb824800 , and nsf grant ast-0708873 ( through l. wang ) .
is supported by nsf grant ast-1008343 .
a.v.f . is grateful for financial assistance from nsf grant ast-1211916 , the tabasgo foundation , and the christopher r. redlich fund .
we greatly appreciate the valuable assistance provided by staff members at the observatories where sn 2009ig was observed .
22.5 & -14.5 & lick / kast & 34009000 & 6700 + aug .
22.6 & -14.4 & keck i / lris & 31007860 & 6700 + aug .
23.4 & -13.6 & het / lrs & 42109000 & 4225 + aug .
23.6 & -13.4 & swift / ugrism & 32004100 & 4070 + aug .
24.5 & -12.5 & lick / kast & 35009000 & + aug .
25.5 & -11.5 & lick / kast & 35009000 & 4000 + aug .
25.6 & -11.4 & swift / ugrism & 32004400 & 4000 + aug .
26.5 & -10.5 & mmt / blue channel & 31908330 & + aug .
27.5 & -9.5 & lick / kast & 34009000 & 6700 + aug .
27.5 & -9.5 & mmt / blue channel & 31808270 & 6700 + aug .
28.5 & -8.5 & lick / kast & 34009000 & 6700 + aug .
28.5 & -8.5 & mmt / blue channel & 31808320 & 6700 + aug .
29.4 & -7.6 & het / lrs & 71709000 & 7210 + aug .
29.5 & -7.5 & mmt / blue channel & 31908270 & 7210 + aug .
30.4 & -6.6 & het / lrs & 47109000 & 6850 + aug .
30.4 & -6.6 & mmt / blue channel & 31808250 & 6850 + aug . 31.4 & -5.6 & het / lrs & 41709000 & + sep . 1.6 & -4.4 & swift / ugrism & 32004900 & + sep .
2.4 & -3.6 & het / lrs & 41609000 & + sep .
3.6 & -2.4 & swift / ugrism & 32004400 & 4200 + sep .
3.5 & -2.5 & het / lrs & 41909000 & 4200 + sep .
4.4 & -1.6 & het / lrs & 41809000 & + sep .
5.4 & -0.6 & het / lrs & 41809000 & + sep .
6.7 & + 0.7 & swift / ugrism & 32004900 & + sep .
9.4 & + 3.4 & het / lrs & 41809000 & + sep .
10.4 & + 4.4 & het / lrs & 41809000 & + sep .
11.5 & + 5.5 & flwo / fast & 33007200 & + sep .
12.5 & + 6.5 & flwo / fast & 33007200 & + sep .
14.4 & + 8.4 & het / lrs & 41809000 & + sep .
15.5 & + 9.5 & flwo / fast & 33007200 & + sep .
16.5 & + 10.5 & flwo / fast & 33007200 & + sep .
17.4 & + 11.5 & flwo / fast & 33007200 & + sep .
18.5 & + 12.5 & flwo / fast & 33007200 & + crcccccccc aug .
22 & -14 & 32.2 & 25.7 & 26.6 & 26.4 & & 23.9 & 22.8 & 31.4 + aug .
23 & -13 & 30.7 & & 25.2 & 24.5 & 23.0 & 23.1 & 22.0 & 30.5 + aug .
24 & -12 & 29.2 & 22.8 & 24.0 & 23.1 & 22.1 & 22.4 & 21.1 & 29.2 + aug .
25 & -11 & 27.9 & 22.2 & 23.2 & 22.8 & 21.9 & 21.8 & 21.0 & 28.4 + aug .
26 & -10 & 26.5 & 22.1 & 22.7 & 22.4 & 21.7 & 21.0 & 20.8 & 27.5 + aug .
27 & -9 & 25.1 & 21.7 & 22.5 & 22.0 & 21.6 & 21.1 & 20.7 & 26.2 + aug .
28 & -8 & 24.0 & 21.7 & 22.2 & 21.9 & 21.5 & 20.8 & 20.8 & 25.0 + aug .
29 & -7 & 23.4 & 21.5 & 22.0 & 21.5 & 21.6 & 20.6 & 20.6 & 24.1 + aug .
30 & -6 & 22.6 & 21.3 & & 21.2 & 21.4 & 20.8 & 20.8 & 23.2 + aug .
31 & -5 & 22.1 & 21.1 & & 21.1 & & 20.4 & & 22.9 + sep . 01 & -4 & 21.3 & 21.0 & & & & & & + sep .
02 & -3 & & & & & & & & 21.8 + sep .
03 & -2 & 21.1 & & & & & & & 21.4 + sep . 04 & -1 & & & & & & & & 21.4 + sep .
05 & 0 & & & & & & & & 21.2 + crcccccccc aug .
22 & -14 & & 16.7 & 18.4 & 18.0 & 15.9 & 17.0 & & + aug .
23 & -13 & & 15.2 & 16.3 & 17.6 & 15.5 & 16.2 & 16.5 & + aug .
24 & -12 & & 14.0 & 15.1 & 16.3 & 13.9 & 15.3 & 15.9 & + aug .
25 & -11 & & 13.6 & 14.5 & 15.9 & 13.7 & 15.2 & 15.6 & + aug .
26 & -10 & 15.8 & 13.1 & 14.3 & 15.1 & 13.5 & 15.0 & 15.2 & + aug .
27 & -9 & 15.4 & 12.9 & 13.8 & 14.8 & 13.2 & 14.6 & 14.5 & 16.0 + aug .
28 & -8 & 15.4 & 12.8 & 13.6 & 14.6 & 13.0 & 14.5 & 14.2 & + aug .
29 & -7 & 15.2 & 12.4 & 13.3 & 14.4 & 12.9 & 14.2 & 14.1 & 15.0 + aug .
30 & -6 & 14.9 & 12.3 & 13.1 & 14.2 & 12.8 & 14.1 & 13.9 & 14.4 + aug .
31 & -5 & 15.1 & 11.9 & 12.9 & 14.0 & 12.4 & 13.7 & 13.6 & 13.9 + sep .
01 & -4 & 14.9 & 12.1 & & & & & & + sep .
02 & -3 & & & 12.0 & 13.5 & & 13.2 & 13.4 & 13.8 + sep .
03 & -2 & 14.8 & & 12.1 & 13.4 & & 13.0 & 13.3 & 13.9 + sep .
04 & -1 & & & 12.3 & 12.7 & & 12.8 & 13.4 & 14.2 + sep .
05 & 0 & & & 12.4 & 12.5 & & 12.9 & 13.4 & 14.6 + sep .
06 & 1 & 14.6 & & & & & & & + sep .
07 & 2 & & & & & & & & + sep .
08 & 3 & & & & & & & & + sep .
09 & 4 & & & 12.4 & 11.9 & & 12.8 & 13.3 & 14.2 + sep .
10 & 5 & & & 12.6 & 11.8 & & 13.4 & 13.3 & 14.0 + sep .
11 & 6 & 14.3 & & 12.7 & 11.9 & & 13.7 & 13.2 & + sep .
12 & 7 & 14.2 & & 12.8 & 12.0 & & 13.7 & 13.1 & + sep .
13 & 8 & & & & & & 13.2 & & + sep .
14 & 9 & & & 12.7 & 11.9 & & 13.2 & 13.2 & 14.0 + sep .
15 & 10 & 13.9 & & 12.8 & 12.0 & & 13.5 & 13.0 & + sep .
16 & 11 & 13.8 & & 12.8 & 11.9 & & 13.0 & 13.0 & + sep .
17 & 12 & 13.5 & & 12.6 & 11.8 & & & 12.9 & + sep .
18 & 13 & 13.5 & & 12.8 & 11.7 & & & 13.0 & + lcccccccc si ii & 3.0 & 16 & 30 & 2.0 & 1.2 & 16 & 30 & 2.0 + si ii(hvf ) & 1.0 & 23 & 50 & 2.0 & 0.1 & 23 & 50 & 2.0 + si iii & 0.8 & 16 & 50 & 2.0 & 0.8 & 13 & 50 & 2.0 + si iii(hvf ) & 0.5 & 23 & 50 & 2.0 & 0.5 & 23 & 50 & 2.0 + mg ii(hvf ) & 0.5 & 22 & 50 & 4.0 & 0.3 & 22 & 50 & 4.0 + s ii & 0.7 & 16 & 50 & 2.0 & 0.5 & 13 & 50 & 2.0 + s ii(hvf ) & 0.1 & 23 & 50 & 2.0 & 0.1 & 23 & 50 & 2.0 + ca ii & 2.0 & 16 & 50 & 8.0 & 2.0 & 13 & 50 & 8.0 + ca ii(hvf ) & 25 & 28 & 50 & 4.0 & 6.0 & 26 & 50 & 4.0 + fe ii(hvf ) & 0.5 & 28 & 50 & 2.0 & 0.05 & 23 & 50 & 2.0 + fe iii & 0.5 & 16 & 50 & 2.0 & 0.5 & 16 & 50 & 2.0 + | we report measurements and analysis of high - velocity ( @xmath0 20,000 km s@xmath1 ) and photospheric absorption features in a series of spectra of the type ia supernova ( sn ) 2009ig obtained between @xmath2 d and @xmath3 d with respect to the time of maximum @xmath4-band luminosity ( @xmath4-max ) .
we identify lines of , , , , and that produce both high - velocity ( hvf ) and photospheric - velocity ( pvf ) absorption features .
sn 2009ig is unusual for the large number of lines with detectable hvf in the spectra , but the light - curve parameters correspond to a slightly overluminous but unexceptional sn ia ( @xmath5 mag and @xmath6 mag ) . similarly , the @xmath76355 velocity at the time of @xmath4-max is greater than normal " for a sn ia , but it is not extreme ( @xmath813,400 km s@xmath1 ) . the @xmath2 d and @xmath9 d spectra clearly resolve hvf from @xmath76355 as separate absorptions from a detached line forming region . at these very early phases , detached hvf are prevalent in all lines . from @xmath10 d to @xmath11 d , hvf and pvf
are detected simultaneously , and the two line forming regions maintain a constant separation of about 8,000 km s@xmath1 .
after @xmath11 d all absorption features are pvf .
the observations of sn 2009ig provide a complete picture of the transition from hvf to pvf .
most sn ia show evidence for hvf from multiple lines in spectra obtained before @xmath12 d , and we compare the spectra of sn 2009ig to observations of other sn .
we show that each of the unusual line profiles for @xmath76355 found in early - time spectra of sn ia correlate to a specific phase in a common development sequence from hvf to pvf . |
magnetic ordering phenomena of both classical and quantum antiferromagnets on non bipartite lattices are a fascinating subject . the simplest and most frequently studied model of this type is the heisenberg antiferromagnet on a triangular lattice ( haft ) . while the question of whether or not the typical features of this model have been observed in experiments is still a controversial issue @xcite , the theoretical understanding of these features has advanced rapidly during recent years @xcite . in contrast to antiferromagnets on bipartite lattices , the haft exhibits non collinear magnetic order in its classical and most likely also in its quantum ground state . as a consequence
, the order parameter of the haft is represented locally by a set of three mutually orthogonal unit vectors or , alternatively , by a rotation matrix which defines the local orientation of this set relative to some fixed frame of reference .
renormalization group ( rg ) studies of appropriate nonlinear sigma ( nl@xmath6 ) models @xcite have revealed a number of interesting properties of the haft .
the symmetry of the model was found to be dynamically enhanced from @xmath7 to @xmath8 , and in a two loop rg calculation for the classical haft @xcite , the temperature dependence of the correlation length @xmath0 was obtained as @xmath9 where @xmath10 is the lattice constant , @xmath11 .
the prefactor @xmath12 is left undetermined by the rg calculation .
it follows from topological considerations that the order
parameter field of the haft allows for excitations of the form of @xmath13 vortices @xcite .
a numerical study of the classical haft @xcite has revealed that these vortices become abundant above a threshold temperature @xmath14 and that they unbind for @xmath15 , similarly as the @xmath16 vortices in the planar @xmath17 model above the kosterlitz
thouless ( kt ) transition temperature @xmath18 @xcite .
further evidence for this similarity beween the dissociation mechanism of the @xmath13 vortices and that of the @xmath16 vortices has been provided by kawamura and kikuchi @xcite . in recent work @xcite , we have studied the influence of the vortices on the partition function of the classical haft on the basis of the nl@xmath6 model . while a true kt
type phase
transition can be ruled out for the haft , our results suggest that for @xmath15 the vortices will affect the properties of the haft rather drastically . in particular , the disorder induced by the unbinding of the vortices can be expected to lead to a crossover from the t dependence eq .
( [ 1 ] ) of the correlation length in the low temperature regime @xmath19 to a kt type behavior @xmath20 in the high temperature regime .
it is the aim of the present paper to supplement our recent analytical study @xcite , which was based on a continuum description of the haft by a monte carlo ( mc ) simulation of the original lattice model .
a similar study has recently been published by southern and young @xcite . while we closely follow the method of these authors , our conclusions will be quite different from theirs . in the next section
, we first give a brief account of the technique we used .
subsequently we present the results for the correlation length @xmath0 and the antiferromagnetic susceptibility @xmath1 .
while these quantities are directly accessible to simulations in the high temperature regime , where the disordering effect of the vortices limits the range of the correlations , the key quantity to be computed in the low temperature regime is the spin stiffness @xcite . our numerical results for this last quantity will be presented and discussed in section iii .
finally , we summarize the evidence for a vortex induced crossover transition in section iv .
the classsical hamiltonian of the triangular heisenberg antiferromagnet can be defined as @xmath21 here , the @xmath22 are three dimensional unit vectors and the sum extends over all distinct pairs of nearest neighbor sites of a triangular lattice of @xmath23 sites .
the exchange constant has been set to unity .
the classical ground state of the hamiltonian eq .
( [ 3 ] ) is a coplanar arrangement in which the spins on the three sublattices are oriented at @xmath24 relative to each other , @xmath25 here , @xmath26 , @xmath27 are a pair of mutually orthogonal unit vectors and @xmath28 can be any one of the six vectors pointing towards the corners of the hexagonal brillouin zone of the triangular lattice , e.g. @xmath29 . the correlation length @xmath0 can be obtained assuming a ornstein zernicke form for the structure factor @xmath30 @xmath31 is then the susceptibility of the system at the ordering wavevector .
the spin correlations @xmath32 in eq .
( [ 5 ] ) can be determined by mc techniques .
we used the local algorithm described by kawamura and miyashita @xcite .
the lattice is divided into independent sublattices .
then , the spins of each of these sublattices are updated sequentially . for a given spin
, a new direction is chosen at random and the standard metropolis rule is used to decide whether the new direction is to be accepted . if it is discarded , a precessional motion through a randomly chosen angle about the direction of the local mean field is performed .
we apply this method to systems of linear sizes @xmath33 . for the smaller systems , @xmath34 , we discard the first @xmath35 sweeps for equilibration and average over the next @xmath36 sweeps . for @xmath37 , we average over @xmath38 sweeps after discarding the initial @xmath39 sweeps , and for @xmath40 , the average is over @xmath41 , and @xmath36 sweeps are discarded .
a selection of results for the correlation length @xmath0 and for the antiferromagnetic susceptibility @xmath1 which have been obtained by averaging over 3 - 5 independent runs of these lengths is tabulated in table [ tab_ksi_chi ] .
as will become apparent shortly , the data shown in this table are crucial in checks of theoretical predictions for the temperature dependence of @xmath0 and @xmath1
. to exhibit possible finite size effects , the table contains two pairs of data for each temperature which correspond to two different system sizes @xmath42 , @xmath43 . in general ,
@xmath0 and @xmath1 decrease with @xmath44 , but increase with the system size @xmath42 .
if , for a given temperature @xmath45 and a given system size @xmath46 , the data for @xmath0 and @xmath1 exhibit no size dependence upon doubling the system size , then one can conclude that the system size @xmath46 suffices to obtain size independent data for all @xmath47 .
data which are size independent by this criterion are marked by an asterisk in table [ tab_ksi_chi ] .
obviously , we can not exclude that the data for the lowest temperature @xmath48 obtained for the @xmath49 system are still size dependent . certainly , however , our data for @xmath48 are lower bounds to the thermodynamic limits of @xmath0 and @xmath1 at this temperature . to facilitate the comparison of our data with the rg predictions we include in table [ tab_ksi_chi ] the values for @xmath0 and @xmath1 which result from fits of the expressions eq .
( [ 1 ] ) and eq . (
[ 6 ] ) to these data @xcite . in figs .
[ ksi - t ] and [ chi - t ] , we show our complete sets of results for the correlation length and for the antiferromagnetic susceptibility as functions of the temperature .
obviously , for any given system size @xmath42 , there is an inflection point in the sequences of data for @xmath0 and @xmath1 .
this point defines a temperature @xmath50 below which both @xmath0 and @xmath1 begin to exhibit finite size effects . in fact , as can be seen in fig .
[ ksi - t ] , the correlation length increases linearly with the system size for sufficiently low temperatures @xmath51 . figs .
[ ksi - t ] and [ chi - t ] also contain fits of different theoretical predictions to the numerical data .
the dashed lines represent fits of the rg result , eq .
( [ 1 ] ) , to our data for @xmath0 and of the form @xmath52 proposed by southern and young @xcite on the basis of rg calculations , to our data for @xmath1 . in these rg predictions , the constants @xmath12 and @xmath53 are the only undetermined parameters . in our fits , we neglect the data points for temperatures @xmath54 which contain finite size effects . in agreement with southern and young
@xcite we find @xmath55 and @xmath56 .
the solid lines represent fits of the kt forms eq .
( [ 2 ] ) and @xmath57 to the data . with the kt form for @xmath0 , eq . ( [ 2 ] ) , the last expression follows from the general relation @xmath58 for the structure factor at the ordering wave vector , if @xmath59 is assumed to take the value @xmath60 as for a proper kt transition . for the threshold temperature ,
we use the value @xmath61 , which we can be inferred from the temperature dependence of the spin stiffness as will be discussed in the next section .
this leaves the constant @xmath62 which is common to the expressions eq .
( [ 2 ] ) and eq . ( [ 7 ] ) and the constant factors @xmath63 and @xmath64 of eq . ( [ 2 ] ) and eq . ( [ 7 ] ) as fit parameters . as before ,
we ignore data points for @xmath54 in our fits .
the results are @xmath65 , @xmath66 and @xmath67 .
obviously , for temperatures @xmath68 , the @xmath69-temperature dependence of the kt forms fits the data better than the @xmath70-temperature dependence predicted by the rg analysis . in figs .
[ ksi - wt ] and [ chi - wt ] we plot @xmath0 and @xmath1 logarithmically against @xmath71 so that the kt forms eq .
( [ 2 ] ) and eq . ( [ 7 ] ) appear as straight lines .
these lines are seen to fit the data quite well in the temperature interval @xmath72 , whereas the agreement between the curves representing the rg forms is restricted to a narrow interval around @xmath73 . in particular
, we emphasize that for @xmath48 , the rg predictions are incompatible with the data points for @xmath0 and @xmath1 which are listed in table [ tab_ksi_chi ] with their respective errors . in this context
, we recall that if our data for @xmath48 do not represent the thermodynamic limits of @xmath0 and @xmath1 , they are certainly lower bounds to these limits . hence , the discrepancy between the rg predictions and the true values of @xmath0 and @xmath1 may even be larger than has been inferred here .
we note that the fit of the kt form , eq .
( [ 7 ] ) , to the data for the susceptibility @xmath1 is better than that of the rg form , eq .
( [ 2 ] ) , to @xmath0 .
this may be attributable to the lower quality of the data for @xmath0 which are obtained indirectly from the ornstein - zernicke expression eq .
( [ 5 ] ) in the limit @xmath74 . in summary
, we observe that our results combined with the earlier findings of kawamura and miyashita @xcite support the view that in the temperature range @xmath15 the spin correlations of the haft are decisively influenced by unbound vortices , so that a perturbative treatment of the model is inadequate in this temperature regime
. it should be obvious , however , that in the above analysis of our numerical results we have been guided by our previous prediction @xcite that , in the case of the haft , the vortex unbinding mechnism leads to a kt - type temperature dependence of the correlation length above a crossover temperature @xmath75 . while we do not claim to have found compelling evidence for this prediction , we regard our numerical results as strong support for it .
to further corroborate the above view and in order to get insight into the low temperature regime , where the correlation length exceeds the accessible system sizes , we also determined the spin stiffness @xmath2 in our simulations .
the diagonal components @xmath76 , of the spin stiffness tensor are the second derivatives of the free energy density @xmath77 with respect to the twist angles @xmath78 of the spins around three mutually orthogonal axes @xmath79 @xcite , @xmath80 ( { \bf u } \cdot { \bf e}_{ij})^2 > \nonumber \\ & - & \frac{2}{\sqrt{3 } t l^2 } < \left [ \sum_{<i , j > } { \bf s}_i \times { \bf s}_j \cdot { \bf e}_{\alpha } ( { \bf u } \cdot { \bf e}_{ij } ) \right]^2 > \;\ ; . \label{8}\end{aligned}\ ] ] here , @xmath81 is the lattice direction along which the twist is applied and @xmath82 is the direction of the bond between nearest neighbor lattice sites @xmath83 and @xmath84 .
the prefactor in eq .
( [ 8 ] ) has been chosen such that eq .
( [ 8 ] ) is the stiffness per unit area . in the simulation , the thermal averages on the right hand side of eq .
( [ 8 ] ) are replaced by averages over configurations which are generated by a large number of successive mc sweeps . for a finite system
, the ordered spin structure will change its orientation in space in the course of the simulation . in a sufficiently large number of sweeps
, one will therefore measure the average spin stiffness @xmath85 since there is no long range order in the haft for any finite temperature , @xmath2 must vanish for the haft in the thermodynamic limit for any finite temperature .
however , for finite system sizes @xmath42 , @xmath2 will be finite for sufficiently low temperatures such that @xmath86 . according to eq .
( [ 1 ] ) , this condition should be satisfied for system sizes @xmath87 up to temperatures of the order of unity , unless the constant @xmath12 is exceedingly small as has been suggested by southern and young @xcite .
as we have argued above , the results for @xmath88 in the high temperature regime , @xmath15 , should not be interpreted as evidence for such a small value of @xmath12 .
in fact , a naive integration of the 2loop renormalization group equations which starts with the microscopic parameters of the haft as initial values yields the result eq .
( [ 1 ] ) with @xmath89 . in our simulations ,
we determine the three @xmath90 separately in each sweep and thus obtain the averages @xmath2 for each sweep . in fig .
[ rho - t ] , we show the average spin stiffness as a function of the temperature for system sizes @xmath91 .
the steep drop in @xmath2 which occurs as @xmath44 increases beyond @xmath5 is consistent with the rapid decrease of @xmath0 in the same temperature regime in which the vortices become unbound @xcite .
in contrast to the spin stiffness of the planar xy model , @xmath2 does not saturate in the low temperature regime @xmath19 with increasing system size @xmath42 but decreases with increasing @xmath42 .
this behavior is to be expected , since in contrast with the correlation length of the xy model , the correlation length of the haft remains finite for low temperatures , where the vortices are bound in pairs . as @xmath92 , our data for @xmath2 approach the correct limiting value @xmath93 . in fig .
[ rho - l ] , we display @xmath2 for various temperatures and system sizes .
the data are averages over 3 - 5 independent runs of lengths comparable to those which we have described above , error bars indicate the standard deviation . in the low temperature regime ,
the thermodynamics of the classical haft should be captured by the appropriate nl@xmath6 model @xcite . on the basis of a renormalization group treatment of this model , azaria _ et al .
_ @xcite have made detailed predictions for the dependence of the spin stiffness on the linear system size @xmath42 and the correlation length @xmath0 . in their mc study of the classical haft
, southern and young @xcite found excellent agreement with the predicted @xmath42 dependence of the spin stiffness tensor at the temperature @xmath94 . in order to be able to compare our numerical results with the predictions of the rg analysis of the nl@xmath6 model
, we integrated the two loop rg equations of azaria _ et al . _
@xcite starting from the initial conditions @xmath95 and @xmath96 . here ,
@xmath97 and @xmath98 are the two main components of the spin stiffness tensor with respect to the reference frame of the local order parameter @xcite . by the above initial conditions we identify @xmath97 and @xmath99 with their microscopic values on the scale of the lattice constant @xmath10 .
we find that in the temperature regime under consideration , @xmath100 , the average stiffness @xmath2 varies linearly with @xmath101 to a very good approximation on the scale @xmath102 covered by our simulations , see fig .
[ rho - lnksi ] below . from the rg equations
, one can also infer that the slope @xmath103 decreases from @xmath104 to @xmath105 , when @xmath42 increases from a value of the order of the lattice constant , @xmath106 , to a value of the order of the correlation length , @xmath107 . in fig .
[ rho - l ] , the straight lines are least squares fits to the data .
the slopes of these lines , normalized to the maximal theoretical slope @xmath104 , are tabulated in table [ gefaelle ] . for @xmath108 ,
the slopes are seen to be rather close to their maximal value which obtains , when @xmath42 is of the order of the lattice constant .
this is not unexpected since according to the rg calculations , the correlation length is many orders of magnitude larger than our maximal system size @xmath49 for these temperatures .
the larger values of @xmath109 which we find for @xmath110 are incompatible with the rg theory .
hence we conclude that for @xmath111 unbound vortices , not being taken into account by the rg analysis , begin to limit the range of the spin correlations in the haft .
if one defines the correlation length @xmath0 through the matching condition @xmath112 , then the result of the integration of the two loop rg equations can be cast into the following form @xmath113 this is shown for two different temperatures in the two graphs in fig .
[ rho - lnksi ] .
obviously , the function @xmath114 depends weakly on l. the relation eq .
( [ 10 ] ) makes it possible to obtain the correlation length from the monte carlo simulation , even in the low temperature regime where @xmath0 is much larger than the system size @xmath42 . inserting our mc data for @xmath115 into eq .
( [ 10 ] ) and solving for @xmath0 , we obtain the data points shown in fig .
[ reklame ] for @xmath116 .
this figure also includes the data for @xmath117 which have already been shown in fig .
[ ksi - t ] .
the dashed and solid lines in fig . [ reklame ] represent fits of the rg and kt forms , eqs .
( 1 ) and ( 2 ) , to the mc data for @xmath116 and @xmath117 , respectively .
we have shown that the mc simulation of the haft yields compelling evidence for the influence of vortices on the spin correlations and hence on the thermodynamics of this model . in agreement with earlier findings by kawamura and miyashita @xcite
we find that the disordering effect of the vortices sets in rather abruptly at a temperature @xmath5 . up to this temperature ,
our data for the spin stiffness @xmath2 and the ensuing temperature dependence of the correlation length @xmath0 are consistent with the predictions of the rg analysis of the haft which ignores the existence of topological defects such as vortices . for @xmath15 , however , the simulation yields temperature dependences of the correlation length @xmath0 and the antiferromagnetic susceptibility @xmath1 which are incompatible with the rg predictions . instead ,
in this temperature regime , the temperature dependences of @xmath0 and @xmath1 which follow from the kt picture of unbinding vortex pairs provide satisfactory fits of the data .
a rapid increase of the density of unbound vortices for @xmath118 had already been found by kawamura and miyashita in their simulation of the haft @xcite .
it had not been clear , however , whether this phenomenon would lead to the same temperature dependences of the correlations of the haft as had been predicted for the planar xy model by kosterlitz and thouless @xcite . in our recent analytical study @xcite ,
we were led to the conclusion that this should indeed be the case .
the present numerical study fully supports this conclusion . as we have discussed in ref .
@xcite , the crossover transition from the rg type behavior to the kt type behavior of the correlations of the haft can not imply a phase transition in the proper sense , because the correlation length is finite both below and above the transition temperature @xmath75 .
the results shown in fig .
[ reklame ] indicate , however , that the transition happens in a very narrow interval around @xmath75 .
therefore we consider it possible that the derivative of the correlation length with respect to the temperature develops a discontinuity at @xmath75 in the thermodynamic limit .
h. kadowaki , h. kikuchi , and y. ajiro , j. phys .
c : solid state phys .
* 2 * , 4485 ( 1990 ) ; k. takeda , k.miyake , k. takeda , and h. hirakawa , j. phys .
. jpn . * 61 * , 2156 ( 1992 ; n. kojima _
et al . _ , j. phys .
. jpn . * 62 * , 4137 ( 1993 ) .
.[tab_ksi_chi ] correlation length @xmath0 and antiferromagnetic susceptibility @xmath1 for system sizes @xmath119 , @xmath120 , and @xmath121 for various temperatures .
the statistical error represents the standard deviation over 3 - 5 independent runs .
the two last columns contain the rg predictions for @xmath0 and @xmath1 .
asterisks indicate agreement of the results for different system sizes , see main text . [ cols=">,>,>,>,>,>,>",options="header " , ] | we have simulated the classical heisenberg antiferromagnet on a triangular lattice using a local monte carlo algorithm .
the behavior of the correlation length @xmath0 , the susceptibility at the ordering wavevector @xmath1 , and the spin stiffness @xmath2 clearly reflects the existence of two temperature regimes a high temperature regime @xmath3 , in which the disordering effect of vortices is dominant , and a low temperature regime @xmath4 , where correlations are controlled by small amplitude spin fluctuations . as has previously been shown , in the last regime , the behavior of the above quantities agrees well with the predictions of a renormalization group treatment of the appropriate nonlinear sigma model . for @xmath3 , a satisfactory fit of the data
is achieved , if the temperature dependence of @xmath0 and @xmath1 is assumed to be of the form predicted by the kosterlitz thouless theory .
surprisingly , the crossover between the two regimes appears to happen in a very narrow temperature interval around @xmath5 .
+ pacs numbers : 75.10.hk , 75.50.ee , 75.40.mg |
the nonlinear electrodynamics was first introduced by born and infeld in 1934 to obtain finite energy density model for the electron @xcite . in recent years
nonlinear electrodynamics models are attracting much interest , too .
the reason is that the nonlinear electrodynamics arises naturally in open strings and @xmath0-branes @xcite-@xcite .
nonlinear electrodynamics models coupled to gravity have been discussed in different aspects ( see for example @xcite-@xcite and references therein ) . in the present work we consider stringy einstein - born - infeld - dilaton ( ebid ) gravity described by the action @xcite-@xcite @xmath1\ ] ] where @xmath2 is ricci scalar curvature with respect to the spacetime metric @xmath3 and @xmath4
is the dilaton field .
the born - infeld ( bi ) part of the action is given by @xmath5.\ ] ] here @xmath6 is the dual to the maxwell tensor and @xmath7 ( @xmath8 ) is the dilaton coupling constant .
in the context of the string theory , the ( bi ) parameter @xmath9 is related to the string tension @xmath10 by @xmath11 .
it should be noted that the ebid action does not posses an electric - magnetic duality .
that is why one should expect that the electrically and magnetically charged solutions will be different .
note that in the @xmath12 limit the action ( [ ebida ] ) reduces to einstein - maxwell - dilaton one .
unfortunately , the field equations yielded by the action ( [ ebida ] ) are too complicated and there are no exact analytical solutions in four or more dimensions .
exact solutions to the ebid equations are known only in three dimensions @xcite .
these solutions are non - asymptotically flat and describe three - dimensional black holes . in the present paper
we derive exact non - asymptotically flat and non-(a)ds black hole solutions to the four - dimensional ebid gravity .
such type solutions , which are non - charged or within the framework of linear electrodynamics have attracted much interest in recent years @xcite-@xcite .
here we consider only magnetically charged case for which @xmath13 and that is why we may restrict ourselves to the truncated bi lagrangian @xmath14.\ ] ] it is also convenient to set @xmath15 where @xmath16 the action ( [ ebida ] ) then yields the following field equations @xmath17g_{\mu\nu } , \\
\nabla_{\mu}\left[e^{-2\gamma\varphi } \partial_{x}{\cal l}_{bi}(x ) f^{\mu\nu}\right]= 0,\\ \nabla_{\mu}\nabla^{\mu}\varphi = 2b\gamma e^{2\gamma\varphi}\left[2x\partial_{x}{\cal l}_{bi}(x ) - { \cal l}_{bi}(x)\right],\end{aligned}\ ] ] where @xmath18 is the covariant derivative with respect to the spacetime metric @xmath3 .
the metric of the static and spherically symmetric spacetime can be written in the form @xmath19 the electromagnetic field is assumed to have the following pure magnetic form @xmath20 where @xmath21 is the magnetic charge . respectively , we obtain for @xmath22 : @xmath23 the field equations reduce to the following system of coupled ordinary differential equations @xmath24r^2 , \nonumber \\ -{1\over r}{d^2r\over dr^2 } & = & \left({d\varphi\over dr } \right)^2 , \\ 1 - { 1\over 2 } { d\over dr } \left(\lambda { dr^2\over dr } \right ) & = & - 2be^{2\gamma\varphi}{\cal l}_{bi}(x ) r^2 , \nonumber \\ { d\over dr } \left(r^2 \lambda { d\varphi\over dr } \right ) & = & 2\gamma be^{2\gamma\varphi } \left[2x\partial_{x}{\cal l}_{bi}(x ) - { \cal l}_{bi}(x)\right]r^2 . \nonumber\end{aligned}\ ] ]
the case @xmath25 is predicted from the ( super)string theory . in order to solve the field equations we make the ansatz @xmath26 where @xmath27 is a constant .
the second equation of ( [ ode ] ) then gives @xmath28 where @xmath29 and @xmath30 are constants .
the consistency condition for the third and the fourth equation of ( [ ode ] ) gives the following algebraic equation for @xmath31 : @xmath32 with @xmath33 . solving this equation with respect to @xmath31 we
obtain @xmath34 where @xmath35 .
therefore the magnetic charge must satisfy the inequality @xmath36 the existence of critical value @xmath37 for the magnetic charge is a pure nonlinear effect which disappears in the limit to linear electrodynamics when @xmath12 .
finally , for the metric function @xmath38 we find @xmath39 where @xmath40r^2_{0 } > 0\ ] ] and @xmath41 is a constant .
below we discuss the physical properties of the solution and , without loss of generality , we set @xmath42 .
the solution is not asymptotically flat and in order to define its mass we use the so - called quasilocal formalism @xcite .
the quasilocal mass is given by @xmath43\ ] ] where @xmath44 is an arbitrary non - negative function which determines the zero of the energy for a background spacetime .
if no cosmological horizon is present , the large @xmath45 limit of ( [ qlm ] ) determines the asymptotic mass @xmath46 . in our case
the natural choice is @xmath47 and we find @xmath48 we first consider solutions with positive mass , @xmath49 .
the kretschmann scalar is @xmath50 where @xmath51\nonumber .\end{aligned}\ ] ] the scalar @xmath52 is singular only for @xmath53 and tends to zero like @xmath54 for @xmath55 .
the solution has a regular horizon at @xmath56 which hides a spacelike singularity located at @xmath53 .
note that the dilaton field is regular on the horizon , too .
the spacial infinity is conformally null and the solution describes a black hole with the same causal structure as the schwarzschild spacetime .
the temperature and the entropy of the black hole are @xmath57 in order to write the first law we should take into account that the constants @xmath58,@xmath59 and @xmath31 are , in fact , related to the background , not to the black hole .
then we find @xmath60 where the variation is with respect to @xmath41 only .
the solution with zero mass ( @xmath61 ) is singular with a null singularity at @xmath53 .
the case @xmath62 corresponds to naked timelike singularities located at @xmath53 .
here we consider solutions with general dilaton coupling constant @xmath7 . as for the particular case @xmath25 we make the ansatz @xmath63 where @xmath27 is a constant
. substituting in the equations ( [ ode ] ) we obtain the following algebraic equation for @xmath31 @xmath64 where @xmath65 . in more explicit form the algebraic equation is @xmath66 where @xmath67 for sufficiently small @xmath68 we have @xmath69 and , therefore , a sufficient ( but not necessary ) condition for the algebraic eq .
( [ ae ] ) to have roots in the interval @xmath70 is @xmath71 . in general
, the function @xmath72 can have zeros also for @xmath73 and , in some cases , more than one zero as it is shown in the figure 1 for @xmath74 .
let us note that the different zeros of @xmath72 , in fact , correspond to different backgrounds rather to different black holes , as it is seen from the metric function @xmath38 presented below . for the metric functions we find
@xmath75 where @xmath30 and @xmath41 are constants and @xmath59 is given by @xmath76r^2_{0 } > 0 .\ ] ] the natural background is given by the metric function @xmath77 for the asymptotic mass we then obtain @xmath78 we first consider solutions with positive mass , @xmath49 .
the kretschmann scalar is given by ( [ ks ] ) where
@xmath79 \nonumber .\end{aligned}\ ] ] the kretschmann scalar is divergent only for @xmath53 and tends to zero like @xmath80 for @xmath55 .
the solution has a regular horizon at @xmath56 hiding a spacelike singularity located at @xmath53 . for @xmath81 ,
spacial infinity is conformally timelike and the causal structure is similar to that of the static btz black hole spacetime @xcite . for @xmath82 ,
spacial infinity is conformally null and the causal structure is just the same as for the schwarzschild spacetime .
the temperature and the entropy of the black holes are given by @xmath83 the first law is written in the form @xmath84 where the variation is with respect to @xmath41 only .
the solutions with zero mass are singular with null singularities .
the case with negative mass is also singular with null singularities for @xmath85 and timelike singularities for @xmath86 .
the stability of the black holes is an important question from physical point of view .
it is well known that there are many black holes solutions which are unstable . here
we show that our black hole solutions are stable against linear radial perturbations . in order to discuss the stability we take the spacetime metric in the form @xmath87 where the functions @xmath88 , @xmath89 and @xmath90 depend on @xmath45 and @xmath91 .
we assume that the metric functions and the dilaton are small perturbations of the static background @xmath92 the convenient gauge is @xmath93 ( i.e. @xmath94 ) .
the electromagnetic field is given by ( [ mfmf ] ) which solves the electromagnetic equations for the time dependent metric ( [ tdm ] ) , too .
the linearized equations for @xmath95 and @xmath96 give @xmath97 @xmath98 \delta\chi - { 1\over 4}\lambda(r)e^{\beta(r ) } \partial_{r}\beta(r ) \left(\partial_{r}\delta\gamma - \partial_{r}\delta\chi\right ) \nonumber \\
= 4abe^{2\gamma\varphi(r)}\left[2x\partial_{x}{\cal l}_{bi}(x ) - { \cal l}_{bi}(x)\right ] e^{\beta}\delta\varphi,\end{aligned}\ ] ] where @xmath99 is the ricci tensor with respect to the static background .
the linearized equation for the dilaton is @xmath100\delta\chi + { 1\over 2}\lambda(r)\partial_{r}\varphi(r )
\left(\partial_{r}\delta\gamma - \partial_{r}\delta\chi \right ) = \nonumber \\ - 4\gamma^2 be^{2\gamma\varphi(r ) } \left[{\cal l}_{bi}(x ) + 4x^2\partial^2_{x}{\cal l}_{bi}(x ) \right]\delta\varphi\end{aligned}\ ] ] where @xmath101 is the coderivative operator with respect to the static background . integrating the equation ( [ pr01 ] )
we obtain @xmath102 eliminating the perturbations @xmath103 and @xmath104 between eqs .
( [ pr22 ] ) , ( [ ccf ] ) and ( [ dper ] ) we find @xmath105 where @xmath106 \right ] .\ ] ] taking into account the explicit form of the bi lagrangian it can be shown that @xmath107 . introducing the new radial coordinate @xmath108 the eq.([gpe ] ) can be written in the explicit form @xmath109 for growing modes @xmath110 we find @xmath111 \delta\varphi(r_{*})=0.\ ] ] for real @xmath112 the effective potential @xmath113 $ ] is positively defined for @xmath114
. therefore , there are no bounded solutions for @xmath115 and we conclude that the black holes are stable against linear radial perturbations .
in this paper we derived exact , magnetically charged , static and spherically symmetric black hole solutions to the einstein - born - infeld - dilaton gravity .
these solutions are neither asymptotically flat nor ( anti)-de sitter . some basic properties of the solutions were discussed .
it was shown that the black holes are stable against linear radial perturbations .
it is worth noting that the black solutions derived in the present paper are solutions not only of the bi electrodynamics but also of general nonlinear electrodynamics described by an arbitrary function @xmath116 provided the corresponding algebraic equations possess roots and the corresponding algebraic inequalities are satisfied . in particular , in the case of the linear electrodynamics with @xmath117the method presented here gives the well - known non - asymptotically flat and non-(a)ds black hole solutions of einstein - maxwell - dilaton ( emd ) gravity @xcite .
moreover , for @xmath117 we have @xmath107 and therefore , the emd black holes are stable was proven in @xcite . ] against linear radial perturbations for arbitrary dilaton coupling constant @xmath7 .
it is puzzling that the nonlinear electrodynamics equations can be solved for arbitrary function @xmath116 .
the cause is that we consider the sector where the theory looses a part of its nonlinearity .
more precisely , the electromagnetic nonlinearity of the differential equations is transformed into algebraic nonlinearity .
this can be explicitly demonstrated as follows . for spherically symmetric magnetic configurations
the ansatz ( [ ansatz1 ] ) and ( [ ansatz2 ] ) are equivalent to consider @xmath22 satisfies @xmath118 the first consequence following from this fact is that the nonlinear maxwell equations ( [ fe ] ) become linear since the nonlinear part can be pulled out of the equations . in second place , redefining the dilaton field as @xmath119 where @xmath120\over x\right)\right|_{x = x_{0}},\ ] ] the dilaton equation becomes the one of the linear emd case .
the only difference appears in the einstein equations , which reduce to @xmath121 where @xmath122 in this way we obtained field equations which are linear in the electromagnetic field .
note however , although linear in the electromagnetic field , these equations , in general , are not the emd equations since @xmath123 for general nonlinear electrodynamics .
for example , in the case of the born - infeld electrodynamics we have @xmath123 for any finite value of @xmath124 .
summarizing , we have shown that the all information about the nonlinearity is encoded in the parameter @xmath125 appearing in the einstein equations and the nonlinear algebraic constraint ( [ nc ] ) .
it can be seen from the exact solutions presented in the previous sections , that the solutions of the `` k - deformed '' emd equations are quite similar to those of the pure emd equations . the parameter @xmath125 influences the background constants in the solutions .
in contrast , the nonlinear constraint ( [ nc ] ) gives severe physical restrictions on the black holes charge .
these restrictions algebraically reflect the nonlinearity of the electromagnetic field .
this can be explicitly seen from the solutions for the born - infeld electrodynamics with @xmath25 where the constraint ( [ nc ] ) gives the existence condition @xmath126 .
i would like to thank a. donkov for reading the manuscript .
this work was partially supported by the bulgarian national science fund under grant mu-408 .
m. born and l. infeld , proc .
r. soc .
london * a143 * , 410 ( 1934 ) . e. fradkin and a. tseytlin , phys . lett . *
b163 * , 123 ( 1985 ) .
r. matsaev , m. rahmanov and a. tseytlin , phys .
* b193 * , 205 ( 1987 ) .
e. bergshoeff , e. sezgin , c. pope and p. townsend , phys .
* b188 * , 70 ( 1987 ) . c. callan , c. lovelace , c. nappi and s. yost , nucl .
* b308 * , 221 ( 1988 ) . o. andreev and a. tseytlin , nucl .
* b311 * , 221 ( 1988 ) .
r. leigh , mod .
. * a4 * , 2767 ( 1989 ) .
m. demianski , found .
* 16 * , 187 ( 1986 ) .
d. wiltshire , phys .
rev . * d38 * , 244500 ( 1988 ) .
h. doliveira , class .
quantum grav .
* 11 * , 1469 ( 1994 ) .
g. gibbons and d. rasheed , nucl .
* b454 * , 185 ( 1995 ) . g. gibbons and d. rasheed , phys .
* b476 * , 515 ( 1996 ) .
e. ayon - beato and a. garcia , phys .
lett . * 80 * , 5056 ( 1998 ) .
e. ayon - beato and a. garcia , gen .
relativ .
grav . * 31 * , 629 ( 1999 ) .
e. ayon - beato and a. garcia , phys .
b464 * , 25 ( 1999 ) . t. tamaki and t. torii , phys .
* d62 * , 061501r ( 2000 ) . g. clement and d. galtsov , phys .
* d62 * , 124013 ( 2000 ) . t. tamaki and t. torii , phys .
* d64 * , 024027 ( 2001 ) . g. gibbons and c. herdeiro , class .
quantum grav .
* 18 * , 1677 ( 2001 ) .
r. yamazaki and d. ida , phys .
* d64 * , 024009 ( 2001 ) . s. yazadjiev , p. fiziev , t. boyadjiev and m. todorov , mod .
* a16 * , 2143 ( 2001 ) . m. gurses and o. sarioglu , class . quantum grav . *
20 * , 351 ( 2003 ) .
t. dey , phys .
* b595 * , 484 ( 2004 ) .
r. cai , d. pang and a. wang , rev .
* d70 * , 124034 ( 2004 ) . s. mignemi and d. wiltshire , class .
quantum grav .
* 6 * , 987 ( 1989 ) .
d. wiltshire , phys .
* d44 * , 1100 ( 1991 ) . s. mignemi and d. wiltshire , phys .
* d46 * , 1475 ( 1992 ) . s. poletti and d. wiltshire , phys .
rev . * d50 * , 7260 ( 1994 ) ; * d52 * , 3753(e ) ( 1995 ) .
k. chan , j. horne and r. mann , nucl .
* b447 * , 441 ( 1995 ) .
r. cai , j. ji and k. soh , phys .
* d57 * , 6547 ( 1998 ) .
r. cai and y. zhang , phys .
* d54 * , 4891 ( 1996 ) . g. clement , d. galtsov and c. leygnac , phys .
rev . * d67 * , 024012 ( 2003 ) .
g. clement and c. leygnac , phys . rev .
* d70 * , 084018 ( 2004 ) .
r. cai and a. wang , phys .
rev . * d70 * , 084042 ( 2004 ) . s. yazadjiev , _ non - asymptotically flat , non - ds / ads dyonic black holes in dilaton gravity _ , gr - qc/0502024 j. brown and j. york , phys .
rev . * d47 * , 1407 ( 1993 ) .
m. banados , m. henneaux and h. zanelli , phys .
* d48 * , 1506 ( 1993 ) . | we derive exact magnetically charged , static and spherically symmetric black hole solutions of the four - dimensional einstein - born - infeld - dilaton gravity .
these solutions are neither asymptotically flat nor ( anti)-de sitter . the properties of the solutions are discussed .
it is shown that the black holes are stable against linear radial perturbations . |
referring to menzel @xcite who confused radiance and irradiance , many astrophysicists believe that coherent interactions are negligible in astrophysical systems .
the best physicists have long had similar ruling , stating for example : townes maser will not work .
while the development of experiments using lasers has led spectroscopists to study the coherent interactions between light and matter , astrophysicists have kept their wrong methods , especially using monte - carlo calculations apparently inspired by nuclear physics .
deducting mainly results from thermodynamics , planck and einstein have obtained very general laws , independent for instance of the use of a classical or quantum theories of light .
a unique formula represents , according to the sign of the amplification coefficient of the radiance of a light beam , absorption and coherent amplification .
section [ optics ] recalls general concepts of coherent optics .
this is not useless , recalling , for instance how planck s and einstein s theories must replace the monte - carlo method to compute the propagation of light in a resonant medium , and that some interpretations of photon - counting signals are wrong .
section [ utile ] describes optical effects well known in laser spectroscopy , which seem apply easily in astrophysics .
it shows that usual monte - carlo computations are wrong in the low density media of astrophysics .
section [ possible ] proposes to apply these optical effects to find simple explanations of many observations .
optical modes are needed to study classical and quantum theories of light .
maxwell s equations are linear in vacuum , so any linear combination of solutions is a solution .
thus , solutions are represented by points in a vector space @xmath1 .
we call mode a set of solutions that differ only by a real multiplicative factor , that is a radius of @xmath1
. the norm of an electromagnetic field in a space _ empty of any other field _ is the energy of this field .
two fields ( consequently two modes ) are orthogonal if the norm of their sum is the sum of the norms of each field . for theory ,
matter is not introduced by permeability and permittivity , but by the elementary sources ( electrons , nucleus ) which , following schwarzschild and fokker are replaced by their advanced fields , so that the field equations remain linear and sets of modes may be defined .
an usual type of mode is defined from a nearly monochromatic , polarized electromagnetic beam limited by two circular holes and diffraction .
its flux of energy by unit of frequency is its spectral radiance .
the modes deduced from the general definition may be extremely complex : their frequency band is not necessarily sharp , they may be polychromatic ; their geometry may be complex , for instance tortuous in opalescent media .
thus , some useful general concepts , for instance radiance , may require new definitions .
we will remind the general concepts using usual modes , else results remain valid while it may be difficult to set their equations .
light may diverge far from a source into a very large region .
thus coherent polarization by a single - mode light beam may concern all atoms in a large volume .
these atoms make a single polarized quantum system which may interact with a single photon . in this large quantum system , in a first approximation close to the classical interpretation , any atom is considered as dressed by the field , its stationary state being slightly mixed with its other states .
polarization can absorb a significant energy to be returned to one or several exciting wave(s ) when light disappears because the atoms must return to their original state .
einstein introduced by thermodynamics the theory of propagation of light in a resonant medium , theory which uses a and b coefficients @xcite . in a large medium , homogeneous enough to use standard optics , the coherent amplification corresponding to b may be easily connected to huygens theory of the propagation of monochromatic waves : huygens construction uses wavelets emitted by all points of a wave surface @xmath2 to generate an other wave surface @xmath3 , envelope of the wavelets .
the retrograde wave is eliminated by the radiation of a wave surface whose distance from @xmath2 is a fourth of wave .
suppose that light propagates in a gas such that the number of molecules interacting in an homogeneous region is large .
wavelets generated by molecules lying on @xmath2 , coherently ( that is with an identical phase ) , at the same frequency ( rayleigh scattering ) generate also @xmath3 .
thus , the amplitudes of the incident and scattered waves may be added , the result being a single wave , with the same frequency . for simpler computations ,
the scattered field is split into two fields : - a field delayed by @xmath4 which produces the refraction ; - a field having the same or an opposite phase which produces a positive or negative amplification of the incident wave .
the scattered amplitude being supposed proportional to the incident amplitude , the amplification of the amplitude ( or of the radiance ) is proportional to the incident amplitude ( or radiance ) .
refraction changes the geometry of the modes while amplification does not .
einstein s theory , restricted to coherent amplification of a light beam propagating in an homogeneous medium , and huygens theory have the same limits of validity based on statistics : between two wave surfaces distant of a wavelength , matter must be homogeneous and the number of molecules having the same properties must be large enough to apply the law of large numbers .
einstein theory applies to all modes .
choosing a particular electric charge , it applies to the advanced and retarded modes of absorption and emission of this charge .
the theory of amplification may become : the ratio of the amplitudes of the advanced and retarded fields is a real number . to introduce also refraction of quasi - monochromatic light , this number is set complex .
the energy exchanged with matter is quantized . in classical and semi - classical representations ,
the energy in a mode may vary continuously . in quantum representation of light ,
the modes have the energies of a quantum oscillator , so that the energy of a mode jumps between levels having exactly energies @xmath5 , with n integer .
atomic theory shows that the distances between elementary charges ( electrons , nuclei ) are much larger than the size of these charges , so the field emitted by a charge @xmath6 in its specific mode is much larger in its neighborhood than the fields created by other charges .
an opposite field is needed to absorb the field emitted by @xmath6 .
the generation of this opposite field , in particular in the vicinity of @xmath6 , requires fields created by numerous other charges : the absorption of all fields is impossible , beyond an absorber it remains a _
residual spectral radiance _ whose mean value @xmath7 in a monomode beam results from the computation of absolute energies needed , at least theoretically , to apply formula @xmath8 @xcite .
consequently the absolute spectral radiance of a monomode beam in a blackbody is @xmath9\}$ ] .
subtracting @xmath7 from an absolute field gives the usual , particular relative field set equal to 0 in the dark .
planck s formula allows to associate , outside of a blackbody , a radiant planck s temperature @xmath10 to a region of radiance @xmath11 of a monomode beam .
_ radiance is the fundamental parameter characterizing the flux of energy in a progressive light mode .
spectral radiance allows to associate a temperature to this flux , then to compute entropy .
_ using the absolute field , einstein s a coefficient is null , so that the theory is simple , therefore reliable : a monomode beam coming from an unknown dark source , has an unknown , therefore stochastic phase , and its spectral radiance is , in the average , @xmath7 .
clausius showed that , in a transparent medium , the radiance is constant along a beam ; in matter , the amplification of radiance depends on @xmath12 coefficient .
let @xmath13 the absolute field resulting from amplification of a residual field @xmath14 , of _ average _ radiance @xmath7 and @xmath15 , the usual relative field .
an opaque photocell absorbs a quantified flux of energy whose mean value @xmath16 is proportional to @xmath17 because it remains the field @xmath14 in the mode . using a relative radiance ,
it is often written that @xmath16 is proportional to @xmath18 , even in photon counting while @xmath19 is predominant .
if two monomode beams are non orthogonal to an initial monomode beam , fluctuations of @xmath14 in the initial beam are involved in the observed beams where they are correlated ( photon bunching ) .
observations of these correlated fluctuations may come from a space cutting of light of a star by a telescope or from a time cutting .
the name monte - carlo means that a random variable is introduced in a calculation , then that the results are averaged .
these computations can be useful when light propagates in a medium too inhomogeneous ( eg opalescent near the critical point of a gas ) for surface wave of radiation can be defined .
but the need of this introduction is necessary only because a too complex system can not be fully known : einstein theory rules fundamentally the light - matter interactions . at medium pressure ,
density fluctuations are responsible for scatterings that are incoherent , for example , the blue sky .
these scatterings are induced by the residual fields of the emission modes of the fluctuations and take their phases .
these phases are usually unknown , but photon bunchings are possible . at low pressure , the fluctuations are produced mainly by binary collisions whose density is proportional to the square of the pressure .
while thermodynamics provides only average values , its use is essential to the interpretation of macroscopic observations .
this involves using radiance of light to get its temperature . in the interstellar medium ,
the gas density is too low for collisions produce a significant incoherent scattering .
thus , using a monte carlo computation is not justified .
boltzman temperature @xmath20 of a set of atoms for a transition of frequency @xmath21 corresponds to an equality between @xmath22 and the ratio of populations of atoms in the upper and lower levels of the transition ( with this definition , @xmath20 may be negative ) .
a positive amplification of a single mode beam can produce a large radiance , if @xmath23 up to a limit corresponding to @xmath24 , without limit if @xmath25 .
if the radiance becomes larger than an arbitrary limit , for instance @xmath26 , the ray is called superradiant .
the gas being isotropic , if the temperatures @xmath27 of two beams at their intersection are much lower than @xmath28 , or if @xmath29 , their amplification coefficients have nearly the same value .
since the exchanged energy is proportional to the incident energy , the more powerful beam exchanges more energy . for a positive amplification , it can absorb the bulk of available energy : this competition of modes leaves only a few modes get a high radiance . dressed by several rays of light , atoms can undergo a series of transitions .
if the radiances of the rays are large , the atoms pass directly from their initial state to their final state : the interaction is called multiphotonic .
few intermediate states may be non - stationary , virtual .
a set of multiphotonic interactions that returns the atoms to their original states , is a parametric interaction , where the atoms act as a catalyst .
it is obviously an increase in the entropy of light rays , computed from planck s law .
several types of atoms can participate in a parametric effect which combines the frequencies , but each atom must complete a full cycle of transitions .
the parametric interactions are especially intense because atoms do not need to acquire an energy that would enable them to reach another stationary state : their initial stationary state is only slightly mixed with other states during polarization .
the coherent rayleigh scattering that produces the refraction can be replaced by a raman scattering if the interaction remains parametric and if the conditions of coherence written by g. l. lamb @xcite are satisfied : the time - coherence ( the length of the pulses which model the time - incoherent light ) must be shorter than all implied time constants .
these constants are the period of a raman resonance and the collisional time . with lamb s conditions , the spectral lines of exciting and scattered beams overlap .
an elementary calculation shows that , for an interaction in an infinitesimal layer of gas , these beams interfere into a single monochromatic beam whose frequency is intermediate , in proportion to the amplitudes .
the infinitesimal shifts of frequency add along the path of the ray . to obtain a parametric interaction ,
the energy transferred to matter by the raman scattering must be compensated by at least an other simultaneous , coherent raman scattering .
the impulsive stimulated raman scattering uses two beams of femtosecond , powerful laser pulses .
it is a nonlinear interaction easily observed in the laboratory @xcite . by using a single lower power , pulsed laser ,
the experiment requires a much larger length of interaction .
the effect appears in long optical fibers used in telecommunications : the pulse frequency is reduced proportionally to the length of the fiber , energy is transferred to thermal radiation .
this effect complicates the multiplexing of information transmitted at several colors by long fibers .
the frequency shift is , in a first approximation , inversely proportional to the cube of the length of pulses @xcite .
natural light is made of pulses of about 1 nanosecond , @xmath30 times longer than the femtosecond laser pulses .
thus , the observation of a frequency shift by a coherent raman effect on incoherent light ( creil ) requires an interaction length @xmath31 times longer than the length of used optical fibers , length not available in a laboratory . with natural light ,
lamb s conditions impose a collisional time larger than 1 nanosecond , so the use of a low pressure gas . in this gas
, there must be a quadrupolar resonance whose period is larger than 1 ns : the resonance at 1420 mhz of the hydrogen atom is not appropriate , but the resonance frequencies 178 mhz in the 2s@xmath32 state , 59 mhz in 2p@xmath32 state , and 24 mhz in 2p@xmath33 are suitable .
strmgren @xcite studied a system composed of an extremely hot source ( temperature larger than @xmath34 ) immersed in a cloud of hydrogen at low pressure , initially cold .
he showed that where the temperature becomes low enough that neutral hydrogen atoms are formed , their intense radiation cools the gas so that it forms a relatively thin shell ( strmgren shell ) of neutral excited atomic hydrogen surrounding an almost completely ionized sphere ( strmgren sphere ) .
the intense superradiance was little known in 1939 , so strmgren did not introduce a competition of modes . let us study the amplification of rays that pass through the system at a distance @xmath35 from the source .
two rays l@xmath36 and l@xmath37 passing at weak distances @xmath38 and @xmath39 from the source cross twice the shell .
shell can be divided into infinitesimal concentric shells .
if @xmath40 , the path of radius l@xmath36 in each infinitesimal shell is shorter than the path of l@xmath37 , thus l@xmath36 is less amplified than l@xmath37 , amplification is thus an increasing function of @xmath35 .
as the amplification is zero if the ray does not penetrate the shell , the amplification function has at least a maximum value for a value @xmath41 of r ( the minimal r is chosen if there are several maxima ) .
the brightest rays are tangent to the sphere of radius @xmath41 . into a given direction , these rays make a cylinder and the system is seen as a ring .
if the superradiance is very intense , there is a competition of the modes on the cylinder , if a spectral line is selected the ring is dotted .
it has the appearance of tem@xmath42 modes of a laser selected for a fixed radial order @xmath43 and the @xmath44 values of angular index @xmath45 .
the rings may appear for several transitions , but these rings and dots are not independent : if a superradiance is very intense , for atoms located on the path of the beam , the upper state of the transition is depleted while the lower is more populated .
this increases the intensity of an other transition which arrives at the upper state , or starts at the lower .
the dots at long wavelengths are large , the dots at shorter wavelengths appear at their inner rim .
this explanation is simpler than the usual explanations of the pearl necklaces of snr1987a , of einstein cross or arcs . in the standard explanation of pearl necklace snr1987a , a shock wave created in h@xmath0 by the explosion of the star hits a ring of relatively dense h@xmath46 that is excited by the collision @xcite .
the propagation time of the shock wave explains that the ring does not appear immediately after the explosion , but the persistence of its brightness is amazing . for much of the energy of the shock wave is not lost , the shock wave must focus on the sources of the rings .
sugerman et al .
@xcite observed by photon echoes , before the onset of the rings , a scattering hourglass of denser matter .
it seems that the rays that form the rings of snr1987a are tangential ( rather slightly secant ) to the hourglass , into our direction , so that a long path in the excited gas close to the surface shows the rings .
burrows et al.@xcite criticized this model , showing that the variation of brightness close to the limb is too slow and concluding : a proper understanding of this system will require new physical insight .
supposing that the hourglass is a distorted strmgren shell , superradiance solves the problem .
the spectrum of the main ring is made of sharp emission lines , with lyman @xmath47 particularly intense .
light from a point on the ring propagates along a column of gas that can be cold , but excited by the hydrogen lines .
this column of gas can be superradiant at its eigenfrequencies , in the same direction . in a strmgren system ,
the temperatures of light beams have four orders of magnitude : \i ) the radial rays , emitted by the star have , at all frequencies , an initial temperature larger than @xmath34 .
\ii ) the superradiant rays considered in the previous subsection are much less hot .
\iii ) spontaneously emitted and not notably amplified lines , in particular lyman lines of atomic hydrogen have lower radiances .
\iv ) the thermal background has the coldest temperatures .
the temperature of the radial rays allows multiphotonic interactions including several multiphoton transitions , even passing through virtual levels .
thus , frequencies may be combined to cool _ at all frequencies _ the temperature of these rays by a pumping of hydrogen atoms to high levels , including ionization levels .
the superradiant rays are hot enough to induce a strong de - excitation of the atoms . pumping and de - excitations
are combined into a parametric effect , a multiphotonic , parametric , induced scattering which transfers most of the energy of the continuous spectrum of radial , extremely hot rays to much cooler energy of the discrete spectrum of hydrogen in the superradiant rays .
this parametric multiphotonic scattering is strong one hand because it corresponds to a large increase of entropy , second because all involved rays are very hot .
the order of magnitude of temperature of radial rays decreases to the order of magnitude of the amplified superradiant rays .
suppose that the order of magnitude of the size of the star is an astronomical unit while the order of magnitude of the source of the superradiant beams is a fraction of light - year .
in despite of the addition of all spectral radiances of the star , its solid angle of observation is so lower than the solid angle of observation of a superradiant dot , that the star is invisible .
the standard theory of snr1987a is unable to explain the total disappearance of the star . the coherent transfer of almost all the energy emitted by the star to the rings solves the problem of their persistent brightness .
as the observation of this small effect requires large column densities of gas , the contributions of gases other than hydrogen seems negligible .
hydrogen must be in an excited state , in practice in the 2s or 2p states because atomic quadrupolar transitions frequencies in more excited states are low and thus provide a priori negligible effects .
this excited hydrogen may result from a thermal heating , from a lyman @xmath47 pumping of too cold ( 10 000 - 40 000k ) atomic hydrogen , or from a cooling of an hydrogen plasma .
the result of the parametric effect catalyzed by the atomic excited hydrogen is frequently a transfer of energy from light which is redshifted to the background thermal radiation which is blueshifted , that is heated .
4.3.1 the frequency shifts of a particular ray depends on energies transferred from or to all crossing beams , that is , assuming a statistical isotropy of the transfer , on their spectral irradiances .
far from a star , the irradiance of its rays is low , we can neglect their contribution to a blueshift .
( a ) and larger ( b ) ; ( c ) experimental and ( d ) theoretical from michael et al .
@xmath48 is ly@xmath49 wavelength.,height=8 ] in the outer regions of a strmgren sphere , it appears excited atoms whose increasing density has an exponential bearing until @xmath35 reaches value @xmath41 .
a spontaneous emission very close to the surface of the sphere is relatively intense , and light outputs the sphere without encountering atoms able to catalyze a redshift .
intensity emitted per unit volume decreases rapidly with depth @xmath50 and becomes negligible .
the column density of excited atoms , thus the redshift of light , grows with @xmath51 , so that the emission volume corresponding to a wavelength shift @xmath52 grows as a function of @xmath51 .
the corresponding compensation of the variation of intensity is not sufficient to prevent its fall to zero on the spectrum ( fig.[rad1]a ) . for @xmath35 slightly larger than @xmath41 , a lot of energy
is transferred from the radial beams to the tangential superradiant beams .
thus , the radial speed of propagation of radiant energy becomes very low , the irradiance of hot beams becomes very large .
the cooler spontaneous beams , are blueshifted ( fig.[rad1]b ) , but along a short path because hydrogen is de - excited by the superradiant beams .
the spectrum observed in the ring of snr 1987a ( fig.[rad1]c ) integrates spectra corresponding to various incidences , so that the break of intensity at short wavelengths is not sharp
. it does not present a high spike which seems cut on the theoretical spectrum obtained from monte - carlo computations ( fig.[rad1]d ) .
the spectrum ( fig.[rad1]a ) is similar to the spectrum of a lyman break galaxy : these galaxies could be relatively close stars surrounded by clouds of hydrogen .
the very hot stars , and stars observed close to very hot objects have often anomalously large redshifts .
between 10 000 k and 40 000 k , hydrogen is hot enough for an atomic state , but too cold for a thermal excitation .
the propagation of far uv , x continuous spectrum light , in atomic hydrogen is remarkable : light is absorbed at ly@xmath49 frequency , 2p hydrogen is produced which catalyzes a transfer of energy to the background , producing a redshift .
an equilibrium is reached in which absorption compensates various de - excitations of the atoms .
there are two cases : \i ) the column density of excited hydrogen resulting from ly@xmath49 absorption along a path @xmath53 would shift the frequencies of light of more than the linewidth of ly@xmath49 line .
the absorption is limited by the redshift to a value @xmath54 .
as long as the initial radiance is larger or equal to @xmath54 , the redshift is permanent .
absorptions and emissions of all lines of the gas follow the redshift .
the linewidths corresponding to the redshift are large , so that the lines are not observable .
\ii ) if the radiance is lower than @xmath54 , the equilibrium @xmath24 is reached without a redshift sufficient to shift an unabsorbed spectral element to ly@xmath49 frequency .
the redshift stops , all lines of the gas are visibly absorbed or emitted . if the low radiance results from a previous absorption of a line , a radiance larger than @xmath54 may be reached after a slow redshift resulting from the excitation of higher states , more probably from a decay of these states which populates the 2p state and , mainly , the metastable 2s state .
several beams may pump the atoms , so that @xmath54 becomes an irradiance . during a stop of the redshift ( case ii )
, the lyman @xmath55 and @xmath56 lines are absorbed making future previous absorptions , so that the rays are multiplied with redshifts @xmath57 and @xmath58 . the remarkable constant 0,061 was observed par burbidge and karlsson @xcite in the lyman forests of the quasars .
the accreting neutron stars have never been observed .
we searched the spectrum of a model consisting of a tiny , very hot source having a strmgren shell so small that its pressure is of the order of 1 000 pa .
the spectrum shows , with decreasing frequency shifts , sharp , superradiant emission lines of the shell , large and saturated lines , then the lyman forest @xcite : some accreting neutron stars may be named quasars . at least a fraction of the cosmological redshift results from a creil effect catalyzed by traces of atomic hydrogen excited by uv radiation in the interstellar and intergalactic regions .
these regions are split into smaller regions where , depending on whether the uv irradiance is , before ly@xmath49 absorption , higher or lower than @xmath54 , light is very weakly or more strongly redshifted . through the regions of high redshift ,
the distance to earth is over - evaluated so that the maps of galaxies are popped : the inner part of a cellulose sponge is surprisingly similar to our maps of galaxies .
peter & judge @xcite have studied the frequency shifts of the uv lines emitted by the sun , using spectra recorded by the solar ultraviolet measurement of emitted radiation ( sumer ) instrument on soho .
they assume that vertical motions of the solar material produce these shifts by a doppler effect .
thus , they assume that the shift is null at the rim of the solar disk , so that they use the observations at the rim to calibrate the spectrometer .
unhappily , they find wavelengths which differ from the laboratory values . the hottest line , fe @xmath59 is blueshifted . in despite of a high temperature , high pressure atomic hydrogen works as a crystal in which the atoms have sharp lines , so that a creil effect is possible .
suppose that the frequency shifts result mainly from creil .
the emission of the uv lines occurs mainly where the phonons have the energy of emitted quanta , deeper for shorter wavelength .
a ray emitted at a certain frequency , then at a certain depth , crosses to the surface a longer distance at the limb than at the center of the disc : thus the shift is larger at the limb than at the center .
it is the opposite of the hypothesis of peter & judge .
creil effect explains the frequency shifts using laboratory frequencies@xcite : at the center of the disk : the shortest wavelength lines fe@xmath59 , mg@xmath60 , .. ) lose energy along their whole path where they cross colder emissions , and background radiation ; they are redshifted .
increasing the wavelength , the colder lines lose less energy along a shorter path , and may receive energy from fe@xmath59 line , for instance , so that their redshift decreases down to zero .
then the redshift is replaced by a blueshift ; o@xmath61 , n@xmath61 lines receive enough energy from the hot lines to be blue - shifted .
the frequency of the longest wavelength line ( he@xmath62 ) is not shifted because hydrogen atoms are not excited along its paths to the surface .
the solar wind cools while it expands , so that neutral , excited hydrogen atoms appear beyond 5 astronomical units . a creil effect transfers energy from the solar light whose temperature is of the order of 5000 k and which is redshifted to colder beams : when pioneer 10 and 11 go beyond 5 au , the very weak radio signals that they exchange with the earth are blue - shifted ( anderson & laing @xcite ) . if the increases of frequency are attributed to a doppler effect , rather than to a creil effect , the probes appear to have an acceleration which can be interpreted only by changing the laws of gravitation , yet well proven .
the very cold background is blue - shifted , that is heated @xcite .
this heating may explain that the low order terms of the spherical development of the cosmological background are bound to the ecliptic .
the geometrical propagation of light in homogeneous media depends only on the indices of refraction .
the variation of the radiance of a monochromatic light beam depends on an amplification coefficient whose sign is the sign of the difference @xmath63 between the boltzman temperature of the involved transition and the planck temperature of the beam .
the usual monte - carlo computations which do not take the phases of the photon pilot waves into account , therefore ignore interferences , are wrong .
they must be replaced by theories using coherent interactions initiated by einstein and developped for laser spectroscopy .
using the absolute radiance whose value is @xmath7 in a blackbody at 0 k , einstein coefficient a is null , so that the risk of phase or energy errors is lowered .
- the strong superradiant emission from the strongly emitting shell found by strmgren produces dotted rings similar to the rings observed around snr 1987a .
the assumed high temperature light emission of the star allows a multiphotonic induced scattering which transfers almost all energy of the continuous spectrum of the star to the superradiant line spectrum , so that the star disappears . for einstein cross and other similar observations , superradiance of an amplifying shell may work more easily than gravitational lensings which require precise alignments of stars . - a coherent parametric effect , made up of several coherent raman effects , exchanges energy between beams of time - incoherent light in media obeying rules explained by g. l. lamb , usually low pressure atomic hydrogen in 2s or 2p states .
the frequency shifts resulting from these exchanges of energy explain many observations : karlsson s periodicities in the spectra of the quasars , anomalous acceleration of the pioneer probes and frequency shifts of the far uv lines of the sun , preserving the laboratory measures of wavelengths .
monte carlo method should be abandoned in astrophysics .
new physics ( mond , variation of fundamental constants , ) should be introduced only after a serious testing of old , well verified physics .
accurate tests and developments of our proposals would be easily done by professional astrophysicists .
d. h. menzel the dilution of radiation in a nebula pasp * 43 * , 70 - 74 ( 1931 ) .
a. einstein , zur quantentheorie der strahlung .
z. * 18 * , 121 - 128 ( 1917 ) .
planck m. , eine neue strahlungshypothese .
. ges . * 13 * , 138 - 175 ( 1911 ) .
a. einstein & o. stern , einige argumente fr die annahme einer molekularen agitation beim absoluten nullpunkt .
annalen der physik * 345 * , 551 - 560 ( 1913 ) g. l. lamb jr . , analytical description of ultra - short optical pulse propagation in a resonant medium , rev .
* 43 * , 99 - 124 ( 1971 ) .
a. m. weiner , d. e. leaird , g. p. wiederrecht , and k. a. nelson , femtosecond multiple - pulse impulsive stimulated raman scattering spectroscopy , j. opt .
* b 8 * , 1264 - 1275 ( 1991 ) j. moret - bailly , structure and light emission of a strmgren system .
arxiv:0905.0554 ( 2009 ) .
b. strmgren , the physical state of interstellar hydrogen , astrophys .
j. , * 89 * , 526 - 547 ( 1939 ) .
s. immler , k. weiler , r. mccray ( eds.)supernova 1987a : 20 years after , aip conf . proc .
, * 937 * , isbn : 978 - 0 - 7354 - 0448 - 9 ( 2007 ) r. a. chevalier & v.v.dwarkadas the presupernova h@xmath0 region around sn 1987a , apj , * 452 * , l45-l48 ( 1995 ) . b. e. k. sugerman , a. p. s. crotts , w. e. kunkel , s. r. heathcote , & s. s. lawrence , the three - dimensional circumstellar environment of sn 1987a .
arxiv : 0502378 ( 2005 ) . c. j. burrows , j. krist , j. j. hester , r .
sahai , j. t. trauger , k. r. stapelfeldt , j. s. gallagher iii , g. r. ballester , s. casertano , j. t. clarke , d. crisp , r.w .
evans , r. e. griffiths , j. g. hoessel , j. a. holtzman , j. r. mould , p. a. scowen , a. m. watson , & j. a. westphal , hubble space telescope observations on the sn 1987a triple ring nebula apj , * 452 * 680 - 684 ( 1995 ) e. michael , r. mccray , r. chevalier , a. v. filippenko , p. lundqvist , p. challis , b. sugerman , s. lawrence , c. s. j. pun , p. garnavich , r. kirchner , a. crotts , c. fransson , w. li , n. panagia , m. phillips , b. schmidt , g. sonneborn , n. suntzeff , l. wang , & j. c. wheeler , hubble space telescope observations of high - velocity ly@xmath47 and h@xmath47 emission from supernova remnant 1987a : the structure and development of the reverse shock , astrophys . j. , * 593 * , 809 - 830 ( 2003 ) .
g.burbidge , the distribution of redshifts in quasi - stellar objects , n - systems , and some radio and compact galaxies , apj , * 154 , * l41-l48 ( 1968 ) . k. g. karlsson , possible discretization of quasar redshifts , a&a , * 13 , * 333 - 335 ( 1971 ) .
j. moret - bailly , propagation of light in low pressure ionized and atomic hydrogen : application to astrophysics , ieeetps , * 31 * , 1215 - 1222 ( 2003 ) .
j. moret - bailly , the parametric light - matter interactions in astrophysics , aip conf .
proc . , * 822 * , 226 - 238 ( 2006 ) . h. peter & p. g. judge , on the doppler shifts of solar ultraviolet emission lines . ,
apj , * 522 * , 1148 - 1166 ( 1999 ) . j.
moret - bailly , anomalous frequency shifts in the solar system , arxiv:0507141 ( 2005 ) .
j. d. anderson , p. a. laing , e. l. lau , a. s. liu , m. m. nieto , & s. g. turyshev , study of the anomalous acceleration of pioneer 10 and 11 , arxiv:0104064 ( 2001 ) . | by confusing the radiance of a single mode light beam , constant in a transparent medium , with the irradiance which decreases away from the source , menzel purports to show that coherent interactions of light with the diluted media of astrophysics , are negligible .
therefore , to study the interaction of light with gases , astrophysicists use monte carlo computations which work to study nuclear systems , but not optics : optical modes which may be defined in inhomogeneous media or for the emissions of single atoms interact coherently with these systems : a unique formula represents , according to the sign of a parameter , absorption and coherent emission .
the optical and spectroscopic properties of a very simple model , an extremely hot source in an isotropic cloud of pure , low pressure , initially cold , huge hydrogen cloud are studied using planck s and einstein s theories .
the similarities of the images and the spectra of this simple model with astronomical observations , for instance of snr1987a , einstein cross , lyman break galaxies , quasars , ... is so large that this model may be an elementary first step in the study of many astrophysical objects .
adaptations of the model to complex astrophysical systems could represent them using only the old , standard theories of physics commonly used in laser spectroscopy .
keywords : radiative transfer ; ism : h@xmath0 regions . 010.5620;030.1640;190.0190 |
the field of transiting extrasolar planets and especially the study of their atmospheres is one of the youngest and most dynamic subjects in current astrophysics .
permanently at the edge of technical feasibility , we have come from the first radial velocity and transit detections @xcite , via the first detections of molecular features in hot - jupiter atmospheres @xcite to ever more detailed characterisations of multitudes of systems @xcite . with
over 700 exoplanets discovered @xcite and over 1200 exoplanetary candidates that await confirmation @xcite , the focus of interest shifts from the detection to the characterisation of smaller and smaller targets .
the governing factor of this progression is the precision at which we can control our instrument and/or stellar systematics , hence the accuracy with which we can analyse the data . to minimise the impact of the systematic noise components , different approaches have been proposed in the past . for space and ground - based observations , eg .
spitzer and hubble ( eg . * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , instrumental systematic noise has been approximated using parametric models , often based on auxiliary information ( optical state vectors ) such as instrumental temperature , orbital inclination , inter and intra - pixel positions of the point - spread - function .
using optical state vectors to de - correlate one s data is an effective technique @xcite .
however for instruments that lack a calibration plan at the precision of 10@xmath0 , the accuracy of the retrieved optical state vectors ( e.g. sensor sampling rates ) and the adequacy of the instrument model s definition itself become difficult to determine .
some of the recent controversy over results reported by various teams can be attributed to this circumstance @xcite .
the situation is even further complicated by brightness variability of the planet hosting star , in particular for small , very active m - stars .
hence , it is important to work towards an alternative route to quantify or remove systematic noise using non - parametric models that work as independent confirmations of existing results .
@xcite , @xcite , @xcite and @xcite have progressed towards non - parametric noise models and signal separation using wavelets , principal component , gaussian processes and fourier based analysis . in this publication
, we propose a new non - parametric method to separate systematic noise from the desired lightcurve signal . given multiple lightcurves , observed simultaneously using spectrographs for example
, we can disentangle our desired astrophysical signal from other time - correlated or non - gaussian systematic noise sources using un - supervised machine learning algorithms .
we furthermore explore the de - correlation of individual time series spanning several consecutive eclipse events .
the importance of this work lies with the fact that no additional knowledge of the system or the star is required , besides the observations themselves .
such non - parametric methods provide a potential new route to de - correlate one s date in the case where the instrument does not feature an adequate calibration plan , the quality of the auxiliary information of the instrument is non - optimal or the host star shows significant activity .
such blind de - convolution techniques provide new insight and powerful validation of the established parametric instrumental models reported in the literature . here
we will briefly introduce the more general theory of blind - source separation and proceed with a description of the algorithm proposed .
the efficiency of said algorithm is tested with two synthetic models and two hst / nicmos data sets available in the public domain and one kepler ( q2 & q3 data release ) time series featuring strong time - correlated stellar variability .
future publications will focus on in - depth discussion of the proposed algorithm to specific data sets .
in this section we will briefly describe the fundamental concepts on which the following analysis is based .
readers familiar with independent component analysis may proceed straight to section [ method ] .
let us consider the analogy of three people talking simultaneously in one room .
the speech signals of these people are denoted by @xmath1 , @xmath2 and @xmath3 . in the same room
are three microphones recording the observed signals @xmath4 , @xmath5 and @xmath6 . the observed signals can be expressed in terms of the original speech signals : @xmath7 instead of assuming @xmath8 and @xmath9 to be proper time signals , we drop the time dependence and assume them to be random variables @xmath10 where @xmath11 is a weighting factor ( in this case the square of the distance of the speakers to the microphone ) and @xmath12 are some real coefficients with @xmath13 being the maximum number of observed signals .
the individual time series can also be expressed in terms of vectors where bold lower - case letters denote column vectors and upper - case letter denote matrices : @xmath14 where the rows of @xmath15 comprise the individual time series , @xmath16 , and similarly @xmath17 is the signal matrix of the individual source signals @xmath18 .
@xmath19 is the mixing matrix consisting of the weights @xmath20 .
equation [ timeseries3 ] is also known as the instantaneous mixing model and often referred to as the classical cocktail party problem @xcite .
the challenge is to estimate the mixing matrix , @xmath19 and its ( pseudo)inverse the de - mixing matrix , @xmath21 , @xmath22 given the observations contained in @xmath15 without any additional prior knowledge of either @xmath19 or @xmath17 , or for some methods without restrictions of @xmath19 & @xmath17 .
several algorithms have been proposed to find the linear transformation of equation [ timeseries3 ] . amongst these are principal component analysis ( pca ) @xcite , factor analysis ( fa ) @xcite , projection pursuit ( pp ) @xcite and the more recently developed independent component analysis ( ica ) @xcite .
the underlying differences between pca and fa on one hand and ica and pp on the other are the underling assumptions on the probability distribution functions ( pdfs ) of the signals comprising @xmath15 .
the former group assumes the signals to follow : 1 ) a gaussian distribution whilst the latter assume the signals to be , 2 ) predominantly non - gaussian or sparse with specific signatures in the spectral domain ( e.g. smica * ? ? ?
this results in significant differences in the way we estimate our signal components .
\1 ) if the observed signals comprising @xmath15 follow gaussian distributions , we can define their statistics by the first and second statistical moments ( mean and covariance ) only .
algorithms such as pca and fa find a linear transformation from the correlated observed signals , @xmath15 , to a set of uncorrelated signals , @xmath17 . in this case
, uncorrelatedness is equivalent to mutual independence and the source signals are at their most separated ( please see appendix [ appendix : independent ] for a more in - depth discussion on independence ) .
such a linear transformation is always possible and easily achieved using , for example , single value decompositions ( svd ) @xcite .
an application of pca to exoplanetary light curve de - trending is given by @xcite .
\2 ) in the case of the observed signals following non - gaussian distributions , significant information is contained in the higher statistical moments ( skew & kurtosis ) and it can be shown that uncorrelated signals ( as produced by pca & fa ) are not necessarily mutually independent and hence not optimally separated from one another . here uncorrelatedness is a weaker constraint than independence and it can be said that independent signals are always uncorrelated but not vice versa ( see appendix [ appendix : independent ] ) . as a consequence using pca or fa algorithms may only yield a partially separated result for non - gaussian sources .
it is important to note that most observed signals , astrophysical or stellar / instrumental noise , are predominantly non - gaussian by nature .
we can also state that most of these signals should be statistically independent from one another ( e.g. an exoplanet light curve signal is independent of the systematic noise of the instrument with which it was recorded ) .
these properties have led to a surge in ica based analysis methods in current cosmology and extra - galactic astronomy . here
ica is used to separate the cosmic microwave background ( cmb ) or signatures of distant galaxies from their galactic foregrounds ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
@xcite furthermore separates instrumental noise from the desired astrophysical signal .
other applications include data - compression of sparse , large data sets to improve model fitting efficiencies ( e.g. * ? ? ?
* ; * ? ? ?
in this publication , we focus on the application of ica to exoplanetary lightcurve analysis .
let us consider multiple time series observations of the same exoplanetary eclipse signal either in parallel , by performing spectrophotometry with a spectrograph , or consecutive in time ( as explained in section [ kepler ] ) . without excluding the most general case ,
let us focus on a time - resolved spectroscopic measurement of an exoplanetary eclipse .
for most observations , the signal recorded is a mixture of astrophysical signal , gaussian ( white ) noise and systematic noise components originating from instrumental defects and other sources such as stellar activity and telluric fluctuations .
we can therefore write the individual time series as sum of the desired astrophysical signal , @xmath23 , systematic ( non - gaussian ) noise components , @xmath24 , and gaussian noise , @xmath25 . we can now define the underlying linear model of our time series data to be @xmath26 or @xmath27 where @xmath28 and @xmath29 are the number of systematic and white noise components respectively and @xmath30 assuming only one component is astrophysical .
the basic assumptions of ica are that the elements comprising @xmath17 , @xmath31 , are mutually independent random variables with probability densities , @xmath32 .
we further assume that all ( or at least one ) of the probability densities , @xmath33 , are non - gaussian .
this non - gaussianity is key since it allows the de - mixing matrix , @xmath21 , to be estimated . from the central limit theorem , we know that a convolution of any arbitrary probability distribution functions ( pdfs ) that feature a formal mean and variance , asymptotically approaches a gaussian distribution in the limit of large @xmath13 @xcite . in other words , the sum of any two non - gaussian pdfs ( ie . @xmath33 and @xmath34 ) is more gaussian than the respective original pdfs .
therefore by maximising the non - gaussianity of the individual signals , we maximise their statistical independence . @xcite .
although several measures of non - gaussianity exist ( we refer the reader to @xcite , @xcite , @xcite and @xcite for detailed summaries ) , we here use the concept of negentropy @xcite .
negentropy is derived from the basic information - theoretical concept of differential entropy . in information theory , in close analogy to thermodynamics , the entropy of a system is at its maximum when all data points are at its most random . in thermodynamics we measure the distribution of particles , in information theory
it is the probability distribution of a random variable . from information theory
we can derive the fundamental result that a gaussian distribution has the largest entropy among all random variables of equal variance and known mean .
hence , by minimising the entropy of a variable , we maximise its non - gaussianity . for a random vector @xmath35 , with random variables @xmath36 , @xmath37 , the entropy is given by @xmath38 where h@xmath39 is the differential or shannon entropy @xcite and @xmath40 is the pdf of the random vector @xmath35 .
h@xmath39 is at its minimum when @xmath41 is at its most non - gaussian .
we can now normalise equation [ shannon ] to yield the definition of negentropy : @xmath42 where @xmath43 is a random gaussian vector with the same covariance matrix as @xmath35 .
now @xmath35 is at its most non - gaussian when @xmath44 is at its maximum .
it is important to note that negentropy is insensitive to a multiplication by a scalar constant .
this has the important consequence of introducing a sign and scaling ambiguity into the retrieved signal components of @xmath17 .
we will discussed the implications of this limitation in section [ simu ] . in practice
it is very difficult to calculate the negentropy of a system and various methods were devised to approximate @xmath44 .
the classic method is to measure the kurtosis of mean - subtracted @xmath35 with unit variance .
however , kurtosis is very prone to distortions by outliers in the data and hence lacks the robustness required as measure of negentropy @xcite . to overcome this limitation , more robust measures of negentropy
have been devised .
one can approximate negentropy by equation [ negentropy2 ] @xcite @xmath45 - \text{e}[g(\nu)])^{2 } \label{negentropy2}\ ] ] where @xmath46 is a random gaussian variable with zero mean and unit variance and @xmath47 is a non - quadratic contrast function chosen to approximate the underlying probability distribution .
there are usually three types of contrast functions : @xmath48 as general purpose function , @xmath49 optimised for super - gaussian ( leptokurtic ) distributions and @xmath50 optimised for sub - gaussian ( platykurtic ) distributions @xcite : @xmath51 \\\nonumber g_{2}(y ) & = -\text{exp}(-y^{2}/2 ) \\\nonumber g_{3}(y ) & = \frac{1}{4}y^{4}\end{aligned}\ ] ] the choice of contrast function is only important if one wants to optimise the performance of the ica algorithm as it is done for the efica @xcite algorithm where choices of contrast functions are tried iteratively . finally , we can maximise the negentropy given in equation [ negentropy2 ] by finding a unit vector @xmath52 that maximises the non - gaussianity of the projection @xmath53 , so that @xmath54 is at its maximum . for the fixed - point fastica algorithm ,
this can be achieved by the simple iterations scheme @xcite : 1 .
choose initial ( i.e. random ) weight vector @xmath52 2 .
increment : @xmath55 - \text e [ g^ { ' } ( { \bf w}^{t } { \bf x } ) ] { \bf w}$ ] 3 .
normalise : @xmath56 4 .
repeat steps 2 & 3 until converged where @xmath57 and @xmath58 are the first and second derivatives of the contrast function @xmath59 .
the iteration scheme above estimates only one weight vector at a time , for estimating all non - gaussian components in parallel the iteration scheme and derivates of @xmath59 are given in appendix [ efica ] and in the standard literature ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
for a comprehensive summary of other ica algorithms please refer to @xcite . throughout this paper , we use a variant of the fastica algorithm .
projection pursuit and independent component analysis are inherently linked as both algorithms try to represent the most non - gaussian components in an multidimensional data set . in this sense
, one can think of ica as a variant of pp with two major differences : 1 ) pp only estimates one non - gaussian component at a time whilst ica is the multivariate definition of pp and estimates all non - gaussian components simultaneously , 2 ) as opposed to ica , pp does not need an underlying data model and no assumptions about independent components are made .
if the ica model holds , optimizing the ica non - gaussianity measures produce independent components ; if the model does not hold , then what we get are the projection pursuit directions @xcite .
this is an important point to make with regards to time - consecutive transit observations , which break the underlying ica assumptions otherwise .
following from the discussion above , we can understand the signal de - mixing to be a two step process : de - correlation of the gaussian components in the observed data using pca , followed by the de - correlation of non - gaussian components using ica and wasobi algorithms .
the de - correlation of gaussian components to form a new uncorrelated vectors set can be understood as pre - processing step to the ica procedure .
the algorithm proposed here consists of five main parts : 1 ) pre - processing of the observed data , 2 ) signal separation , 3 ) signal reconstruction 4 ) lightcurve fitting and 4 ) post - analysis .
figure [ flow ] lays out the individual processing steps of the algorithm . similarly to section [ independent ] , the observed signals can be expressed as @xmath60 dimensional matrix * x * where each row constitutes an individual time series , @xmath16 , with @xmath61 number of data points .
multiple time series observations are needed to separate the instantaneously mixed non - gaussian components .
the process of identifying statistical independent components is greatly simplified if the input signals to any ica algorithm have previously been whitened ( also referred to as sphering ) . whitening is essentially a transformation of our input data matrix * x * into a mean subtracted , @xmath62 , orthogonal matrix @xmath63 , where its auto - covariance matrix , @xmath64 , equals the identity matrix , @xmath65 = i$ ] . the instantaneous mixing model for the whitened data is now given by @xmath66 where @xmath67 is the inverse square root of @xmath68 and @xmath69 the corresponding mixing matrix of @xmath63 . for a more detailed explanation see appendix .
this whitening is easily achieved by performing a principal component analysis on the data ( see appendix ) .
this step has two distinct advantages : \1 ) it reduces the complexity of the un - whitened mixing matrix , @xmath19 , from @xmath70 parameters , to @xmath71 for a whitened , orthogonal matrix @xmath69 @xcite .
2 ) using whitening by principal components , we can reduce the dimensionality of the data - set by only maintaining a sub - set of eigenvectors .
this reduces possible redundancies of the components comprising the data and prevents the later to be employed ica algorithm from over - learning for over - complete sets .
we also like to note that any type of additional linear signal cleaning or pre - processing step , such as those described by @xcite , are allowed .
linear data filtering or cleaning can be understood as multiplying equation [ timeseries3 ] from the left with a linear transformation @xmath72 to get : @xmath73 .
the underlying data model assumed in this paper is hence not affected .
after the observed signals have successfully been whitened ( @xmath63 ) , we estimate the mixing matrix of the whitened signal , @xmath69 , using the multi - combi algorithm @xcite .
multi - combi comprises two complimentary algorithms , efica @xcite and wasobi @xcite .
efica , an asymptotically efficient variant of the fastica algorithm @xcite , is designed to separate non - gaussian , instantaneously mixed signals .
wasobi , on the other hand , is an asymptotically efficient version of the sobi algorithm @xcite , and is geared towards separating gaussian auto - regressive ( ar ) and time - correlated components .
it uses second - order statistics and can be understood to be similar to principal component analysis .
the use of both algorithms is necessary since a real life data set will always contain a mixture of both , non - gaussian and gaussian ar processes . for a more in - depth discussion of the algorithms employed here , we like to refer the interested reader to the appendices [ efica ] & [ wasobi ] and the original publications .
the efica and wasobi algorithm can be shown to be asymptotically efficient , i.e. the estimators approach the cramr - rao lower bound @xcite . in other words ,
the algorithms employed here can be shown to converge to the correct solution given the original source signals and in the limit of @xmath74 iterations . in reality
the number of iterations is finite and and imperfect convergence results in traces of other sources to remain in the individual signals comprising @xmath75 .
we can hence state that equation [ demix ] becomes @xmath76 a measure of this error is the deviation of @xmath77 ( or @xmath78 for the whitened case ) from the unity matrix by inspecting the variance of its elements @xcite .
this leads to the concept of the interference - over - signal ratio ( isr ) matrix .
the isr is the standard measure in signal processing of how well a given signal has been transmitted or de - convolved from a mixture of signals .
it can be understood as the inverse of the signal - to - noise ratio ( snr ) .
the higher the isr for a specific signal , the less well has it been separated from the original mixture . for a real case example , @xmath19 is unknown and the isr needs to be estimated .
an analytic approximation to the isrs for the efica and wasobi algorithms are found in the appendices [ efica ] & [ wasobi ] .
finally , we check the stability of the signal separation by perturbing the input matrix @xmath63 by a random and known matrix @xmath79 to give @xmath80 we now re - run the multi - combi procedure using @xmath81 as input and estimate @xmath82 . knowing @xmath79 we can work backwards to obtain @xmath69 as we are dealing with a linear transformation .
this step is repeated several times to check the convergence of the algorithm by inspecting the variation on the mean isr values of each separation attempt and the variations in consecutive estimations of @xmath69 directly . once the mixing matrix ,
@xmath69 is estimated , we need to identify which signals are astrophysical , which ones are white and which are systematic noise .
this is done in a two step process : \1 ) we construct the estimated signal matrix , @xmath83 , and for its individual components @xmath84 compute the pearson correlation coefficient between @xmath84 and the first principal component of the pca decomposition in section [ pca ] . for medium
signal to noise ( snr ) observations , the first principal component ( pc ) , ie .
the one with the highest eigenvalue associated to it , will contain the predominant lightcurve shape .
as previously discussed , the first pc is not perfectly separated from the systematic signals and hence can not be used directly for further analysis but it is good enough to use it as lightcurve identification .
the identified lightcurve signal is labeled @xmath85 .
\2 ) once the lightcurve signal is identified , we exclude this row from @xmath83 and proceed to classify the remaining signals with respect to their non - gaussianity ( ie .
systematic noise sources ) . here
we use the ljung - box portmanteau test ( see appendix and * ? ? ?
* ) to test for the hypothesis that the time series is statistically white ( ie .
gaussian ) .
this test was originally designed to check the residuals of auto - regressive moving - average ( arma ) models for significant departures from gaussianity .
it is hence ideally suited for our need to identify which estimated signal components are the desired non - gaussian ones .
the identified non - gaussian , systematic noise , signals are hence labeled @xmath86 and the remaining white noise signals @xmath87 to give @xmath88 where @xmath83 has dimensions @xmath89 and @xmath85 , @xmath86 , @xmath87 , have dimensions of @xmath90 , @xmath91 and @xmath92 respectively where @xmath93 .
the de - mixing matrix is given by @xmath94 .
as previously mentioned , the components of @xmath83 have ambiguities in scaling and sign and can be thought to be similar to the eigenvectors of a principal component analysis with missing eigenvalues .
fortunately there exist two approaches to resolving this degeneracy : 1 .
in the case of @xmath85 being well separated as individual component , we can take @xmath85 and the de - mixing matrix @xmath95 and only retain the row containing the astrophysical signal component forming the row - vector @xmath96 .
we then reconstruct the original data @xmath63 using only the separated signal component : + @xmath97 + where @xmath98 is the reconstructed whitened data with all but the astrophysical signal components removed .
using equation [ whitening ] , we can now calculate the un - whitened matrix @xmath99 .
+ @xmath100 + hence we can think of @xmath101 as a linear , optimal filter for the signal component in @xmath102 .
please note that this linear filtering does not impair the scaling information as this is re - instated going from @xmath85 to @xmath103 .
2 . in the case of @xmath85 not being well separated but other systematic noise components are , a different , more indirect approach can be used . here , the systematic noise components , @xmath86 which do not contain sign or scaling information , are simultaneously fitted to the time series data ( preferably out - of - transit data ) , @xmath16 .
we therefore define the systematic noise model for an individual time series by @xmath104 , + @xmath105 + where @xmath106 is a k @xmath107 k diagonal scaling matrix of @xmath86 , which needs to be fitted iteratively as free parameters in the following section .
having either filtered the data to obtain @xmath99 or constructed the noise model @xmath104 , we can now fit the original time series , @xmath16 using the standard analytical lightcurve models @xcite in addition to the diagonal matrix @xmath106 , if necessary .
for the purpose of this paper , which focuses on blind - source - separation , we will restrict ourselves to demonstrating the feasibility of estimating @xmath106 and only leave the transit depth as variable lightcurve parameter .
we use the analytic lightcurve model by @xcite and a nelder - mead minimisation algorithm @xcite . for real data applications ,
we advise the reader to use markov chain monte carlo methods , or similar , which have become standard in the field of exoplanets and allow the estimation of the posterior probability distributions and their associated errors @xcite .
once the model fitting stage has been completed , we are left with fitting residual , @xmath108 , i.e. @xmath109 .
several tests are useful to determine how well the signals have been removed from the original time series , @xmath16 . in the case of a full markov chain monte carlo fitting , the posterior distributions of the fitting parameters
may be used to asses the impact of the remaining systematic noise in the data when compared to a simulated data set only containing white noise .
portmanteau tests on individual time series are useful to test for non - white noise signals and allow a measure of the overall performance of the algorithm @xcite .
additionally , we can determine the kullback - leibler divergence @xcite of our residual s probability distribution function ( pdf ) to an idealised gaussian case .
for the simulations and real - data examples presented in the following section , we have merely plotted the autocorrleation functions ( acf ) of the residuals obtained to determine whether for a given lag , these are within the 3@xmath110 confidence limit of the residual being dominated by white - noise @xcite . here the acf is given by : @xmath111 where @xmath61 is the number of data points in the time series ,
@xmath112 the specific lag and the confidence intervals are given by @xmath113 .
in order to test the behaviour and efficiency of the algorithm described above , we produced a toy model simulation with five observed signals : 1 ) a secondary eclipse @xcite lightcurve ; 2 ) a sinusoidal signal ; 3 ) a sawtooth function ; 4 ) a fourth order auto - regressive signal to simulate time - correlated signals ; 5 ) gaussian noise with a full width half maximum ( fwhm ) of 0.01 magnitudes .
the premixed signals are displayed in figure [ ex1premix ] .
this gives us our signal matrix , @xmath75 , which needs to be recovered later on .
we have then proceeded to mix the signals in figure [ ex1premix ] using a random mixing matrix , @xmath19 , to obtain our observed signals , @xmath102 , in figure [ ex1mix ] . for the sake of comparability
we keep the mixing matrix @xmath19 to be the same for all simulations .
we now subdivide the simulations to illustrate the two possible methods of the signal reconstruction .
_ method _ 1 computes @xmath114 using equation [ sigcomp2 ] whilst _ method _ 2 fits the noise model @xmath104 ( equation [ noisemodel2 ] ) simultaneously with the @xcite lightcurve .
these two examples demonstrate that both techniques work equally well for a well behaved data set .
in this example , we use the observed signals in figure [ ex1mix ] as input to the algorithm .
we however do not perform a dimensionality reduction using pca since we are not dealing with an over - complete set in this example .
the results of the separation are shown in figure [ ex1postmix ] . here the top three , red lightcurves are the estimated systematic noise components as identified by the algorithm .
the fourth component is gaussian noise and the bottom is an inverse of the lightcurve signal .
it should again be noted here that the blind - source separation does not preserve the scaling nor the signs of the signals in @xmath83 .
however , when the original data is reconstructed using only the signal component , @xmath85 , to obtain @xmath114 ( equation [ sigcomp2 ] ) , the scaling and sign informations are re - instated . for a well behaved data set , i.e. one that obeys the instantaneous mixing model and has negligible gaussian noise in their signal components , it is therefore possible to re - construct the lightcurve signal from the raw data as explained in section [ nmodel ] .
figure [ ex1filter ] shows the top lightcurve of figure [ ex1mix ] ( blue circles ) and overplotted the retrieved signal component ( red crosses ) and offset below the systematic noise component ( black squares ) . as a useful by - product of the algorithm
, we obtain the interference over signal matrices ( isr , equations [ isref ] @xmath115 [ isrwa ] in the appendix [ bss ] ) for both the efica and wasobi algorithms .
these give us valuable information on the efficiency at which the signals have been separated .
figure [ ex1isrs ] shows the hinton diagrams of the efica and wasobi @xmath116 matrices . here , the smaller the off - diagonal elements of the matrix , the better the signal separation .
in this example , the efica algorithm outperforms the wasobi one , which is to be expected since all signals but one are non - gaussian . in the previous section
, we have shown that in the case that the astrophysical component @xmath85 is well separated as individual signal , we can create a filter for the raw data that directly filters the lightcurve signal from the noise . however , in most real data applications , @xmath85 , is not perfectly separated but the components of @xmath86 may be more so . in this case
we can construct the noise model @xmath104 given by equation [ noisemodel2 ] and the diagonal elements of @xmath106 are fitted as described in section [ fitting ] .
the starting position of the algorithm is the same as for the previous example ( figure [ ex1mix ] ) .
the model fit of the first lightcurve in figure [ ex1mix ] and its residuals are shown in figure [ ex1fitted ] . the autocorrelation function for 100 lags
is plotted in figure [ ex1auto ] .
all but two lags are within the 3@xmath110 confidence limit that the residual is white noise dominated , indicating that all signals have been removed effectively .
finally we simulate the convergence properties of both efica and wasobi under varying white noise conditions . here
we repeatedly run the algorithm until signal separation is completed and record the mean isrs of the source separation .
we performed this simulation 300 times for gaussian noise fwhms varying from 0.0 - 0.3 magnitudes ( figure [ ex1fitted ] has a fwhm@xmath117 = 0.01 ) and every isr measurement reported is the mean of 10 iterations .
figure [ ex1isr2 ] summarises the results . here
, the red circles represent the mean isr or the efica algorithm and the blue crosses that of wasobi .
it can clearly be seen that for this example the efica algorithm outperforms wasobi and on average reaches lower isr values .
we can further note that the blind source separation is not significantly affected by the magnitude of the white noise and performs well under difficult signal to noise conditions .
the previous examples illustrated the two methods available to correct the observed data : _ method 1 _ : filtering the astrophysical signal out of the systematic noise or _ method 2 _ : fitting a systematic noise model to the raw data .
here we apply these techniques to two primary eclipse data sets of hd 189733b and xo1b recorded by the nicmos instrument on the hubble space telescope as well as a single time - correlated time series obtained by the kepler spacecraft .
first presented by @xcite , this data set of the primary eclipse of hd 189733b was recorded using hst / nicmos in the g206 grism setting spanning five consecutive orbits .
the hst - pipeline calibrated data were downloaded from the mast archive and the spectrum was extracted using both standard iraf routines as well as a custom built routine for optimal spectral extraction .
both extractions are in good accord with each other but the custom built routine was found to yield a better signal to noise and was subsequently used for all further analysis .
a binning of 10 spectral channels ( @xmath118 0.08@xmath119 m ) was used resulting in 10 light curves across the g206 grism band .
figure [ hd189rawlc ] shows the obtained time series which serve as input to the algorithm .
it can be seen that each time series is strongly affected by instrument systematics propagating from the blue side of the spectrum ( bottom light curve ) to the red with varying intensity and even sign .
@xcite showed that these systematics are correlated to instrument state vectors such as orbital phase , relative positions and angles of the spectrum on the detector , instrument temperature , etc .
we can hence expect that these systematics are statistically independent from the recorded astrophysical signal ( the light curve ) and it should therefore be in principle possible to de - correlate the signal .
we here demonstrate the de - trending on an individual light curve at @xmath1181.694@xmath119 m ( 8@xmath120 one down in figure [ hd189raw ] ) .
all time series in figure [ hd189raw ] were taken as input to the algorithm described above to estimate the de - mixing matrix @xmath95 , the astrophysical signal vectors , @xmath121 and the systematic noise vectors , @xmath122 .
the interference over signal ( isr ) matrix indicated the good separation of four main components figure [ hd189hinton ] with the rest of the components being classified as predominantly gaussian or weakly systematic .
the existence of more than one gaussian component ( @xmath123 ) indicates that the set is overcomplete .
however since the data - set is small enough , no pca dimensionality reduction was performed .
after the algorithm has identified the correct astrophysical signal , it proceeded to reconstruct the light curve using both methods described above .
_ method 1 _ : the astrophysical signal was filtered using equations [ sigcomp ] @xmath115 [ sigcomp2 ] .
figure [ hd189componly ] shows the raw light curve ( blue circles ) with the de - trended time series , @xmath114 underneath ( green squares ) .
superimposed light curves were computed using @xcite with orbital parameters were taken from @xcite and limb - darkening parameters from @xcite .
it is clear that the de - trended light curve is an improvement to the raw time series but that systematics still remain in the data .
this is further illustrated by plotting the autocorrelation function of the model - fit residual in figure [ hd189acf ] ( red circles ) . here , residual correlation can be observed in particular at low lags .
this is a consequence of the astrophysical signal , @xmath121 , being well separated but as shown in figure [ hd189hinton ] ( component 1 ) , there remains some weak interference between the @xmath121 and other vectors , which is a consequence of equation [ demix2 ] and to be expected for real data - sets .
_ method 2 _ : the second method is a less direct approach . instead of filtering for the astrophysical signal directly , we try to construct a systematic noise model that is then subtracted off the raw data . using equation [ noisemodel2 ] and a simplex downhill algorithm @xcite we estimated the scaling matrix , @xmath106 , by fitting the the systematic noise vectors , @xmath122 to the four out of transit orbits .
the scaled systematic noise vectors are shown in figure [ hd189snvec ] which combine to form the systematic noise model , @xmath104 , in figure [ hd189raw ] .
it should be noted that @xmath106 is only a scaling matrix of the individual vectors as the scaling information is not preserved by the independent component analysis .
hence , relative intra and inter - orbit variations are preserved .
figure [ hd189corr ] shows the corrected data by subtracting the systematic noise model off the raw data . inspecting the fitting residual s autocorrelation function in figure [ hd189acf ] ( black circles ) indicates the residual to be statistically white and
a maximal de - correlation of the data has been achieved .
originally presented by @xcite , the primary eclipse of xo1b was observed using the hst / nicmos instrument in the g141 grism setting .
the hst - pipeline calibrated data was downloaded and the spectra extracted using the same settings as for section [ hd189sect ] .
this yielded 10 light curves and which serve as input to the algorithm , see figure [ xo1rawlc ] .
similar to the hd 189733b the algorithm retrieved four main components , the light curve signal and three main systematic noise components .
the isr matrix can is shown in figure [ xo1hinton ] .
we now proceeded to de - trending the light curve at the very red end of the spectrum ( first from top in figure [ xo1rawlc ] ) as it , after visual inspection , exhibits the most prominent systematics of the 10 time series .
light curve fits assumed limb - darkening and orbital parameters by @xcite .
_ method 1 _ : figure [ xo1componly ] shows the raw time series and the de - trended light curve using equation [ sigcomp2 ] .
the light curve is significantly de - trended but systematics remain in the data as also shown by the autocorrelation function ( red circles ) in figure [ xo1acf ] .
_ method 2 _ : as described in previous sections , figure [ xo1snvec ] shows the retrieved systematic noise vectors and figure [ xo1raw ] features the raw data with the combined systematic noise model ( red ) underneath . the autocorrelation function of the model fit residual is shown in figure [ xo1acf ] ( black crosses ) and shows a factor 2 improvement on the de - correlation in the lower lags .
figure [ xo1comp2 ] compares the de - trended light curves of _ method 1 _ and _ method 2 _ and shows the residual of _ method 1 _ - _ method 2 _ ( black crosses ) .
there is little difference between both methods indicating that the signal separation for this data - set is close to its maximum with the data being partially de - correlated .
this is in contrast to the hd 189733b example where a near perfect de - correlation was achieved and can be attributed to the systematics being mostly wavelength invariant in the case of xo1b .
in other words , systematic noise components which have a constant weighting throughout the data set can not be de - correlated using ica or pca methods , which is to be expected following equations [ intro1 ] - [ timeseries3 ] .
in the previous examples we have shown that spectroscopic datasets can be de - correlated effectively .
we test here how well the proposed algorithm can de - correlate consecutively observed data - sets .
this is of particular interest in cases where no multi - channel data are available and the time series data are contaminated with time - correlated noise , be it stellar or instrumental . as opposed to the previous examples , where several time series , @xmath16 , were observed simultaneously , here we take a single time series covering several consecutive eclipse features and cut the time series into segments spanning equal lengths over each eclipse event . using these segments as inputs to the algorithm clearly violates the underlying assumptions of the independent component analysis , as the mixing is not instantaneous . in this case , the ica analysis can be understood as a projection pursuit ( pp ) analysis , see section [ projectionpursuit ] and @xcite . here
the ica algorithm , in the absence of a working ica data - model , will try to extract as many non - gaussian components as possible and return the rest of the data in its original form .
this is very similar to projection pursuit , where the data is not described by an underlying data model at all but only the most non - gaussian component is retrieved . in other words , we can only expect to retrieve the eclipse signal component , @xmath85 , with any degree of accuracy . as a result we will not be able to retrieve systematic noise components , @xmath86 , and we can only use @xmath124 ( in section [ nmodel ] ) to de - trend the data .
we have downloaded data observed by the kepler space telescope @xcite for a planet - hosting candidate star observed over the second and third data - release quarters ( q2 & q3 ) .
the time series , with the kepler i d : 10118816 , exhibits highly variable features and significant time - correlated noise ( see figure [ kepler3 ] , blue crosses ) .
given kepler s superb instrument calibration , we can assume this time - correlated noise to be due to stellar variability .
using the periodogram calculator on the nsted database ] , we identified four main periodically recurring signals in the data - set . choosing the second strongest feature with a period of 0.040915 days , we phase folded the data and cut the time series in 10 equally sized segments .
as for the previous examples we now took these time series segments as input to the algorithm .
we performed our de - correlation as for the previous examples but using _ method 1 _ only .
figure [ kepler4 ] shows the isr matrix of the separation indicating a relatively poor separation of the components but two ( components 4 & 9 ) . as discussed above , this behaviour is to be expected with the breaking of the instantaneous mixing model .
nonetheless , we obtained a clear feature ( component 4 ) in our analysis which is over plotted ( red circles ) on the mean , phase - folded data ( blue crosses ) in figure [ kepler1 ] . here
the de - correlated signal has a much reduced scatter compared to the mean of the phase - folded feature , which indicates that much of the unwanted stellar variability has been removed .
it is also clear from this figure that we are not dealing with an exoplanetary light curve but a stellar pulsation feature .
as expected , the remaining components returned from the algorithm ( figure [ kepler2 ] ) are the residuals of the input data minus the component shown in figure [ kepler1 ] .
hence we only used the component in figure [ kepler1 ] to re - construct the original time series .
this was done by using equation [ sigcomp2 ] on each segment of the time series , followed by adding the segments back together in the order they were originally split up .
figure [ kepler3 ] shows the original input data ( blue crosses ) with the filtered signal ( red circles ) over plotted on top . in the bottom plot ( a zoomed in version of the time series ) , it is clear that the desired feature remains in the filtered time series whilst the contribution of other time - correlated stellar noise is substantially reduced .
in the previous sections we have shown that for a set of simultaneously observed time series data ( e.g. following an exoplanetary eclipse with a spectrograph ) we can describe the data by an instantaneous mixing model ( equation [ timeseries3 ] ) .
this allows the separation of non - gaussian , time and spatially - correlated signals from one another .
the degeneracy caused by not being able to retrieve the component s signs or amplitudes can be circumvented in two ways : _ method 1 _ ) the separated signals are used to construct a linear transformation to filter the astrophysical signal from the originally observed data and hence preserve all scaling information ; _ method 2 _ ) the separated astrophysical signal is not used directly but instead all systematic noise components are combined to form a ` systematic noise model ' which can then be used to correct the original observed data . we have explored the efficiency of the signal de - trending on two simulated and two hst / nicmos data sets with different types of systematic noise due to different grisms .
the simulations demonstrate the two methods of de - trending the data in an idealised case and explore the efficiency of the signal separation in the presence of varying gaussian noise in the data . in the instantaneous mixing model employed here ,
gaussian noise sources are only indirectly allowed and can interfere with the effectiveness of separating non - gaussian vectors .
we tested this point by adding additional gaussian noise components of variable amplitude to the simulations but did not observe any significant reductions in the signal separation efficiency .
we proceeded to analyse two hst / nicmos data sets : the primary eclipses of hd189733b and xo1b . for both data sets we find the _ method 2 _ to yield better results .
in the case of hd189733b , we can achieve a near perfect de - correlation of astrophysical signal and systematic noise and no further steps are necessary to the de - correlation process . a more in depth discussion of this data set and hst / nicmos systematics
is beyond the scope of this publication . in the case of xo1b
the de - correlation is significant but incomplete .
the difference in maximum de - correlation achievable can be attributed to the systematic noise sources being strong functions of wavelength in the case of hd189733b whilst almost with constant weighting ( @xmath125 in equation [ intro2 ] ) in the case of xo1b .
whenever systematics have constant weighting per channel observed ( @xmath16 ) and/or time , it becomes very difficult for pca or ica based approaches to de - correlate the signal from the systematics .
here auxiliary information of the instrument is needed to proceed with the de - correlation process .
this is very well possible in the case of dedicated instruments such as kepler , as these have been specifically designed with such high precision and stability measurements in mind @xcite . for instruments that do not feature the calibration plan required to further de - correlate with instrument state parameters ,
the solution is far less obvious .
we furthermore explored the de - correlation of eclipse signals observed consecutively rather than in parallel . we demonstrated , using kepler data , that despite the formal violation of the ` instantaneous mixing model ' , the proposed algorithm is able to retrieve the desired signal component with good accuracy .
such an application is particularly important for treating variability of the host - star which can significantly impair the quality of the final science result ( eg .
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is furthermore interesting to note that pre and post - processing steps ( e.g. wavelets @xcite , fourier based techniques @xcite , de - correlation using instrument state parameters @xcite ) , do not break the instantaneous mixing model and can be run in conjunction with ica methods .
this makes independent component analysis a very powerful and versatile tool for non - parametric de - correlation of exoplanetary data sets .
in the light of searching and characterising ever smaller and fainter exoplanetary targets , the development of novel de - trending routines becomes increasingly critical . based on the concepts of blind source deconvolution of instantaneously mixed signals ,
we have presented a first step towards non - parametric corrections and data filters that do not require additional information on the systematic noise of the instrument or stellar activity .
such algorithms have two important applications : \1 ) for instruments that lack a calibration plan at the accuracy of 10@xmath0 in flux variation , which is required for spectroscopy of exoplanetary atmospheres , the spectroscopic signatures become inherently entangled and dependent on the method used to correct instrument and other systematics in the data .
the de - correlation of spectroscopic data was demonstrated using two hst / nicmos data sets .
\2 ) detections of faint exoplanetary eclipses are often made difficult by time - correlated activity of the host star .
we demonstrated , using a single kepler time series , that much of the stellar variability can be removed in time series that span several exoplanetary eclipse events .
the algorithm proposed is a powerful tool for lightcurve de - trending , which can be used by its own or in conjunction with any other type of data filtering or cleaning technique .
this becomes an invaluable advantage for data analysis when the instrument s response function is unknown or poorly characterised .
i.p.w . would like to thank prof .
e. feigelson , the referee , dr .
g. tinetti , dr .
f. abdalla and dr .
s. fossey for comments and suggestions that helped to greatly improve this paper .
i.p.w . is supported by an stfc studentship .
aigrain , s. , pont , f. , zucker , s. , 2011 , arxiv : 1110.1034 , e. , cowan , n. b. , knutson , h. a. , deming , d. , et al . , 2010 , apj , 721 , 1861 aumont , j. & macas - prez , j. , f. , 2007 , mnras , 376 , 739 bakos , g. . , shporer , a. , pl , a. , torres , g. , kovcs , g. et al . , 2007 , apjl , 671 , l173 ballerini , p. , micela , g. , lanza , a. f. , pagano , i. , 2011 , a&a , submitted bean , j. l. , , 2011 , nature , 478 , 41 , j. l. , dsert , j .-
m . , kabath , p. , et al . , 2011 ,
apj , 743 , 92 beaulieu , j. p. ; carey , s. ; ribas , i. and tinetti , g. , 2008 , apj , 677 , 1343 beaulieu , j. p. , kipping , d. m. , batista , v. , tinetti , g. , ribas , i. , carey , s. et al . , 2010 , arxiv,0909.0185 belouchrani a. , abed - meraim k. , cardoso j.f . , moulines e. , 1997 , ieee trans .
signal processing , vol .
45 , 434 boisse , i. , bouchy , f. , hbrard , g. , et al . , 2011 ,
a&a , 528 , a4 borucki , w. j. , dunhm , e. w. , koch , d. g. , cochran , w. d. , et al . , 1996 ,
astrophys . & space science 241 , 111 , w. j. , koch , d. , basri , g. , et al . , 2010 , science , 327 , 977 , w. j.,koch , d. g. , basri , g. et al .
2011 , apj , 736 , 19 brillouin , l. , 1953 , j. of applied physics , 24 , 1152 brockwell p.j .
, richard a.d . , 2006 ,
time series : theory and methods ( second edition) , springer verlag , isbn : 1 - 4419 - 0319 - 8 , h. , deleuil , m. , fridlund , m. , alonso , r. , et al .
, 2010 , a&a , 519 , a51 burke , c. j. , mccullough , p. r. , valenti , j. a. , johns - krull , c. m. , janes , k. a. et al . , 2007 , apj , 671 , 2115 burke , c. j. , mccullough , p. r. , bergeron , l. e. , et al . , 2010 , apj , 719,1796 , d. a. , kolodziejczak , j. j. , van cleve , et al . , 2010 , apjl , 713 , l92 carter j.a .
, winn j.n . ,
apj , 2009 , apj , 704 , 51 t collier cameron , a. , wilson , d. m. , west , r. g. , hebb , l. , et al .
, 2007 , mnras , 380 , 1230 charbonneau , d. , brown , t. , m. , latham , d. , w. , mayor , m. , 2000 , apj , 529 , l45 charbonneau , d. , brown , t. m. , noyes , r. w. and gilliland , r. l. , 2002 , apj , 568 , 377 , d. and allen , l. e. , megeath , s. t. , torres , g. , alonso , r. , brown , t. m. , et al . , a. , 2005 , apj , 626,523 , d. , knutson , h. a. , barman , t. , allen , l. e. , mayor , m. , et al . ,
s. , 2008 , apj , 686,1341 cichocki a. , amari s. , 2002 , john wiley & sons inc .
, adaptive blind signal and image processing , isbn : 0471607916 claret , a. , 2000 , a@xmath115a , 363 , 1081 comon , p. , 1994
, signal processing , 36 , 287 comon , p. , & jutten , c. , 2000 , handbook of blind source separation : independent component analysis and applications , academic press czesla , s. , huber , k. , f. , wolter , u. , schrter , s. , schmitt , h. , m. , m. , 2009 , a&a , 505 , 1277 davison , a. c. , 2009 , cambridge university press , statistical models , isbn : 9780521734493 , j. , cardoso , j .- f . & patanchon , g. , 2003 , mnras , 346 , 1089 , d. , richardson , l. j. & harrington , j. , 2007 , mnras , 378 , 148 doron e. , yeredor a. , 2004 , proc . of ica 2004 ,
390 ford e. b. , 2006 , apj , 642 , 505 friedman , j. , h. , 1987 , j. of the american statistical association , 82 , 249 , n. p.,pont , f. & aigrain , s. , 2011 , mnras , 411 , 2199 gibson , n. p. , aigrain , s. , roberts , s. , et al . , 2011 , arxiv:1109.3251 , m. and lanotte , a. a. , barman , t. , miller , n. , demory , b .- o . , et al . , j. , 2010 , a&a , 511 , 3 , p. c. , 2011 , mnras , 410 , 94 grillmair , c. j. , burrows , a. , charbonneau , d. , armus , l. , stauffer , j. , meadows , v. , van cleve , j. , von braun , k. & levine , d. , 2008 , nature , 456 , 767 harman , h. , h. , 1967 , modern factor analysis , university of chicago press , 2nd edition , a. p. , fridlund , m. , nachmani , g. , mazeh , t. , et al .
2010 , arxiv : 1105.3372v1 huber , p. , j. , 1985 , the annals of statistics , 13 , 435 hyvrinen a.,1999 , ieee trans . on neural networks , 10(3 ) ,
626 hyvrinen a. , pajunen , p. , 1999
, neural networks , 12 , 429 hyvrinen a. , oja , e. , 2000 , neural networks , 13 , 411 hyvrinen a. , karhunen j. , oja e. , 2001 , john wiley & sons inc . , independent component analysis , isbn : 0 - 471 - 40540-x , j. m. , caldwell , d. a. , chandrasekaran , et al .
, 2010 , apjl , 713 , l87 jolliffe i.t . , 2002
, springer verlag new york inc .
, isbn : 0 - 387 - 95442 - 2 , h. a. , charbonneau , d. , allen , l. e. , fortney , j. j. , et al . , 2007 , nature , 447,183 , h. a. , charbonneau , d. , noyes , r. w. , brown , t. m. , gilliland , r. l. , 2007 , apj , 655,564 , h. a. , madhusudhan , n. , cowan , n. b. , christiansen , j. l. , agol , e. , deming , d. , et al .
, 2011 , arxiv : 1104.2901 , d. g. , borucki , w. j. , basri , g. , et al . , 2010 , apjl , 713 , l79 koldovsk z. , tichavsk p. , oja e. , ieee trans .
on neural networks , 17(5),1265 koldovsk z. , tichavsk p. , oja e. , 2005 , proc.of ieee / sp 13th workshop on stat .
signal processing kullback , s. , leibler , r.a . , 1951 , annals of mathematical statistics 22 , 79 lagarias , j.c . ,
j. a. reeds , m. h. wright , & p. e. wright , 1998siam journal of optimization , 9(1 ) , 112 lu , h. , zhou , h. , wang , j. , et al . , 2006 , apj , 131 , 790 , d. , farusi , a. , baccigalupi , c. , et al .
, 2002 , mnras , 334 , 53 maino , d. , donzelli , s. , banday , a. , j. , stivoli , f. , baccigalupi , c. , 2007 , mnras , 374 , 1207 mandel , k.and agol , e. , 2002 , apjl , 580 , l171 manly , b. , j. , f. , 1994 , multivariate statistical methods - a primer , 2nd ed . ,
chapman & hall marcy , g. , w. , butler , r. , p. , vogt , s. , s. , fischer , d. , lissauer , j. , j. , 1998 , apjl , 505 , 147 mayor , m. , queloz , d. , 1995 , nature , 378 , 355 nadaraya , e. a. , 1964 , prob . & its applic . , 10 , 186 nelder , j. a.m mead , r. , 1965 , computer journal , 7 , 308 oja , e. , 1992 , neural networks , 5 , 927 pearson , k. , 1901 , phil .
, 2 , 559 , f. , aigrain , s. & zucker , s. , 2010 , mnras , 411 , 1953 press , w. h. , teukolsky , s. a. , vetterling , w. t. , flannery , b. p. , 2007
, numerical recipes , cambridge uni . press , isbn :
978 - 0 - 521 - 88407 - 5 redfield , s. , endl , m. , cochran , w. d. and koesterke , l. , 2008 , apjl,673,l87 riley , k. f. , hobson , m. , p. , bence , s. j. , 2002 , mathematical methods for physics and engineering , cambridge uni . press , isbn : 0 - 521 - 89067 - 5 , j. , dedieu , c. , le sidaner , p. , savalle , r. , zolotukhin , i. , 2011 , a&a , 532 , 79 , s. and malln - ornelas , g. , 2003 , apj , 585 , 1038 shannon , c. , 1948 , bell system tech . j. , 27 , 379 , d. k. , pont , f. , aigrain , s. , et al . , 2011 , mnras , 416 , 1443 , i. a. g. and albrecht , s. and de mooij , e. j. w. and le poole , r. s. , 2008 , a&a , 487 , 357 , i. a. g. and de kok , r. j. and de mooij , e. j. w. and albrecht , s. , 2010a , nature , 465 , 1049 , i. a. g. and de mooij , e. j. w. and burrows , a , 2010b , a&a , 513 , 76 , k. b. , harrington , j. , nymeyer , s. , madhusudhan , n. , seager , s. , et al .
, 2010 , nature , 464 , 1161 stivoli , f. , baccigalupi , c. , maino , d. , stomper , r. , 2006 , mnras , 372 , 615 stone , j. , v. , 2004 , independent component analysis : a tutorial introduction , a bradford book swain , m. r.,vasisht , g. and tinetti , g. , 2008 , nature , 452 , 329 , m. r. , vasisht , g. , tinetti , g. , bouwman , j. , chen , p. , et al , 2009 , apjl , 690 , l114 , m. r. , tinetti , g. , vasisht , g. , deroo , p. , griffith , c. , et al . ,
d. , 2009 , apj , 704 , 1616 tichavsk p. , doron e. , yeredor a. , gomez - herrero g. , 2006 , proc .
eusipco-2006 tichavsk p. , doron e. , yeredor a. , nielsen j. , 2006 , proc .
eusipco-2006 , g.,vidal - madjar , a. , liang , m .- c . , beaulieu , j .-
p . , yung , y. , et al . , f. , 2007 , nature , 448 , 169 , g. , deroo , p. , swain , m. r. , griffith , c. a. , vasisht , g. , et al . , p. , 2010
, apjl , 712 , l139 , a. and deroo , p. and swain , m. r. , 2010 , a&a , 523 , 35 waldmann i.p . , drossart p. , tinetti g. , griffith c.a . , swain m. , deroo p. , apj , in press wang , j. , xu , h. , gu , j. , an , t. , 2010 , arxiv : 1008.3391v1 watson , g.s . , 1964 , sanky series a , 26 , 359 winn , j. n. , holman , m. j. , henry , g. w. , et al . , 2007 , aj , 2007 , 133 , 1828 yeredor a. , 2000 , ieee sig . proc
. letters , 7 , 197 this appendix provides some additional notes to the methods employed in this paper . for a more in - depth discussion of the topics presented here , please refer to the cited publications .
in gaussian statistics , our probability densities are fully defined by the first and second statistical moments , i.e. their means and covariances .
two random vectors , @xmath126 and @xmath127 , are said to be uncorrelated when their covariance ( @xmath128 ) is zero : @xmath129 ) ( { s_{l+1 } } - e [ { s_{l+1 } } ] ) ] \\\nonumber & = & e[{s_l } , { s_{l+1 } } ] - e [ { s_l}]e [ { s_{l+1 } } ] = 0\end{aligned}\ ] ] where @xmath130 $ ] is the expectation value of @xmath31 which can be approximated by the mean in this case by @xmath131 \approx \frac{1}{m } \sum_{t=1}^{m}s_{l}(t ) \label{expectation}\ ] ] with @xmath132 being the number of data points in the time series .
furthermore , we define two random variables ( @xmath31 and @xmath133 ) to be orthogonal , when both their expectation values , in addition to their covariance are zero : @xmath134 = \text{e}[{s_{l+1 } } ] = { \bf c}_{{s}_l , { s}_{l+1 } } = 0\ ] ] we can always find an affine , linear transformation from a correlated set of variables to an orthogonal one . finally , our two random vectors @xmath31 and @xmath133 are independent from one another if and only if the joined probability distribution @xmath135 of both signals are factorizable into the product of their marginal pdfs , @xmath136 and @xmath137 : @xmath138 and satisfy the property @xmath139 = \text{e } [ g ( { s_{l } } ) ] \text{e } [ h ( { s_{l+1 } } ) ] \label{independence2}\ ] ] where @xmath140 and @xmath141 are absolutely integrable functions of @xmath142 and @xmath143 respectively . from the definition of independence in equation [ independence2 ] , we obtain the definition of uncorrelatedness ( equation [ ortho ] ) in the special case where both @xmath144 and @xmath145 are linear and are only defined by their covariances ( i.e. no higher order statistical moments ) @xcite . in other words , uncorrelatedness is a special case of independence .
uncorrelated gaussian random variables are always also independent and the definitions of uncorrelatedness , orthogonality ( for zero mean ) and statistical independence become identical .
the covariance matrix of @xmath102 , @xmath146 , is given by @xmath147 , where @xmath148 is the matrix of eigenvectors and @xmath149 the diagonal matrix of eigenvalues , @xmath150 .
using principal component analysis ( pca ) , we compute * e * and * d * and the whitening matrix is hence the inverse square root covariance matrix @xmath67 is then given by equation [ vwhite ] @xcite .
@xmath151 @xmath152 where @xmath153 and is the de - mixing matrix of the whitened observed signals @xmath63 .
at the heart of the algorithm lies the blind - source separation routine . to attain the demixing matrix @xmath95 ,
many different types and varieties of algorithms are being used in the literature . here
we will use the multi - combi algorithm developed by @xcite combining a fixed point high - order ica algorithm to separate non - gaussian sources with a second - order statistics blind - source - separation ( bss ) algorithm for separating auto - regressive ( ar ) sources .
we will briefly outline these algorithms and explain how it is applied to the whitened data @xmath63 obtained in section [ pca ] .
in section [ icadef ] we briefly outlined the measures of non - gaussianity , negentropy ( equation [ negentropy ] ) , used throughout this paper and stated that negentropy can be approximated via the kurtosis of the random vector @xmath35 or via the use of contrast functions , equation [ negentropy2 ] & [ nonlin1 ] . in section [ fastica ]
we showed stated the iteration scheme to obtain a single independent component ( ic ) at a time .
this is called a deflationary algorithm where the computed ic is subtracted from the data before the second ic is computed .
such an iteration scheme has the property of finding the ics in the order of decreasing non - gaussianity .
however , the main drawback of a serial computation of ics is that estimation errors in the first ics propagate in the extraction of later ics via the orthogonalization step . this effect is cumulative and may signficantly impair weaker ics .
this predicament can be circumvented by estimating all ics in parallel . similar to the single unit iteration ,
the whitened demixing matrix @xmath95 is at its most mutually independent when the projection @xmath154 is at its most non - gaussian .
the fastica fixed point iteration step is then given by @xmath155\bf{\tilde{w } } \label{ica1}\ ] ] where @xmath156 is the unnormalised next iteration of @xmath95 , @xmath157 is an n x 1 vector of 1 s and @xmath158 and @xmath159 are the first and second order derivatives of the nonlinear function @xmath160 : @xmath161 this is followed by a symmetric orthogonalisation step : @xmath162 equations [ ica1 ] & [ ica2 ] are iterated until the result has converged . for a full derivation
we refer you to @xcite and @xcite .
whereas the convergence of the fastica algorithm is often dependent on the non - linearity chosen by the user , the efica @xcite algorithm employed here is a variant of the above iteration scheme and allows for different non - linearities to be assigned adaptively to different sources .
@xcite showed that efica is asymptotically efficient , ie .
reaches the cramer - rao lower bound ( crlb ) in an ideal case where the nonlinearity @xmath160 equals the score function . to assert a good degree of separation
, we can define @xmath163 as the gain matrix . for a perfectly estimated de - mixing matrix , @xmath21 ,
the gain matrix is equal to its identity matrix @xmath164 in signal processing , the performance of blind - source separation algorithms is usually measured by the interference over signal ratio matrix , @xmath116 @xmath165 where @xmath166 and @xmath167 denote the observed and estimated sources
. the isr for and individual observed signal @xmath166 is given by @xmath168 however , the original mixing matrix , * a * , is not generally known for real data sets and equations [ isr ] & [ isrk ] are only useful in the case of simulations .
@xcite have shown that the whole * isr * matrix for the efica algorithm can be approximated by @xmath169 @xmath170 \\\nonumber \tau_{k } & = |\mu_{k } - \rho_{k}| \\
\nonumber \rho_{k } & = e[g_{k}'(\hat{s}_{k } ] \\\nonumber \beta_{k } & = e[g_{k}^{2 } ( \hat{s}_{k } ] \label{isref}\end{aligned}\ ] ] where @xmath171 and @xmath172 are the kth and lth observed and estimated signals of * s * in equation [ timeseries3 ] , @xmath173 and @xmath174 the first and second derivative of @xmath160 for signal @xmath166 and @xmath13 is the number of signals estimated . here it should be mentioned that , of course , the true realisation of each * isr * component is unknown and a mean-*isr * is computed leading to the best on average separation of the signals .
whilst efica is optimised for the separation of instantaneously mixed , non - gaussian sources , second - order statistics bss algorithms rely on time - structure in the sources correlation function to estimate @xmath95 .
a variety of algorithms exist in the literature , here we use a derivative of the popular sobi algorithm @xcite , wasobi @xcite to separate gaussian auto - regressive ( ar ) sources in the input data @xmath63 . here
, the blind source separation follows the same linear model as in equation [ timeseries3 ] and the mixing matrix @xmath69 is estimated by a joint diagonalisation of the signals autocorrelation matrices .
the unknown correlation matrices of the observed signals for a given lag @xmath112 , @xmath175 $ ] @xmath176 \overset{\triangle}{= } \frac{1}{n } \sum_{n=1}^{n } \textbf{x}[n ] \textbf{x}^{t}[n+\tau ] , ~~ \tau = 0 , ... , m-1 \label{wasobi1}\ ] ] satisfies the relation where @xmath177 \overset{\triangle}{= } e [ \textbf{s}[n ] \textbf{s}^{t}[n+\tau]]$ ] are the source signals diagonalised correlation matrices @xcite .
hence , if the correlation matrices are diagonal , ie .
the off - diagonal components are zero , the separated signals can be said to be independent from each other .
the sobi & wasobi algorithms estimate @xmath69 as the joint diagonoliser of a set of correlation matrices .
similar to the efica code , we can define an asymptotic estimate of the * isr * matrix @xmath178}{\sigma_{l}^{2}r_{k}[0 ] } \label{isrwa}\ ] ] @xmath179 \label{phi}\ ] ] where @xmath166 and @xmath167 denote the observed and the estimated sources , @xmath180\}^{m-1}_{\tau=0}$ ] is the covariance sequence of the @xmath166-th source , @xmath181 is the variance of the source and @xmath182 are the auto - regression coefficients of the @xmath167-th source @xcite .
the algorithms introduced above are highly complementary to each other . whilst efica has an asymptotically efficient performance in separating non - gaussian instantaneous mixtures , wasobi is asymptotically efficient in separating gaussian time - correlated signals .
both these properties are necessary since a real data set will have both of the aforementioned properties and its components would hence not be optimally de - mixed if one would only employ one type of algorithm .
multi - combi @xcite uses a clustering technique in which both algorithms are run on the set of unseparated sources @xmath63 and their interference over signal matrices , @xmath183 and @xmath184 , are estimated .
the signals are then clustered depending on whether their specific @xmath185 is lower for the efica or wasobi case .
then , the process is repeated until all clusters are singeltons , ie . only contain one signal per cluster , and the signals are hence optimally separated .
from the multi - combi algorithm , we obtain the estimated signal matrix @xmath83 , an overall * isr * matrix as well as final @xmath183 and @xmath184 . since the algorithms used here use fixed - point convergence techniques , the problem of non - repeatability of the separation process is less than for neural network based approaches .
however , it is common sense to check the stability of the result obtained and to estimate the error on @xmath83 . in order to estimate the stability of the convergence , we perturb the unknown mixing matrix * a * with a random and known
mixing matrix @xmath79 to give a new mixing matrix @xmath186 and equation [ timeseries3 ] becomes : @xmath187 .
this is equivalent to multiplying the whitened signal @xmath63 with * p * @xmath188 we re - run the separation step and estimate @xmath189 . since * p * is known , we can reconstruct the original mixing - matrix and compare it with the new result . in the scope of an automated algorithm ,
the sum of all terms of @xmath190 is compared to the sum of @xmath191 and the result is reported . to identify the stochastic nature of the retrieval we furthermore re - run the separation step with the same whitened signal , @xmath63 , akin to a monte carlo simulation .
we perform @xmath192 realisations ( where @xmath193 typically ) and use the de - mixing matrices @xmath194 to construct mean noise models later on . this way , we propagate the signal separation error to the model - fitting in a coherent manner .
in order to identify the non - white ( i.e. systematic ) signals in our estimated signal matrix @xmath83 , we use the ljung - box portmanteau test @xcite .
the test statistic , usually denoted by @xmath195 , is defined by summing the normalised autocorrelations of the individual time series , @xmath84 over a range of lags : where @xmath197 is the autocorrelation at lag @xmath112 and @xmath61 is the number of observations in the time series .
the hypothesis of the time series being solely white noise is rejected if @xmath195 is bigger than a pre - specified fraction of the chi - squared distribution | the characterisation of ever smaller and fainter extrasolar planets requires an intricate understanding of one s data and the analysis techniques used .
correcting the raw data at the 10@xmath0 level of accuracy in flux is one of the central challenges .
this can be difficult for instruments that do not feature a calibration plan for such high precision measurements . here
, it is not always obvious how to de - correlate the data using auxiliary information of the instrument and it becomes paramount to know how well one can disentangle instrument systematics from one s data , given nothing but the data itself .
we propose a non - parametric machine learning algorithm , based on the concept of independent component analysis , to de - convolve the systematic noise and all non - gaussian signals from the desired astrophysical signal .
such a ` blind ' signal de - mixing is commonly known as the ` cocktail party problem ' in signal - processing . given multiple simultaneous observations of the same exoplanetary eclipse , as in the case of spectrophotometry , we show that we can often disentangle systematic noise from the original light curve signal without the use of any complementary information of the instrument . in this paper , we explore these signal extraction techniques using simulated data and two data sets observed with the hubble - nicmos instrument . another important application is the de - correlation of the exoplanetary signal from time - correlated stellar variability . using data obtained by the kepler mission we show that the desired signal can be de - convolved from the stellar noise using a single time series spanning several eclipse events .
such non - parametric techniques can provide important confirmations of the existent parametric corrections reported in the literature , and their associated results .
additionally they can substantially improve the precision exoplanetary light curve analysis in the future . |
gw gem ( @xmath5 , @[email protected] , @[email protected] ) was announced by hoffmeister ( 1949 ) to be an algol - type variable .
subsequently , eclipse timings of the system have been reported assiduously by numerous workers , and history is now long enough to understand the binary s period behavior .
light curves have been made only by broglia & conconi ( 1981a , hereafter b&c ) from photoelectric observations during the seasons of 1978 and 1979 .
unfortunately , the comparison star ( @xmath8 ) used in their observations turned out to be a @xmath9 scuti - type variable with a peak - to - peak amplitude of about 0.035 mag ( broglia & conconi 1981b ) .
b&c analyzed their @xmath10 light curves , and concluded that gw gem is a semi - detached system with the secondary star filling its inner roche lobe .
they also suggested that the period of the binary system be considered constant .
most recently , shaw ( 1994 ) included the system in his second catalog of near - contact binaries but with an unknown subclass . in this article ,
we report and analyze multiband light curves and present the first detailed analysis of the @xmath11@xmath12 diagram .
our ccd photometry of gw gem was performed on 13 nights from 20 december 2007 through 1 march 2008 in order to obtain multicolor light curves .
the observations were taken with a site 2k ccd camera and a @xmath13 filter set attached to the 61-cm reflector at sobaeksan optical astronomy observatory ( soao ) in korea .
the instrument and reduction method are the same as those described by lee et al .
( 2007 ) . since the field of view of an individual ccd image was large enough to observe a few tens of nearby stars simultaneously , we monitored many of them on each frame .
2mass 07530545 + 2716551 ( @[email protected] , @[email protected] from the nomad catalogues ; zacharias et al . 2005 ) was used as a comparison star and no brightness variation of it was detected against measurements of the other monitoring stars during our observing runs .
a total of 2,078 individual observations was obtained among the four bandpasses ( 508 in @xmath14 , 511 in @xmath6 , 532 in @xmath15 , and 527 in @xmath16 ) and a sample of them is listed in table 1 .
the light curves are plotted in figure 1 as differential magnitudes of @xmath17 _ versus _ orbital phase .
from our ccd observations , times of minimum light in each filter have been determined with the method of kwee & van woerden ( 1956 ) . weighted mean
timings from these determinations and their errors are given in table 2 , together with all other photoelectric and ccd timings .
the second column gives the standard deviation of each timing . in all , there were assembled 155 eclipse timings ( 48 photographic plate , 80 visual , 2 photographic , 10 photoelectric and 15 ccd ) from the literature and from our ccd measures .
except for the photoelectric and ccd minima , the other timings were extracted from the modern database published by kreiner et al .
an error for each method was assigned : @xmath20.0230 d for photographic plate , @xmath20.0059 d for visual , @xmath20.0022 d for photographic , and @xmath20.0012 d for photoelectric and ccd minima .
relative weights were then scaled from the inverse squares of these values ( lee et al .
2007 ) .
first of all , we constructed an @xmath11@xmath12 diagram for gw gem using the light elements of kreiner et al .
: @xmath18 the resulting @xmath11@xmath19 residuals calculated with equation ( 1 ) are listed in the fourth column of table 2 and drawn in the upper panel of figure 2 , where the timings are marked by different symbols according to observational method .
the general trend of these @xmath11@xmath12 residuals from 1929 to 2008 is a curvilinear pattern .
thus , by introducing all times of minimum light into a parabolic least - squares fit , we obtained the following quadratic ephemeris : @xmath20 the 1@xmath21-value for the last decimal place of each ephemeris parameter is given in parentheses .
the result is represented as a continuous curve in the upper panel of figure 2 .
the @xmath11@xmath22 residuals from this equation are given in the fifth column of table 2 and are plotted in the lower panel of figure 2 .
as can be seen in figure 2 , the upward - parabola ephemeris provides a good fit to the @xmath11@xmath12 residuals and signifies a continuous period increase at a rate of @xmath1([email protected])@xmath310@xmath23 d yr@xmath24 , corresponding to a fractional period change of @xmath1([email protected])@xmath310@xmath4 . within its error
, this agrees well with the value of @xmath11.1@xmath310@xmath4 calculated later in this paper with the wilson - devinney binary code ( wilson & devinney 1971 , hereafter wd ) .
because our light - curve synthesis shows that gw gem is in a semi - detached configuration with the less massive secondary filling its inner roche lobe , such a positive quadratic term can be produced by conservative mass transfer from the secondary to the primary component . from the masses of both components listed in table 6 ,
we calculate a modest mass transfer rate of 3.2@xmath310@xmath23 m@xmath25 yr@xmath24 .
as shown in figure 1 , the morphology of the light curve of gw gem resembles that of @xmath26 lyr ( dissimilar eclipse depths and light variability continuous with phase ) and therefore indicates a significant temperature difference between the two components and significant distortion of the photospheres . in order to derive reasonable representations of the binary system
, we analyzed simultaneously both our and the b&c light curves by applying the wd synthesis code to all individual observations .
for this purpose , we established unit light level at phase 0.75 and used an observational weighting scheme identical to that for the eclipsing binary ru umi ( lee et al .
table 3 lists the light - curve sets for gw gem analyzed in this paper and the standard deviations ( @xmath21 ) of a single observation .
the noise of the soao light curves is somewhat larger than that of the b&c ones despite the variability of their comparison star .
this is presumably due to the smaller telescope used at soao , and/or because our comparison star is significantly fainter than gw gem . the effective temperature of the brighter , and presumably more massive , star was initialized at 7700 k from flower s ( 1996 ) table , because ( @xmath7)[email protected] at min ii ( i.e. , phase 0.50 ) as given by b&c and because of the small reddening , @xmath27(@xmath7)[email protected] , calculated following schlegel et al .
there is a small systematic error made in the temperature assignment because the eclipses are not complete and a minor lune of the cool star is visible at secondary minimum
. the temperature of the hot star could be as much as 100 k hotter than has been assigned but not nearly so hot as if the simbad spectral type of a4 were chosen .
it is appropriate to regard the hot star envelope as a radiative atmosphere .
the logarithmic bolometric ( @xmath28 , @xmath29 ) and monochromatic ( @xmath30 , @xmath31 ) limb - darkening coefficients were interpolated from the values of van hamme ( 1993 ) and were used in concert with the model atmosphere option .
the light curves were analyzed in a manner similar to those of the near - contact binaries ax dra ( kim et al . 2004 ) and ru umi ( lee et al .
2008a ) . in this section and the next one , subscripts 1 and 2
refer to the stars eclipsed at primary and secondary minima , respectively .
although a photometric solution of gw gem was reported by b&c from the analysis of their own light curves , there is no spectroscopic mass ratio ( @xmath32 ) for the system .
we therefore conducted an extensive @xmath32-search procedure for various modes of the wd code so as to understand generally the geometrical structure and the photometric parameters of the system ( cf .
lee et al .
in this procedure , the method of multiple subsets ( wilson & biermann 1976 ) was used to ensure the stability of the result . only the photometric solutions for mode 5
( semi - detached systems in which the secondary components fill their inner critical surfaces ) are acceptable for gw gem .
this configuration is also consistent with mass transfer from the secondary star to the hotter , more massive primary star inferred from our period study . to confirm the preliminary result
, @xmath32-searches were repeated for each data set individually .
the weighted sums of the squared residuals ( @xmath33 = @xmath34 ) as a function of @xmath32 are displayed in figure 3 .
here , circles , squares , and diamonds represent the search results for b&c , soao , and all data sets , respectively .
the optimal solution is close to @xmath32=0.46 .
the value of @xmath32 was treated as an adjustable parameter in all subsequent calculations to derive photometric solutions .
the best result is listed in table 4 and plotted in figure 4 , where for clarity individual observations have been compiled into 200 mean points using bin widths of 0.005 in phase for each filtered light curve . as seen in the figure , the computed light curves describe the b&c data satisfactorily , but do not fit the soao data around phase 0.25 as well as can be wished .
the discrepancies increase toward longer wavelengths and these can be modelled by a cool spot on the secondary star which must have a convective atmosphere . however , this representation is not unique for one can not exclude a possible hot spot on the primary star due to impact from mass transfer between the components .
thus , two different model spots were postulated to reanalyze the soao light curves .
we used the unspotted photometric parameters as initial values and included @xmath35 and @xmath36 as additional free variables .
final results are given in table 5 and , within errors , each of these is in excellent agreement with the geometry of the unspotted solution and fits the soao light curves better than the unspotted model .
the light residuals from the spot models are plotted in figure 5 and it can be seen that there is no systematic trend among them . during the evaluation with spot parameters , we searched for a possible third light source but none was detected in the light - curve analysis .
it is numerically impossible to discriminate between the two spot models but there is more information to be weighed .
it is noteworthy that the older b&c light curves show no obvious asymmetry but the evidence of the period study is that mass transfer was ongoing for at least the last 80 years .
we could therefore expect an impact hot spot to have also existed during the 1978 and 1979 seasons but the light curve asymmetry in the b&c data did not appear .
the easiest way to reconcile this limited and discontinuous information is to postulate that an inconspicuous impact hot spot has always existed due to the feeble mass transfer .
such a spot would be relatively inconspicuous since the hot star is large with respect to its roche lobe and the free - fall height onto its photosphere is not great .
kinetic heating from the transferring gas is therefore modest and the effect in the light curves below the level of precision of the measures .
finally , for reasons that can not be known , the cool star had developed a conspicuous cool spot by the time of the soao observations .
therefore , the cool spot model would be a more reasonable interpretation , and what has been modelled is the differential effect between a postulated meager hot spot and the more extensive transient cool spot .
nothing can be known of photometric activity between the older and newer light curves but it would have to be true that a trifling hot spot continued to be present in that interim .
based on the historical and new photometry , we have presented the first detailed period analysis and light - curve synthesis for gw gem .
the absolute stellar parameters for the system can be computed from our complete photometric solutions with the cool - spot model of table 5 and from harmanec s ( 1988 ) relation between spectral type and physical parameters ( effective temperature , mass , and radius ) . the dereddened color ( @xmath7)@[email protected] and temperature of the primary component correspond to a normal near - main - sequence star with a spectral type of about a7 .
the astrophysical parameters are listed in table 6 , where the radius ( @xmath38 ) of the secondary star results from the ratio ( @xmath39=0.856 ) of the mean - volume radii for each component .
the bolometric corrections ( bcs ) were obtained from the scaling between @xmath40 and bc recalculated by kang et al .
( 2007 ) from flower s table .
an apparent visual magnitude of @xmath6=+10.49 at maximum light ( b&c ) , appropriate dereddening , and our computed light ratio at phase 0.75 lead to @xmath41=+10.59 and @xmath42=+13.05 for the primary and secondary stars , respectively . using the interstellar reddening of @xmath43=0.10 , the calculated value of @xmath41 , and the expected value of @xmath44 for the primary star
, we calculated an approximate distance to the system of about 120 pc .
a comparison of the gw gem parameters with the mass - radius , mass - luminosity , and hertzsprung - russell diagrams ( hilditch et al . 1988 ) clearly demonstrates that the primary component lies between the zero - age and terminal - age main - sequence loci , while the secondary is oversized and overluminous for its mass . in these diagrams , the locations of the two component stars fall amid those of previously - known near - contact binaries .
thus , the system is a semi - detached and fo vir - subtype , near - contact binary consisting of a detached main - sequence primary component with a spectral type of a7 and an evolved secondary component with a spectral type of approximately k1 which fills its limiting lobe completely .
such a configuration supports the concept of mass transfer from the secondary to the primary component as indicated by our period analysis .
our results suggest that gw gem currently is in a state of broken contact evolving from a contact configuration as predicted by thermal relaxation oscillation theory ( lucy 1976 , lucy & wilson 1979 ) .
high - resolution spectroscopy will determine the astrophysical parameters and evolutionary status of the system better than is possible with photometry alone .
the authors wish to thank prof .
chun - hwey kim for his help using the @xmath11@xmath12 database of eclipsing binaries , and ms . in kyung
baik for collecting the b&c observations on gw gem .
we also thank the staff of the sobaeksan optical astronomy observatory for assistance with our observations .
we have used the simbad data base maintained at cds many times and appreciate its availability .
brt , l. , zejda , m. , & svoboda , p. 2007
, open european journal on variable stars , 74 , 1 broglia , p. , & conconi , p. 1981a
, a&as , 46 , 185 broglia , p. , & conconi , p. 1981b
, a&a , 100 , 201 dvorak , s. w. 2003 , inf . bull .
variable stars , no .
5378 dvorak , s. w. 2006 , inf .
variable stars , no .
5677 flower , p. j. 1996 , apj , 469 , 355 harmanec , p. 1988 , bull .
czechoslovakia , 39 , 329 hilditch , r. w. , king , d. j. , & mcfarlane , t. m. 1988 , mnras , 231 , 341 hoffmeister , c. 1949 , erg .
, 12 , 1 hbscher , j. 2007 , inf .
variable stars , no .
5802 kang , y .- w . ,
hong , k .- s . , & lee , j. 2007 , in asp conf .
362 , the seventh pacific rim conference on stellar astrophysics , ed .
kang et al .
( san francisco : asp ) , 19 kim , h .-
i . , lee , j. w. , kim , c .- h . , youn , j .- h . ,
kwon , s .-
koch , r. h. 2004 , pasp , 116 , 931 kreiner , j. m. , kim , c .- h .
, & nha , i .- s . 2001 , an atlas of o - c diagrams of eclipsing binary stars ( krakow : wydawn . nauk .
pedagogicznej ) kwee , k. k. , & van woerden , h. 1956 , bull .
netherlands , 12 , 327 lee , j. w. , kim , c .- h . , & koch , r. h. 2007 , mnras , 379 , 1665 lee , j. w. , kim , c .- h . , kim , s .- l . , lee , c .- u . , han , w. , & koch , r. h. 2008a , pasp , 120 , 720 lee , j. w. , youn , j .- h . ,
kim , c .- h . ,
lee , c .- u .
, & kim , h .-
2008b , aj , 135 , 1523 lucy , l. b. 1976 , apj , 205 , 208 lucy , l. b. , & wilson , r. e. 1979 , apj , 231 , 502 nagai , k. 2003 , vsolj variable star bull .
, 40 , 3 nagai , k. 2004 , vsolj variable star bull . , 42 , 3 nagai , k. 2006 , vsolj variable star bull . , 44 , 5 nelson , r. h. 2001 , inf .
variable stars , no .
5040 nelson , r. h. 2003 , inf . bull .
variable stars , no .
5371 ogloza , w. , niewiadomski , w. , barnacka , a. , biskup , m. , malek , k. , & sokolowski , m. 2008 , inf . bull .
variable stars , no .
5843 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , apj , 500 , 525 shaw , j. s. 1994 , mem .
italiana , 65 , 95 van hamme , w. 1993 , aj , 106 , 2096 wilson , r. e. , & biermann , p. 1976 , a&a , 48 , 349 wilson , r. e. , & devinney , e. j. 1971 , apj , 166 , 605 zacharias , n. , monet , d. g. , levine , s. e. , urban , s. e. , gaume , r. , & wycoff , g. l. 2005 , vizier online data catalog , 1297 , 0 cccccccc 455.20948 & @xmath451.321 & 455.20594 & @xmath451.001 & 455.20703 & @xmath450.756 & 455.21267 & @xmath450.510 + 455.22320 & @xmath451.317 & 455.21069 & @xmath450.985 & 455.21171 & @xmath450.779 & 455.21700 & @xmath450.526 + 455.22764 & @xmath451.332 & 455.21502 & @xmath450.966 & 455.21604 & @xmath450.796 & 455.22628 & @xmath450.525 + 455.23177 & @xmath451.342 & 455.21933 & @xmath450.994 & 455.22032 & @xmath450.759 & 455.23065 & @xmath450.521 + 455.23582 & @xmath451.302 & 455.22436 & @xmath450.995 & 455.22535 & @xmath450.771 & 455.23476 & @xmath450.517 + 455.24820 & @xmath451.335 & 455.22876 & @xmath450.995 & 455.23386 & @xmath450.742 & 455.24262 & @xmath450.502 + 455.26056 & @xmath451.320 & 455.23685 & @xmath450.967 & 455.24172 & @xmath450.748 & 455.24656 & @xmath450.510 + 455.26953 & @xmath451.288 & 455.24079 & @xmath450.977 & 455.24564 & @xmath450.766 & 455.25500 & @xmath450.521 + 455.27716 & @xmath451.317 & 455.25708 & @xmath450.987 & 455.25016 & @xmath450.779 & 455.26343 & @xmath450.501 + 455.28098 & @xmath451.301 & 455.26159 & @xmath450.993 & 455.25409 & @xmath450.754 & 455.26724 & @xmath450.507 + llrrrccl 43,514.53438 & @xmath20.00003 & 27,097.0 & @xmath450.00343 & @xmath450.00011 & pe & i & b&c ( 1981a ) + 43,543.55006 & @xmath20.00004 & 27,141.0 & @xmath450.00330 & @xmath450.00001 & pe & i & b&c ( 1981a ) + 43,544.5411 & @xmath20.0005 & 27,142.5 & @xmath450.00143 & 0.00186 & pe & ii & b&c ( 1981a ) + 43,577.5120 & @xmath20.0004 & 27,192.5 & @xmath450.00274 & 0.00051 & pe & ii & b&c ( 1981a ) + 43,589.3798 & @xmath20.0010 & 27,210.5 & @xmath450.00494 & @xmath450.00170 & pe & ii & b&c ( 1981a ) + 43,849.53175 & @xmath20.00006 & 27,605.0 & @xmath450.00378 & @xmath450.00080 & pe & i & b&c ( 1981a ) + 43,876.56881 & @xmath20.00006 & 27,646.0 & @xmath450.00394 & @xmath450.00098 & pe & i & b&c ( 1981a ) + 43,905.58451 & @xmath20.00001 & 27,690.0 & @xmath450.00379 & @xmath450.00086 & pe & i & b&c ( 1981a ) + 43,926.3566 & @xmath20.0003 & 27,721.5 & @xmath450.00419 & @xmath450.00129 & pe & ii & b&c ( 1981a ) + 51,650.7741 & @xmath20.0001 & 39,435.0 & 0.01215 & 0.00172 & ccd & i & nelson ( 2001 ) + 52,297.0294 & & 40,415.0 & 0.01200 & @xmath450.00001 & ccd & i & nagai ( 2003 ) + 52,585.8672 & @xmath20.0003 & 40,853.0 & 0.01319 & 0.00044 & ccd & i & nelson ( 2003 ) + 52,628.7304 & @xmath20.0002 & 40,918.0 & 0.01251 & @xmath450.00036 & ccd & i & dvorak ( 2003 ) + 52,912.2932 & & 41,348.0 & 0.01424 & 0.00064 & ccd & i & nagai ( 2004 ) + 52,976.2574 & & 41,445.0 & 0.01234 & @xmath450.00142 & ccd & i & nagai ( 2004 ) + 53,677.9092 & @xmath20.0001 & 42,509.0 & 0.01538 & @xmath450.00029 & ccd & i &
dvorak ( 2006 ) + 53,679.2286 & & 42,511.0 & 0.01589 & 0.00022 & ccd & i & nagai ( 2006 ) + 53,683.1850 & & 42,517.0 & 0.01562 & @xmath450.00006 & ccd & i & nagai ( 2006 ) + 53,730.9978 & & 42,589.5 & 0.01871 & 0.00290 & ccd & ii & nagai ( 2006 ) + 54,085.4471 & @xmath20.0002 & 43,127.0 & 0.01668 & @xmath450.00012 & pe & i & hbscher ( 2007 ) + 54,127.6524 & @xmath20.0003 & 43,191.0 & 0.01754 & 0.00062 & ccd & i & ogloza et al . (
2008 ) + 54,188.3217 & @xmath20.0013 & 43,283.0 & 0.01796 & 0.00087 & ccd & i & brt et al .
( 2007 ) + 54,461.33036 & @xmath20.00036 & 43,697.0 & 0.01667 & @xmath450.00121 & ccd & i & this paper + 54,497.27047 & @xmath20.00035 & 43,751.5 & 0.01707 & @xmath450.00092 & ccd & ii & this paper + 54,512.10740 & @xmath20.00003 & 43,774.0 & 0.01650 & @xmath450.00153 & ccd & i & this paper + 54,523.97755 & @xmath20.00007 & 43,792.0 & 0.01665 & @xmath450.00141 & ccd & i & this paper ccc @xmath46 ( hjd ) & + @xmath47 ( d ) & + d@xmath47/d@xmath48 & + @xmath32 & + @xmath49 ( deg ) & + @xmath50 ( k ) & 7700 & 5007@xmath27 + @xmath51 & [email protected] & 2.799 + fill - out ( % ) & 85.2 & 100 + @xmath28 , @xmath29 & 0.670 , 0.210 & 0.642 , 0.165 + @xmath52 , @xmath53 & 0.818 , 0.328 & 0.853 , @xmath450.016 + @xmath54 , @xmath55 & 0.712 , 0.284 & 0.799 , 0.117 + @xmath56 , @xmath57 & 0.592 , 0.249 & 0.714 , 0.176 + @xmath58 , @xmath59 & 0.483 , 0.225 & 0.617 , 0.195 + @xmath60 & 0.9482(7 ) & 0.0518 + @xmath61 & 0.9072(8 ) & 0.0928 + @xmath60 & 0.9482(9 ) & 0.0518 + @xmath61 & 0.9072(10 ) & 0.0928 + @xmath62 & 0.8658(11 ) & 0.1334 + @xmath63 & 0.8255(11 ) & 0.1745 + @xmath64 ( pole ) & 0.3508(7 ) & 0.2933(4 ) + @xmath64 ( point ) & 0.3835(11 ) & 0.4210(16 ) + @xmath64 ( side ) & 0.3628(8 ) & 0.3060(4 ) + @xmath64 ( back ) & 0.3737(10 ) & 0.3386(4 ) + @xmath64 ( volume)@xmath65 & 0.3628 & 0.3139 + cccccc @xmath66@xmath67 & & & + @xmath32 & & & + @xmath49 ( deg ) & & & + @xmath50 ( k ) & 7700 & 5004(6 ) & & 7700 & 5004(6 ) + @xmath51 & 3.276(4 ) & 2.797 & & 3.257(4 ) & 2.795 + fill - out ( % ) & 85.7 & 100 & & 86.1 & 100 + @xmath52 & 0.847(17 ) & 0.486(23 ) & & 0.846(17 ) & 0.552(23 ) + @xmath54 & 0.740(18 ) & 0.722(11 ) & & 0.745(17 ) & 0.736(11 ) + @xmath56 & 0.628(23 ) & 0.772(9 ) & & 0.646(23 ) & 0.744(9 ) + @xmath58 & 0.471(23 ) & 0.615(6 ) & & 0.490(22 ) & 0.583(6 ) + @xmath60 & 0.940(1 ) & 0.060 & & 0.942(1 ) & 0.058 + @xmath61 & 0.904(1 ) & 0.096 & & 0.906(1 ) & 0.094 + @xmath62 & 0.868(1 ) & 0.132 & & 0.868(1 ) & 0.132 + @xmath63 & 0.827(1 ) & 0.173 & & 0.827(1 ) & 0.173 + @xmath64 ( pole ) & 0.3518(15 ) & 0.2932(7 ) & & 0.3540(16 ) & 0.2930(8 ) + @xmath64 ( point ) & 0.3850(24 ) & 0.4208(31 ) & & 0.3882(26 ) & 0.4205(33 ) + @xmath64 ( side ) & 0.3640(18 ) & 0.3058(8 ) & & 0.3665(19 ) & 0.3056(8 ) + @xmath64 ( back ) & 0.3750(20 ) & 0.3384(8 ) & & 0.3778(22 ) & 0.3382(8 ) + @xmath64 ( volume ) & 0.3639 & 0.3137 & & 0.3664 & 0.3135 + colatitude ( deg ) & 86.3(7 ) & & & & 80.3(2.4 ) + longitude ( deg ) & 123.2(7 ) & & & & 112.7(3.7 ) + radius ( deg ) & 10.2(4 ) & & & &
16.2(1.4 ) + @xmath50@xmath68/@xmath50@xmath69 & 1.060(6 ) & & & & 0.754(32 ) + @xmath33 & & & + ccc @xmath70 ( m@xmath25 ) & 1.74 & 0.80 + @xmath15 ( r@xmath25 ) & 1.75 & 1.50 + @xmath71 @xmath72 ( cgs ) & 4.19 & 3.99 + @xmath73 ( l@xmath25 ) & 9.65 & 1.26 + @xmath74 ( mag ) & + 2.23 & + 4.44 + bc ( mag ) & @xmath10.03 & @xmath450.30 + @xmath75 ( mag ) & + 2.20 & + 4.74 + distance ( pc ) & + @xmath12 diagram of gw gem . in the upper panel
constructed with the light elements of kreiner et al . , the continuous curve represents the quadratic term of equation ( 2 ) .
the residuals from the quadratic ephemeris are plotted in the lower panel .
cc , pe , pg , vi , and p represent ccd , photoelectric , photographic , visual , and photographic plate minima , respectively . ] for gw gem as a function of assumed mass ratio @xmath32 .
the circles , squares , and diamonds represent the @xmath32-search results for b&c , soao , and all data sets , respectively .
the arrow indicates the minimum value of @xmath34 close to @xmath32=0.46 . ] | new multiband ccd photometry is presented for the eclipsing binary gw gem ; the @xmath0 light curves are the first ever compiled .
four new minimum timings have been determined .
our analysis of eclipse timings observed during the past 79 years indicates a continuous period increase at a fractional rate of @xmath1([email protected])@xmath310@xmath4 , in excellent agreement with the value @xmath11.1@xmath310@xmath4 calculated from the wilson - devinney binary code .
the new light curves display an inverse oconnell effect increasing toward longer wavelengths .
hot and cool spot models are developed to describe these variations but we prefer a cool spot on the secondary star .
our light - curve synthesis reveals that gw gem is in a semi - detached , but near - contact , configuration .
it appears to consist of a near - main - sequence primary star with a spectral type of about a7 and an evolved early k - type secondary star that completely fills its inner roche lobe .
mass transfer from the secondary to the primary component is responsible for the observed secular period change . |
the sagdeev potential approach @xcite has been considered to investigate the existence of solitary wave and double layer at @xmath0 , where @xmath1 is the lower bound of the mach number @xmath2 , i.e. , solitary wave and/or double layer solutions of the energy integral start to exist for @xmath3 . in
most of the earlier works , @xcite solitary wave and/or double layer solutions have been investigated for @xmath3 .
however , some recent investigations @xcite have shown that finite amplitude solitary wave can exist at @xmath0 in the parameter regime where solitons of both polarities exist .
the numerical observations @xcite of the solitary wave solution of the energy integral at @xmath0 , motivate us to set a general analytical theory for the existence of solitary wave and double layer solution of the energy integral at @xmath19 .
the present paper is purely analytical study on the existence of solitary wave and double layer solution of the energy integral at @xmath19 . without this analytical study , it is difficult to predict the existence of double layer solution at @xmath19 .
the analytical study also give the point in the parameter regime where the double layer solution at @xmath19 exists .
so , we consider the following well known energy integral , which has been commonly used in various plasma environments . @xcite
@xmath20 with @xmath21 is the sagdeev potential . here
, @xmath6 is same as @xmath22 , i.e. , the mach number @xmath2 is omitted from the notation @xmath6 when no particular emphasis is put upon it and let @xmath23 is the domain of definition of the function @xmath6 , i.e. , @xmath6 is well defined as a real number for all @xmath7 and for all @xmath8 .
hereafter , by the phrases `` for all @xmath24 '' , `` for all @xmath25 '' , and `` for all @xmath26 '' , we mean , respectively , `` all those @xmath8 which are strictly positive '' , `` all those @xmath8 which are strictly negative '' , and `` all those @xmath7 which are strictly positive '' .
if @xmath27 " indicates differentiation of @xmath6 with respect to @xmath5 , then with the help of some definitions and some well - known theorems of real analysis , we have proved the following two theorems , which are necessary to prove the main theorems regarding the existence of solitary wave and double layer at @xmath0 . + * theorem 1 :* if @xmath28 , then there exists a strictly positive real number @xmath29 ( @xmath30 ) such that @xmath31 for all @xmath32 . +
* theorem 2 :* if @xmath33 and @xmath28 , then there exists a strictly positive real number @xmath29 ( @xmath30 ) such that @xmath34 for all @xmath35 and @xmath36 for all @xmath37 .
if @xmath9 , using * theorem 1 * , * theorem 2 * , and some well - known theorems of real analysis , we have proved the following theorems to confirm the existence of solitary wave and double layer solution of the energy integral ( [ energy int ] ) at @xmath0 . +
* theorem 3 :* if @xmath10 , then at @xmath0 , there may exist either a solitary wave or a double layer in the positive potential side but there does not exist any solitary wave or double layer in the negative potential side . + * theorem 4 :* if @xmath11 , then at @xmath0 , there may exist either a solitary wave or a double layer in the negative potential side but there does not exist any solitary wave or double layer in the positive potential side . + * theorem 5 :* if @xmath28 , then it is not possible to have coexistence of both positive and negative potential solitary structures ( including double layers ) at @xmath0 . +
* theorem 6 :* if @xmath10 , @xmath38 for all @xmath39 and for all @xmath40 , and if there exists at least one value @xmath17 of @xmath2 such that the system supports positive potential solitary waves ( ppsws ) for all @xmath18 , then there exist either a ppsw or a positive potential double layer ( ppdl ) at @xmath0 . + * theorem 7 :* if @xmath11 , @xmath38 for all @xmath39 and for all @xmath41 , and if there exists at least one value @xmath17 of @xmath2 such that the system supports negative potential solitary waves ( npsws ) for all @xmath18 , then there exist either a npsw or a negative potential double layer ( npdl ) at @xmath0 . + * theorem 8 :* if @xmath10 , @xmath38 for all @xmath39 and for all @xmath40 , and if the system supports only npsws for @xmath3 , then there does not exist any ppsw at @xmath0 . + * theorem 9 :* if @xmath11 , @xmath38 for all @xmath39 and for all @xmath41 , and if the system supports only ppsws for @xmath3 , then there does not exist any npsw at @xmath0 . + * theorem 10 :* if @xmath28 , @xmath38 for all @xmath39 and for all @xmath42 , and if there exists at least one value @xmath17 of @xmath2 such that the system supports both ppsws and npsws for all @xmath18 , i.e. , if the system supports coexistence of both ppsws and npsws for all @xmath18 , then there exist either a ppsw or a ppdl at @xmath0 if @xmath10 , whereas for @xmath11 , there exist either a npsw or a npdl at @xmath0 .
apart from the conditions of * theorem-6 * ( * theorem-7 * ) , the ppdl ( npdl ) solution at @xmath0 is possible only when there exists a ppdl ( npdl ) solution in any right neighborhood of @xmath1 , i.e. , ppdl ( npdl ) solution at @xmath0 is possible only when the curve @xmath43 tends to intersect the curve @xmath0 at some point in the solution space of the energy integral , where each point of the curve @xmath43 corresponds to a ppdl ( npdl ) solution of the energy integral whenever @xmath44 .
so , without going through all qualitatively different solution spaces of the energy integral by considering the entire range of parameters involved in the system , it is not possible to make a systematic investigation of the solitary wave solution and/or double layer solution of the energy integral at @xmath0 .
the most interesting fact is the existence of the double layer solution at @xmath0 .
the existence of solitary structures at @xmath0 indicates that where and when one can apply small amplitude theory to find the weakly nonlinear solitary structures because of the following reasons : \(a ) if in a portion of the solution space , there exist solitary waves or double layers of finite amplitude at @xmath0 , then the amplitude of solitary waves corresponding to all @xmath3 is greater than that finite value . then it is unwise to study weakly nonlinear solitary structures by means of any small amplitude methods ( viz . , replacing @xmath22 by @xmath45 , where @xmath45 represents the power series expansion of the sagdeev potential or pseudo - potential @xmath22 in @xmath5 keeping terms up to @xmath46 ) in that portion of the solution space , since these methods may lead to erroneous results .
actually , it is not possible to apply small amplitude method to study the weakly nonlinear solitary structures in a parameter regime where finite amplitude solitary structures exist at @xmath0 , because in any small amplitude method , amplitude of solitary structure always tends to zero as @xmath47 which contradicts the occurrence of finite amplitude solitary structure at @xmath0 .
therefore , to study the weakly nonlinear solitary structures by means of small amplitude methods for @xmath3 , the simple prescription is to investigate the nonlinear solitary structures at @xmath0 .
so , one can not apply small amplitude theory to find weakly nonlinear solitary structures without investigating the existence of solitary structures at @xmath0 in the entire compositional parameter space .
\(b ) if in a portion of the solution space , there exist solitary structures having a finite amplitude at @xmath0 , then it is simple to find the amplitude as well as width of the solitary structure for @xmath48 ( for sufficiently small strictly positive quantity @xmath49 ) and this can be easily done by plotting @xmath22 against @xmath5 at @xmath48 .
however , if in a portion of the solution space , there does not exist solitary structures at @xmath0 , then it is very difficult to find the amplitude as well as width of the solitary wave for @xmath48 by plotting @xmath22 against @xmath5 .
in this situation , the general prescription is to use the small amplitude method for finding the weakly nonlinear solitary structure at @xmath48 .
\(c ) considering any specific problem , it can be easily checked that the solitary structure along the curve @xmath0 is same as the solitary structure for @xmath3 , if one can move the solution space through the family of curves parallel to the curve @xmath0 .
finally , we have used the analytical results as presented in this paper to a specific problem of dust acoustic wave in nonthermal plasma to investigate new results and new physical ideas .
the present paper is organized as follows : physical interpretation of the energy integral for the existence of solitary wave and double layer solutions is given in sec .
[ sec : physics ] .
sec : math ] , some mathematical preliminaries ( including definitions , properties and some basic theorems ) have been given .
using these definitions , properties and basic theorems we have proved ten important theorems in sec . [ sec : proof1 - 10 ] for the existence of solitary wave and/or double layer solution at @xmath0 .
the analytical results have been verified numerically by considering a problem of dust acoustic wave in nonthermal plasma in sec .
[ sec : verify ] .
finally , a brief conclusion has been given in sec .
[ sec : conclusion ] .
the energy integral ( [ energy int ] ) can be regarded as the one - dimensional motion of a particle of unit mass whose position is @xmath5 at time @xmath50 with velocity @xmath51 in a potential well @xmath22 .
the first term of the energy integral can be regarded as the kinetic energy of a particle of unit mass at position @xmath5 and time @xmath50 whereas @xmath22 is the potential energy of the same particle at that instant . since
kinetic energy is always a non - negative quantity , @xmath52 for the entire motion , i.e. , zero is the maximum value for @xmath22 . again from ( [ energy int ] ) , we find @xmath53 , i.e. the force acting on the particle at the position @xmath5 is @xmath54 .
suppose , @xmath55 , therefore , the particle is in equilibrium at @xmath56 because the velocity as well as the force acting on the particle at @xmath56 are simultaneously equal to zero .
now if @xmath57 can be made an unstable position of equilibrium , the energy integral can be interpreted as the motion of an oscillatory particle if @xmath58 for some @xmath59 , i.e. , if the particle is slightly displaced from its unstable position of equilibrium then it moves away from its unstable position of equilibrium and it continues its motion until its velocity is equal to zero , i.e. , until @xmath5 takes the value @xmath60 .
now the force acting on the particle of unit mass at position @xmath61 is @xmath62 . for @xmath63 , the force acting on the particle at the point @xmath64
is directed towards the point @xmath56 if @xmath65 , i.e. , if @xmath66 .
on the other hand , for @xmath67 , the force acting on the particle at the point @xmath64 is directed towards the point @xmath56 if @xmath68 , i.e. , if @xmath69 .
therefore , if @xmath69 ( for the positive potential side ) or if @xmath66 ( for the negative potential side ) then the particle reflects back again to @xmath57 . again , if @xmath70 then the velocity @xmath71 as well as the force @xmath72 both are simultaneously equal to zero at @xmath61 .
consequently , if the particle is slightly displaced from its unstable position of equilibrium ( @xmath57 ) it moves away from @xmath57 and it continues its motion until the velocity is equal to zero , i.e. , until @xmath5 takes the value @xmath61 .
however it can not be reflected back again at @xmath57 as the velocity and the force acting on the particle at @xmath61 vanish simultaneously . actually , if @xmath69 ( for @xmath67 ) or if @xmath66 ( for @xmath63 ) the particle takes an infinite long time to move away from the unstable position of equilibrium .
after that it continues its motion until @xmath5 takes the value @xmath60 and again it takes an infinite long time to come back its unstable position of equilibrium .
therefore , for the existence of a ppsw ( npsw ) solution of the energy integral ( [ energy int ] ) , we must have the following : ( a ) @xmath56 is the position of unstable equilibrium of the particle , ( b ) @xmath58 , @xmath73 @xmath74 for some @xmath75 @xmath76 , which is nothing but the condition for oscillation of the particle within the interval @xmath77 and ( c ) @xmath78 for all @xmath79 @xmath80 ) , which is the condition to define the energy integral ( [ energy int ] ) within the interval @xmath77 . for the existence of a ppdl ( npdl ) solution of the energy integral ( [ energy int ] )
, the conditions ( a ) and ( c ) remain unchanged but here ( b ) has been modified in such a way that the particle can not be reflected again at @xmath57 , i.e. , the condition ( b ) assumes the following form : @xmath81 , @xmath82 for some @xmath75 @xmath83 ) .
the above discussions for the existence of solitary waves and double layers are valid if @xmath57 is an unstable position of equilibrium , i.e. , if @xmath84 along with @xmath55 .
in other words , @xmath56 can be made an unstable position of equilibrium if the potential energy of the said particle attains its maximum value at @xmath56 .
now , the condition @xmath85 gives a lower bound @xmath1 of @xmath2 , i.e. , @xmath86 , @xmath87 , and @xmath88 .
this @xmath1 is , in general , a function of the parameters involved in the system , or a constant .
therefore , if @xmath89 , the potential energy of the said particle attains its minimum value at @xmath56 , and consequently , @xmath56 is the position of stable equilibrium of the particle , and in this case , it is impossible to make any oscillation of the particle even when it is slightly displaced from its position of stable equilibrium , and consequently there is no question of existence of solitary waves or double layers for @xmath89 .
in other words , for the position of unstable equilibrium of the particle at @xmath56 , i.e. , for @xmath90 , the function @xmath22 must be convex within a neighborhood of @xmath56 and in this case both type of solitary waves ( negative or positive potential ) may exist if other conditions are fulfilled .
now suppose that @xmath91 and also @xmath92 , then if @xmath93 , the potential energy of the said particle attains its maximum value at @xmath56 and consequently , @xmath56 is the position of unstable equilibrium . on the other hand if @xmath91 , @xmath92 , and @xmath94 , the potential energy of the said particle attains its minimum value at @xmath56 and consequently , @xmath56 is the position of stable equilibrium of the particle and in this case there is no question of existence of solitary wave solution and/or double layer solution of the energy integral ( [ energy int ] ) .
but if @xmath28 along with @xmath95 , then without going through the complete analytical investigation , it is difficult to predict the existence of solitary wave and/or double layer solution of the energy integral ( [ energy int ] ) at @xmath0 . in this situation ,
i.e. , when @xmath28 along with @xmath95 , to confirm the existence of solitary wave and/or double layer solution of the energy integral ( [ energy int ] ) at @xmath0 , some definitions and well - known theorems of real analysis are necessary .
so , in the next section , we consider only those definitions and theorems of real analysis for continuation of physical interpretation of the existence of solitary wave and/or double layer solution of the energy integral ( [ energy int ] ) at @xmath0 when @xmath28 along with @xmath95 .
* d1 :* let @xmath96 be any point of @xmath97 , the set of all real numbers and @xmath98 is a strictly positive real number , then the @xmath99 neighborhood of the point @xmath96 is denoted by @xmath100 and is defined as a set of all real numbers @xmath101 such that @xmath102 , i.e. , @xmath103 , i.e. , @xmath104 * p1 :* if @xmath105 , using eq . ( [ open nbd ] ) it is simple to check the following property : @xmath106 * d2 :* let @xmath107 be any subset of @xmath97 and @xmath108 , then @xmath96 is said to be an interior point of @xmath109 if there exists a @xmath98 such that @xmath110 , i.e. , any point of @xmath100 is also a point of @xmath109 . +
* d3 :* @xmath111 is the set of all two - dimensional points @xmath112 ,
i.e. , @xmath111 is the set of all ordered pairs @xmath112 such that @xmath113 and @xmath114 , i.e. , @xmath115 if @xmath109 and @xmath116 are any two subsets of @xmath97 , then we define @xmath117 as the set of all ordered pairs @xmath112 such that @xmath118 and @xmath119 , i.e. , @xmath120 * p2 :* if @xmath109 , @xmath116 , @xmath121 and @xmath122 are any four subsets of @xmath97 and @xmath123 , then it is simple to prove the following property : @xmath124 * p3 :* for any four subsets @xmath109 , @xmath116 , @xmath121 and @xmath122 of @xmath97 , it is simple to prove the following property : @xmath125 .
+ * d4 :* let @xmath126 be any point of @xmath111 and @xmath127 , @xmath128 are two strictly positive real numbers , then the @xmath129- rectangular neighborhood of the point @xmath126 is denoted by @xmath130 and is defined as a set of all points @xmath112 of @xmath111 such that @xmath131 and @xmath132 , i.e. , @xmath133 , i.e. , @xmath134 * p4 :* @xmath135 + * d5 :* let @xmath117 be any subset of @xmath111 and @xmath126 be a point of @xmath117 , i.e. , @xmath108 and @xmath136 , then @xmath126 is said to be an interior point of @xmath117 if there exists a @xmath129- rectangular neighborhood of the point @xmath126 with @xmath127 , @xmath128 such that @xmath137 , i.e. , any point of @xmath130 is also a point of @xmath117 , i.e. , @xmath138 , i.e. , @xmath139 and @xmath140 . *
bth1 :* let @xmath141 be a real valued function defined on a subset @xmath122 of the set of real numbers @xmath97 and @xmath141 be continuous at @xmath142 . if @xmath143 , then there exists a real number @xmath144 such that @xmath145 for all @xmath146 .
furthermore , if @xmath96 is an interior point of d , then there exists a real number @xmath147 such that @xmath145 for all @xmath148 .
+ * bth2 :* let @xmath149 be a real valued function defined on a subset @xmath150 of @xmath151 and @xmath149 be continuous at @xmath152 . if @xmath153 , then there exist two real numbers @xmath144 and @xmath154 such that @xmath155 for all @xmath156 . furthermore ,
if @xmath126 is an interior point of @xmath122 , then there exist two real numbers @xmath157 and @xmath158 such that @xmath155 for all @xmath159 . + * bolzano s theorem :* let @xmath141 be a real valued function defined on a closed interval @xmath160 $ ] ( @xmath161 ) . if @xmath141 is continuous on @xmath160 $ ] and @xmath162 , then there exists a point @xmath163 such that @xmath164 . + * lagrange s mean - value theorem :* let @xmath141 be a real valued function defined on a closed interval @xmath160 $ ] ( @xmath161 ) . if @xmath141 is continuous on @xmath160 $ ] and if it is derivable on @xmath126 , then there exists a point @xmath163 such that @xmath165/(b - a)=f'(c),\nonumber\end{aligned}\ ] ] where `` @xmath27 '' indicates derivative with respect to @xmath101 . +
* ift ( the implicit function theorem of two variables ) :* let @xmath6 be a real valued function of two variables @xmath2 and @xmath5 defined on @xmath166 @xmath167 @xmath168 , i.e. , for each @xmath7 and for each @xmath169 , there exists a unique real number @xmath6 .
let @xmath170 be an interior point of @xmath23 and + * a1 :* @xmath171 , + * a2 :* @xmath172 is continuous in a rectangular neighborhood of @xmath173 , + * a3 :* @xmath174 , + there exists a rectangular neighborhood @xmath175 of @xmath173 such that corresponding to every @xmath176 , @xmath177 can be expressed uniquely as a function of @xmath2 , for which the following conditions are simultaneously satisfied . + * c1 :* @xmath178 , + * c2 :* @xmath179 for all @xmath176 , + * c3 :* @xmath180 for all @xmath181 . to prove theorem 1 - theorem 10 , it is helpful to remember that if a property is true for each point of a set then that property is also true for each point of any subset of that set . without any loss of generality , we can assume @xmath182 as an interior point of @xmath23 .
* theorem 1 :* if @xmath28 , then there exists a strictly positive real number @xmath29 ( @xmath30 ) such that @xmath31 for all @xmath32 . + *
proof :* here @xmath183 is continuous at @xmath56 and consequently from * bth1 * , we can conclude that there exists a strictly positive real number @xmath29 ( @xmath30 ) such that @xmath184 * theorem 2 :* if @xmath33 and @xmath28 , then there exists a strictly positive real number @xmath29 ( @xmath30 ) such that @xmath34 for all @xmath35 and @xmath36 for all @xmath37 . + * proof :* from * theorem 1 * , we see that if @xmath28 , then there exists a strictly positive real number @xmath29 ( @xmath30 ) such that @xmath185 again as @xmath186 and its derivatives of all order with respect to @xmath5 exist finitely , @xmath187 is a continuously derivable function of @xmath8 .
now , for any @xmath188 , we have ( a ) the real valued function @xmath189 is well - defined in the closed interval [ @xmath190 , @xmath191 , ( b ) the real valued function @xmath189 is continuous in the closed interval [ @xmath190 , @xmath191 , ( c ) the real valued function @xmath189 is derivable in the open interval ( @xmath190 , @xmath192 ) .
therefore , from the lagrange s mean - value theorem of differential calculus , we can conclude that there exists a @xmath193 such that @xmath194 as @xmath195 and @xmath42 , ( [ lagrange ] ) assumes the following form : @xmath196 suppose that @xmath41 , then @xmath197 and @xmath198 , and consequently , @xmath199 , but since @xmath188 we have @xmath200 .
next suppose that @xmath40 , then @xmath201 and @xmath202 , and consequently , @xmath203 , but since @xmath188 we have @xmath204 .
therefore in any case ( @xmath41 or @xmath40 ) , we always have @xmath205 , and consequently , from eq .
( [ theorem1 ] ) , we get @xmath206 from eq .
( [ v3dash10 ] ) , we get @xmath207 from eq .
( [ v2dash ] ) , using eqs .
( [ phiv3dash10 lt 0 ] ) and ( [ phiv3dash gt 0 ] ) , we get @xmath208 and consequently , theorem 2 is proved . from theorem 2
, we observe that whenever @xmath9 and @xmath10 , then @xmath209 for all @xmath37 and @xmath210 for all @xmath35 , i.e. , @xmath186 is a concave function within the interval @xmath37 and @xmath186 is a convex function within the interval @xmath35 , where @xmath29 is strictly positive quantity but it can be made sufficiently small .
therefore , if the particle placed at @xmath56 be slightly displaced towards the positive potential side , it falls within the interval @xmath35 and due to the convexity of the function @xmath186 within the interval @xmath35 , it moves away from @xmath57 and it continues its motion until its velocity is equal to zero , i.e. , until @xmath5 takes the value @xmath211 , where @xmath212 and in this case if @xmath213 ( @xmath214 and @xmath215 ) , one can easily get a ppsw ( ppdl ) as a solution of the energy integral ( [ energy int ] ) .
again , as @xmath186 is a concave function within the interval @xmath37 , if the particle placed at @xmath56 be slightly displaced towards the negative potential side , it falls within the interval @xmath37 and due to the concavity of the function @xmath186 within the interval @xmath37 , it moves towards the point @xmath57 and consequently , there does not exist any solitary wave or double layer solution in the negative potential side . in true physical sense
, @xmath216 is an unstable position which lies in a small right neighborhood of the equilibrium position @xmath56 , whereas @xmath217 is a stable position which lies in a small left neighborhood of the equilibrium position @xmath56 , where @xmath218 and @xmath5 approaches to zero , i.e. , @xmath5 approaches to zero from the right side of zero and @xmath219 and @xmath5 approaches to zero , i.e. , @xmath5 approaches to zero from the left side of zero .
hence , as @xmath216 is an unstable position which lies in a small right neighborhood of the equilibrium position @xmath56 , there may exist either a solitary wave or a double layer in the positive potential side . on the other hand ,
as @xmath217 is a stable position which lies in a small left neighborhood of the equilibrium position @xmath56 , there does not exist any solitary wave or double layer in the negative potential side .
hence the above theoretical discussions can be summarized as follows . +
* theorem 3 :* if @xmath9 and @xmath10 , then at @xmath0 , there may exist either a solitary wave or a double layer in the positive potential side but there does not exist any solitary wave or double layer in the negative potential side . again from theorem 2
, we observe that whenever @xmath9 and @xmath11 , then @xmath210 for all @xmath37 and @xmath209 for all @xmath35 , i.e. , @xmath186 is a convex function within the interval @xmath37 and @xmath186 is a concave function within the interval @xmath35 .
therefore , following the same argument as given in the proof of theorem 3 , we have + * theorem 4 :* if @xmath9 and @xmath11 , then at @xmath0 , there may exist either a solitary wave or a double layer in the negative potential side but there does not exist any solitary wave or double layer in the positive potential side . from theorem 3 and theorem 4
, we note that if @xmath220 and @xmath221 , the point @xmath56 is the point of inflexion that seperates the convex part of @xmath186 from the concave part and if @xmath57 is a point of inflexion then there may exist either a solitary wave or a double layer in either positive potential side or negative potential side , but the solitary structures ( including double layers ) of both polarities can not coexist at @xmath0 . therefore ,
+ * theorem 5 :* if @xmath9 and @xmath28 , it is not possible to have coexistence of both positive and negative potential solitary structures ( including double layers ) at @xmath0 .
next to confirm the existence of the solitary wave and/or double layer solution of the energy integral ( [ energy int ] ) at @xmath0 , we have proved the following theorems .
+ * theorem 6 :* if @xmath9 , @xmath10 , @xmath38 for all @xmath39 and for all @xmath40 , and if there exists at least one value @xmath17 of @xmath2 such that the system supports positive potential solitary waves ( ppsws ) for all @xmath18 , then there exist either a ppsw or a positive potential double layer ( ppdl ) at @xmath0 . + * proof :* from * theorem 3 * , we have seen that if @xmath9 and @xmath10 , there may exist either a solitary wave or a double layer in the positive potential side at @xmath0 but there does not exist any solitary wave or double layer in the negative potential side at @xmath0 .
now with the help of ift , we shall confirm the existence of either a solitary wave or a double layer in the positive potential side at @xmath0 .
let @xmath222 , then according to the condition of the present theorem , there exists a solitary wave in the positive potential side at @xmath223 .
let the amplitude of this solitary wave is @xmath224 , then from the conditions for existence of ppsw we have sm1 - 1 : : : @xmath171 , sm1 - 2 : : : @xmath225 , sm1 - 3 : : : @xmath226 for all @xmath227 .
again , as @xmath228 is a continuous function of @xmath229 , we get sm1 - 4 : : : @xmath172 is continuous on a rectangular neighborhood of @xmath173 .
therefore , * sm1 - 1 * , * sm1 - 4 * , and * sm1 - 2 * are , respectively , same as the conditions * a1 * , * a2 * , and * a3 * of ift , and consequently , we can apply ift in a rectangular neighborhood of the point @xmath173 .
therefore , from ift , we can conclude that there exists a rectangular neighborhood @xmath175 of @xmath173 such that corresponding to every @xmath176 , @xmath177 can be expressed uniquely as a function of @xmath2 , for which the following conditions are simultaneously satisfied .
ift - c1 : : : @xmath178 , ift - c2 : : : @xmath179 for all @xmath176 , ift - c3 : : : @xmath180 for all @xmath230 .
again from * sm1 - 2 * , using theorem * bth2 * , we have strictly positive real numbers @xmath231 and @xmath232 such that sm1 - 5 : : : @xmath233 for all @xmath234 .
again from * ift - c3 * , we see that @xmath5 is derivable function of @xmath2 for @xmath235 , and consequently , it is continuous in @xmath236 .
as @xmath5 is continuous in @xmath236 and @xmath237 , using theorem * bth1 * , we have a strictly positive real number @xmath238 such that sm1 - 6 : : : @xmath239 for all @xmath240 .
so , if we take @xmath241 and @xmath242 , then @xmath243 for all @xmath244 , and @xmath245 for all @xmath246 , consequently , @xmath247 for all @xmath244 and @xmath246 .
therefore , we can also apply ift in the rectangular neighborhood @xmath248 , and consequently , with respect to the rectangular neighborhood @xmath249 , * ift - c1 * ,
* ift - c2 * , * ift - c3 * , * sm1 - 5 * , and * sm1 - 6 * can be put in the following form : msm1 - 1 : : : @xmath178 , msm1 - 2 : : : @xmath179 for all @xmath250 , msm1 - 3 : : : @xmath180 for all @xmath251 , msm1 - 4 : : : @xmath233 for all @xmath251 , msm1 - 5 : : : @xmath239 for all @xmath250 .
now choose two strictly positive real numbers @xmath252 and @xmath253 such that @xmath254 and @xmath255 .
these two conditions are simultaneously satisfied if we choose @xmath256 and @xmath257 . as @xmath258 for every @xmath259 and as @xmath260 ,
then we must have @xmath261 @xmath262 @xmath263 , where we set @xmath264 .
now choose two strictly positive real number @xmath265 and @xmath266 such that @xmath267 and @xmath268 .
these two conditions are simultaneously satisfied if we choose @xmath269 and @xmath270
. therefore , @xmath255 and @xmath268 @xmath262 @xmath271 @xmath262 @xmath272 , and consequently , we can also apply ift in the rectangular neighborhood @xmath273 .
so , with respect to the rectangular neighborhood @xmath274 , * ift - c1 * , * ift - c2 * , *
ift - c3 * , * msm1 - 4 * , and * msm1 - 5 * can be put in the following form : fmsm1 - 1 : : : @xmath275 , fmsm1 - 2 : : : @xmath179 for all @xmath276 , fmsm1 - 3 : : : @xmath180 for all @xmath277 , fmsm1 - 4 : : : @xmath233 for all @xmath277 , fmsm1 - 5 : : : @xmath239 for all @xmath276 . as @xmath278 and @xmath279 , from *
fmsm1 - 1 * , * fmsm1 - 2 * , * fmsm1 - 4 * , and * fmsm1 - 5 * , we get the following conditions : smc-1 : : : @xmath212 , where @xmath264 , smc-2 : : : @xmath280 , smc-3 : : : @xmath281 .
it is important to note that if we take @xmath282 , then also the above three conditions are simultaneously satisfied but @xmath283 , and consequently , @xmath56 is a position of stable equilibrium .
so , there is no question of existence of any solitary wave or double layer solution of the energy integral ( [ energy int ] ) for all @xmath282 . again ,
from the condition of the present theorem , we have @xmath38 for all @xmath39 and for all @xmath284 . as our discussion
is restricted to positive potential side ( @xmath285 ) only , from * fmsm1 - 3 * and * fmsm1 - 4 * , we get mc : : : @xmath286 for all @xmath276 , which shows that @xmath5 is an strictly increasing function of @xmath2 for all @xmath287 .
now we shall prove that smc-4 : : : @xmath288 for all @xmath289 , and the following analysis have been developed to prove * smc-4*. consider a value @xmath290 of @xmath2 such that @xmath291 , then there exists a ppsw of amplitude @xmath292 , and consequently we have sm2 - 1 : : : @xmath293 , sm2 - 2 : : : @xmath294 , sm2 - 3 : : : @xmath295 for all @xmath296 .
as @xmath297 , we have @xmath298 .
again as @xmath5 is an increasing function of @xmath2 for all @xmath287 , and @xmath299 with @xmath300 , we have @xmath301
. we can write * sm2 - 3 * as follows msm2 - 3 : : : @xmath295 for all @xmath302 .
if possible , let there exists a value @xmath303 of @xmath5 such that @xmath304 and q1 : : : @xmath305 .
again from * msm2 - 3 * , as @xmath306 , we get q2 : : : @xmath307 . considering @xmath308 as a function of @xmath2 only , from * q1 * , *
q2 * and * bolzano s theorem * , we can conclude that there exists a value @xmath309 of @xmath2 such that @xmath310 where @xmath311 .
but as @xmath311 , there exists a ppsw at @xmath312 of amplitude @xmath313 , and consequently we have sm3 - 1 : : : @xmath314 , sm3 - 2 : : : @xmath315 , sm3 - 3 : : : @xmath316 for all @xmath317 , as @xmath318 , we have @xmath319 .
again as @xmath5 is an increasing function of @xmath2 for @xmath287 , and @xmath320 with @xmath321 , we have @xmath322 .
so we can write * sm3 - 3 * as follows msm3 - 3 : : : @xmath316 for all @xmath323 , again from * msm3 - 3 * , as @xmath324 , we get q3 : : : @xmath325 , which contradicts eq .
( [ m3p30 ] ) , i.e. , @xmath326 .
therefore , our supposition is wrong , and consequently , there does not exist any @xmath327 such that @xmath305 , i.e. , @xmath328 and consequently , we get msmc-4 : : : @xmath288 for all @xmath329 . now * msmc-4 * suggests the following two possible cases : case - 1 : : there does not exist any @xmath330 such that @xmath331 , i.e. , @xmath332 for all @xmath329 .
case - 2 : : there exists at least one @xmath330 such that @xmath331 . combining * case - 1 * with * smc-1 * , * smc-2 * and * smc-3 *
, we get ( 1 ) : : @xmath212 , where @xmath264 , ( 2 ) : : @xmath333 , ( 3 ) : : @xmath281 , ( 4 ) : : @xmath334 for all @xmath289 . these four conditions confirm the existence of ppsw solution of the energy integral ( [ energy int ] ) at @xmath0 .
for * case - 2 * , there exists at least one @xmath330 such that @xmath331 .
let @xmath335 be the smallest strictly positive root of the equation @xmath331 such that @xmath336 and consequently , for this case , we have the following subcases .
subcase-1 : : @xmath334 for all @xmath337 and @xmath338 for all @xmath339$](see fig .
[ fig : appendix](a ) ) .
subcase-2 : : @xmath334 for all @xmath337 and @xmath338 for all @xmath340 $ ] with @xmath341(see fig .
[ fig : appendix](b ) ) .
subcase-3 : : @xmath334 for all @xmath337 and @xmath334 for all @xmath342 as well as @xmath343 along with @xmath344 for some @xmath345 ( see fig . [
fig : appendix](c ) ) . as @xmath186 is a continuously differentiable function of @xmath5 for all @xmath8 and for the * subcase-1
* , it is simple to check that @xmath186 is not differentiable at @xmath346 , * subcase-1 * is not possible . with the help of same argument
, it is simple to check that * subcase-2 * is not possible when @xmath347 . when @xmath348 , then either we have * subcase-3 * or we have two different branches of the curve @xmath186 originated from @xmath346 .
the later case is not possible because @xmath186 is not differentiable at @xmath346 .
so , * subcase-3 * is the only possible case of * case-2 * and for this subcase we have the following conclusions .
* subcase-3 :* in this case , as @xmath186 is a continuously differentiable function of @xmath5 , @xmath186 is not only attains its maximum value at @xmath349 , @xmath186 is convex for all @xmath350 .
therefore , from the elementary theory of convexity of a function , we have @xmath351 combining these two conditions along with the first condition of * subcase-3 * , i.e. , @xmath334 for all @xmath337 , we get ( 1 ) : : @xmath344 , ( 2 ) : : @xmath352 , ( 3 ) : : @xmath353 , ( 4 ) : : @xmath354 , ( 5 ) : : @xmath334 for all @xmath337 .
these five conditions confirm the existence of ppdl solution of the energy integral ( [ energy int ] ) at @xmath0 .
therefore , under the conditions of the present theorem , one can get either a ppsw or a ppdl as a solution of the energy integral ( [ energy int ] ) at @xmath0 . + * theorem 7 :* if @xmath9 , @xmath11 , @xmath38 for all @xmath39 and for all @xmath41 , and if there exists at least one value @xmath17 of @xmath2 such that the system supports negative potential solitary waves ( npsws ) for all @xmath18 , then there exist either a npsw or a negative potential double layer ( npdl ) at @xmath0 . + * proof :* same as theorem 6 with a slight modification for the case of negative potential side . + * theorem 8 :* if @xmath9 , @xmath10 , @xmath38 for all @xmath39 and for all @xmath40 , and if the system supports only npsws for @xmath3 , then there does not exist any ppsw at @xmath0 .
+ * proof :* from * theorem 3 * , we have seen that if @xmath9 and @xmath10 , there may exist solitary wave in the positive potential side at @xmath0 but there does not exist any solitary wave in the negative potential side at @xmath0 .
if possible let there exists a ppsw at @xmath0 and for @xmath3 , the system supports only npsws .
therefore , if @xmath355 is the amplitude of the ppsw at @xmath0 , the following conditions are satisfied : mc1 : : : @xmath212 , mc2 : : : @xmath356 , mc3 : : : @xmath280 , mc4 : : : @xmath334 for all @xmath289 .
as @xmath228 is a continuous function in its domain of definition , we have , mc5 : : : @xmath172 is continuous in some rectangular neighborhood of @xmath357 .
therefore , all the assumptions , viz .
, * mc1 * , * mc5 * and * mc3 * , of * ift * at @xmath357 are satisfied by the function @xmath6 .
therefore , from ift , we can conclude that there exists a rectangular neighborhood @xmath358 of @xmath357 such that corresponding to every @xmath359 , @xmath360 can be expressed uniquely as a function of @xmath2 , for which the following conditions are simultaneously satisfied .
ift mc1 : : : @xmath275 , ift mc2 : : : @xmath179 for all @xmath359 , ift mc3 : : : @xmath180 for all @xmath361 .
again from * mc3 * , using * bth2 * , we have two strictly positive real numbers @xmath362 and @xmath363 such that mc6 : : : @xmath233 for all @xmath364 .
so , if we take @xmath365 and @xmath366 , then @xmath367 for all @xmath368 , and @xmath369 for all @xmath370 , consequently , @xmath371 for all @xmath368 and @xmath370 .
therefore , we can also apply ift in the rectangular neighborhood @xmath372 , and consequently , * ift mc1 * , * ift mc2 * , * ift mc3 * , and * mc6 * can be re - written as follows : mmc1 : : : @xmath275 , mmc2 : : : @xmath179 for all @xmath373 , mmc3 : : : @xmath180 for all @xmath374 , mmc6 : : : @xmath233 for all @xmath374 .
again , from the condition of the present theorem , we have @xmath38 for all @xmath39 and for all @xmath284 . as our discussion
is restricted to positive potential side ( @xmath285 ) only , from * mmc3 * and * mmc6 * , we get @xmath375 is an increasing function of @xmath2 for all @xmath376 . consider a value @xmath377 of @xmath2 such that @xmath378 , and consequently , from * mmc2 * , the condition that the function @xmath375 is an increasing function of @xmath2 for all @xmath376 and * mmc6 * , we get the following three conditions : m1 - 1 : : : @xmath171 , with @xmath379 , m1 - 2 : : : @xmath380 , m1 - 3 : : : @xmath225 , [ it is important to note that if we take @xmath381 , then also the above three conditions are simultaneously satisfied but @xmath283 , and consequently , @xmath56 is a position of stable equilibrium .
so , there is no question of existence of any solitary wave solution of the energy integral ( [ energy int ] ) for all @xmath381 . ]
now if @xmath226 for all @xmath227 , then from * m1 - 1 * , * m1 - 2 * , and * m1 - 3 * , we can conclude that there exists a ppsw at @xmath382 but this is impossible as the system supports only npsws for @xmath3 .
therefore , there exists at least one value @xmath292 of @xmath5 such that @xmath383 now we shall prove that actually @xmath292 is restricted by the following inequality : @xmath384 .
now from the condition @xmath38 for all @xmath39 and for all @xmath40 , we have @xmath6 is a strictly decreasing function of @xmath2 for all @xmath385 and for any given value of @xmath285 . as @xmath386
, we get m1 - 4 : : : @xmath387 for all @xmath285 . again , from * mc4 *
, we get @xmath334 for all @xmath388 @xmath262 @xmath389 for all @xmath388 , where we have used the inequality ( @xmath387 for all @xmath285 ) as given in * m1 - 4*. therefore , we have m1 - 5 : : : @xmath226 for all @xmath390 . again , as @xmath356 and @xmath212 , putting @xmath391 in * m1 - 4 * , we get m1 - 6 : : : @xmath392 . combining * m1 - 5 * and * m1 - 6 * , we get m1 - 7 : : : @xmath226 for all @xmath393 . therefore , from * m1 - 7 *
, we can conclude that @xmath384 .
now we shall prove that there exists a value @xmath290 of @xmath2 such that @xmath298 for @xmath394 .
let us define the function @xmath395 for all @xmath396 $ ] .
now as @xmath375 is a continuously differentiable function of @xmath2 for all @xmath373 , then it is automatically continuous for all @xmath373 .
again as @xmath397 \subset n(m_{c},\delta)$ ] , @xmath375 is also continuous for all @xmath398 $ ] , and consequently , @xmath399 is continuous for all @xmath398 $ ] .
now as @xmath264 and @xmath379 , the condition @xmath384 gives @xmath400 @xmath401 @xmath402 and @xmath403 @xmath401 @xmath404 and @xmath405 .
therefore , we have a continuous function @xmath406\longmapsto \re$ ] such that @xmath404 and @xmath405,and consequently , by * bolzano s theorem * , we can conclude that there exists a value @xmath290 of @xmath2 such that @xmath407 for @xmath394 .
now as @xmath6 is strictly decreasing function of @xmath385 for any fixed value of @xmath40 , and consequently , as @xmath408 , we have @xmath409 for any given value of @xmath24 .
if we take @xmath410 , we get @xmath411 because @xmath412 , where we have used the condition * mmc2 * to get @xmath413 , so , we have @xmath414 which again contradicts eq .
( [ vm1phi20 ] ) , i.e. , @xmath415 .
thus , our original assumption is wrong , and consequently , theorem is proved , i.e. , if the system supports only npsws for @xmath3 , then there does not exist any ppsw at @xmath0 provided the conditions of the present theorem are fulfilled . + * theorem 9 :* if @xmath9 , @xmath11 , @xmath38 for all @xmath39 and for all @xmath41 , and if the system supports only ppsws for @xmath3 , then there does not exist any npsw at @xmath0 .
+ * proof :* same as theorem 8 with a slight modification for the negative potential side . + * theorem 10 :* if @xmath9 , @xmath28 , @xmath38 for all @xmath39 and for all @xmath42 , and if there exists at least one value @xmath17 of @xmath2 such that the system supports both positive and npsws for all @xmath18 , i.e. , if the system supports the coexistence of both ppsws and npsws for all @xmath18 , then if @xmath10 , there exist either a ppsw or a ppdl at @xmath0 whereas if @xmath11 there exist either a npsw or a npdl at @xmath0 .
+ * proof :* follows from * theorem 6 * and * theorem 7*. from theorem 6 -theorem 10 , it is simple to see that the critical mach number alone can not determine the existence of particular nonlinear structure and its polarity .
this is not even the motivation of our present manuscript .
however , our intension is to determine the conditions under which nonlinear structure exists at the critical mach number .
the existence of nonlinear structure at the critical mach number and its polarity is determined by the nature of the solitary structures for at least one right neighbourhood of the point @xmath0 together with the sign of the function @xmath416 . except theorem 1 and theorem 2 ,
the proof of rest of the theorems ( theorem 3 - theorem 10 ) carry physical sense inherently , since the proofs are based on the physical interpretation of the energy integral as discussed in sec .
ii . for example
, theorem 6 is proved with the help of theorem 3 while theorem 3 is proved on the basis of physical interpretation of the energy integral .
one can get new result behind the nonlinear structures at @xmath0 by considering any specific problem , only when these theorems have been used on the qualitatively different solution spaces or compositional parameter spaces with respect to any parameter of the system ( keeping arbitrary fixed values of other parameters in their respective physically admissible range ) showing the existence region of different types of solitary structures for @xmath3 . for this purpose
, we have considered the same problem of dust acoustic wave in nonthermal plasma of das _ _ et al.__@xcite in sec .
[ sec : verify ] so , without going through all qualitatively different solution spaces of the energy integral by considering the entire range of parameters involved in the system , it is not possible to make a systematic investigation of the solitary structures at @xmath0 .
to construct the solution space or compositional parameter space with respect to any parameter of the system ( keeping arbitrary fixed values of other parameters in their respective physically admissible range ) showing the existence region of different types of solitary structures for @xmath3 , it is not necessary to know the existence of solitary structure at @xmath0 . for @xmath3 , using the physical interpretation as given in first para of sec .
ii , one can easily set up a numerical scheme to construct the compositional parameter space with respect to any parameter of the system ( keeping arbitrary fixed values of other parameters in their respective physically admissible range ) showing the existence region of different types of solitary structures .
so , to construct the solution space , the existence of solitary structure at @xmath0 does not play any role .
here we consider the specific problem of das _ _ et al__@xcite to apply the analytical results ( theorem 3 - theorem 10 ) as presented in this paper . in this
paper,@xcite a computational scheme has been developed to study the arbitrary amplitude dust acoustic solitary waves and double layers in a nonthermal plasma consisting of negatively charged dust grains , nonthermal ions and isothermal electrons including the effect of dust temperature .
the sagdeev potential approach , which is valid to study the arbitrary amplitude solitary waves and double layers , has been employed .
four basic parameters of the system are @xmath417 and @xmath418 , which are respectively the ratio of unperturbed number density of electrons to that of nonthermal ions , the ratio of average temperature of nonthermal ions to that of isothermal electrons , a parameter of the nonthermal distribution of ions and the ratio of average temperature of dust particles to that of ions divided by the number of negative charges residing on a dust grain surface . the sagdeev potential @xmath4 as given by eq .
( 21 ) of das _
et al._,@xcite is well defined as a real number for all @xmath7 and for all @xmath8 , where @xmath166 is the set of all stictly positive real number and @xmath419 , and @xmath420 is given by the first equation of ( 17 ) of das _ _
et al.__@xcite so , except the point @xmath421 , any @xmath8 is an interior point of @xmath168 and any @xmath39 is an interior point of @xmath166 .
the solution spaces of the energy integral ( 20 ) with respect to @xmath422 has been shown in fig . 3 of das _
_ et al.__@xcite for fixed values of other parameters . from this solution space
, they have found that for any fixed values of the parameters @xmath423 , @xmath424 and @xmath418 the entire interval of @xmath422 can be broken up into four disjoint subintervals i : @xmath425 , ii : @xmath426 , iii : @xmath427 , iv : @xmath428 , where @xmath429 .
fig . 3 of das _ _
et al.__@xcite also shows that ( i ) in subinterval i , only npsws can exist and the mach number @xmath2 for these waves lies within the interval @xmath430 , where @xmath1 is the lower bound of @xmath2 and @xmath431 is the upper bound of @xmath2 which is defined only when the system can support npsws , ( ii ) in subinterval ii , both npsws and ppsws can coexist and the mach number @xmath2 for these waves lies within the interval @xmath432 , whereas only npsws are possible if @xmath433 , ( iii ) in subinterval iii , both npsws and ppsws can coexist and mach number for these waves lies within @xmath430 , whereas only ppsws are possible if @xmath434 , ( iv ) in subinterval iv , only ppsws can exist and mach number @xmath2 for these waves lies within @xmath435 . in subintervals
ii , iii and iv , only ppdl solution is possible along the curve @xmath43 . again from this figure
, we see that the curve @xmath43 tends to intersect the curve @xmath0 at @xmath436 , i.e. , there always exists a ppdl solution in any right neighborhood of @xmath1 for some @xmath437 .
therefore , at @xmath438 , we can expect a ppdl solution at @xmath0 , if the other conditions of theorem 6 have been fulfilled . before going to discuss the existence of solitary wave and/or double layer solutions of the present problem at @xmath0 , it is necessary to know the sign of @xmath439 , @xmath440 and the sign of @xmath441 when @xmath92 . from the expression of @xmath4 as given by eq .
( 21 ) of das _
et al._,@xcite we get the following equations .
@xmath443 where @xmath444 from eq .
( [ delvdelm ] ) , we see that @xmath38 for all @xmath39 and for all @xmath445 , and consequently , for all @xmath39 , the condition @xmath38 is satisfied for all @xmath24 as well as @xmath25 . now from the expression of @xmath440 as given by eq .
( [ v3 at mc non ] ) of present paper , it can be easily checked that @xmath446 for all @xmath447 and for any admissible values of the other parameters , consequently , @xmath416 is strictly increasing function of @xmath422 for all @xmath447 .
again , it can be easily checked that @xmath448 for @xmath449 and @xmath450 for @xmath451 .
therefore , @xmath416 strictly increases with increasing @xmath422 starting from a negative value and ending with a positive value .
so , @xmath416 intersects the axis of @xmath422 at @xmath452 and consequently , we have @xmath448 for all @xmath453 , @xmath450 for all @xmath454 and @xmath455 at @xmath456 .
it is also interesting to note that the equation @xmath455 has one and only one real root @xmath456 of @xmath422 within the admissible range of @xmath422 and for any fixed values of the other parameters .
again , it is a simple task to check that @xmath457 at @xmath456 , i.e. , @xmath457 when @xmath455 .
therefore , when @xmath455 , @xmath56 is the position of stable equilibrium and consequently , there does not exist any solitary structure of the energy integral ( 20 ) of das _
et al._,@xcite at @xmath0 .
now we are in a position to discuss the existence of solitary wave and/or double layer solution of the energy integral ( 20 ) of das _
et al._,@xcite at @xmath0 for any admissible value of @xmath422 and fixed values of the other parameters . in subinterval i ( @xmath425 )
, one can expect only npsw at @xmath0 , provided @xmath458 ( from theorem 4 ) .
but it can be easily checked that @xmath459 for any value of @xmath422 lying within subinterval i. so from theorem 8 , we can conclude that there does not exist any solitary wave solution at @xmath0 for any value of @xmath422 lying within subinterval i. again , in subinterval iv ( @xmath428 ) , one can expect only ppsw at @xmath0 , provided @xmath459 ( from theorem 3 ) .
but it can be easily checked that @xmath458 for any value of @xmath422 lying within subinterval iv .
so from theorem 9 , we can conclude that there does not exist any solitary wave solution at @xmath0 for any value of @xmath422 lying within subinterval iv .
as the system supports coexistence of both npsws and ppsws in subintervals ii and iii , i.e. , for @xmath422 lying within the interval @xmath460 , from theorem 10 , we can conclude that there must exist solitary wave or double layer solution at @xmath0 when @xmath460 and the nature of the solitary wave or double layer solution depends on the sign of @xmath416 . in fig .
[ fig : verify ] , @xmath1 , @xmath461 and @xmath416 have been plotted against @xmath422 for @xmath460 . from this figure
we see that the double layer solution is not possible for @xmath462 , as the curve @xmath43 does not tend to intersect the curve @xmath0 for any point of @xmath422 lying within the interval @xmath462 .
so , for any @xmath422 lying within the interval @xmath462 , we get either a npsw or a ppsw at @xmath0 . again from fig .
[ fig : verify ] , we see that the curve @xmath43 tends to intersect the curve @xmath0 at @xmath436 , i.e. , there always exists a ppdl solution in any right neighborhood of the point @xmath1 for some @xmath437 .
therefore , at @xmath438 , we have a ppdl solution at @xmath0 . the existence of solitary wave solution at @xmath0 for any @xmath422 lying within the interval @xmath462 and the existence of ppdl solution at @xmath0 when @xmath438 can easily be verified through fig .
[ fig : verify ] .
specifically , we have @xmath459 for any @xmath422 lying within the interval @xmath463 and @xmath458 for any @xmath422 lying within the interval @xmath464 .
consequently , we have ppsw at @xmath0 for any @xmath422 lying within the interval @xmath463 , whereas we get npsw at @xmath0 for any @xmath422 lying within the interval @xmath464 and at @xmath456 , there does not exist any solitary structure at @xmath0 . to be more specific ,
we draw fig . [
fig : positive ] .
this figure shows that the amplitude of the ppsw increases with decreasing @xmath1 and ultimately this sequence of ppsws end with a ppdl solution at @xmath0 when @xmath422 assumes the value @xmath465 .
however , we have seen in the literature that when all the parameters involved in the system assume fixed values in their respective physically admissible range , the amplitude of solitary wave increases with increasing @xmath2 and these solitary waves end with a double layer of same polarity , if exists . here
it is important to note that @xmath2 is not a function of the parameters involved in the system but is restricted by the inequality @xmath466 , where @xmath43 corresponds to a double layer solution .
so we can not compare this case with the case of @xmath1 , since @xmath1 is a function of the parameters involved in the system and consequently , monotonicity of @xmath1 entirely depends on a parameter when the other parameters assume fixed values in their respective physically admissible range . but the solitons and double layer are not two distinct nonlinear structures even when @xmath0 .
actually , double layer solution , if exists , must be the limiting structure of at least one sequence of solitons of same polarity even when @xmath0 .
more specifically , existence of double layer solution implies that there must exists a sequence of solitary waves of same polarity having monotonically increasing amplitude converging to the double layer solution , i.e. , the amplitude of the double layer solution acts as an exact upper bound or least upper bound ( _ lub _ ) of the amplitudes of the sequence of solitary waves .
therefore , if the double layer solution exists at @xmath0 then this double layer solution is also a limiting structure of a sequence of solitary waves of same polarity .
so , double layer solution can be regarded as a source or sink of solitary waves of same polarity for increasing mach number .
particularly , for the problem of das _
@xcite , the double layer solution at @xmath0 when @xmath438 can be regarded as a source for ppsws for increasing @xmath1 along the curve @xmath0 whereas the same double layer solution at @xmath0 when @xmath438 can be regarded as a sink of ppsws for decreasing @xmath1 along the curve @xmath0 .
however , for any other double layer solution , i.e. , for the double layer solution at @xmath467 can be regarded as sink of ppsws for increasing @xmath2 restricted by the inequality : @xmath466 . on the other hand ,
if we take the interval @xmath468 and move along the curve @xmath0 , then we have seen from fig .
[ fig : negative ] that the amplitude of npsw increases with increasing @xmath1 and it attains its maximum value at @xmath469 .
this is , of course , expected since @xmath470 is the point of intersection of @xmath1 and @xmath431 and for @xmath471 there does not exist any solitary structure at @xmath0 . for this interval ,
i.e. , @xmath468 , the solitary wave solution at @xmath0 is similar to the solitary wave solution for @xmath3 , because for this interval the amplitude of npsw increases with increasing @xmath1 .
moreover , it can be easily checked that the solitary structures along the curve @xmath0 is same as the solitary structures for @xmath3 , if one can move the solution space through the family of curves parallel to the curve @xmath0 .
to be more specific , solution space for the present system with respect to @xmath422 has been presented in fig .
[ fig : sol space amp](a ) , in which the curve @xmath0 is omitted from the solution space as presented in fig .
3 of ref [ 26 ] . now consider the family of curves parallel to @xmath0 .
for instance , consider one such parallel curve for @xmath472 as shown in fig .
[ fig : sol space amp](a ) . in this figure
, @xmath473 is the value of @xmath422 where the curve @xmath43 intersects the curve @xmath472 , whereas @xmath474 is the value of @xmath422 where the curve @xmath475 intersects the curve @xmath472 .
[ fig : sol space amp](a ) can be interpreted in the same way of fig .
3 of ref [ 26 ] with @xmath465 replaced by @xmath473 and @xmath470 replaced by @xmath474 . however , in fig .
[ fig : sol space amp](a ) the solitary structures exist along the curve @xmath472 , specifically , ( i ) only npsw exists for @xmath476 , ( ii ) both npsw and ppsw coexist for @xmath477 , ( iii ) only ppsw exists whenever @xmath478 and at @xmath479 , a npsw coexists with a ppdl .
the variation of amplitude of these solitary waves ( exist along the curve @xmath472 ) have been shown in fig .
[ fig : sol space amp](b ) .
this figure shows that the amplitude of npsws increase with @xmath422 having maximum amplitude at @xmath480 .
again , the amplitude of ppsws decreases with increasing @xmath422 for @xmath481 and the amplitudes of ppsws are bounded by the amplitude of ppdl at @xmath479
. moreover , the amplitudes of the ppdl solutions increases with increasing @xmath461 along the curve @xmath43 .
[ fig : sol space amp](b ) also shows that along the curve @xmath472 , the amplitude of ppsw increases with decreasing @xmath2 along the curve @xmath472 and ultimately , these ppsws end with a ppdl at @xmath479 .
this fact is similar for the case , when we move along the curve @xmath0 .
so , if we move the solution space along the family of curves parallel to the curve @xmath0 , then one can not distinguish between the solitary structures , exist along the curve @xmath0 and any curve parallel to @xmath0 .
for @xmath482 , @xmath483 and @xmath484 , the values of @xmath473 and @xmath474 lie in the neighborhood of @xmath485 and @xmath486 , respectively . to form a ppdl
, the positive energy must dominate the negative energy in some region of the system and the ppdl will occur at some point where the potential drop is maximum . from fig .
[ fig : sol space amp](b ) , it is clear that the positive potential dominates the negative potential in a right neighborhood of @xmath479 and again the potential drop is maximum thereat .
this figure shows that physical reasons also suggest the existence of ppdl solution at @xmath479 , which is already confirmed in fig .
[ fig : sol space amp](a ) .
this is the reason for the coexistence of npsw and ppdl at @xmath479 .
actually , the same phenomena happens for the occurrence of ppdl solution at @xmath0 , although there does not exist any npsw along the curve @xmath0 for @xmath487 .
we ll see later , that weakly nonlinear negative potential solitary structure exist along the curve @xmath0 for @xmath487 and also the amplitude of the npsw tends to zero as @xmath488 , whereas ppsw of finite amplitude exist along the curve @xmath0 in the right neighborhood of @xmath436 and consequently , the potential drop is maximum at @xmath436 .
this is the reason behind the occurrence of ppdl at @xmath0 when @xmath436 .
this can also be interpreted as follows : since there is no npsw solution along the curve @xmath0 for @xmath487 , one can simply think the occurrence of npsw of amplitude zero along the curve @xmath0 for @xmath487 and in this case also the potential drop is maximum at @xmath436 and consequently , we have a ppdl solution at @xmath0 .
so the solitary structures at @xmath0 is not localized . actually , the solitary structure along the curve @xmath0 is same as the solitary structures for @xmath3 , if we move the solution space through the family of curves parallel to the curve @xmath0 .
again , at @xmath456 , @xmath455 and @xmath457 , and consequently , @xmath56 is the position of stable equilibrium .
therefore , it is not possible to get any solitary structure of the energy integral ( 20 ) of das _
et al._,@xcite at @xmath0 when @xmath456 . from fig .
[ fig : positive ] , we see that along the curve @xmath0 the amplitude of ppsw decreases with increasing @xmath422 for @xmath463 and ultimately it collapses at @xmath456 . on the other hand , from fig .
[ fig : negative ] , we see that along the curve @xmath0 the amplitude of npsw decreases with decreasing @xmath422 for @xmath464 and ultimately it also collapses at @xmath456 .
so , the figures fig .
[ fig : positive ] and fig .
[ fig : negative ] are consistent with the physical interpretation for nonexistence of solitary structure at @xmath0 when @xmath456 .
again , ppsws and npsws at @xmath0 behave kdv like solitons in the left and right neighbourhood of the point @xmath456 , respectively .
again , the existence of solitary structures at @xmath0 indicates the applicability of small amplitude theory to study weakly nonlinear solitary structures . to be more specific , if we replace @xmath22 by @xmath489 to study weakly nonlinear solitary structures on the basis of the assumption @xmath490 , the energy integral ( 20 ) of das _
@xcite assumes the following form : @xmath491 where @xmath489 is given by @xmath492 @xmath493 @xmath494.\label{b}\end{aligned}\ ] ] we do not consider the case @xmath495 , since the expression for @xmath496 , in this case , does not contain any nonlinear term and therefore no solitary structures exist . for @xmath497 , we can write @xmath489 as @xmath498 where @xmath499 now applying the conditions for existence of the solitary structures of the energy integral ( [ energy int3 ] ) as discussed in first para of sec .
ii , we have found that weakly nonlinear das wave exists if and only if @xmath500 and @xmath497 . the condition @xmath500 gives @xmath501 so , @xmath1 is the lower bound of @xmath2 from which solitary wave solution of ( [ energy int3 ] ) starts to exist .
the condition @xmath497 determines the nature of the solitary wave solution of ( [ energy int3 ] ) . in particular
, the solitary wave solution of ( [ energy int3 ] ) describes either a ppsw or a npsw according to @xmath502 or @xmath503 .
the amplitude of ppsw or npsw is given by @xmath504 as @xmath505 .
furthermore , solution of ( [ energy int3 ] ) is unable to describe the existence of both ppsw and npsw simultaneously , and consequently , @xmath489 is representative of @xmath22 to study weakly nonlinear solitary structures only when @xmath422 lies either within the interval @xmath425 or within the interval @xmath428 since the amplitude of the solitary wave solution of ( [ energy int3 ] ) tends to zero as @xmath505 . in fig . [
fig : b ] , @xmath116 is plotted against @xmath422 for @xmath506 .
although , from fig .
[ fig : b](a ) , we see that @xmath116 is equal to zero at @xmath436 , one can easily verify that @xmath116 becomes zero in the neighborhood of @xmath507 when @xmath2 is close to @xmath1 .
therefore fig .
[ fig : b](a ) and fig .
[ fig : b](b ) actually shows the existence of weakly nonlinear npsw or ppsw according to @xmath422 lies either within the interval @xmath425 or within the interval @xmath428 .
it is also important to note that for @xmath425 or @xmath428 , there does not exist any solitary structure at @xmath0 . to recover the coexistence of small amplitude npsws and ppsws along with ppdls when @xmath422 lies within the interval @xmath460
, we consider the following modified energy integral with @xmath508 .
@xmath509 here @xmath510 is given by @xmath511 where @xmath512,\end{aligned}\ ] ] @xmath109 and @xmath116 are given by eqs .
( [ a ] ) and ( [ b ] ) , respectively . for @xmath513 , @xmath510 is same as @xmath489 , so , we consider the case for @xmath514 . for @xmath514 , we can write @xmath510 as @xmath515 where @xmath516 @xmath517 from ( [ phim1 ] ) and ( [ phim2 ] ) , we note the following facts .
+ ( i ) @xmath518 and @xmath519 both are real and distinct if @xmath520
. + ( ii ) @xmath521 . + ( iii ) @xmath522 if and only if @xmath523 .
+ ( iv ) @xmath524 and @xmath525 as @xmath505 .
+ the necessary conditions for the existence of solitary structures of the energy integral ( [ energy int4 ] ) are @xmath526 and @xmath527 .
the conditions @xmath528 and @xmath529 are trivially satisfied whereas the condition @xmath527 gives @xmath530 . here also the condition @xmath500 gives @xmath3 .
therefore , as @xmath530 , from ( ii ) , we see that @xmath531 if @xmath532 and in this case , @xmath533 for all @xmath534 as well as @xmath533 for all @xmath535 and consequently , both npsw and ppsw exist simultaneously . on the other hand , if @xmath536 , one can not get coexistence of weakly nonlinear npsws and ppsws , since for @xmath536 , both @xmath518 and @xmath519 are either positive or negative . again as @xmath524 and @xmath525 as @xmath505 ,
@xmath510 effectively define the coexistence of both npsws and ppsws when @xmath422 lies within the interval @xmath460 provided that in this subinterval of @xmath422 , @xmath532 and @xmath520 . in fig .
[ fig : c](a ) , @xmath121 and @xmath537 are plotted against @xmath422 ( @xmath460 ) for @xmath506
. a quick look of this figure may suggest that @xmath538 at some point of @xmath422 . to remove this illusion , we draw @xmath537 against @xmath422 ( in a small neighborhood of @xmath539 ) for three different values of @xmath2 in fig . [
fig : c](b ) .
this figure clearly shows that @xmath540 at @xmath507 only when @xmath0 .
however , we have already found that at @xmath507 , @xmath56 is a position of stable equilibrium , and consequently there does not exist any solitary structure of the energy integral at @xmath0 .
therefore , for @xmath3 , @xmath541 and consequently , these two figures actually show the coexistence of weakly nonlinear npsws and ppsws when @xmath422 lies within the interval @xmath542 .
so , @xmath510 is unable to produce coexistence of both npsws and ppsws for all @xmath422 within @xmath460 , even when we take values of @xmath2 smaller than @xmath506 .
again , in this interval of @xmath422 , i.e. , @xmath543 , the amplitude of both npsws and ppsws tends to zero as @xmath505 .
but for @xmath544 , we get ppsws of finite amplitude at @xmath0 and for @xmath545 , we get npsws of finite amplitude at @xmath0 . however , both ppsws and npsws collapse at @xmath456 , i.e. , except for a very small neighborhood of @xmath456 , ppsws having finite amplitude exist at @xmath0 and npsws of finite amplitude exist at @xmath0 .
so , it is not possible to apply small amplitude method to study the weakly nonlinear solitary structures in a parameter regime where finite amplitude solitary structure occurs at @xmath0 , because in any small amplitude method , amplitude of solitary wave always tends to zero as @xmath505 which contradicts the occurrence of finite amplitude solitary wave at @xmath0 .
there may be another possibility that @xmath510 is not the perfect match of @xmath22 to study weakly nonlinear solitary structures in @xmath460 . in any case
, one can find inconsistency to study weakly nonlinear solitary structures for @xmath460 .
therefore , our conclusion is that if there exists a region in the parameter space where finite amplitude solitary structures exist at @xmath0 , it is not possible to study weakly nonlinear solitary structures by means of any small amplitude method in that portion of the parameter space .
the paper deals with the existence of solitary wave and double layer at @xmath0 .
the work is relevant to some recent numerical observations .
@xcite however , the existence of solitary wave at @xmath0 is limited by means of these numerical observations only . in this paper ,
our motivation is to find the clarifications behind these numerical observations .
the analytical theory for the existence of solitary wave and double layer at @xmath0 has been provided for the first time in the literature . moreover ,
this analytical theory is purely based on the physical interpretation of energy integral as discussed in sec .
ii . without having the prior knowledge of physics of energy integral ,
it is not possible to give the proper mathematical clarification of the above mentioned numerical observations .
therefore , the theorems presented in this manuscript are based on the physical consequences of energy integral . on investigating das waves in sec.[sec :
verify ] , we have found the existence of double layer solution at @xmath0 .
we have also discussed in our theoretical section , the conditions under which double layer solution occurs at @xmath0 .
this observation is not specific for this problem only . actually , in any plasma environments
, if the curve @xmath43 tends to intersect the curve @xmath0 at some point of the compositional parameter space or solution space , then we can always get a double layer solution at @xmath0 , where @xmath546 corresponds to a double layer solution .
we have mentioned this result without proof .
the existence of double layer solution at @xmath0 has been observed for the first time in literature and thus , this is quiet a new result . from the investigation of solitary structures at @xmath0 of dust acoustic waves in sec.[sec : verify ] , we find that the solitons and double layer are not two distinct nonlinear structures even when @xmath0 . actually , double layer solution , if exists , must be the limiting structure of at least one sequence of solitons of same polarity even when @xmath0 .
the following simple theorem of real analysis gives the actual logic behind the numerical observation that solitary waves end with a double layer , if exists .
+ _ if @xmath547 is a monotonically increasing sequence of real numbers , then @xmath547 converges to its _
lub_. furthermore , if the sequence is bounded above , then _ _
lub__@xmath547 is finite . _
+ this theorem gives the actual logic behind the numerical observation of baboolal _
@xcite actually , baboolal _
@xcite numerically observed that the double layer solution is the ultimate position of solitons , i.e. , solitary waves end with a double layer , if exists .
this numerical observation of baboolal _
@xcite has been referred by several authors in different plasma environments .
if we explain any fact with proper justification by means of mathematics , we can apply it in any specific problem . in the present paper , with the help of simple theorems
, we have extended the physical interpretation for the existence of the solitary structures of the well - known energy integral even when the mach number @xmath2 assumes its smallest possible value @xmath1 .
the authors are grateful to prof .
k. p. das for his valuable suggestions to prepare this manuscript .
one of the authors ( animesh das ) is thankful to state government departmental fellowship scheme for providing research support .
50 r. z. sagdeev , reviews of plasma physics , vol .
4 , ed m a leontovich ( consultant bureau , new york ) 1966 .
p. k. shukla and m. y. yu , j. math .
* 19 * , 2506 ( 1978 ) .
m. y. yu , p. k. shukla , and s. bujarbarua , phys . fluids . *
23 * , 2146 ( 1980 ) .
r. bharuthram and p. k. shukla , phys . fluids .
* 29 * , 3214 ( 1986 ) .
s. baboolal , r. bharuthram , and m. a. hellberg , j. plasma .
* 40 * , 163 ( 1988 ) .
s. baboolal , r. bharuthram , and m. a. hellberg , j. plasma . phys . *
41 * , 341 ( 1989 ) .
s. baboolal , r. bharuthram , and m. a. hellberg , j. plasma . phys . * 44 * , 1 ( 1990 ) .
s. baboolal , r. bharuthram , and m. a. hellberg , j. plasma . phys . * 46 * , 247 ( 1991 ) .
r. a. cairns , a. a. mamun , r. bingham , r. b@xmath548strom , r. o. dendy , c. m. c. nairn , p. k. shukla , geophys .
* 22 * , 2709 ( 1995 ) .
r. a. cairns , r. bingham , r. o. dendy , c. m. c. nairn , p. k. shukla and a. a. mamun , j. phys .
iv * 5 * , c6 - 43 ( 1995 ) .
r. a. cairns , a. a. mamun , r. bingham and p. k. shukla , physica scripta * t63 * , 80 ( 1995 ) .
s. i. popel , m. y. yu , and v. n. tsytovich , phys .
plasmas * 3 * , 4313 ( 1996 ) .
a. a. mamun , r. a. cairns and p. k. shukla , phys .
plasmas * 3 * , 702 ( 1996 ) .
a. a. mamun , r. a. cairns , and p. k. shukla , phys .
plasmas * 3 * , 2610 ( 1996 ) .
b. xie , k. he , and z. huang , phys .
a * 247 * , 403 ( 1998 ) . c. a. mendoza - brice@xmath549o , s. m. russel and a. a. mamun , planet .
space sci . * 48 * , 599 ( 2000 ) .
t. s. gill , h. kaur , and n. s. saini , phys .
plasmas * 10 * , 3927 ( 2003 ) .
s. k. maharaj , s. r. pillay , r. bharuthram , s. v. singh and g. s. lakhina , physica scripta * t113 * , 135 ( 2004 ) .
s. k. maharaj , s. r. pillay , r. bharuthram , r. v. reddy , s. v. singh and g. s. lakhina , j. plasma phys . * 72 * , 43 ( 2006 ) .
m. a. hellberg , f. verheest , and t. cattaert , j. phys .
gen * 39 * , 3137 ( 2006 ) .
f. verheest and s. r. pillay , phys .
plasmas * 15 * , 013703 ( 2008 ) . t. k. baluku and m. a. hellberg , phys .
plasmas * 15 * , 123705 ( 2008 ) . f. tanjia and a. a. mamun , phys .
* 78 * , 065502 ( 2008 ) .
f. verheest , phys .
plasmas * 16 * , 013704 ( 2009 )
. m. djebli and h. marif , phys .
plasmas * 16 * , 063708 ( 2009 ) .
a. das , a. bandyopadhyay , and k. p. das , phys .
plasmas * 16 * , 073703 ( 2009 ) .
a. das , a. bandyopadhyay , and k. p. das , phys .
plasmas * 17 * , 014503 ( 2010 ) .
f. verheest , phys .
plasmas * 17 * , 062302 ( 2010 ) .
f. verheest and m. a. hellberg , phys .
plasmas * 17 * , 023701 ( 2010 ) . t. k. baluku , m. a. hellberg , i. kourakis , and n. s. saini , phys .
plasmas * 17 * , 053702 ( 2010 ) . f. verheest , phys .
plasmas * 17 * , 062302 ( 2010 ) .
t. k. baluku , m. a. hellberg , and f. verheest , epl * 91 * , 15001 ( 2010 ) . | a general theory for the existence of solitary wave and double layer at @xmath0 has been discussed , where @xmath1 is the lower bound of the mach number @xmath2 , i.e. , solitary wave and/or double layer solutions of the well - known energy integral start to exist for @xmath3 .
ten important theorems have been proved to confirm the existence of solitary wave and double layer at @xmath0 .
if @xmath4 denotes the sagdeev potential with @xmath5 is the perturbed field or perturbed dependent variable associated with the specific problem , @xmath6 is well defined as a real number for all @xmath7 and for all @xmath8 , and @xmath9 , @xmath10 ( @xmath11 ) , @xmath12 for all @xmath13 and for all @xmath14 ( @xmath15 ) , where @xmath16 " , the main analytical results for the existence of solitary wave and double layer solution of the energy integral at @xmath0 are as follows .
* result-1 : * if there exists at least one value @xmath17 of @xmath2 such that the system supports positive ( negative ) potential solitary waves for all @xmath18 , then there exist either a positive ( negative ) potential solitary wave or a positive ( negative ) potential double layer at @xmath0 . *
result-2 : * if the system supports only negative ( positive ) potential solitary waves for @xmath3 , then there does not exist positive ( negative ) potential solitary wave at @xmath0 . *
result-3 : * it is not possible to have coexistence of both positive and negative potential solitary structures ( including double layers ) at @xmath0 .
apart from the conditions of * result-1 * , the double layer solution at @xmath0 is possible only when there exists a double layer solution in any right neighborhood of @xmath1 .
finally , these analytical results have been applied to a specific problem on dust acoustic waves in nonthermal plasma in search of new results . |
in this paper we consider the following system of equations:@xmath2{c}\frac{\partial u\left ( x , t\right ) } { \partial t}=-u\left ( x , t\right ) + \int_{-\infty}^{\infty}j\left ( x - y\right ) q\left ( y , t\right ) s\left ( u\left ( y , t\right ) \right ) dy\\ \frac{1}{\varepsilon}\frac{\partial q\left ( x , t\right ) } { \partial t}=1-q\left ( x , t\right ) -\beta q\left ( x , t\right ) s\left ( u\left ( x , t\right ) \right ) \end{array } \right . , \ \label{-1}\ ] ] where @xmath0 is a normalized exponential @xmath3 and the firing rate function @xmath1 is given by @xmath4 for certain positive parameters @xmath5 @xmath6 @xmath7 @xmath8 and @xmath9 the variable @xmath10 is the synaptic input current for a neural network with synaptic depression , the effect of which is represented by the scaling factor @xmath11 these equations were proposed and studied by g. faye in @xcite .
the faye model is a simplified version of one first introduced by kilpatrick and bressloff in @xcite .
these authors included a variable and equation to allow for spike frequency adaptation .
however they show by numerical computation that adaptation has little effect on the resulting waves .
faye dropped the adaptation equation and variable in @xcite to get his system ( [ -1 ] ) .
see @xcite and @xcite for further information on the physical background of ( [ -1 ] ) . in @xcite
the author proves two interesting results about the system ( [ -1 ] ) , namely the existence of a travelling pulse solution and the stability of this solution .
a travelling pulse solution of ( [ -1 ] ) is a non - constant solution @xmath12 of the form @xmath13 such that both @xmath14 and @xmath15 exist and these limits are equal . in this paper
we are interested in the existence of values of @xmath16 for which ( [ -1 ] ) has such a solution . as we describe briefly below ,
using ( [ -2 ] ) leads to a set of four ode s in which @xmath16 is a parameter . to show that a travelling pulse exists for some @xmath17
, faye uses the theory of geometric singular perturbation initiated by fenichel in @xcite and extended by jones and kopell in @xcite .
the blowup method is also employed @xcite . here
we extend the existence result in @xcite in several ways .
we show that for sufficiently small @xmath18 there are at least two travelling pulses , hence a fast pulse and a slow pulse , for speeds @xmath19 . also , we remove an important hypothesis used in @xcite , one which can only be verified by numerical integration of a related ode system .
( this hypothesis is stated and discussed in section [ discussion ] . )
our proof is for a general class of firing functions which includes the specific @xmath1 for which faye states his theorem .
further , we use a method which allows , in some sense , a larger range of @xmath20 than seems possible with geometric perturbation .
this will be made precise in the statements of our theorems .
we believe , based on our past experience with a similar problem , that it is feasible to check existence rigorously for particular positive values of @xmath21 using precise numerical analysis based on interval arithmetic , but we have not carried out such a check . this will be explained further in section [ discussion ] .
we now mention two well - known predecessors of the kilpatrick - bressloff and faye models . in 1992 ,
ermentrout and mcleod studied the equation @xmath22 as above , @xmath1 is positive , bounded , and increasing . since there is no feedback in the equation , ( [ emc ] )
supports only traveling fronts , where @xmath10 is monotone . in the landmark paper @xcite ermentrout and mcleod proved the existence of fronts for a wide variety of symmetric positive weight functions @xmath0 and firing rates @xmath23 (
their work applied to a more general equation ) subsequently , in @xcite , pinto and ermentrout introduced the needed negative feedback in order to get pulses .
their system is@xmath2{c}\frac{\partial u\left ( x , t\right ) } { \partial t}=-u - v+\int_{-\infty}^{\infty } j\left ( x - y\right ) s\left ( u\left ( t , y\right ) \right ) dy\\ \frac{1}{\varepsilon}\frac{\partial q\left ( x , t\right ) } { \partial t}=u-\gamma v \end{array } \right . .
\label{pe}\ ] ] they analyzed this system primarily for the case @xmath24 where @xmath25 is the heaviside function and @xmath26 is a constant representing a firing threshold . while some partial results have been obtained recently by scheel and faye ( see section [ discussion ] ) , we are not aware of any existence proof for pulses which covers all reasonable smooth functions @xmath23 we discuss what we mean by reasonable in section [ discussion ] , where we also indicate why our method does not appear to apply to this model , and why we expect that ( [ pe ] ) supports a richer family of bounded traveling waves than exist for ( [ -1 ] ) .
travelling pulse solutions of ( [ -1 ] ) with ( [ -2 ] ) are shown to satisfy a system of ode s by letting @xmath27 and computing @xmath28 and @xmath29 .
we find that @xmath2{c}u^{\prime}=\frac{v - u}{c}\\ v^{\prime}=w\\ w^{\prime}=b^{2}\left ( v - qs\left ( u\right ) \right ) \\ q^{\prime}=\frac{\varepsilon}{c}\left ( 1-q-\beta qs\left ( u\right ) \right ) .
\end{array } \right . \label{1}\ ] ] we will denote solutions of this system by @xmath30 and we look for values of @xmath16 for which there is a non - constant solution @xmath31 such that @xmath32 and @xmath33 both exist and are equal .
the orbit of such a solution of ( [ 1 ] ) is called homoclinic . in the language of dynamical systems
, @xmath34 is a pulse solution of ( [ -1 ] ) if and only if the orbit of @xmath31 is homoclinic .
we make the following assumptions on @xmath1 .
[ c0 ] the function @xmath1 is positive , increasing , bounded , and has a continuous first derivative @xmath35 [ c0a]the function @xmath36 has one local maximum followed by one local minimum , and no other critical points .
[ c1]@xmath1 is such that the system ( [ 1 ] ) has exactly one equilibrium point , say @xmath37 .
[ c2]the function @xmath1 is also such that the fast system@xmath2{c}u^{\prime}=\frac{v - u}{c}\\ v^{\prime}=w\\ w^{\prime}=b^{2}\left ( v - q_{0}s\left ( u\right ) \right ) \end{array } \right . \label{2}\ ] ] has three equilibrium points , @xmath38 @xmath39 , and @xmath40 with @xmath41 [ c3]@xmath42 for convenience we will assume that @xmath43 on @xmath44 then conditions [ c0]-[c2 ] imply that @xmath45 @xmath46 , and @xmath47 we will denote solutions of ( [ 2 ] ) by @xmath48 the local minimum of @xmath49 will be denoted by @xmath50 in @xcite specific ranges of @xmath26 and @xmath51 are given so that these conditions are satisfied by the function given in ( [ -3 ] ) . in figure [ figurea ]
we show the graphs of @xmath49 , @xmath52 ( the @xmath53 nullcline ) , and @xmath54 , when @xmath1 is given by ( [ -3 ] ) .
we use the same parameter values as were chosen for illustration in @xcite . ]
we can now state our first main result .
[ thm1a]if conditions [ c0]- [ c3 ] are satisfied , and @xmath20 is positive and sufficiently small , then there are at least two positive values of @xmath55 say @xmath56 such that ( [ 1 ] ) has a non - constant solution @xmath31 satisfying @xmath57 in order to state our remaining theorems it is convenient first to give some basic information about the fast system , ( [ 2 ] ) .
we state this information as a pair of lemmas , which will be used in proving our theorems .
their proofs are given in the appendix .
[ lem1]if conditions [ c2 ] and [ c3 ] are satisfied , then for each @xmath17 the equilibrium point @xmath58 of ( [ 2 ] ) is a saddle point , with a one dimensional unstable manifold @xmath59 and a two dimensional stable manifold @xmath60 there is , for each @xmath61 a unique solution @xmath62 of ( [ 2 ] ) with @xmath63 for all @xmath64 and satisfying the conditions @xmath2{c}u_{0,c}\left ( 0\right ) = u_{m}\\ w_{0,c}>0\text { on } ( -\infty,0 ] .
\end{array } \right .
\label{3}\ ] ] further , there is a unique @xmath65 such that @xmath66 on @xmath67 and @xmath68 in other words , the branch @xmath69 of @xmath70 pointing into the positive octant @xmath71 is a heteroclinic orbit connecting @xmath72 to @xmath73 also , @xmath74 on @xmath75 , which implies that @xmath76 and @xmath77 .
this solution is called a front for ( [ 2 ] ) .
a front for ( [ 2 ] ) can be characterized as a solution of this equation which exists on @xmath78 is nonconstant and bounded , and satisfies @xmath79 on @xmath44 [ lem2]if @xmath80 then @xmath81 on @xmath67 , and @xmath82 if @xmath83 then @xmath84 is initially positive and has a unique zero .
also , @xmath85 has a unique zero , and @xmath86 if @xmath87 and @xmath88 is the zero of @xmath85 , where @xmath89 is a maximum , then @xmath90 and @xmath91 suppose finally that for some @xmath92 @xmath93 on an interval @xmath94 @xmath95 and @xmath96 then for any @xmath97 , @xmath98 [ rem1]we conjecture that the condition @xmath99 would imply the same conclusion , but we have not been able to prove this .
the positive number @xmath100 defined in lemma [ lem1 ] plays an important role throughout this paper .
[ thm1]suppose that conditions [ c0]- [ c3 ] are satisfied .
suppose also that there is a @xmath101 , such that if @xmath102 is the unique zero of @xmath103 ( which exists by lemma [ lem2 ] ) , then @xmath104 assume as well that for some @xmath18 there is a solution @xmath105 of ( [ 1 ] ) with @xmath106 which has the following properties:@xmath107@xmath108{c}u_{\varepsilon , c_{1}}^{\prime}>0\text { on some interval \thinspace}(-\infty , t_{1})\text { and } u_{\varepsilon , c_{1}}^{\prime\prime}\left ( t_{1}\right ) < 0\\ u_{\varepsilon , c_{1}}^{\prime}<0\text { on some interval } ( t_{1},t_{3}]\text { and } u_{\varepsilon , c_{1}}\left ( t_{3}\right ) = 0 \end{array } \right . \tag{ii}\label{ii}\ ] ] then for the given @xmath20 there are two values of @xmath55 say @xmath109 and @xmath110 such that ( [ 1 ] ) has a homoclinic orbit .
figure [ fig2a ] below includes a graph of the orbit of a solution satisfying ( [ i ] ) and ( [ ii ] ) projected onto the @xmath111 plane , with the points @xmath112 and @xmath113 marked ( as well as an additional point @xmath114 which is explained later ) . the other solution shown in that figure satisfies ( [ i ] ) but not ( [ ii ] ) .
theorem [ thm1a ] is implied by theorem [ thm1 ] and the following result .
[ thm2]if conditions [ c0]- [ c3 ] are satisfied then there is a @xmath115 satisfying the conditions in the second sentence of theorem [ thm1 ] .
further , with this @xmath116 if @xmath20 is sufficiently small , then the solution @xmath117 of ( [ 1 ] ) with @xmath106 satisfies ( [ i ] ) and ( [ ii ] ) of theorem [ thm1 ]
. it will follow from the proofs of these results that as @xmath118 @xmath119 .
the following result is all we have proved about the asymptotic behavior of @xmath120 [ thm3]@xmath121 however there is an @xmath122 independent of @xmath20 such that if there is a homoclinic orbit for @xmath123 then @xmath124 for a given pair @xmath125 , the hypotheses of theorem [ thm1 ] can be verified by checking one solution of ( [ 2 ] ) at @xmath106 and one solution of ( [ 1 ] ) , with the given @xmath20 and @xmath126 for the specific model considered in @xcite , standard numerical analysis ( non - rigorous ) easily finds specific values of @xmath127 where these hypotheses are apparently satisfied satisfies the conditions in theorem [ thm1 ] .
if the conjecture in remark [ rem1 ] is true then it appears that @xmath128 would work . ] . in the discussion section
we describe how this could , in principle , be checked rigorously using uniform asymptotic analysis near the equilibrium points of ( [ 2 ] ) and [ 1 ] ) , and then a rigorous numerical ode solver ( using interval arithmetic ) over two compact intervals .
we cite a paper where a similar procedure was followed successfully , but we have not attempted it here . in @xcite only one homoclinic solution is found , and there is an extra hypothesis about the system ( [ 2 ] ) .
( hypotheses 3.1 ) as far as we know , this hypothesis can only be checked by numerically solving the system ( [ 2 ] ) . we discuss this further in section [ discussion ] .
we need two simple preliminary results about the behavior of solutions . [ prop1]for any @xmath129 the regions @xmath130 and @xmath131 are positively invariant open sets for the system ( [ 1 ] ) .
we are assuming that @xmath132 for all @xmath133 hence , @xmath134 if @xmath135 and @xmath136 if @xmath137 therefore @xmath138 is positively invariant .
further , if @xmath139 then @xmath140 if @xmath141 and @xmath142 if @xmath143 the result follows .
note as well that because @xmath1 is bounded , all solutions of ( [ 1 ] ) exist on @xmath144 [ prop0 ] if @xmath145 is a solution of ( [ 1 ] ) , and @xmath146 for some @xmath147 then either @xmath148 or @xmath149 this follows from condition [ c1 ] , which implies that the graph of the decreasing function @xmath150 in the @xmath151 plane , where @xmath152 , passes under the point @xmath153 .
( see figure [ figurea ] . ) in the first , and longest , part of the proof of theorem [ thm1 ] we show that there is a fast pulse , with speed @xmath154 which tends to @xmath100 as @xmath20 tends to zero . in the second part
we look for a slow pulse , with a speed @xmath155 which tends to zero as @xmath20 tends to zero .
we will show that for any possible homoclinic orbit , @xmath156 .
we look for homoclinic orbits such that , as well , @xmath157 in @xmath158 . in searching for the fast solution we will consider for each @xmath17 a certain uniquely defined solution @xmath159 such that @xmath160 .
we will show that there is a nonempty bounded set of positive values of @xmath16 , called @xmath161 , such that , among other properties of @xmath162 , either @xmath163 exceeds @xmath164 at some point , or @xmath165 becomes negative .
we then examine the behavior of @xmath166 where @xmath167 .
the goal is to show that @xmath168 .
this is done be eliminating all the other possible behaviors of @xmath169 , often by showing that a particular behavior implies that all values of @xmath16 close to @xmath154 are not in @xmath161 .
the following result is basic to our analysis of the full system ( [ 1 ] ) .
the proof is routine and again left to the appendix .
[ lem2b ] suppose that conditions [ c0]- [ c3 ] hold , and let @xmath37 be the unique equilibrium point of ( [ 1 ] ) .
then for any @xmath129 and @xmath17 the system ( [ 1 ] ) has a one dimensional unstable manifold at @xmath170 , say @xmath171 with branch @xmath172 starting in the region @xmath173 if @xmath174 is a solution lying on this manifold , then for large negative @xmath147 @xmath175 and @xmath176 also , @xmath177 while if @xmath18 then @xmath178 for large negative @xmath179 the invariant manifold @xmath172 depends continuously on @xmath180 in @xmath181 @xmath182 ( the meaning of continuity here is made clear in the text below . ) finally , if @xmath183 is the positive eigenvalue of the linearization of ( [ 1 ] ) around @xmath184 then @xmath185 for each @xmath17 and @xmath186 the following proposition follows trivially from ( [ 1 ] ) and will be used a number of times , often without specific mention .
[ prop2 ] @xmath187 we use the fourth item in this list to prove [ lem7]for any @xmath18 and @xmath61 if @xmath31 is a solution on @xmath172 and @xmath188 on an interval @xmath189 $ ] , then @xmath190 on @xmath191 .
if @xmath192 never changes sign , let @xmath193 denote @xmath194 .
otherwise , suppose that @xmath192 first changes sign at @xmath195 if @xmath196 for some @xmath197 and @xmath198 is the first zero of @xmath199 , then @xmath200 and by the fourth item of proposition [ prop2 ] , @xmath201 .
from the definitions of @xmath193 and @xmath202 @xmath203 and so @xmath204 since @xmath192 does not change sign at @xmath202 @xmath205 and so @xmath206 hence at @xmath198 , @xmath207 if @xmath208 then @xmath209 is an equilibrium point , a contradiction . if @xmath210 then @xmath198 is a local maximum of @xmath211 which
is inconsistent with the assumption that @xmath212 on @xmath213.$ ] hence @xmath214 but then @xmath215 this again implies that @xmath216 on some interval to the left of @xmath202 contradicting the definition of @xmath217 this completes the proof of lemma [ lem7 ] .
[ lem2c ] if @xmath145 is a solution on @xmath172 then @xmath218 on an interval @xmath189 $ ] with @xmath219 observe that @xmath220 for @xmath221 .
it follows that if @xmath222 and @xmath157 on an interval @xmath223 , then @xmath142 on this interval .
hence @xmath218 as long as @xmath224 and @xmath157 .
( that is , if @xmath224 and @xmath157 on @xmath225 $ ] , then @xmath218 on this interval . ) since @xmath226 as long as @xmath227 , lemma [ lem7 ] implies that @xmath218 as long as @xmath224 , proving lemma [ lem2c ] .
hence the conditions @xmath228 and @xmath218 on @xmath229 $ ] determine a unique solution @xmath230 on @xmath231 let @xmath232 since @xmath233 it follows from proposition [ prop1 ] that if @xmath172 is a homoclinic orbit , then it lies entirely in @xmath234 in the rest of this subsection let @xmath115 and @xmath20 be chosen as in theorem [ thm1 ] .
[ lem3a ] if @xmath235 then either @xmath236 on @xmath237 or at the first zero of @xmath238 @xmath239 by the hypotheses on @xmath115 in theorem [ thm1 ] , lemma [ lem2 ] implies that @xmath240 to extend this to @xmath18 a comparison result is needed .
let @xmath241 lemma [ lem7 ] implies that if @xmath18 then @xmath157 on any interval @xmath225 $ ] where @xmath242 since in such an interval @xmath243 and @xmath244 also , as long as @xmath218 we can consider @xmath10 , @xmath245 , and @xmath53 as functions of @xmath246 say that @xmath247 , @xmath248 , and @xmath249 .
then@xmath2{c}u^{\prime}(v)=\frac{v - u(v)}{cw(v)}\\ w^{\prime}(v)=\frac{b^{2}(v - q(v)s(u(v))}{w(v)}\end{array } \right .
\label{a1}\ ] ] we compare @xmath248 with the solution when @xmath250 let @xmath251 then we can write @xmath252 , @xmath253 , and @xmath54 .
the equations become @xmath2{c}u_{1}^{\prime}(v)=\frac{v - u_{1}(v)}{cw_{1}(v)}\\ w_{1}^{\prime}(v)=\frac{b^{2}(v - q_{0}s(u_{1}(v))}{w_{1}(v)}\end{array } \right .
\label{a2}\ ] ] since @xmath254 ( lemma [ lem2b ] ) , it is seen by considering eigenvectors of the linearization of ( [ 1 ] ) around @xmath170 is given in appendix b. ] that for @xmath255 sufficiently close to @xmath256 ( i.e. for large negative @xmath64),@xmath108{c}u\left ( v\right ) < u_{1}\left ( v\right ) \\
w\left ( v\right ) > w_{1}\left ( v\right ) \end{array } \right . .
\label{a3}\ ] ] if , at some first @xmath257 one of these inequalities should fail while the other still holds , then a contradiction results from comparing ( [ a1 ] ) and ( [ a2 ] ) , because @xmath157 and @xmath1 is increasing .
for example , if @xmath258 and @xmath259 , then ( [ a1 ] ) and ( [ a2 ] ) imply that @xmath260 , a contradiction because @xmath261 on @xmath262 . also ,
if @xmath263 and @xmath264 then @xmath265 , since @xmath157 as long as @xmath266 .
this is also a contradiction of the definition of @xmath267 .
if both inequalities fail at the same @xmath257 then there is still a contradiction because @xmath268 hence , if @xmath269 for @xmath270 then ( [ a3 ] ) holds in this interval .
this implies that for any @xmath271 if @xmath272 has a first zero at @xmath273 then @xmath274 in the proof of lemma [ lem2 ] it is shown that @xmath275 and combining these shows that if @xmath276and @xmath277 then @xmath278 this contradiction completes the proof of lemma [ lem3a ] .
[ lem3]if @xmath279 then @xmath280 , @xmath281 @xmath282 and @xmath283 as @xmath284 this follows from lemma [ lem2 ] and the comparison used to prove lemma [ lem3a ] .
we are now ready to apply a shooting argument to obtain the fast pulse . still with @xmath20 as in theorem [ thm1 ] , for each @xmath17 let @xmath285 and set @xmath286\text { and } u < u_{0}\text { on } ( t_{2},t_{3}].\right\}\end{aligned}\ ] ] ( see figures [ fig2a ] . ) [ lem5 ] @xmath287 is an open subset of the half line @xmath288 .
suppose that @xmath289 and choose @xmath290 as in the definition of @xmath291 note from ( [ 1 ] ) that if @xmath292 then there is a @xmath293 such that @xmath294 @xmath295 also , @xmath296 if @xmath297 hence @xmath298 and @xmath299 also , ( [ 1 ] ) implies that if @xmath300 and @xmath301 then @xmath302 .
since @xmath303 is a smooth function of @xmath16 , uniformly for @xmath64 in , say , @xmath304,$ ] it follows that for @xmath16 in some neighborhood of @xmath305 , @xmath306 is defined for @xmath307 and all the inequalities in the definition of @xmath287 continue to hold , so that this neighborhood lies in @xmath291 this proves lemma [ lem5 ] .
the hypotheses of theorem [ thm1 ] imply that @xmath308 , while by lemma [ lem3 ] , if @xmath279 then @xmath309 the numbers @xmath310 depend on @xmath55 and when we need to emphasize this we will denote them by @xmath311 for @xmath312 .
we now let @xmath313 ( this is finite , by lemma [ lem3 ] . )
[ lem4]with @xmath20 as in theorem [ thm1 ] , @xmath314 is a homoclinic orbit of ( [ 1 ] ) .
the proof depends on the fact that @xmath305is a boundary point of @xmath287 and lies in @xmath315 in figure 4 we show several graphs of @xmath111 which , if they occurred for @xmath316 would suggest ( without quite implying ) that @xmath16 was on the boundary of @xmath291 we must eliminate these and some other possibilities , and this will imply that @xmath317 the reader may want to review the definition of @xmath287 before examining these figures .
[ figcc ] we need additional lemmas . recalling that @xmath20 was chosen in the statement of [ thm1 ] , we will now drop the @xmath20-dependence of @xmath318 and its components from our notation , writing @xmath162 and @xmath287 .
when the dependence of @xmath162 on @xmath16 is not crucial to an argument we will use @xmath319 for its components .
[ lem5a]suppose that @xmath145 is a non - constant solution of ( [ 1 ] ) satisfying one of the following sets of conditions at some @xmath320@xmath321 then @xmath322 suppose that ( i ) holds .
then @xmath323 and @xmath324 can not both vanish .
if @xmath325 and @xmath326 then @xmath327 hence in some interval @xmath328 , @xmath329 if @xmath330 and @xmath331 then @xmath332 once again we see that ( [ 6aa ] ) holds on some interval @xmath328
. consider the backward system satisfied by @xmath333 if @xmath334 then @xmath2{c}u^{\prime}=\frac{u - v}{c}\\ v^{\prime}=-w\\ w^{\prime}=b^{2}\left ( qs\left ( u\right ) -v\right ) \\
q^{\prime}=\frac{\varepsilon}{c}\left ( q+\beta qs\left ( u\right ) -1\right ) \end{array } \right . .
\label{6}\ ] ] also , @xmath335 from ( [ 6aa ] ) and ( [ 6a ] ) it follows that on some interval @xmath336 , @xmath337 we claim that these inequalities hold for all @xmath338 if , on the contrary , one of them fails at a first @xmath339 then @xmath340 but ( [ 8 ] ) , ( [ 6 ] ) , and ( [ 6a ] ) imply that at @xmath341 all of the inequalities in ( [ 7 ] ) still hold , because @xmath342 this contradiction implies that @xmath343 , @xmath344 , and @xmath345 continue to increase , and @xmath346 continues decrease on @xmath347 and in particular , @xmath343 does not tend to @xmath256 as @xmath348 thus , @xmath322 the proofs in cases ( ii ) , ( iii ) and ( iv ) are similar and left to the reader .
this completes the proof of lemma [ lem5a ] .
we now begin our study of the properties of @xmath349 lemma [ lem5a ] will assist us in proving the following result .
[ lem6]the number @xmath350 is still defined , as the first zero of @xmath351 and @xmath352 either @xmath314 is homoclinic or @xmath353 is still defined , as the first zero of @xmath354 . also , if @xmath314 is not homoclinic then @xmath355 on @xmath356.$ ] this lemma eliminates the graph in figure [ figcc]-a .
suppose that @xmath350 is not defined .
then @xmath357 on @xmath358 since @xmath170 is the only equilibrium point of ( [ 1 ] ) , this implies that for some @xmath198 , @xmath359 and @xmath360 then these inequalities hold at @xmath198 for nearby @xmath55 and by proposition [ prop1 ] , @xmath361 for @xmath362 hence @xmath363 on @xmath364 and so @xmath365 contradicting the definition of @xmath366 therefore @xmath350 is defined .
we now show that @xmath367 again assume that @xmath368 and suppose that @xmath369 if @xmath370 then @xmath112 is a local maximum of @xmath211 which is not possible because @xmath112 is the first zero of @xmath371 hence at @xmath112 , @xmath372 if @xmath373 then at @xmath374 @xmath375 by lemma [ lem7 ] .
this implies that @xmath192 changes sign from negative to positive at @xmath374 again a contradiction of the definition of @xmath376 hence at @xmath374 @xmath377 or @xmath378 , since @xmath379 also , @xmath380 in some interval @xmath381 however @xmath10 is bounded by @xmath382 and does not tend to a limit above @xmath383 therefore @xmath192 changes sign at some @xmath384 .
since @xmath212 on @xmath189,$ ] @xmath385 on this interval . at @xmath198 , @xmath386 and
so there is a point @xmath193 in @xmath387 such that @xmath388 and @xmath389 hence at @xmath390 @xmath391 but by lemma [ lem2a ] @xmath392 and @xmath393 is increasing in @xmath394 so @xmath395 again a contradiction of lemma [ lem7 ] .
we have therefore proved the first sentence of lemma [ lem6 ] .
[ lem8]if @xmath396 then @xmath397 as long after @xmath112 as @xmath398 suppose instead that there is a first @xmath399 such that @xmath400 on @xmath189 $ ] but @xmath401 .
then @xmath402 first consider the case @xmath136 on @xmath189.$ ] if @xmath403 then @xmath404 changes from negative to positive before @xmath405 and this will be true as well for @xmath16 close to @xmath406 contradicting the definition of @xmath407 hence suppose that @xmath408 if @xmath409 then @xmath192 has a local minimum at @xmath202 contradicting the definition of @xmath217 hence@xmath410 but now the conditions in ( i ) of lemma [ lem5a ] are satisfied , giving a contradiction .
we have left to consider the case that @xmath411 then @xmath412 if @xmath403 then @xmath413 so @xmath136 in an interval @xmath414 hence in this case , @xmath192 changes sign ( from negative to positive ) before @xmath415 for @xmath16 close to @xmath305 there are two possibilities : either @xmath416 changes sign from negative to positive before @xmath417 and so before @xmath418 , or else @xmath419 in a neighborhood of @xmath202 but in a neighborhood of , say , @xmath420 , @xmath79 and @xmath421 .
( see figure [ figure3 ] ) neither of these possibilities occurs if @xmath289 so once again , @xmath422 a contradiction .
this proves lemma [ lem8 ] .
it follows that there is a first @xmath423 such that @xmath424 also , @xmath425 and ( equivalently by the fourth item of proposition [ prop2 ] ) @xmath426 [ lem9]@xmath427 and @xmath355 as long after @xmath428 as @xmath429 this lemma eliminates the graph in figure [ figcc]-b .
let @xmath430 since @xmath431 @xmath216 and @xmath397 on some interval @xmath432 $ ] with @xmath433 we claim that @xmath216 on any such half - closed interval in which @xmath434 this follows because , by proposition [ prop2 ] , @xmath435 at any point where @xmath436 and @xmath434 we next show that @xmath397 on any interval @xmath432 $ ] in which @xmath216 and @xmath437 if not , then there is a first @xmath438 with @xmath439 @xmath440 and @xmath441 on @xmath213.$ ] then @xmath442 if @xmath443 then @xmath357 in some interval @xmath444 in this case , for @xmath16 close enough to @xmath406 @xmath416 changes sign after @xmath112 but before @xmath445 or else @xmath165 crosses @xmath256 and back again , and such @xmath16 can not lie in @xmath446 a contradiction . hence ,
@xmath447 but then , because @xmath448 @xmath449 in the region where @xmath216 and @xmath450 @xmath451 hence , @xmath452 and again @xmath357 in an interval to the right of @xmath193 but before @xmath453 a contradiction as before .
the only other possibility contradicting lemma [ lem9 ] is that there is a first @xmath454 where @xmath455 and @xmath456 we consider two cases : ( a ) @xmath457 and @xmath458 and ( b ) @xmath459 first consider ( a ) . in an interval
@xmath460 @xmath427 , @xmath461 and @xmath462 and so at @xmath202 if @xmath368 then @xmath463 also , @xmath464 but @xmath465 is impossible because it means that even for nearby @xmath55 @xmath79 after @xmath112 but before @xmath466 therefore at @xmath198 , @xmath467 and @xmath468 then @xmath469 but on the nullcline @xmath405 with @xmath421 , @xmath470 and @xmath471 @xmath472 this implies that @xmath192 has a local minimum at @xmath198 , whereas we know that @xmath397 in @xmath473 this contradicts the definition of @xmath217 turning to case ( b ) , we now have that at @xmath202 @xmath474 thus @xmath475 if @xmath476 then @xmath226 to the right of @xmath217 as before , if @xmath16 is close to @xmath305 then either @xmath165 crosses @xmath256 twice , or @xmath162 does nt reach the region @xmath477 before @xmath478 both of which mean that @xmath479 .
if @xmath480 then @xmath209 is again an equilibrium point .
the third possibility , @xmath481 implies that ( ii ) of lemma [ lem5a ] is satisfied , and thus again gives a contradiction .
this completes the proof of lemma [ lem9 ] .
if @xmath421 on @xmath67 then @xmath355 and @xmath482 on @xmath483 , and @xmath314 is homoclinic .
this proves lemma [ lem6 ] .
thus , for @xmath484 if @xmath314 is not homoclinic then @xmath114 exists with @xmath485 and @xmath397 on @xmath356.$ ] however , there is no @xmath113 such that @xmath477 on @xmath486 $ ] and either @xmath487 or @xmath300 , for otherwise @xmath488 and this has already been ruled out . therefore if @xmath489 then @xmath490 and @xmath491 for otherwise nearby values of @xmath16 are once again not in @xmath291 if @xmath314 is not homoclinic and @xmath484 , then there must be a first @xmath492 with @xmath493 , @xmath494 @xmath495 and @xmath496 suppose that this is the case and also @xmath497 ( this is pictured in figure [ figcc]-c . ) then @xmath498 if @xmath499 then ( iii ) of lemma ( [ lem5a ] ) applies and gives a contradiction .
hence @xmath500 then at @xmath202 @xmath501 and @xmath502 ( this is pictured in figure [ figcc]-d . but
this is case ( ii ) of lemma [ lem5a ] and so also impossible .
we have established that if @xmath314 is not homoclinic ( with @xmath421 on @xmath67 ) then for large @xmath147 @xmath503 and @xmath504 this is only possible if @xmath314 is homoclinic ( with @xmath505 and @xmath357 for large @xmath64 ) .
this proves lemma [ lem4 ] . to complete the proof of theorem [ thm1 ] we look for a second homoclinic orbit , with @xmath506 again
we adapt the method in @xcite .
it is stated so as to be useful in the proofs of theorems [ thm2 ] and [ thm3 ] , as well as theorem [ thm1 ] .
[ lem10]there are @xmath507 and @xmath508 both independent of @xmath5 such that if @xmath509 and @xmath510 then the solution @xmath162 remains in the region @xmath511 on @xmath512 and @xmath10 crosses @xmath513 since the proof uses some of the easier parts of the proof of lemma [ lem2 ] , it is included in the appendix . from here to the end of this section the parameters @xmath20 and @xmath115 remain as in the previous subsection .
lemma [ lem10 ] implies that if we extend @xmath287 to @xmath514 with otherwise the same definition as above , then @xmath515 this suggests that @xmath516 corresponds to a homoclinic orbit .
the problem with this argument is that the concept of front in the sense used in the method of geometric perturbation , breaks down for small @xmath517 the slow homoclinic orbit is not close , even up to the first zero of @xmath518 to the front found when @xmath250 more precisely , our proof of the first sentence of lemma [ lem6 ] is no longer valid , because we can not assert that the first zero of @xmath416 occurs with @xmath519 hence we must modify our shooting set on the @xmath16 axis .
this requires several steps .
our argument from here no longer refers to a point @xmath112 where @xmath416 changes sign , but instead considers solutions such that @xmath520 changes sign .
let @xmath521~|~\text{there is a } \tau_{1}>0\text { such that } q_{c}^{\prime}<0\text { on } \left ( -\infty,\tau_{1}\right ) , \right .
\\ & \left .
\text { } q_{c}^{\prime}\left ( \tau_{1}\right ) = 0\text { , and } u_{c}^{\prime}\left ( \tau_{1}\right ) < 0\right\ } .\end{aligned}\ ] ] our argument does not require that if @xmath522 then @xmath416 has only one zero in @xmath523 though numerically this appears to be the case .
@xmath524 recall that @xmath20 and @xmath115 were chosen so that @xmath525 has a unique zero . as in the proof of lemma [ lem9
] , this implies that @xmath526 has a unique zero , say @xmath527 and @xmath528 hence @xmath529 .
also , lemma [ lem10 ] shows that there is an interval @xmath530 which contains no points of @xmath531 let @xmath532 [ lem11]there is a @xmath428 such that @xmath533 on @xmath523 @xmath534 @xmath535 and @xmath536 if @xmath533 on @xmath67 then there is a @xmath537 such that @xmath538 and @xmath539 from the continuity of @xmath540 with respect to @xmath55 the same is true for @xmath165 if @xmath16 is sufficiently close to @xmath541 in particular , again @xmath542 on @xmath543 .
but then @xmath544 contradicting the definition of @xmath545 .
therefore a first @xmath428 is defined such that @xmath546 .
then @xmath547 also , by proposition [ prop2 ] , @xmath548 if @xmath549 then by the implicit function theorem , @xmath550 is defined for nearby @xmath16 as the first zero of @xmath551 with @xmath552 and @xmath542 on @xmath553 contradicting the definition of @xmath541 hence @xmath554 . if @xmath555 then @xmath428 is a local minimum of @xmath556 contradicting the definition of @xmath557 if @xmath558 then @xmath559 and since @xmath560 and @xmath561 @xmath562 and @xmath563 this implies that @xmath564 on an interval @xmath565 again a contradiction .
hence @xmath566 completing the proof of lemma [ lem11 ] .
thus , @xmath533 in some interval @xmath567 this result implies that @xmath568 however the interval @xmath569\subset\sigma.$ ] lemma [ lem11 ] also implies that points in @xmath569 $ ] near to @xmath545 are not in @xmath570 since the corresponding solutions on @xmath172 must have a change of sign of @xmath520 from positive to negative after @xmath571 let @xmath572 we claim that @xmath573 is a homoclinic orbit .
the proof uses techniques very similar to those above .
first observe that @xmath574 and @xmath575 therefore @xmath576 is defined as in the definition of @xmath531 then use the following result .
[ lemslow2]if @xmath577 then @xmath419 on any interval @xmath578 $ ] in which @xmath579 if @xmath580 at some first @xmath438 with @xmath581 then @xmath582 but in the region where @xmath441 and @xmath583 @xmath29 is positive , and this implies that @xmath162 crosses into @xmath584 a contradiction of the definition of @xmath585 if @xmath586 is not homoclinic then there is a @xmath587 such that @xmath588 and @xmath589 on @xmath590 .$ ] further , @xmath591 on @xmath592 .
let @xmath593 lemma [ lemslow2 ] implies the existence of @xmath594 suppose there is a first @xmath595 with @xmath596 from the definitions of @xmath597 and @xmath598 @xmath599 on @xmath600 since @xmath601 is decreasing and @xmath216 if @xmath602 @xmath10 can not increase indefinitely .
hence there is a @xmath603 with @xmath604 a contradiction then results from ( iv ) of lemma [ lem5a ] .
now apply the technique of lemma [ lem5 ] , including use of proposition [ prop2 ] and lemma [ lem5a ] , to show that @xmath605 and @xmath606 on @xmath607 .
in particular , lemma [ lem5a ] is used to show that there is no @xmath608 ( in fact , no @xmath64 at all ) with @xmath609 it follows that on @xmath610 @xmath611 and so indeed , @xmath586 is homoclinic .
this completes the proof of theorem [ thm1 ] .
as mentioned above , theorem [ thm1a ] follows from theorems [ thm1 ] and [ thm2 ] . in theorem [ thm2
] @xmath20 is not fixed . also , @xmath115 is any number in @xmath514 which however is fixed at this stage for the rest of this section .
let @xmath612 $ ] .
we note that the unstable manifold @xmath172 varies continuously with @xmath180 for @xmath129 and @xmath613 to be more precise , if @xmath614 exists on @xmath615 $ ] then in some neighborhood of @xmath616 @xmath617 exists for @xmath618 and is a continuous function of @xmath619 from the third sentence of lemma [ lem2 ] it follows that @xmath115 can be chosen in @xmath620 such that if @xmath621 then @xmath622 where @xmath623 is the unique zero of @xmath624 this proves the first assertion of theorem [ thm2 ] .
we now choose @xmath115 in this way .
[ lem2a]there is an @xmath625 and a @xmath626 such that if @xmath627 and @xmath628 $ ] , then @xmath629 for some first @xmath630 and @xmath236 on @xmath189.$ ] ( hence , @xmath318 satisfies the the first condition of theorem [ thm1 ] . )
further , @xmath631 can be chosen so that @xmath117 satisfies conditions ( [ i ] ) and ( [ ii ] ) of theorem [ thm1 ] . from the choice of @xmath116 lemma [ lem2 ]
implies that for any @xmath632 there is a @xmath633 such that if @xmath634 then @xmath635 in the interval @xmath636 $ ] where @xmath637 from @xmath638 it follows that for some @xmath639 if @xmath640 then @xmath641 before @xmath642 the uniform continuity of @xmath643 in @xmath644 for @xmath645 and @xmath646 in any compact interval @xmath647 $ ] with @xmath648 then implies the first conclusion of the lemma .
the remaining assertion of lemma [ lem2a ] follows by similar arguments .
we have now proved theorems [ thm1 ] , [ thm2 ] , and [ thm1a ] , in that order . to prove theorem [ thm3 ] apply a continuity argument similar to that just used to show that for any @xmath633 there is an @xmath649 such that if @xmath650 and @xmath651 the pair @xmath652 satisfy the hypotheses on @xmath125 in theorem [ thm1 ] .
it follows that pulses exist for some @xmath653 and some @xmath654 but lemma [ lem10 ] implies that @xmath124 theorem [ thm3 ] follows .
as stated earlier , there is an additional hypothesis in the existence result given in @xcite , namely hypothesis 3.1 in that paper . this hypothesis is interesting in a broader context , and we will include some comments on its relation to the well - known pde model of fitzhugh and nagumo . to state this hypothesis we need to introduce a basic tool in the method of geometric perturbation , the so - called singular solution . the singular solution of ( [ 1 ] ) is a continuous piecewise smooth curve in @xmath655 consisting of four smooth pieces .
the first piece is the front with speed @xmath100 found in lemma [ lem1 ] .
( recall that fronts were defined just after the statement of this lemma . ) in figure [ figfront ] the green line segment is the projection of the graph of the front onto the @xmath111 plane .
the second piece of the singular solution is a segment of the nullcline @xmath656 ( with @xmath657 @xmath658 ) as shown in figure [ figslow ] .
it is obtained from ( [ 1 ] ) by letting @xmath659 formally setting @xmath660 in the resulting system of ode s for @xmath661 and solving the resulting set of one differential equation and three algebraic equations , one of which is @xmath662 for more information on this segment , and the singular solution in general , see @xcite .
we do nt need to say more about this segment here . to define the third part of the singular solution ( crucial in hypothesis 3 of @xcite )
, we consider the fast system ( [ 2 ] ) , but with @xmath164 replaced by @xmath663 as a parameter ranging between @xmath664 and @xmath665 @xmath2{c}u^{\prime}=\frac{v - u}{c}\\ v^{\prime}=w\\ w^{\prime}=b^{2}\left ( v - qs\left ( u\right ) \right ) \end{array } \right .
\label{fast}\ ] ] for each @xmath666 $ ] there is a unique @xmath667 such that ( [ fast ] ) has a bounded non - constant solution @xmath668 the graph of this solution is a heteroclinic orbit connecting the left and right branches of @xmath669 in the @xmath12 plane .
there is at least one value of @xmath670 such that @xmath671 this will be true for any @xmath53 such that @xmath672 for @xmath673 and sufficiently close to @xmath664 , the integral above is negative , and whenever this is the case , the connecting heteroclinic orbit exists for some @xmath674 but unlike the front defined earlier , @xmath675 is decreasing , from @xmath676 to @xmath677 such a solution is called a back . if @xmath678 then there is a back for any @xmath679 all this can be proved using methods from the appendix , or see @xcite .
the third part of the singular solution of ( [ 1 ] ) is a back , at a value @xmath680 with speed @xmath681 there must be at least one @xmath682 for which such a back exists .
as far as we know there is no proof that @xmath16 is a monotone function of @xmath53 in @xmath683 so possibly there could be more than one such @xmath684 in this case we can require that the jump down is at the largest possible @xmath682 supporting a back with speed @xmath685 the singular solution is said to jump down above the knee if @xmath686 numerical computations suggest that often no such @xmath682 exists in @xmath687 . in this case , there is still a traveling back with speed @xmath688 but it is at @xmath689 the fourth part of the singular solution is a slow return along the left branch of the nullcline .
hypotheses of faye : \(i ) _ the system ( [ 1 ] ) has a unique equilibrium point _ @xmath690 . _ _ _ ( ii ) if _ @xmath36 _ then the equation _
@xmath691 _ has exactly three solutions , _
@xmath692 _ with _ @xmath693 _ _ @xmath694 _ and _ @xmath695 _ _ _ _ ( iii)__@xmath696 _ ( iv ) the speed of any back with _
@xmath697 _ is less than _
it follows that the back of the singular solution is required to be at the knee .
this condition can only be verified by numerical integration of the fast system . * * ( * * faye * ) : * _ under hypotheses ( i ) , ( ii ) , ( iii ) , and ( iv ) , if _
@xmath20 is sufficiently small then _ the system ( [ 1 ] ) has a homoclinic orbit for at least one positive value of _
geometrical perturbation , based on work of fenichel and others , is a technique for showing that the singular solution is close to a real solution if @xmath20 is sufficiently small .
certain transversality conditions can be complicated to check , requiring a technique called blow - up .
we can contrast the existence of a back at the knee with the well - known behavior of the pde model of fitzhugh and nagumo .
( see @xcite for a presentation of this model and a proof that it has two traveling pulses . )
the fast fitzhugh - nagumo pulse can be described as having a jump up , ( close to a front ) during which @xmath10 increases rapidly while @xmath245 is nearly zero , followed by a slow increase in @xmath699 and then a jump down ( close to a back ) , with @xmath245 again nearly constant ( but positive ) , and @xmath10 decreasing rapidly . in this case , the back occurs before @xmath700 reaches the knee .
see [ figure5 ] . we have done some preliminary numerical investigation to test whether it is possible , in the model studied in this paper and with the particular function @xmath1 in ( [ -3 ] ) , to adjust the parameters @xmath51 and @xmath26 so that the jump down occurs before @xmath245 reaches the knee .
we have not found such a pair @xmath701 but we can not assert that none exists . for the fitzhugh - nagumo model , however , it is clear that the jump down is always before reaching the knee .
this follows because the reaction term in equations is a cubic polynomial , @xmath702 where @xmath703 this function is symmetric around its inflection point , which leads to the `` before the knee '' behavior of the singular solution .
so we searched numerically for alternative functions to use for @xmath704 which , while still cubic like , permit the down jump of the singular solution to be at the knee .
we found such a function , as illustrated in figure [ figure6 ] .
we are not aware of a method which determines analytically where the downjump occurs , either for the model of faye or that of fitzhugh - nagumo when @xmath704 is asymmetric .
the existence proof in @xcite for fast and slow homoclinic orbits of the fitzhugh - nagumo system also applies to functions @xmath704 such as that pictured in figure [ figure6 ] . on the other hand it appears that the proof by geometric perturbation in this case , while probably still basically valid , requires a more complicated analysis because the downjump of the singular solution may occur at the knee .
it appears that the fast - slow analysis used to get the singular solution , for any of the models we have discussed , does not apply to the slow pulse .
hence it appears difficult to use geometric perturbation to obtain this solution . in [ @xcite this solution
was obtained in the fitzhugh - nagumo case using dynamical systems methods , but only for @xmath705 sufficiently close to @xmath706 the analysis in @xcite was in the region @xmath705 close to @xmath707 @xmath20 and @xmath16 small .
indeed , if @xmath708 then there is only one pulse and it is a standing wave ( @xmath709 ) .
it is proved in @xcite that for @xmath710 positive but small there is a smooth curve @xmath711 corresponding to pulse solutions and connecting @xmath712 to @xmath366 the speeds @xmath712 and @xmath406 and the maximum of @xmath713 in @xmath714 $ ] all tend to zero as @xmath715 we suspect that a similar picture holds for the model of faye , but it is not clear that our analysis is able to prove this much . the fitzhugh - nagumo condition that @xmath716 is small would be replaced here by requiring that @xmath717 is small .
we said in the introduction that there was only a partial existence result for the model in @xcite .
we were referring there to a recent paper by faye and scheel @xcite , which uses an interesting extension of the geometric perturbation method to infinite dimensional spaces to handle this kind of problem .
the method is powerful because it allows an extension beyond the sorts of kernel which reduce the problem to an ode .
faye and scheel remark that their paper appears to apply to the pinto - ermentrout model .
this is true , but with a limitation .
a key hypothesis in their paper is that for the singular solution , as described above , the jump down occurs above the knee . in a private communication professor ermentrout
has observed that while this is true for the pinto - ermentrout model in some parameter ranges , it is also common for the down - jump to occur at the knee .
this is why we characterized their result as " partial for pinto - ermentrout .
unfortunately , we have not been able to make our approach work for pinto - ermentrout .
the reason may be related to an important difference between ( [ 1 ] ) and the equivalent set of ode s obtained from ( [ pe ] ) .
the linearization of ( [ 1 ] ) around its equilibrium point has only real eigenvalues , for any @xmath21 while the equivalent linearization for ( [ pe ] ) has complex eigenvalues for a range of positive @xmath20 .
thus , a homoclinic orbit would oscillate around equilibrium .
a few oscillations could occur even for the very small values of @xmath20 where the eigenvalues are real .
the final steps in our proof above clearly do not allow such oscillations .
in [ @xcite ] we showed that there was co - existence of complex roots and a homoclinic orbit for fitzhugh - nagumo , and observed that work of evans , fenichel and feroe then implied the existence of many periodic solutions and a form of chaos .
this leads to a conjecture that the pinto - ermentrout model supports a richer variety of bounded solutions than the model of faye .
the new solutions are probably unstable , however , so their physical importance is unclear .
a local stability result for the fast solution was proved by faye .
his proof depends crucially on analysis of both the front and back of his solution , as described above .
the analysis of the back is less standard because of the assumption that the jump down is at the knee .
here he relies on previous work on similar problems .
presumably if the jump were above the knee ( where , however his existence proof is not claimed to apply ) , the stability analysis would be easier .
condition ( [ i ] ) says that the solution @xmath117 is on the unstable manifold @xmath718 of ( [ 1 ] ) at the equilibrium point @xmath719 to check condition ( [ ii ] ) we must follow @xmath117 until a point where @xmath720 our proposal for doing this is based on @xcite , where a similar procedure was followed for the well known equations of lorenz . using a standard ode solver we can arrive at a conjectured value for @xmath721 to begin analyzing @xmath117
numerically we would expand the solutions around @xmath719 a high order expansion of @xmath172 results in algebraic expressions which are then evaluated using rigorous numerical analysis based on interval arithmetic . with this technique
one hopes to show that @xmath172 enters a very small box near @xmath170 . for the example in @xcite this box had a diameter of about @xmath722 this gives us an initial estimate accurate to ( say ) 68 significant digits .
from there , a rigorous ode solver , as described for example in @xcite , would be used to continue @xmath117 until @xmath723 whether this can be done can not be determined ahead of time .
one has to run the solver .
the number of guaranteed accurate digits decreases as the integration proceeds .
we then hope that some significant digits would be maintained long enough to reach @xmath724 it must be checked along the way that @xmath725 based on the great sensitivity of the lorenz equations to the initial conditions , we expect that this would be easier for the faye model than it was in @xcite . we are not aware of any proposal to try to estimate @xmath20 for the method of geometric perturbation . 99 alefeld , g. and mayer , g. , interval analysis , theory and applications , _ j. comp
. appld . math _
* 121 * ( 2000 ) , 421 - 464 .
dumortier , f. , llibre , c. , and arts , c. _ qualitative behavior of planar systems _ , springer , 2006 .
ermentrout , g. b. and mcleod , j. b. , existence and uniqueness of waves for a neural network , _ proc .
edinburgh sect . a _ * 123 * , 451 - 478 . faye , g , existence and stability of travelling pulses in a neural field equation with synaptic depression , _ siam j. of dynamical systems _ * 10*(2013 ) , 147 - 160 .
faye , g. and scheel , a. , existence of pulses in excitable media with nonlocal coupling , _ advances in mathematics _ * 270 * 2015 , 400 - 456 .
fenichel , n. , geometric perturbation theory for ordinary differential equations , _
* 31 * ( 1979 ) , 53 - 98 .
hassard , b. , hastings , s. p. , troy , w. c. , zhang , j. , a computer proof that the lorenz equations have chaotic solutions .
math . letters _ * 7 * 1994 , 79 - 63 .
hastings , s. p. , on wave solutions of the hodgkin - huxley equations , _ arch .
anal _ * 60 * 1972 , 229 - 257 .
hastings , s. p. , single and multiple pulse waves for the fitzhugh - nagumo equations , _
siam j. applied math _ * 42 * 1982 , 247 - 260 .
hastings , s. p. and mcleod , j. b. , _ classical methods in ordinary differential equations , _ amer .
hartman , p. , _ ordinary differential equations _ , classics in applied mathematics , siam , 2002 .
hodgkin , a.l . ,
huxley , a.f .
, a quantitative description of membrane current and its application to conduction and excitation in nerve , _ journal of physiology _ * 117 * 1952 , 500 - 544 .
jones , c. k. r. t. , kopell , n. , langer , r. , construction of the fitzhugh - nagumo pulse using differential forms , _ patterns and dynamics in reactive media _ , i m a volumes in mathematics and its applications * 37 * ( 1991 ) , 101 - 116 .
kilpatrick , z. and bressloff , p. , effects of synaptic depression and adaptation on spatio - temporal dynamics of an excitatory neural network .
_ physica d. _ * 239 * 2010 , 547 - 560 .
krupa , m. , sandstede , b. , and szmolyan , p. , fast and slow waves in the fitzhugh - nagumo equations , _
* 133 * 1997 , 49 - 97 .
pinto , d. , ermentrout g. b. , spatially structured activity in synaptically coupled neuronal networks : i traveling fronts and pulses , _
siam j. applied math _ * 62 * 2001 , 206 - 225 .
rauch , j. , smoller , j. , qualitative theory of the fitzhugh - nagumo equations , _ advances in mathematics _ * 27 * 1978 , 12 - 44 .
we prove these results together .
the linearization of ( [ 2 ] ) around @xmath726 is the system @xmath727 with @xmath728{ccc}-\frac{1}{c } & \frac{1}{c } & 0\\ 0 & 0 & 1\\ -b^{2}q_{0}s^{\prime}\left ( u_{0}\right ) & b^{2 } & 0 \end{array } \right ) .\ ] ] the characteristic polynomial of @xmath729 is @xmath730 recall that @xmath731 .
condition [ c2 ] implies that the equation @xmath732 has three solutions , @xmath692 and by condition [ c0a ] , @xmath733 it follows that @xmath734 therefore @xmath735 also , @xmath736 and both @xmath737 and @xmath738 are positive for @xmath739 hence @xmath729 has one real positive eigenvalue . also , @xmath740 which implies that @xmath729 has two real negative eigenvalues .
further , it is easily seen that there is an eigenvector corresponding to the positive eigenvalue of @xmath729 which points into the positive octant .
if @xmath741 is a solution lying on the branch @xmath742 of the unstable manifold of ( [ 2 ] ) at @xmath743 , then initially , @xmath211 @xmath744 and @xmath29 are positive .
it follows from the first two equations of ( [ 2 ] ) that @xmath511 as long as @xmath218 .
also , @xmath745 for @xmath746 and so @xmath142 while @xmath10 is in @xmath747 .$ ] hence there is a first @xmath748 such that @xmath749 and we can assume that @xmath750 we have now proved the assertions of the first and second sentences of lemma [ lem1 ] . for the third sentence of lemma [ lem1 ] , and for all of lemma [ lem2 ] , we need a comparison lemma . for each @xmath61 let @xmath751 be the unique solution of ( [ 2 ] ) on @xmath742 such that @xmath752 on @xmath229 $ ] and @xmath753 suppose that @xmath752 on a maximal interval @xmath754 where possibly @xmath755 then in @xmath756 we can consider @xmath10 and @xmath245 as functions of @xmath255 , letting @xmath757 and @xmath758 this defines the functions @xmath759 and @xmath760 on the interval @xmath761 and for @xmath255 in this interval , @xmath762 [ lem12 ] if @xmath763 then @xmath764 in the interval @xmath765 , @xmath2{c}u_{d_{2}}<u_{d_{1}}\\ w_{d_{2}}>w_{d_{1}}\end{array } \right . .
\label{a1a}\ ] ] the first sentence follows by proving ( [ a1a ] ) on the smaller of the two intervals .
we first show that these inequalities hold on some initial interval @xmath766 .
this is seen by comparing unit eigenvectors corresponding to the positive eigenvalues @xmath767 and @xmath768 of the linearizations of ( [ 2 ] ) around @xmath743 .
suppose that for a particular @xmath16 the eigenvector corresponding to @xmath769 is @xmath770 then @xmath771 inequalities ( [ a1a ] ) follow near @xmath743 if @xmath772 for this we turn to the characteristic polynomial of @xmath773 given in ( [ 17 ] ) but now denoted by @xmath774 .
it is easier to work with @xmath775 noting that @xmath182 the positive eigenvalue of @xmath729 is determined by the equation @xmath776 and the condition @xmath777 then @xmath778 since @xmath779 @xmath780 and @xmath781 for @xmath782 @xmath783 .
@xmath784also , @xmath785 @xmath786 it follows that @xmath787 if @xmath788 but @xmath789 so indeed , @xmath790 therefore ( [ a1a ] ) holds on some interval @xmath791 suppose that the first inequality fails at a first @xmath792 while the second holds over @xmath793.$ ] then at @xmath267 , @xmath794 @xmath795 but then , @xmath796 a contradiction since @xmath797 on @xmath798.$ ] a similar argument eliminates the other possibilities , using the fact that @xmath1 is increasing , and this completes the proof of the lemma [ lem12 ] .
[ cor1]if @xmath799 then @xmath800 [ lem12b]if @xmath801 , @xmath802 and @xmath803 then either @xmath804 or @xmath805 suppose that @xmath806 and @xmath807 lemma [ lem12 ] implies that @xmath808 and so @xmath809 a contradiction because @xmath810 is the first zero of @xmath811 next we must show the existence of @xmath812 [ lem13a ] for sufficiently large @xmath55 @xmath755 from ( [ a01 ] ) , in the interval @xmath813 where @xmath814 is increasing , @xmath815 recall that @xmath816 in @xmath817 as long as @xmath818 and @xmath819 @xmath820 this implies that for large @xmath16 , in the interval where @xmath821 @xmath822 grows rapidly and so , in turn , does @xmath823 . in particular , @xmath824 before @xmath825 and this implies that @xmath826 now we wish to show that for small @xmath61 @xmath827 before @xmath828 it is in this step that condition [ c3 ] is used .
[ lem14 ] there is a @xmath829 such that for any @xmath61 if @xmath830 and @xmath831 then @xmath832 for @xmath833 and @xmath814 leaves the interval @xmath834 . if @xmath835 then @xmath814 crosses @xmath382 , while if @xmath836 then @xmath814 crosses @xmath837 .
let @xmath838 .
since @xmath839 if @xmath840 then for @xmath841 @xmath842 from which follows that @xmath255 must leave @xmath843 before @xmath844 [ lem15]if @xmath845 on @xmath189 $ ] then @xmath846 on this interval .
if @xmath847 on @xmath213,$ ] then @xmath848 on this interval . with @xmath849 ,
@xmath850 so if @xmath851 then @xmath852 also , if @xmath853 in @xmath854 then@xmath855 since @xmath856 as @xmath857 , the first sentence of the lemma follows and the second is similar
. based on this lemma , we consider , in addition to ( [ 2 ] ) , the system@xmath2{c}v^{\prime}=w\\ w^{\prime}=b^{2}\left ( v - q_{0}s\left ( v\right ) \right ) \end{array } \right .
, \label{10}\ ] ] this system has equilibrium points at @xmath858 @xmath859 and @xmath860 and a standard phase plane analysis , assuming condition [ c3 ] , shows that the positive branch @xmath861 of unstable manifold of ( [ 10 ] ) at @xmath862 is homoclinic .
also we consider the system @xmath2{c}v^{\prime}=w\\ w^{\prime}=b^{2}\left ( v - q_{0}s\left ( v-\hat{c}\right ) \right ) \end{array } \right . ,
\label{11}\ ] ] for small @xmath863 choose @xmath864 so small that this system also has three equilibrium points , and a homoclinic orbit based at the left most of these .
this orbit entirely encloses the homoclinic orbit of ( [ 10 ] ) .
finally we consider the system @xmath2{c}v^{\prime}=w\\ w^{\prime}=b^{2}\left ( v - q_{0}s\left ( v+\hat{c}\right ) \right ) \end{array } \right . , \label{12a}\ ] ] for sufficiently small @xmath864 this system also has a homoclinic orbit .
this orbit lies entirely inside the homoclinic orbit of ( [ 10 ] ) .
however , the lower left branch @xmath865 of the unstable manifold of this system crosses the homoclinic orbits of ( [ 10 ] ) and ( [ 11 ] ) , and this branch will play a role below .
( see figure 12 . )
[ figure77 ] from now on , @xmath866 @xmath867 and @xmath868 will denote the unique solutions of the systems ( [ 10 ] ) , ( [ 11 ] ) , and ( [ 12a ] ) respectively which lie on the homoclinic orbits of those systems and satisfy @xmath869 in each of these cases , if @xmath870 is homoclinic then @xmath871 is bounded by @xmath872 .
this follows from the definition of @xmath872 in lemma [ lem14 ] , the results of which also apply to ( [ 11 ] ) and ( [ 12a ] ) , with the same proofs .
if @xmath871 exceeds @xmath872 then @xmath31 is not bounded .
recall that in lemmas [ lem1 ] and [ lem2 ] , @xmath873 denoted the unique solution on the unstable manifold @xmath59 such that @xmath874 and @xmath81 on @xmath229.$ ] in the rest of this proof we will denote this solution by @xmath875 by lemma [ lem15 ] we can choose @xmath16 so small that if @xmath112 is the first zero of @xmath876 then @xmath877 on @xmath878.$ ] [ lem13]condition [ 13 ] implies that if @xmath752 on @xmath225,$ ] then @xmath879 is in the annular region between the orbit of @xmath880 and the orbit of @xmath881 the proof is similar to the proof of lemma [ lem12 ] . as long as @xmath882 @xmath165 and @xmath822 can be considered functions of @xmath883 also , @xmath884 by considering the eigenvalues of the linearizations of ( [ 10 ] ) as functions of @xmath16 we can show , using ( [ 17a ] ) , that for large negative @xmath147 @xmath885 lies in the claimed annular region .
suppose that for some first @xmath202 @xmath886 lies on the upper boundary of this region , that is , on the homoclinic orbit of ( [ 11 ] ) , at a point where @xmath887 the slope of this homoclinic orbit at this point is @xmath888 but @xmath889 ( since @xmath890 as long as @xmath891 ) , and since @xmath1 is increasing and @xmath892 , it follows from ( [ 14 ] ) that @xmath893 and so the curve @xmath894 arrives at this point from outside of the annular region , contradicting the definition of @xmath198 .
in a similar manner it is shown that @xmath894 lies above the orbit of @xmath880 as long as @xmath895 this uses the bound @xmath890 as long as @xmath752 .
a similar comparison shows that if @xmath896 is the first point where @xmath897 then @xmath898 is an increasing function of @xmath55 for @xmath899 and that for @xmath279 @xmath752 on @xmath900 this shows the uniqueness of @xmath812 to complete the proof of lemma [ lem2 ] we show that from the first @xmath112 where @xmath901 the curve @xmath902 lies either to the right or below the orbit @xmath903 and also below the left branch of the unstable manifold of ( [ 12a ] ) , at least up to the point where @xmath904 ( @xmath905 ) .
( if @xmath906 then , as in lemma [ lem14 ] , @xmath814 becomes negative , which is what we are trying to show .
see figure 8 . )
this follows by the same sort of comparison as above , now comparing @xmath894 with the lower half of the unstable manifold of ( [ 12a ] ) .
this is possible because by lemma [ lem15 ] , @xmath907 as long as @xmath908 . to prove that @xmath100 as defined in lemma [ lem2 ] exists
, we note that the set of @xmath16 such that @xmath909 for some @xmath910 is open , as is the set of @xmath16 such that @xmath911 for some @xmath912 and @xmath913 on @xmath914 $ ] .
this follows because @xmath162 is a continuous function of @xmath517 lemmas [ lem13a ] and [ lem13 ] imply that these sets are nonempty , and their definitions and proposition [ prop1 ] imply that they are disjoint . since the interval @xmath915 is connected , the existence of some positive @xmath100 which is not in either set .
its uniqueness follows from the corollary to lemma 12 . from the definition of @xmath916 @xmath917 on @xmath900
suppose that there is an @xmath918 with @xmath919 and @xmath920 on @xmath921 then @xmath922 @xmath923 and @xmath924 because @xmath925 on @xmath926 @xmath927 giving @xmath928 this contradiction completes the proof of lemma [ lem1 ] . to complete lemma [ lem2 ] we must prove the assertions in the third and fourth sentences . in the third sentence , @xmath929 when @xmath470 and at the first zero of @xmath211 @xmath930 so @xmath931 the implicit function theorem and the comparison ( [ a1a ] ) imply the limit statement .
for the last sentence of lemma [ lem2 ] , it suffices to prove that @xmath932 suppose instead that @xmath933 since @xmath632 lemma [ lem12 ] and the hypotheses of lemma [ lem2 ] imply that @xmath934 hence @xmath935 which contradicts the definition of @xmath623 .
this completes the proof of lemma [ lem2 ] .
this result is about system ( [ 1 ] ) .
however the argument in lemma [ lem14 ] , initially about system ( [ 2 ] ) , applies equally well to ( [ 1 ] ) , so if @xmath822 increases monotonically to above @xmath872 then @xmath255 crosses @xmath936 followed by @xmath133 hence we can assume that if @xmath910 is the first zero , if any , of @xmath822 , then @xmath937 on @xmath938 .
as earlier in obtaining ( [ 13 ] ) , it follows that if @xmath939 , then @xmath940 on @xmath938 .
therefore , @xmath941 uniformly on @xmath942 $ ] and for @xmath18 . also , @xmath943 uniformly on @xmath944 $ ] and for @xmath61 @xmath18 .
this is proved by the same argument which lead to ( [ x1 ] ) , .
now consider the equation obtained from ( [ 1 ] ) by formally setting @xmath709 in ( [ 1 ] ) , namely @xmath945 where @xmath946 because ( [ 1 ] ) has only one equilibrium point , @xmath947 if @xmath948 hence there is a @xmath949 such that if for some @xmath147 @xmath950 and @xmath951 , then @xmath952 ( here @xmath953 is independent of the particular solution involved . )
lemma [ lem10 ] then follows from ( [ x1 ] ) and ( [ x2 ] ) .
suppose that the linearization of ( [ 1 ] ) around @xmath170 is @xmath954 then @xmath955{rrrr}-\frac{1}{c } & \frac{1}{c } & 0 & 0\\ 0 & 0 & 1 & 0\\ -b^{2}q_{0}s^{\prime}\left ( u_{0}\right ) & b^{2 } & 0 & -b^{2}s\left ( u_{0}\right ) \\
-\frac{\varepsilon}{c}\beta q_{0}s^{\prime}\left ( u_{0}\right ) & 0 & 0 & -\frac{\varepsilon}{c}\left ( 1+\beta s\left ( u_{0}\right ) \right ) \end{array } \right ) .
\label{16}\ ] ] the characteristic polynomial of @xmath956 is@xmath957 while proving lemma [ lem1 ] we showed that if @xmath958 then one of the non - zero eigenvalues of @xmath956 is positive and two are real and negative .
we also saw that @xmath959 and therefore , @xmath960 if @xmath186 since the trace of @xmath956 is also negative , if @xmath18 then @xmath956 has either one or three eigenvalues with negative real part , and for sufficiently small @xmath961 it has three , all of which are real . in fact , since @xmath962 , @xmath963 , and @xmath964 if @xmath965 @xmath956 has exactly one real positive eigenvalue for every @xmath180 in the positive quadrant @xmath966 for each @xmath61 as @xmath20 increases the other roots of @xmath967 remain in the left hand plane unless , for some @xmath5 two of them are pure imaginary .
consideration of the characteristic polynomial in this case ( one negative , one positive , and two pure imaginary roots ) shows that the coefficients of @xmath968 and @xmath969 have the same sign .
this is not the case with @xmath970 because the coefficient of @xmath969 is positive and the coefficient of @xmath968 is negative .
hence , as asserted in lemma [ lem2b ] , the unstable manifold @xmath971 of ( [ 1 ] ) at @xmath170 is one dimensional .
further , because @xmath959 it follows from ( [ 16 ] ) that if @xmath972 is the unit eigenvector of @xmath956 with @xmath973 then @xmath974 and @xmath975 also ( [ 16 ] ) implies that if @xmath660 then @xmath976 and if @xmath18 then @xmath977 the claimed behavior for large negative @xmath64 of solutions on @xmath971 follows .
the continuity of @xmath971 for @xmath129 follows from theorem 6.1 in chapter 6 in the text of hartman , @xcite . in this case
the unstable manifold @xmath978 all we need , is the set of all solutions @xmath979 which tend to @xmath170 at an exponential rate as @xmath980 to obtain the desired continuity of @xmath172 with respect to @xmath20 and @xmath55 apply hartman s theorem to ( [ 2 ] ) augmented with equations @xmath981 and @xmath982 this is the closest we come to center manifolds in our approach . ]
the final assertion of the lemma , that @xmath983 if @xmath17 and @xmath18 follows by writing the characteristic polynomial of @xmath956 in the form @xmath984 we see that since @xmath985 , @xmath986 hence @xmath987 completing the proof of lemma [ lem2b ] . | in 1992 g. b. ermentrout and j. b. mcleod published a landmark study of travelling wavefronts for a differential - integral equation model of a neural network . since then a number of authors have extended the model by adding an additional equation for a recovery variable , thus allowing the possibility of travelling pulse type solutions . in a recent paper g. faye
gave perhaps the first rigorous proof of the existence ( and stability ) of a travelling pulse solution for a model of this type , treating a simplified version of equations originally developed by kilpatrick and bressloff .
the excitatory weight function @xmath0 used in this work allowed the system to be reduced to a set of four coupled odes , and a specific firing rate function @xmath1 , with parameters , was considered .
the method of geometric singular perturbation was employed , together with blow - ups . in this paper
we extend faye s results on existence by dropping one of his key hypotheses , proving the existence of pulses at at least two different speeds , and in a sense , allowing a wider range of the small parameter in the problem .
the proofs are classical , and self - contained aside from standard ode material . |
among numerous different nuclear many - body models , the nuclear density functional theory ( dft ) @xcite is the only one , which can describe nuclear properties microscopically throughout the entire nuclear landscape @xcite .
the cornerstone of the nuclear dft is the energy density functional ( edf ) , which incorporates nucleonic interactions and many - body correlations into a functional constructed from one - body densities and currents .
the skyrme edf , for its part , relies on local nuclear densities and currents , together with a set of coupling constants as model parameters . due to the lack of suitable _
ab - initio _ methods to compute these coupling constants , they must be determined through adjustment to experimental data , such as nuclear binding energies and radii . during the last couple of decades ,
numerous skyrme parameterizations have been obtained from various adjustment schemes , see e.g. the list in @xcite .
the standard skyrme edf has proven to be quite successful , but its limitations have also become apparent . for the sake of better accuracy , more reliable predictive power , and for a spectroscopic - quality level , one has to move beyond the the standard skyrme edf @xcite .
nevertheless , by studying the performance and predictive power of the present edfs , valuable information can be obtained which can be used and applied in the work towards forthcoming novel edfs @xcite .
as with every model parameters optimization procedure , one of the main challenges is to find the best set of input observables , in order to constrain the parameter space of the model .
the predictive power of the edf highly depends on the input data .
therefore , a comprehensive analysis of the impact of input observables on the parameter space and on the model predictions provides valuable information . naturally , the nuclear bulk properties are crucial for general constraining and the data especially relating to odd - mass nuclei is important for spectroscopical properties @xcite .
binding energies , surface thickness , charge radii , single particle energies and energies of giant resonances are essential properties of nuclei , and used in various edf optimization schemes . all model predictions contain several sources of uncertainties .
roughly speaking , these can be divided into two main categories , the systematic model uncertainties and the statistical model uncertainties .
the systematic model uncertainty stems from sources like the model deficiency and input data bias .
the statistical uncertainty results from the model parameter optimization process . despite the importance of uncertainty analysis ,
error estimate is a rather novel topic in low - energy nuclear physics @xcite . during the last few years
, efforts have been made to improve this situation in the edf calculations @xcite , as well as in the domain of _ ab - initio _
calculations @xcite .
various statistical tools have been applied from traditional methods to more modern ones ( e.g. the bayesian framework @xcite ) .
apart from the fact that uncertainty quantification is an important topic in itself , with the help of statistical analysis , information about shortcomings of theoretical models and optimization procedures is also obtained . in this work ,
we present the quantitative results for statistical uncertainty propagation for three existing unedf skyrme edf models : the unedf0 @xcite , the unedf1 @xcite , and the unedf2 @xcite .
in particular , we quantify contributions from the model parameters to the total error budget of binding energy in isotopic and isotonic chains of nuclei . by analyzing the obtained information we may recognize potential frailties of these models . in the present study ,
two - neutron separation energies are also considered and the related uncertainties are worked out . in addition to
even - even nuclei , uncertainties related to odd - even nuclei are studied .
this paper is organized as follows . in sec .
[ sec : theorframe ] we briefly review the theoretical framework related to the topic : namely , the skyrme energy density functional and the error propagation . in sec . [ sec : results ] we present our results and , finally , conclusions and future perspectives are given in sec .
[ sec : concl ] .
the unedf models are based on the skyrme energy density .
the ground state of a nucleus is determined in the framework of the hartree - fock - bogoliubov ( hfb ) theory @xcite .
the three parameterizations considered in this work , unedf0 , unedf1 and unedf2 , were adjusted on a experimental data consisting of binding energies for deformed and spherical nuclei , odd - even mass differences , and charge radii .
in addition , latter parameterizations include data on fission isomer excitation energies and single - particle energies .
the skyrme edf has a form of local energy density functional , stemming from skyrme energy density .
it can be written as @xmath0 \,.\end{aligned}\ ] ] where the energy density @xmath1 is a time - even , scalar , isoscalar and real function of local densities and their derivatives . in the equation ( [ eq : skyrmee ] ) , the skyrme energy density has been split into kinetic term @xmath2 , isoscalar ( @xmath3 ) and isovector ( @xmath4 ) particle - hole skyrme energy densities @xmath5 , pairing energy density @xmath6 and coulomb term @xmath7 .
the time - even part of the isoscalar and isovector particle - hole skyrme energy densities is given by @xmath8 in the equation ( [ eq : phterm ] ) , @xmath9 is the isoscalar or isovector kinetic density and @xmath10 is the spin - current density tensor .
definitions of these densities can be found in reference @xcite . with the unedf models ,
only the time - even part of the total energy density was defined and time - odd part of the energy density was set to zero .
the energy density is always time - even , also the part called time - odd - the time - odd energy density means that this part of the energy density is built by using time - odd densities .
the pairing energy density @xmath6 used here has the form of @xmath11 \widetilde{\rho}_q^2 ( \mathbf{r } ) \,,\ ] ] where @xmath12 ( @xmath13 ) are the pairing strength parameters for neutrons and protons , respectively , and @xmath14 was set to the equilibrium density @xmath15@xmath16 .
all the coefficients @xmath17 and @xmath12 are real constants , except the coefficients @xmath18 which depend on the isoscalar density so that @xmath19 altogether , there are 13 independent constants from skyrme energy density and two constants from pairing , namely @xmath20 seven of these parameters for @xmath3 and @xmath4 , @xmath21 , @xmath22 , @xmath23 and @xmath24 , can be written with the help of the infinite nuclear matter parameters @xcite .
all in all , the model depends on 15 independent parameters , namely @xmath25 which were optimized in the previous works , in references @xcite . here
@xmath26 is the saturation density , @xmath27 represents the total energy per nucleon at equilibrium , @xmath28 is the nuclear matter incompressibility , @xmath29 is the symmetry energy coefficient , @xmath30 describes the slope of the symmetry energy , @xmath31 is the isoscalar effective mass and the last one , @xmath32 , is the isovector effective mass . in the present work ,
we have compared calculated theoretical binding energies to the experimental ones from @xcite . in order to obtain experimental nuclear binding energies ,
the experimental atomic masses were corrected by taking into account the electron binding energies , approximated as @xmath33 lccccccccccc edf & @xmath34 & @xmath35 & @xmath28 & @xmath29 & @xmath30 & @xmath36 & @xmath37 & @xmath38 & @xmath39 & @xmath40 & @xmath41 + unedf0 & x & x & & x & x & & & x & x & x &
+ unedf1 & x & & & x & x & x & & x & x & x & + unedf2 & x & & x & x & & x & & x & x & x & x
+ [ tab : unedfpara ] the unedf parameterizations were accompanied by sensitivity analysis , providing covariance matrix of the model parameters .
this allows to calculate the standard deviation of any observable predicted by the model . in the present work we consider the statistical errors on binding energies and on two neutron separation energies .
the statistical standard deviation @xmath42 of an observed variable @xmath43 is given by @xmath44 \left[\frac{\partial y}{\partial x_j}\right ] \,,\ ] ] where @xmath45 is the covariance matrix element between the model parameters @xmath46 and @xmath47 , and @xmath48 is the number of model parameters .
the covariance matrix @xmath45 is related to the corresponding correlation matrix @xmath49 as @xmath50 where @xmath51 and @xmath52 are the standard deviations of parameters @xmath46 and @xmath47 , respectively .
the correlation matrices of the unedf models and the standard deviations of the model parameters are given in [ app : appendixa ] .
some of the unedf model parameters have been excluded from this sensitivity analysis .
table [ tab : unedfpara ] lists those unedf parameters which were included in the sensitivity analysis @xcite ( marked with @xmath53 ) , those which were fixed during the whole optimization procedure ( - ) and those which ended up at their boundaries during the optimization ( empty space ) .
sensitivity analysis can not be performed for fixed parameters or those which drifted onto the boundary during the optimization .
however , those parameters which were included in sensitivity analysis have a visible contribution to the statistical error of an observable and their contribution to the total error budget can be calculated from equation ( [ eq : sigma2 ] ) . in the present work we used the code hfbtho @xcite to calculate observables and their statistical errors .
the program solves the hartree - fock - bogoliubov equations for skyrme edfs in the axially symmetric harmonic oscillator basis .
time - reversal and parity symmetries were assumed . because particle number is not a good quantum number in hfb theory
, we used the lipkin - nogami method to restore it approximately .
the hfb equations were solved in a basis consisting of 20 oscillator shells and the convergence criteria was set to @xmath54 .
this means that the desired accuracy has been reached when the norm of the hfb matrix difference between two iterations is less than @xmath54 .
both of the coulomb terms , @xmath55 and @xmath56 were used , but the exchange term was calculated by using slater approximation . a rough position of the energy minimum , with respect of quadrupole deformation , was first located from a constrained hfb calculation .
then , an unconstrained hfb calculation was performe in order to converge to the precise position of the energy minimum . in order to obtain standard errors
one has to calculate the derivatives of an observable @xmath57 with respect to the model parameters @xmath46 , as in equation ( [ eq : sigma2 ] ) . in the present work
these derivatives were approximated by a finite differences .
that is @xmath58 where value of @xmath59th parameter has been shifted by amount of @xmath60 from the model base values .
the rounded values of unedf parameters and corresponding shifts @xmath60 have been listed on table [ tab : parsteps ] .
we tested that the computed statistical errors remained essentially the same when shift parameters @xmath60 were slightly varied .
lastly , we recall that standard deviation does not measure the total uncertainty of a model . another main ingredient , namely the systematic error , is much more challenging to assess . it can be addressed e.g. by studying a dispersion of different predictions given by various skyrme edf models @xcite .
however , due to lack of exact reference model , precise systematic errors are not within one s reach .
the differences between the theoretical and experimental binding energies for even - even gadolinium and dysprosium isotopes are shown in figure [ fig : gddy_residuals ] .
the error bars represent the calculated theoretical standard deviations , and they are also given as a function of the mass number in the graphs on the right .
the uncertainties are given for all the calculated binding energies , including also those nuclei for which the experimental binding energy is not currently known .
as it can be seen , unedf0 gives more consistent results with the measured experimental energies for lighter isotopes , whereas unedf1 and unedf2 seem to improve their accuracy considerably for heavier isotopes .
most of the theoretical results do not overlap with the experimental values , even when including error bars .
two interesting points can be seen in the graphs on the left : @xmath61gd and @xmath62dy . both of these nuclei have neutron number of @xmath63 , that is , one of the magic numbers . here , the theoretical predictions for the binding energies given by unedf1 and unedf2 are comparatively farther away from the experimental results .
however , there is no visible increase in the standard deviation of the binding energy of these two nuclei .
this suggests that the increased residual is due to underlying model deficiency , and not due to the parameter optimization procedure .
the calculated standard deviations of binding energies are found to be around 0.53.0mev , 0.41.7mev , and 0.31.5mev for unedf0 , unedf1 and unedf2 , respectively .
even though the standard deviations have a magnitude of one thousandth of the total binding energy , the theoretical uncertainties are still far larger compared to experimental precision , which can be of the order of few kev s @xcite .
however , uncertainties of the unedf models have decreased after every model : the obtained standard deviation for unedf0 is larger compared to two later parameterizations .
the behavior of uncertainty is relatively smooth and the uncertainty of binding energy grows quickly when going towards neutron rich nuclei .
in addition , one can also see that uncertainties grow when going towards the other extreme , namely proton rich nuclei .
this is an indication that isovector part of the edfs is not as well constrained as the isoscalar part .
but for two neutron separation energy @xmath64.,scaledwidth=80.0% ] the residuals of the two neutrons separation energy , @xmath65 , are shown in figure [ fig : gddy_sep_res ] .
two neutron separation energies were calculated for even - even dy and gd isotopes .
similarly to the previous figure , theoretical errors are marked as error bars in the graphs on the left hand side panels , and also given as a function of mass number on the right hand side panels .
the theoretical statistical error is calculated similarly , through finite differences of @xmath65 values to compute the derivatives . for neutron
rich nuclei , all three parameterizations give essentially the same result for @xmath65 , within the error bars .
otherwise , the latest unedf2 parameterization seems to differ most from the experimental results when compared to previous two parameterizations .
since @xmath64 is defined as a difference of two binding energies , the partial derivatives are also calculated from energy difference between two nucleus . as a consequence
, some of the parameter uncertainties can cancel each other . in particularly , the uncertainty coming from a relatively less constrained isovector part of the edf is now partly canceled , resulting to more moderate uncertainty in the neutron rich region compared to the uncertainty of binding energy .
similar observation was also done at @xcite . as a function of mass number @xmath66 .
right panel : calculated statistical error of @xmath67 when using either lowest or secondary energy minimum of @xmath68gd.,scaledwidth=70.0% ] one common feature for all unedf models is an existence of a few high peaks in the statistical error of @xmath64 .
these peaks are located mainly around the same neutron numbers for gd and dy isotopes .
for instance , the two highest peaks given by unedf2 are located at the nuclei @xmath68gd and @xmath69dy .
the explanation for all of the high peaks can be found in a sudden change in deformation .
figure [ fig : betas_minimum ] , in panel a , shows how deformation parameter @xmath70 varies with mass number @xmath66 for gd . when comparing this to that of @xmath67 , shown in figure [ fig : gddy_sep_res ]
, one can notice similarity between uncertainty peaks and large change in the deformation .
if there is a significant difference in deformation between two consecutive even - even nuclei , this results to a larger statistical error of two neutron separation energy . .
, scaledwidth=45.0% ] the relationship between @xmath67 and a sudden large change of @xmath70 can be tested by by looking at the secondary local minimum of the deformation energy landscape . for the calculated gd and dy nuclei , there usually exists two energy minima , the oblate one and the prolate one , as shown in figure [ fig : dy_defe_behav ] . by picking always the lowest minimum results to large statistical error of @xmath64 when deformation has a large change between two even nucleus .
however , if one uses the secondary minimum , in which case the two nuclei appearing in the expression of @xmath64 have similar deformation compared to the each other , one obtains substantially smaller @xmath67 . in figure
[ fig : betas_minimum ] , panel b , the black line describes the same peak at @xmath71 given by unedf2 as in figure [ fig : gddy_sep_res ] , calculated with the lowest energy minima .
the second ( red ) line corresponds to case where second minimum of @xmath68gd was used , resulting a much smaller @xmath67 for @xmath68gd .
indeed , a large difference in deformation of involved nuclei seems to give large uncertainty on two neutron separation energy .
a possible explanation is considerably different shell structure between these two nuclei due to deformation .
the largest impact on the extremely high peaks in @xmath67 given by unedf2 is due to the parameter @xmath72 , whereas for unedf1 the main contributors are @xmath31 , both @xmath73 together with @xmath74 and @xmath75 parameters . , but for the isotonic chain with the neutron number @xmath76.,scaledwidth=80.0% ] the binding energy residuals between theory and experiment for isotonic chain of @xmath76 can be found in figure [ fig : isotonic_residuals ] .
only even - even nuclei are studied .
the unedf1 and unedf2 parameterizations give rather similar results , but the binding energy behavior of unedf0 parameterization is notably different .
compared to the unedf0 optimization procedure , in the optimization of unedf1 the same set of 12 edf parameters were optimized but seven additional data points were included in the database and the center of mass correction was neglected @xcite . the other important remark is the fact that even though the trend of unedf1 and unedf2 models is incorrect the calculated binding energies for proton rich nuclei are getting further far away from the experimental ones when mass number increases the uncertainties do not nevertheless become larger .
but for even - even and even - odd sn isotopes.,scaledwidth=80.0% ] we have also considered the statistical error of odd - even nuclei . the binding energy residual for sn isotopic chain
was computed by using all three unedf models .
the results are shown in figure [ fig : sn_residual ] .
the odd - even nuclei were calculated by using the quasiparticle blocking procedure with the equal filling approximation @xcite .
the same blocking configuration , which corresponded the lowest energy with unshifted parameterization , was used for calculation of all partial derivatives .
the same set of results for even sn isotopes , with unedf0 , was calculated in @xcite .
the results show that the binding energy residuals of even - even nuclei are relatively greater compared to those of odd nuclei . on this account
, the binding energy residuals stagger between the odd and even nuclei .
nevertheless , there are no visible odd - even effects in the standard deviations of binding energy .
this can be explained by the lack of time - odd part in the used edfs .
one of the goals of present work is to study contributions of the model parameters to the total error budget of binding energy . in practice ,
the components of the total error are calculated by using equation ( [ eq : sigma2 ] ) so that contributions from each 2nd summation index are given term by term whereas the 1st summation index is summed over all parameters .
thus the total squared standard deviation is a sum of all these components .
some of these components have a negative sign , due to a negative partial derivative or a negative covariance matrix element .
the total squared standard deviation is always , nevertheless , a positive number .
one should also bear in mind that the contribution of a parameter to the standard deviation is visible only if this parameter was included in the sensitivity analysis , as mentioned before .
for even - even dy isotopes with unedf0 , unedf1 and unedf2 edfs.,scaledwidth=75.0% ] but for the isotonic chain @xmath76.,scaledwidth=75.0% ] the results for the model parameters contributions are shown in figure [ fig : dycomponents ] for the binding energies of even dy isotopes and in figure [ fig : isotonic_components ] for the binding energies of even @xmath76 isotones .
the error budget for unedf0 binding energy is mainly composed of only a few contributing parameters . for the nuclei close to the valley of stability ,
two dominant sources of uncertainty are the parameters @xmath27 and @xmath77 , whereas in the neutron rich dy isotopes parameters @xmath30 and @xmath29 dominate .
in other words , model parameters related to symmetry energy become more important with neutron rich nuclei .
it was found earlier that @xmath30 has also a strong impact on the statistical error of neutron root - mean - square radii and neutron skin thickness @xcite .
the most dominant sources of uncertainty are the same for isotopic and isotonic chains . contrary to unedf0 ,
the error budget of latter two parameterizations is more split among various model parameters . generally speaking
, @xmath78 , both @xmath79 parameters , @xmath80 , and @xmath81 have a significant impact on the total error budget with unedf2 .
it should be noticed that the correlation between @xmath77 and @xmath81 is strong : in principle , if one can reduce the uncertainty on @xmath77 , it should also reduce the uncertainty of @xmath81 . the isovector parameters , for their part , are difficult to constrain , but they impact on the stability of the functional : for instance @xmath82 is the coupling constant of the gradient term , which has been found to trigger scalar - isovector instabilities @xcite .
in the present work we have calculated statistical errors of the unedf models for binding energy and for two neutron separation energy by using the knowledge of the covariance matrix of the model parameters . the standard deviation has been interpreted as a statistical error .
we have also quantified the contributions of each model parameters to the total error budget and checked if there are any visible odd - even effects in the uncertainties of the unedf models .
we found out that the standard deviation grows quickly when going away from the valley of stability towards proton rich or neutron rich nuclei .
that is , the predictive power of unedf models becomes weaker when going further away from known nuclei to the region where the models are used for extrapolating .
for the sn isotopic chain , even though there exists odd - even staggering in the residuals of binding energies , no visible odd - even effect is seen in the related errors .
the error budget of the unedf parameterizations becomes more evenly split among various model parameters with unedf1 and unedf2 parameterizations .
the most dominant contributors to the error budget of neutron rich nuclei with unedf0 were @xmath83 and @xmath78 , that is to say , coefficients related to the symmetry energy . in the case of unedf1 , @xmath83 and @xmath78 still have a significant contribution and , in addition , the role of @xmath84 and @xmath81 becomes important .
with unedf2 , the role of other model parameters becomes even more prominent .
a comparison of the standard deviations to the residuals seems to point out that the underlying theoretical model is missing some important physics .
one clear indication of this is the increase of the binding energy residuals close to the semi - magic nuclei .
however , calculated standard deviation does not usually reflect this kind of behavior .
similar observation was also done in @xcite .
also , the systematically incorrect trend of unedf1 and unedf2 with the binding energy residual in the @xmath76 isotonic chain does not appear in theoretical uncertainties , as shown in figure [ fig : isotonic_residuals ] .
the trend in the unedf statistical errors is such that the calculated standard deviation decreases as more data is included in the optimization procedure . from a statistical point of view , the larger uncertainty of the isovector parameters reflects to the increasing uncertainty when going towards both isospin extremes .
unfortunately , new data may not always help to reduce uncertainties .
a sophisticated parameter optimization within bayesian framework showed that new data points on nuclear binding energies at neutron rich region were not able provide a better constraint for the unedf1 model parameters @xcite . on the other hand , data on the neutron skin thickness
could potentially help to reduce uncertainties related on isovector model parameters @xcite . as was concluded for the unedf2 parameterization ,
the limits of standard skyrme - edfs have been reached and novel approaches are called for @xcite .
information about the shortcomings and uncertainties of present models will provide valuable input for development of novel edf models .
a comprehensive sensitivity analysis of a novel edf parameterization is essential when addressing its predictive power .
t.h . thanks karim bennaceur and nicolas schunck for great advice and remarks .
this work was supported by the academy of finland under the centre of excellence programme 2012 - 2017 ( nuclear and accelerator based physics programme at jyfl ) and under the fidipro programme .
was also supported by a grant ( 55151927 ) of the finnish cultural foundation , north karelia regional fund .
[ tab : unedf0corr ] [ tab : unedf1corr ] [ tab : unedf2corr ]
[ tab : parsteps ] | the parameters of the nuclear energy density have to be adjusted to experimental data . as a result
they carry certain uncertainty which then propagates to calculated values of observables . in the present work
we quantify the statistical uncertainties on binding energies for three unedf skyrme energy density functionals by taking advantage of the knowledge of the model parameter uncertainties .
we find that the uncertainty of unedf models increases rapidly when going towards proton or neutron rich nuclei .
we also investigate the impact of each model parameter on the total error budget .
_ keywords _ : skyrme energy density functional , uncertainty quantification , error propagation |
globular clusters ( gcs ) are among the oldest surviving stellar objects in the universe .
they provide collections of population ii stars with homogeneous abundances and histories , and unique stellar dynamical conditions .
the milky way s globular cluster system ( gcs ) is the prototypical one , and its study has contributed much to our knowledge of stellar evolution and galactic structure .
it is important to make sure that conclusions drawn from this study are not biased either because the milky way s gcs is somehow unusual or because our location in the galaxy prevents us from fully characterizing its properties .
globular clusters in local group galaxies are particularly valuable for comparison with milky way globular clusters .
m31 has the local group s largest globular cluster population , so it is a natural starting place for studies of extragalactic globular clusters .
the first m31 globular cluster catalog was published by @xcite , followed by @xcite , @xcite , @xcite , and @xcite .
the most comprehensive recent catalog is that of @xcite ; recent works by @xcite and @xcite cover only portions of m31 .
all of these catalogs contain objects which are not m31 globular clusters : for example , table 2 of @xcite lists 199 cluster candidates later shown to be non - clusters .
the existing catalogs are also likely to be missing clusters due to magnitude , spatial coverage , and/or resolution limits .
@xcite defined several samples of m31 globular clusters , including a ` confirmed ' sample ( 199 objects ) , an ` adopted best ' sample ( 298 objects ) , and an ` extended ' sample ( 356 objects ) . in @xcite
we compiled a list of clusters and plausible candidates containing 435 objects .
quantifying the extent of incompleteness and contamination in m31 globular cluster catalogs is extremely important for the interpretation of globular cluster system properties .
for example , the spatial distribution of known clusters is flatter , and their globular cluster luminosity function ( gclf ) brighter , near the nucleus @xcite is this because the clusters there are truly fewer and brighter , or because existing surveys have not detected the entire cluster population ?
even the census of milky way clusters is likely to be incomplete : @xcite estimates that 1030 milky way globulars may be hidden behind the galactic bulge and therefore missing from current catalogs , which list about 150 objects @xcite .
two such clusters were found by @xcite .
it is not unreasonable to suspect that the m31 cluster catalogs could be incomplete by at least a similar fraction .
ground - based high - resolution imaging and spectroscopy have been used to distinguish m31 globular clusters from interlopers such as foreground stars , background galaxies , and other objects belonging to m31 ( e.g. , h ii regions and open clusters ) .
the bright ( @xmath3 ) portion of m31 globular cluster catalogs has been fairly thoroughly examined using one or both of these methods . @xcite and @xcite used short - exposure ccd images taken in excellent seeing to determine if cluster candidates in the m31 halo were resolved into stars ; they found that majority of the halo cluster candidates were background galaxies , not clusters .
radial velocities from optical spectroscopy have also been used by several groups ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) to eliminate background galaxies and foreground stars from cluster candidate lists . neither method is infallible , however : compact clusters may be mistaken for background galaxies if not resolved into stars , or for stars if they have a small radial velocity .
km s@xmath4 .
the velocity range of m31 globular clusters is about + 70 to @xmath5 km s@xmath4 , and the galactic models of @xcite predict that the radial velocities of milky way stars with similar colors and magnitudes to m31 globulars are in the range @xmath6 to @xmath7 km s@xmath4 . ] hst imaging , with its superior spatial resolution , is a useful tool for removing some of the ambiguities inherent in the ground - based studies . at the distance modulus of m31 given by @xcite and @xcite ( @xmath8 , @xmath9 kpc ) , the angular resolution of hst s wfpc2 camera
is equivalent to a spatial resolution of 0.38 pc . is very helpful for the identification of globular clusters in m31 .
the differences between globular clusters and contaminating objects are much more obvious than with ground - based imaging .
m31 has been a popular target for hst : as of 1 december 2000 , the hubble data archive contained almost 1100 wfpc2 images within 150 of the center of the galaxy .
as of the same date , about two dozen m31 globular clusters had been specifically targeted for observation with hst , and the images of these clusters comprise about 20% of all the m31 images .
the goal of most targeted hst observations of m31 globular clusters has been the production of color - magnitude diagrams for the clusters and surrounding stellar populations .
hst programs which specifically targeted m31 globular clusters include gos 5112 , 5420 , 5464 , 5907 , 6477 , 6671 , 7826 , 8296 , and 8664 .
our study uses the publicly - available archival data from these programs and many others . in mid-2000
, we began a project to search for globular clusters in archival hst images for the purpose of quantifying the incompleteness of existing cluster catalogs ; preliminary results were described in @xcite .
the present paper report the results of our efforts to find globular and other star clusters in archival hst / wfpc2 images and their implications for catalog completeness and contamination .
a companion paper @xcite presents measurements of the structural parameters of the clusters and their implications .
we do not attempt to construct cmds for the clusters , since this work is already being carried out by other groups .
we searched the hst archive for all wpfc2 observations with the following properties : * center of field within @xmath10 of the center of m31 * broadband filter with central wavelength 300 nm or longer * total exposure time longer than 100 s. these parameters were chosen to ensure that we would have a reasonable chance of detecting globular clusters if they were in the image fields .
many images met the requirements , but since most positions had more than one observation per filter and observations in more than one filter , the images comprised only 157 separate fields .
some of these fields were known to contain m31 globular clusters ; we retained these fields in our search as a check on our ability to identify clusters .
we searched the images in only one filter per field .
if more than one filter was available , we chose filters in the following order : f555w , f814w , f606w , f450w , f439w , f336w , f300w .
( this ordering reflects the distribution filters used for the images combined with our desire to examine as many fields as possible in the same filter . ) information on the fields searched , including dataset name , location , filter , and exposure time is given in table [ tbl - flds ] .
the images searched are mostly in f555w and f814w , although there is at least one image in each of the filters listed above .
the exposure times ranged from 100 to 8400 s. figure [ field - pic ] shows the location of all fields on the sky .
we retrieved the images from either the space telescope science institute or the canadian astronomy data centre . in both cases
the images were pipeline - processed from the raw data at the time of retrieval with the best available calibration images . from stsci we retrieved individual hst images ; when multiple images existed for a single field ( e.g. , in the case of ` cosmic ray - split ' images ) , we combined the images using the iraf task crrej . from cadc we retrieved ` wfpc2 associations ' ; these are coadded images produced by the cadc pipeline , which combines multiple cr - split images with the gcombine task .
we found that gcombine did not adequately remove cosmic - ray hits when only two images were co - added , so in that case we retrieved the individual images and combined them with crrej .
there were no obvious differences in the images produced using the two methods we used both since we became aware of availability of association images from the cadc part way into the project .
once the images were processed , we began the search for star clusters .
the first step was carried out ` blind ' , that is , without any knowledge of the positions of cataloged clusters .
working independently , each of us visually examined each image .
pb used sextractor @xcite to automatically identify objects with large areas and/or extended profiles , then visually checked the sextractor candidates ( many of these were actually bright stars ) and searched for additional candidates .
jh used only visual examination of the images .
bright m31 globular clusters can be visually distinguished from stars and elliptical galaxies because they appear more ` ragged ' at the edges ( being resolved into stars ) and do not have the diffraction spikes seen around bright stars .
faint or small clusters are distinguished more by their image shapes larger than the point spread function , and less smooth than a galaxy than by resolution into stars .
clusters are distinguished from h ii regions or nebulae by the fact that the latter are much more diffuse are show few individual stars .
the visual classification is not completely objective , but it was the best method we could contrive for dealing with the large number of images to be examined and the large number of potential contaminating objects .
our confidence in the visual classification was bolstered by the fact that we only disagreed on the classification of about 10% of the objects .
we re - examined these together to make a final classification .
we combined our two lists of cluster candidates to make a final list .
although we were interested primarily in globular clusters , we recorded positions of possible open clusters as well .
following previous authors ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , we classified our globulars in classes a through d , where a is ` very likely to be a globular cluster ' and d is ` likely not a globular cluster ' .
we refer to objects in classes a and b as good candidates , and objects in classes c and d as marginal cluster candidates . after generating our final list of cluster candidates
, we checked the image positions against existing catalogs of m31 globular clusters .
this allowed us to gauge our detection efficiency and locate objects we would otherwise have missed .
the globular cluster list used was a ` master list ' of globular clusters and candidates , produced by combining the lists of @xcite , @xcite , @xcite , @xcite , and @xcite ; it includes all the objects listed in the @xcite catalog , plus additional low - probability candidates and non - clusters .
we consider the low and high - probability globular clusters separately .
` high probability ' are clusters a or b class clusters from @xcite , @xcite , or @xcite ; all other objects are ` low probability ' .
@xcite showed that the @xcite classification correlates well with the probability that a candidate will be subsequently shown to be a cluster .
75 high - probability clusters from our master list were located in the hst fields ; we detected 71 , and some images of previously - cataloged clusters are shown in figure [ fig - gc ] .
three of the four non - detections ( 138000 , 166000 , and 133191 ) appeared to be stars or blends of stars rather than globular clusters ; the fourth object was dao040 and we did not detect any object at the coordinates given by @xcite . of the 72 low - probability ( class c or d ) cluster candidates in our hst fields , we found 7 good candidates ( 000d038 , 000m91 , 020d089 , 097d000 , 132 - 000 , 264nb19 , and nb39 ) , 4 marginal candidates ( 000m045 , 257000 , nb41 and nb86 ) , and 45 objects which did not appear to be clusters .
we did not detect the other 14 objects in our visual search . on
re - examining the positions of these objects , we found that none were good or even marginal cluster candidates .
several were clearly stars , and the others were blends of stars or blank fields .
table [ tbl - noncl ] gives a list of the non - clusters and their classifications .
our visual search of the hst fields produced 32 objects not included in any cluster catalog .
10 of these were good candidates , although only about half are as obviously clusters as most of the brighter objects .
the good candidates images are shown in figure [ fig - newclust ] .
the nature of the remaining 22 objects is unclear .
they are clearly not stars ; all are at least marginally resolved ( fwhm @xmath11 )
. however , most are quite faint , and they are not obviously resolved into stars as is the case for most of the globular clusters .
they may be blended stars in m31 , compact background galaxies , or compact star clusters .
we show images of these low - quality objects in figure [ fig - myst ] .
table [ tbl - newgc ] gives the location and quality of all the new cluster candidates .
the dividing line between open and globular clusters is somewhat blurred , even in the milky way . in their compilation of data on milky way
globular clusters , @xcite note that there are several globulars ( bh 176 , uks 2 ) which could instead be open clusters . in our search , we noted several concentrated objects which could be m31 open clusters .
their nature is uncertain : they could also be low - concentration globulars , or just chance superpositions of stars .
their images are shown in figure [ fig - oc ] .
we checked the cluster coordinates against those given in @xcite s ( @xcite ) list of m31 open clusters .
the coordinates in that catalog have rather low precision ( 20 in both right ascension and declination ) , so we searched for coordinate matches within an error circle of radius 30 .
we found 5 matches and attempted to confirm these by comparing the finding charts in @xcite to our images .
the results were inconclusive : either the objects were not clearly identified on the charts , or they were located too close to the edge of the hst image to make a positive identification .
we note the possible matches in table [ tbl - newgc ] . to see if any of our newly - proposed globular and open cluster candidates had been previously cataloged as background galaxies , we checked their positions against those of galaxies listed in ned .
none of the new clusters matched the position of any galaxy listed in ned , although one is listed as a possible h ii region @xcite and two others may contain radio and x ray sources @xcite .
the matches are also noted in table [ tbl - newgc ] .
however , the matches are uncertain since positional uncertainties for the other surveys are large .
figure [ posplot ] shows the positions on the sky of all the m31 clusters , both previously - known and newly - discovered .
the ` open cluster ' near ngc 205 is well outside the disk and is probably not a real cluster .
after the m31 clusters had been identified on the ` search ' images , we retrieved images of their fields in other available filters to extract the most photometric information from the hst archive .
all but 18 clusters had been imaged by wfpc2 in more than one filter .
we combined images for cosmic - ray rejection in the same manner used for the search images .
additional processing steps included removing cosmic rays interactively using the iraf task imedit ( this was especially important for non - cosmic - ray - split images ) and correcting for warm pixels using the iraf task cosmicrays . while the stsdas task warmpix is the preferred method of dealing with warm pixels , it is slow and requires correction of individual images before they are combined for cosmic ray rejection . since we had hundreds of individual images to deal with , we chose the more expedient method of treating the warm pixels as if they were as cosmic ray hits on ground - based images .
nearby bright stars and ccd flaws were masked out of the images to prevent contamination of the photometry .
a few images were not useful for photometry at all : the globular clusters were either very faint ( mostly in the f300w and f336w filters ) , too close to an image edge , or saturated .
photometry of extragalactic globular clusters is unfortunately not as simple as photometry of isolated stars or galaxies .
there are two key steps in integrated photometry of m31 clusters : measuring the background light , and identifying an appropriate aperture size .
the background light consists of two components : unresolved light from the sky and m31 , and light from resolved stars in m31 ( the latter are a lesser problem in ground - based photometry of m31 gcs since many fewer m31 stars are resolved ) .
standard background estimators are usually designed to determine the sky background level by rejecting the stars in the background annulus .
since we expect there to be stars overlapping our clusters as well , we estimated the background value for each image as the mean ( rather than the more commonly used median or mode ) of the pixel values around the image edge , and subtracted it from the image before doing photometry .
determining the ` correct ' aperture size to be used for integrated photometry is non - trivial since the clusters are not all the same size .
we estimated the total flux for each object by measuring aperture magnitudes in concentric apertures spaced 0.15 apart , plotting magnitude growth curves , and noting where the flux stopped increasing . using these measurements of the total flux of each cluster , we determined the half - light radius is that which contains half of the integrated cluster light .
it should not be confused with the radius at which the surface brightness drops to half of its central value , called variously the core radius @xmath12 , the half - intensity radius , or the half - width at half maximum ( hwhm ) . ] by interpolating the aperture magnitude curves .
we calibrated the instrumental magnitudes from the wfpc2 system to the standard system by iteratively solving the equations given in @xcite , using the charge - transfer - efficiency corrections given by @xcite .
the iterative solution of the calibration equations requires instrumental magnitudes in at least two filters ; for objects with only one instrumental magnitude , we fixed the ` standard color ' as either the measured ground - based color from @xcite , or if that was unavailable , the average m31 gc color .
the results for integrated magnitudes and half - light radii are given in table [ tbl - phot ] . in figure
[ fig - photcomp ] , we compare the new hst photometry to the ground - based measurements compiled in @xcite .
the agreement is gratifying : the median offset in @xmath13 is @xmath14 magnitudes , and in @xmath15 is @xmath16 mag .
most of the large offsets are for objects near the edge of a wfpc2 chip , or whose previous photometry was estimated from photographic plates .
to estimate the completeness of globular cluster catalogs in m31 , we first need to understand our own detection efficiency .
we estimated this by inserting artificial globular clusters into the inner images , for which the distance from the center of m31 @xmath17 was less than 30 .
the artificial clusters were actually images of the brightest real globular clusters we detected . to insert the artificial clusters , we scaled the image fluxes 04 magnitudes fainter , adjusted for the exposure time of the inserted image , rotated the cluster to a random position angle , and applied a random axial ratio from 0.85 to 1.0 .
this may not have been an entirely correct method of generating artificial clusters , since cluster size , surface brightness , and integrated magnitude are known to be correlated for milky way clusters . however , we decided it was better not to introduce additional assumptions about the correlation of these parameters into our detection test . once an artificial cluster was inserted into a copy of each hst frame , we extracted a @xmath18 region around the inserted cluster , and examined only that portion of the image .
this cut - out procedure was similar to the procedure used for the re - examination of ` problem ' images .
in fact the visual examination of the two groups of images was done at the same time , with no reference to which were the inserted clusters and which were the real objects . for each cut - out image
, we decided whether or not it was a _
bona fide _ globular cluster .
the results of our search for the inserted globular clusters are in figure [ fig - incomp ] , where detection of each inserted cluster is indicated as a function of @xmath13 magnitude and @xmath19 .
the figure shows , as expected , that our detection efficiency was generally worse for fainter objects and objects near the center of m31 .
faint clusters are more difficult to find against the bright background of the m31 disk and nucleus
. we also failed to detect a few bright objects , mostly in short exposures or in the near - uv filters f300w and f336w .
overall , we correctly identified 80% of the inserted clusters , and 92% of the objects which appeared in long f555w and f814w exposures .
the distribution of the real globular clusters and candidates detected in the hst images is shown in figure [ fig - gcdet ] .
the number of newly - detected objects increases at fainter magnitudes ; there is no clear trend in the number of new objects with @xmath19 .
we use the data in figures [ fig - incomp ] and [ fig - gcdet ] to estimate the completeness of existing catalogs .
while it would be desirable to estimate the completeness as a joint function of magnitude and position , the small number of objects we have to work with makes deriving @xmath20 difficult .
instead we summed over one variable to produce separate functions @xmath21 and @xmath22 , which are plotted in figure [ fig - compfn ] .
the catalog completeness is computed by dividing the number of cataloged objects in a given bin by the true number of objects : @xmath23 where the ` true ' object total is the sum of the number of cataloged objects and the number of new objects divided by our detection efficiency @xmath24 .
the number of new objects includes the marginal objects . from the results of @xcite , only a fraction ( @xmath25 ) of the marginal objects
are likely to be true globular clusters .
we therefore give a range of solutions for the completeness functions in figure [ fig - compfn ] , corresponding to @xmath26 and 1.0 .
the figure shows , as expected , that existing catalogs are reasonably complete to @xmath27 , after which the completeness drops drastically . to compute the completeness as a function of @xmath19
, we assumed that detection efficiency at @xmath28 was the same as that in the @xmath29 bin .
the completeness as a functions of @xmath19 does not follow any particular pattern ; the most important point is the low completeness in the innermost bin .
@xmath22 can only be measured out to about @xmath30 , and averaging over this region yields values for the overall completeness of 5085% .
the small number of objects per radial bin and the uncertainty about the nature of the marginal objects make this estimate rather imprecise .
it is important to know the total number of globular clusters in m31 , since it is one of the few spiral galaxies with well - studied gcss .
existing surveys ( summarized in * ? ? ?
* ) and our new hst survey bring the number of confirmed m31 gcs to over 250 . the most comprehensive attempt to estimate the total number of m31 gcs @xcite gives population ratios @xmath31 ; with @xmath32 , this gives @xmath33 , or @xmath34 .
we can use the results of our completeness study to attempt a new estimate of the total number of m31 gcs .
we take two approaches , which use the completeness data somewhat differently .
one approach is to use the result that for @xmath35 and @xmath36 , the existing sample is close to complete .
we can therefore use the results of gclf fitting in this region to estimate the number of clusters fainter than @xmath27 .
the gclfs computed in @xcite give a total number of clusters @xmath37 in the range 394417 ; the midpoint of the range is 406 .
in the region @xmath38 , there are 37 cataloged clusters or candidates and the catalog completeness is about 70% .
this implies that the true number of clusters is about 53 , so the total number of gcs in m31 is approximately @xmath39 .
a reasonable estimate of the error in this value is 15% , or @xmath40 .
we can also use the completeness estimates directly , to estimate @xmath41 where @xmath42 and @xmath43 are the number of good and marginal cataloged clusters in a given @xmath19 bin .
the @xmath44 used in the computation is the value plotted in figure [ fig - compfn ] for the appropriate value of @xmath45 .
the catalog used is that given by @xcite with likely ngc 205 clusters and ( likely young ) blue clusters removed .
the 299 good clusters are those confirmed by spectroscopy or high - resolution imaging and/or members of the ` adopted best sample ' of @xcite ; the other 130 clusters are considered marginal .
the resulting @xmath37 is sensitive to the value of @xmath45 and ranges from @xmath46 for @xmath47 to @xmath48 for @xmath49 .
a value of @xmath50 , which we believe is reasonable , gives @xmath51 .
our two estimates of the total number of gcs in m31 are compatible both with each other and with the results of @xcite .
the precision of our results is not much better than that of previous estimates , and improvement will require a wide - field , deep ccd survey of m31 which can be used to find gcs in a uniform manner across the galaxy ; such a survey is currently being carried out @xcite .
we now consider implications of our estimated value of m31 s @xmath37 for its specific frequency @xmath52 and ` mass specific frequency ' @xmath53 @xcite . to do this we need values for m31 s luminosity and mass - to - light ratio .
@xcite gives the total magnitude of m31 as @xmath54 .
correcting for foreground extinction @xmath55 ( see * ? ? ?
* ) and our adopted value of @xmath8 gives @xmath56 , which is bracketed by the values given by @xcite , @xmath57 , and @xcite , @xmath58 .
for ease of comparison we use @xmath59 as do @xcite . with @xmath60 ,
this gives @xmath1 and @xmath2 .
@xcite give @xmath52 and @xmath53 values for seven sb - sc spirals in addition to m31 .
the mean and dispersion of @xmath52 and @xmath53 for these galaxies are @xmath61 and @xmath62 . for four
sa and sab spirals the ( highly uncertain ) mean values are @xmath63 and @xmath64 .
m31 s values of @xmath52 and @xmath53 fall at the large end of the range of observed values for sb and sc spirals , and well within the range observed for sa and sab spirals .
while it has about twice as many clusters per unit mass or luminosity than the milky way , m31 is within the range of variation seen in other spirals gcss .
it is interesting to speculate on the difference between the milky way in m31 in terms of differences in the two galaxies histories , since obviously environmental differences can not be a major player .
@xcite suggested that perhaps m31 suffered an early major merger ; perhaps this was responsible for the creation of extra gcs in m31 , as in the picture of @xcite .
this would be consistent with the suggestion that some of the metal - rich gcs in m31 are younger than the rest of the population @xcite , although there are not enough metal - rich clusters to account for the entire ` cluster excess ' in m31 .
further explanation of the total number of clusters in m31 awaits both a larger spiral comparison sample and more detailed theoretical picture of gcs formation .
using the hubble space telescope archive to search for m31 globular clusters in wfpc2 images , we present the discovery of many previously - known clusters , a number of new cluster candidates , and some 20 objects which may be m31 open clusters .
we use the discovery data , together with an estimate of our discovery efficiency , to estimate the completeness of existing cluster catalogs .
as expected , the existing catalogs are least complete for faint clusters and clusters very near the center of m31 . as we found in a preliminary version of this analysis in @xcite ,
the completeness is very high to the magnitude limit we used for computing the globular cluster luminosity function .
this validates our finding that the m31 gclf varies with both radial distance from the center of m31 and with metallicity .
we use the completeness results to estimate the total number of globular clusters in m31 and derive values in the range 450500 , consistent with or somewhat higher than previous estimates .
the specific frequency of gcs in m31 is @xmath1 and the mass specific frequency @xmath2 .
m31 has more clusters per unit mass or luminosity than the milky way , but is within the range of specific frequencies seen in the limited number of other spirals studied to date .
lee , m. g. , kim , s. c. , geisler , d. , seguel , j. , sarajedini , a. , & harris , w. e. 2001 , in extragalactic star clusters , iau symposium 207 , ed .
e. k. grebel , d. geisler , & d. minniti ( san francisco : asp ) , 0 isophotes of ngc 205 ( nw ) and m32 ( se ) .
the wfpc2 symbols are drawn about @xmath65 actual size to make them easier to see .
the group of fields at @xmath66 , @xmath67 is part of a snapshot survey of field galaxies ( go-6354 ) .
[ field - pic ] ] magnitude vs. @xmath19 for artificial clusters .
symbol type indicates whether an object was detected and how it was classified .
the histograms are the fraction of inserted objects detected ; solid lines include only a or b class ( ` good ' ) gcs , and dashed lines include marginal objects .
[ fig - incomp ] ] vs. @xmath19 space .
symbols indicate object quality and presence in existing catalogs .
histograms estimate the existing catalogs completeness by showing ( number of previously known objects per bin)/(number of known @xmath68 number of new objects per bin ) .
solid line histograms include only a or b class gcs , and dashed line histograms include marginal objects .
[ fig - gcdet ] ] , is summed over entire magnitude range , and bottom panel , @xmath21 , is summed over entire radial range .
different line types reflect different assumptions about how many marginal objects are true m31 clusters .
solid lines : none , dotted lines : half , and dashed lines : all .
[ fig - compfn ] ] lllll 00 32 36.21 & + 39 27 43.4 & f606w & 1400 & u4k2oi01r + 00 32 36.62 & + 39 27 42.0 & f606w & 1500 & u4k2oi02r + 00 32 49.01 & + 39 35 00.4 & f555w & 1600 & u2e20709 t + 00 34 13.68 & + 39 23 26.5 & f814w & 2800 & u2ta0501 t + 00 34 13.26 & + 39 23 48.4 & f555w & 600 & u4490401r + 00 34 13.46 & + 39 24 40.5 & f702w & 600 & u27l0501 t + 00 36 59.20 & + 39 52 21.3 & f555w & 800 & u4710201 m + 00 37 43.08 & + 39 58 00.6 & f336w & 200 & u4f50907r + 00 37 49.14 & + 40 06 29.2 & f555w & 600 & u2782x01 t + 00 37 58.50 & + 39 58 32.8 & f606w & 2100 & u67ffp01r + 00 38 32.54 & + 41 28 45.4 & f555w & 830 & u39i0104 t + 00 38 55.51 & + 40 20 41.1 & f606w & 800 & u2804i01 t + 00 39 32.23 & + 40 30 48.1 & f555w & 5300 & u4ca0701r + 00 39 47.35 & + 40 31 58.0 & f555w & 1200 & u5bj0101r + 00 39 53.99 & + 41 47 19.2 & f555w & 2600 & u3kl1004r + 00 40 01.58 & + 40 34 14.8 & f555w & 1200 & u5bj0201r + 00 40 10.11 & + 40 46 08.9 & f814w & 200 & u4woah05r + 00 40 14.10 & + 40 37 11.4 & f555w & 160 & u2ye0703 t + 00 40 14.86 & + 40 49 02.8 & f814w & 200 & u4woc605r + 00 40 15.76 & + 40 36 48.1 & f300w & 1200 & u2m80c01 t + 00 40 22.15 & + 41 41 38.4 & f336w & 400 & u2gh020ct + 00 40 23.16 & + 41 40 55.6 & f555w & 2600 & u3kl0704 m + 00 40 23.66 & + 41 41 55.2 & f555w & 2600 & u3kl0804r + 00 40 23.77 & + 41 41 40.8 & f555w & 100 & u2ee0506 t + 00 40 25.50 & + 41 42 25.7 & f555w & 2600 & u3kl0904r + 00 40 26.84 & + 41 27 27.3 & f555w & 2000 & u2830103 t + 00
40 29.18 & + 41 36 31.9 & f814w & 5400 & u3kl0501 m + 00 40 29.40 & + 40 43 58.3 & f814w & 400 & u2ab0103 t + 00 40 30.61 & + 40 44 50.5 & f336w & 400 & u4f51107r + 00 40 31.26 & + 40 42 59.6 & f336w & 400 & u4f51007r + 00 40 33.17 & + 40 45 39.0 & f606w & 350 & u2g20701 t + 00 40 33.81 & + 41 39 40.2 & f555w & 2600 & u3kl0604 m + 00 40 34.22 & + 41 22 06.5 & f555w & 5300 & u4ca0201r + 00 40 39.54 & + 40 33 25.5 & f555w & 520 & u34l6903r + 00 40
39.75 & + 40 53 24.0 & f555w & 300 & u2g20e03 t + 00 40 46.06 & + 39 35 01.0 & f814w & 130 & u4woau05r
+ 00 40 50.80 & + 40 41 16.7 & f555w & 400 & u2q00101 t + 00 40 56.68 & + 40 35 29.0 & f555w & 5300 & u4ca0101r + 00 40 59.08 & + 40 46 42.1 & f606w & 1050 & u581ol01r + 00 40 59.54 & + 41 03 38.4 & f439w & 800 & u2tr0804 t + 00 41 16.28 & + 40 56 12.6 & f555w & 5300 & u4ca0301r + 00 41 17.85 & + 41 09 00.7 & f814w & 3700 & u2ot0o01 t + 00
41 22.08 & + 40 37 06.7 & f555w & 1200 & u5bj0301r + 00 41 38.85 & + 39 35 39.8 & f814w & 130 & u4wobk05r + 00 41 42.21 & + 40 12 22.4 & f814w & 2000 & u2830201 t + 00 41 43.30 & + 41 34 20.4 & f555w & 2000 & u2830303 t + 00 41 53.85 & + 40 50 30.2 & f814w & 1200 & u2806a02 t + 00 41 55.58 & + 40 47 15.0 & f555w & 5300 & u4ca0601r + 00 42 05.02 & + 41 12 14.9 & f300w & 2300 & u2ou7501 t + 00 42 05.27 & + 40 57 33.9 & f555w & 520 & u34l7003r + 00 42 06.07 & + 41 07 55.5 & f814w & 6200 & u3b83y01 t + 00 42 14.14 & + 41 10 22.6 & f606w & 160 & u2ourq01 t + 00 42 14.36 & + 41 06 24.7 & f606w & 1800 & u581r201r + 00 42 14.40 & + 41 10 11.7 & f555w & 8400 & u3d90207 t + 00 42 18.01 & + 40 45 03.7 & f555w & 900 & u3dg0107 t + 00
42 27.21 & + 41 08 28.0 & f300w & 1600 & u2ouut01 t + 00 42 28.88 & + 41 03 05.2 & f606w & 800 & u4k2rg01r + 00 42 31.00 & + 41 10 12.2 & f814w & 400 & u4wo9n05r + 00
42 32.47 & + 41 13 39.5 & f555w & 5200 & u2y30204 t + 00 42 32.70 & + 40 33 55.5 & f555w & 1200 & u3yk0101r + 00 42 35.13 & + 41 10 35.1 & f555w & 5200 & u2y30305 t + 00 42 38.97 & + 41 15 29.2 & f555w & 1680 & u2kj0109 t + 00 42 39.28 & + 40 51 42.2 & f814w & 600 & u2e20401 t + 00 42 39.49 & + 40 51 46.9 & f555w & 104 & u2lg0101 t + 00 42 39.88 & + 41 10 48.9 & f814w & 400 & u42z2302r + 00 42 40.85 & + 41 15 51.2 & f555w & 2500 & u5lt0104r + 00 42 40.96 & + 40 51 07.3 & f555w & 110 & u2ee0405 t + 00 42 41.68 & + 40 51 04.3 & f555w & 2000 & u2880704 t + 00 42 41.74 & + 41 15 57.9 & f814w & 600 & u2e20201 t + 00 42 42.21 & + 40 52 22.4 & f555w & 1200 & u2e20307 t + 00 42 44.64 & + 41 16 39.2 & f555w & 1680 & u2e2010bt + 00
42 46.91 & + 41 16 15.9 & f555w & 2200 & u2lg0201 t + 00 42 47.63 & + 41 16 11.0 & f336w & 460 & u2lh0103 t + 00 42 50.04 & + 41 36 17.8 & f814w & 700 & u4woc805r + 00 42 50.15 & + 40 59 56.0 & f814w & 8000 & u2oq3201 t + 00 42 50.34 & + 41 17 54.4 & f814w & 7000 & u2oqf801 t + 00 42 51.45 & + 41 06 52.9 & f814w & 400 & u4woa305r + 00 42 52.07 & + 41 24 53.4 & f814w & 6700 & u2oqf301 t + 00 42 52.26 & + 41 08 06.8 & f814w & 4500 & u26kcz01 t + 00 42 52.37 & + 41 10 31.7 & f814w & 400 & u42z3402r + 00 42 52.73 & + 40 56 30.4 & f814w & 2000 & u4xi0101r
+ 00 42 53.03 & + 41 14 23.4 & f814w & 400 & u42z1102r + 00 42 54.11 & + 41 08 08.9 & f814w & 800 & u4woa205r + 00 42 54.87 & + 41 10 35.0 & f814w & 700 & u4wobc05r
+ 00 42 58.84 & + 40 50 34.3 & f814w & 4400 & u3vj0103r + 00 43 00.93 & + 41 13 17.7 & f300w & 2600 & u31k0109 t + 00 43 01.94 & + 41 19 19.9 & f555w & 5200 & u38k0103 t + 00 43 04.61 & + 40 54 33.0 & f555w & 2000 & u2880801 t + 00 43 05.28 & + 40 50 37.8 & f300w & 5400 & u27h0f01 t + 00 43 05.38 & + 40 56 40.4 & f606w & 350 & u4g40104r + 00 43 06.12 & + 41 12 59.7 & f555w & 2000 & u5850103r + 00 43 06.14 & + 41 13 00.0 & f555w & 2000 & u5850107r + 00 43 06.19 & + 40 56 52.0 & f606w & 350 & u4g40103r + 00 43 06.40 & + 40 56 31.2 & f814w & 350 & u4g40101r + 00 43 07.21 & + 40 56 42.8 & f814w & 260 & u4g40102r + 00 43 07.85 & + 40 53 32.8 & f814w & 1100 & u4c80403r + 00 43 08.59 & + 41 14 51.7 & f300w & 8400 & u27h0e01 t + 00 43 09.06 & + 40 51 18.6 & f814w & 1000 & u4wobj05r + 00 43 12.49 & + 41 02 02.7 & f606w & 350 & u4g40204r + 00 43 12.82 & + 41 02 17.0 & f606w & 350 & u4g40203r + 00 43 13.76 & + 41 01 59.0 & f814w & 350 & u4g40201r + 00 43 14.09 & + 41 02 13.3 & f814w & 260 & u4g40202r + 00 43 18.05 & + 39 49 13.1 & f814w & 2000 &
u2830401 t + 00 43 20.30 & + 41 05 36.2 & f814w & 400 & u42z1202r + 00 43 20.80 & + 41 06 14.5 & f814w & 300 & u42z4602r + 00 43 22.39 & + 41 13 53.8 & f814w & 1500 & u2ot0s01 t + 00 43 25.28 & + 41 04 02.2 & f814w & 4400 & u3vj0203r + 00 43 36.72 & + 41 26 15.4 & f555w & 350 & u2kw0601 t + 00 43 43.26 & + 41 00 32.4 & f814w & 400 & u4woa105r + 00 43 46.58 & + 41 11 14.7 & f814w & 300 & u42z5702r + 00 43 47.85 & + 41 11 00.6 & f814w & 300 & u42z5802r + 00 43 54.55 & + 41 24 10.8 & f300w & 600 & u2m80g01 t + 00 43 57.06 & + 41 25 33.4 & f300w & 600 & u2m80h01 t + 00 44 14.38 & + 41 20 45.2 & f336w & 400 & u4f51207r + 00 44 23.45 & + 41 20 40.8 & f336w & 320 & u4f51307r + 00 44 23.74 & + 41 45 16.3 & f555w & 800 & u4710101 m + 00 44 34.81 & + 41 38 38.4 & f336w & 800 & u5750101r + 00 44 35.25 & + 41 31 21.6 & f814w & 500 & u4wobh08r + 00 44 36.29 & + 41 35 05.6 & f814w & 500 & u4wobs05r + 00 44 42.45 & + 41 44 24.2 & f555w & 5300 & u4ca0401 m + 00 44 42.52 & + 41 44 24.1 & f555w & 5300 & u4ca0501r + 00 44 42.59 & + 41 14 30.3 & f814w & 500 & u4woa005r + 00 44 44.23 & + 41 27 33.9 & f555w & 140 & u2y20106 t + 00 44 46.19 & + 41 51 33.4 & f555w & 1200 & u5bj0401r + 00 44 49.34 & + 41 28 59.0 & f555w & 140 & u2y20206 t + 00 44 50.61 & + 41 19 11.1 & f555w & 3800 & u2gv0401 t + 00 44 51.22 & + 41 30 03.7 & f555w & 160 & u2ye0603 t + 00 44 57.63 & + 41 30 51.7 & f555w & 140 & u2y20306 t + 00 45 00.36 & + 41 31 55.1 & f300w & 600 & u2m80a01 t + 00 45 03.79 & + 41 31 09.6 & f300w & 600 & u2m80e01 t + 00 45 05.66 & + 41 38 00.4 & f336w & 320 & u4f51407r + 00 45 07.76 & + 41 36 46.8 & f336w & 280 & u4f51507r + 00 45 09.25 & + 41 34 30.7 & f555w & 140 & u2y20406 t + 00 45 11.95 & + 41 36 57.0 & f555w & 140 & u2y20506 t + 00 45 28.46 & + 41 05 53.9 & f555w & 4320 & u2ug010at + 00 45 36.98 & + 41 42 17.3 & f606w & 2300 & u5hnm301r + 00 45 38.15 & + 41 36 02.4 & f300w & 650 & u4wobg0er + 00 45 39.00 & + 41 36 36.3 & f814w & 580 & u4wobg09r + 00 45 39.25 & + 41 36 32.4 & f606w & 1140 & u4wobg01r + 00 46 01.64 & + 40 41 58.3 & f555w & 4320 & u2ug020at + 00 46 20.46 & + 40 16 34.1 & f555w & 5300 & u4ca0801r + 00 46 24.33 & + 42 07 01.7 & f814w & 350 & u4woc905r + 00 46 24.56 & + 42 01 38.7 & f555w & 5300 & u4ca0901r + 00 46 29.85 & + 42 04 50.4 & f336w & 320 & u4f51707r + 00 48 15.15 & + 40 26 31.0 & f606w & 1000 & u36z7801r + 00 48 21.29 & + 40 29 02.4 & f606w & 1000 & u36z8401r
+ 00 48 42.83 & + 40 24 47.2 & f606w & 1000 & u36z7701r + 00 48 44.22 & + 40 33 06.0 & f606w & 1000 & u36z8201r
+ 00 49 08.22 & + 40 30 33.9 & f606w & 1000 & u36z8301r + 00 49 13.48 & + 40 21 51.1 & f606w & 1000 & u36z8001r + 00 49 18.35 & + 40 30 12.0 & f606w & 1000 & u36z8501r + 00 49 28.48 & + 40 21 54.4 & f606w & 1000 & u36z8101r + 00 49 31.26 & + 40 27 57.9 & f606w & 1000 & u36z8601r + 00 50 06.13 & + 41 33 56.0 & f606w & 2100 & u67fgy01r lll 000 - 253 & star & u2m80h01t[4 ] + 000-m046 & hii region ? & u2y20106t[2 ] + 000-m050 & hii region ? & u4wobh08a[4 ] + 000-m068 & stars & u4f51407a[4 ] + 000-v211 & stars & u2tr0804b[4 ] + 000-v212 & star & u2tr0804b[1 ] + 000-v298 & star & u2y20206t[4 ] + 064d - nb80 & star & u2kj0109a[3 ] + 074d - nb88 & star & u2lh0103b[4 ] u2lg0201b[4 ] u42z1102r[3 ] + 075d - nb96 & star & u2lh0103b[4 ] u2lg0201b[4 ] + 080d - nb93 & star & u2lh0103b[3 ] u2lg0201b[3 ] + 084d-000 & star & u2oqf801a[3 ] + 086d-000 & star & u3vj0103a[3 ] + 093d-000 & star & u27h0f01b[3 ] + 114d-000 & star & u4710101a[4 ] + 133 - 191 & stars & u2oqf301a[4 ] + 138 - 000 & star & u2oqf801a[3 ] + 166 - 000 & star & u2ot0s01a[4 ] + 185d-000 & star & u5bj0201a[3 ] + 193d-055 & star & u2ee0506t[3 ] u3kl0804a[4 ] u3kl0904a[1 ] + 254d-000 & star & u5bj0401a[4 ] + 285 - 000 & galaxy & u2ug010aa[2 ] + 326 - 000 & star & u2ee0506t[1 ] u3kl0804a[2 ] u2gh020ca[1 ] + 332 - 000 & star & u3kl0704a[4 ] + 444 - 000 & star & u2ee0506t[2 ] u3kl0704a[2 ] u3kl0804a[3 ] + 446 - 000 & galaxy & u3kl0501a[4 ] + 501 - 345 & galaxy & u36z8401a[1 ] + nb100 & star & u31k0109b[3 ] + nb103 & star & u42z1102r[4 ] + nb104 & star & u42z1102r[1 ] + nb106 & star & u31k0109b[4 ] + nb26 & star & u31k0109b[3 ] + nb27 & star ( pr ) & u2y30204a[2 ] + nb30 & star ( pr ) & u2e2010ba[3 ] + nb37 & star & u2kj0109a[4 ] + nb42 & star ( pr ) & u2e2010ba[3 ] u2lh0103b[2 ] u2oqf801a[1 ] + nb44 & star & u2e2010ba[3 ] + nb45 & star & u42z1102r[3 ] + nb49 & star & u42z1102r[2 ] u31k0109b[3 ] + nb51 & stars ( pr ) & u2y30204a[3 ] + nb53 & star & u2kj0109a[3 ] + nb54 & star & u2y30204a[3 ] + nb56 & star & u2y30204a[4 ] + nb75 & star & u2kj0109a[4 ] + nb76 & star & u2kj0109a[4 ] + nb78 & star & u2kj0109a[3 ] u5lt0104a[3 ] + nb82 & stars ( pr ) & u2y30204a[3 ] + nb84 & star ( pr ) & u2y30204a[3 ] + nb85 & star ( pr ) & u2y30204a[3 ] + nb94 & star & u2lh0103b[4 ] u2lg0201b[4 ] + nb95 & star & u2lh0103b[4 ] u2lg0201b[4 ] + nb97 & star & u42z1102r[1 ] + nb99 & star & u31k0109b[3 ] u42z1102r[2 ] + 000-d040 & blank & u4ca0101a[4 ] + 000-m023 & blank & u4f51307b[4 ] + 000-m056 & blank & u4f51407a[3 ] + 000-m060 & blank & u4f51407a[4 ] + 038d-000 & blank & u2ou7501t[3 ] + 092d-000 & blank & u31k0109b[2 ] + 353 - 000 & blank & u2lh0103b[4 ] u2lg0201b[4 ] u5lt0104a[2 ] + nb18 & blank & u2oqf801a[2 ] + nb38 & blank & u2y30204a[2 ] + nb40 & blank & u42z1102r[2 ] + nb57 & blank & u2y30204a[2 ] + nb58 & blank & u5lt0104a[2 ] u2e20201a[2 ] u2kj0109a[2 ] + nb59 & blank & u2lh0103b[4 ] u42z1102r[4 ] + nb74 & blank & u2e2010ba[2 ] u2e20201a[4 ] u5lt0104a[4 ] + nb87 & blank & u2oqf801a[2 ] + lllll m31gc j003411 + 392359 & 00 34 11.48&39 23 59.1 & c / d & + m31gc j004010 + 403625 & 00 40 10.33&40 36 24.7 & c / d & + m31gc j004023 + 414045 & 00 40 22.68&41 40 44.5 & c & + m31gc j004027 + 414225 & 00 40 27.25&41 42 24.8 & b & + m31gc j004030 + 404530 & 00 40 30.46&40 45 29.6 & b & + m31gc j004031 + 404454 & 00 40 30.63&40 44 54.3 & c & + m31gc j004031 + 412627 & 00 40 30.68&41 26 27.1 & c & + m31gc j004034 + 413905 & 00 40 34.42&41 39 04.8 & c / d & + m31gc j004037 + 403321 & 00 40 37.15&40 33 21.4 & c & + m31gc j004045 + 405308 & 00 40 44.92&40 53 07.6 & c & + m31gc j004051 + 404039 & 00 40 50.68&40 40 38.6 & b / c & + m31gc j004103 + 403458 & 00 41 02.88&40 34 57.9 & b & hodge 119 ? + m31gc j004146 + 413326 & 00 41 45.57&41 33 26.2 & c & + m31gc j004200 + 404746 & 00 42 00.39&40 47 45.8 & c & + m31gc j004228 + 403330 & 00 42 27.56&40 33 29.8 & c / d & + m31gc j004246 + 411737 & 00 42 46.01&41 17 36.5 & c & + m31gc j004251 + 405841 & 00 42 50.80&40 58 40.7 & c & + m31gc j004251 + 411035 & 00 42 50.78&41 10 34.7 & a & + m31gc j004257 + 404916 & 00 42 57.05&40 49 16.4 & c & hodge 195 ?
+ m31gc j004258 + 405645 & 00 42 58.02&40 56 45.4 & a & + m31gc j004301 + 405418 & 00 43 01.35&40 54 17.5 & b & + m31gc j004304 + 405129 & 00 43 04.27&40 51 29.2 & c & + m31gc j004304 + 412028 & 00 43 03.75&41 20 28.2 & a & + m31gc j004312 + 405303 & 00 43 11.86&40 53 02.8 & b & + m31gc j004312 + 410249 & 00 43 11.99&41 02 49.1 & c & + m31gc j004424 + 414502 & 00 44 23.71&41 45 02.3 & c & x - ray src : shp278 ?
+ m31gc j004425 + 414529 & 00 44 25.21&41 45 29.1 & c / d & + m31gc j004439 + 414426 & 00 44 39.07&41 44 26.3 & c & + m31gc j004537 + 413644 & 00 45 37.25&41 36 44.3 & b & + m31gc j004537 + 414332 & 00 45 36.75&41 43 32.2 & c & + m31gc j004622 + 420631 & 00 46 21.80&42 06 30.8 & c & + m31gc j004624 + 420059 & 00 46 23.50&42 00 58.5 & c & + m31oc j003836 + 412739 & 00 38 35.73&41 27 39.3 & b & + m31oc j003941 + 403154 & 00 39 40.52&40 31 53.6 & c & + m31oc j003943 + 403116 & 00 39 43.21&40 31 15.6 & c & + m31oc j004000 + 403326 & 00 39 59.99&40 33 25.9 & c & + m31oc j004008 + 403507 & 00 40 07.55&40 35 06.6 & b & + m31oc j004027 + 404524 & 00 40 27.26&40 45 23.7 & c & + m31oc j004031 + 404537 & 00 40 30.51&40 45 37.4 & c & + m31oc j004053 + 403519 & 00 40 52.94&40 35 19.2 & d & + m31oc j004054 + 404625 & 00 40 54.14&40 46 24.7 & c & + m31oc j004057 + 403425 & 00 40 56.62&40 34 24.7 & c & + m31oc j004119 + 403608 & 00 41 18.69&40 36 08.2 & b / c & + m31oc j004123 + 403756 & 00 41 23.30&40 37 56.1 & c & + m31oc j004421 + 414516 & 00 44 21.44&41 45 15.9 & c & + m31oc j004442 + 415122 & 00 44 41.84&41 51 22.4 & c & + m31oc j004442 + 415237 & 00 44 42.25&41 52 36.7 & c & + m31oc j004449 + 414430 & 00 44 48.83&41 44 30.3 & c / d & + m31oc j004450 + 415211 & 00 44 50.27&41 52 11.1 & c & + m31oc j004510 + 413646 & 00 45 10.45&41 36 46.3 & c & hodge 311 ? + m31oc j004512 + 413712 & 00 45 11.81&41 37 11.6 & c & h ii region , hodge 313 ?
+ m31oc j004539 + 414220 & 00 45 38.88&41 42 20.4 & c & radio src my0042 + 414 ?
+ lllllll 000001 & & & 13.807 & & 12.684 & 0.40 + 000d38 & & & 19.247 & & 18.276 & 0.37 + 000m045 & & 19.391 & 18.723 & & 17.446 & 1.49 + 000m91 & & & 19.143 & & & 0.89 + 006058 & & & 15.463 & & 14.354 & 0.38 + 009061 & & & & & 15.809 & 0.71 + 011063 & & & 16.578 & & 15.624 & 0.31 + 012064 & & & 15.042 & & 13.979 & 0.42 + 018071 & & & 17.533 & & 16.385 & 1.14 + 020d089 & & & & & 16.039 & 0.80 + 027087 & & & 15.559 & & 14.409 & 0.41 + 030091 & & & 17.377 & & 15.593 & 0.59 + 045108 & & & 15.784 & & 14.477 & 0.42 + 058119 & & & 14.925 & & 13.837 & 0.36 + 064125 & 17.461 * & & & & & 0.69 + 068130 & & 17.575 & 16.407 & & 14.849 & 0.68 + 070133 & & & 17.262 & & 16.165 & 0.20 + 071000 & 22.716 * & & & & & 0.07 + 076138 & 17.720 & 17.483 & 16.927 & & 15.626 & 0.53 + 077139 & & 18.829 & 17.734 & & 16.125 & 0.42 + 092152 & 18.766 * & & & & & 0.41 + 097d000 & & & & 17.878 & 17.121 & 1.08 + 101164 & 18.523 * & & & & & 0.42 + 109170 & & 17.407 & 16.197 & & 14.936 & 0.61 + 110172 & & & 15.355 & & & 0.66 + 114175 & & & 17.439 & & 15.940 & 0.47 + 115177 & & & 15.997 & & 14.560 & 0.25 + 118nb6 & & & 16.431 & & 15.207 & 0.52 + 123182 & & & 17.416 & 16.754 & 16.126 & 0.58 + 124nb10 & 16.094 & & 14.777 & & 13.631 & 0.53 + 127185 & 15.756 & & 14.467 & & 13.239 & 0.75 + 128187 & & & & 16.441 & 15.764 & 0.41 + 132000 & & & 17.739 & 17.244 & 16.440 & 0.34 + 134190 & & & & 16.064 & 15.502 & 0.52 + 143198 & & & 15.954 & & 14.731 & 0.25 + 145000 & 19.901 * & & & & & 0.47 + 146000 & 18.458 * & & & & & 0.59 + 148200 & & & 16.110 & & & 0.51 + 153000 & 18.220 * & & & & & 0.38 + 155210 & & & 18.011 & & 16.672 & 0.40 + 156211 & & & 16.969 & & 15.873 & 0.64 + 160214 & & & 18.076 & & 17.075 & 0.52 + 167000 & & & & & 16.109 & 0.33 + 205256 & 16.938 * & & & & & 0.32 + 231285 & & 18.227 & 17.248 & & & 0.48 + 232286 & 16.440 & 16.391 & 15.646 & & 14.543 & 0.66 + 233287 & & & 15.718 & & 14.585 & 0.41 + 234290 & & 17.780 & 16.783 & & & 0.51 + 240302 & & & 15.181 & & 14.230 & 0.80 + 257000 & & 11.907 & 20.960 & & 16.312 & 0.65 + 264000 & 18.652 * & & 17.577 & & 16.811 & 0.67 + 268000 & & & 18.314 & & 16.880 & 0.39 + 279d068 & & & 18.549 & & 16.964 & 0.68 + 311033 & & & 15.445 & & 14.215 & 0.38 + 315038 & & 16.548 & 16.473 & & & 0.56 + 317041 & & & 16.573 & & 15.713 & 0.73 + 318042 & & 17.234 & 17.047 & & & 0.63 + 319044 & & 18.333 & 17.608 & & & 0.41 + 324051 & & & 18.446 & & 17.633 & 0.30 + 328054 & & & 17.861 & & 16.918 & 0.85 + 330056 & & & 17.724 & & 16.555 & 0.56 + 331057 & & & 18.191 & & 17.076 & 0.56 + 333000 & & & 18.840 & & 17.711 & 0.90 + 338076 & & & 14.195 & & 13.174 & 0.56 + 342094 & & 18.033 & 17.733 & & & 0.92 + 343105 & & & 16.310 & & 15.274 & 0.36 + 358219 & & & 15.219 & & 14.122 & 0.55 + 368293 & & 18.189 & 17.924 & & & 0.54 + 374306 & 19.128 * & & 18.319 & & & 0.68 + 379312 & & & 16.183 & & 14.936 & 0.65 + 384319 & & & 15.752 & & 14.564 & 0.36 + 386322 & & & 15.547 & & 14.393 & 0.36 + 468000 & & & 17.788 & & 16.626 & 1.95 + nb21 & & & 17.865 & & 16.771 & 0.41 + nb39 & 18.551 & & 17.941 & & 17.876 & 0.38 + nb41 & & & 18.097 & & 17.183 & 0.43 + nb81 & & & 17.025 & & & 0.35 + nb83 & & & 17.585 & & 16.599 & 0.25 + nb86 & & & 18.522 & & 17.446 &
0.17 + nb89 & & & 17.965 & & 16.888 & 0.34 + m31gc j003411 + 392359 & & & & 22.302 & & 4.44 + m31gc j004010 + 403625 & & 18.906 & 18.505 & & & 0.24 + m31gc j004023 + 414045 & & & 18.289 & & 16.990 & 1.08 + m31gc j004027 + 414225 & & & 19.691 & & 19.138 & 0.76 + m31gc j004030 + 404530 & & & 16.064 & & & 0.37 + m31gc j004031 + 404454 & & 15.090 & 22.708 & & 20.337 & 0.29 + m31gc j004031 + 412627 & & & 20.930 & & 19.477 & 0.57 + m31gc j004034 + 413905 & & & 18.666 & & & 7.52 + m31gc j004037 + 403321 & & & 19.773 & & 18.773 & 0.49 + m31gc j004051 + 404039 & & & 19.862 & & 18.616 & 0.73 + m31gc j004103 + 403458 & & & 18.487 & & 17.920 & 0.38 + m31gc j004146 + 413326 & & & 20.716 & & 18.826 & 0.39 + m31gc j004200 + 404746 & & & 20.327 & & 19.636 & 0.45 + m31gc j004228 + 403330 & & & 21.195 & & 19.477 & 0.51 + m31gc j004246 + 411737 & 19.507 * & & 18.111 & & 17.706 & 0.43 + m31gc j004251 + 405841 & & & 20.296 & & 19.309 & 0.20 + m31gc j004251 + 411035 & & 19.177 & 18.178 & & 16.886 & 0.21 + m31gc j004257 + 404916 & & & 20.271 & & 18.648 & 1.18 + m31gc j004258 + 405645 & & & 18.066 & & & 0.31 + m31gc j004301 + 405418 & & & 18.613 & & 17.290 & 0.45 + m31gc j004304 + 405129 & & & 19.666 & & 18.392 & 1.12 + m31gc j004304 + 412028 & & & 18.828 & & 16.857 & 0.38 + m31gc j004312 + 405303 & & & 20.670 & & 19.150 & 0.73 + m31gc j004312 + 410249 & & & 18.810 & & 18.816 & 0.52 + m31gc j004424 + 414502 & & & 21.143 & & 19.232 & 0.31 + m31gc j004425 + 414529 & & & 20.597 & & 20.580 & 0.99 + m31gc j004439 + 414426 & & & 19.840 & & 18.746 & 0.67 + m31gc j004537 + 413644 & & 20.350 & 19.648 & & 18.724 & 0.92 + m31gc j004537 + 414332 & 22.529 * & & 19.759 & & & 0.47 + m31gc j004622 + 420631 & & 20.085 & 18.614 & & 17.135 & 2.86 + m31gc j004624 + 420059 & & & 20.445 & & 19.188 & 0.70 + m31oc j004539 + 414220 & 20.812 * & & 20.163 & & & 0.61 + m31oc j004027 + 404524 & & & 18.148 & & 18.046 & 0.55 + m31oc j004512 + 413712 & & 18.352 & 17.466 & & & 0.65 + m31oc j003941 + 403154 & & 20.398 & 20.002 & & & 0.68 + m31oc j004123 + 403756 & & 21.752 & 19.965 & & & 0.78 + m31oc j004442 + 415122 & & 19.788 & 19.533 & & & 1.08 + m31oc j004442 + 415237 & & 19.997 & 19.812 & & & 0.50 + m31oc j004450 + 415211 & & 20.599 & 20.439 & & & 0.40 + m31oc j004449 + 414430 & & & 20.229 & & 19.455 & 0.75 + m31oc j004054 + 404625 & & 23.432 & 22.082 & & 21.504 & 0.18 + m31oc j003943 + 403116 & & 20.854 & 21.014 & & & 0.45 + m31oc j004000 + 403326 & & 19.232 & 19.006 & & & 1.09 + m31oc j004031 + 404537 & & & 17.806 & & & 4.75 + m31oc j004057 + 403425 & & & 18.756 & & 18.675 & 0.86 + m31oc j004053 + 403519 & & & 18.579 & & 18.634 & 0.99 + m31oc j003836 + 412739 & & & 20.406 & & & 1.43 + m31oc j004008 + 403507 & & 20.773 & 20.298 & & & 0.84 + m31oc j004119 + 403608 & & & 19.501 & & & 1.88 + | globular clusters at the distance of m31 have apparent angular sizes of a few arcseconds .
while many m31 gcs have been detected and studied from ground - based images , the high spatial resolution of hst allows much more robust detection and characterization of star cluster properties .
we present the results of a search of 157 hst / wfpc2 images of m31 , we found 82 previously - cataloged globular cluster candidates as well as 32 new globular cluster candidates and 20 open cluster candidates .
we present images of the new candidates and photometry for all clusters .
we assess existing cluster catalogs completeness and use the results to estimate the total number of gcs in m31 as @xmath0 .
the specific frequency is @xmath1 and the mass specific frequency @xmath2 ; these values are at the upper end of the range seen for spiral galaxies . |
since veselago @xcite first pondered creating artificial materials to induce unorthodox physical properties in 1968 , metamaterials have been at the forefront of interdisciplinary research to develop novel applications . after the turn of the new millennium , groundbreaking progresses on invisibility and cloaking @xcite , slow - light @xcite and superlensing effects @xcite
have been reported .
this scientific paradigm has spawned the subfield of nanostructure - based plasmonic metamaterials , which exploit the properties of the surface plasmon , a collective oscillation of valence electronic charges @xcite .
the ability for plasmonic materials to focus light at sub - wavelength scales and to generate a large optical density of states are two significant physical properties which may be utilized in future applications @xcite .
there have already been numerous proposed implementations of plasmonic metamaterials , including spasers @xcite , optical transmission modulators @xcite , bio - sensers @xcite , switches @xcite and optical storage devices @xcite .
there is continued and active interest in schemes enabling long - range propagation in plasmonic waveguides @xcite since they are the building blocks of plasmonic circuitry @xcite , where it is hoped that a similar large bandwith of information as in conventional photonics can be achieved , but without the limitation of diffraction at submicron cross sections .
linear , one - dimensional ( 1d ) chains of regularly - spaced metallic nanoparticles , which support localized surface plasmons ( lsps ) confined to each nanoparticle , have been proposed almost two decades ago to act as plasmonic waveguides @xcite . when irradiated with light , the lsp excitations in the nanoparticles forming the array couple to each other through the optical near - field dipole - dipole interaction , which induces collective plasmons @xcite that extend over the whole chain and which guide electromagnetic radiation with strong lateral confinement . since the original proposal by quinten _
et al . _
@xcite , plasmon propagation in nanoparticle chains has been extensively investigated both theoretically @xcite and experimentally @xcite .
recently , the potentially nontrivial topological properties of collective excitations in zigzag @xcite and ` diatomic ' @xcite chains of nanoparticles were uncovered .
in particular , it has been shown in ref .
@xcite by means of a classical quasi - static calculation based on the macroscopic maxwell equations that the collective plasmonic oscillations in the linear array become decidedly topologically nontrivial when the chain becomes dimerized .
other 1d or quasi-1d artificial systems such as photonic lattices @xcite , dielectric resonators @xcite , silicon waveguides @xcite , polariton cavities @xcite and metallic nanowires @xcite were also shown to present interesting nontrivial topological features , paving the way to a new exciting field of research coined ` topological photonics ' @xcite . in this work
, we present a fully quantum - mechanical treatment of collective plasmons in a linear bipartite chain of metallic nanoparticles , which provides a clear analogy with the nontrivial electronic properties of fermionic topological insulators @xcite .
such an analogy remains only elusive within a classical treatment @xcite .
moreover , a quantum treatment becomes essential when the size of the nanoparticles composing the chain ( below ca .
@xmath0{nm}$ ] in diameter ) is such that quantum - size effects appear in their optical response @xcite .
the formed bipartite lattice , given by a regular array with a unit cell consisting of two identical nanoparticles , presents a small gap at the edge of the first brillouin zone in the two - band collective plasmonic bandstructure .
we find an effective hamiltonian for the collective plasmons , which at the edge of the brillouin zone corresponds to a 1d dirac equation for massive bosonic quasiparticles .
these bogoliubov excitations are accompanied with a nontrivial vacuum state , which is different from the vacuum of the original plasmonic excitations .
the dirac - like nature of the quasiparticles associated with our system allows us to propose a setup to observe the plasmonic analogue of klein tunneling @xcite , which is celebrated for enabling transmission with highly suppressed backscattering , as is the case in two - dimensional graphene sheets @xcite .
we show that depending on the strength of the dimerization , the system may be in either a topologically trivial or nontrivial phase , which is codified by a topological invariant @xcite .
this invariant governs the existence of special , topologically - protected modes at the ends of the chain and identifies the bipartite nanoparticle chain as a ` bosonic topological insulator ' for plasmonic excitations .
we unveil several types of edge states , which appear at the midgap frequency of the gapped plasmonic bandstructure .
we further investigate the topologically - protected midgap state formed between two dimerized chains , which is reminiscent of the jackiw - rebbi solitonic state @xcite , and which may be utilized to store light at the nanoscale .
a major challenge for the effective use of plasmonic circuitry is how to effectively combat losses necessarily present in a system with a high scattering rate of electrons , radiation losses into the far - field and dissipation due to quantum - size effects @xcite .
here we estimate , with analytical expressions for both radiation and landau @xcite damping decay rates , the plasmonic losses in the bipartite chain .
in particular , we find that the two - band collective plasmonic bandstructure is comprised of one bright and one dark band per polarization , with drastically different responses to radiative losses .
this analysis enables us to comment on the experimental challenges to observe topological collective plasmons .
the sequel of this paper is organized as follows : section [ sec2 ] presents our model of collective plasmonic excitations in a bipartite chain of metallic nanoparticles coupled to a bath of electron - hole pairs ( responsible for the nonradiative landau damping ) as well as to a photonic bath ( which leads to radiation damping of the collective excitations ) . in sec .
[ sec3 ] , we diagonalize the purely plasmonic problem and discuss the properties of the associated plasmonic bandstructure .
we detail in sec .
[ sec4 ] the mapping of the plasmonic hamiltonian to a 1d dirac equation for massive bosonic quasiparticles , and consider the plasmonic version of klein tunneling . in sec .
[ sec5 ] , we study the topological properties of the bipartite chain , codified by a nontrivial zak phase , and illustrate how various types of edge states and topologically - protected midgap states may form . an analysis of plasmonic losses due to both the size - dependent radiation and landau damping is presented in sec
. [ sec6 ] , which enables us to comment on the experimental realization of our proposal .
finally , we draw conclusions in sec .
a few appendices complement the discussion presented in the main text .
we consider a linear chain composed of @xmath1 dimers , which comprise the unit cell of two identical spherical metallic nanoparticles .
the two corresponding inequivalent sublattices are denoted by @xmath2 and @xmath3 , respectively .
this geometry is sketched in fig .
[ fig1 ] , showing the nanoparticle radius @xmath4 and the interparticle separations @xmath5 and @xmath6 which build a bipartite chain with a period @xmath7 . along the chain
each nanoparticle supports an lsp resonance @xcite , a collective excitation of the electronic center of mass , which couples to the neighboring nanoparticles via the dipole - dipole interaction @xcite .
this approximation is valid for interparticle distances @xmath8 @xcite . in what follows ,
we denote by @xmath9 the plasmonic hamiltonian describing the near - field - coupled lsps along the bipartite chain .
each lsp is coupled to vacuum photonic modes described by the hamiltonian @xmath10 through the light - matter interaction @xmath11 ( which includes retardation effects ) , leading to the radiative decay of the collective excitations .
moreover , the coupling described by the hamiltonian @xmath12 of the lsps to electron - hole pairs ( with hamiltonian @xmath13 ) in each of the nanoparticles leads to the size - dependent landau damping of the collective modes @xcite .
the hamiltonian of our system hence reads @xmath14 dimers , comprised of a pair of metallic nanoparticles denoted @xmath2 and @xmath3 , each supporting localized surface plasmon resonances . ] since the nanoparticles we consider are small enough to be subject to quantum - size effects @xcite , and to elucidate the connection to fermionic quasiparticles in dirac materials , we quantize the classical plasmonic hamiltonian @xmath9 @xcite in the coulomb gauge @xcite , leading to @xmath15 where @xmath16 is an index which denotes the dimer number in the chain ( see fig .
[ fig1 ] ) , and the summation over @xmath17 counts the two transverse ( @xmath18 ) and the single longitudinal ( @xmath19 ) polarizations of the collective plasmonic modes .
the factor @xmath20 ( @xmath21 ) accounts for the repulsive ( attractive ) interaction between transverse ( longitudinal ) nearest neighbor lsps .
we neglect interactions beyond those between nearest neighbors since such interactions only lead to a small quantitative change of the plasmonic dispersion ( see appendix [ appenda ] for details ) .
the bosonic operators @xmath22 and @xmath23 ( @xmath24 and @xmath25 ) in eq . annihilate ( create ) an lsp with polarization @xmath26 in the @xmath16th dimer in nanoparticle @xmath2 and @xmath3 , respectively .
the hamiltonian displays terms which are reminiscent of a 1d tight - binding model .
firstly , the role of the onsite energy is played by the lsp resonance frequency @xmath27 [ here we refer to the first and second terms on the right - hand side of eq . ] . for the case of alkaline nanoparticles in vacuum
, this resonance frequency corresponds to the mie frequency @xmath28 , where @xmath29 is the charge on an electron with effective mass @xmath30 and the number of electrons in each nanoparticle is @xmath31 @xcite .
the plasma frequency is @xmath32 , where @xmath33 is the electronic density of the considered metal .
secondly , the analogue of the hopping terms in eq . , namely those @xmath34 and @xmath35 , are due to the nearest neighbor dipole - dipole interaction .
the dipolar interaction gives rise to the coupling constants @xmath36 which alternate along the chain .
however , this is where the comparison to a standard tight - binding model ends , due to the appearance of nonresonant terms @xmath37 and @xmath38 which enter the hamiltonian and are also due to the dipolar interaction .
such nonrotating - wave terms can not be neglected when calculating plasmon eigenstates and , hence , quantities dependent on the latter , such as damping rates ( see sec .
[ sec6 ] ) @xcite . without the nonresonant terms
mentioned above , the model encapsulated in the hamiltonian can be seen as the bosonic counterpart to the su - schrieffer - heeger @xcite and rice - mele models @xcite , where electrons hopping in 1d lattices with staggered hopping amplitudes were treated for the spinless and spinful cases , respectively .
these simple , though physically rich models were originally developed in studies of conducting polymers such as polyacetylene , and are known to exhibit rather exotic physics including fractionally charged excitations and edge states @xcite .
how the presence of the bosonic nonresonant terms in the hamiltonian will affect the formation and properties of edge states is not immediately clear .
this issue is addressed in the sequel of our paper . in eq .
, the vacuum modes of the electromagnetic field , to which the lsps are coupled to , lead to the radiative damping of the plasmonic collective modes and are described by the photonic hamiltonian @xmath39 where @xmath40 ( @xmath41 ) annihilates ( creates ) a photon with momentum @xmath42 , transverse polarization @xmath43 , and dispersion @xmath44 , with @xmath45 the speed of light in vacuum . in the long - wavelength limit , the plasmon - photon coupling hamiltonian entering eq .
has the expression @xmath46 , \end{aligned}\ ] ] where the position of the center of the nanoparticle belonging to sublattice @xmath2 ( @xmath3 ) in the @xmath16th dimer is denoted by @xmath47 ( @xmath48 ) , and where @xmath49 is the quantization volume of the photonic modes .
we note that the coupling , together with the static dipole - dipole interaction in eq . , result in a fully retarded description of the plasmon - photon interaction @xcite .
the hamiltonian describing the bath of electron - hole pairs in eq . reads as @xmath50 where @xmath51 labels the electron and hole states with energy @xmath52 in nanoparticle @xmath53 , and where @xmath54 annihilates ( creates ) the one - body state @xmath55 . in the hard - wall approximation for the mean - field potential felt by the electrons @xcite , the collective plasmon - electron - hole coupling leading to the landau damping of the collective plasmons reads @xcite @xmath56,\end{aligned}\ ] ] where @xmath57 .
we start by fully characterizing the fundamental properties of the plasmonic bipartite chain in absence of the damping mechanisms . to this end , we diagonalize exactly in the sequel the hamiltonian by means of bosonic bogoliubov transformations .
we provide analytic expressions for the plasmonic dispersion relation and its associated eigenstates .
using periodic born - von karman boundary conditions and introducing the bosonic operators in momentum space @xmath58 and @xmath59 through the fourier transforms @xcite [ eq02 ] @xmath60 the real space hamiltonian becomes @xmath61,\end{aligned}\ ] ] where we have introduced the quantity @xmath62 the quadratic hamiltonian is diagonalized with the bosonic bogoliubov transformation @xcite @xmath63 where the index @xmath64 ( @xmath65 ) corresponds to the upper ( lower ) collective plasmon band . in eq . , the bogoliubov coefficients read [ eq06 ] @xmath66 and @xmath67 and the auxiliary bosonic operators are @xmath68 thus we arrive at the diagonalized form of eq .
@xcite , @xmath69 where the eigenfrequencies of the collective plasmonic modes are @xmath70 ) and longitudinal ( blue and cyan solid lines , @xmath71 ) modes for both the upper ( @xmath72 ) and the lower ( @xmath73 ) bands . the undimerized case ( thinner lines ) with @xmath74 and a typical dimerized case ( thicker lines ) with @xmath75 and @xmath76 are both displayed . ] the dispersion relation is plotted in fig .
[ fig2 ] for the regular chain limit where @xmath77 ( thinner lines ) and a typical dimerized chain with @xmath78 ( thicker lines ) for both the transverse ( dashed red and orange lines ) and longitudinal ( solid blue and cyan lines ) collective plasmons .
one notices an immediate analogy to the energy bands in a peierls chain @xcite , as the dimerization has the effect of opening up a small gap of the order of @xmath79 at the edges of the first brillouin zone ( @xmath80 ) .
the longitudinal collective plasmon mode has a gap which is about twice as large as compared to the transverse modes because the attractive interaction between pairs of dipoles parallel to the chain is twice as much as between dipoles perpendicular to the chain which experience a repulsive interaction [ cf .
the factor @xmath81 appearing in eq . ] . in appendix [ appenda ]
we calculate the collective plasmon dispersion where the dipole - dipole interaction is considered beyond nearest neighbors .
such additional terms in the near - field interaction only lead to a minor correction to the dispersion relation shown in fig .
[ fig2 ] , justifying the nearest - neighbor approximation to the near - field interaction which we use in the remainder of this work .
now that the plasmonic hamiltonian has been diagonalized in sec .
[ sec3 ] , we are in a position to study in more details the properties of its associated eigenstates . in this section we show that at the edge of the brillouin zone , the system can effectively be described by a 1d dirac - like hamiltonian , which gives rise to a klein tunnelling phenomenon for the collective plasmons .
we further comment on the experimental setups required to observe the aforementioned tunneling effect . to analyze the physics occurring at the edge of the brillouin zone , we taylor - expand the plasmon dispersion in the vicinity of @xmath82 , with @xmath83 . in the regime @xmath84 , we find that the collective excitations are governed by a pseudo - relativistic spectrum @xmath85 which is reminiscent of the spectrum of massive dirac particles , @xmath86 , where @xmath87 distinguishes the electron and positron particle species , and @xmath88 and @xmath89 are the momentum and mass of the particle , respectively .
the pseudo - relativistic spectrum indicates a connection to a dirac - like hamiltonian , as was first noticed in the continuum quantum field theory study of the ( fermionic case ) of diatomic polymers @xcite . working in the regime of first order in @xmath84 ,
one finds from eq . that @xmath90 and @xmath91 , where @xmath92 .
it follows that for small dimerizations @xmath93 , @xmath94 , such that the plasmonic hamiltonian can be expressed in terms of the auxiliary operators of eq . ,
@xmath95 . performing the summation over @xmath96
, we obtain a @xmath97 matrix hamiltonian @xmath98 , with @xmath99 where @xmath100 and where @xmath101 .
the associated eigenvectors are in the form of a dirac spinor @xmath102 which is reminiscent of eigenvectors found in 1d dirac materials such as single - wall carbon nanotubes @xcite .
the index @xmath96 denotes whether the excitation can be thought of ( in the language of semiconductor physics ) as hole - like ( @xmath103 ) or electron - like ( @xmath104 ) , as is the case with the spinor of graphene @xcite . for the dirac - like nature of the hamiltonian to become apparent ,
we utilize the unitary transform @xmath105 , to reach @xmath106 , with @xmath107.\ ] ] here , @xmath108 is the @xmath97 identity matrix and @xmath109 , @xmath110 and @xmath111 are pauli s spin matrices .
the parallels with the 1d form of the dirac hamiltonian @xmath112 are evident , up to the unimportant constant energy shift @xmath113 .
a striking feature in the effective hamiltonian is that the role of the mass term in a typical dirac hamiltonian is played by a quantity @xmath114 which can be modified in both magnitude and sign through the interparticle distances @xmath115 and @xmath116 [ see eq . ] .
this tunability has significant implications for both pseudo - relativistic klein tunneling ( see appendix [ appendb ] ) and the topological properties of bipartite metallic nanoparticle chains ( cf . sec .
[ sec5 ] ) .
the present realization of 1d dirac - like bosons encapsulated in the hamiltonian complements the recent discovery of their counterparts in 2d honeycomb metasurfaces @xcite .
a prominent feature of excitations described by a dirac equation is their ability to be transmitted through a potential barrier with highly suppressed backscattering @xcite , the so - called klein paradox . for the system of nanoparticles we investigate here , a plasmonic version of klein tunneling
can be realized by arranging a dimerized chain into two distinct sections , which we distinguish with the notation @xmath117 ( for left ) and @xmath118 ( for right ) .
these two sections have different resonance frequencies @xmath119 to mimic the application of an electrostatic potential step , as can be inferred from the first term in the right - hand side of the dirac - like hamiltonian ( see appendix [ appendb ] for a detailed calculation of the corresponding transmission probability ) .
such a scenario may be experimentally achieved by the three following different means : ( i ) using two different metals ( for example silver and gold ) ; ( ii ) modifying the electrostatic environment of one chain with respect to the other , using an embedding dielectric medium ; or ( iii ) using nanoparticles of different sizes , for which quantum - size effects such as the spill - out of the electronic density outside the nanoparticle @xcite as well as coupling to the electronic environment @xcite influence the position of the surface plasmon resonance .
the appearance of a dirac equation with its associated spinor wavefunctions in the description of the collective plasmons in the dimerized chain ( see sec . [ sec4 ] ) indicates that a nontrivial berry phase may arise in this system . in this section
we study the implications of such physics on bosonic quasiparticles , classify the topological properties which are found to arise and investigate the existence of edge and soliton - like states . as a result
, we uncover our system as a ` plasmonic topological insulator ' , which has the ability to both harbor and exploit topologically - protected midgap states .
the berry phase @xcite when applied to the dynamics of electrons in periodic solids in 1d is known as the zak phase @xcite .
the latter is connected to a @xmath120 topological invariant , which characterizes exotic properties such as fermionic charge fractionalization @xcite . for our purposes here ,
its utility arises because it is related to the existence of edge states , in a manifestation of a bulk - edge correspondence principle @xcite .
the zak phase is defined as an integration over the first brillouin zone of the berry connection @xmath121 , explicitly @xmath122 which requires knowledge of the plasmonic eigenstates @xmath123 .
we may write the plasmonic eigenstates via identification with the bogoliubov coefficients of eq . , namely @xmath124 in the basis @xmath125 @xcite . then , proceeding to calculate the zak phase from eq . with eq . , we find @xmath126 defined modulo @xmath127 ( since the berry connection @xmath128 is not gauge invariant ) .
thus the dimerized chain has a trivial zak phase @xmath129 when @xmath130 [ @xmath131 , cf .
eq . ] and a nontrivial one @xmath132 when @xmath133 ( @xmath134 ) , with a topological phase transition at the crossover point of the regular chain @xmath77 , where the bandgap closes ( see fig .
[ fig2 ] ) .
equation alludes to the presence of highly confined edge states in the nontrivial phase , corresponding to topologically - protected localized collective plasmons , which we investigate in detail in the remainder of this section .
( red dots , thin gray line : guide for the eye ) at each site in a chain of @xmath135 nanoparticles , where @xmath136 .
inset : sketch of a typical chain comprised of an even number of nanoparticles , where the loops depict the favored dimerization . ]
we now search for a midgap collective plasmon state at the frequency @xmath27 supported by the finite bipartite chain described by the hamiltonian and composed of an even number of nanoparticles ( for completeness , we consider in appendix [ appendd ] the case where the chain is composed of an odd number of constituants ) . to this end
, we work in real space with open boundary conditions . due to the requirement of preserving bosonic statistics , diagonalization of eq .
requires the framework of paraunitary transforms @xcite .
the result of this numerical procedure is shown in fig .
[ fig3 ] , where we plot the probability density associated with the midgap state at each nanoparticle site in the regime @xmath137 ( @xmath133 ) . as foreshadowed by the nontrivial zak phase @xmath138 found in sec .
[ sec : zak ] in that case , both edges are highly plasmonically active .
we have also checked numerically that , as expected , there is no midgap state when @xmath139 ( @xmath130 ) for which one finds a vanishing zak phase [ cf .
eq . ] . to better understand the critical factor determining the formation of the midgap edge states , we neglect the nonresonant terms in eq . and maintain only the regular ` hopping ' terms in the following analytic analysis .
this rotating wave approximation ( rwa ) is justified post hoc due to the excellent comparison to the results found via full numerical diagonalization of the hamiltonian in the limit of a large number of nanoparticles .
we thus start from the real space hamiltonian in the form @xmath140 , with @xmath141 writing the above hamiltonian as @xmath142 where @xmath143 , @xmath144 is an even , symmetric , tridiagonal matrix .
the diagonal elements of @xmath144 are all equal to @xmath27 , while the off - diagonal elements alternate between @xmath145 and @xmath146 . diagonalizing the matrix @xmath144 amounts to the following system of equations for the coefficients @xmath147 associated with the operator @xmath148 : [ eqbig1 ] @xmath149\ ] ] @xmath150\ ] ] @xmath151 @xmath152 searching for midgap states at @xmath27 produces from eqs . and the boundary conditions @xmath153 at the chain edges . in the bulk , the recurrence relations formed from eqs . and
asymptotically satisfy these boundary conditions as long as @xmath137 ( @xmath133 ) , with the probability densities on each site given by [ eq88 ] @xmath154}\left ( \frac{\omega_1}{\omega_2 } \right)^{2 ( n-1 ) } , \\
\left|b^\sigma_{n}\right|^2 & = \frac{(\omega_1/\omega_2)^2 - 1}{2\left[(\omega_1/\omega_2)^{2\mathcal{n}}-1\right]}\left ( \frac{\omega_1}{\omega_2 } \right)^{2 ( \mathcal{n}-n)},\end{aligned}\ ] ] with @xmath155 $ ] .
a plot of the expressions above is indistinguishable from the one of the full numerical solution in fig .
[ fig3 ] for long enough chains and so is not displayed in the figure @xcite .
the exponential behavior of the plasmonic probability density associated with the midgap state at frequency @xmath27 can be understood with the help of the simple picture in the inset of fig .
[ fig3 ] , where we sketch a representative chain consisting of an even number of nanoparticles .
the @xmath137 dimerization favors a pairing - up of nanoparticles as depicted with the drawn loops , leaving seemingly isolated nanoparticles at both ends of the chain which are both free to harbor the majority of the plasmonic activity . a completely different edge state as the one previously discussed may form at the interface between two topologically distinct bipartite chains ( both considered for simplicity to have the same resonance frequency @xmath27 ) . at the formed ` domain wall ' , a topologically protected kink - like bound state arises at the midgap frequency @xmath27 .
in this way , the state is reminiscent of a jackiw - rebbi solitonic mode @xcite , an exponentially pinned eigenstate which arises at the boundary where the mass term changes sign in a dirac equation with position dependent mass .
let us consider bringing together two bipartite chains , identified by their different zak phases @xmath156 and @xmath157 , and characterized by their coupling constants @xmath158 we choose the left - hand side chain to be topologically trivial @xmath159 and the right - hand side chain to be topologically nontrivial @xmath160 .
furthermore , the connection of these two chains is described with the interface coupling constant @xmath161 . for simplicity
, we investigate the case of an odd total number of nanoparticles , comprised of @xmath162 nanoparticles in the left hand side chain and @xmath163 nanoparticles in the right hand side chain .
the resulting real space hamiltonian can be written down in analogy to the hamiltonian . at each site in a chain of @xmath164 nanoparticles , for three different choices of the coupling constants . in the two symmetric setups @xmath165 , with the interface coupling constant being either @xmath166 ( red dots ) or @xmath167 ( blue squares ) . in the asymmetric setup @xmath168 , @xmath169 , and @xmath166 ( green triangles ) .
the thin gray lines are guides for the eye . ] as alluded to in sec .
[ sec : edge_states ] , one may diagonalize quadratic , bosonic hamiltonians by employing paraunitary transforms @xcite . following this approach ,
we obtain the probability density of the midgap state at frequency @xmath27 as a function of the nanoparticle site , which is plotted in fig .
we consider three fundamental varieties of solitonic state which can be obtained by expedient choices of inter - nanoparticle separations .
firstly , we consider symmetric states with coupling constants @xmath165 . with the interface coupling constant @xmath166 ( red dots )
the soliton is single - peaked and centered on the nanoparticle corresponding to the @xmath2 sublattice in the @xmath170st dimer ; whilst for @xmath167 ( blue squares ) the state is double - peaked , with equal maxima on the nanoparticles corresponding to the @xmath2 sublattice in the @xmath171th and @xmath172st dimer .
finally , a typical asymmetric midgap state is shown in fig .
[ fig4 ] with the parameters @xmath168 , @xmath169 , and @xmath166 ( green triangles ) . in this case
, the significant asymmetry causes the midgap state to be highly confined to the nanoparticle in sublattice @xmath2 in the @xmath172st dimer .
notably , due to the experimental possibility to position metallic nanoparticles in any desired location , the system provides some tunability over the size of the localization length of the soliton - like state .
the topologically - protected midgap states above can be well described using the rwa , which disregard the nonresonant terms which appear in the governing hamiltonian .
the validity of this conjecture is , as in sec .
[ sec : edge_states ] , proved after the fact by comparing the analytic expressions obtained in the rwa with the outcome of the numerical diagonalization of the full hamiltonian .
such an rwa analysis proceeds in a similar manner to that carried out in sec .
[ sec : edge_states ] .
one finds the complete extinction of plasmonic activity on the @xmath3 sublattice , @xmath173 , with @xmath174 $ ] .
meanwhile on the @xmath2 sublattice , we find an exponentially pinned probability density profile [ eq : probsoliton ] @xmath175 for @xmath176 $ ] and @xmath177 for @xmath178 $ ] in terms of the probability density on the final @xmath2 site of the left hand side chain , @xmath179^{-1}. \end{aligned}\ ] ] since these analytic expressions are imperceptible from the numerically exact solution displayed in fig .
[ fig4 ] , they are not plotted in the figure .
equation defines a bosonic version of a solitonic midgap state .
its existence at the domain wall boundary only requires the connection of two chains of different zak phases .
the precise details of the chain are unimportant , so it corresponds to a topologically - protected plasmonic bound state .
the analysis above provides a physical realization of various bosonic jackiw - rebbi - like states , and gives rise to the chance of witnessing such exotic states experimentally , especially when compared to fermionic systems such as conducting polymers where such an observation is a challenging task .
now that the topological properties of the purely plasmonic problem encapsulated in the hamiltonian have been discussed , here we analyze in detail the influence of both the photonic and electronic environments [ eqs . and , respectively ] onto plasmon lifetimes .
in particular , we calculate the radiative and size - dependent nonradiative ( landau damping ) losses encountered by the collective plasmons in bipartite chains of metallic nanoparticles , and comment on the various challenges to experimentally observe their topological properties .
the plasmonic system sketched in fig .
[ fig1 ] is subject to radiative losses , which arise due to the coupling of the collective plasmons along the nanoparticle chain to vacuum electromagnetic field modes described by eq . and
entering the system hamiltonian . treating the plasmon - photon coupling as a perturbation , the fermi golden rule result for the radiative decay rate of the collective plasmon with momentum @xmath180 in the branch @xmath96 and polarization @xmath26 reads @xmath181 the above expression describes the transitions between the plasmonic excited state @xmath182 and the ground state @xmath183 by spontaneous emissions of photons ( @xmath184 ) . using eq
. together with the bogoliubov transformation , we arrive at @xmath185 where the array factor is defined as @xmath186d},\ ] ] with @xmath187 being the @xmath19 component of the photon wavenumber @xmath42 .
equation can easily be evaluated , yielding @xmath188d/2)}}{\sin^2{([q - k_z]d/2)}},\ ] ] which , in the large chain limit ( @xmath189 ) , simplifies to @xmath190 d \right ) $ ] . using that @xmath191 , working in the continuum limit ( @xmath192 ) , and replacing the sum over @xmath42 by a three - dimensional integral in eq .
, we finally arrive at the result @xmath193 \nonumber\\ & \times \left [ 1 + \tau \operatorname{sgn}{\{\eta_{\sigma}\ } } \frac{\omega_1 \cos ( q d_1 ) + \omega_2 \cos ( q d_2)}{\sqrt{\omega_1 ^ 2 + \omega_2 ^ 2 + 2 \omega_1 \omega_2 \cos
( q d ) } } \right ] \nonumber\\ & \times \theta \left ( \omega_q^{\sigma \tau } - c |q| \right)\end{aligned}\ ] ] for the radiation damping of the collective plasmons in terms of the single particle result @xmath194 , where @xmath195 . ) and longitudinal ( blue and cyan solid lines , @xmath71 ) collective plasmonic modes for @xmath196 .
the undimerized case ( thinner lines ) with @xmath74 and a typical dimerized case ( thicker lines ) with @xmath197 and @xmath198 are both displayed . only the results for the bright bands , i.e.
, the lower @xmath73 band for the longitudinal polarization and the upper @xmath72 band for the transverse modes , are displayed in the figure as the dark bands provide vanishingly small contributions . ]
we plot in fig .
[ fig5 ] the radiation damping decay rate .
the results for the regular chain limit with @xmath77 ( thinner lines ) and a typical dimerized chain with @xmath199 ( thicker lines ) are both plotted , each for the transverse ( dashed red / orange lines ) and longitudinal ( solid blue / cyan lines ) polarizations .
notably , even when the chain is in the dimerized regime , there remains only one band per polarization that provides a non - negligible contribution to the decay rate .
namely , the longitudinal modes in the upper @xmath72 band and the transverse modes in the lower @xmath73 band are weakly coupled to light and are thus referred to as ` dark bands ' .
these polarization - dependent low - loss bands will inevitably have an increased plasmonic lifetime due to being protected against radiation damping .
the dark modes arise due to destructive interferences in the far field , with an almost vanishing net dipole moment .
in contrast the bright modes ( the lower @xmath73 band for the longitudinal polarization and the upper @xmath72 band for the transverse modes ) contribute a non - vanishing dipole moment and are thus subject to radiation damping .
as can be seen in fig .
[ fig5 ] , dimerization has a weak effect on radiation damping , as the latter follows qualitatively the @xmath180-dependence of the simple chain result @xcite . outside of the light cone
@xmath200 the decay rate is vanishing , whilst inside the light cone @xmath201 the modes are superradiant , with a magnitude significantly larger than that of a single nanoparticle .
the coupling of the collective plasmons to a bath of electron - hole pairs , described by the hamiltonian and appearing in the system hamiltonian , gives rise to a size - dependent nonradiative decay of a purely quantum mechanical origin : the landau damping @xcite . assuming that the plasmon - electron - hole hamiltonian acts perturbatively , one can write down the following fermi golden rule for the nonradiative decay rate of the collective plasmon with momentum @xmath180 in the branch @xmath96 and with the polarization @xmath26 , @xmath202 with @xmath203 denoting electron - hole states in the nanoparticle belonging to the dimer @xmath16 and the sublattice @xmath204 . the above expression describes the transitions between the plasmonic excited and ground states by nonradiative emission of electron - hole pairs ( @xmath205 ) .
evaluating eq .
with the help of the bogoliubov transformation yields @xmath206 in terms of the summation over electron and hole states ( assumed to be the same in each nanoparticle ) @xmath207 the quantity @xmath208 can be analytically calculated using a semiclassical expansion of the angular - momentum - restricted density of states @xcite , leading to @xmath209 where @xmath210 and @xmath211 are the fermi velocity and energy of the metal , respectively .
an explicit expression of the monotonically decaying function @xmath2120,1]$ ] can be found in eqs . ( 107 ) and ( 108 ) of ref .
@xcite or in eq .
( b2 ) of ref .
@xcite . ) and longitudinal ( blue and cyan solid lines , @xmath71 ) collective plasmonic modes for @xmath213 .
the thinner ( thicker ) lines correspond to @xmath74 ( @xmath197 and @xmath198 ) . ]
the landau damping decay rate is plotted in fig .
[ fig6 ] , scaled with the single nanoparticle result @xmath214 @xcite .
the results for the regular chain with @xmath77 ( thinner lines ) and a typical dimerized chain with @xmath199 ( thicker lines ) for both the transverse ( dashed red / orange lines ) and longitudinal ( solid blue / cyan lines ) polarizations are shown .
the landau damping linewidths of the coupled plasmons are significantly different in magnitude depending on the band index @xmath96 due to the quartic dependence on the spectrum . whilst the @xmath103 bands show decreasing landau damping with increasing magnitude of the momentum , the opposite is true for the @xmath104 bands .
therefore the center of the brillouin zone displays the greatest disparity in decay rate between each of the two bands per polarization , being either appreciably higher or lower than the result of a single nanoparticle .
the preceding quantitative studies of both the radiative and landau decay rates in secs .
[ sec : radiative ] and sec .
[ sec : landau ] , respectively , allow us to investigate the optimal parameters of the bipartite chain of nanoparticles in order to maximize the lifetime of the collective plasmonic excitations . in fig .
[ fig6b ] , we plot these aforementioned decay rates , scaled by the mie frequency @xmath27 , as a function of both @xmath215 ( @xmath216 denotes the fermi wavevector ) and the scaled momentum @xmath217 .
the parameters used in the figure correspond to a chain of silver nanoparticles with @xmath218 , @xmath219{\aa}$ ] , @xmath220{nm}$ ] , and with interparticle separations @xmath221 and @xmath222 .
the panels ( a)-(d ) in fig .
[ fig6b ] display the competition between the radiative and landau losses for each of the four plasmonic bands : ( a ) the upper transverse band ( @xmath223 , @xmath104 ) , ( b ) the lower transverse band ( @xmath223 , @xmath103 ) , ( c ) the upper longitudinal band ( @xmath224 , @xmath104 ) and ( d ) the lower longitudinal band ( @xmath224 , @xmath104 ) .
the drastic effects of radiative losses scaling with the nanoparticle size as @xmath225 [ see eq . ] can be seen in the bright plasmonic bands , shown in panels ( a ) and ( d ) , in stark contrast to the dark plasmonic bands , shown in panels ( b ) and ( c ) .
the two bright bands also exhibit the striking feature of a complete supression of radiative losses outside the light cone @xmath226 , as was explicitly derived in eq . .
the common feature across all four types of plasmonic band is the quantum size effect of strong landau damping for small nanoparticles , which graphically illustrates the result of eq . and its distinctive @xmath227 scaling .
it may be seen from figs .
[ fig6b](a),(d ) that the optimal size of nanoparticle to probe bright modes of collective plasmonic excitations ( of either polarization ) is of the order of @xmath228 , corresponding to @xmath229{nm}$ ] .
an estimate of the total plasmonic lifetime can be made by considering @xmath230 , where the total damping rate @xmath231 includes the momentum - independent bulk ohmic decay rate @xmath232 . with a typical ohmic damping rate @xmath233 for silver nanoparticles @xcite ,
the bright band plasmonic lifetime is approximately @xmath234{fs}$ ] for @xmath235{nm}$ ] , such that experimental detection is a somewhat challenging task .
however , an enhancement of the lifetime by a factor of two can be achieved by employing metallic nanoparticles with gain in order to compensate for ohmic losses @xcite .
various methods of introducing gain into plasmonic systems exist , including embedding the nanoparticles in a dielectric environment with gain @xcite or by cladding the nanoparticle core with a gain material shell @xcite .
therefore , we conclude that the phenomena discussed in this work are within experimental accessibility , particularly for setups with optimally sized metallic nanoparticles and appropriate gain media .
in this theoretical work , we have analyzed in detail the fundamental properties of collective plasmonic excitations in bipartite chains of spherical metallic nanoparticles .
we have revealed that this system harbours an unusual form of collective excitations : topological bosonic bogoliubov quasiparticles . near the edge of the brillouin zone , we have shown that these quasiparticles are described by a one - dimensional massive dirac hamiltonian .
therefore , we suggest bipartite chains of plasmonic nanoparticles as an ideal testbed to experimentally probe 1d dirac - like bosons , which undergo exotic phenomena such as klein tunneling . whilst our proposal complements the hosting of 2d dirac - like bosons in honeycomb lattices @xcite , it also has the notable advantage that the bipartite linear lattice may be easier to realize experimentally .
having derived the full quantum eigenstates describing the bosonic bogoliubov excitations , we were able to calculate the associated zak phase of the system . this has allowed us to characterize two topologically distinct regimes , depending on the dimerization of the chain , and foreshadowed the presence of plasmonic edge states .
we have studied the various types of topologically - protected edge states that may be formed and derived transparent formulas describing their behavior , which opens up a new avenue to manipulate light at the nanoscale . since combating radiative and nonradiative losses
is a considerable challenge in plasmonics , we have undertaken a detailed study of the decay rates of plasmonic excitations in bipartite nanoparticle chains .
we have provided analytic formulas describing losses in the system due to both radiation damping , which has uncovered the existence of both bright and dark plasmonic bands , and landau damping , a quantum size effect fundamental to the study of quantum metamaterials .
we have therefore been able to estimate the lifetime of the collective plasmonic excitations in the system and hence the optimal parameters required for experimental detection . with the current trend of technological advances in the fabrication of increasing small metallic nanoparticle arrays with gain media ,
the topological and dirac - like phenomena we have discussed should be within experimental reach .
we are grateful to d. weinmann for illuminating discussions and for his careful reading of the manuscript .
we thank t. j. sturges for technical support .
we acknowledge financial support from the cnrs through the pics program ( contract no .
6384 apag ) and from the anr ( grant no . anr-14-ce26 - 0005 q - metamat ) .
in the plasmonic hamiltonian , we consider the interaction between nearest neighbors on the 1d bipartite lattice .
however , as the near - field dipolar interaction decays as one over the cube of the interparticle distance , it is necessary to check the robustness of our results against the inclusion of interactions beyond nearest neighbors only . in this appendix
, we show that the interaction between the nearest neighbors alone captures the relevant physics of the problem and that the plasmon dispersion is only slightly quantitatively modified by interactions beyond nearest neighbors . in the following ,
we evaluate the collective plasmon dispersion classically by taking into account the interaction between an arbitrary number of neighbors on the bipartite lattice .
specifically , we consider the dipole - dipole interaction between a nanoparticle on site @xmath236 belonging to sublattice @xmath204 and dimer @xmath16 up to sites @xmath237 and @xmath238)$ ] with @xmath239 and where @xmath240 @xmath241 for @xmath242 @xmath243 .
the plasmonic dispersion can thus be straightforwardly evaluated to yield the general result @xmath244 with @xmath245 and @xmath246 ^ 3 \mathrm{e}^ { \mathrm{i } ( m-1 ) q d } \right .
\nonumber\\ & + \left .
\left[\frac{a}{(m-1)d+d_2 } \right]^3 \mathrm{e}^ { - \mathrm{i } m q d}\right\}.\end{aligned}\ ] ] , @xmath247 and @xmath248 , corresponding to the longitudinal collective plasmon polarization . ]
we plot the bandstructure in fig .
[ fig7 ] for different values of the number of neighbors taken into account ( for clarity , we show only the longitudinal mode @xmath224 , since similar conclusions can be drawn from the two transverse modes @xmath223 ) .
the nearest neighbor case @xmath249 [ corresponding to eq . ] is shown by solid blue lines , whilst the cases with successively more nearest neighbors included , i.e. , @xmath250 , @xmath251 , and @xmath252 are represented by dashed green , dotted orange and dash - dotted magenta lines , respectively . as can be seen from the figure , dipole - dipole interactions beyond the nearest neighbor contribution
do not lead to significant qualitative changes in the collective plasmon dispersion relation .
quantitatively , when we discuss physics occurring close to the edge of the brillouin zone we should renormalize the resonance frequency by a small amount @xmath253 . for simplicity ,
we thus limit ourselves to the discussion of dipole - dipole interaction effects between nearest neighbors in the main text .
in his appendix we consider the transmission problem of massive dirac - like collective plasmons at the edge of the first brillouin zone between two semi - infinite bipartite chains of metallic nanoparticles , denoted by l ( for left ) and r ( for right ) , and described by two different resonance frequencies @xmath119 , respectively .
therefore , the four associated coupling constants arising from the dipole - dipole interaction @xmath254 are also ( in general ) different . for simplicity we here consider the case where the interparticle spacings @xmath115 and @xmath116 are the same in the left and right chains .
the dirac - like collective plasmons at the edge of the first brillouin zone @xmath255 are described by the spinor , which pick up a label @xmath256 in each section of the bipartite chain , such that the quantity entering eq
. becomes @xmath257 in our proposed setup , a notable contrast to the traditional klein potential step scenario is that the change in on - site energy between the left and right parts of the chain @xmath119 is intrinsically linked to a change in the mass term , as the coupling constants are directly proportional to @xmath119 .
therefore , we effectively deal here with a combined ` electrostatic - mass step ' , meaning in the hamiltonian both the customary diagonal terms @xmath258 and the mass term @xmath259 change as the plasmonic excitation propagates from the left to the right section of the chain . the transmission probability @xmath260 can be calculated using the spinor wavefunctions describing incident , reflected and transmitted plasmonic excitations , respectively . from conservation of energy , the planewaves wavevectors are @xmath261 here @xmath262 is the frequency of the excitation propagating from left to right in a chain with @xmath263 , and @xmath264 arises to ensure conservation of the sign of the group velocity , which is not necessarily aligned with the wavevector .
applying the boundary conditions of continuity of both wavefunction components of the eigenspinors at the domain interface ( @xmath265 ) and solving for the transmission coefficient , we arrive at the transmission probability @xmath266 , @xmath75 , @xmath267 , and @xmath248 .
inset : sketch of the spectrum of the quasiparticles .
the excitations may be transmitted in three regimes ( dotted lines ) corresponding to : ( i ) electron - like to electron - like , ( ii ) electron - like to hole - like and ( iii ) hole - like to hole - like states . ]
we plot in fig .
[ fig8 ] the transmission probability of a longitudinal excitation ( @xmath224 ) as a function of excitation frequency in a chain with @xmath263 .
the problem can be partitioned into three different regimes of interest , as sketched in the inset of fig .
[ fig8 ] . in region
i ( iii ) , above @xmath268 ( below @xmath269 ) the tunneling is from electron ( hole)-like to electron ( hole)-like states .
however , in the interval ii , where @xmath270 , the effective quasiparticle species changes from electron - like to hole - like , thus relying on the klein physics to transmit .
most notably , due to the small effective masses the transmission probabilities are nearly always approaching unity , whilst no transmission takes place in the gapped regions .
in the main text , we consider bipartite chains consisting of an even number of nanoparticles and described by the hamiltonian . while bulk properties do not depend on this restriction ( see secs .
[ sec3 ] and [ sec4 ] ) , the formation of edge states is highly influenced by the parity of the chain . for the sake of completeness , we derive in this appendix expressions for the probability densities at each nanoparticle site corresponding to the midgap state at frequency @xmath27 for the case of a chain comprised of an odd number of nanoparticles . proceeding as in the case of an even number of nanoparticles ( cf .
[ sec : edge_states ] ) but adding an @xmath2 site on the right side of the chain , we now deal with a symmetric tridiagonal matrix @xmath271 of odd order , which is always guaranteed to support one eigenvalue at the midgap frequency @xmath27 @xcite . here
the boundary conditions arising at the edges of the chain are @xmath272 , which together with the recurrence relation for @xmath3 sites imply the complete absence of a midgap plasmonic mode on one half of the nanoparticle sites . on the @xmath2 sites ,
nonvanishing probability densities arise as follows : @xmath273 @xmath274 with @xmath275 $ ] .
thus , highly localized collective plasmons occur at one edge only : either the left [ eq . ] or right [ eq . ]
end of the chain , determined by the critical ratio @xmath276 , as exemplified in fig .
such an outcome can be naturally explained with the aid of the sketches in the inset of of the figure .
as one nanoparticle is always left over when the chain coalesce into dimers , depending on the coupling ratio @xmath276 one can choose the remaining unpaired nanoparticle , which primarily supports the edge state , to be at either end of the chain .
at each site in a chain of 101 nanoparticles , for @xmath136 ( red dots ) and @xmath277 ( blue crosses ) .
the thin gray lines are guides for the eye .
inset : sketches of a typical chain comprised of an odd number of nanoparticles , where the loops depict the favored dimerization for @xmath278 ( top ) and @xmath139 ( bottom ) . ] d. schurig , j. j. mock , b. j. justice , s. a. cummer , j. b. pendry , a. f. starr , and d. r. smith , metamaterial electromagnetic cloak at microwave frequencies , http://dx.doi.org/10.1126/science.1133628[science * 314 * , 977 ( 2006 ) ] . d. j. bergman and m. i. stockman , surface plasmon amplification by stimulated emission of radiation : quantum generation of coherent surface plasmons in nanosystems , http://dx.doi.org/10.1103/physrevlett.90.027402[phys .
lett . * 90 * , 027402 ( 2003 ) ] .
r. f. oulton , v. j. sorger , d. a. genov , d. f. p. pile , and x. zhang , a hybrid plasmonic waveguide for subwavelength confinement and long - range propagation , http://dx.doi.org/10.1038/nphoton.2008.131[nat . photon . * 2 * , 496 ( 2008 ) ] .
m. quinten , a. leitner , j. r. krenn , and f. r. aussenegg , electromagnetic energy transport via linear chains of silver nanoparticles , http://dx.doi.org/10.1364/ol.23.001331[opt . lett . * 23 * , 1331 ( 1998 ) ] .
m. l. brongersma , j. w. hartman , and h. a. atwater , electromagnetic energy transfer and switching in nanoparticle chain arrays below the diffraction limit , http://dx.doi.org/10.1103/physrevb.62.r16356[phys .
b * 62 * , r16356 ( 2000 ) . ]
s. y. park and d. stroud , surface - plasmon dispersion relations in chains of metallic nanoparticles : an exact quasistatic calculation , http://dx.doi.org/10.1103/physrevb.69.125418[phys .
b * 69 * , 125418 ( 2004 ) ] .
j. del pino , j. feist , f. j. garcia - vidal , and j. j. garcia - ripoll , entanglement detection in coupled particle plasmons , http://dx.doi.org/10.1103/physrevlett.112.216805[phys .
lett . * 112 * , 216805 ( 2014 ) ] .
a. brandstetter - kunc , g. weick , c. a. downing , d. weinmann , and r. a. jalabert , nonradiative limitations to plasmon propagation in chains of metallic nanoparticles , phys .
b ( to appear ) , https://arxiv.org/abs/1607.02999[arxiv:1607.02999 ] .
j. r. krenn , a. dereux , j. c. weeber , e. bourillot , y. lacroute , j. p. goudonnet , g. schider , w. gotschy , a. leitner , f. r. aussenegg , and c. girard , squeezing the optical near - field zone by plasmon coupling of metallic nanoparticles , http://dx.doi.org/10.1103/physrevlett.82.2590[phys .
82 * , 2590 ( 1999 ) ] .
s. a. maier , m. l. brongersma , p. g. kik , and h. a. atwater , observation of near - field coupling in metal nanoparticle chains using far - field polarization spectroscopy , http://dx.doi.org/10.1103/physrevb.65.193408[phys .
b * 65 * , 193408 ( 2002 ) ] .
s. a. maier , p. g. kik , h. a. atwater , s. meltzer , e. harel , b. e. koel , and a. a. g. requicha , local detection of electromagnetic energy transport below the diffraction limit in metal nanoparticle plasmon waveguides , http://dx.doi.org/10.1038/nmat852[nat .
* 2 * , 229 ( 2003 ) ] .
d. solis jr .
, b. willingham , s. l. nauert , l. s. slaughter , j. olson , p. swanglap , a. paul , w - s .
chang , and s. link , electromagnetic energy transport in nanoparticle chains via dark plasmon modes , http://dx.doi.org/10.1021/nl2039327[nano lett .
* 12 * , 1349 ( 2012 ) ] .
a. apuzzo , m. fevrier , r. salas - montiel , a. bruyant , a. chelnokov , g. lerondel , b. dagens , and s. blaize , observation of near - field dipolar interactions involved in a metal nanoparticle chain waveguide , http://dx.doi.org/10.1021/nl304164y[nano lett . * 13 * , 1000 ( 2013 ) ] .
s. j. barrow , d. rossouw , a. m. funston , g. a. botton , and p. mulvaney , mapping bright and dark modes in gold nanoparticle chains using electron energy loss spectroscopy , http://dx.doi.org/10.1021/nl5009053[nano lett . * 14 * , 3799 ( 2014 ) ] .
a. poddubny , a. miroshnichenko , a. slobozhanyuk , and y. kivshar , topological majorana states in zigzag chains of plasmonic nanoparticles , http://dx.doi.org/0.1021/ph4000949[acs photonics * 1 * , 101 ( 2014 ) ] . c. w. ling , m. xiao , c. t. chan , s. f. yu , and k. h. fung , topological edge plasmon modes between diatomic chains of plasmonic nanoparticles , http://dx.doi.org/10.1364/oe.23.002021[opt . express * 23 * , 2021 ( 2015 ) ] .
a. kanshu , c. e. rter , d. kip , v. shandarov , p. p. believ , i. ili , and m. stepi , observation of discrete gap solitons in one - dimensional waveguide arrays with alternating spacings and saturable defocusing nonlinearity , http://dx.doi.org/10.1364/ol.37.001253[opt . lett . * 37 * , 1253 ( 2012 ) ] .
c. poli , m. bellec , u. kuhl , f. mortessagne , and h. schomerus , selective enhancement of topologically induced interface states in a dielectric resonator chain , http://dx.doi.org/10.1038/ncomms7710[nat .
* 6 * , 6710 ( 2015 ) ] .
a. p. slobozhanyuk , a. n. poddubny , a. e. miroshnichenko , p. a. belov , and y. s. kivshar , subwavelength topological edge states in optically resonant dielectric structures , http://dx.doi.org/10.1103/physrevlett.114.123901[phys .
114 * , 123901 ( 2015 ) ] .
a. blanco - redondo , i. andonegui , m. j. collins , g. harari , y. lumer , m. c. rechtsman , b. i. eggleton , and m. segev , topological optical waveguiding in silicon and the transition between topological and trivial defect states , http://dx.doi.org/10.1103/physrevlett.116.163901[phys .
lett . * 116 * , 163901 ( 2016 ) ] .
d. d. solnyshkov , a. v. nalitov , and g. malpuech , kibble - zurek mechanism in topologically nontrivial zigzag chains of polariton micropillars , http://dx.doi.org/10.1103/physrevlett.116.046402[phys .
lett . * 116 * , 046402 ( 2016 ) ] .
g. weick , c. woollacott , w. l. barnes , o. hess , and e. mariani , dirac - like plasmons in honeycomb lattices of metallic nanoparticles , http://dx.doi.org/10.1103/physrevlett.110.106801[phys . rev
110 * , 106801 ( 2013 ) ] .
g. weick , r. a. molina , d. weinmann , and r. a. jalabert , lifetime of the first and second collective excitations in metallic nanoparticles , http://dx.doi.org/10.1103/physrevb.72.115410[phys .
b * 72 * , 115410 ( 2005 ) ]
. a. brandstetter - kunc , g. weick , d. weinmann , and r. a. jalabert , decay of dark and bright plasmonic modes in a metallic nanoparticle dimer , http://dx.doi.org/10.1103/physrevb.91.035431[phys .
b * 91 * , 035431 ( 2015 ) ] ; http://dx.doi.org/10.1103/physrevb.92.199906[*92 * , 199906(e ) ( 2015 ) ] .
it is known , e.g. , from studies of crystalline structure with more than one atom per unit cell [ see c. bena and g. montambaux , remarks on the tight - binding model of graphene , http://dx.doi.org/10.1088/1367-2630/11/9/095003[new j. phys . *
11 * , 095003 ( 2009 ) ] ] , that whilst the expectation values of operators describing physical quantities is independent of the basis chosen , the form of the operators may depend on the basis .
here we made this choice of transformation to ensure that we arrive at the most transparent form of both the zak phase , which is dependent on the choice of the unit cell , and also the effective dirac hamiltonian which both arise later in this work .
the very form of the bogoliubov transformation appearing in the hamiltonian exposes the unusual nature of the excitations as a nontrivial mixture of both creation and annihilation operators of the original plasmonic particles .
indeed , the vacuum of the bogoliubov excitations @xmath281 is not the vacuum of the original @xmath282 plasmonic excitations .
instead , @xmath283 involves pairs of @xmath282 excitations which are virtually excited by the dipolar interaction .
r. jackiw and g. semenoff , continuum quantum field theory for a linearly conjugated diatomic polymer with fermion fractionization , http://dx.doi.org/10.1103/physrevlett.50.439[phys .
lett . * 50 * , 439 ( 1983 ) ] .
s. banerjee , j. fransson , a. m. black - schaffer , h. agren , and a. v. balatsky , granular superconductor in a honeycomb lattice as a realization of bosonic dirac material , http://dx.doi.org/10.1103/physrevb.93.134502[phys .
b * 93 * , 134502 ( 2016 ) ] .
g. weick , g .-
ingold , r. a. jalabert , and d. weinmann , surface plasmon in metallic nanoparticles : renormalization effects due to electron - hole excitations , http://dx.doi.org/10.1103/physrevb.74.165421[phys .
b * 74 * , 165421 ( 2006 ) ] .
we note that these bosonic eigenstates are normalized according to an inner product with a metric , @xmath284 , where @xmath285 .
this prescription arises due to the necessity of observing bosonic statistics , see refs .
@xcite .
a method to diagonalize a general bosonic quadratic hamiltonian , which utilizes paraunitary matrices in order to conform to the bosonic commutation relations , is described in refs .
the required paraunitary matrices are obtained by a cholesky decomposition of the bosonic hamiltonian written in bogoliubov-de gennes form .
charl , w. schulze , and b. winter , the size dependent shift of the surface plasmon absorption band of small spherical metal particles , http://dx.doi.org/10.1007/bf01427000[z.phys .
d * 12 * , 471 ( 1989 ) ] .
m. a. noginov , g. zhu , m. bahoura , j. adegoke , c. small , b. a. ritzo , v. p. drachev , and v. m. shalaev , the effect of gain and absorption on surface plasmons in metal nanoparticles , http://dx.doi.org/10.1007/s00340-006-2401-0[appl .
b * 86 * , 455 ( 2007 ) ] .
m. a. noginov , g. zhu , a. m. belgrave , r. bakker , v. m. shalaev , e. e. narimanov , s. stout , e. herz , t. suteewong , and u. wiesner , demonstration of a spaser - based nanolaser , http://dx.doi.org/10.1038/nature08318[nature * 460 * , 1110 ( 2009 ) ] . | we study a bipartite linear chain constituted by spherical metallic nanoparticles , where each nanoparticle supports a localized surface plasmon .
the near - field dipolar interaction between the localized surface plasmons gives rise to collective plasmons , which are extended over the whole nanoparticle array .
we derive analytically the spectrum and the eigenstates of the collective plasmonic excitations . at the edge of the brillouin zone ,
the spectrum is of a pseudo - relativistic nature similar to that present in the electronic band structure of polyacetylene .
we find the effective dirac hamiltonian for the collective plasmons and show that the corresponding spinor eigenstates represent one - dimensional dirac - like massive bosonic excitations .
therefore , the plasmonic lattice exhibits similar effects to those found for electrons in one - dimensional dirac materials , such as the ability for transmission with highly suppressed backscattering due to klein tunnelling .
we also show that the system is governed by a nontrivial zak phase , which predicts the manifestation of edge states in the chain . when two dimerized chains with different topological phases are connected , we find the appearance of the bosonic version of a jackiw - rebbi midgap state .
we further investigate the radiative and nonradiative lifetimes of the collective plasmonic excitations and comment on the challenges for experimental realization of the topological effects found theoretically . |
since the birth of nonlinear optics@xcite as a discipline in physics nonlinear optical processes have been of great interest to scientists , for instance to help describe surfaces and interfaces of condensed matter.@xcite studies of surfaces and interfaces of condensed media belong to the regime of mesoscopic physics , where also quantum wells , -wires and -dots can be found.@xcite among the many nonlinear optical phenomena that has been studied in the regime of mesoscopic physics are ( i ) second harmonic generation from magnetic@xcite as well as nonmagnetic systems , ( ii ) sum- and difference frequency generation , of which one of the most prominent applications today is sisyphus cooling of atoms , ( iii ) photon drag , ( iv ) dc - electric - field induced second harmonic generation,@xcite ( v ) the second - order kerr effect , ( vi ) electronic and vibrational surface raman scattering , ( vii ) two - photon photoemission , ( viii ) generation of higher harmonics,@xcite ( ix ) the second - order lorenz - mie scattering,@xcite and ( x ) degenerate four - wave mixing . in the present communication we study phase conjugation by
degenerate four - wave mixing in a quantum - well structure , where both interband and intraband transitions are allowed .
phase conjugation is a nonlinear process where the response field is counterpropagating to an incoming probe field .
the usual descriptions ( see refs . and references therein ) of degenerate four - wave mixing ( dfwm ) are based on the assumption that the field amplitudes are slowly varying on the optical wavelength scale [ slowly varying envelope ( sve ) approximation ] , and thus also on the electric dipole ( ed ) approximation .
we have previously presented the reasons ( in refs . and ) why these approximations are invalid when considering optical interactions with matter of mesoscopic size , especially when evanescent components of the optical field are present .
four - wave mixing in media with two - dimensional translational invariance has so far been studied by other authors in the context of phase conjugation of electromagnetic surface waves,@xcite and of a bulk wave by surface waves . in these investigations
macroscopic electrodynamic approaches were used . in order to go beyond the sve and ed approximations a nonlocal microscopic theoretical model for optical phase conjugation by dfwm has been constructed ( see ref . ) for nonmagnetic media .
in addition to avoiding the sve and ed approximations , other usually made approximations when considering optical phase conjugation are avoided in our model , namely ( i ) the paraxial approximation , ( ii ) the assumption of a lossless medium , ( iii ) the assumption of a weak probe field , and ( iv ) the requirement of phase matching between the interacting fields . in a two - dimensional translationally invariant
system the change in energy of an electron due to an electric field can either involve a change of momentum along the translationally invariant plane ( intraband transition ) , a change of energy eigenstate perpendicular to the translationally invariant plane ( interband transition ) , or both .
the change of momentum along the translationally invariant plane occurs as an addition of the momentum parallel to the plane of the interacting field component ( denoted by @xmath0 ) to ( or subtraction from ) the momentum of the electron parallel to the surface ( denoted @xmath1 ) .
it is thus convenient to divide the photon momentum @xmath2 into its components parallel and perpendicular to that plane , i.e. , @xmath3 .
then the vacuum dispersion relation @xmath4 provides us with an extra degree of freedom , since @xmath5 can be larger than @xmath6 . using the vacuum dispersion relation we find that @xmath7 becomes imaginary in that case . in the following ,
propagating field components thus shall refer to the case where @xmath7 is a real quantity , and evanescent field components to the case where @xmath7 is a purely imaginary quantity .
if we want to get a broad understanding of the phase conjugated response of a probe containing both propagating and evanescent field components from a quantum - well phase conjugator , two cases are of fundamental interest , namely ( i ) the pure intraband case and ( ii ) the case where also an interband transition is involved .
the phase conjugated response from a pure intraband quantum well we have described in ref . .
this analysis revealed that the phase conjugation reflection coefficient is not only highly nonuniform in the @xmath0-spectrum , but also that the coupling efficiency is several orders of magnitude larger in part of the evanescent regime than in the propagating regime .
since evanescent waves are strongly decaying in space we further concluded that if one wants to see the phase conjugation of evanescent modes , both excitation and observation should take place close to the surface of the quantum well .
furthermore was discussed the problems of excitation of the near - field regime , and the consequences from choosing a broadbanded ( with respect to @xmath0 ) two - dimensional point source ( quantum wire ) revealed that parts of the evanescent spectrum could be excited , and in ref .
that phase conjugation of evanescent fields can lead to a focus of the phase conjugated field substantially below the so - called diffraction limit.@xcite since this has also been experimentally observed,@xcite we judge that it is highly relevant also to give an account of how evanescent fields are phase conjugated in a system where not only one electronic level is present . since including more than one interband transition
will be necessary for most practical applications , we present in this paper the complete solution to the theoretical model of ref . in the case of two - dimensional translational invariance , although a description based on the self - field approximation according to the feibelman theory@xcite would be sufficient in order to determine the dominating response . giving a complete solution also allows us to comment on what we would lose using the self - field approximation .
the solution is based on a discretization in the energy levels of the two - dimensionally tranlational invariant medium .
contrary to discretization schemes performed in real space or fourier space , our discretization does not in itself imply an approximation .
thus , once the complete orthonormal set of wavefunctions for the phase conjugating medium has been determined , the phase conjugated response can in principle be calculated from the solution presented in this paper . how to find the proper set of wavefunctions for a given material system is another problem , which for example can be treated using one of several band - structure methods,@xcite e.g. , the korringa - kohn - rostoker ( kkr),@xcite the linearized augmented plane - wave ( lapw),@xcite or the linear muffin - tin orbital ( lmto ) method.@xcite these methods are based on an atomic description of the potential in a certain radius of each atom , adding exchange- and correlation terms@xcite and different approximations in the regions between the atomic boundaries .
using such a method one will probably be able to give more accurate numerical results for specific materials , but at the cost of the ( relatively ) analytical simplicity .
therefore , we will not elaborate further on this point here , but in stead resort to using a simple description of the matter wavefunctions . doing so
, we will be able to present a qualitative discussion based on analytical expressions .
using a two - level quantum - well phase conjugator , it is also possible to study resonant four - wave mixing , which until now has been studied only without spatial dispersion ( at the point @xmath8 in the @xmath9-@xmath10-plane , @xmath11 being the interband transition frequency ) , as described in , e.g. , refs . .
thus , in sec . [ sec : ii ] we present the theory in the form of a local - field formalism , we choose a scattering geometry , and the solution is presented as a discretization in the energy eigenstates . in sec .
[ sec : iii ] we prepare for a numerical calculation .
we start by adopting the simple infinite barrier ( ib ) model to describe the quantum well .
furthermore we define the phase conjugation reflection coefficient , and the section is concluded with a specific choice of a convenient system to investigate .
to give an impression of the implications of our theoretical model we have presented in sec .
[ sec : iv ] numerical calculations for a two - level quantum - well phase conjugator .
the calculation is supplemented by a discussion of the results , in particular an identification of the different resonances appearing when the wavenumber along the surface plane as well as the frequency varies . in sec .
[ sec : v ] we widen our discussions , with emphasis on ( i ) the interband resonance , ( ii ) the validity of the self - field approximation , and ( iii ) the choice of appropriate relaxation times . finally , in sec .
[ sec : vi ] , we conclude .
as a forerunner for the analysis of the optical phase conjugation from a two - level quantum well we briefly sketch how a local - field calculation allows one to determine the so - called degenerate four - wave mixing response of a mesoscopic metallic film deposited on a dielectric substrate . to create a phase conjugated field , which in the plane of the film propagates in a direction opposite to that of the probe field , two counterpropagating pump fields must be present inside the phase conjugating medium .
although the theoretical model developed in ref . allows us to make almost arbitrary choices of the interacting optical fields , we will in the present work assume for simplicity that the pump fields ( i ) propagate parallel to the plane of the film , and ( ii ) have constant amplitude across the film .
the scattering geometry is shown in fig .
[ fig:1 ] together with the chosen coordinate system .
we will further limit our study to the case where ( iii ) scattering takes place in the @xmath12-@xmath13-plane , and ( iv ) the interacting fields are linearly polarized , either in ( p - polarized ) or perpendicular to ( s - polarized ) the scattering plane . since it is necessary in a study of nonlinear optical phenomena in mesoscopic interaction volumes to abandon macroscopic electrodynamics , the starting point is the microscopic maxwell - lorentz equations . the phase conjugated field from a quantum well exhibiting free - electron - like dynamics in the plane of the well ( @xmath12-@xmath14-plane ) can then be described using the single - coordinate ( @xmath13 ) loop equation@xcite ( 0,0)(85,25 ) ( 0,0 ) ( 2,15)(83,15 ) ( 2,0)(83,10 ) ( 0,7)(30,7 ) ( 0.35,7)(29,7 ) ( 24,7)(0,0)[c]2 ( 85,7)(55,7 ) ( 84.65,7)(56,7 ) ( 61,7)(0,0)[c]1 ( 15,24 ) ( 0,0)(20,0 ) ( 0.25,0)(19.25,0 ) ( 0,-1)(0,0)[tl]probe ( 69.25,0.25)(82.75,3.25 ) ( 82.25,0.25)(-12.5,3)[c]substrate ( 82.75,14.5)(0,0)[tr]quantum well ( 82.75,16)(0,0)[br]vacuum ( 36,0 ) ( 3,10)1.75 ( -0.5,9.875)(15.5,9.875 ) ( 3,25)(3,-1 ) ( 1.5,15)(4.5,15 ) ( 12,8.375)(15.5,9.375)(12,9.375 ) ( -0.5,0.25)(3.625,0.25)(4.625,3)(-0.5,3 ) ( 0,9.875)(15,9.875 ) ( 3,25)(3,0 ) ( 3,10)1.0 ( 3,10)0.25 ( 15,11)(0,0)[rb]@xmath12 ( 2,11)(0,0)[rb]@xmath14 ( 1.75,0.75)(0,0)[rb]@xmath13 ( 2,15)(4,15 ) ( 3.5,15.5)(0,0)[bl]@xmath15 @xmath16 where @xmath10 is the common angular frequency of the participating fields , and @xmath0 is the component of the probe field in the film plane .
it is the so - called background field , @xmath17 , which makes the loop problem different for the various nonlinear ( and linear ) problems .
it is here given by @xmath18 where @xmath19 is the current density driving the nonlinear process .
the pseudo - vacuum propagator @xmath20 is given by @xmath21 + { e^{-{\rm{i}}q_{\perp}(z+z')}\over2{\rm{i}}q_{\perp}}\left [ r^s\vec{e}_{y}\otimes\vec{e}_{y}+r^p\vec{e}_{r}\otimes\vec{e}_{i}\right ] \nonumber\\ & & + { 1\over{}q^2}\delta(z - z')\vec{e}_{z}\otimes\vec{e}_{z } , \label{eq:5}\end{aligned}\ ] ] where the first term describes the direct propagation of the electromagnetic field from a source plane at @xmath22 to the observation plane at @xmath13 , the second term accounts for the reflection at the quantum - well / substrate interface , and the third term characterizes the field generated at the observation plane by the current density prevailing in the same plane ( thus named the self - field term ) . above , @xmath23 , and @xmath24 , taking @xmath25 .
the quantities @xmath26 and @xmath27 are the amplitude reflection coefficients at the vacuum / substrate interface in the absence of the quantum well for s- and p - polarized fields , respectively
. both of these are in general functions of @xmath28 . moreover , the vectors @xmath29 , @xmath30 , and @xmath31 are unit vectors along the principal axes in the cartesian @xmath12-@xmath14-@xmath13-coordinate system , @xmath32 is the heaviside unit step function , and @xmath33 is the dirac delta function .
the @xmath34th tensor element of the linear response tensor @xmath35 , appearing in eq .
( [ eq:1 ] ) , is given by@xcite @xmath36 provided the set of wavefunctions is complete . in eq .
( [ eq:7 ] ) we have introduced the transition current density in the mixed fourier space , namely @xmath37 . \label{eq:8}\end{aligned}\ ] ] in relation to eq .
( [ eq:7 ] ) , @xmath38 is equal to @xmath39 , where @xmath1 is the wavevector of the given electron in the plane of the well .
the transition current density also occurs in the nonlinear response tensor ( see appendix [ app : a ] ) and in this context various combinations of @xmath0 , @xmath40 , and @xmath1 appear in @xmath38 .
the quantities @xmath41 , @xmath42 , are the one - dimensional electronic energy eigenstates of the quantum well belonging to the @xmath13-direction , and they satisfy the field - unperturbed schrdinger equation @xmath43 .
the quantity @xmath44 denote the fermi - dirac distribution for the eigenstate @xmath45 , where also the solution to the schrdinger equation along the quantum well is taken into account .
it is given by @xmath46^{-1}$ ] , where @xmath47 is the boltzmann constant , @xmath48 is the chemical potential of the electron system , and @xmath49 the absolute temperature . for the various cartesian components of the transition current density , we use the notation @xmath50 , @xmath51 .
the complex cyclic transition frequency is defined by @xmath52 -i\tau_{nm}^{-1 } , \label{eq:9}\end{aligned}\ ] ] where @xmath53 and @xmath54 are the eigenenergies of the quantum well states belonging to the @xmath13-direction , and @xmath55 and @xmath56 can be any of the relevant combinations of @xmath0 , @xmath40 , and @xmath1 .
the quantity @xmath57 is the relaxation time .
the nonlinear current density , @xmath19 , is related to the pump and probe fields by a constitutive relation of the form @xmath58 where @xmath59 is the relevant nonlinear response tensor when the pump fields are essentially constant ( slowly varying ) across the quantum well , i.e. , @xmath60 and @xmath61 in eq .
( [ eq:6 ] ) . within the framework of a single - electron random - phase - approximation approach
an explicit expression for @xmath62 has been established in ref . .
the term `` i.t . '' denotes the so - called `` interchanged term '' , which takes into account the symmetry of the pump fields .
it is obtained from the first term by interchanging the two pump fields ( the pump field wavevector @xmath40 is replaced by @xmath63 ) . the explicit expression for the simplified nonlinear conductivity tensor , @xmath64 ,
can be found in appendix [ app : a ] , while the more general conductivity tensor @xmath62 has been given in ref . .
we have , however , in appendix [ app : a ] only listed one of the seven parts , namely part g , of the nonlinear conductivity tensor that appears in ref .
, since when interband transitions are strong , it is dominating the response by several orders of magnitude compared to the other six ( a f ) . as a consequence of the above - mentioned choice ( but independent of the direction in which the pump fields propagate ) the number of terms in the nonvanishing elements of the nonlinear response tensor
is further reduced , since the orthonormality of the @xmath13-dependent parts of the wavefunction gives @xmath65 where @xmath66 is the kronecker delta . also
, by integration of the microscopic transition current density given by eq .
( [ eq:8 ] ) over @xmath13 one finds @xmath67 , \label{eq:38}\ ] ] where @xmath68 is proportional to the @xmath13-component of the electric dipole moment related to the @xmath69-transition.@xcite the conductivity tensor @xmath62 has in general 81 nonzero tensor elements ( 3@xmath703@xmath703@xmath703 ) and consists of seven different parts ( a g ) after the seven different physical processes contributing to the response ( see ref . for details ) .
when scattering takes place in the @xmath12-@xmath13-plane with linearly polarized light the general treatment can be split into eight separate parts related to the possible combinations of polarization of the three different incident fields . in this scattering geometry @xmath0 and @xmath40 lie along the @xmath12-axis , giving a mirror plane at @xmath71 . consequently , only tensor elements in the nonlinar response tensor with a cartesian index even numbered in @xmath14 contributes , and the @xmath72 tensor elements generally appearing are reduced to @xmath73 .
the separation of the tensor elements into the eight sets of elements contributing in these configurations follows in a straight forward manner from the definition of the sum - product operator `` @xmath74 '' between the nonlinear current density and the interacting electric fields , i.e. , @xmath75_{i}= \sum_{jkh}{\xi}_{ijkh}{e}_{h}{e}_{k}{e}^{*}_{j}$ ] .
the added restriction of letting the pump fields travel along the @xmath12-axis then reduces the number of contributing matrix elements from @xmath73 to @xmath76 , since when traveling along the @xmath12-axis , the pump fields are polarized in either the @xmath14-direction or the @xmath13-direction .
the resulting sets of tensor elements we have presented in table [ tab : i ] . to solve eq .
( [ eq:1 ] ) , we can establish a so - called coupled antenna loop .
first , we notice that each matrix element of the linear conductivity tensor [ eq . ( [ eq:7 ] ) ] with the insertion of eq . ( [ eq:8 ] )
can be written as a product of a @xmath13-independent term and two terms depending on @xmath13 and @xmath22 , respectively .
element @xmath34 then takes the form @xmath77 8.6cmcc|@ccc input pol . & output pol . &
nonlinear tensor elements + sss & s & @xmath78 + pps & s & @xmath79 + ssp & p & @xmath80 , @xmath81 , @xmath82 , @xmath83 + ppp & p & @xmath84 , @xmath85 , @xmath86 , @xmath87 + spp , psp & s & @xmath88 , @xmath89 , @xmath90 , @xmath91 + sps , pss & p & @xmath92 , @xmath93 , @xmath94 , @xmath95 + where @xmath96 .
the various @xmath97-quantities can readily be identified from eq .
( [ eq:7 ] ) , and the integrals can be solved using the method described in appendix [ app : b ] .
inserting eq .
( [ eq:10 ] ) into eq .
( [ eq:1 ] ) , we get @xmath98 omitting the reference to @xmath0 and @xmath10 for brevity . in eq .
( [ eq:18 ] ) we have introduced the @xmath99 tensor @xmath100 with the nonzero elements @xmath101 and the elements of the vector @xmath102 are written @xmath103 to determine the phase conjugated field the quantity @xmath102 must be calculated .
this is done by multiplication of each element @xmath104 , @xmath51 of the phase conjugated field in eq .
( [ eq:1 ] ) by the relevant @xmath105 followed by an integration over the @xmath13-coordinate .
hence , when the phase conjugated light is s - polarized , eq .
( [ eq:18 ] ) is transformed into the following set of linear algebraic equations : @xmath106 i.e. , @xmath107 equations with just as many unknowns . in the case of p - polarized light
, we obtain @xmath108 which are @xmath109 equations with just as many unknowns . in eqs .
( [ eq:32])([eq:34 ] ) above , the elements of the vectorial quantity @xmath110 are given by @xmath111 and the @xmath99 tensorial quantity @xmath112 has the five nonzero elements @xmath113 where the indices @xmath114 and `` @xmath115 '' can take the values of @xmath116 . by means of the procedure sketched above ,
we have been able to transform the integral - equation problem for the phase conjugated field , @xmath117 , [ eq . ( [ eq:1 ] ) ] to a matrix problem for the @xmath102-vectors .
this discretization in the energy levels is exact , and once the linear algebraic set of equations for the @xmath102-vectors , truncated so as to keep only the subspace of relevant energy levels , has been solved ( numerically ) the phase conjugated field can be obtained from eq .
( [ eq:18 ] ) .
integral equations of the type given in eq .
( [ eq:1 ] ) is often solved ( numerically ) by discretization in the real space coordinate . by such a procedure
one has to worry about how small discretization lengths one may dare to take from a physical point of view .
the discretization in energy levels used here does not suffer from this uncertainty .
our description of the phase conjugated field has until now been independent of the actual wavefunctions in the quantum well , and thus also independent of the form the potential takes across the active medium .
however , if we want to perform a numerical calculation of the phase conjugated field we have to choose a definite potential across the quantum well , giving us a set of wavefunctions to work with .
below we use the infinite barrier ( ib ) model potential for the numerical study , since this model is sufficient for a qualitative study . as shown in fig .
[ fig:2 ] , in this model the one - dimensional potential @xmath118 is zero inside the quantum well ( in the interval @xmath119 ) and infinite everywhere else .
the stationary state wavefunctions inside the quantum well are given by @xmath120 and outside the quantum well , @xmath121 .
the associated eigenenergies are @xmath122 . within the ib model , eq . ( [ eq:39 ] ) gives @xmath123\over(n^2-m^2)d } \label{eq:40}\ ] ] for @xmath124 , and @xmath125 for @xmath126 . for a metallic quantum
well one may even at room temperature approximate the fermi - dirac distribution functions by their value at zero temperature , i.e. , ( 0,0)(80,60 ) ( 15,50)(15,10)(45,10)(45,50 ) ( 15,51)(30,0)2(0,0)[b]@xmath127 ( 14,20)(46,20 ) ( 10,30)(50,30 ) ( 14,40)(46,40 ) ( 47,20)(0,0)[l]@xmath128 ( 47,40)(0,0)[l]@xmath129 ( 5,10)(5,55 ) ( 4,10)(6,10 ) ( 4,20)(6,20 ) ( 4,30)(6,30 ) ( 4,40)(6,40 ) ( 0,10)(0,0)[l]@xmath130 ( 0,20)(0,0)[l]@xmath131 ( 0,30)(0,0)[l]@xmath132 ( 0,40)(0,0)[l]@xmath133 ( 5,56)(0,0)[b]@xmath134 ( 0,5)(55,5 ) ( 15,4)(15,6 ) ( 45,4)(45,6 ) ( 45,0)(0,0)[b]@xmath130 ( 15,0)(0,0)[b]@xmath135 ( 56,5)(0,0)[l]@xmath13 ( 11.25,36.46)(15,40)(18.75,43.54)(22.5,45)(26.25,43.54)(30,40)(33.75,36.46)(37.5,35)(41.25,36.46)(45,40)(48.75,43.54 ) ( 7.5,23.54)(15,20)(22.5,16.46)(30,15)(37.5,16.46)(45,20)(52.5,23.54 ) ( 65,20)(65,40 ) ( 75,40)(75,20 ) ( 64,20)(76,20 ) ( 64,40)(76,40 ) ( 66,41)(70,43)(74,41 ) ( 66,19)(70,17)(74,19 ) ( 64,30)@xmath136 ( 76,30)@xmath137 ( 70,44)(0,0)[b]@xmath138 ( 70,16)(0,0)[t]@xmath139 @xmath140\right\ } , \label{eq:41}\ ] ] where @xmath141 is the fermi energy of the system . in the low - temperature limit it is possible to find analytical solutions to the integrals over @xmath1 appearing in eq .
( [ eq : a5 ] ) .
the explicit calculations are tedious but trivial to carry out , and since the final expressions are rather long we do not present them here . for the interested reader some steps in the calculations
are reproduced in appendix [ app : b ] .
the fermi energy is calculated from the global charge neutrality condition,@xcite and for a quantum well described by the ib model , it becomes@xcite @xmath142 , \label{eq:42}\ ] ] where @xmath143 is the number of positive ions per unit volume , @xmath144 is the valence of these ions , and @xmath145 is the quantum index of the highest occupied level . from eq .
( [ eq:42 ] ) , the number of occupied levels can be calculated if the thickness is known , and vice versa . the minimal thickness for the quantum well to have @xmath146 levels below the fermi level can be determined from the relation @xmath147 , and the maximal thickness from the condition @xmath148 .
thus for @xmath146 bound states below the fermi energy we find the minimal and maximal thicknesses @xmath149{{\pi{}n\over2zn_+}\left[n^2-{(n+1)(2n+1)\over6}\right ] } , \label{eq:43}\ ] ] i.e. , a result that depends on the number of levels below the fermi energy and the number of conduction electrons in the film .
to estimate the amount of phase conjugated light , we use the phase conjugation ( energy ) reflection coefficient defined as @xmath150 in which @xmath151 , @xmath152 , @xmath153 , and @xmath154 are the intensities of the two pump beams , the probe and the phase conjugated field , respectively . each of the intensities are given by @xmath155 .
the factor of @xmath156 originates from the manner in which we have introduced the fourier amplitudes of the fields . for the remaining part of this work we choose a copper quantum well with @xmath157m@xmath158 and @xmath159 ( data taken from ref . ) . the cu quantum well
is assumed to be deposited on a glass substrate for which we use a refractive index @xmath146 of 1.51 . with this substrate ,
the linear vaccum / substrate amplitude reflection coefficients can be obtained by use of the classical fresnel formulae @xmath160/ [ q_{\perp}+(n^2q^2-q_{\|}^2)^{1/2}]$ ] and @xmath161/[n^2q_{\perp}+(n^2q^2-q_{\|}^2)^{1/2}]$ ] . having the pump wavevectors parallel to the @xmath12-axis then gives a pump wavenumber of @xmath162 .
to calculate the phase conjugated response from a quantum well with an arbitrary number of bound eigenstates one would have to superimpose interband and intraband contributions .
thus in a study of the complete response where local - field effects are neglected one basically would have to add the contributions from the various pairs of levels located in different subbands or in the same band . seen in this light , thorough treatments of the single - level case , where only intraband transitions are allowed , and the two - level case , where transitions between two eigenstates located in different bands occur , would form a good qualitative starting point for analyses of multi - level quantum - well systems .
the single - level case we have studied before,@xcite and the following treatment will thus be directed towards a description of the phase conjugated response from a two - level quantum well .
thus , we choose the simplest possible configuration in which interband transitions can occur , i.e. , a quantum well with only one bound state below the fermi energy . above the fermi energy we also assume that only one bound state can be reached , and thus the wavefunctions are @xmath163 and @xmath164 .
the associated energies then become @xmath165 , and @xmath166 , respectively . the quantum well with the various relevant energies and wavefunctions , as well as the electronic excitations are shown in schematic form in fig . [ fig:2 ] . in the present two - level case eq .
( [ eq:40 ] ) becomes @xmath167 where @xmath168 .
if just the ground state should have an energy less than the fermi energy , we see from eq .
( [ eq:43 ] ) that the film thickness must be less than @xmath169{3\pi/(2zn_+)}$ ] .
the minimal thickness is in the ib model zero , but in reality the smallest thickness is a single monolayer . using eq .
( [ eq:43 ] ) the maximal thickness for a two - level cu quantum well then becomes @xmath170 , which is more than two monolayers and less than three .
thus we have two obvious choices for the thickness of the quantum well , namely a single monolayer or two monolayers .
we choose two monolayers , since by this choice the two energies @xmath131 and @xmath133 are closest to each other , and thus the energy needed for a resonant transition to occur is lowest .
two monolayers of copper roughly corresponds to a thickness of @xmath171 ( bulk value ) . with this choice ,
the energy difference between the two states is @xmath172 ev , and the corresponding resonance in the optical spectrum is found at the wavelength @xmath173 nm . among the eight possible ways of using linearly polarized light in our chosen scattering configuration , two combinations give an s - polarized response when using an s - polarized probe field , the pump fields being either s - polarized or p - polarized , but with the same polarization for both pump fields .
when the pump fields are s - polarized , the nonlinear conductivity tensor element that contributes to the response is @xmath78 . altogether
the phase conjugated response in this purely s - polarized case ( called `` sss '' ) is negligible , since it is tens of orders of magnitude less than those of the other combinations . if , on the other hand , the pump fields are p - polarized ( pps ) , @xmath79 is the element of the nonlinear conductivity tensor that contributes
. plotted as isophotes ( contours of equal intensity ) in the normalized @xmath10-@xmath9-plane ( @xmath10 normalized to the interband transition frequency @xmath11 and @xmath9 normalized to the vacuum wavenumber @xmath6 ) , the result is shown in fig .
[ fig:3 ] .
two other combinations of polarization give p - polarized response using a p - polarized probe field .
as above , the pump fields have to be of the same polarization , and can either be s- or p - polarized . with s - polarized pump fields ( ssp ) , four tensor elements of the nonlinear conductivity tensor
contribute to the phase conjugated response ( see table [ tab : i ] ) .
the phase conjugated response is shown in the normalized @xmath10-@xmath9-plane in fig .
[ fig:4 ] . in the other case ,
another four tensor elements of the nonlinear conductivity tensor contribute to the phase conjugated response when the pump fields are p - polarized ( see table [ tab : i ] ) .
we have in fig . [ fig:5 ]
shown the phase conjugated response for this configuration ( ppp ) in the normalized @xmath10-@xmath9-plane . in the remaining four cases ,
the response has a different polarization than the probe field .
this is obtainable by the use of differently polarized pump fields . in order to achieve an s - polarized response from a p - polarized probe field one
makes use of two differently polarized pump fields , and four tensor elements of the dfwm response tensor contribute to the solution , cf .
table [ tab : i ] .
similarly , two differently polarized pump fields are needed in order to produce a p - polarized response from an s - polarized source . for this process
, another four tensor elements of the nonlinear conductivity tensor contributes according to table [ tab : i ] .
since the resonance structure of these last four cases are similar , it is sufficient here to discuss the result obtained for just one of those cases .
thus , in fig . [ fig:6 ]
the result is shown for the case where pump field @xmath174 is s - polarized and pump field @xmath175 and the probe are p - polarized ( spp ) .
-2 the ib model only offers a crude description of the electronic properties of a quantum well , since , for example , the electron density profile at the ion / vacuum edge is poorly accounted for .
this gives too sharp a profile and underestimates the spill - out of the wavefunction .
altogether one should be careful to put too much reality into the ib model when treating local - field variations ( related to , say , @xmath9 or @xmath7 ) on the atomic length scale .
furthermore , neclecting the bloch character of the wavefunctions accounting for the dynamics in the plane of the well is doubtful in investigations of the local field among the atoms of the quantum well .
the crucial quantity in the above - mentioned context is the fermi wave number @xmath176 , and in relation to figs .
[ fig:3][fig:6 ] , only results for @xmath177 ratios less than approximately @xmath178 appears reliable .
thus we have cut off our results at the line @xmath179 in the @xmath180-@xmath177-plane in figs .
[ fig:3][fig:6 ] . in many theoretical studies of the properties of phase conjugated fields
it is assumed that the phase conjugator is ideal.@xcite by this is meant that the phase conjugation reflection coefficient is independent of the angle of incidense of the ( propagating ) probe field ( and maybe also of the state of polarization ) . as we concluded for the single - level quantum well,@xcite and as we can now see for the two - level quantum well in figs .
[ fig:3][fig:6 ] this assumption is not such a good approximation , at least not for a metallic quantum well system .
looking at figs .
[ fig:3][fig:6 ] , a number of resonances occur .
they can all be accounted for from the analytic solution to eq .
( [ eq : a5 ] ) by looking at the denominators appearing in the analytic decomposition of the products , as given by eqs .
( [ eq : b3 ] ) and ( [ eq : b4 ] ) in appendix [ app : b ] .
these resonances are shown on the scale of figs .
[ fig:3][fig:6 ] in fig .
[ fig:7 ] . in the analytic solution of the integrals over @xmath1 shown in appendix [ app : b ] , the solution to the terms with three multiplied denominators
is reduced in eq .
( [ eq : b4 ] ) to the problem of finding a basic solution to the integrals over @xmath1 for each of these denominators multiplied by a @xmath1-independent factor .
the resulting integrals do not contain sharp resonances , but the factors in front of them do , when @xmath181 , for @xmath182 and @xmath183 . in order to make an analytical treatment of the resonances appearing in the nonlinear conductivity tensor we in the following define a term of the nonlinear conductivity tensor as a product of three denominators in eq .
( [ eq : a5 ] ) , and number them @xmath184
. however , not all terms gives contributions to the result in a two - level quantum well .
the terms that does not give any contributions are the terms with a @xmath185-contribution in the denominator , i.e. , terms 12 and 1112 .
when the denominators of the rest of the terms ( 310 ) are put into the form of eq .
( [ eq : b1 ] ) , a total of four different @xmath186 s and nine different @xmath187 s appear .
they are listed in appendix [ app : c ] .
since we are looking for the location of the resonances in the system it is reasonable in the following analysis to let the respective relaxation times @xmath57 in eqs .
( [ eq : c5])([eq : c13 ] ) be infinite .
( 0,-2)(75,163 ) ( 0,0 ) ( 10,10)[1 ] ( 0,0 ) ( 10,10 ) ( 0.699,0)(0.699,150 ) ( 10,45.369)(70,45.369 ) ( 10,10)(30,0)3 ( 0,0)(0,-2 ) ( 0,-2.5)(0,0)[t]@xmath188 ( 10,10)(30,0)2 ( 0.301,0)(0.301,-1 ) ( 0.477,0)(0.477,-1 ) ( 0.602,0)(0.602,-1 ) ( 0.699,0)(0.699,-1 ) ( 0.778,0)(0.778,-1 ) ( 0.845,0)(0.845,-1 ) ( 0.903,0)(0.903,-1 ) ( 0.954,0)(0.954,-1 ) ( 10,10)(0,30)6 ( 0,0)(-2,0 ) ( -2.5,0)(0,0)[r]@xmath188 ( 10,10)(0,30)5 ( 0,0.301)(-1,0.301 ) ( 0,0.477)(-1,0.477 ) ( 0,0.602)(-1,0.602 ) ( 0,0.699)(-1,0.699 ) ( 0,0.778)(-1,0.778 ) ( 0,0.845)(-1,0.845 ) ( 0,0.903)(-1,0.903 ) ( 0,0.954)(-1,0.954 ) ( 40,0)(0,0)[b]@xmath180 ( 0,85)(0,0)[t]@xmath177 ( 10,147)(70,87)(70,160)(10,160 ) ( 10,10)(70,10)(70,160)(10,160)(10,10 ) ( 5,151.5)(73.5,84 ) ( 0,154)(0,0)[l]@xmath189 ( 5,45.369)(9.5,45.369 ) ( 4,45.369)(0,0)[r]@xmath146 ( 17,20)(0,0)[tl]a ( 15,23)(0,0)[br]b ( 19.5,62)(0,0)[tr]c ( 22,67)(0,0)[bl]d ( 48.5,56)(0,0)[bl]e ( 46.5,33)(0,0)[tl]f ( 30,75)(0,0)[br]g ( 40.5,60)(0,0)[bl]h ( 51,58)(0,0)[bl]i ( 49,30.5)(0,0)[tl]j ( 21,40)(0,0)[bl]k ( 19,52)(0,0)[bl]l ( 27,125.5)(0,0)[tr]m in terms of the @xmath186 s and @xmath187 s listed in appendix [ app : c ] , we observe that the third term of @xmath190 has resonances at ( i ) @xmath191 , ( ii ) @xmath192 , and ( iii ) @xmath193 .
after insertion of the relevant @xmath186 s and @xmath187 s , substitution of @xmath194 in favor of @xmath195 ( since @xmath196 in our treatment ) , and a normalization of @xmath9 to @xmath197 , i.e. , @xmath198 , we may solve the resulting second order equations with respect to @xmath10 as a function of @xmath177 .
then resonance condition ( i ) gives @xmath199 } , \label{eq:49}\end{aligned}\ ] ] resonance condition ( ii ) becomes @xmath200 } , \label{eq:50}\end{aligned}\ ] ] and condition ( iii ) is @xmath201 } .\label{eq:51}\end{aligned}\ ] ] in some of the above equations , some of the solutions can be ruled out immediately , since , for example , in eq .
( [ eq:50 ] ) the minus in front of the square root gives only rise to negative values of @xmath10 in the `` interchanged term '' ( when @xmath202 ) . in the fourth term of
@xmath190 we observe that in addition to a resonance of type ( ii ) , resonances appear at ( iv ) @xmath203 and ( v ) @xmath204 .
again inserting the respective @xmath186 s and @xmath187 s from appendix [ app : c ] , substituting @xmath195 for @xmath194 , and normalizing @xmath9 to the vacuum wavenumber , resonance condition ( iv ) becomes @xmath205 and resonance condition ( v ) is equivalent to eq .
( [ eq:50 ] ) , taking into account the interchanged term . in our configuration , the choice of a two - level quantum well puts some restrictions on the values of the quantum numbers @xmath146 , @xmath206 , @xmath207 , and @xmath208 in order to get a nonzero result .
comparing eqs .
( [ eq : a5 ] ) and ( [ eq:38 ] ) we observe that if pump field one ( indexed @xmath209 ) is s - polarized then @xmath210 , while @xmath211 if it is p - polarized .
similarly , if the other pump field ( indexed @xmath212 ) is s - polarized we get @xmath213 , while we get @xmath214 if it is p - polarized .
these conditions are summarized in table [ tab : ii ] , and the contributions from eqs .
( [ eq:49])([eq:52 ] ) to the resonances in fig .
[ fig:7 ] are shown in table [ tab : iii ] for the valid combinations of quantum numbers . 8.6cmcc|@ccc @xmath209 & @xmath212 & @xmath215 terms 34 & @xmath215 terms 58 & @xmath215 terms 910 + s & s & @xmath216 & @xmath217 & @xmath218 + s & p & @xmath219 & @xmath220 & @xmath221 + p & p & @xmath222 & @xmath223 & @xmath224 + the resonances conditions in the fifth term of @xmath190 are ( vi ) @xmath225 , ( vii ) @xmath226 , and ( viii ) @xmath227 . by insertion of the respective @xmath186 s and @xmath187 s from appendix [ app
: c ] , substitution of @xmath194 by @xmath195 , and normalization of @xmath9 to the vacuum wavenumber , resonance condition ( vi ) becomes @xmath228 } , \label{eq:54}\end{aligned}\ ] ] condition ( vii ) gives @xmath229 } , \label{eq:55}\end{aligned}\ ] ] and case ( viii ) becomes @xmath230}. \label{eq:56}\end{aligned}\ ] ] the sixth term of @xmath190 has a resonance of the type ( vi ) , and further resonances at ( ix ) @xmath231 and ( x ) @xmath232 .
insertion of the different @xmath186 s and @xmath187 s , @xmath196 , and nomalizing @xmath9 to the the vacuum wavenumber gives ( ix ) resonances at @xmath233 } , \label{eq:57}\end{aligned}\ ] ] and ( x ) resonances equivalent to those given in eq .
( [ eq:55 ] ) . in the seventh term of @xmath190
there is a resonances of the type of case ( ix ) , and furthermore at ( xi ) @xmath192 and ( xii ) @xmath234 . as in the previous cases we insert the different @xmath186 s and @xmath187
s found in appendix [ app : c ] , replace @xmath194 with @xmath195 , and nomalize @xmath9 to the the vacuum wavenumber .
then case ( xi ) gives resonances at @xmath235 } , \label{eq:59}\end{aligned}\ ] ] and case ( xii ) the resonances are equivalent to eq .
( [ eq:50 ] ) .
the eighth term of @xmath190 has a resonance of the type given in case ( xi ) , and additional resonances at ( xiii ) @xmath236 and ( xiv ) @xmath237 . repeating the procedure from above ,
we get for case ( xiii ) the solution @xmath238 } , \label{eq:61}\end{aligned}\ ] ] and in case ( xiv ) gives resonances equivalent to the result given in eq .
( [ eq:51 ] ) . again , when considering a two - level quantum well in our configuration , some restrictions apply to the quantum numbers .
if we again compare eqs .
( [ eq : a5 ] ) and ( [ eq:38 ] ) we see that if pump field one ( index @xmath209 ) is s - polarized , then @xmath239 , and if it is p - polarized , then @xmath240 . additionally ,
if pump field two ( index @xmath212 ) is s - polarized , @xmath213 , and if it is p - polarized , @xmath214 .
this has the consequences that ( i ) the quantum numbers @xmath146 and @xmath206 can be chosen arbitrarily when both pump fields are s - polarized , ( ii ) when both pump fields are p - polarized we either get @xmath241 and @xmath210 , or we get @xmath242 and @xmath243 , ( iii ) when pump field one is s - polarized and the other one p - polarized we get either @xmath242 or @xmath210 , and ( iv ) in the opposite case we get either @xmath242 or @xmath241 . these conditions are summarized in table [ tab : ii ] , and the contributions from eqs . ( [ eq:54])([eq:61 ] ) to the resonances in fig .
[ fig:7 ] are shown in table [ tab : iii ] for the valid combinations of quantum numbers .
it should be noted that in eq .
( [ eq:54 ] ) , the combinations of quantum numbers that give rise to the resonances `` b '' , `` e '' , `` i '' , `` h '' , and `` l '' are going into resonance `` m '' after they have reached the line at @xmath244 . none of the other equations contributes to resonance `` m '' . for the ninth term of @xmath190
the resonances are at ( xv ) @xmath245 , ( xvi ) @xmath226 , and ( xvii ) @xmath246 .
after insertion of the relevant @xmath186 s and @xmath187 s from eqs .
( [ eq : c1])([eq : c13 ] ) , @xmath196 and a normalization of @xmath9 to the vacuum wavenumber , we resulting second order equations can be solved with respect to @xmath10 as a function of @xmath177 .
then case ( xv ) is equivalent to eq .
( [ eq:52 ] ) , and cases ( xvi ) and ( xvii ) to eq .
( [ eq:55 ] ) . finally , in the tenth term of @xmath190 a resonance of the type given by case ( xvi ) occur .
two other resonances are located at ( xviii ) @xmath247 and at ( xix ) @xmath248 , respectively . inserting the @xmath186 s and @xmath187 s given in appendix [ app : c ] and
using the same substitution and normalization as above , case ( xviii ) gives .resonances generated by eqs .
( [ eq:49])([eq:66 ] ) are shown as a function of the valid combinations of quantum numbers @xmath249 and the sign appearing in front of the square roots . in each of upper and lower parts of the table ,
the upper row shows the generating equation and the next four rows show the values of the quantum numbers , which can take the value @xmath174 or @xmath175 in a two - level quantum well .
the last four rows show the resonances resulting from use of the quantum numbers in the respective equations for each sign @xmath250 and @xmath251 , the first two of these rows being associated with the normal term , and the last two with the interchanged term .
a zero in the last four rows refers to @xmath252 , and the letters a
l refers to the resonances shown in fig .
[ fig:7 ] .
a `` @xmath253 '' is used when the value of a quantum number is indifferent , and a `` - '' in the output field appears when the result is outside the shown range in fig . [ fig:7 ] . since eq .
( [ eq:52 ] ) is a linear solution in @xmath10 the sign does not apply , and the result is listed under the sign `` + '' for simplicity .
it should be noted that in eq .
( [ eq:54 ] ) , that the combinations of quantum numbers that give rise to the resonances `` b '' , `` e '' , `` i '' , `` h '' , and `` l '' are going into resonance `` m '' after they have reached the line at @xmath244.[tab : iii ] [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] @xmath254 } , \label{eq:66}\end{aligned}\ ] ] and case ( xix ) has a solution equivalent to the one given in eq .
( [ eq:56 ] ) . as before we find by a comparison of eqs .
( [ eq : a5 ] ) and ( [ eq:38 ] ) that some selection rules appear when choosing a two - level quantum well in our configuration , since when pump field one ( indexed @xmath209 ) is s - polarized we get @xmath239 , and when it is p - polarized , @xmath240 . similarly , when pump field two ( indexed @xmath212 ) is s - polarized we get @xmath210 , and when it is p - polarized , @xmath211 . then ,
if both pump fields are s - polarized we may in a two - level quantum well choose @xmath241 or @xmath255 .
in the case where both pump fields are p - polarized , the result is identically zero . in the case where pump field one is s - polarized and pump field
two is p - polarized we may choose either @xmath213 or @xmath242 , while in the opposite case we may choose either @xmath213 or @xmath241 .
as before , these conditions are summarized in table [ tab : ii ] , and the contributions from eq .
( [ eq:66 ] ) to the resonances in fig .
[ fig:7 ] are shown in table [ tab : iii ] for the valid combinations of quantum numbers . in the linear conductivity tensor [ eq .
( [ eq:7 ] ) ] resonances occur when @xmath256 , where @xmath257 the solutions are @xmath258 or @xmath252 , independent of the values of @xmath146 and @xmath206 . adding this resonance to the ones we found in eqs .
( [ eq:49])([eq:66 ] ) the resonances appearing in figs .
[ fig:3][fig:6 ] have been identified .
q. e. d. while most of the resonances described above and shown in fig .
[ fig:7 ] are clearly pronounced in figs .
[ fig:3][fig:6 ] , the resonance named `` m '' does not appear so clearly , although in figs .
[ fig:5 ] and [ fig:6 ] the curves indicate that something is present around the position of `` m '' .
this resonance is striking by the fact that it approaches the fermi wavenumber when the frequency approaches zero .
it might also be appropriate here to mention that the resonances named `` a '' and `` b '' have the asymptotic value of @xmath259 in the low end of the normalized @xmath9-@xmath10-spectrum , and that the resonances named `` c '' and `` d '' approaches @xmath260 for high values of @xmath177 and low values of @xmath180 . the resonance named `` h '' is the interband resonance .
to give an impression of the magnitude of the phase conjugated response , we have in figs . [ fig:3][fig:6 ]
highlighted the isophotes with magnitude of @xmath261m@xmath262/w@xmath263 and @xmath264m@xmath262/w@xmath263 by drawing them with a long - dashed curve and a short - dashed curve , respectively .
their positions in the normalized @xmath9-@xmath10-plane shows quite clearly that most of the area reachable within a single - mode experiment should produce a phase conjugated response of a magnitude comparable to what one gets from second - harmonic generation ( compare refs . ) .
knowing the positions of the resonances in the normalized @xmath9-@xmath10-plane , one could of course be tempted to plot the magnitude of the phase conjugated response along paths following each of the resonances ( e.g. , following the path of resonance `` i '' , and its continuation into `` m '' ) in order to give an improved understanding of the importance of the different resonances . however , since it would be rather difficult in an experiment to follow such a path , and since the exact positions of the resonances probably will be shifted in a practical situation , we have chosen not to do so .
we have in stead in figs .
[ fig:8 ] , [ fig:9 ] , and [ fig:10 ] plotted the intensity of the phase conjugated field along linear cuts in the normalized @xmath10-@xmath9-plane at @xmath265 , @xmath266 , and @xmath267 , respectively .
following the curves in figs .
[ fig:8][fig:10 ] along their respective path on figs .
[ fig:3][fig:6 ] , the appearance and dissapearance of each resonance along the path is easily identified . from figs .
[ fig:8][fig:10 ] it also appears that some of the regions in figs .
[ fig:3 ] and [ fig:4 ] with high density of isophotes are zeros rather than resonances .
( -5,0)(80,95 ) ( -5,2 ) ( 6,50)@xmath268 [ m@xmath262/w@xmath263 ] ( 47,0)(0,0)[b]@xmath180 ( 20,12)(38,27 ) one of the resonances are of special interest , namely the resonance at the interband transition frequency , which experimentally is rather easy to tune into . until now
, resonant four - wave mixing has been studied in other contexts,@xcite but always at the point @xmath270 in the @xmath9-@xmath10-plane . to go beyond that ,
we have plotted the phase conjugated response in the case where the interband transition is resonant ( along the linear path in the normalized @xmath9-@xmath10-plane where @xmath271 ) in fig .
[ fig:11 ] .
( 0,0)(80,90 ) ( -5,2 ) ( 6,50)@xmath268 [ m@xmath262/w@xmath263 ] ( 47,0)(0,0)[b]@xmath180 ( 20,12)(38,27 ) in configurations with only a single source field in the field - matter interaction , such as , e.g. , in linear response , second - harmonic generation , photon drag , and photoemission the so - called self - field approximation has proven to be quite effective . the founding argument to use the self - field approximation is that the dynamics across the quantum well ( in the @xmath13-direction here ) are dominating over motion in the plane of the quantum well ( @xmath12-@xmath14-plane here ) .
let us as a test in the following look at the consequences of applying the self - field approximation in the present case of degenerate four - wave mixing , where three incident fields are present .
working within the framework of the self - field approximation , we observe from eq .
[ eq:5 ] that the phase conjugated response would have been limited to the cases where nonlinear and linear current densities is produced in the @xmath13-direction .
hence , only tensor elements with @xmath272 would contribute .
then , from table [ tab : i ] we observe that the contributions from ( i ) the two cases where the pump fields have the same polarization and the probe field is s - polarized ( sss and pps ) , and ( ii ) the mixed - pump configurations spp and psp would have been neglected .
thus , the data presented in figs .
[ fig:3 ] and [ fig:6 ] would have been absent .
while this is certainly a good aproximation in the pure s - polarized case , the argument is not so good in cases with pump or probe dynamics in the @xmath13-direction . using
the argument of the dominating @xmath13-dynamics , it is striking that the mixed - pump configurations with s - polarized probe field survives the self - field approximation while the two others do not , because we would expect more dynamices in the @xmath13-direction from the latter two .
another interesting conclusion is that with the loss of fig .
[ fig:3 ] we would also lose the resonances named `` j '' , `` k '' , and `` l '' in fig .
[ fig:7 ] . at the same time
we would keep the essentially nonresonant ssp case . comparing the raw amplitudes of the different configurations we can see from figs .
[ fig:3][fig:6 ] and [ fig:8][fig:10 ] that in most regions of the @xmath9-@xmath10-plane , the ppp configuration gives a response that is a few orders of magnitude larger than the other configurations , but we also observe that the three other cases have resonances around @xmath273 , while the ppp - case do not .
thus , for near - normal incidense of the probe , the phase conjugated reflection coefficient is larger for some of the mixed modes than for the pure p - polarized configuration , indeed leaving room for experiments that can not be described within the framework of the self - field approximation .
( -4,0)(80,91.5 ) ( -5,2 ) ( -0,2 ) ( 0,50)@xmath268 [ m@xmath262/w@xmath263 ] ( 47,0)(0,0)[b]@xmath177 ( 69.1,4)(69.1,91.5 ) ( 69,0)(0,0)[b]@xmath274 all in all , we may conclude from the above discussion that although the argument behind the self - field approximation remains intact , when one allows more than one incident field to participate in the interaction ( as in , e.g. , sum- and difference frequency generation , or degenerate four - wave mixing ) , one should be careful in applying the self - field approximation in cases where mixed polarizations of the incident fields are allowed . outside the resonances the influence of the relaxation time is insignificant , but around the resonances the choice of relaxation time has a great influence on the width ( in the @xmath9-space ) and amplitude of each resonance . choosing adequate relaxation times
@xmath57 is a difficult problem and it appears from fig .
[ fig:12 ] how big impact the relaxation time has on the phase conjugation reflection coefficient .
the intraband relaxation time in the occupied state ( @xmath275 ) has been chosen in accordance with ref . to be 3fs . for the unoccupied state the relaxation time @xmath276 ( see fig . [ fig:2 ] ) has been chosen to approach infinity . in the present case where also interband transitions contribute to the phase conjugated response ,
the intraband relaxation time is of little importance , and thus it is the choice of interband relaxation times ( here @xmath277 and @xmath278 ) that are critical . in the present calculation
we assume no relaxation from state @xmath128 to state @xmath129 , letting @xmath279 .
( -4,0)(80,91.5 ) ( -5,2 ) ( -0,2 ) ( 0,50)@xmath268 [ m@xmath262/w@xmath263 ] ( 47,0)(0,0)[b]@xmath177 ( 71.75,4)(71.75,91.5 ) ( 71.75,0)(0,0)[b]@xmath274 the phase conjugation reflection coefficent has in fig .
[ fig:12 ] been plotted for four values of the relaxation time from state @xmath129 to state @xmath128 , namely ( i ) 30fs and ( ii ) 200fs , which are typical values one would find for bulk copper@xcite at ( i ) room temperature and ( ii ) at 77k , ( iii ) 3fs , and ( iv ) 2ps .
the value in case ( iii ) is obtained by a conjecture based on the difference between measured data for a lead quantum well@xcite and the bulk value for lead at room temperature .
the difference between the relaxation times measured by jalochowski , stroak , and zdyb@xcite is for two monolayers approximately one order of magnitude .
case ( iv ) is included to see the effect of raising the value of the relaxation time one order of magnitude , thus essentially assuming a better conductance than in case ( ii ) .
the values ( i)(iii ) are the same values as we chose in our description of the single - level quantum - well case where only intraband transitions were allowed,@xcite but since the interband transition is of a more bulk - like character we have in the present calculations chosen @xmath280fs .
we notice that in the case where both pump fields are s - polarized ( polarized in the plane of the quantum well ) , the phase conjugated response does not vary as a function of the interband relaxation time , whereas in the other three cases the general tendency is that they have larger magnitudes for larger values of the relaxation time .
( -0,0)(80,110 ) ( -5,80 ) ( -5,54 ) ( -5,28 ) ( -5,2 ) ( 6,54)@xmath268 [ m@xmath262/w@xmath263 ] ( 47,0)(0,0)[b]@xmath177 ( 19,90)(0,0)[bl]pps ( 19,64)(0,0)[bl]ssp ( 19,38)(0,0)[bl]ppp ( 19,12)(0,0)[bl]spp ( 55.7,93.2)(77,106 ) ( 71.75,4)(71.75,109.5 ) ( 71.75,0)(0,0)[b]@xmath274
our main conclusion from this work is that dfwm in a thin metallic film gives rise to several resonance structures even in the propagating regime of the @xmath9-spectrum .
furthermore the coupling by the phase conjugation reflection coefficient is of a magnitude that is well within experimental reach .
thus , also single mode excitation in the experimentally feasible regime ( up to around @xmath281 ) should be possible by use of the standard otto@xcite or kretschmann@xcite techniques , and a qualitative comparison with the present work should be possible .
however , for a better quantitative comparison in a specific system , it will be necessary to refine the numerical calculation by , e.g. , abandoning the ib model in favor of one of the flavors of the kkr , lapw or lmto models , although such a task may prove to be strenuous .
[ app : a ] under the assumption that the electron dynamics is free - electron - like in the plane of the quantum well the nonlinear response function @xmath64 [ given by eq .
( [ eq:36 ] ) , and with tensor elements @xmath190 ] can be obtained from the results established for @xmath283 in ref . .
upon integration over @xmath284 and @xmath285 [ and use of eq .
( [ eq:37 ] ) ] one gets @xmath286 \nonumber\\ & & \times \int{}j_{h , ml}(z''';2\vec{\kappa}_{\|}-\vec{k}_{\|})dz ' '' \int{}j_{k , lv}(z'';2\vec{\kappa}_{\|}-\vec{k}_{\|})dz '' j_{j , v{}n}(z';2\vec{\kappa}_{\|}+\vec{q}_{\| } ) \nonumber\\ & & % 2 + \left [ \left({f_{l}(\vec{\kappa}_{\|}-\vec{k}_{\| } ) -f_{m}(\vec{\kappa}_{\|})\over\tilde{\omega}_{l{}m}(\vec{\kappa}_{\| } -\vec{k}_{\|},\vec{\kappa}_{\|})-\omega } + { f_{l}(\vec{\kappa}_{\|}-\vec{k}_{\|})-f_{v}(\vec{\kappa}_{\| } -\vec{k}_{\|}+\vec{q}_{\|})\over \tilde{\omega}_{vl}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\| } , \vec{\kappa}_{\|}-\vec{k}_{\|})+\omega}\right ) { 1\over\tilde{\omega}_{v{}m}(\vec{\kappa}_{\|}-\vec{k}_{\| } + \vec{q}_{\|},\vec{\kappa}_{\| } ) } \right.\nonumber\\ & & \left .
+ \left({f_{l}(\vec{\kappa}_{\|}-\vec{k}_{\| } ) -f_{v}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\|})\over \tilde{\omega}_{vl}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\| } , \vec{\kappa}_{\|}-\vec{k}_{\|})+\omega } + { f_{n}(\vec{\kappa}_{\|}+\vec{q}_{\| } ) -f_{v}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\|})\over \tilde{\omega}_{nv}(\vec{\kappa}_{\|}+\vec{q}_{\|},\vec{\kappa}_{\| } -\vec{k}_{\|}+\vec{q}_{\|})-\omega } \right ) { 1\over\tilde{\omega}_{nl}(\vec{\kappa}_{\| } + \vec{q}_{\|},\vec{\kappa}_{\|}-\vec{k}_{\| } ) } \right ] \nonumber\\ & & \times \int{}j_{h , ml}(z''';2\vec{\kappa}_{\|}-\vec{k}_{\|})dz ' '' \int{}j_{k , v{}n}(z'';2\vec{\kappa}_{\|}-\vec{k}_{\|}+2\vec{q}_{\|})dz '' j_{j , lv}(z';2\vec{\kappa}_{\|}-2\vec{k}_{\|}+\vec{q}_{\| } ) \nonumber\\ & & + \left [ \left({f_{l}(\vec{\kappa}_{\|}+\vec{q}_{\|})-f_{m}(\vec{\kappa}_{\| } ) \over\tilde{\omega}_{l{}m}(\vec{\kappa}_{\|}+\vec{q}_{\| } , \vec{\kappa}_{\|})+\omega } + { f_{l}(\vec{\kappa}_{\|}+\vec{q}_{\|})-f_{v}(\vec{\kappa}_{\| } -\vec{k}_{\|}+\vec{q}_{\|})\over \tilde{\omega}_{vl}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\| } , \vec{\kappa}_{\|}+\vec{q}_{\|})-\omega}\right ) { 1\over\tilde{\omega}_{v{}m}(\vec{\kappa}_{\|}-\vec{k}_{\| } + \vec{q}_{\|},\vec{\kappa}_{\| } ) } \right.\nonumber\\ & & \left
. + \left({f_{l}(\vec{\kappa}_{\|}+\vec{q}_{\| } ) -f_{v}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\|})\over \tilde{\omega}_{vl}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\| } , \vec{\kappa}_{\|}+\vec{q}_{\|})-\omega } + { f_{n}(\vec{\kappa}_{\|}+\vec{q}_{\| } ) -f_{v}(\vec{\kappa}_{\|}-\vec{k}_{\|}+\vec{q}_{\|})\over \tilde{\omega}_{nv}(\vec{\kappa}_{\|}+\vec{q}_{\|},\vec{\kappa}_{\| } -\vec{k}_{\|}+\vec{q}_{\|})-\omega}\right ) { 1\over\tilde{\omega}_{nl}(\vec{\kappa}_{\| } + \vec{q}_{\|},\vec{\kappa}_{\|}+\vec{q}_{\|})-2\omega } \right ] \nonumber\\ & & \times\left .
\int{}j_{h , lv}(z''';2\vec{\kappa}_{\|}-\vec{k}_{\|}+2\vec{q}_{\|})dz ' '' \int{}j_{k , v{}n}(z'';2\vec{\kappa}_{\|}-\vec{k}_{\|}+2\vec{q}_{\|})dz '' j_{j , ml}(z';2\vec{\kappa}_{\|}+\vec{q}_{\| } ) \right\ } j_{i , nm}(z;2\vec{\kappa}_{\|}+\vec{q}_{\| } ) d^2\kappa_{\|}. \label{eq : a5}\end{aligned}\ ] ]
in this appendix we discuss how analytical solutions to the integrals over the electronic wavevector , @xmath1 , appearing in the linear and nonlinear conductivity tensor may be obtained , and for simplicity the discussion is limited to cover the low - temperature limit .
these integrals can , when scattering takes place in the @xmath12-@xmath13-plane , be expressed as a sum over terms of the general type @xmath287 } d\kappa_{x}d\kappa_{y } , \label{eq : b1}\ ] ] where @xmath288 are nonnegative integers , and @xmath197 is an even nonnegative integer .
the functions depends on ( i ) the quantum number @xmath146 , which is a positive nonzero integer , ( ii ) a set of real quantities , @xmath289 appearing in front of the integration variable @xmath290 in the denominator , ( iii ) a set of complex nonzero quantities , @xmath291 appearing also in the denominator , and ( iv ) the real quantity s representing the displacement ( in the @xmath12-direction ) of the center of the fermi - dirac distribution function from @xmath292 .
the quantity s together with each element in the set @xmath293 are in general functions of the parallel components of the probe and pump wavevectors , @xmath0 and @xmath40 .
each element in the set @xmath294 is furthermore a function of @xmath295 , the relaxation time . the combinations of @xmath296 and @xmath197 needed in eq .
( [ eq : b1 ] ) in order to solve the integrals over @xmath1 in the nonlinear conductivity tensor are @xmath297 , @xmath298 , @xmath299 , @xmath300 , @xmath301 , @xmath302 , @xmath303 , @xmath304 , @xmath305 , and @xmath306 .
however , functions with @xmath307 and @xmath308 can be expressed in terms of functions with @xmath309 in the following way : @xmath310 eqs .
( [ eq : b3 ] ) and ( [ eq : b4 ] ) are given with the provision that the values of the different @xmath311 are nonzero , @xmath312 in eq .
( [ eq : b3 ] ) and @xmath313 in eq .
( [ eq : b4 ] ) . if any @xmath311 , for instance @xmath314 , becomes zero , we see from eq .
( [ eq : b1 ] ) that the order ( in @xmath290 ) of the denominator becomes smaller by one .
this implies that @xmath315 in eq .
( [ eq : b3 ] ) . the similar conclusion for eq .
( [ eq : b4 ] ) is @xmath316 .
analogous reductions applies for any other @xmath317 . in the low - temperature limit
the fermi - dirac distribution function is zero outside the fermi sphere and equal to one inside , and it is therefore advantageous to shift @xmath290 by @xmath318 , and afterwards carry out the integrations in polar @xmath319 coordinates . using @xmath320 , @xmath321 , and @xmath322 , the integrals to be solved are of the type @xmath323 dropping the now superfluous index on @xmath186 and @xmath187 .
the upper limit of the radial integration is @xmath324 , @xmath325 . if @xmath326 , the fermi - dirac distribution function is zero , and thus the integral vanishes .
physically , @xmath327 may be characterized as the two - dimensional fermi wavenumber for electrons in subband @xmath146 . since the following treatment is a formal solution of eq .
( [ eq : b5 ] ) , we will also drop the reference to @xmath146 for brevity , letting @xmath328 . to solve eq .
( [ eq : b5 ] ) , let us make the substitutions @xmath329 and thereby turn eq .
( [ eq : b5 ] ) into @xmath330 i.e. , compared to the possible values of @xmath296 and @xmath197 , an expression where the angular integral is expressed as a sum of terms of the form @xmath331 in the nominator , where @xmath332 .
to carry out the angular integrals we put @xmath333 so that the integrals become of the type @xmath334 in eq .
( [ eq : b17 ] ) , the poles at @xmath335 in the @xmath336-plane are located at @xmath337 and the integration runs along the unit circle . since @xmath338
, one of these poles is located inside the unit circle while the other is outside .
when @xmath339 there is an additional pole of order @xmath212 at @xmath340 . using the unit circle as contour ,
residue calculations give the nontrivial solutions @xmath341 , \label{eq : b20}\\ \int_{0}^{2\pi}{\cos^2\theta\over\eta+u\cos\theta}d\theta & = & { 2\pi\eta\over{}u^2}\left[{\eta\over\sqrt{\eta^2-u^2}}-1\right ] , \label{eq : b21}\\ \int_{0}^{2\pi}{\cos^3\theta\over\eta+u\cos\theta}d\theta & = & { \pi\over{}u}+{2\pi\eta^2\over{}u^3 } \left[1-{\eta\over\sqrt{\eta^2-u^2}}\right ] , \label{eq : b22}\\ \int_{0}^{2\pi}{\cos^4\theta\over\eta+u\cos\theta}d\theta & = & { 2\pi\eta^3\over{}u^4}\left[{\eta\over\sqrt{\eta^2-u^2}}-1\right ] -{\pi\eta\over{}u^2}. \label{eq : b23}\end{aligned}\ ] ] to finish the formal solution , ( i ) insert these results into eq .
( [ eq : b7 ] ) , ( ii ) carry out the elementary radial integrations ( see , e.g. , ref . , sec . 2.27 ) , ( iii ) backsubstitute @xmath342 , and ( iv ) check convergence for @xmath343 .
step ( iv ) can be carried out by use of a binomial series expansion of the square roots appearing , and a comparison the the result one gets by setting @xmath344 already in eq .
( [ eq : b5 ] ) .
the solution to the integrals appearing in eqs .
( [ eq:7 ] ) and ( [ eq : a5 ] ) are then found in a straightforward manner , but since the algebraic expressions are rather long , we will omit presenting them here [ they can be found in ref . together with explicit expressions for the case where @xmath344 ]
@xmath346 100 present address : fachbereich physik , freie universitt berlin , arnimallee 14 , d14195 berlin ; and max - planck - institut fr mikrostrukturphysik , weinberg 2 , d06120 halle / saale , germany .
e - mail : [email protected] . c. tomas and o. keller , phys
b ( 1999 ) , to be published .
[ see also c. tomas , _ etude de la corrlation entre emission photolectrique et gnration dharmoniques partir de mtaux _
( universit claude bernard
lyon 1 , lyon , 1998 ) , ph .
d. thesis ] . | the phase conjugated response from nonmagnetic multi - level metallic quantum wells is analyzed and an essentially complete analytical solution is presented and discussed . the description is based on a semi - classical local - field theory for degenerate four - wave mixing in mesoscopic interaction volumes of condensed media developed by the present authors [ t. andersen and o. keller , phys .
scripta * 58 * , 132 ( 1998 ) ] .
the analytical solution is supplemented by a numerical analysis of the phase conjugated response from a two - level quantum well in the case where one level is below the fermi level and the other level is above .
this is the simplest configuration of a quantum well phase conjugator in which the light - matter interaction can be tuned to resonance .
the phase conjugated response is examined in the case where all the scattering takes place in one plane , and linearly polarized light is used in the mixing . in the numerical work we study a two - monolayer thick copper quantum well using the infinite barrier model potential .
our results show that the phase conjugated response from such a quantum - well system is highly dependent on the spatial dispersion of the matter response .
the resonances showing up in the numerical results are analytically identified from the expressions for the linear and nonlinear response tensors .
in addition to the general discussion of the phase conjugated response with varying frequency and parallel component of the wavevector , we present the phase conjugated response in the special case where the light is in resonance with the interband transition . times.sty = 0 2 |
globular clusters have attracted considerable attentions for the understanding of the process of galaxy formation because they are thought to be fossil relics from the early formation history of the galaxy .
it is now well known that the globular clusters we see today may be only the survivors of an initial population . from the calculations of destruction
rate , lee & goodman ( 1995 ) and gnedin & ostriker ( 1997 ) predicted that possibly as many as half of the present day galactic globular clusters would be destroyed in the next hubble time .
two body relaxation makes the velocity distribution of stars in a globular cluster toward a maxwellian . as some stars acquire enough energy to escape from the cluster , the total mass slowly decreases ( spitzer & thuan 1972 ) .
the evaporation is also accelerated by the energy equipartition process which makes all stars have the same kinetic energy .
it causes the mass segregation through which high - mass stars sink toward the central region of the cluster while low - mass stars have higher velocities and tend to occupy the outer region of the cluster ( spitzer 1987 ) .
therefore mass segregation leads to the preferential loss of low - mass stars .
in addition , clusters experience shocks when the external tidal field varies rapidly . passage through the galactic disk and close to the galactic bulge also accelerate the destruction of clusters via gravitational shocks .
gnedin & ostriker ( 1997 ) found that tidal shocks contribute at least as much as two body relaxation to the destruction of the current globular clusters .
the escaped stars may remain in the vicinity of the cluster for several galactic orbits . as a result
, the cluster is expected to have tidal tails .
observational studies about tidal tails of globular clusters have been carried out only recently because of difficulties in wide field photometry .
the signs of the existence of tidal tails around globular clusters were found in many previous studies using radial density profiles or two - dimensional density maps ( grillmair et al .
1995 ; grillmair et al .
1996 ; holland , fahlman , & richer 1997 ; lehmann & scholz 1997 ; leon , bergond , & vallenari 1999 ; leon , meylan , & combes 2000 ; testa et al .
2000 ; odenkirchen et al .
2001 ; sohn et al . 2003
; lee et al . 2003 ; odenkirchen et al .
the studies from radial density profiles suggest that many clusters have weak halo or tails of unbound stars which might result from tidal stripping .
however , two - dimensional density maps obtained in these studies did not clearly confirm this suggestion because these are too complex and diffuse to be regarded as tidal tails .
more convincing evidences for the tidal tails came out of deep and wide ccd observations . recently , using wide field photometric data from the sloan digital sky survey ( sdss ) , odenkirchen et al .
( 2001 , 2003 ) detected clear tidal tails around the galactic globular cluster palomar 5 , and showed that this cluster is being tidally disrupted .
the tails of pal 5 extend over an arc of @xmath2 on the sky , corresponding to a projected length of 4 kpc at the distance of the cluster .
they also suggested that the pal 5 would be destroyed after the next disk crossing , which will happen in about 100 myr . from the wide field photometry using cfh12k mosaic ccd , lee et al .
( 2003 ) showed that the presence of an extra - tidal profile extending out to at least @xmath3 from the center of galactic globular cluster m92 . using cfh12k wide field photometry , sohn et al .
( 2003 ) found very weak tidal halos around the remote young globular clusters pal 3 and pal 4 .
in this paper we study the mass segregation effect and spatial distribution of stars around the galactic globular cluster ngc 7492 using the chf12k @xmath0 photometry data .
ngc 7492 is one of the most sparse globular clusters which is located far from the galactic center and plane ( @xmath4 kpc , @xmath5 kpc ) and at 26.2 kpc from the sun .
the fundamental parameters of ngc 7492 are listed in table 1 .
the first ccd photometry for this cluster was presented by buonanno et al .
they performed bv ccd photometry for ngc 7492 down to @xmath6 roughly 2 mag fainter than the main - sequence turn - off .
ct , richer & fahlman ( 1991 , hearafter crf91 ) carried out ccd photometry for this cluster covering a field of @xmath7 containing cluster center .
they presented a color - magnitude diagram ( cmd ) for ngc 7492 that traces the main sequence to about one magnitude fainter than that of buonanno et al .
they confirmed the cluster metallicity to be [ fe / h ] = @xmath8 , and derived a distance modulus of @xmath9 mag by comparing the fiducial isochrone of ngc 6752 given in vandenberg , bolte , & stetson ( 1990 ) .
they also reported the discovery of 27 blue straggler candidates which are found to be more centrally concentrated than the cluster subgiants of similar brightness .
if the blue stragglers are more massive than the stars at the main sequence turnoff , it would be the result of cluster mass segregation .
they also presented star counts for ten annuli centered on the cluster core .
however , they found no evidence for mass segregation in either the surface density profiles or in the luminosity functions at different radial positions .
possible reasons of failure in finding mass segregation are that the stellar mass range sampled by their data and/or the observed region is too small to find the change of luminosity function ( lf ) or mass function ( mf ) .
the small variation in the lf of ngc 7492 as a function of radius also can be expected from the study by pryor , smith , & mcclure ( 1986 ) .
recently leon , meylan , & combes ( 2000 ) investigated the presence of tidal tails around 20 galactic globular clusters including ngc 7492 . except for a tiny extension pointing toward the galactic center
, they did not find any other signs of tidal tail in ngc 7492 . in this study
, we present the investigation of mass segregation and tidal tails of ngc 7492 using deeper and wider field ccd photometry than that of crf91 . in the following section
, we present the observation and data reduction process . in section 3 , we show cmds of ngc 7492 and describe cmd - mask algorithm which select the sample of cluster member candidates . in section 4 , we examine the mass segregation effect of the cluster using luminosity and mass functions of the cluster . in section 5 , we examine the spatial distribution of stars around the cluster using surface density map and luminosity functions .
final results are summarized in section 6 .
the observations were made with the 3.6 m canada - france - hawaii telescope ( cfht ) during 1999 oct .
17 - 18 using cfh12k , a @xmath10 mosaic of @xmath11 ccds .
the cfh12k has an angular scale of @xmath12/pixel at the @xmath13 prime focus of the cfht , giving a field of view of @xmath14 .
each chip covers an area of @xmath15 .
two adjacent fields were observed using @xmath16 and @xmath17 filters , covering an area of @xmath1 , which extends beyond 3 times larger than the known tidal radius of this cluster , @xmath18 ( harris 1996 ) .
the positions of the observed fields relative to the center of ngc 7492 are shown in figure 1 , and observational information is given in table 2 .
the points in figure 1 represent the stars brighter than @xmath16 = 24 and the point size is proportional to the brightness of stars .
each field was properly dithered to fill in the gaps between chips .
we also obtained frames of fields containing landolt ( 1992 ) standard stars for photometric calibration .
all of the science images were obtained under good seeing conditions ( fwhm of @xmath19 ) . several twilight flat - field , bias and dark frames were also taken .
pre - processing of the raw data , bias and dark subtraction and flat fielding , was done using the fits large images processing software ( flips ) , a very efficient package for the reduction of ccd mosaic images developed by jean - charles cuillandre at cfht ( kalirai et al .
flips is designed to operate on individual chips within the cfh12k mosaic .
first , the good exposures for each of the bias and dark exposures were median combined . for the flat - field images of each filter we averaged and sigma clipped the flats taken from observing run .
median combined bias and dark frames were then subtracted from the flat - field and object images .
after the object images were debiased , the images were flat - fielded to account for pixel - to - pixel variations .
flips normalizes the background sky values to the chip with the highest sky level ( lowest gain ) and provides for a scaled data set with a smooth background on all chips .
this makes the instrumental zero points for each chip almost equal .
the statistics at various positions in the mosaic show that the flat - fielding to be typically better than 1% in both @xmath16 and @xmath17 filters .
instrumental magnitudes of the point sources in the images were derived using the programs daophot ii / allstar ( stetson 1994 ) . for long exposure data ( 600s ) we averaged the images of each filter using the align and imcombred commands within flips .
short exposure images were used to get the instrumental magnitudes for stars saturated in long exposure images .
the analysis is done separately for each chip because the point - spread function ( psf ) is different for each of the 12 ccds on the mosaic .
we used a variable psf to account for small changes in the profiles of stars over the large range of each ccd .
we used the stellarity index of sextractor ( bertin & arnouts 1996 ) for separating stars from background galaxies .
as an example , in figure 2 we show this parameter for sources detected in f1 . those objects with a stellarity index greater than 0.8
were considered to be stars .
we applied a little more strict cut to avoid overestimation of faint stars which could exaggerate the mass segregation effect and tidal extensions of the cluster .
this process also eliminates bad pixels caused by cosmic - ray hits .
after further cuts , based on the two daophot parameters , the magnitude error ( @xmath20 ) and the psf fitting parameter , @xmath21 ( @xmath22 ) , ( we did nt consider sharpness parameter because we already used stellarity index ) were made to eliminate spurious detections , the data sets in each filter were matched and the positions , instrumental magnitudes , and colors were obtained for all the stars . a number of landolt ( 1992 ) standard stars in sa-92 were obtained during the observing run in order to convert the instrumental magnitudes to standard photometric magnitudes .
the transformation equations to derive standard magnitudes are as follows ; @xmath23 @xmath24 in these equations @xmath25 and @xmath26 are instrumental magnitudes and @xmath27 , @xmath28 are the extinction coefficients of the airmass term @xmath29 , @xmath30 , @xmath31 are the transformation coefficient of the color term @xmath32 , and @xmath33 , @xmath34 are the zero - point shift for the @xmath16 and @xmath17 bands .
since we could get only @xmath35 standard stars on each chip , we applied the standard values for the color term and the atmospheric extinction coefficient given on the cfht home page .
separate calibrations are actually required for each chip , because there are systematic differences between chips .
however , the color term does not significantly change with different exposures or nights of observations and the extinction coefficient is very stable at cfht site ( kalirai et al .
we used standard stars to get the zero points of each chip which are the most critical part of the transformation equations .
the instrumental zero points for the chips were almost identical because flips normalizes the background sky value .
the mean value of the differences between our photometry of the landolt ( 1992 ) standard stars and the standard values is 0.01 . to check the accuracy of photometry and calibration we used stars in the dithered frames in each field and overlapped region between the two fields .
the same stars in the different chips and at different locations in the field of view of the instrument showed differences less than limiting magnitude error ( 0.2 ) in all chips .
figure 3 shows the magnitude differences between the same stars in the different chips .
the derived cmd based on the long and short exposure data for the entire observed region of the cluster is shown in figure 4 .
well - defined cluster sequence as well as field stars scattered over the entire cmd can be easily seen . in figure 4 , we can see a well - defined blue horizontal branch ( bhb ) , and the absence of a red horizontal branch ( rhb ) .
also clearly visible are the red giant branch ( rgb ) , subgiant branch ( sgb ) , the main sequence ( ms ) turnoff at @xmath36 mag , and a ms that is extended down to about @xmath37 mag which is roughly 3.5 mag fainter than turnoff .
the characteristics of cmd morphology are described by crf91 in detail . since our main concern is the distribution of cluster stars , the contamination from field stars needs to be removed .
to select cluster stars , we used the technique of the cmd - mask algorithm , which was introduced by grillmair et al .
( 1995 ) . in this way we can differentiate cluster members from the background and foreground field stars by means of comparing stars inside the tidal radius of the cluster with the field stars in a distant part of the field .
the filtering mask is empirically chosen so as to optimize the ratio of cluster stars to field stars in the relatively sparsely populated outer region of the cluster .
we regarded the regions beyond 2.5 times of tidal radius ( @xmath38 ) as background regions which is shown in figure 1 . in figure 5 , we show cmds of the cluster region inside the tidal radius(@xmath39 ) , and outer region ( @xmath40 ) . the procedure was performed in the region @xmath41 and @xmath42 which is indicated by rectangular boxes in figure 5 .
the region was subdivided into a @xmath43 array in which the individual subgrid elements were 0.05 mag wide in @xmath44 and 0.2 mag high in @xmath16 . assuming that the color - magnitude distribution of the field stars is constant across the observed field , a color - magnitude sequence for the cluster can be estimated from ( grillmair et al .
1995 ) : @xmath45 where @xmath46 and @xmath47 refer to the number of stars with color index @xmath48 and magnitude index @xmath49 , counted in an area of circle with radius @xmath50 around the cluster center , and around the background field region , respectively .
the factor @xmath51 is the ratio of the area of the cluster region to that of the background field region
. the signal - to - noise ratio of the expected true number of cluster stars for each subarea was computed from : @xmath52 from the signal - to - noise @xmath53 we obtain a filtering mask by isolating the region in the color - magnitude subgrid with @xmath54 . to select the optimal range of color and magnitude , the elements of @xmath55 were sorted into a gradually descending order over the one - dimensional index @xmath56 first . from the subgrid element with the highest @xmath53 value ,
cumulative star counts were carried out in the cluster region using progressively larger areas of the cmd , @xmath57 , where @xmath58mag@xmath59 is the area of a single element in the color - magnitude array .
then the cumulative signal - to - noise ratio , @xmath60 , can be computed from : @xmath61 where @xmath62 and @xmath63 are the cumulative number of stars in the corresponding subarea in the cluster region and in the outer background field region , respectively .
now , @xmath64 and @xmath65 refer to the number of stars within the cluster region and background field region , having ordered color - magnitude index @xmath56 .
this cumulative function reaches a maximum for a particular subarea of the cmd plane .
based on the peak value of @xmath66 , a threshold value of @xmath67 is determined .
the heavy lines in figure 5 indicate the filtering mask differentiating cluster members from the field stars . by extracting stars outside the filtering mask in the cmd ,
the field contamination was reduced by a factor of @xmath68 .
it is well known that the radial variation of luminosity functions ( lfs ) and mass functions ( mfs ) give a crucial hint of the mass segregation effect in a globular cluster .
lfs were constructed using all cmd - selected stars within the cluster s tidal radius . to build a reliable lf especially in the crowded field
, incompleteness corrections must be applied correctly . to correct for the incompleteness of our photometry , we ran artificial star tests on the @xmath16 frames , using addstar routine in iraf / daophot .
first , we added artificial stars in each 0.5 mag bin randomly on all original images .
the number of added stars was designed not to exceed 10% of the total number of stars that is actually present in that bin so as not to enhance original crowding .
the new frames obtained in this way were then reduced in an identical manner used on the original frame .
figure 6 shows a sample of the plots of the differences between input and output magnitudes for recovered artificial stars selected by the same criteria applied to original image reduction . since most of values do not exceed 0.3 mag , the stars which have magnitude differences smaller than 0.3 mag are considered as recovered stars .
we repeated these procedures 10 times on each chip to obtain meaningful statistics in each magnitude bin .
finally , the incompleteness correction factor @xmath69 is obtained by @xmath70 , where @xmath71 is the number of added stars and @xmath72 is the number of recovered stars . to examine the accuracy of the incompleteness correction , we ran artificial star tests in another way .
we selected two chips in each field of @xmath16 and @xmath17 frames .
then the artificial stars which have similar color band of the cluster ms and rgb stars have been added on the @xmath16 and @xmath17 images with the same method described above .
the new frames were then reduced in an identical manner used on the original frame .
the stars , which have magnitude differences smaller than 0.3 mag and color differences smaller than 0.2 mag are considered as recovered stars .
the incompleteness correction factors obtained from this way were in accord with the previous results within 5% on the same chips .
the incompleteness correction factors depended on the distance from the cluster center and did not show any differences between chips .
the resulting incompleteness correction factors were then applied to the raw lfs . to investigate any spatial variation of lf ,
the lfs for the inner region ( @xmath73 ) and the outer region ( @xmath74 ) are constructed separately .
the regions are divided to contain similar number of stars in each region .
the corrected number of stars in each magnitude bin and the incompleteness factors are given in table 3 , and the corrected lfs are plotted in figure 7 .
the histogram in figure 7 is the lf for all selected stars within tidal radius ( @xmath75 ) .
the solid line overlaid on the inner lf is that of ngc 7492 for @xmath76 derived by crf91 , which agrees very well with ours .
this agreement implies that the incompleteness corrections are properly applied .
the presence of mass segregation be seen from the radial variation of lfs .
the slope of lfs of globular clusters are known to increase toward the outer parts of the cluster where low mass stars preferentially occupy as a result of mass segregation .
however , we could not find any definite evidence from figure 7 alone .
as discussed below , the small variation in the lf of ngc 7492 as a function of radius is not an unexpected result .
the lfs were converted into mass functions ( mfs ) using the mass - luminosity relation of baraffe et al .
( 1997 ) and a distance modulus of @xmath77 ( crf91 ) .
the calculated mfs are shown in figure 8 . due to the small number of stars brighter than the ms turn - off , only stars with @xmath78 mag have been used . filled circles represent the mf for the inner region ( @xmath79 ) while open circles for outer region ( @xmath80 ) .
open triangles represent the mf derived by crf91 . because of our lf extends only to @xmath81 the mf is defined only in the range @xmath82 . for this reason ,
it is difficult to say anything definitely about the slope of the main sequence mf based solely on figure 8 .
crf91 derived a slope of the mf @xmath83 using the theoretical lf from drukier et al .
( 1988 ) . here
@xmath84 represents the slope of a mf of the form @xmath85 we insert lines in figure 8 which show several power law indices covering the slope from crf91 .
although it is difficult to fit a single power law to mfs of this cluster , we can conclude that the @xmath84 value lies within the range from previous study .
it is noteworthy that the cluster has a very flat mf and the variation of the mfs from inner to outer regions is very small . although the slope changes by very small amount , we can see a clear tendency of increase in the slope of mf from inner to outer regions in figure 8 .
it could be a sign of mass segregation in this cluster .
small change in the mf slope does not mean that this cluster is not fully relaxed .
pryor , smith & mcclure ( 1986 ) have investigated the effects of mass segregation on main sequence mfs for globular clusters with various central concentrations .
they found that mass segregation depends on core concentration parameter ( @xmath86 ) and the slope of mass functions ( @xmath84 ) .
they showed that the mass segregation effect appears more clearly in models with large concentrations and steep ( larger @xmath84 ) mass functions .
for a fully relaxed cluster having @xmath87 and @xmath88 , there will be very little change in the mfs due to mass segregation . using the hst photometry ,
grillmair & smith ( 2001 ) investigated the main - sequence lf of pal 5 . according to their results , pal 5 has a very flat mf ( @xmath89 ) and shows very little evidence for mass differentiation between the core of the cluster and the half - mass radius .
they suggested that pal 5 has lost a large fraction of its original stellar content as a result of tidal shocking .
as mentioned in section 1 , pal 5 appears to be in the final phase of tidal disruption ( odenkirchen et al .
2003 ) . in any globular cluster
the form of the present day mass function depends on the initial mass function and the process of dynamical evolution of the cluster .
mcclure et al . ( 1986 ) suggested that the slope of mf is related with metallicity of cluster .
they proposed that the observed strong dependence of mf slopes on metallicity reflects , at least in part , properties of the initial mass functions with which the clusters formed .
later , djorgovski , piotto , & capaccioli ( 1993 ) demonstrated that the mf slopes are determined not only by the metallicity but also by the location in the galaxy .
they showed that the clusters closer to the galactic center have flatter mfs . at a given galactocentric distance ( @xmath90 ) , clusters with a smaller distance from the galactic plane ( @xmath91 )
have flatter mfs , and at a given position , clusters with lower metallicity have steeper mfs . from n - body simulations , capaccioli , piotto , & stiavelli ( 1993 ) interpreted the dependence on position as the effect of tidal shocks .
disk and bulge shocking acts preferentially on stars located in the outer region of the cluster and leads to a loss of low mass stars . as a result
the clusters near the galactic center and plane , tend to have flatter mfs .
since ngc 7492 has intermediate metallicity ( [ fe / h ] = @xmath8 ) among the galactic globular clusters and large distant from galactic center and plane ( @xmath4 kpc ; @xmath5 kpc ) , flat mf of ngc 7492 can not be interpreted neither by its metallicity nor by galactic position .
however , dynamical evolution of a globular cluster can not be determined only by present position but also by cluster s orbit .
piotto , cool , & king ( 1997 ) presented a comparison of deep hst lfs of four galactic globular clusters .
they showed that three of four clusters ( m15 , m30 , and m92 ) have nearly identical lfs , whereas ngc 6397 has a distinctly different lf , especially in the fainter part .
they suggested that the three globular clusters which have similar lfs were formed with similar mfs and suffered very little changes ( or experienced similar changes regardless of the location ) , and that the deficiency of low - mass stars in ngc 6397 was due to tidal shocks , stellar evaporation through internal relaxation , or a combination of the two .
relatively flat mf of ngc 6397 could be explained by its orbit given by dauphle et al .
( 1996 ) which is very vulnerable to tidal shocks .
later , comparing mf slopes for 7 galactic globular clusters piotto & zoccali ( 1999 ) suggested that the flattening of mf slopes might be related to the cluster s dynamical evolution . from the n - body simulations ,
baumgardt & makino ( 2003 ) predicted that the slope of mfs of the clusters should decrease constantly as the clusters evolve , and there is a good correlation between the slopes and the cluster lifetimes .
if the flat mf of ngc 7492 is not primordial it might be a result of dynamical evolution of this cluster . when the galactic orbit of ngc 7492 is determined the influence of galactic tidal shocks on this cluster
could be known .
in order to characterize the distribution of stars around the cluster , we constructed surface number density map . the sample of cluster member candidates obtained by cmd - mask process in the previous section was used to construct surface density map .
initially , the candidate cluster stars in the surveyed region were binned on a grid of @xmath92 . unfortunately
, our surveyed region is not wide enough to get proper background maps .
however , we used field stars having color greater than 0.9 to check the gradient of distribution of field stars and possible photometric biases in the entire surveyed region . after constructing the map of field stars by binning @xmath93 , we smoothed it to get a surface map dominated only by the smooth gradient of field stars .
we subtracted this field star map from the cmd - selected one .
we then convolved the map with a gaussian kernel of width @xmath94 .
the resulting smoothed surface density map is shown in figure 9 .
we overlaid contour levels and marked the tidal radius with a thick circle .
the arrow indicates the direction to the galactic center .
we can see a marginal extension toward the galactic anticenter ( northeast ) in figure 9 .
also there seems to be a small extension to the northwest from the cluster center .
many n - body simulations of globular cluster tidal tails show that the stars evaporated from the cluster form a twisted , two - sided lobe distribution of cluster member stars ( combes , leon , & meylan 1999 ; yim & lee 2002 ; lee et al .
2004 ) . however , the shape of such lobes depends on the orientation of the cluster s orbit to our line of sight .
combes , leon , & meylan ( 1999 ) showed that the lobes can be asymmetric and even seen one - sided according to the cluster s orbits and projection effects in their figure 13 and figure 15 .
the extensions shown in figure 9 can not be regarded as a real feature solely based on the surface density map .
we will check the significance of the extensions using luminosity functions in the following subsection . from figure 9
, we can see that the cluster ngc 7492 has very flattend shape .
white & shawl ( 1987 ) have measured the projected axial ratio @xmath95 of 100 galactic globular clusters , where @xmath96 and @xmath97 denote semi - major and semi - minor radii , respectively .
they obtained the value of @xmath98 from ngc 7492 while mean axial ratio was @xmath99 .
we used ellipse task in iraf / stsdas to fit the elliptical shape of the cluster .
figure 10 shows the axial ratio of the cluster as a function of the semi - major axes of the ellipses .
the axial ratio decreased and reached to a value of 0.77 .
white & shawl ( 1987 ) argued that the flattened shape of the globular clusters can be caused by either anisotropy in velocity dispersion or rotation .
the tidally truncated clusters tends to become quickly isotropic ( takahashi & lee 2000 ) , the flattening is likely due to rotation ( combes , leon & meylan 1999 ) .
kinematical data also showed that the flattening could be explained by rotation , and that the minor axes are nearly coincident with the rotation axes ( meylan & mayor 1986 ) .
the initial angular momentum of a cluster is expected to disappear with time by the escape of stars carrying some angular momentum .
kim et al . ( 2002 ) showed using the fokker - plank method that the initial rotation decreases with time after core - collapse and finally disappear . as a result
, the flattening of a globular cluster could be related with its dynamical age . from a compilation of data on globular cluster flattenings in our galaxy and m31 ,
davoust & prugniel ( 1990 ) showed that globular clusters with shorter relaxation times tend to be rounder .
they suggested a scenario which globular clusters are initially flattened and become rounder as they lose stars and angular momentum . according to this scenario ,
ngc 7492 would be the one which has the youngest dynamical age among the galactic globular clusters .
however considering the flat mf of ngc 7492 , it is unlikely that this cluster maintain initial dynamical status .
if the flattening of ngc 7492 can not be explained only by initial rotation of this cluster , we should consider some alternative sources of globular cluster rotation .
one possible alternative is tidal interaction with the galaxy . from n - body simulations
lee et al .
( 2004 ) showed that the globular clusters gain angular momentum from tidal interactions with host galaxy .
if a cluster rotates by the tidal interaction with the galaxy , the orientation of the rotation axis would be the same as its revolution axis .
when we get the orbit of ngc 7492 , the influence of tidal interaction on its rotation could be examined .
figure 9 shows marginal extensions of stellar distributions around ngc 7492 toward northeast and northwest from the cluster center .
however we can not regard the extensions shown in figure 9 as a real feature using only the surface density map .
we therefore analyzed the lfs of the stars around the clusters to check the significance of the extensions . in figure 11 , we show the cumulative lf for stars ( solid line ) located in outer northern part ( @xmath100 ) where extensions appear . the dashed line indicates the lf for stars within the tidal radius ( @xmath101 ) of ngc 7492 .
for comparison , we also show the lf for stars in southern part of the cluster as a dotted line .
these lfs are constructed using cmd - selected stars .
incompleteness corrections are properly applied to each lf as described in the previous section .
the lfs are normalized by the total number of stars for each selected group .
a kolmogorov - smirnov ( k - s ) test of the lfs for stars in the outer northern part ( solid line ) compared to that for stars within the tidal radius ( dashed line ) gives a significance level of 75% probability that both cumulative lfs follow the same distribution .
on the other hand , the k - s significance level of cumulative lf for stars in the southern part ( dotted line ) is less than 1% . since mass segregation effect of this cluster
is very small , we can expect that the lf for stars in the tidal tails of the cluster would be almost the same with that for stars within the tidal radius in the limit of our analysis ( grillmair & smith 2001 ) .
odenkirchen ( 2003 ) also showed that the lf for the stars in the tidal tails of pal 5 is in very good agreement with the lf in the cluster .
theses indicate that the extensions shown in the surface density map ( figure 9 ) could be a real feature .
we have carried out a wide field ccd photometry of ngc 7492 and investigated the dynamical structure of the cluster .
the observations with the cfh12k mosaic ccd cover the area of @xmath102 , about 3 times larger than the known tidal radius of this cluster .
we used the technique of the cmd - mask algorithm to select cluster member star candidates , and we have used these stars to examine the characteristics of the spatial distribution of stars around the clusters . the mass segregation effect and stellar distribution around the cluster
have been investigated by means of a completeness - corrected lfs and mfs and the surface density map .
the slope of cluster main sequence mf slightly increases toward the outer part of the cluster , as expected from mass segregation .
it could be a verification of mass segregation effect of this cluster which was suggested doubtfully only from the distribution of blue stragglers by crf91 .
although the change in the mf slope is very small it does not mean that this cluster is not fully relaxed .
the relatively flat mf slope of ngc 7492 and its small mass might be a result of dynamical evolution of this cluster .
surface density map shows possible extensions of spatial stellar distributions beyond the tidal radius of ngc 7492 .
the extensions of the tails of ngc 7492 are oriented toward the galactic anticenter ( northeast ) and northwest from the cluster center .
a comparison of lf for stars in the tails with that for stars within the tidal radius suggests that the extensions shown in the surface density map could be a real feature .
surface density map also shows that the ngc 7492 has very flat shape .
the axial ratio ( @xmath103 ) of this cluster decrease with semi - major axis and reached to a value of 0.77 which is one of the smallest values the galactic globular clusters have .
if the flattened shape of this cluster is caused by its rotation , it might have been much affected by galactic tidal field .
pryor et al . ( 1991 ) argued that low concentration ( low-@xmath86 ) clusters lost much of their original stellar mass over extended periods of time through the evaporation and stripping of stars .
recently , odenkirchen et al .
( 2003 ) showed clear evidences that the low - mass , low-@xmath86 galactic halo cluster pal 5 is being tidally disrupted .
in addition to previous studies , flat mass function , tidal extensions , and flat shape of the cluster which are investigated in this study suggest that ngc 7492 might have experienced a lot of dynamical processes through tidal interaction with the galaxy .
however , kinematic data of ngc 7492 are not obtained yet .
the information on the orbits of individual halo clusters can produce powerful constraints on the galactic potential .
we need spectroscopic observations and proper motion studies about ngc 7492 to get the information about the rotation and the galactic orbit of the cluster .
we also encourage deeper and wider ccd photometry of this cluster to find possible tidal tails extending over the coverage of this study .
hml was supported by the kosef grant no .
r01 - 1998 - 00023 .
h.s . acknowledges the support of the korea science and engineering foundation ( kosef ) to the astrophysical research center for the structure and evolution of the cosmos ( arcsec@xmath104 ) at sejong university .
baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. , 1997 , a&a , 327 , 1054 baumgardt , h. , & makino , j. , 2003 , mnras , 340 , 227 bertin , e. , & arnouts , s. , 1996,a&as , 117 , 393 buonanno , r. , corsi , c. e. , ferraro , i. , & fusci pecci , f. , 1987 , a&as , 67 , 327 capaccioli , m. , piotto , g. , & stiavelli , m. , 1993 , mnras , 261 , 819 combes f. , leon s. , & meylan g. , 1999 , a&a , 352 , 149 ct , p. , richer , h. b. , & fahlman , g. g. , 1991 , aj,102 , dauphole , b. , geffert , m. , colin , j. , ducourant , c. , odenkirchen , m. , & tucholke , h. -j . , 1996 , a&a , 304 , 202 davoust , e. , & prugniel , p. , 1990
, a&a , 230 , 67 djorgovski , s. , piotto , g. , & capaccioli , m. , 1993 , aj , 105 , 2148 drukier , g. a. , fahlman , g. g. , richer , h. b. , & vandenberg , d. a. , 1988 , apj , 95 , 1415 gnedin , o. y. , & ostriker , j. p. , 1997 ,
apj , 474 , 223 grillmair , c. j. , freeman , k. c. , irwin , m. , & quinn , p. j. , 1995 , aj , 109 , 2553 grillmair , c. j. , ajhar , e. a. , faber , s. m. , baum , w. a. , holtzman , j. a. , lauer , t. r. , lynds , c. r. , & oneil , e. j , 1996 , aj , 111 , 2293 grillmair , c. j. , & smith , g. h. , 2001 , aj , 122 , 3231 harris w. e. , 1996 , aj , 112 , 1487 holland s. , fahlman g. g. , & richer h. b. , 1997 , aj , 114 , 1488 kalirai , j. s. , richer , h. b. , fahlman , g. g. , cuillandre , j .- c . , ventura , p. , dantona , f. , bertin , e. , marconi , g. , & durrell , p. r. , 2001 , aj , 122 , 257 kim , e. , einsel , c. , lee , h. m. , spurzem , r. , & lee , m. g. 2002 , mnras , 334 , 310 landolt , a. u. , 1992 , aj , 104 , 340 lee , h. m. , & googman , g. , 1995 , apj , 443 , 109 lee , k. h. , lee , h. m. , fahlman , g. g. , & lee , m. g. , 2003 , aj , 126 , 815 lee , k. h. , lee , h. m. , yim , k. j. , & sung , h. , 2004 , in preparation lehmann i. , & scholz r. -d . , 1997 , a&a , 320 , 776 leon , s. , bergond , g. , & vallenari , a. , 1999 , a&a , 344 , 450 leon , s. , meylan , g. , & combes , f. 2000 , a&a , 359 , 907 mcclure , r. d. , vandenberg , d. a. , smith , g. h. , fahlman , g. g. , richer , h. b. , hesser , j. e. , harris , w. e. , stetson , p. b. , & bell , r. a. , 1986 , apj , 307 , l49 meylan , g. , & meyor , m. , 1986 , a&a , 166 , 122 odenkirchen m. , grebel e. k. , rockosi c. m. , dehnen w. , ibata r. , rix h. w. , stolte a. , wolf c. , anderson , jr
. j. e. , bahcall n. a. , brinkmann j. , csabai i.,hennessy g. , hindsley r. b. , ivezic z. , lupton r. h. , munn j. a. , pier j. r. , stoughton c. , & york d. g. , 2001 , apj , 548 , l165 odenkirchen , m. , grebel , e. k. , dehnen , w. , rix , h .- w . , yanny , b. , newberg , h. j. , rockosi , c. m. , martinez - delgado , d. , brinkmann , j. , & pier , j. r. , 2003 , aj , 126 , 2385 piotto , g. , cool , a. m. , & king , i. r. , 1997 , aj , 113 , 1345 piotto , g. , & zoccali , m. , 1999 , a&a , 345 , 485 pryor , c. , smith , g. h. , & mcclure , r. d. , 1986 , aj , 92 , 1358 pryor , c. , mcclure , r. d. , fletcher , j. m. , & hesser , j. e. , 1991 , aj , 102 , 1026 sohn , y .-
j , park , j .- h . ,
rey , s .- c .
, kim , h .-
, oh , s. j. , lee , s .-
, lee , m. g. , & han , w. , 2003 , aj , 126 , 803 spitzer , l. , 1987 , dynamical evolution of globular clusters ( princeton : princeton univ . press ) spitzer , l. , & thuan , t. x. , 1972 , apj , 175 , 31 stetson , p. b. , 1994 , pasp , 106 , 250 takahashi , k. , & lee , h. m. , 2000 , mnras , 316 , 671 testa v. , zaggia s. r. , andreon g. l. , scaramella r. , djorgovski s. g. , & de carvalho r. , 2000 , a&a , 356 , 127 vandenberg , d. a. , bolte , m. , & stetson , p. b. , 1990 , aj , 100 , 445 white , r. e. , & shawl , s. j. , 1987 , apj , 317 , 246 yim , k. j. , & lee , h. m. , 2002 , j. korean astron .
soc . , 35 ,
75 @xmath105 & @xmath106 & harris ( 1996 ) + @xmath107 & @xmath108 & harris ( 1996 ) + @xmath109(deg ) & 53.39 & harris ( 1996 ) + @xmath110(deg ) & -63.48 & harris ( 1996 ) + log @xmath111 ( @xmath112 /pc@xmath113 ) & 0.97 & harris ( 1996 ) + @xmath101(arcmin ) & 8.35 & harris ( 1996 ) + @xmath86 = log @xmath114 & 1.00 & harris ( 1996 ) + [ fe / h ] & -1.51 & crf91 + @xmath115 & 17.09 & crf91 + @xmath116 & 0.00 & crf91 + f1 & @xmath117s & oct .
17 1999 & @xmath118s & oct .
17 1999 + f1 & @xmath119s & oct .
17 1999 & @xmath119s & oct .
17 1999 + f1 & @xmath120s & oct .
17 1999 & @xmath120s & oct .
17 1999 + f2 & @xmath118s & oct .
18 1999 & @xmath118s & oct .
18 1999 + f2 & @xmath119s & oct .
18 1999 & @xmath119s & oct .
18 1999 + f2 & @xmath121s & oct .
18 1999 & @xmath121s & oct .
18 1999 + 15.5 - 16.5 & 4.0 & 1.0 & 3.0 & 1.0 & 7.0 & 1.0 + 16.5 - 17.5 & 10.0 & 1.0 & 12.0 & 1.0 & 22.0 & 1.0 + 17.5 - 18.5 & 32.0 & 1.0 & 26.0 & 1.0 & 58.0 & 1.0 + 18.5 - 19.5 & 28.0 & 1.0 & 22.0 & 1.0 & 50.0 & 1.0 + 19.5 - 20.5 & 73.2 & 0.99 & 49.5 & 0.99 & 122.7 & 0.98 + 20.5 - 21.5 & 490.0 & 0.90 & 310.4 & 0.96 & 800.4 & 0.94 + 21.5 - 22.5 & 963.3 & 0.79 & 669.2 & 0.91 & 1632.5 & 0.87 + 22.5 - 23.5 & 1263.8 & 0.69 & 943.3 & 0.90 & 2207.1 & 0.80 + 23.5 - 24.5 & 1274.4 & 0.43 & 1068.6 & 0.70 & 2343.0 & 0.58 + | we present a wide field ccd photometric study of galactic globular cluster ngc 7492 .
the derived @xmath0 color - magnitude diagram extends down to about 3.5 mag below the cluster main sequence turn - off .
the field covers @xmath1 about 3 times larger than the known tidal radius of this cluster .
the sample of cluster member candidates obtained by cmd - mask process has been used to construct luminosity and mass functions and surface density map .
ngc 7492 has a very flat mass function with very little variation in the slope with the distance from the cluster center .
however , there is a clear evidence for the increase of the slope of the mass function from inner to outer regions , indicating a mass segregation of the cluster .
the surface density map of ngc 7492 shows extensions toward the galactic anticenter ( northeast ) and northwest from the cluster center .
a comparison of luminosity function for stars in the tails with that for stars within the tidal radius suggests that the extensions shown in the surface density map could be a real feature .
the overall shape of ngc 7492 is significantly flattened .
if the flattened shape of the ngc 7492 is caused by its rotation , galactic tidal field must have given important influences , since the initial rotation would have been almost completely removed by dynamical relaxation . |
polar ring galaxies ( prg ) are composed of an early - type host , surrounded by a gaseous and stellar perpendicular ring which has the characteristic of a late - type galaxy : large hi amount , young stars , blue colors .
if the polar ring is in equilibrium , it could provide clues on the 3d - shape of dark matter in the host galaxy .
for that , the polar ring must not be too massive , in order to keep the initial gravitational potential unperturbed ; or in any case , we must know the formation mechanism of the polar rings , to be able to generalise the results to normal haloes .
the probability to have a prg has been estimated to @xmath0 5% from the observations ( whitmore et al 1990 ) .
several mechanisms have been proposed in the literature to form polar ring galaxies , that can be gathered into two kinds : merger between two galaxies , of which one at least is gas rich , or mass accretion by the original host galaxy .
the first kind can be a major merger between two perpendicularly oriented systems , or the disruption of a small companion on a polar orbit ( the debris align then on the polar ring ) .
the second kind can be either gas transfer between two galaxies , in an hyperbolic passage , or gas accretion from cosmic filaments in a polar ring by the host . in all cases , the bulk of the stars in the polar plane form after the event , from the gas settled in the polar plane .
the collision scenario has first been simulated by bekki ( 1997 , 1998 ) .
bournaud & combes ( 2003 ) have explored the geometrical parameters of the collision to determine the stability of the final system , and the actual probability to form a prg from such a merger .
the mechanim is perfectly able to form polar rings , and even double rings ( see the eso 474-g26 system in figure 1 , reshetnikov et al 2005a ) .
however the parameter space is smaller than for accretion , so the majority of prgs must not have been formed through mergers . [ cols="^,^ " , ] the tidal accretion scenario , proposed by schweizer et al ( 1983 ) and first simulated by reshetnikov et al ( 1997 ) , was long thought to form only very light polar rings , since the donor galaxy exchange only part of its mass , before leaving the interaction scene . when the geometrical parameters are explored , they are able to form quite inclined polar ring , such as the prototype ngc 660 , which makes the continuity with highly warped systems ( see also ngc 3718 , krips et al 2005 ) . from the statistics of initial conditions ,
the accretion scenario was estimated to be 3 - 5 times more likely to form polar rings ( bournaud & combes 2003 ) . the configuration of particular systems tends to orient towards this accretion scenario , which provides the best fit , ( see for instance am1934 - 563 , reshetnikov et al 2005b ) .
very massive polar rings can form through accretion , with polar components of comparable mass ( or even larger ) to the host component the prototype ngc 4650a is a good case for accretion , since no stellar halo is detected around the galaxy .
such a stellar halo is expected in the merging scenario to come from the dispersed stars from the destroyed satellite .
the polar component is quite massive , 8 10@xmath1 m@xmath2 of hi gas and 4 10@xmath1 m@xmath2 in stars .
the formation of polar rings could also occur without any interaction or merger , through gas infalling from cosmic filaments , with inclined angular momentum ( katz & rix 1992 , semelin & combes 2005 ) .
this scenario has been simulated by maccio & moore ( 2005 ) : the gas is acquired from a filament of @xmath0 1 mpc long , and is essentially cold gas accretion , with small impact parameter .
this kind of cold gas accretion might be the more realistic way by which galaxies get their gas ( e.g. keres et al 2005 ) .
one case happens to be quite similar to the prototype ngc 4650a , where the stars are essentially in the host and gas in the polar plane ; several small companions are also found all along the filament .
dark matter is quite round in the visible part after the infall of gas .
an alternative scenario , requiring no external gas accretion or merger for prg formation , has been proposed by tremaine & yu ( 2000 ) .
it assumes that a disk galaxy lies in the symmetry plane of a triaxial dark matter halo .
the dark halo tumbles with a retrograde pattern speed with respect to the disk .
when this tumbling slows down , through dark matter infall , the stars in the disk could get trapped in a vertical resonance with the halo . indeed , a stellar orbit slightly tilted with respect to the plane , precess at a slow retrograde rate , and when the resonance progressively reaches stars at larger and larger radii , the stars are propulsed into the polar plane .
with this mechanism , the polar ring would contain two equal - mass counter - rotating stellar disks .
this mechanism also explains the splitting of the stars in two counter - rotating streams in the disk , when the tumbling halo slows down from retrograde , to zero and then prograde ( evans & collett 1994 ) .
this mechanism however has not been confirmed in observations , when velocity fields of the stars in the polar ring have been studied ( e.g. swaters & rubin 2003 ) .
the stellar kinematics in ngc 4650a for instance does not reveal counter - rotation . a flat rotation curve in observed in the polar ring , which confirms that it is more a full plane than a ring . with the help of large telescopes , it is possible to have now much more precise and accurate kinematical information on the remote polar ring , as shown by the new vlt - fors2 data on ngc 4650a ( iodice et al 2005 ) .
the detailed stellar kinematics on the major and minor axis of the host galaxy reveal a flat velocity dispersion , with a slight drop in the center , and will allow in the near future a better determination of the gravitational potential and its 3d shape .
since a complete velocity field is difficult to obtain for many prgs , a single tully - fisher allows to gather more statistics ( iodice et al 2003 ) .
the position of 15 polar rings revealed quite surprising in such a plot , since the hi total widths of the gas in the polar plane is clearly larger than expected from the normal spiral tf relation .
because of the superposition of the two perpendicular systems , the orbits are expected to be non circular in the polar rings . in general , and for obvious selection effects , both components are seen nearly edge - on . then the maximum velocity in projection in the ring is coming from the top of the ring , corresponding to the largest distance from the center of the galaxy . the observed velocity is then expected to be lower than circular , and then lower than the maximal velocity observed in the host galaxy plane .
this is already the case in absence of dark matter , or if the dark matter is spherical .
if the dark halo is flattened and oblate , parallel to the host disk , this effect will be accentuated , and the hi width in the polar plane expected even lower .
but the contrary is observed .
is this due to the fact that the host component is also perturbed , and does not fall on the tf relation ?
figure 2 shows the tf diagram for the prgs where a detailed velocity width have been obtained in the two perpendicular components .
it shows that the host galaxy obeys the normal tf relation , while there is a large difference of velocity between the two perpendicular planes , the velocity being larger in the polar ring .
the only solution to the problem is to assume that the dark matter halo is oblate and aligned with the polar ring itself . in this case , the orbits in the host are then non - circular , and we observe the maximum velocity when the matter is farther from the center , and lower than circular . in the polar plane , the orbits will be circular , and larger than in the host plane .
it does not help to assume that the dark halo is prolate , keeping the orbits circular in the host .
is there a formation scenario able to explain that most prgs have their dark matter halo flattened along the polar disk ?
with dissipationless dark matter , both merging and accretion scenarios produce either spherical haloes , or halos flattened along the host . however ,
if a large fraction of the dark matter around galaxies is dissipative ( dark gas ) , it will settle in the polar plane , and it is possible to account for the flattening along the polar disk . between the velocities in the host galaxy equatorial plane and in the polar ring for each system .
from iodice et al ( 2003).,width=340 ]
pure cdm simulations predict triaxial shapes for collapsed structures , so that the haloes are not axisymmetric even in the plane of the baryonic galactic disk . the axisymmetry in galactic disks
have been checked through the orbits of the baryons , and in particular the hi gas , with low velocity dispersion .
of course , inclination effects have to be taken into account , as well as flaring , warps or other distortions , due to the spiral , bars or ring features in the galaxy disks .
the result of these investigations is that galaxies are actually axisymmetric in their planes , with a very low upper limit for the excentricity : below 0.1 with the isophote shape versus hi velocity widths method ( merrifield 2002 ) , or even less than 0.045 , when using near - infrared data to avoid extinction ( rix & zaritsky 1995 ) . on special cases ,
the limit can be better , excentricity of the order of 0.012 in potential in the very regular early - type galaxy ic2006 with an hi ring ( franx et al 1994 ) .
this axisymmetric shape of galactic haloes is confirmed by the low scatter observed for the tully - fisher relation .
as for the flattening , pure cdm haloes are predicted rather flattened in numerical simulations ; they are half oblate and half prolate , with axis ratios of the order of c / a = 0.5 , b / a = 0.7 .
it is interesting to note that the dark haloes are predicted more flattened then observed elliptical galaxies ; the distribution peaks at e5 ( while elliptical galaxies peak at e2 ( c / a = 0.8 ) ) cf dubinski & carlberg ( 1991 ) .
however , when the dissipative infall of gas is taken into account , the dark matter haloes are concentrated , through adiabatic contraction , and are also forced to an oblate shape , the prediction now becomes in average : c / a = 0.5 , b / a @xmath3 0.7 ( dubinski 1994 ) .
this depends a lot on the actual physics of baryons .
when the cooling is taken into account , the dark haloes are obtained much rounder , since the baryons are more concentrated to the center , in the simulations of kazantzidis et al ( 2004 ) .
these predictions are roughly compatible to the flattening observed statistically by weak lensing measurements ( hoekstra et al 2004 ) , where the average ellipticity of the haloes is 0.33 , if the dark matter is assumed aligned with the baryons .
* method of hi flaring * one of the methods most frequently used is the flaring of the hi planes : the thickness of the gaseous layers is a a function of the velocity dispersion , perpendicular to the plane and of the gravity restoring force towards the plane , which depends on the mass contained within the hi layer .
the z - velocity dispersion is not known in the edge - on galaxies where the hi thickness can be measured , but it is assumed the same as in similar face - on galaxies , where it is measured to be about constant with radius , and around 10km / s .
the thickness that can be derived by this method is sensitive to the radial distribution of the dark matter and mainly to its extension or truncation at large radius , which is not known precisely . olling ( 1995 , 1996 ) , and becquaert & combes ( 1997 ) found quite flattened halos for ngc 224 , 891 , 3198 , 4013 and 4244 ( see merrifield 2002 ) . from an nfw profile of the dark matter halo in the outer parts ( density in @xmath4 )
narayan et al ( 2005 ) find as the best fit a spherical shape . from the flaring also , kalberla ( 2004 )
favors a model where most of the dark matter is in the disk . * polar rings * as was shown above , the polar ring method is potentially powerful , but many more systems should be studied in details .
there is evidence for substantial dark matter , relatively flattened , although more along the polar plane than along the host plane .
only the case of ngc 4650a is reported in figure 3 .
* x - ray isophotes * to determine the amount and the flattening of dark matter in elliptical galaxies , the method of the shape of x - ray isophotes have been used ( e.g. buote et al 1998 , 2002 ) .
the method relies on the comparison between the observed ellipticities and orientations of x - ray isophotes , with those expected if the gravitating mass traces the same shape as the stellar light .
this method was applied to two isolated e4 ellipticals ( ngc 720 and ngc 3923 ) , and one isolated s0 galaxy ( ngc 1332 ) .
the results were dark matter ellipticities between 0.4 and 0.6 , and the indication in ngc 720 that the dark matter is flatter than the stellar distribution ( cf fig 3 ) .
* tidal streamers in the dark halo of the mw * most sagittarius streams are too young to constrain the shape of the milky way dark halo .
either oblate or prolate , with c=0.6 are compatible with the data .
only one older stream and its kinematics give a prolate shape of 0.6 helmi ( 2004a , b ) . a more spherical fit is found by johnston et al ( 2005 ) .
warps in stellar disks are very frequent ( e.g. reshetnikov & combes 1998 ) .
if the disk is self - gravitating , spontaneous bending instabilities can explain these warps ( revaz & pfenniger 2004 ) .
the disks are unstable to bending , if the z - velocity dispersion is below araki limit @xmath5/@xmath6 = 0.293 .
if the optical disk warps have all to be explained through this instability , then the fraction of dark matter in the disk should be above a certain threshold , according to the flattening assumed for the spheroidal halo . for a typical galaxy as the milky way , the dark matter should be equally distributed in the disk and the spheroidal halo ( revaz & pfenniger 2004 ) .
the kinematical observations of polar rings are precious tools to probe the 3d shape of dark haloes .
their tully - fisher diagram until now reveal that most prgs have their dark matter flattened in the polar planes .
more detailed data on more systems are required to confirm this result , which would imply that dark baryons could play a role in their formation .
small stellar warps that are present in most spiral disks put constraints in the dark matter flattening , if it is assumed that they are excited by bending waves .
large gas warps might be a phenomenon akin to polar rings , formed through late accretion of external gas .
the frequency of these pehnomena confirm the important role of late gas accretion in galaxy formation .
becquaert , j. , combes f. : 1997 , a&a 325 , 41 bekki , k. 1997 : apj , 490l , 37 bekki , k. 1998 , apj , 499 , 635 bournaud , f. , combes , f. 2003 , a&a , 401 , 817 buote d.a .
, canizares c.r . : 1998 , aspc 136 , 289 ( astro - ph/9710001 ) buote d.a . ,
jeltema t.e . ,
canizares c.r .
, garmire g.p . : 2002 , apj 577 , 183 dubinski j. , carlberg r.g . : 1991 , apj 378 , 496 dubinski j. : 1994 , apj 431 , 617 evans , n. w. , collett , j. l. : 1994 apj 420 , l67 franx , m. , van gorkom , j. h. , de zeeuw , t. : 1994 , apj 436 , 642 helmi a. : 2004a , mnras 351 , 643 helmi a. : 2004b , apj 610 , l97 hoekstra h. , yee h.k.c . ,
gladders m.d . : 2004 , apj 606 , 67 iodice e. , arnaboldi m. , bournaud f. et al : 2003 , apj 585 , 730 iodice e. , arnaboldi m. , saglia r.p . , et al : 2005 , apj submitted johnston k.v . , law d.r . , majewski s.r . : 2005 , apj 619 , 800 kalberla p.m.w . : 2004 , in `` baryons in dark matter haloes '' , sissa , dettmar r. et al ( eds ) katz , n. , rix , h - w . : 1992 apj 389 , l55 kazantzidis , s. , kravtsov a , zentner a. etg al . : 2004 , apj 611 , l73 keres d. , latz n. , weinberg d. , dave r. : 2005 , mnras , in press ( astro - ph/0407095 ) krips m. , eckart a. , neri r. et al . : 2005 , a&a , in press ( astro - ph/0507403 ) maccio a.v .
, moore b. : 2005 , mnras , submitted merrifield m.r . :
2002 in `` the shapes of galaxies and their halos '' yale , may 2001 , ed world scientific ( astro - ph/0107291 ) narayan c.a .
, saha k. , jog c.j . : 2005 , a&a in press ( astro - ph/0508358 ) olling , r. p. : 1995
, aj 110 , 591 olling , r. p. : 1996
, aj 112 , 457 reshetnikov , v. , sotnikova , n. 1997 : a&a , 325 , 933 reshetnikov , v. , combes f. : 1998 , a&a 337 , 9 reshetnikov v. , bournaud f. , combes f. , et al : 2005a , a&a 431 , 503 reshetnikov v. , bournaud f. , combes f. , et al : 2005b , a&a in press revaz , y. , pfenniger , d. : 2004 a&a 425 , 67 schweizer , f. , whitmore , b. c. , rubin , v. c. 1983 : aj , 88 , 909 semelin b. , combes f. : 2005 , a&a in press ( astro - ph/0506589 ) swaters , r. a. , rubin , v. c. : 2003 apj 587 , l23 tremaine , s. , yu , q. : 2000 mnras 319 , 1 whitmore , b. c. , lucas , r. a. 1990 : aj , 100 , 1489 | polar ring galaxies , where matter is in equilibrium in perpendicular orbits around spiral galaxies , are ideal objects to probe the 3d shapes of dark matter halos .
the conditions to constrain the halos are that the perpendicular system does not strongly perturb the host galaxy , or that it is possible to derive back its initial shape , knowing the formation scenario of the polar ring . the formation mechanisms are reviewed : mergers , tidal accretion , or gas accretion from cosmic filaments .
the tully - fisher diagram for polar rings reveals that the velocity in the polar plane is higher than in the host plane , which can only be explained if the dark matter is oblate and flattened along the polar plane .
only a few individual systems have been studied in details , and 3d shapes of their haloes determined by several methods .
the high frequency of warps could be explained by spontaneous bending instability , if the disks are sufficiently self - gravitating , which can put constraints on the dark matter flattening . |
this work has initiated during a visit to ufmg .
i thank r. dickman introducing me to the field of nonequilibrium phase transitions .
this work was partially supported by cnpq and faperj . | in cond - mat/0107371 , de mendona proposes that diffusion can change the universality class of a parity - conserving reaction - diffusion process . in this comment
we suggest that this can not happen , due to symmetry arguments .
we also present numerical results from lattice simulations which support these arguments , and mention a previous result supporting this conclusion .
although there are a large variety of nonequilibrium reaction / reaction - diffusion models which present a transition from an active to an absorbing state , their critical behaviors fall within a small number of universality classes . the simplest one , which can be viewed as the nonequilibrium counterpart of the ising universality class , is the contact process ( cp ) , or directed percolation ( dp ) universality class @xcite .
another well - established class is the parity - conserving ( pc ) universality class @xcite .
more recently , binary spreading ( bs ) processes ( with or without parity conservation ) , have appeared as representative of a new universality class @xcite .
it appears that the nature of the creation and annihilation processes is fundamental on determining to which universality class a given process might belong , that is , the number of particles necessary for a reaction to occur seems to play a key role .
one way to assess the importance of such terms is via numerical simulations or , whenever possible , by studying the analytical properties of the equations governing the dynamical process . for any markovian process ,
one can write a master equation describing how the probability @xmath0 of a given configuration @xmath1 evolves in time @xcite . a technique first introduced by doi @xcite allows us to derive a field - theoretic representation for the master equation , and then to handle fluctuations systematically .
moreover , with the field - theoretic action in hand , one can write down the langevin equation which governs the stochastic evolution of the field operators , with no ambiguity on the definition of the noise @xcite . even though the noise is always multiplicative in absorbing state phase transitions
, it can have different functional forms , which can change entirely the long - term behavior of the system being described .
for instance , when two particles are necessary for annihilation to occur , the noise in the langevin equation is , counter - intuitively , purely _ imaginary _ ( as in representatives of the pc class ) . on the other hand ,
when both creation and annihilation require a pair of particles , there is a competition between `` real '' and `` imaginary '' noise ( as in representatives of the bs class ) .
.nature of creation / annihilation processes within each universality class , that is , the number of particles needed for a reaction to occur . [ cols="^,^,^",options="header " , ] the cp(m ) process , invented by inui and tretyakov @xcite , is a parity - conserving dynamical process that can be summarized as follows @xmath2 it can be interpreted as a process where a particle generates @xmath3 others with probability @xmath4 , while @xmath3 particles , upon contact , are annihilated with probability @xmath5
. an isolated cluster of @xmath6 particles can not diffuse or annihilate .
it has been shown by the authors that cp(2 ) , starting with an even number of particles , belongs to the pc universality class @xcite .
recently , mendona @xcite studied the cp(2 ) process with diffusion of solitary particles . on the basis of finite size scaling of the exact numerical diagonalization of the evolution operator , he reports that diffusion changes the universality class from pc to dp .
numerical estimates for @xmath7 , the dynamic exponent governing spatial fluctuations of the cluster of alive sites , seem to converge to @xmath8 . at a first glance , the claim that the addition of diffusion might alter the dynamical critical behavior of the cp(2 ) process does not seem plausible , since it does not change the symmetry of the problem : the number of particles
is still conserved modulo @xmath9 and the absorbing state is still unique and devoid of particles . moreover , as in the analysis of simpler models @xcite , it can be seen that a coarse - grained description of non - diffusive lattice models generally includes a diffusive term , and even without coarse - graining , an effective diffusion is already present in cp(2 ) , for instance , in the sequence @xmath10 . to verify this ,
we have developed a probabilistic cellular automaton for the @xmath11 cp(2)d ( with diffusion rate @xmath12 ) with even and odd sites constituting sub - lattices @xmath13 and @xmath14 @xcite .
starting from a fully occupied lattice of @xmath15 sites , we have measured the density of active sites @xmath16 , which should decay as @xmath17 at criticality . from the data we find @xmath18 .
we have also performed dynamical simulations at the critical point , starting from a single pair of particles and measuring the accumulated number of active sites , @xmath19 , from which we find @xmath20 .
these results suggest that diffusion does no alter the critical behavior of cp(2 ) , and only confirm what has already been obtained by zhong and ben - avraham , who studied a branching - annihilating random walk with finite annihilation rate @xcite .
this system is equivalent to the cp(2)d with high diffusion probability , and thus should have the same critical behavior .
we point out to cond - mat/0207720 for recent simulations of the diffusive pair contact process @xcite , a lattice model which ( possibly ) has a diffusion dependent universality class . |
the authors acknowledge the assistance of the entire fermi team during the preparation and tuning of the fel and also wish to thank c. svetina , m. zangrando and the padres group for the support on the fermi photon diagnostics .
p.r.r . acknowledges financial support from the talents up programme ( 7th fp , specific programme people , marie curie actions - cofund ) g.a .
b.m . acknowledges financial support from the labex palm ( ramses project ) .
the authors also acknowledge useful discussions with g. stupakov and k. prince .
99ifxundefined [ 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 * * , ( ) the experimental spectral intensity in the central part of the muti - peak region is lower compared to the theoretical values , indicating a reduction in electron rebunching of about 30% .
possible reasons for this effect may include a non - gaussian initial electron energy distribution , a correlated energy spread or a microbunching instability in the electron beam .
the experimental spectral maps were reproduced using eqs .
[ bunching_factor ] and [ fel_phase ] in the spectral domain .
the additional dispersive strengths of the modulator and the radiators were taken into account in the calculations .
the only free parameter was the amplification chirp ( the values used for each case were 2.0 , 4.0 and @xmath35 rad / fs@xmath19 respectively , in agreement with theory @xcite and simulations performed with perseo @xcite ) .
@noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | we demonstrate the ability to control and shape the spectro - temporal content of extreme - ultraviolet ( xuv ) pulses produced by a seeded free - electron laser ( fel ) .
the control over the spectro - temporal properties of xuv light was achieved by precisely manipulating the linear frequency chirp of the seed laser .
our results agree with existing theory , which allows retrieving the temporal properties ( amplitude and phase ) of the fel pulse from measurements of the spectra as a function of the fel operating parameters .
furthermore , we show the first direct evidence of the full temporal coherence of fel light and generate fourier limited pulses by fine - tuning the fel temporal phase . the possibility to tailor
the spectro - temporal content of intense short - wavelength pulses represents the first step towards efficient nonlinear optics in the xuv to x - ray spectral region and will enable precise manipulation of core - electron excitations using the methods of coherent quantum control . [ [ section ] ] the development of the first high - power pulsed lasers in the 1960s marks an important milestone in the long - standing effort to actively control the temporal evolution of quantum - mechanical systems .
however , it was not until the early 1990s , when ultrashort pulse shaping techniques became practical , that coherent control of quantum phenomena finally became a reality @xcite . tailoring
the spectro - temporal content of intense laser pulses in the visible range opened up countless possibilities for manipulating the quantum state of matter , with examples ranging from control of population transfer in optical transitions @xcite and currents in semiconductors @xcite to control of chemical reactions @xcite and energy flow in biomolecular complexes @xcite , and many others ( see , e.g. ref . @xcite and references therein ) . undoubtedly , the ability to shape the spectro - temporal content of powerful extreme - ultraviolet ( xuv ) and x - ray pulses would trigger widespread efforts to extend the concepts of coherent control into the short - wavelength regime , leading ultimately to the development of new methods for probing and manipulating core electrons in atoms , molecules and materials . as a more immediate application ,
it may find use in numerous advanced spectroscopic techniques such as resonant inelastic x - ray scattering @xcite and coherent ultrafast core - hole correlation spectroscopy , offering unique capabilities for probing elementary excitations @xcite . in the xuv / x - ray spectral region , free - electron lasers ( fels )
are currently the only devices that can deliver femtosecond laser - like pulses with peak powers in the gigawatt range @xcite .
however , the ability to generate fully coherent pulses and to shape their spectro - temporal content with high stability on a shot - to - shot basis is extremely challenging , due to the difficulties in precisely controlling the light generation process . in this letter
we show that the spectro - temporal content of powerful ultrashort xuv pulses can be precisely shaped using a laser - seeded fel . by tuning the seed laser operating parameters we generate intense femtosecond xuv pulses with a controllable amount of linear frequency chirp , which can be ultimately reduced to zero .
our results constitute the first experimental evidence of fourier limited pulses from an xuv fel and pave the way to full spectro - temporal shaping of powerful ultrafast pulses in the xuv to soft - x - ray spectral region . a natural approach towards fully controlling the spectro - temporal profile of powerful xuv / x - ray pulses
is to exploit the unique capabilities of a seeded fel .
whereas in fels based on sase ( self - amplified spontaneous emission ) the amplification is triggered by spontaneous undulator emission @xcite , a seeded fel relies on a coherent input signal .
we employed the high - gain harmonic generation ( hghg ) seeding scheme @xcite available at the fermi fel @xcite . in hghg , see fig .
[ hhgscheme ] , a seed laser is used to imprint a periodic energy modulation at the seed wavelength @xmath0 ( typically in the uv ) onto a relativistic electron beam ( e - beam ) in the modulator . this energy modulation is converted into a current density modulation when the e - beam interacts with the magnetic field of the dispersive section .
such a microbunched e - beam emits coherent light at @xmath1 ( @xmath2 integer ) in the xuv / x - ray region as the electrons traverse the periodic magnetic field of the radiator . the fel output at the @xmath2th harmonic
is driven by the electron bunching factor @xcite @xmath3 e^{i n[\phi_s(t)+\phi_e(t ) ] } \mbox { , } \label{bunching_factor}\ ] ] where @xmath4 is the dimensionless strength of the dispersive section ( dispersive strength ) @xcite , @xmath5 is the @xmath2th order bessel function , and @xmath6 is the time - dependent energy modulation ( normalized to the electron energy spread @xmath7 ) , which is proportional to the envelope @xmath8 of the seed laser electric field @xmath9}$ ] , @xmath10 being the central frequency . the exponent in the last factor , where @xmath11 , accounts for the slowly varying phase of the seed @xmath12 and a possible time - dependent energy profile @xmath13 imprinted onto the e - beam by the linear accelerator ( linac ) , which produces the relativistic electrons . as the microbunched e - beam is injected into the radiator it starts emitting coherent light .
initially the e - beam is rigid and the fel electric field is directly proportional to the bunching factor in eq .
[ bunching_factor ] . in the linear regime ( before saturation ) @xcite , the field envelope is preserved despite amplification in the radiator , as demonstrated in the following .
however , a small additional phase @xmath14 is introduced @xcite , so that the total fel phase becomes : @xmath15+\phi_a(t ) \mbox{. } \label{fel_phase}\ ] ] the above equations provide the basis for fel pulse shaping through the manipulation of the seed envelope @xmath8 and phase @xmath12 . both the e - beam time - dependent energy profile and the phase developed during amplification affect the pulse properties @xcite .
however , as we show in the following , these effect can be fully compensated for by properly tuning the temporal phase of the seed laser .
for the cases considered here it suffices to expand each of the individual phase contributions @xmath12 , @xmath16 and @xmath14 into a power series in time up to the second order . while a linear - term coefficient @xmath17 results in an absolute frequency ( wavelength ) shift @xcite , a quadratic - term coefficient @xmath18 , called the chirp rate , gives rise to a linear frequency chirp in the pulse @xcite . a suitable seed laser chirp
can then be used to counter the combined effects due to the e - beam quadratic energy curvature and the chirp developed during amplification .
fig .
[ bunching_factor_figure ] highlights how the spectro - temporal content of fel light can be shaped by tuning the operating parameters .
first , fig .
[ bunching_factor_figure ] a ) , the bunching envelope , generated here by a gaussian seed , can be modified by changing the strength @xmath4 of the dispersive section .
with increasing @xmath4 , the bunching develops modulations as a function of time ( eq . [ bunching_factor ] ) due to the process of electron overbunching and rebunching @xcite , leading to a pulsed structure .
as emphasized above , the fel pulse temporal shape directly corresponds to the bunching envelope .
second , manipulating the fel temporal phase using a chirped seed , fig .
[ bunching_factor_figure]b ) , leads to a drastic modification of the fel spectral content ; see fig .
[ bunching_factor_figure ] c ) .
while the spectral map ( spectrum vs. @xmath4 ) of a significantly chirped fel pulse directly corresponds to its temporal map @xcite , a fourier limited pulse with a flat temporal phase shows a distinctively different spectral signature . in the following
we experimentally verify the above predictions and demonstrate the power of the hghg scheme in controlling the spectro - temporal content of intense xuv pulses .
figure [ spectra ] a ) shows the experimental spectral map , when the relatively high positive frequency chirp on the fel , dominated by the positive chirp due to a stretched seed , dictates the radiation properties . here
, we use the group delay dispersion ( gdd ) as a measure for the degree of chirp on the seed laser .
the strong linear frequency chirp , characterized by a gdd of 7390 fs@xmath19 , was achieved by putting a 20 mm thick plate of fused silica into the seed laser ( frequency up - converted pulses from an optical parametric amplifier , opa ) path , inducing a positive chirp rate of @xmath20 rad / fs@xmath19 and stretching the seed pulse duration to 250 fs ( fwhm ) , at the operating wavelength of 258 nm .
the situation corresponds to the one in the top part of fig .
[ bunching_factor_figure]c ) .
the fel spectrum develops intensity modulations with increasing @xmath4 , which directly correspond to the intensity modulations in the temporal domain .
excellent agreement @xcite between experiment and theory ( inset ) @xcite demonstrates that , despite amplification in the radiator , the fel pulse envelope is preserved , justifying the use of eqs .
[ bunching_factor ] and [ fel_phase ] to describe the spectro - temporal content of fel light . and [ fel_phase ] in the spectral domain .
the wavelength shift ( spectral map tilting ) in c ) is due to the contribution of a linear - term coefficient @xmath17 from the e - beam energy profile . the experiment in a ) was done at @xmath21 using all of the six 2.4-m - long radiator sections , while for b ) and c ) @xmath22 and the number of active radiator sections was 4 and 5 , respectively . in all cases ,
the e - beam energy , energy spread , current , bunch charge , and emittance were @xmath23 gev , @xmath24 kev , @xmath25 a , @xmath26 pc , and @xmath27 mm mrad , respectively . for a ) and b ) the e - beam ( positive ) quadratic energy curvature was @xmath28 5 mev / ps@xmath19 , while for c ) the value was @xmath28 12 mev / ps@xmath19 .
the data was acquired with a dedicated spectrometer @xcite .
the maps in a ) and b ) consist of 10 accumulated single - shot spectra for each value of @xmath4 , while data in c ) were taken by summing 3 multi - shot spectra ( 10 shots ) .
the data within each spectral slice ( for fixed @xmath4 ) are normalized in amplitude for better visualization . ,
scaledwidth=40.0% ] the next experiment , fig .
[ spectra]b ) , was performed with a moderate positive chirp ( gdd = 4520 fs@xmath19 ) on the seed laser . for this purpose
the direct output of the opa at 255 nm , with a pulse duration of 160 fs and a chirp rate of @xmath29 rad / fs@xmath19 , was used to seed the fel . because the total fel chirp is not high enough to satisfy the condition for the spectro - temporal equivalence @xcite ,
the central part of the spectral map is significantly modified .
again , a remarkably good correspondence is obtained between experiment and theory ( inset ) , demonstrating the predictive power of eqs .
[ bunching_factor ] and [ fel_phase ] . based on the two spectral signatures in figs .
[ spectra]a ) and b ) the next set of experiments was carried out by putting a negative chirp rate on the seed ( third harmonic of a ti : sapphire laser , operating at 261.6 nm ) ; in this case the gdd was -1100 fs@xmath19 .
a fine control of the negative chirp rate was possible using an optical compressor based on a pair of gratings .
the data in fig .
[ spectra]c ) correspond to a chirp rate of @xmath30 rad / fs@xmath19 , compressing the seed pulse down to 120 fs .
a strong modification of the spectral content vs. @xmath4 with respect to the previous two cases is seen in fig .
[ spectra ] c ) , confirming the extremely high sensitivity of the spectral maps to the fel phase .
experimental data once again fit well with calculations ( inset ) .
remarkably , the spectral signature corresponds to that of a fourier limited pulse with a flat temporal phase ( cf .
fig .
[ bunching_factor_figure]c ) , bottom ) .
the negative chirp rate on the seed compensates the positive chirp due to the quadratic curvature ( @xmath31 mev / ps@xmath19 , measured at the end of the linac using a radio - frequency deflecting cavity in combination with an energy spectrometer @xcite ) on the e - beam time - dependent energy profile and the chirp developed during the amplification stage @xcite .
more precisely , because the phase contribution from the e - beam @xmath16 is a function of @xmath4 , the fel chirp rate varies linearly from about @xmath32 to @xmath33 rad / fs@xmath19 in the range of dispersive strengths used in fig .
[ spectra ] c ) , going through zero at @xmath34 , where full chirp compensation is achieved .
several important conclusions can be drawn from the above results .
the experimental spectral maps in fig .
[ spectra ] and their excellent agreement with theory based on eqs .
[ bunching_factor ] and [ fel_phase ] imply that the fel output can be fully manipulated by controlling the seed laser envelope and phase , where the dispersive strength can act as an additional tuning knob ( e.g. , an fel pulse train with a fixed phase relationship between individual pulses is obtained simply by increasing the strength of the dispersive section ) .
our results demonstrate , to the best of our knowledge , the first compelling evidence that the radiation produced by a seeded fel is temporally fully coherent and that by adjusting the seed laser phase , a controllable amount of linear frequency chirp can be transferred to the fel pulse .
moreover , a careful tuning allows generation of fourier limited pulses with a flat temporal phase .
such pulses are typically used as references in the field of coherent quantum control .
our results therefore open the door to full spectro - temporal shaping of intense ultrashort pulses in the xuv to soft - x - ray region using methods similar to the ones developed in the visible spectral region @xcite . in combination with the possibility to engineer the fel transverse radiation profile using hghg @xcite ,
this sets the stage for entirely new experiments with pulsed light at short wavelengths . |
early x - ray imaging observations with the _ einstein _ observatory and _ rosat _ showed that in the central regions of clusters of galaxies the radiative cooling time is shorter than the age of the universe ( e.g canizares , stewart , & fabian 1983 ) . as a result ,
the intracluster medium ( icm ) should cool down to form a cold ( @xmath6k ) gas phase inducing a global inflow of gas .
this `` cooling flow '' ( see fabian 1994 for a review ) picture has been extensively discussed and formed a basic assumption in many arguments .
the low resolution spectroscopy in 0.5 - 2 kev by _ rosat _ showed that in some clusters the icm temperature actually decreases towards the center ( e.g. bhringer et al .
1994 ; david et al .
1994 ; allen & fabian 1994 ) .
higher resolution spectroscopy in 0.5 - 10 kev with _ asca _
, however , can not be fully understood with the conventional cooling flow model .
_ asca _ spectra of cooling flow clusters can be well explained by a two ( hot and cool ) phase plasma without significant excess absorption features ( e.g. ikebe et al .
1999 ; makishima et al .
a naive cooling flow model predicting a range of temperatures with intrinsic absorption could also fit the _ asca _ data but generally produce worse chi - square results ( e.g. allen et al .
most recently , very high resolution spectroscopy with xmm - newton / rgs unambiguously show that there is very little x - ray emission from gas cooler than certain lower cut - off temperatures of @xmath7 kev ( tamura et al . 2001 ; kaastra et al .
2001 ; peterson et al .
2001 ) . unless a large amount of cooled gas or the metals in the cold gas are hidden ( fabian et al .
2001a ) , there must exist a heating mechanism that prevents the icm from radiative cooling .
the necessity of global heat input into the icm in addition to gravitational heating has also been pointed out from the break of the self - similarity between dark matter and icm , which is most clearly demonstrated in the x - ray luminosity - temperature relation ( @xmath8 relation ) . a simple scaling analysis ( kaiser 1986 )
suggests a relation of @xmath9 , while observation shows @xmath10 ( e.g. edge & stewart 1991 ; david et al .
1993 ; white et al .
1997 ; wu et al .
the break of the self - similarity is also seen in the entropy vs temperature relation .
ponman , cannon , & navarro ( 1999 ) showed that cooler systems ( @xmath11 kev ) have entropies higher than achievable through gravitational collapse alone .
simulation studies show that the observed relations can be reproduced , if there is enough non - gravitational heat input into the icm by feedback from galaxies ( e.g. metzler & evrard 1994 ; bower et al .
2001 ) or preheating before the cluster formation ( e.g. navarro et al . 1995 ; tozzi & norman 2000 ) . in order to shed some new light onto these `` cooling flow phenomena '' and `` break of self similarity '' , we , in the present paper , perform a comparison of the temperature distribution of the icm to the distribution of the velocity dispersion of the dark matter .
a parameter , @xmath12 , is often used as a measure of the average kinetic energy per unit mass in galaxies relative to that in the icm . from observations of many clusters ,
the mean @xmath3 is @xmath13 with large scatter ( e.g. wu et al . 1999 ) , indicating that the energy equipartition between galaxies and icm is roughly achieved on average . in analogy with @xmath3
, we can introduce @xmath14 for comparison between the mean kinetic energy of the dark matter and that of the icm , and define the dark matter `` temperature '' as @xmath15 .
note that the above definition of `` temperature '' , by means of proton mass instead of the actual mass of dark matter particles , is only for the sake of comparison with the gas temperature .
therefore we put `` temperature '' in quotes .
we obtain , in this paper , the radial profile of the @xmath2 value observationally for the first time . in sect .
2 , we describe the method of measuring the dark matter velocity dispersion in a cluster of galaxies from an x - ray observation .
we applied the method to the xmm - newton data of a prototypical cooling flow cluster , abell 1795 ( hereafter a1795 ) , which is located at z=0.0616 ( struble & rood 1987 ) .
the x - ray data analysis and results are presented in sect .
3 . discussion and summary are found in sect . 4 and sect .
5 , respectively . throughout the paper , the hubble constant
is given as 70 @xmath16 km s@xmath17 mpc@xmath17 , and a flat universe ( @xmath18 , @xmath19 ) is assumed .
at the redshift of a1795 , 1 arcsec corresponds to 1.19 kpc .
we use a simplified model cluster being composed of a hot plasma icm with a temperature of @xmath20 k , and dark matter made of collisionless particles . under the assumptions of spherical symmetry and hydrostatic equilibrium ,
the icm distribution is described by @xmath21 where @xmath22 is the total gravitating mass within a sphere of radius @xmath23 , @xmath24 is the boltzmann constant , @xmath25 is the gravitational constant , @xmath26 and @xmath27 is the density and temperature of the icm , respectively . from an x - ray observation , @xmath26 and @xmath27
are measured and @xmath22 can be obtained via eq .
( [ eq : hydro ] ) .
when the dark matter particles tracing the same gravitational field are in steady state , they obey the jeans equation @xmath28 where @xmath29 is the one - dimensional radial velocity dispersion , and @xmath30 is the mass density , which is given as @xmath31 .
therefore , once the total gravitating mass , @xmath32 , and the gas density , @xmath33 , profiles are obtained from an x - ray observation , eq .
( [ eq : jeans ] ) contains only one unknown parameter , @xmath29 .
equation ( [ eq : jeans ] ) can be rewritten in the form of an ordinary differential equation for @xmath34 as @xmath35 which can be solved numerically under a given boundary condition to derive the velocity dispersion profile of the dark matter .
although a steady state dark matter distribution is assumed above , a dark matter density profile should actually be growing as matter is falling onto the system from outside .
a numerical simulation by fukushige & makino ( 2001 ) shows that the dark matter halo grows in a self - similar way , keeping the density profile in the central region unchanged .
this justifies the assumption at least in the central region .
a similar technique as described above was used in determining the icm temperature profile , using x - ray imaging data without spectroscopic information taken e.g. with the _ einstein _ observatory .
a model profile for the total mass is assumed and the temperature profile is derived from the equation of hydrostatic equilibrium so that the observed brightness profile is explained ( e.g. fabian et al .
1981 ; hughes 1989 ) . here
we solve the jeans equation instead in order to derive the dark matter velocity dispersion .
an ideal opportunity to study the mass profile as well as velocity dispersion profile of dark matter is provided by the xmm - newton observation of a prototypical cooling flow cluster , a1795 .
early x - ray imaging observations with the _ einstein _ observatory and _ rosat _ showed that the x - ray emission from a1795 has almost circular symmetry with a small elongation along the north - south direction , which indicates that the cluster is well relaxed dynamically ( jones & forman 1984 ; boute & tsai 1996 ) .
the radial brightness profile shows a huge central excess above a prediction from an isothermal @xmath36-model profile , and the central excess luminosity gives a mass deposition rate of @xmath37 m@xmath38 yr@xmath17 based on the standard cooling flow interpretation ( edge et al .
1992 ; briel & henry 1996 ) . from the 0.5 - 10 kev spectrum taken with _
asca _ , however , a significantly smaller mass deposition rate is obtained ( @xmath39 m@xmath38 yr@xmath17 fabian et al . 1994
; @xmath40 m@xmath38 yr@xmath17 xu et al . 1998 ; @xmath41 m@xmath38 yr@xmath17 allen et al . 2001 ) . instead of the cooling flow model , the _ asca _ spectrum can be best described with a two temperature model without excess absorption ( xu et al . 1998 ) .
the xmm - newton observation of a1795 was carried out during the performance - verification phase on june 26 2000 with total observing time of @xmath42 ksec . the icm temperature profile obtained with epic mos as well as epic pn
were already reported by tamura et al .
( 2001 ) and by arnaud et al .
( 2001 ) , showing that it is almost isothermal at @xmath436 kev in the 2 - 10 arcmin radius region , while it decreases towards the center reaching the minimum temperature at @xmath44 kev .
the chandra observation gives consistent results ( ettori et al .
a high resolution spectrum from the central region obtained with xmm - newton / rgs has been analyzed by tamura et al .
( 2001 ) , showing a lack of signature of cool gas components below @xmath44 kev and an upper limit of 77 @xmath45 m@xmath38 yr@xmath17 for the mass deposition rate is obtained if an isobaric cooling flow model is applied . in this paper
, we analyze the data taken with the xmm - newton / epic pn , mos1 and mos2 to measure the total mass profile of a1795 and to derive the dark matter velocity dispersion profile using the method described in sect .
a significant fraction of any xmm - newton observation is often contaminated by the huge background count rate by soft proton flares . in order to derive time intervals with stable background
, we made the 0.510 kev band lightcurves individually with the pn , mos1 , and mos2 data , and eliminated time periods where the count rate deviates from the mean value during quiescent periods by @xmath46 .
the total usable exposure time thus left is 23 ksec , 31 ksec , and 34 ksec for pn , mos 1 and mos 2 , respectively . in the present analysis ,
the single and double pixel events of the pn data and the single , double , triple , and quadruple events of the mos data are used .
the background consisting of the cosmic x - ray background and the high energy particle events was estimated from the lockman hole data taken during the revolution # 70 . in this sky region ,
the galactic column density is as low as the a1795 field . after eliminating the contamination sources from the lockman hole data as well as the a1795 data ,
the lockman hole data are scaled so that the count rates of the background and the cluster data becomes the same in the 7 - 12 kev energy band , and in the radius range of 575 - 925 arcsec for pn and of 500 - 850 arcsec for mos , where no significant x - ray signal from the cluster was detected .
the background data thus obtained were subtracted from the cluster data . for pn , the out - of - time events ( strder et al .
2001 ) were also subtracted prior to the subtraction of the background . defining the center as the x - ray peak position , we have produced a projected radial brightness profile in the 0.8 - 10 kev band .
the x - ray count rate was converted to the x - ray flux by correcting each x - ray photon with the instruments responses that include the effective area and vignetting of the x - ray telescopes , and the quantum efficiency of the ccd cameras . combining the data taken with pn , mos1 and mos2 , we thus derived the brightness profile , which is illustrated in fig .
[ fit_2beta ] . for comparison with the other x - ray observations
, we quantified the radial profile using the @xmath36-model profile given by @xmath47^{-3\beta+0.5}\ , \label{eq : beta}\ ] ] where @xmath48 is a core radius and @xmath36 is a beta parameter .
the @xmath36 model has been often used to fit x - ray brightness profiles and generally gives a good representation for non cooling flow clusters . since the x - ray brightness profile of a1795
has long been known to show a central excess above a @xmath36-model profile , we employed here a double @xmath36-model profile that is the sum of two @xmath36-model profiles to fit the 0.8 - 10 kev brightness profile within 848 arcsec . in the actual fitting
, the model brightness profile is convolved with the point spread function of the x - ray telescope,^{-\alpha}$ ] , where @xmath49 and @xmath50 depend on off - axis angle as well as x - ray energy ( see ghizzardi 2001 ) .
we assumed that the psf is constant in the entire fov and independent of the x - ray energy .
as the representative , we employed the on - axis psf at 1.5 kev . using the @xmath49 and @xmath50 parameters of ( @xmath49[arcsec ] , @xmath50 ) = ( 5.37 , 1.5 ) , ( 4.72 , 1.457 ) , and ( 4.4545 , 1.4035 ) , for pn , mos1 , and mos2 , respectively
, we obtained the average psf , which is then used in the brightness profile fittings .
] whose half - power - radius is @xmath51 arcsec at 1.5 kev on - axis and significantly affects several of the central bins .
evaluating the goodness of the fit with a chi - square statistic , we obtained a good fit as shown in fig .
[ fit_2beta ] with the best - fit parameters as summarized in table [ tab:2beta ] .
the temperature and metallicity profiles given by tamura et al .
( 2001 ) were derived from the conventional annular spectral analysis , accumulating spectra from concentric annular regions in the projected 2-dimensional space .
we derive here the temperature and metallicity profiles instead in 3-dimensional form from deprojected spectra . with each pn , mos1 , and mos2 data set ,
we first accumulate 13 annular spectra from the vignetting corrected count rate profiles in individual energy channels .
the outer radii are 16 , 32 , 48 , 64 , 96 , 128 , 192 , 256 , 384 , 512 , 640 , 768 , and 848 arcsec .
assuming spherical symmetry and that there is no cluster emission beyond 848 arcsec from the cluster center , we deprojected the annular spectra to spectra for 13 spherical shell regions whose radii correspond to the annular radii . since no intrinsic absorption is found even from the central region ( tamura et al .
2001 ) , the x - ray emission from the cluster is entirely optically thin , and we performed the deprojection process as follows .
the relation between the energy spectra observed in the @xmath52 annular regions , @xmath53 ( @xmath54 ) , and those emitted from the @xmath52 shell regions , @xmath55 ( @xmath56 ) , is given with a matrix relation as @xmath57 where @xmath58 in the present case .
the matrix elements are given as @xmath59 / ( r_{\rm i+1}^3 - r_{\rm i}^3 ) & { \rm if}\ i > j,\\ ( r_{\rm i+1}^2 - r_{\rm i}^2)^{3/2 } / ( r_{\rm i+1}^3 - r_{\rm i}^3 ) & { \rm if}\ i = j,\\ 0 & { \rm if}\ i < j , \end{array } \right.\ ] ] where @xmath60 ( @xmath61 ) and @xmath62 ( @xmath63 ) are the inner and the outer radius of the ith shell ( jth annular ) , respectively .
each element represents the volume fraction of the ith shell contributing to the emission detected in the jth annular .
since @xmath64 is a triangular matrix , it can be easily inverted to @xmath65 , and the deprojected shell region spectra can be derived from the annular spectra by @xmath66 statistical errors in the individual energy bins are properly propagated through this operation . out of the 13 deprojected spectra
thus derived from each pn , mos1 , and mos2 data set , we used only the inner 10 deprojected spectra within a 512 arcsec radius in the analysis below so that the artificial cut - off of the x - ray emission outside the maximum radius of 848 arcsec does not affect the resulting deprojected spectra within 512 arcsec .
in addition , the signal to noise ratios of the spectra within 512 arcsec are high enough , and some 10% systematic error of the predicted background count rates introduce no significant systematic errors on temperature measurements . in order to derive the temperature and metallicity profile , we simultaneously fitted the pn , mos1 , and mos2 spectra with a single - temperature plasma model in each deprojected shell region .
we employed the mekal model ( mewe et al .
1985 , 1986 ; kaastra 1992 ; liedahl et al . 1995 ) and assumed the abundance ratio among different elements to have solar values .
the fits are acceptable for all the ten deprojected shell regions .
the deprojected temperature and metallicity profiles thus derived are illustrated in fig .
[ tpro ] and fig .
[ apro ] , respectively . compared with the result from the conventional annular spectral analysis
, the effect of the projection is clearly seen inside the @xmath4360 arcsec radius .
the temperature is consistent with being isothermal in the outskirts beyond @xmath43200 kpc , while towards the center it decreases monotonically and reaches 2.8 kev in the innermost shell .
note that the central minimum temperature agrees with the cut - off temperature in a cooling flow model that describes the rgs spectrum ( tamura et al . 2001 ) and the temperature of the x - ray filament detected with chandra ( fabian et al .
for usage in the following sections , we obtained an analytical formula that approximates the temperature profile .
we employed a formula given as @xmath67^{-1 } \
, \label{eq : tpro}\ ] ] which has been proposed by allen , schmidt , and fabian ( 2001 ) to give a good description to radial temperature profiles in relaxed clusters . leaving all the four parameters of @xmath68 , @xmath69 , @xmath70 , and @xmath71 as free parameters , we fitted the temperature profile derived from the deprojected spectra .
the parameters obtained from a @xmath72 minimization are @xmath73 ( kev ) , @xmath74 ( kev ) , @xmath75 ( arcsec ) , and @xmath76 , where the errors give the 90% confidence range for four parameters of interest ( @xmath77 ) .
the best - fit function and the error ranges for the individual radii are overlaid in fig .
[ tpro ] , showing that the function gives a good representation of the temperature profile . in the metallicity profile ( fig .
[ apro ] ) , the strong central concentration of metals found in the previous works ( tamura et al .
2001 ; ettori et al .
2002 ) is confirmed . as for the temperature profile ,
the metallicity profile was also fitted with a similar function given as @xmath78^{-3/2 } \ .
\label{eq : apro}\ ] ] we derived @xmath79 ( solar ) , @xmath80 ( solar ) , and @xmath81 ( arcsec ) , and the best - fit function is overlaid in fig .
[ apro ] .
an immediate way to derive the mass profile is using eq .
( [ eq : hydro ] ) with the temperature profile measured in sect .
4.3 and the density profiles of the icm which can be obtained from the double @xmath36-model profile fitted to the observed x - ray brightness profile in sect . 4.2 .
assuming spherical symmetry , the projected count rate profile in the 0.8 - 10 kev band , @xmath82 , gives the deprojected radial emissivity profile , @xmath83 , which can be converted to the icm density via the relation @xmath84 where @xmath85 is the emissivity coefficient defined by temperature ( @xmath86 ) , metallicity ( @xmath87 ) , energy range ( 0.8 - 10 kev ) , and the instrument responses .
taking into account the error ranges in the temperature profile , we thus derived the total gravitating mass profile , which is shown in fig .
[ masspro ] .
the result roughly agrees with the mass profile given by xu et al .
( 1998 ) derived from the _ asca _ data with the same method , although a shoulder - like structure found in their profile at @xmath88 kpc is less prominent in our profile .
the difference in the mass profiles is mainly attributed to the different parameters of the double @xmath36-model profile fitted to the brightness profile .
the total mass profile thus derived can not be conveniently used in eq .
( [ eq : diffeq ] ) for determining the dark matter velocity dispersion , however . the term of @xmath89 in eq .
( [ eq : diffeq ] ) is calculated from the third derivative of the icm density profile , i.e. the third derivative of the radial brightness profile which is approximated with the double-@xmath36 profile , and it gives an uneven profile at the radius where the two @xmath36 profiles cross over .
this causes an unstable behavior in the solution of eq .
( [ eq : diffeq ] ) and introduces non - negligible systematic errors in the results .
moreover , the double-@xmath36 modeling on the brightness profile can only reproduce a mass profile with the flat core as long as the temperature profile is virtually flat near the center , and provides rather restrictive mass profile . here
we make an alternative approach from the derivation above .
we start modeling the total mass profile with an analytical formula with a few free parameters that assures a smooth profile with eq .
( [ eq : diffeq ] ) . combined with the temperature profile derived in sect .
4.3 , the mass profile is used to predict the x - ray surface brightness profile of the cluster to be compared with the data to adjust the free parameters . in the actual calculation , a given total mass profile , @xmath22 , and a temperature profile , @xmath90 , is converted to the icm density profile from eq .
( [ eq : hydro ] ) as @xmath91\ , \label{eq : m2n}\ ] ] where @xmath92 is the central icm density . the emissivity profile in a given energy range , @xmath93 ,
is then obtained by eq .
( [ eq : n2epsilon ] ) , where @xmath33 , @xmath86 , @xmath87 are substituted by eqs .
( [ eq : m2n ] ) , ( [ eq : tpro ] ) and ( [ eq : apro ] ) , respectively , which is then converted to the brightness profile , @xmath94 .
being convolved with the psf , the model brightness profile is fitted to the 0.8 - 10 kev count rate profile observed to determine the best - fit parameters in the mass profile model .
although the observational constraint on the icm temperature profile is limited within 512 arcsec from the center , we apply eq .
( [ eq : tpro ] ) beyond the radius , where it describes a practically isothermal model .
as for the mass profile , we apply two representative models having a flat core and a cuspy core . the king model ( king 1966 ) is a classic model for the gravitational potential structure characterized by a flat core .
we used an approximated formula of the king model , in which the density profile is given as @xmath95^{-3/2 } \
, \label{eq : king_rho}\ ] ] while the integrated mass profile is given as @xmath96\ , \label{eq : king_mass}\ ] ] where @xmath97 is the central density and @xmath48 is the core radius ( see e.g. binney and tremaine 1987 ) .
the predicted x - ray brightness profile from the king model is fitted to the 0.8 - 10 kev count rate profile .
fitting parameters are the core radius ( @xmath48 ) and the central density ( @xmath97 ) of the approximated king model , and the central icm density ( @xmath92 ) . the best - fit model and the fit residuals are shown in fig . [ fit3 ] and
the parameters derived are summarized in table [ tab : mass_model ] .
the fit is not acceptable . in fig .
[ masspro ] , the best - fit king model is compared with the mass profile obtained in sect .
4.4 , showing clearly that a central mass excess is necessary in additional to the king model mass .
we then added another king - model component to account for the central mass excess to construct a double approximated king model , which still has a flat core profile .
fitting parameters are now two sets with a core radius and a central density for each ( @xmath98 , @xmath99 , @xmath100 , and @xmath101 ) , and the central icm density ( @xmath92 ) . as shown in fig .
[ fit3 ] and summarized in table [ tab : mass_model ] , a good fit was obtained .
another analytical formula for the total mass profile we employed is the universal halo profile by navarro et al .
( 1995 , 1996 , 1997 ; nfw model ) , which is characterized by a sharp central cusp .
the density profile of the nfw model is given as @xmath102^{-2}\ , \ ] ] where @xmath103 is called scale radius .
the integrated mass profile of the nfw model is given as ( makino et al .
1998 ; suto et al .
1998 ) @xmath104\ .\ ] ] the x - ray count rate profile predicted from the nfw model is fitted to the data , where the free parameters are @xmath97 , @xmath103 , and @xmath92 . as illustrated in fig .
[ fit3 ] and summarized in table [ tab : mass_model ] , the nfw model reproduces the data very well . for each best - fit mass profiles , the double approximated king model and the nfw model , the corresponding virial mass , @xmath105 , and the virial radius , @xmath106 ,
are calculated and given in table 2 .
the concentration parameters of the nfw model defined as @xmath107 is obtained to be 6.1 .
it is worth mentioning that the obtained @xmath108 and @xmath105 are consistent with the theoretically predicted @xmath109 relation ( e.g. bullock et al .
2001 ) . the double approximated king model and the nfw model obtained above
are compared in fig .
[ masspro_comp ] , together with the mass profile obtained in sect .
the three mass profiles are all consistent within the 20 - 600 arcsec radius region , and they approximately follow @xmath110 in this zone . in fig .
[ dens_pro ] , the double king model and the nfw model are compared in the mass density profile , together with the corresponding icm density profile . as clearly seen in fig .
[ masspro_comp ] and [ dens_pro ] , the two theoretical model profiles differ most significantly in the very central region within @xmath4320 arcsec radius , where xmm can not well resolve the spatial structure .
ettori et al .
( 2002 ) , using chandra data , obtained the total mass profile in a1795 with finer spatial resolution and found that the density approximately follows @xmath111 in the central region .
this profile is steeper than the king model ( @xmath112 ) and flatter than the nfw model ( @xmath113 ) near the center .
therefore , the two mass profiles give a conservative mass range . in the next subsection , for determining the dark matter velocity dispersion profile , we use both the double approximated king model and the nfw model for the total mass profile model , which
would also be expected to give a conservative range of the velocity dispersion representing systematic uncertainty involved in the usage of a specific mass profile model .
we now calculate the dark matter velocity dispersion profile by solving the differential equation given in eq .
( 3 ) . as the total mass profile , @xmath22 , the double approximated king model or the nfw model profile obtained in sect .
4.5 is employed , while the dark matter density profile is obtained from the total mass and the corresponding icm density profile by @xmath114 .
both @xmath22 and @xmath115 are substituted in eq .
( [ eq : diffeq ] ) and the differential equation is solved with the fourth - order runge - kutta method for various boundary conditions at the center , i.e. @xmath116 . figure [ setofsol ] shows a set of solutions of eq .
( [ eq : diffeq ] ) , when the best - fit nfw model is applied to @xmath22 .
note that there is only a weak observational constraint on the mass profile beyond @xmath117 mpc and the solution at the large radii depends on the validity of the assumed mass profile . depending on the inner boundary condition , the dark matter velocity dispersion may fall to zero at small radii , or it may diverge to infinity . in order to select physically plausible solutions among them , we placed a conservative boundary condition that the velocity dispersion at the virial radius is greater than 0 and less than the free fall velocity at this radius .
the constraints on the velocity dispersion profile thus obtained are shown in fig .
[ sigma_dm ] individually for different mass models applied .
the errors on the icm temperature profile are properly taken into account to obtain the error bands . in the central region , there is a clear difference between the results from the double approximated king model and the nfw model as expected from their different behavior .
the velocity dispersion is then converted to the `` temperature '' of the dark matter ( @xmath118 ) , which is compared with the icm temperature in fig .
[ dm_tpro ] . the icm temperature is greater than the dark matter `` temperature '' everywhere . even in the central region
where radiative cooling is expected to be most effective , the icm temperature is significantly higher than that of the dark matter .
the comparison of the temperatures can be more directly described by means of the @xmath2 value defined as @xmath119 , which ranges @xmath430.3 - 0.8 ( fig .
[ beta_pro ] ) .
there is no sign that the icm is cooled significantly below the dark matter `` temperature '' . in other words ,
the dark matter looks to form the temperature floor that limits the icm temperature . as a matter of fact ,
what we derived by solving the jeans equation is the velocity dispersion of the collisionless particles that includes the dark matter as well as the stellar component . as shown in fig .
[ masspro ] , the stellar component makes a minor contribution to the total mass and the derived velocity dispersion profile shown in fig .
[ dm_tpro ] is virtually that of the dark matter , except in the very central region , where the two components are comparable . if the galaxies alone are in a steady state , they should also obey the jeans equation under the same gravitational potential . in fig .
[ dm_tpro ] , we overlay the velocity dispersion profile of the member galaxies measured by den hartog and katgert ( 1996 ) , and the stellar velocity dispersion of the cd galaxy provided by blakeslee and tonry ( 1992 ) , after converted to `` temperature '' via @xmath120 .
they show roughly a consistent profile with that of the dark matter at least in the case when the total mass profile is given by the nfw model .
equation ( [ eq : jeans ] ) that we have used to calculate the dark matter velocity dispersion profile is based on the assumption of isotropic motion of dark matter particles .
however , simulation studies ( e.g. eke , navarro , and frenk 1998 ; coln , klypin , & kravtsov 2000 ) indicate that the radial velocity dispersion , @xmath121 , should be rather larger than the tangential velocity dispersion , @xmath122 .
the degree of anisotropy is often measured with @xmath123 and the jeans equation is modified as @xmath124 employing @xmath125 derived by coln , klypin , & kravtsov ( 2000 ) , which is as large as 0.5 at the virial radius , @xmath106 , and converging to 0 at the center , we solved eq .
[ eq : jeans2 ] as done in sect .
the @xmath0 profile thus derived for the case of the best fit nfw mass profile is shown in fig .
[ sigma_dm2 ] , overlaid with the @xmath126 solution given in fig . [ sigma_dm ] .
the @xmath127 value with the anisotropy is greater than that without the anisotropy by only @xmath436% at 100kpc and at most @xmath4340% around 1mpc .
the @xmath2 profile determined here from observations should provide information on the thermal history of the icm . from numerical simulation studies ,
@xmath128 is expected , if there is no cooling or additional heating ( e.g. metzler and evrard 1994 ; navarro et al .
1995 ; bryan & norman 1998 ) .
an obvious way to explain the @xmath2 value smaller than unity in a1795 is heating of the icm .
as suggested from the break of the self - similarity between dark matter and icm , there should have been non - gravitational heating acting globally .
we , from our results , estimated the excess energy of the icm over that of the dark matter particles as @xmath129 where @xmath130 denotes mass weighted mean within radius , @xmath23 .
as shown in fig .
[ mean_de_pro ] , the excess energy , @xmath131 , thus derived is found to be @xmath7 kev per particle , which may be compared with theoretical model predictions .
the amount of energy injection to the gas phase that explains e.g. the observed @xmath8 relation depends on the period when the heating occurred .
heating prior to cluster collapse , `` preheating '' , needs @xmath132 kev per particle ( e.g. navarro et al . 1995 ; tozzi & norman 2001 ) , while heating after a cluster formation requires higher values of @xmath133 kev per particle ( e.g. metzler & evrard 1994 ; loewenstein 2000 ; wu , fabian , & nulsen 2000 ; bower et al .
our results given above may indicate that the global non - gravitational heating that may cause the break of self similarity has happened mainly within a collapsed cluster .
however , even if such non - gravitational heating that explains the global x - ray feature of the cluster is provided , the central region of a1795 has a short radiative cooling time ( fig .
[ coolt ] ) , and there must be another significant energy input to the icm in the central region at the present epoch to prevent the icm from cooling . as a possible energy source in the central region ,
we first consider gravitational energy of the member galaxies and stellar components therein .
the kinetic energy of the random motion of stars can be partially transferred to the icm by stellar mass loss .
gas supplied by stellar mass loss has velocities of the bulk motions of stars relative to the icm , which is the sum of a galaxy motion and the motions of the stars in the galaxy , and is likely to be thermalized by interactions with the ambient gas . if the stellar component moving in the same gravitational potential has a similar velocity dispersion profile as that of the dark matter ( fig .
[ dm_tpro ] ) , the gas temperature achieved from this process is expected to be comparable to the velocity dispersion of the dark matter .
this process nicely accounts for x - ray emission from isolated x - ray compact elliptical galaxies ( matsushita 2001 )
. the input rate of the kinetic energy of the gas from stellar mass loss may be simply estimated as @xmath134 = @xmath135 ergs / s , where @xmath136 is the total stellar mass of @xmath137 m@xmath38 , @xmath138=@xmath139 m@xmath38 yr@xmath17 ( @xmath140m@xmath38)@xmath17 is the stellar mass loss rate in unit time , and @xmath141(=580@xmath14280 km / s ) is the velocity dispersion of the galaxies from den hartog & katgert ( 1996 ; fig .
[ dm_tpro ] ) .
( the stellar velocity dispersions in the galaxies are smaller and neglected here . )
this is much smaller than the output energy in the central region by x - ray radiation in galaxy clusters like a1795 , however .
thus for the case of galaxy clusters , we have to look for additional heat sources .
kinetic energy of the stellar component might be more efficiently transferred to icm via magnetic fields . as pointed out by e.g. makishima et al .
( 2001 ) , the motion of stars may amplify interstellar magnetic fields and reconnections of the fields may heat up the icm rather efficiently .
the galaxies must have lost their kinetic energies through interactions with the icm and have gradually fallen inwards accumulating onto the central galaxy to form the cd galaxy . a deep optical image of the cd galaxy in a1795 derived by johnstone et al .
( 1991 ) shows a concentration of elliptical galaxies of various sizes and stars forming a largely extended envelope with 131 @xmath143 kpc effective radius , which strongly suggests the on - going formation process of the cd galaxy .
quantitatively , the total amount of dynamical energy of the stellar component in the member galaxies that has been lost in the past is estimated . the stellar component in the galaxies
is assumed to have a negligible small potential ( @xmath144 ) and kinetic energy ( @xmath145 ) before the formation of the cluster , and the current energy of the stars is estimated to be @xmath146 ergs .
if the energy has been released over the last 10 gyr and has been spent in icm heating , the heating luminosity is expected to be @xmath147 ergs s@xmath17 .
this amounts to the bolometric luminosity of the icm within 60@xmath143 kpc ( fig .
[ coolt ] ) , and may be sufficient to sustain the thermal energy of the icm against radiative cooling .
this model predicts that the stellar velocity dispersion became smaller than that of icm , i.e. @xmath148 , which is consistent with the actual observed value in the central region ( fig .
[ dm_tpro ] ) .
alternatively , there may be sufficient non - gravitational heat input from an active galactic nucleus ( agn ) powered by an accretion of low entropy gas at a cluster center ( churazov et al .
2002 ; bhringer et al . 2002 ) .
numerical simulations show that an outflow from an agn may form hot bubbles of relativistic plasma rising with buoyancy , and the bubbles may uplift cold gas mixing with ambient icm ( churazov et al .
2001 ; quilis , bower , & balogh 2001 ; brggen & kaiser 2002 ; basson & alexander 2003 ) .
high resolution imaging observations revealed ripples and shock features in the central region of perseus cluster ( fabian et al .
2003 ) and m87 ( forman et al .
2004 ) , suggesting that the bubble energy may also be transfered by sound wave to larger distances , and that shocks may be the major contribution of the energy dissipation .
the heating mechanism is self regulated : the lower the entropy , the higher the accretion rate onto the central engine . a portion of the accretion power is dissipated back into the icm to make its entropy high and regulate the accretion rate to achieve an equilibrium state .
this process automatically prevents the persistence of cold and hence dense clouds .
the cd galaxy of a1795 has a radio source , 4c26.42 , and the existence of an agn is evident . using the physical state of the icm in the center we can actually estimate the energy provided by the agn by applying the bondi accretion model .
according to the well - known correlation of the black hole mass with the mass of the bulge component ( magorrian et al .
1998 ) , the black hole mass is expected to be @xmath149 m@xmath38 .
assuming that the gas profile is flat in the center , we can use the measured values of @xmath33=0.1 @xmath150 and @xmath68=2.8 kev to obtain the bondi mass accretion rate @xmath151 where @xmath152 and @xmath153 are the density and sound velocity outside the bondi accretion radius .
we find @xmath154 . under the standard assumption of 10% of the accretion energy to be dissipated ,
the output energy is found to be @xmath155 ergs / s .
this amounts to the x - ray luminosity within the inner @xmath156 kpc region only .
the bondi accretion radius is estimated to be @xmath157 30 pc , much smaller than the resolution of the temperature and density structure that can be measured with xmm - newton .
if the icm density is not uniform but is clumpy , the bondi accretion rate should be significantly larger and the heating rate could also be larger than the above estimation .
we can note that the icm conditions might temporally vary and that we currently see a relatively low state .
note that there are other heating mechanisms discussed , which include e.g. a classical idea of thermal conduction from the hot outer regions ( e.g. takahara & takahara 1979 ; tucker & rosner 1983 ; bregman , & david 1988 ; gaetz 1989 ) , and the acoustic wave heating recently proposed by fujita , suzuki , & wada ( 2004 ) .
comparison of those model predictions with our results presented in the current paper would be very interesting .
we derived the dark matter velocity dispersion profile from an x - ray observation for the first time . using the xmm - newton epic data of a1795 , we derived @xmath2 of @xmath158 , indicating that the icm temperature is larger than the dark matter `` temperature '' everywhere .
we also derived the excess energy in the icm , @xmath131 , which is found to be @xmath5 kev per particle .
these observational quantity should be useful to provide a new clue to solve the heating mechanism in the cooling core and to understand the thermal history of a galaxy cluster .
we thank the xmm team for providing the data and the data analysis tools .
in particular , we express our gratitude to michael freyberg for helping the data analysis .
we also acknowledge ftools .
we thank kuniaki masai for helpful discussions .
we are also grateful to paul lynum for critically reading this manuscript .
allen , s. w. , fabian , a. c. , johnstone , r. m. , arnaud , k. a. & nulsen , p. e.
j. 2001 , , 322 , 589 allen , s. w. , schmidt , r. w. , & fabian , a. c. 2001 , , 328 , l37 allen , s. w. & fabian , a. c. 1994 , , 269 , 409 arnaud , m. , neumann , d. m. , aghanim , n. , gastaud , r. , majerowicz , s. , & hughes , j. p. 2001
, , 365 , l80 basson , j. f. , & alexander , p. 2003
, , 339 , 353 binney , j. , & tremaine , s. 1987 , galactic dynamics , princeton university press blakeslee , j. p. , & tonry , j. l. 1992 , , 103 , 1457 bhringer , h. , briel , u. g. , schwarz , r. a. , voges , w. , hartner , g. & trumper , j. 1994 , nature , 368 , 828 bhringer , h. , matsushita , k. , churazov , e. , ikebe , y. , & chen , y. 2002 , , 382 , 804 boute , d. a. , & tsai , j. c. 1996 , , 458 , 27 bower , r. g. , benson , a. j. , lacey , c. g. , baugh , c. m. , cole , s. , & frenk , c. s. 2001 , , 325 , 497 bregman , j. n. , & david , l. p. 1988
, , 326 , 639 briel , u. g. , & henry , j. p. 1996
, , 472 , 131 brggen , m. , & kaiser , c. r. 2002 , nature , 418 , 301 bryan , g. l. & norman , m. l. 1998 , , 495 , 80 bullock , j. s. , kolatt , t. s. , sigad , y. , somerville , r. s. , kravtsov , a. v. , klypin , a. a. , primack , j. r. , & dekel , a. 2001 , , 321 , 559 burkert , a. 1995 , apj , 447 , l25 canizares , c. r. , stewart , g. c. , & fabian , a. c. 1983 , , 272 , 449 churazov , e. , brggen , m. , kaiser , c. r. , bhringer , h. , & forman , w. 2001 , , 554 , 261 churazov , e. , sunyaev , r. , forman , w. , bhringer , h. 2002 , , 332 , 729 coln , p. , klypin , a. a. , kravtsov , a. v. 2000 , , 539 , 561 cowie , l. l. , & binney , j. 1977 , , 215 , 723 david , l. p. , slyz , a. , jones , c. , forman , w. , vrtilek , s. d. , & arnaud , k. a. 1993 , , 412 , 479 david , l. p. , jones , c. , forman , w. & daines , s. 1994 , , 428 , 544 den hartog , r. , & katgert , p. 1996
, , 279 , 349 edge , a. c. , & stewart , g. c. 1991 , , 252 , 414 edge , a. c. , stewart , g. c. , & fabian , a. c. 1992 , , 258 , 177 eke , navarro , & frenk ( 1998 ) , apj , 503 , 569 ettori , s. , fabian , a. c. , allen , s. w. , & johnstone , r. m. 2002 , , 331 , 635 fabian , a. c. , hu , e. m. , cowie , l. l. , & grindlay , j. 1981 , , 248 , 47 fabian , a. c. , nulsen , p. e. j. , & canizares , c. r. 1984 , nature , 310 , 733 fabian , a. c. 1994 , ara&a , 32 , 277 fabian , a. c. , arnaud , k. a. , bautz , m. w. , & tawara , y. 1994 , , 436 , l63 fabian , a. c. , mushotzky , r. f. , nulsen , p. e. j. , & peterson , j. r. 2001a , , 321 , l20 fabian , a. c. , sanders , j. s. , ettori , s. , taylor , g. b. , allen , s. w. , crawford , c. s. , iwasawa , k. , & johnstone , r. m. 2001b , , 321 , l33 fabian , a. c. , sanders , j. s. , allen , s. w. , crawford , c. s. , iwasawa , k. , johnstone , r. m. , schmidt , r. w. , and taylor , g. b. 2003 , , 344 , l43 forman , w. et al .
astro - ph/0312576 fujita , y. , suzuki , t. k. , & wada , k. 2004 , , 600 , 650 fukushige , t. , & makino , j. 2001 , , 557 , 533 gaetz , t. j. 1989 , , 345 , 666 ghizzardi , s 2001 , xmm - soc - cal - tn-0022 , http://xmm.vilspa.esa.es/docs/documents/cal-tn-0022-1-0.ps.gz hughes , j. p. 1989
, , 337 , 21 ikebe , y. , makishima , k. , fukazawa , y. , tamura , t. , xu , h. , ohashi , t. , & matsushita , k. 1999 , , 525 , 58 johnstone , r. m. , naylor , t. , & fabian , a. c. 1991 , , 248 , 18p jones , c. , & forman , w. 1984 , , 276 , 38 kaastra , j.s .
1992 , an x - ray spectral code for optically thin plasmas ( internal sron - leiden report , updated version 2.0 ) kaastra , j. s. , ferrigno , c. , tamura , t. , paerels , f. b. s. , peterson , j. r. , & mittaz , j. p. d. 2001 , , 365 , l99 kaiser , n. 1986 , , 222 , 323 kitayama , t. , & suto , y. 1996 , , 469 , 480 king , i. r. 1966 , , 71 , 276 liedahl , d. a. , osterheld , a. l. & goldstein , w. h. 1995 , , 438 , l115 loewenstein , m. 2000 , , 532 , 12 magorrian , j. , tremaine , s. , richstone , d. , bender , r. , bower , g. , dressler , a. , faber , s. m. , gebhardt , k. , green , r. , grillmair , c. , kormendy , j. , & lauer , t. 1998 , , 115 , 2285 makino , n. , sasaki , s. , & suto , y. 1998 , , 497 , 555 makishima , k. et al .
2001 , pasj , 53 , 401 matsushita , k. , 547 , 693 metzler , c. a. , & evrard , a. e. 1994 , , 437 , 564 mewe , r. , gronenschild , e. h. b. m. , & van den oord , g. h. j. 1985 , a&as , 62 , 197 mewe , r. , lemen , j. r. , & van den oord , g. h. j. 1986 , a&as , 65 , 511 moore , b. , quinn , t. , governato , f. , stadel , j. , lake , g. et al . 1999 , , 310 , 1147 navarro , j. f. , frenk , c. s. , & white , s. d. w. 1995 , , 275 , 720 navarro , j. f. , frenk , c. s. , & white , s. d. m. 1996 , , 462 , 563 navarro ,
j. f. , frenk , c. s. , & white , s. d. m. 1997 , , 490 , 493 peterson , j. r. et al .
2001 , , 365 , l104 ponman , t. j. , cannon , d. b. , & navarro , j. f. 1999 , nature , 397 , 135 quilis , v. , bower , r. g. , & balogh , m. l. 2001 , , 328 , 1091 suto , y. , sasaki , s. , & makino , n. 1998 , , 509 , 544 struble , m. f. , & rood , h. j. 1987 , , 63 , 543 strder et al .
2001 , , 365 , 18l tamura , t. et al .
2001 , , 365 , l87 takahara , m. , & takahara , f. 1979 , prog .
, 62 , 1253 tozzi , p , & norman , c. 2001 , , 546 , 63 tucker , w. h. , & rosner , r. 1983 , , 267 , 547 white , d. a. , jones , c. , & forman , w. 1997 , , 292 , 419 wu , k. k. s. , fabian , a. c. , & nulsen , p. e.
j. 2001 , , 318 , 889 wu , x. , xue , y. , & fang , l. 1999 , , 524 , 22 xu , h. , makishima , k. , fukazawa , y. , ikebe , y. , kikuchi , k. , ohashi , t. , & tamura , t. 1998 , , 500 , 738 lll @xmath159 & @xmath160 & [ c / s / arcsec@xmath161 + @xmath98 & 26.3 @xmath142 1.0 & [ arcsec ] + @xmath162 & 0.562 @xmath142 0.015 & + @xmath163 & @xmath164 & [ c / s / arcsec@xmath161 + @xmath100 & 106 @xmath142 4 & [ arcsec ] + @xmath165 & 0.633 @xmath142 0.010 & + @xmath166 & 83.2/74 & + ccccccccc 1 king & @xmath167 & 94.6 & & & 0.063 & 1124/77 & - & - + 2 king & @xmath168 & @xmath169 & @xmath170 & @xmath171 & @xmath172 & 88.3/75 & @xmath173 ( @xmath174 ) & @xmath175 + & @xmath97 & @xmath103 & & & & & & + nfw & @xmath176 & @xmath177 & & & @xmath178 & 88.7/77 & @xmath179 ( @xmath180 ) & @xmath181 + | we present a method from an x - ray observation of a galaxy cluster to measure the radial profile of the dark matter velocity dispersion , @xmath0 , and to compare the dark matter `` temperature '' defined as @xmath1 with the gas temperature .
the method is applied to the xmm - newton observation of abell 1795 .
the ratio between the specific energy of the dark matter and that of the intracluster medium ( icm ) , which can be defined as @xmath2 in analogy with @xmath3 , is found to be less than unity everywhere ranging @xmath4 . in other words , the icm temperature is higher than the dark matter `` temperature '' , even in the central region where the radiative cooling time is short .
a @xmath2 value smaller than unity can most naturally be explained by heating of the icm .
the excess energy of icm is estimated to be @xmath5 kev per particle . |
_ ginga _ , _ asca _ and recent _ bepposax _ x - ray observations of seyfert 1 galaxies strongly support a general _ observational _ framework that includes : ionized absorption by a warm absorber ( wa ) , a steep intrinsic power - law continuum , a compton reflection hump and associated neutral fek@xmath7 line ( narrow and/or broad ) , and a high - energy cutoff ( @xmath8 ) .
these observational results have raised the following questions : * what is the location of the x - ray reprocessor ?
is it the accretion disk , the broad - line region ( blr ) and/or the molecular torus ? is the answer the same for all seyfert 1s ? * where is the location of the wa ?
is it within the blr and/or the molecular torus or is it external to them ?
is there only one wa ?
* does @xmath8 differ from source - to - source ( as currently shown by _
bepposax _ ,
piro 1999 , matt 2000 ) or is it variable with time and/or flux in any single object ?
in order to address these questions , we have performed multiple ( 4 ) observations of the prototypical seyfert 1 galaxy ic4329a to obtain time - resolved spectral constrains on the wa , reflection continuum , fe k@xmath9 line and , possibly , @xmath8 .
this is potentially the best way ( see also weaver 2000 ) to establish whether these observational features are produced from material near the source ( e.g. the accretion disk , in which case short lags , @xmath101000 s , are expected between the reflection and continuum variations ) or farther away ( e.g. in the blr or torus , in which case longer lags of at least @xmath0 10@xmath11 s are expected ) .
ic 4329a @xmath12 is a seyfert 1 galaxy well studied in x - rays ( miyoshi et al . 1988 ; piro , yamauchi & matsuoka 1990 ; fiore et al . 1992 ; cappi et al .
1996 ) and hard x - rays ( fabian et al . 1993 ; madejski et al . 1995
; zdziarski et al .
it offers the best opportunity to address the above mentioned variability issues because it is among the brightest known seyferts in x - rays ( f(2 - 10kev)@xmath01.6@xmath210@xmath3erg @xmath4s@xmath5 ) and shows only _ moderate _
variability on short ( @xmath10 day ) timescales , but large variability on longer timescales ( a factor 2 - 3 in a few days ) , both below and above 10 kev ( done , madejski & zycki 2000 ; fabian et al .
1993 ) .
_ bepposax _ observed ic 4329a in 1996 for 100 ks ( perola et al .
1999 ) , and then subsequently 3 times in 1998 , with the 3 observations spaced approximately 5 days apart .
as expected , the source did _ not _ vary much _ within _ each observation .
unfortunately , it did not vary much _ between _ the 4 observations either ( see fig .
1 ) , despite the large and systematic flux variations typically observed in ic4329a by rxte ( done , madejski & zycki 2000 ) .
spectral ratios were produced between all observations , in search of any model - independent spectral variations ( fig .
we find marginal evidence for variability of the spectral features at @xmath13 kev and at @xmath1410 kev , but we find no clear variations in the fek@xmath9 and reflection component intensities , as well as in the high-@xmath8 .
the variations at low energies are most likely due to variability of the ionization state of the wa responding to continuum variations , but complications due to the presence of a possible scattered component and/or intrinsic soft ( excess ) component can not be ruled out .
spectral fitting of each observation with a complex model including 2 absorption edges , a steep power - law continuum , a reflection component with associated fek@xmath6 , and a high-@xmath8 gave best - fit parameters similar to those reported by perola et al .
( 1999 ) with somewhat larger errors ( e.g. @xmath15 0.1 , r @xmath16 , @xmath8 @xmath0 250 @xmath17 100 kev ) .
3 shows the marginal variations of the fek line intensity and equivalent width .
4 shows the ( similar ) best - fit broad - band spectra of two different observations .
the most interesting result of the present analysis is that despite a 30% flux increase between obs . 1 and obs
. 2 , the line intensity became surprisingly weak ( compare fig . 1 and fig .
3 ) . during the higher flux obs .
3 the line re - established its typical value of @xmath0 100 - 150 ev , though . in other words , there is some evidence ( at a @xmath0 2 @xmath18 significance level ) that the fe k@xmath6 line _ does not _ follow `` instantaneously '' ( at least within 40 ks ) the continuum variations , but it _ does _ on a timescale between 410 days .
* moderate variations of the absorption structure , most likely related to the warm absorber , are apparently detected but require further investigation for more quantitative conclusions . * despite the unprecedented statistics of these observations , we find _ no _ significant spectral variations of the compton reflection continuum , the power law spectral index and the high - enegy cutoff .
this is probably because _ bepposax _ unfortunately caught ic 4329a at a not too different flux level ( 30% variation ) in all four observations hampering , thereby , any spectral variability .
* there is marginal evidence that the fek@xmath6 line is being produced by the outer regions of the accretion disk in ic 4329a , thereby suggesting a possible explanation for the lack of a clear broad fek@xmath6 line in this object .
cappi m. , mihara t. , matsuoka m. , et al .
, 1996 , apj , 458 , 149 done , c. , madejski , g. m. and ycki , p. t. 2000 , apj , 536 , 213 fabian , a.c .
et al . , 1993 ,
apj , 416 , l57 fiore , f. , perola , g.c . , matsuoka , et al .
, 1992 , a&a , 262 , 37 miyoshi , s. , et al .
, 1988 , pasj , 40 , 127 madejski , g.m . , et al .
1995 , apj , 438 , 672 matt , g. 2000 , these proceedings , astro - ph/0007105 perola , g.c . , matt , g. , cappi , m. , et al . , a&a , in press piro , l. , yamauchi , m. , & matsuoka , m. , 1990 , apj , 360 , l35 piro , l. 1999 , in `` heating and acceleration in the universe '' , astron .
, in press weaver , k. 2000 , these proceedings , astro - ph/0007327 | ic 4329a is the brightest known seyfert galaxy in hard ( @xmath0 2 - 30 kev ) x - rays and is likely to be representative of seyfert 1 galaxies as a class .
a recent 100 ks _ bepposax
_ observation ( perola et al . 1999 ) clearly confirmed the presence of a warm absorber , a reflection component ( r @xmath0 0.6 ) , and a high - energy cut - off in the power law at e@xmath1 @xmath0 270 kev .
its richness in spectral features , combined with its large flux ( @xmath0 1.6 @xmath2 10@xmath3 erg @xmath4 s@xmath5 between 2 - 10 kev ) , make this target ideal for multiple observations ( in particular with _ bepposax _ ) to search for spectral variations .
results obtained from 3 follow - up observations ( 40 ks each ) are presented here .
the first and most important goal of this study was to probe the origin of the compton reflection component observed in seyfert galaxies by monitoring the variability of the reflection continuum and fe k@xmath6 line in response to primary continuum variations .
the second goal was to search for variability in the high energy cutoff .
we obtain however no conclusive results on any of these issues .
in fact , all four observations unfortunately caught the source at almost the same flux , showing only little , and marginal , spectral changes between different observations . |
the advent of new and high statistics deeply inelastic neutrino scattering experiments apart from providing valuable data on @xmath1 and @xmath2@xcite has shown again the importance of nuclear effects that render difficult the extraction of the nucleon structure functions .
furthermore , the use of nuclear targets is necessary due to the smallness of the cross sections .
qualitatively , the nuclear effects are well known .
shadowing , antishadowing , fermi motion , binding , ... have all been widely studied for charged leptons in the context of the european muon collaboration ( emc ) effect . for a review ,
see @xcite .
moreover , it is well known that nuclear effects are substantially different for neutrino reactions due to the presence of the axial current and the different valence and sea quark contributions for each observable @xcite .
this situation asks for a detailed and quantitative microscopical understanding of the nuclear effects , rather than the parametrizations that have been used sometimes , like recently by the nutev collaboration @xcite .
one of the basic ingredients in all calculations is the nuclear spectral function .
this presents some serious difficulties , as these functions are not so well known to the precision level reached by current experiments .
therefore , the analysis might introduce unwanted model dependences .
however , it was soon noticed that under some approximations the nuclear structure functions could be written as simple expansions on the nucleon structure functions and their derivatives .
all the nuclear information would then be encoded in the expected values of some nuclear magnitudes , like the average kinetic energy of the nucleons or the mean nucleon removal energy .
these kind of approximations , both for charged lepton and for neutrino induced reactions , have been widely used in the literature @xcite .
the aim of this paper is to investigate the quality of these expansions and to what extent they could be used in the analysis of lepton deep inelastic scattering experiments . in order to do that ,
once the formalism is established in the next section , we will make a comparative study for @xmath3 and for a few typical nuclear spectral functions used in the literature .
the nuclear structure functions can be written as a convolution of the nuclear spectral functions and nucleon structure functions .
see for instance @xcite and references therein . in the rest frame of the nucleus ,
the @xmath1 and @xmath4 structure functions are @xmath5 where @xmath6 is the nuclear spectral function , normalized to the number of protons ( neutrons ) in the nucleus , and describes the probability of finding a proton ( neutron ) with momentum @xmath7 and removal energy @xmath8 .
the four - momentum of the nucleon can be written as @xmath9 , with @xmath10 .
the _ z _ axis is oriented in such a way that @xmath11 lies on it , @xmath12 is the transverse momentum of the nucleon and @xmath13 . here , @xmath14 is the natural bjorken variable for the nucleon in the nuclear medium , i.e. @xmath15 ; while @xmath0 is the bjorken variable in the nucleon rest frame , @xmath16 .
they are related by @xmath17 for isoscalar nuclei such as @xmath18 , only the isoscalar component of the spectral function and the structure function have to be accounted for . when this is done ( [ eq:1 ] ) and ( [ eq:2 ] ) ,
these read as @xmath19 where @xmath20 , which is the isoscalar part of the nuclear spectral function , is now normalized to unity and we perform the calculations for the nuclear structure functions averaged over neutrinos and antineutrinos , i.e. , we only consider the symmetric @xmath21 combination @xcite . in a nucleus
, the expected values of @xmath22 and @xmath23 averaged with the nuclear spectral function are much smaller than unity .
thus , @xmath24 and @xmath25 . under these assumptions
, we can perform a taylor expansion of the integrands in expressions ( [ eq:4 ] ) and ( [ eq:5 ] ) around @xmath26 , keeping terms up to order @xmath27 and @xmath28 . in this way we will be able to take out of the integral the structure functions and their derivatives and
we will be left with expected values of the removal energy @xmath8 and momentum squared @xmath29 .
this statement is true if the nuclear spectral functions only depend on the modulus of the momentum @xmath30 ( as it is in the case of the nuclear spectral functions we will consider ) and not upon its direction . under this assumption
we can drop the expected values of @xmath31 ( or any other component of the momentum ) and @xmath32 because they are identically zero due to symmetry considerations .
we will begin with the structure function @xmath33 .
after performing the taylor expansion , keeping terms up to order @xmath27 and @xmath28 , and dropping those terms which go with @xmath34 or @xmath35 ( because they are identically zero as stated above ) , we obtain @xmath36 -x\frac{\partial f_2^n\left ( x , q^2\right ) } { \partial x}\left [ \frac{\left\langle \epsilon\right\rangle } { m}+\left ( 2-\gamma^2\right)\frac{\left\langle \mathbf{p}^2\right\rangle } { 3m^2 } \right ] \nonumber\\ \fl & + & \left ( x\frac{\partial f_2^n\left ( x , q^2\right ) } { \partial x}+\frac{x^2}{2}\frac{\partial^2f_2^n\left ( x , q^2\right ) } { \partial x^2}\right ) \gamma^2\frac{\left\langle \mathbf{p}^2\right\rangle } { 3m^2}\label{eq:6},\end{aligned}\ ] ] where @xmath37 is the expected value of the operator @xmath38 averaged with the isoscalar spectral function @xmath20 : @xmath39 we have also used the fact that , with a nuclear spectral function which depends only on the modulus of the momentum , the expected value of momentum squared is shared equally among every squared component , i.e : @xmath40 . in the bjorken limit , ( @xmath41 ) , ( [ eq:6 ] ) coincides with the expansions used by frankfurt _
_ @xcite and ciofi degli atti _
. if we do the same for the nuclear structure function @xmath42 , we obtain @xmath43+\gamma^2\left [ x\,\frac{\partial f_3^n}{\partial x}+\frac{x^2}{2}\,\frac{\partial^2f_3^n}{\partial x^2}\right ] \right\rbrace \label{eq:9},\end{aligned}\ ] ] a similar result to that kulagin @xcite . to allow for an easier comparison , in the above expressions
, @xmath44 can be rewritten as @xmath45 then , the only difference between ( [ eq:9 ] ) and that of @xcite is that , for simplicity , we do not consider the off - shell dependence in the nucleon structure function @xmath46 . in our calculation , the only source of off - shell dependence is through the nuclear spectral function @xmath47 .
thus , we do not have the term @xmath48 that appears in @xcite . for our study
we have selected the @xmath18 nucleus , which is isoscalar and already will show important medium effects .
the free nucleon structure functions have been taken from @xcite and we have chosen several different nuclear spectral functions .
the first one ( labelled i ) is a phenomenological model of the spectral function which has a mean field part and high - momentum components coming from nn - correlations and it is described in kulagin _
et al . _
@xcite where it was used in a global study of nuclear structure functions .
the second one ( labelled ii ) , that also contains correlations is taken from @xcite where it has been tested in the calculation of several electron scattering observables .
the third spectral function is taken from the semiphenomenological model based on a many body calculation and the local density approximation that is described in @xcite ( labelled iii ) .
finally , we also consider the simple mean field spectral function of @xcite(labelled iv ) , which was used in the study of the emc effect .
in figure [ fig1 ] , we show the results of the ratio @xmath49 for the nuclear spectral functions ( i ) , ( ii ) and ( iii ) .
this gives us some idea of the uncertainties related to these functions .
the differences are small , even when the spectral funtions have been obtained with diverse methods and are in fact quite different if one studies in detail their energy and momentum dependence
. however , .expected values of the nucleon removal and kinetic energies for the nuclear spectral functions of kulagin _ et al . _
@xcite ( i ) , ankowski _ et al . _
@xcite ( ii ) and fernandez de cordoba _ et al . _
@xcite ( iii ) . [ cols="<,^,^,^",options="header " , ] in particular , spectral functions ( i ) and ( iii ) that have quite close expected values also produce very similar ratios , whereas ( ii ) which has an appreciably smaller binding energy gives slightly larger values for the ratio . in figure [ fig2 ]
, we compare the full results for the same ratio ( [ eq:4 ] ) whith those obtained making use of the approximation of ( [ eq:6 ] ) .
the use of this expansion has been assumed to be a good approximation for @xmath53 @xcite .
obviously , the series expansion agrees well at low values of @xmath0 with the full results .
however , in the three cases produces lower values for the ratio at intermediate @xmath0 showing a maximum deviation of a 3 - 4@xmath54 around @xmath55 .
this region , with a dip in the ratio , is dominated by the mean removal energy ( or equivalently the binding energy ) per nucleon .
although this could look a small error , we should remark that it means increasing the deviation due to nuclear effects from the value @xmath56 by around a 30% . at higher @xmath0 s , where fermi motion of the nucleons provides the dominant effect , the series expansions grow faster than the full results and
become larger for @xmath57 .
thus , we find that for typical nuclear spectral functions the convergence of the series expansion is not so good except at very low @xmath0 , where in any case other effects not considered here , like shadowing play a major role .
the results for the @xmath58 ratio are shown in figure [ fig3 ] . the full model for @xmath59 is given by ( [ eq:5 ] ) and the series expansion by ( [ eq:9 ] ) .
as it was the case for @xmath60 , there is a dip region dominated by the binding energy and the fermi motion of the nucleons produces the large rising at high values of @xmath0 .
the comparison of the full results and the series expansions shows the same features as for @xmath60 .
the expansions systematically underestimate the ratios at intermediate values of @xmath0 and overestimate them for @xmath61 .
this overestimation of the effect of the fermi motion was already discussed in @xcite .
there , it was also claimed that in the limit of high @xmath62 and for heavy nuclei the expansion should be a good approxiamtion up to @xmath63 .
however , after studying @xmath60 and @xmath59 at different @xmath62 values , apart from the one shown in this paper , we have found that for medium nuclei the convergence of the series is only good at very low @xmath0 where other nuclear effects are very relevant . in order to obtain a better convergence one would need to reach a higher order in the expansions .
this implies also the sensitivity to nuclear expected values of higher powers of the nucleons momenta as @xmath64 .
one should notice that due to correlations these expected values are not negligible @xcite .
the three spectral functions considered above contain high momentum components and one may expect any expansion up to order @xmath65 to fail to have a good convergence .
to test this point , we have also calculated @xmath60 for the case of a simpler mean field spectral function which does not incorporate nucleon correlations and does not have those high momentum components @xcite .
this function was used to analyse the a dependence of the position and magnitude of the dip . here
, we have used the same parton distribution functions as in the original reference @xcite .
the results are shown in figure [ fig4 ] . in this case
, there is an almost perfect agreement between the full results and the expansion . in summary , we have studied the quality of some series expansions commonly used to incorporate approximately the nuclear effects in the analysis of dis processes .
we have found that for realistic enough nuclear spectral functions , that include nucleon correlations and have high momentum components , the convergence of the series to the full result is poor except at very low values of @xmath0 , where in fact , other nuclear effects , like shadowing or antishadowing , are more relevant . at high @xmath0 values , in the fermi motion region ,
the expansions clearly overestimate the full result .
this was known and expected . however , even at relatively low @xmath0 values , where the expansions were suppossed to provide a good approximation , we have found that they systematically underestimate the value of @xmath1 and @xmath4 , artificially increasing the size of the nuclear corrections .
this work was partially supported by the mec contract fis2006 - 03438 and the eu integrated infrastructure initiative hadron physics project contract rii3-ct- 2004 - 506078 . i.r.s .
acknowledges mec for additional support .
d. f. geesaman , k. saito and a. w. thomas , ann .
nucl . part .
* 45 * ( 1995 ) 337 .
n. armesto , j. phys .
g * 32 * ( 2006 ) r367 [ arxiv : hep - ph/0604108 ] .
i. schienbein , j. y. yu , c. keppel , j. g. morfin , f. olness and j. f. owens , phys .
d * 77 * ( 2008 ) 054013 [ arxiv:0710.4897 [ hep - ph ] ] .
j. r. smith and g. a. miller , phys .
c * 65 * ( 2002 ) 055206 [ arxiv : nucl - th/0202016 ] . c. ciofi degli atti
, l. l. frankfurt , l. p. kaptari and m. i. strikman , phys .
c * 76 * ( 2007 ) 055206 [ arxiv:0706.2937 [ nucl - th ] ] .
s. a. kulagin and r. petti , phys .
d * 76 * ( 2007 ) 094023 [ arxiv : hep - ph/0703033 ] .
s. a. kulagin and r. petti , nucl .
a * 765 * ( 2006 ) 126 [ arxiv : hep - ph/0412425 ] .
m. gluck , e. reya and a. vogt , eur .
j. c * 5 * ( 1998 ) 461 [ arxiv : hep - ph/9806404 ] . | we study the convergence of the series expansions sometimes used in the analysis of the nuclear effects in deep inelastic scattering ( dis ) proccesses induced by leptons .
the recent advances in statistics and quality of the data , in particular for neutrinos calls for a good control of the theoretical uncertainties of the models used in the analysis . using realistic nuclear spectral functions which include nucleon correlations , we find that the convergence of the derivative expansions to the full results is poor except at very low values of @xmath0 . |
the internal antarctic plateau is , at present , a site of potential great interest for astronomical applications .
the extreme low temperatures , the dryness , the typical high altitude of the internal antarctic plateau ( more than 2500 m ) , joint to the fact that the optical turbulence seems to be concentrated in a thin surface layer whose thickness is of the order of a few tens of meters do of this site a place in which , potentially , we could achieve astronomical observations otherwise possible only by space . in spite of the exciting first results @xcite the effective gain that astronomers might achieve from ground - based astronomical observation from this location still suffers from serious uncertainties and doubts that have been pointed out in previous work @xcite
. a better estimate of the properties of the optical turbulence above the internal antarctic plateau can be achieved with both dedicated measurements done simultaneously with different instruments and simulations provided by atmospheric models .
simulations offer the advantage to provide volumetric maps of the optical turbulence ( @xmath0 ) extended on the whole internal plateau and , ideally , to retrieve comparative estimates in a relative short time and homogeneous way on different places of the plateau . in a previous paper @xcite
our group performed a detailed analysis of the meteorological parameters from which the optical turbulence depends on , provided by the general circulation model ( gcm ) of the european center for medium - range weather forecasts ( ecmwf ) . in that work we quantified the accuracy of the ecmwf estimates of all the major meteorological parameters and , at the same time , we pointed out which are the limitations of the gcms . in contexts in which the gcms fail , mesoscale models can supply more accurate information because they are conceived to reconstruct phenomena that develop at a too small spatial and temporal scale to be described by a gcm . in spite of the fact that mesoscale models can attain higher resolution than the gcms ,
some parameters , such as the optical turbulence , are not explicitly resolved but are parameterized , i.e. the fluctuations of the microscopic physical quantities are expressed as a function of the corresponding macroscopic quantities averaged on a larger spatial scale ( cell of the model ) . for classical meteorological parameters the use of a mesoscale model should be useless if gcms such as the one of the ecmwf could provide estimate with equivalent level of accuracy .
for this reason the hagelin et al . paper ( 2008 ) has been a first step towards the exploitation of the mesoscale meso - nh model above the internal antarctic plateau . in that study
we retrieved an exhaustive characterization of all the meteorological parameters from the ecmwf analyses ( wind field , potential temperature , absolute temperature ... ) and , at the same time , we defined the ecmwf s analyses limitations : we concluded that in the first 10 - 20 m , the ecmwf analyses show a discrepancy with respect to measurements of the order of 2 - 3 m.s@xmath3 for the wind speed and of 4 - 5 k for the temperature .
[ fig : oro ] the meso - nh model has been proven to be reliable in reproducing 3d maps of optical turbulence @xcite and it has been statistically validated @xcite above mid - latitude astronomical sites .
preliminary tests concerning the optimization of the model configuration and sensitivity to the horizontal and the vertical resolution have already been conducted by our team @xcite for the internal antarctic plateau . in this paper
we intend to quantify the performances of the model above this peculiar environment .
more precisely , our goals are : * to compare the performances of the mesoscale meso - nh model and the ecmwf gcm in reconstructing wind speed and absolute temperature ( main meteorological parameters from which the optical turbulence depends on ) with respect to the measurements .
this analysis will quantify the performances of the meso - nh model with respect to the gcm from the ecmwf . * to perform simulations of the optical turbulence above dome c ( 75@xmath40604``s , 123@xmath42048''e ) employing different model configurations and compare the typical simulated thickness of the surface layers well as the seeing in the free atmosphere with the one measured by @xcite ( hereafter tr2008 ) . in this way we aim to establish which configuration
is necessary to reconstruct correctly the @xmath0 . in summary
we aim to validate the meso - nh model on the antarctic site . the two issues : ( 1 ) the surface layer thickness @xmath1 and ( 2 ) the typical seeing in the free atmosphere are certainly the two main features that might get this place on the earth extremely appealing for astronomers and it might be extremely useful to have an independent confirmation from models of the typical values measured on the site . this study is focused on the winter season . in section 2
we present the meso - nh model and the different configurations that were used to perform numerical weather simulations above the internal antarctic plateau .
section 3 is devoted to a statistical comparison of standard meteorological parameters ( wind speed and temperature ) deduced from meso - nh simulations , ecmwf analyses and radiosoundings . in section 4
we present the results of the computation with meso - nh of the surface layer thickness for 15 nights in winter time and a comparison with the observed surface layer thickness from tr2008 .
finally conclusions are drawn in section 5 .
meso - nh @xcite is the non - hydrostatic mesoscale research model developed jointly by mto - france and laboratoire darologie .
it can simulate the temporal evolution of the three - dimensional atmospheric flow over any part the globe .
the prognostic variables forecasted by this model are the three cartesian components of the wind @xmath5 , @xmath6 , @xmath7 , the dry potential temperature @xmath8 , the pressure @xmath9 , the turbulent kinetic energy @xmath10 .
the system of equation is based upon an anelastic formulation allowing for an effective filtering of acoustic waves . a gal - chen and sommerville@xcite
coordinate on the vertical and a c - grid in the formulation of arakawa and messinger@xcite for the spatial digitalization is used .
the temporal scheme is an explicit three - time - level leap - frog scheme with a time filter @xcite .
the turbulent scheme is a one - dimensional 1.5 closure scheme @xcite with the bougeault and lacarrre@xcite mixing length .
the surface exchanges are computed in an externalized surface scheme ( surfex ) including different physical packages , among which isba @xcite for vegetation .
masciadri et al .
( 1999a , b ) implemented the optical turbulence package to be able to forecast also the optical turbulence ( @xmath0 3d maps ) and all the astroclimatic parameters deduced from the @xmath0 .
we will refer to the astro - meso - nh code to indicate this package . to compare simulations with measurements
the integrated astroclimatic parameters are calculated integrating the @xmath0 with respect to the zenith in the astro - meso - nh code .
the parameterization of the optical turbulence and the reliability of the astro - meso - nh model have been proved in successive studies in which simulations have been compared to measurements provided by different instruments @xcite .
this has been achieved thank to a dedicated calibration procedure that has been proposed and validated by the same authors @xcite .
the atmospheric meso - nh model is conceived for research development and for this reason is in constant evolution .
one of the major advantages of meso - nh that was not available at the time of the masciadri s studies is that it allows now for the use of the interactive grid - nesting technique @xcite .
this technique consists in using different imbricated domains with increasing horizontal resolutions with mesh - sizes that can reach 10 meters .
we use in this study the astro - meso - nh package , implemented in the most recent version of the atmospheric meso - nh model . to facilitate the put in the context of this work , the differences that have been implemented in the model configuration with respect to the previous masciadri s studies are listed here : * a higher vertical resolution near the ground has been selected .
we still work with a logarithmic stretching near the ground up to 3.5 km but we start with a first grid point of 2 m ( instead of 50 m ) with 12 points in the first hundred meters . this configuration has been allowed thanks to the extremely smooth orography of this region of the earth .
it is obviously preferable because it permits to better quantify the turbulence contribution that typically develops in the thin vertical slabs in the first hundred of meters above the internal antarctic plateau .
above 3.5 km the vertical resolution is constant and equal to @xmath11=600 m as well as in masciadri s previous work .
the maximum altitude is 22 kilometers . *
the grid - nesting ( see table [ tab1 ] ) is implemented with 3 imbricated domains allowing a maximum horizontal resolution of 1 km in a region around the concordia station ( 80 km @xmath12 80 km ) . *
the simulations are forced at synoptic times ( every 6 hours ) by analyses from the ecmwf .
this permits to perform a real forecast of the optical turbulence . to avoid misunderstandings , we highlight indeed that , as it has been extensively explained in previous studies ( masciadri et al .
2004 , masciadri & egner 2006 ) , the meso - nh model has been used so far for simulations of the optical turbulence in a configuration permitting a quantification of the mean optical turbulence during a night and not a forecast of the optical turbulence .
we perform therefore a step ahead with respect to results obtained so far with the astro - meso - nh code . in spite of the fact that the orographic morphology is almost flat above antarctica
, it is known that even a weak slope can be an important factor to induce a change in the wind speed at the surface in these regions .
the physics of the optical turbulence strongly depend on a delicate balance between the wind speed and temperature gradients . in order to study the sensitivity of the model to the horizontal resolution and to identify
which configuration provides more reliable estimates we performed two sets of simulations with different model configurations . in the first configuration ( that we will call _ monomodel _ ) we used an horizontal resolution @xmath13=100 km covering the whole antarctic continent ( figure [ fig : oro]a , b and table [ tab1 ] ) .
we selected this configuration because it permits us to discuss , where it is possible , our results with respect to those obtained by @xcite ( hereafter sg2006 ) with the regional atmospheric model mar above the antarctic plateau . in that case , indeed , the authors used this extremely low horizontal resolution that has the advantage to be cheap from a computational point of view .
this model configuration permits fast simulations but it is certainly necessary to verify that it is high enough to correctly resolve the most important features of the optical turbulence near the ground and in the high part of the atmosphere . .meso - nh model configuration . in the second column
the horizontal resolution @xmath14x , in the third column the number of grid points and in the fourth column the horizontal surface covered by the model domain . [ cols="^,^,^,^",options="header " , ] [ see_fa ] again we observed that results are weakly dependent on the temporal range on which the means values are calculated and for this reason we report just the 12 - 16 utc case .
[ fig : corr_see ] shows the correlation between the observed and simulated values for the seeing in the free atmosphere and in the whole atmosphere .
the median of the observed seeing in the free atmosphere for the 15 nights is @xmath15@xmath16 [email protected] arcsec ; the median seeing in the free atmosphere simulated by meso - nh with the high horizontal resolution is @xmath18@xmath16 [email protected] arcsec and with the low horizontal resolution is @xmath18@xmath16 [email protected] arcsec .
both median simulated values ( with low and high resolution ) match the median value obtained with observation within the statistical error even if the high resolution is much better correlated ( relative error of 16@xmath19 , @xmath14@[email protected] " ) .
if we look at the total seeing developed on the whole atmosphere it is well visible ( table [ see_fa ] and fig.[fig : corr_see ] ) that the model overestimates the measurements with both resolutions .
we have a simulated median @xmath21@xmath16 [email protected] arcsec and @xmath22@xmath16 [email protected] arcsec versus an observed median @xmath23@xmath16 [email protected] arcsec .
even if we take into account the more accurate estimates ( high resolution ) we obtain a dispersion simulations / observations @xmath14@xmath24 @xmath25 0.7 arcsec . the excess of optical turbulence reconstructed by the meso - nh model is clearly concentrated in the surface layer .
we can not exclude an underestimate from measurements but there are , at present time , no major elements that lead to this assumption .
we are working , on the contrary , on a paper to explain this discrepancy and overcome this limitation . considering that we proved that the meteorological parameters are well reconstructed by the meso - nh model near the surface ( section 3 ) and that the surface numerical scheme ( interaction soil biosphere atmosphere isba ) responsible of the control of the budget of the turbulent ground / air fluxes has been recently optimized for antarctic applications ( @xcite ) in the context of our project
, we concentrated our attention on the dynamical and optical numerical turbulence schemes . in terms of comparison with the sg2006 study we note that the latter study indicates a typical underestimated total seeing of 1.16 arcsec with respect to the observed one ( 1.6 arcsec ) .
the discrepancy is smaller from a quantitative point of view ( @xmath14@[email protected] arcsec ) with respect to what we find and it is in the opposite direction .
the questionable issue in the sg2006 study is that the turbulence kinetic energy provided by sg2006 in the first levels of the mar model is often of the order of 10@xmath27 m@xmath28s@xmath29 ( @xcite ) .
this values is extremely low and it basically indicates no turbulent kinetic energy on the first level of the model and such a condition is contrary to what observed with measurements .
this is consistent with the fact that the mar model underestimates the seeing in the surface layer .
we conclude that the meso - nh model , in the present configuration , reconstructs with good statistical reliability the @xmath1 and the seeing in the free atmosphere while shows a tendency in overestimating the strength of the seeing in the surface layer .
the interesting result of this paper is therefore the fact that the most important features for astronomical interest ( the surface layer thickness and the typical seeing in the free atmosphere ) observed with measurements are confirmed with mesoscale atmospherical model .
we note that the this is the first confirmation made by a mesoscale model of the typical seeing in the free atmosphere . besides it is worth to highlight that these are the first @xmath2 simulations ever done above the internal antarctic plateau and extended all along the whole atmosphere .
figure [ see_free ] shows the temporal evolution of the @xmath2 profile in the free atmosphere ( more precisely in the ( 1,12 ) km vertical slab ) related to three selected nights in the sample of the 15 simulated nights . in all of the three nights
it is well visible that , even at such high altitudes , the model is active and the vertical distribution of the optical turbulence changes in time with a not negligible dynamic from a quantitative point of view .
the @xmath2 values extend , indeed , on the logarithmic scale ( -18,-16.5 ) . in all the 3 cases it appears clearly that the high - horizontal resolution provides a better temporal variability as expected .
these results are therefore very promising in terms of predictions of the @xmath2 3d maps on long time scales .
[ see_free ]
in this paper we study the performances of the meso - nh mesoscale meteorological model in reconstructing meteorological parameters ( wind speed and temperatures ) as well as the optical turbulence above concordia station in the dome c area , a site in the internal antarctic plateau .
this is , at our knowledge , the first study concerning the optical turbulence reconstructed with an atmospherical mesoscale model above antarctica on the whole atmosphere .
this study is concentrated on the winter season i.e. the most interesting for stellar astronomical applications .
the validation of the model for the meteorological parameters has been done comparing measurements ( radiosoundings ) and simulations on a sample of 47 nights .
the validation of the model for the optical turbulence has been done comparing simulations with measurements on a sample of 15 nights .
two different model configurations were tested : monomodel simulations using a low horizontal resolution ( @xmath13=100 km ) and grid - nesting simulations with high horizontal resolution ( @xmath13=1 km for the innermost domain ) .
the low resolution model permitted us to discuss the results obtained previously in the literature .
the observations used for the validation are , for the meteorological parameters , the analyses from the ecmwf global circulation model and radiosoundings ( 47 nights ) and , for the optical turbulence , the @xmath2 and seeing values ( 15 nights ) measured in situ ( trinquet et al . 2008 ) .
the main conclusions of this study are : * we showed that near the surface , meso - nh retrieved better wind speed vertical gradient ( wind shear ) than the ecmwf analyses from a qualitative as well as quantitative point of view , thanks to the use of a highest vertical resolution .
we expect therefore a better reconstruction of the katabatic winds typical of these regions by the meso - nh model than the gcm models .
also meso - nh better reconstructs the thermal stability near the surface than the gcms .
the analysis of the first vertical grid point permits us to conclude that the meso - nh model surface temperature is closest to the observations @xmath14t@[email protected] k than the ecmwf general circulation model ( @xmath14t@[email protected] k ) which is too warm .
the improvement for the estimate of the wind speed is even more evident ( @xmath14v@xmath30@xmath16 0.04 m.s@xmath3 versus @xmath14v@xmath31@xmath16 2.49 m.s@xmath3 ) .
+ * for what concerns the parameters concerning the optical turbulence , again the results are resolution dependent .
the simulations with low resolution provides a too thick surface layer ( almost double of the observed one ) while those with high resolution provide a mean h@[email protected]@xmath177.6 m versus an equivalent observed h@[email protected]@xmath175.1 m. taking into account the statistical error we observe that the high horizontal mode provides a surface layer thickness that is statistically just 6 m higher than the observed one but within the dispersion @xmath34 of the observations . +
* the integral of the @xmath2 above the h@xmath35 i.e. the seeing in the free atmosphere @xmath15@xmath16 [email protected] arcsec is reconstructed with an excellent level of reliability ( @xmath14@[email protected] arcsec ) by the model used with the high resolution configuration @xmath18@xmath16 [email protected] arcsec .
the low resolution provides a worse estimate even if within the @xmath34 of the observations . +
* the model still shows a tendency in overestimating the turbulence in the surface layer .
for an observed @xmath23@xmath16 1.6 arcsec we have a simulated @xmath36@xmath16 2.29 arcsec with the model in high horizontal resolution mode .
this is the subject of an on - going study conceived to answer to this open question . + * the results concerning the computation of the mean thickness of the surface layer as well as the seeing in different vertical slabs are not very dependent of the time interval used to average it .
this widely simplifies the analysis of simulations .
+ * estimates obtained with the grid - nested simulations are closer to the observations than those obtained with monomodel simulations .
this study highlighted the necessity of the use of high horizontal resolution to reconstruct a good meteorological field as well as the parameters characterizing the optical turbulence in antarctica , even if the orography is almost flat over the internal antarctic plateau .
the employment of the low resolution ( 100 km ) alone can hardly be used to identify _ the best site on the antarctic plateau_. however , it can be used to identify rapidly , on the whole antarctic plateau , the most interesting regions in which to focus , successively , simulations at high horizontal resolutions on smaller surfaces domains . with `` the most interesting regions '' we mean those with the lowest surface layer thickness for example .
+ * the meso - nh model is able to reconstruct a mean @xmath2 profile well fitting the vertical optical turbulence distribution measured in the first 20 km from the ground .
the model also shows a not negligible temporal variability in the whole 20 km from the ground in a very small dynamic range .
the latter is to be considered a very interesting feature because it is known that this is a region of the atmosphere in which in general the mesoscale models are less sensible than near the ground .
it is therefore a further indication that the meso - nh model is well placed to forecast the turbulence evolution at these time scales .
+ once the tendency in overestimating the strength of the turbulence in the surface layer will be solved ( forthcoming paper ) we plan to run the meso - nh model in other regions of the internal antarctic plateau to identify the best locations for astronomical observations i.e. the places with the best turbulence characteristics from an astronomical point of view .
this study has been funded by the marie curie excellence grant ( forot ) - mext - ct-2005 - 023878 .
ecmwf analyses are extracted from the catalog mars , @xmath37 .
radiosoundings come from the progetto di ricerca `` osservatorio meteo climatologico '' of the programma nazionale di ricerche in antartide ( pnra ) , _
http://www.climantartide.it_. agabi , a. , aristidi , e. , azaouit , m. , fossat , e. , martin , f. , sadibekova , t. , vernin , j. , ziad , a. , 2006 , pasp , 118 , 344 .
arakawa , a. , messinger , f. , 1976 , garp tech .
, 17 , wmo / icsu , geneva , switzerland aristidi , e. , agabi , k. , azouit , m. , fossat , e. , vernin , j. , travouillon , t. , lawrence , j. s. , meyer , c. , storey , j. w. v. , halter , b. , roth , w. l. , walden , v. , 2005 , a&a , 430 , 739 asselin , r. , 1972 , mon . weather
, 100 , 487 bougeault , p. , lacarrre , p. , 1989
, mon . weather .
, 117 , 1872 cuxart , j. , bougeault , p. , redelsperger , j .- l . , 2000 , q. j. r. meteorol .
soc . , 126 , 1 gal - chen , t. , sommerville , c. j. , 1975 , j. comput . phys . , 17 , 209 galle , h. , arena site testing workshop , june 2007 , _ http://www.concordiastation.org/domec/pdf/20060426 /programma.pdf _ geissler , k. , masciadri , e. , 2006 , pasp , 118 , 1048 hagelin , s. , masciadri , e. , lascaux , f. , stoesz , j. , 2008 , mnras , 387 , 1499 lafore , j .-
p . , stein , j. , asencio , n. , bougeault , p. , ducrocq , v. , duron , j. , fischer , c. , hereil , p. , mascart , p. , masson , v. , pinty , j .-
p . , redelsperger , j .- l . , richard , e. , vil - guerau de arellano , j. , 1998 , annales geophysicae , 16 , 90 lascaux , f. , masciadri , e. , stoesz j. , hagelin , s. , 20 - 22 march 2007 , symposium on seeing , kona - hawaii , proceeding available at _ http://weather.hawaii.edu/symposium/publications _ lawrence , j. , ashley , m. , tokovinin , a. , travouillon , t. , 2004 , nature , 431 , 278 le moigne , p. , noilhan , j. , masciadri , e. , lascaux , f. , pietroni , i. , 2008 , optical turbulence - astronomy meets meteorology , 15 - 18 september 2008 , alghero , sardegna , italy , @xmath38 , in press masciadri , e. , vernin , j. , bougeault , p. , 1999a
, a&ass , 137 , 185 masciadri , e. , vernin , j. , bougeault , p. , 1999b
, a&ass , 137 , 203 masciadri , e. , vernin , j. and bougeault , p. 2001
, a&a , 365 , 699 masciadri , e. , jabouille , p. , 2001
, a&a , 376 , 727 masciadri , e. , avila , r. , sanchez , l. j. , 2004 , rmxaa , 40 , 3 masciadri , e. , egner , s. , 2006 , pasp , 118 , 849 , 1604 noilhan , j. , planton , s. , 1999 , mon .
, 117 , 536 stein , j. , richard , e. , lafore , j .-
p . , pinty , j .-
, asencio , n. , cosma , s. , 2000 , meteorol .
phys . , 72 , 203 stoesz , j. , masciadri , e. , lascaux , f. , hagelin , s. , spie , 7012 , 70124c , 2008 swain , m. , galle , h. , 2006 , pasp , 118 , 1190 trinquet , h. , agabi , k. , vernin , j. , azouit , m. , aristidi .
e. , fossat , e. , 2008 , pasp , 120,864 , 203 | these last years ground - based astronomy has been looking towards antarctica , especially its summits and the internal continental plateau where the optical turbulence appears to be confined in a shallow layer close to the icy surface .
preliminary measurements have so far indicated pretty good value for the seeing above 30 - 35 m : 0.36 `` @xcite , 0.27 '' @xcite and 0.3 " @xcite at dome c. site testing campaigns are however extremely expensive , instruments provide only local measurements and atmospheric modeling might represent a step ahead towards the search and selection of astronomical sites thanks to the possibility to reconstruct 3d @xmath0 maps over a surface of several kilometers .
the antarctic plateau represents therefore an important benchmark test to evaluate the possibility to discriminate sites on the same plateau .
our group @xcite has proven that the analyses from the ecmwf global model do not describe with the required accuracy the antarctic boundary and surface layers in the plateau . a better description could be obtained with a mesoscale meteorological model .
the mesoscale model meso - nh has proven to be reliable in reproducing 3d maps of optical turbulence above mid - latitude astronomical sites @xcite . in this paper
we study the ability of the meso - nh model in reconstructing the meteorological parameters as well as the optical turbulence above dome c with different model configurations ( monomodel and grid - nesting ) .
we concentrate our attention on the model abilities in reproducing the optical turbulence surface layer thickness ( @xmath1 ) and the integral of the @xmath2 in the free atmosphere and in the surface layer .
it is worth to highlight that these are the first estimates ever done so far with a mesoscale model of the optical turbulence above the internal antarctic plateau .
[ firstpage ] site testing atmospheric effects turbulence |
recent observations of pulsars revealed that some of them emit thermal - like radiation in the soft x - ray range ( gelman 1995 ) .
these observations stimulated further investigations of neutron star ( ns ) photospheres responsible for the properties of the emitted radiation .
one of important and hitherto untouched problems associated with modeling of the ns photospheres is the problem of convective energy transport which can affect the temperature distribution and the emergent spectra . due to huge gravitational accelerations , @xmath2 cm s@xmath3 ,
the ns photospheres are much denser , @xmath4 g @xmath5 , than those of usual stars .
the increased densities shift ionization equilibrium : the nonionized fraction grows with @xmath6 at moderate densities , @xmath7 g @xmath5 , and sharply decreases at @xmath8 g @xmath5 due to pressure ionization . as a result , zones of increased opacity ( increased radiative gradient @xmath9 ) and reduced adiabatic gradient @xmath10 develop in the ns photospheres at temperatures much higher than in photospheres of usual stars , which may cause convective instability at depths where the emergent spectrum is formed .
so far the ns photospheres have been considered either for stars with very strong surface magnetic fields @xmath11 g ( see , for example , pavlov @xmath12 . 1994 ) where existing convection theories are not applicable , or for ` nonmagnetic ' ( low - field ) photospheres , @xmath13 g , with high surface effective temperatures @xmath14 k ( romani 1987 ) , where the convection can hardly be expected . since thermal ns radiation can be detected for @xmath15 k in the soft x - ray range , and for even lower temeratures , @xmath16 k , in the uv / optical range ( pavlov @xmath12 . 1995 ) , the study of the convective ns photospheres is important for the proper interpretation of these observations .
the convective flow in stellar photospheres is turbulent and imposes many complicated problems .
a common practice , which we follow here , is to use the phenomenological mixing - length theory with the traditional schwarzschild criterion for convective instability , @xmath17 .
this theory has been widely implemented and testified .
details of our numerical calculations will be described elsewhere .
generally , we employ the complete linearization method for computing the photosphere models and include the convective energy transfer as described by mihalas ( 1978 ) .
we calculated the model of nonmagnetic photospheres for different chemical compositions . here
we present examples for pure hydrogen , helium and iron compositions at the gravitational acceleration @xmath18 cm s@xmath3 which corresponds to standard ns mass @xmath19 and radius @xmath20 km .
the raditive opacities and equation of state for the iron composition were taken from the opal library ( iglesias @xmath12 .
the results can be directly applicable to very old nss with low magnetic fields ( e. g. , millisecond pulsars ) .
our results show that , similar to the case of usual stellar photospheres , the convective energy transfer begins to play a role at lower surface temperatures when atoms are not fully ionized and the radiative opacities are strongly increased by contribution of the bound - free and bound - bound transitions .
the increased opacities result in high values of @xmath9 the depth dependence of the temperature becomes steeper in order to transfer the energy flux throughout the photosphere . on the other hand , @xmath10 in the dense partially ionized layers
can be much smaller than its limiting value 0.4 for an ideal fully ionized or fully nonionized gas ( see , for example , cox and guili 1968 ) . as a result ,
superficial convection zones in ns photospheres arise and , as in usual stars , they are associated with layers of partial ionization .
our computations show that @xmath10 can drop down to @xmath21 when the nonionized fractions are @xmath2270% .
the actual temperature gradient @xmath23 in the convection zones satisfies the following relation : @xmath24 .
1 shows the temperature distributions in photospheres with and without allowance for convection .
one can see that convection leads to more gradual profiles , in accordance with the above relation .
both in very surface layers , where @xmath9 is too low , and in deep layers , where @xmath10 is close to its maximum value , convection is absent , and the temperature profiles remain the same .
the convective transfer affects not only the structure of photosphere but also the spectra of the ns thermal radiation ( fig .
2 ) because the temperature profiles are changed in the layers where the radiation escapes from . in particular , convection substantially ( up to two orders of magnitide ) lowers the flux from h and he photospheres at photon energies above the main photoionization edges , so that the high - energy spectral tails become softer .
the spectra remain the same at low and very high energies since both shallow and very deep layers are not affected by convection . in the case of fe composition , the convective zone lies so deep that only a high - energy tail ( @xmath25 kev ) is affected .
the effect of convection on the spectra disappears with increasing effective temperature ( _ e .
g. _ , at @xmath26 k for h , at @xmath27 k for he and at @xmath28 k for fe photospheres ) .
the presented results correspond to the convective efficiency @xmath29 .
acceptable values of this parameter can vary from 0.3 to 2.5 .
bergeron @xmath12 .
( 1992 ) showed that higher efficiency enhances convection and smoothes the spectra of white dwarfs .
the same effect can be expected in the ns photospheres and , consequently , the convection there may develop at higher effective temperatures than in the models presented .
convection in ns photospheres is important because it can mix the material in convective zones , bringing heavier elements from bottom to surface layers which would otherwise contain only light elements due to gravitational stratification .
our calculations show that this may happen in cold nss with @xmath30 k , whereas radiating layers of hotter nss can be expected to consist mainly of hydrogen and helium . since the presence of convection softens the high - energy tails of the spectra , this effect should be taken into account for the proper interpretation of euv / soft - x - ray observations of thermal radiation from nss .
bergeron p. @xmath12 . , 1992 , apj , 387 , 288 cox j.p . and giuli r.t . , 1968 , principles of stellar structure , new york : cordon and breach iglesias c.a
. , 1992 , apj , 397 , 717 mihalas d. , 1978 , stellar atmospheres , san francisco : freeman gelman h. , 1995 , in the lives of the neutron stars , eds .
a. alpar , . kizilolu and j. van paradijs , kluwer academic publisher , p.101 pavlov g.g .
@xmath12 . , 1994 , a&a 289 , 837 pavlov g.g .
@xmath12 . , 1995 ,
apj , submitted romani r. 1987 , apj , 313 , 718 | we present first results of modeling convective photospheres of neutron stars .
we show that in photospheres composed of the light elements convection arises only at relatively low effective temperatures ( @xmath0 k ) , whereas in the case of iron compositon it arises at @xmath1 k. convection changes the depth dependence of the photosphere temperature and the shapes of the emergent spectra .
thus , it should be taken into account for the proper interpretation of euv / soft - x - ray observations of the thermal radiation from neutron stars . |
one from the most important physical processes is electrodiffusion .
it describes both diffusional motion of mass and charge flow due to applied electric field .
the electric potential distribution is govern by the poisson equation and total transport of particles is given in terms of the continuity equation @xcite .
the significance of this equation is broadly described in existing physical , chemical and biological literature @xcite and lots of scientific articles , particularly , those which concern properties of nano and micro transport .
+ mathematically , equations of electrodiffusion constitute a set of coupled nonlinear equations where the laplace operator @xcite appears together with the first order partial time derivative .
the laplace operator is the basic operator met in many physical situations .
thus the first step to deal with the electrodiffusional problem is to approximate the solution of the laplace equation with help of the finite elements method .
practically , it means that an appropriate mesh should be designed for a prescribed 3d domain .
the mesh must fit well to the physical conditions like e. g. symmetry of the problem .
therefore , different mesh shapes could be desired ( spherical , cylindrical , conical or cubic ) up to problem . after having accurate basic spatial solutions on appropriate meshes ,
the problem should be extended to the time
dependent case of the diffusion equation by finding discrete approximation in time .
it could be done by means of truncated taylor series or other single step procedures like the crank nicolson scheme @xcite or the gurtin s approach to finite element approximation in terms of variational principle @xcite . from now , further extension of above - presented computations involving non - linear terms could be easily implemented and numerically solved using the newton s method @xcite .
the equation of electrodiffusion @xcite has the form @xmath1 where @xmath2 is a number of @xmath3-th ions , @xmath4 - electric potential , @xmath5 - the diffusion coefficient of @xmath3-th particles , @xmath6 - boltzmann constant , @xmath7 - temperature , @xmath8 - valence of the @xmath3-th kind of ions , @xmath9 - electric charge . to find the electric potential @xmath4 the poisson equation
must be solved @xmath10 where @xmath11 , @xmath12 .
thus equations ( [ electrodiffusion ] ) and ( [ poisson ] ) both constitute the system of coupled equations .
next , they can be solved numerically using the finite element method @xcite where the problem is represented as @xmath13d\omega = 0 \\ \displaystyle \int_\gamma \mathbf{\tilde{v}}^t \mathbf{b(u ) } d\gamma \equiv \int_\gamma [ \tilde{v}_1 b_1(\mathbf{u } ) + \tilde{v}_2 b_2(\mathbf{u } ) + \tilde{v}_3 b_3(\mathbf{u } ) ] d\gamma = 0 \end{array}\ ] ] where @xmath14\ ] ] and @xmath15 and @xmath16 are sets of arbitrary functions equal in number to the number of equations ( or components of @xmath17 ) involved .
@xmath18 and @xmath19 are given by the following formulas , respectively @xmath20 \nabla \mathbf{u}^t\left [ \begin{array}{c } % \vspace{-10pt } 0\\ % \vspace{-10pt } 0\\ 1 \end{array } \right ] - \tilde{d_+ } \nabla\mathbf{u}^t \left [ \begin{array}{c } % \vspace{-10pt } 1\\ % \vspace{-10pt } 0\\ 0 \end{array } \right ] \right ) + [ 1 , ~0 , ~0 ] \displaystyle\frac{\partial}{\partial t}\mathbf{u}\\ \vspace{10pt } a_2(\mathbf{u } ) = \nabla^t \left(-k_-\mathbf{u}^t \left [ \begin{array}{c } % \vspace{-10pt } 0 \\ % \vspace{-10pt } 1 \\ 0 \end{array } \right ]
\nabla \mathbf{u}^t\left [ \begin{array}{c } % \vspace{-10pt } 0\\ % \vspace{-10pt } 0\\ 1 \end{array } \right ] - \tilde{d}_- \nabla\mathbf{u}^t \left [ \begin{array}{c } % \vspace{-10pt } 0\\ % \vspace{-10pt } 1\\ 0 \end{array } \right ] \right ) + [ 0,~1,~0 ] \displaystyle\frac{\partial}{\partial t}\mathbf{u}\\ a_3(\mathbf{u } ) = \epsilon_0\epsilon\nabla^t\nabla\mathbf{u}^t \left [ \begin{array}{c } % \vspace{-10pt } 0\\ % \vspace{-10pt } 0\\ 1 \end{array } \right ] + [ 1 , ~-1 , ~0]ze\mathbf{u}. \end{array}\ ] ] where @xmath21 , @xmath22 .
an expression @xmath23 gives the boundary conditions on @xmath24 , however , we choose a _ forced _ type of boundary conditions on @xmath24 i. e. @xmath25 let us substitute @xmath26 = \mathbf{n}\mathbf{\tilde{u } } $ ] and put @xmath27 where @xmath28 . in that way , we end up with the galerkin formulation of the problem . in turn
, we can approximate the nodal electric potential @xmath29 and number of particles @xmath30 at a time @xmath31 by @xmath32 and taking advantage from an evaluation of @xmath33 $ ] in the taylor series we obtain @xmath34 where @xmath35 takes values from @xmath36 $ ] and @xmath37 denotes time step . after incorporating it into a general form of time
dependent equations @xmath38 @xmath39 where @xmath40 represents these parts of @xmath38 with a space dependent operator we get time approximation for a given node @xmath41 when @xmath42 an approximate solution to the semi - discrete equations at each time @xmath31 is given by the euler , , backward scheme @xmath43 otherwise , the expression for @xmath44 is as follows @xmath45 after integration @xmath46 by parts , one obtains a mixed set of linear and nonlinear equations @xmath47 and a corresponding set of boundary terms for @xmath4 and @xmath48 where @xmath22 @xmath49 where @xmath50 denotes derivative normal to @xmath24 . presented
above spatially temporal discretization has been done for the case with @xmath51 at each node . if the _ forced _
boundary conditions ( see eq .
( [ boun ] ) ) are imposed on @xmath52 and @xmath53 , respectively , then all terms in eq .
( [ gamma ] ) can be neglected by restricting the choice of @xmath54 functions to those which equal 0 on @xmath24 .
let s denote integrals from eq .
( [ omega ] ) as @xmath55 where @xmath56 with @xmath57 denotes sum over elements . in 3d space
tetrahedral elements seem to be a natural choice of finite volume elements
. then indices @xmath58 take four values each ( an element has four nodes ) from the set of @xmath59 values . for tetrahedral linear elements shape functions @xmath60
can be assumed as equal area coordinates @xmath61 given by the formula @xmath62 where @xmath63 is a volume of tetrahedron .
the following integration formula can be useful @xmath64 in order to calculate integrals @xmath65 , @xmath66 and @xmath67 . shape functions for linear elements are @xmath68 for @xmath69 .
this gives @xmath70 [ b_c , c_c , d_c]^t = \\
\vspace{10pt } \displaystyle \frac{1}{4!6v^e } \left ( b^bb_c + c^bc_c + d^bd_c \right)\\ \vspace{10pt } \displaystyle \tilde{k}^{b , e}_{a } = \int_{\omega^e}\left ( \nabla l^b \right)^t\nabla l_a d\omega^e = \frac{1}{36v^e } \left ( b^bb_a + c^bc_a + d^bd_a \right)\\ \vspace{10pt } \displaystyle k^{b , e}_a = \int_{\omega^e } l_a l^b d\omega^e = \frac{6v^e}{5 ! } ~~\rm{when}~~ a\ne b \\
\displaystyle k^{a , e}_a = \int_{\omega^e } l_a l^a d\omega^e = \frac{12v^e}{5!}. \label{integrals } \end{array}\ ] ] finally , we have @xmath71 where @xmath72 with @xmath73 and only these nodes @xmath74 and @xmath75 that participate in the particular element @xmath9 can give a non
zero contribution to the sums of the general type @xmath76 .
+ let s assume that all values of @xmath77 are known at a time @xmath31 . then we can solve the third equation in ( [ pnp_final ] ) obtaining the result @xmath78 where @xmath79 denotes elements of the inverse matrix to @xmath80 .
after substitution the solution for @xmath81 to eq .
( [ pnp_final ] ) we get @xmath82 where @xmath73 .
the above written set of equations is of a nonlinear type .
let s denote all of them as @xmath83 thus to solve it we have to employ the iterative newton s method @xcite .
it means that we have to start from an initial guess of @xmath84 values . and
during next iterations for @xmath85 @xmath86 the solution should be achieved .
@xmath87 denotes the following matrix of partial derivatives : @xmath88\ ] ] where @xmath89 where @xmath90 .
putting @xmath91 in the equation of electrodiffusion we neglect the electrostatic term .
it leads to the following equation describing diffusion in @xmath92 @xcite @xmath93 where @xmath94 denotes a diffusion coefficient .
this kind of equation represents an initial value problem . assuming that the considered domain @xmath95 in @xmath96 is of a cubic type @xmath97\times[y_{min } , y_{max}]\times[z_{min } , z_{max}]$ ]
let us take @xmath98 as a boundary condition .
now we seek a solution of the equation ( [ diffusion ] ) which satisfies this boundary condition and prescribed initial condition at the time @xmath99 .
the solution of the equation is approximated by the triple sum @xmath100 where @xmath101 are unknown coefficients that must be determinated from the initial condition : @xmath102 in the case of the domain being @xmath103\times[0~ \pi]\times[0~ \pi]$ ] and @xmath104 the solution has the form @xmath105 in the case of cylindrical domain defined by @xmath106 $ ] , @xmath107 and @xmath108 $ ] , and with the boundary condition of the form @xmath98 the solution of eq .
( [ diffusion ] ) can be expanded in an absolutely and uniformly convergent series of the form @xmath109 where @xmath110 are the zeros of the bessel functions and @xmath111 are constants that must be found from the initial condition by making use of the _ orthogonality relation _ for the trigonometric and bessel functions .
on the other hand , assuming that the time derivative in the diffusion equation ( [ diffusion ] ) equals 0 and putting @xmath112 we end up with the boundary value problem of the laplace type @xcite @xmath113 let s consider @xmath95 being a cubic domain i. e. @xmath103\times[0~ \pi]\times[0~ \pi]$ ] . and for the @xmath114 function equals 0 everything on @xmath115 apart from @xmath116 one can approximate the exact solution by @xmath117 for @xmath118 where @xmath119 the laplace equation has the solution defined in @xmath96 for @xmath120 @xmath121 where @xmath122 denotes volume of b(0,1 ) in @xmath96 and equals @xmath123 .
below are listed a few technical remarks referring to the mesh generation routine applied to obtain a designed 3d mesh .
an initial mesh is built on the basis of main surface nodes ( _ outer _ nodes ) which define a figure s shape .
the whole figure is considered as divided into perpendicular to @xmath124-axis layers .
thus the _ outer _ nodes are distributed on the edges of layers . in the center of each layer and also in the middle between two layers are located inner nodes .
they are connected with _
outer _ nodes creating in this way the main figure s construction .
initial mesh elements obtained in such a manner are of tetrahedral shape .
the boundary of the figure is defined by set of _ surface _
equations for vertical and horizontal segment lines linking _ outer _ nodes . after each mesh iteration
new nodes are created and labeled as _ outer _ or _
ones according to surface equations .
moreover , the location of each node ( i. e. on which exactly vertical , horizontal line or surface patch the node is lying ) is also stored .
new elements are created by a division of already existing elements . at the beginning of the routine
, the surface of division mainly connects a new node born on the longest element edge with two other nodes belonging to that mesh element and one node from the divided edge .
the procedure constitutes a 3d extension of the 2d mesh generation routine described already in @xcite .
however , during the routine a number of small elements is increasing , and the division of the longest bar is not anymore the optimal way of proceeding .
that is why , before choosing an edge to the division volume of elements common to it is checked .
the edge that will not produce new elements having its volume smaller than an assumed * critical volume * is chosen to be cut .
the optimization is done with help of the metropolis algorithm .
the system energy is calculated as a sum of discrepancies between an element volume @xmath63 and assumed element volume @xmath125 where @xmath126 denotes a prescribed length of the edge @xmath127 thus the smaller a degeneracy from a designed volume distribution the more optimal state .
the metropolis routine starts from a nodal configuration given by described above procedure .
the main point is to reach the optimal global configuration by ascertaining local optimal states .
they arise from such a configuration of @xmath3-th node and its neighboring nodes which gives smaller energy @xmath128 .
this partial energy is calculated from the sum eq .
( [ energy ] ) taken over elements containing the node of interest . to compute new positions for each node ( giving new configuration ) the following expression is put forward @xmath129 where @xmath130 denotes a shifting strength and @xmath131 is the length of @xmath132 edge .
the value of @xmath130 determines the strength of a nodal shift and varies from 0 to 1 .
it is also worthy considering to choose its value as a random number from uniform distribution @xmath133 . for meshes in a cubic domain @xmath134\times[0
, \pi]\times[0 , \pi]$ ] with a ) unique elements created for @xmath135 and b ) meshes optimized with the metropolis algorithm with @xmath136.,title="fig : " ] for meshes in a cubic domain @xmath134\times[0 , \pi]\times[0 , \pi]$ ] with a ) unique elements created for @xmath135 and b ) meshes optimized with the metropolis algorithm with @xmath136.,title="fig : " ] within the metropolis routine the transition probability is calculated by the formula @xmath137 where @xmath6 is a boltzmann constant ( here set as 1 ) , @xmath7 temperature and @xmath138 is a difference between energies of these two states .
if a value of @xmath139 is greater than a random number from @xmath133 new state is accepted .
otherwise the old one is preserved .
+ all above - described local metropolis steps can lead to different global configurations .
therefore , for each division number this distribution of nodes which gives a lower energy of the total system should be kept . to find this ,
again the metropolis rule is employed , but this time changes in the total energy of the whole system are examined . to estimate maximal temperature @xmath140 ( in expression ( [ tran_prob ] ) ) a range of changes in potential energy corresponding to current number of elements must be found . moreover , in each global metropolis step temperature might decrease according to @xmath141 where parameter @xmath142 .
for the same domain as in fig .
[ length ] and for a ) unique elements ; b ) elements with @xmath143 optimized with the metropolis algorithm.,title="fig : " ] for the same domain as in fig . [ length ] and for a ) unique elements ; b ) elements with @xmath143 optimized with the metropolis algorithm.,title="fig : " ] to improve mesh quality the following transformations are applied @xcite . * three elements common to an edge
are transformed to two elements , when one of the elements fails to satisfy the delaunay criterion @xcite . *
four elements common to an edge are transformed to a new configuration of four elements , when one of the elements does not meet the delaunay criterion @xcite .
the new pattern is chosen from two different possibilities @xcite .
additionally , too small boundary elements could be destructed by a projection of its internal node to the center of the outer patch of element that is opposite to it .
such approach is justified in the case of boundary elements , however , in the case of internal ones leads to creation of so called irregular nodes .
the accuracy of fem approximation of the laplace equation on different meshes were examined .
numerical results vs. analytical ones for cubic and spherical domain are presented in fig .
( [ laplace_cube ] ) .
the relative difference between both analytical and numerical solutions has been calculated as @xmath144 the laplace equation has been solved for the cubic domain @xmath134\times[0 , \pi]\times[0 , \pi]$ ] with potential function @xmath145 everywhere on the boundary @xmath24 apart from one its side at @xmath146 where potential @xmath147 and for the spherical domain with the boundary conditions imposed by putting an elementary charge outside the sphere in @xmath148 $ ] .
the exact solutions for both considered cases are evaluated precisely in sec .
fem approximation has been computed for the , , linear order of tetrahedron @xcite .
however , the comparison between both orders of approximation i .e , , linear and , , quadratic for the uniform mesh with @xmath149 has been performed .
the formulas for higher orders of approximation i. e. quadratic and cubic can be found in @xcite .
the results show that mean discrepancy between numerical and analytical solutions calculated according to eq .
( [ difference_eq ] ) for the laplace equation equal @xmath150 in the linear case and @xmath151 in quadratic approximation , respectively .
thus in further studies linear approximation will be used as being sufficiently accurate . in the case of cubic domain a mesh of unique element volume ( non optimized ) has been applied in contrast to the spherical domain where mesh has been used after its enhancement with both the metropolis algorithm and the delaunay routine .
( as described in sec .
4 ) vs. the analytical result together with b ) a distribution of differences between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) vs. the exact solution of the laplace equation for the spherical domain with the boundary values defined by putting an elementary charge outside the sphere in @xmath148 $ ] b ) the volume distribution in the spherical domain optimized with the metropolis algorithm with elements of a prescribed volume @xmath0 equals @xmath152.,title="fig : " ] ( as described in sec .
4 ) vs. the analytical result together with b ) a distribution of differences between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) vs. the exact solution of the laplace equation for the spherical domain with the boundary values defined by putting an elementary charge outside the sphere in @xmath148 $ ] b ) the volume distribution in the spherical domain optimized with the metropolis algorithm with elements of a prescribed volume @xmath0 equals @xmath152.,title="fig : " ] ( as described in sec .
4 ) vs. the analytical result together with b ) a distribution of differences between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) vs. the exact solution of the laplace equation for the spherical domain with the boundary values defined by putting an elementary charge outside the sphere in @xmath148 $ ] b ) the volume distribution in the spherical domain optimized with the metropolis algorithm with elements of a prescribed volume @xmath0 equals @xmath152.,title="fig : " ] ( as described in sec .
4 ) vs. the analytical result together with b ) a distribution of differences between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) vs. the exact solution of the laplace equation for the spherical domain with the boundary values defined by putting an elementary charge outside the sphere in @xmath148 $ ] b ) the volume distribution in the spherical domain optimized with the metropolis algorithm with elements of a prescribed volume @xmath0 equals @xmath152.,title="fig : " ] to test accuracy of discrete approximation in time the equation of diffusion ( see sec .
3 ) has been examined . to find @xmath153 particles distribution at @xmath154 moment in time the taylor expansion ( see eq .
[ discrete_in_time ] ) with @xmath155 has been applied .
the diffusion equation with the following initial condition @xmath156 for the cubic domain and @xmath157 where @xmath158 denotes a radius of the cylindrical domain has been solved .
the boundary value of the @xmath153 is set as 0 . results for both domains : cubic ( with uniform elements see fig .
[ volume]a ) and cylindrical ( see fig . [ regular ] ) , this one tuned to the designed element volume with the metropolis recipe , are shown in fig .
[ diffusion_pic ] . in the case of
cylindrical domain mean values of the ratio @xmath159 calculated at each point of domain for times @xmath160 have the average value @xmath161 . and
at the time @xmath162 with @xmath163 [ units ] vs. the analytical result together with b ) a distribution of differences ( eq . [ difference_eq ] ) between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) of the diffusion equation for the cylindrical domain at @xmath164 and at the following times : @xmath165 ( blue ) , @xmath166 ( black ) and @xmath167 ( red ) with @xmath163 [ units ] ; d ) volume profile of elements within the cylindrical domain where @xmath168 and @xmath169 ; mesh quality were enhanced with help of the metropolis algorithm.,title="fig : " ] and at the time @xmath162 with @xmath163 [ units ] vs. the analytical result together with b ) a distribution of differences ( eq . [ difference_eq ] ) between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) of the diffusion equation for the cylindrical domain at @xmath164 and at the following times : @xmath165 ( blue ) , @xmath166 ( black ) and @xmath167 ( red ) with @xmath163 [ units ] ; d ) volume profile of elements within the cylindrical domain where @xmath168 and @xmath169 ; mesh quality were enhanced with help of the metropolis algorithm.,title="fig : " ] and at the time @xmath162 with @xmath163 [ units ] vs. the analytical result together with b ) a distribution of differences ( eq .
[ difference_eq ] ) between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) of the diffusion equation for the cylindrical domain at @xmath164 and at the following times : @xmath165 ( blue ) , @xmath166 ( black ) and @xmath167 ( red ) with @xmath163 [ units ] ; d ) volume profile of elements within the cylindrical domain where @xmath168 and @xmath169 ; mesh quality were enhanced with help of the metropolis algorithm.,title="fig : " ] and at the time @xmath162 with @xmath163 [ units ] vs. the analytical result together with b ) a distribution of differences ( eq . [ difference_eq ] ) between numerical and analytical solutions obtained for each node in @xmath95 domain ; c ) fem approximation ( linear ) of the diffusion equation for the cylindrical domain at @xmath164 and at the following times : @xmath165 ( blue ) , @xmath166 ( black ) and @xmath167 ( red ) with @xmath163 [ units ] ; d ) volume profile of elements within the cylindrical domain where @xmath168 and @xmath169 ; mesh quality were enhanced with help of the metropolis algorithm.,title="fig : " ] the system of coupled equations describing process of electrodiffusion ( [ electrodiffusion ] ) written in terms of the fem method ( see eq .
( [ pnp_final ] ) ) with the following values of constants : @xmath170 and the time step equals @xmath171 has been numerically solved by using the newton s algorithm .
the boundary values of @xmath172 are set as 1 at @xmath173 , @xmath146 , @xmath174 , @xmath175 , and @xmath176 and equal 2 at @xmath177 .
an initial guess of @xmath178 and @xmath4 distributions has been chosen as 0 everywhere in the domain apart from its boundaries .
the system of equations has been computed up to the final time @xmath179 .
[ electrodiffusion_pic ] presents obtained profiles of cations @xmath180 and the potential @xmath4 at the center of the domain i. e. at @xmath181 .
there is no visible difference between cations @xmath180 and anions @xmath182 distributions so the latter is not shown .
the maximum of @xmath183 is decaying from 0.023 for @xmath184 ( @xmath99 ) to 0.0093 for @xmath185 ( @xmath186 ) .
the maximum of difference between @xmath187 and @xmath188 distributions computed at each node equals @xmath189 . employing physical notion
it means that the system of charged particles is electroneutral .
moreover , the system of particles is tending to its stationary state . and
at the time @xmath179 with the values of parameters @xmath163 [ units ] , @xmath190 , @xmath191 [ units ] a ) the distribution of cations @xmath180 ; b ) the profile of the potential @xmath4 ; computations have been performed on the uniform mesh with the volume of tetrahedron @xmath192 and the element size @xmath193.,title="fig : " ] and at the time @xmath179 with the values of parameters @xmath163 [ units ] , @xmath190 , @xmath191 [ units ] a ) the distribution of cations @xmath180 ; b ) the profile of the potential @xmath4 ; computations have been performed on the uniform mesh with the volume of tetrahedron @xmath192 and the element size @xmath193.,title="fig : " ] additionally , components @xmath194 and @xmath195 of the total flux of @xmath187 particles flowing through the domain @xmath95 have been computed .
they are shown in fig .
[ electrodiffusion_current ] .
presence of a difference in an amount of @xmath180 particles at the both sides of @xmath95 in the @xmath124direction i. e. at @xmath196 and @xmath197 causes a non zero flow along @xmath124 axis whereas a lack of such a difference in the two other directions i. e. @xmath198 and @xmath199 leads to the vanishing flows@xmath200 and @xmath201 in the center of the domain . and at the time @xmath179 .
computations have been done with the following values of parameters : the time step @xmath163 [ units ] , nonlinear term multipliers @xmath190 [ units ] , diffusion coefficients @xmath191 [ units ] .
this figure shows approximated solutions for flux components for cations @xmath187 in the @xmath199 direction a ) and in the @xmath124 direction b ) .
the computations have been performed on the uniform mesh with the volume of tetrahedron @xmath192 and the element size @xmath193.,title="fig : " ] and at the time @xmath179 .
computations have been done with the following values of parameters : the time step @xmath163 [ units ] , nonlinear term multipliers @xmath190 [ units ] , diffusion coefficients @xmath191 [ units ] .
this figure shows approximated solutions for flux components for cations @xmath187 in the @xmath199 direction a ) and in the @xmath124 direction b ) .
the computations have been performed on the uniform mesh with the volume of tetrahedron @xmath192 and the element size @xmath193.,title="fig : " ]
the presented software offers a 3d mesh generation routine as well as its further application to the 3d electrodiffusional problem . + the proposed mesh generator offers a confident way to creature a quite uniform mesh built with elements having desired volume .
mesh elements have been adjusted to assumed sizes by making use of both the metropolis algorithm and the delaunay criterion .
mesh quality depicted in histograms occurs to be fairly satisfactory .
moreover , goodness of obtained meshes together with robustness of their applications to the finite element method have been also tested by solving the 3d laplace problem and the 3d diffusion equation on them .
comparison between these numerical solutions and analytical results shows very good agreement .
+ to find solutions to a nonlinear problem defined by a system of coupled equations describing electrodiffusion the fem approach and the newton method have been jointly applied .
analysis of obtained results confirms usefulness of the presented solver to deal with nonlinear differential problems . c. t. kelley , _ solving nonlinear equations with newton s method in fundamentals of algorithms _ siam , 2003 ; j. brzzka , l. dorobczyski , _ matlab .
rodowisko oblicze naukowo - technicznych .
_ ( pwn , warszawa , 2008 ) | the presented article contains a 3d mesh generation routine optimized with the metropolis algorithm .
the procedure enables to produce meshes of a prescribed volume @xmath0 of elements .
the finite volume meshes are used with the finite element approach .
the fem analysis enables to deal with a set of coupled nonlinear differential equations that describes the electrodiffusional problem .
mesh quality and accuracy of fem solutions are also examined .
high quality of fem type space
dependent approximation and correctness of discrete approximation in time are ensured by finding solutions to the 3d laplace problem and to the 3d diffusion equation , respectively .
their comparison with analytical solutions confirms accuracy of obtained approximations . |
there is a need for broadband and intense light sources in the mid ir as most molecules display a distinctive spectral fingerprint in this spectral region . in particular ,
some spectral sub - bands within this region can not be addressed by ir sources such as quantum cascade lasers ( qcls ) @xcite .
this problem can be tackled by supercontinuum generation through an adequate pumping scheme and choice of transmission medium .
intense and short pump pulses are best , although high cw powers can trigger sc generation as well @xcite .
pulsed co~2~ lasers are readily available and can provide intense nanosecond pulses around 10 m , and at 5 m through second - harmonic generation @xcite .
tabletop cw co~2~ lasers can easily achieve kilowatt levels .
co lasers can also generate high power levels around 5 m @xcite .
this leaves us with the choice of an appropriate parametric medium .
unfortunately , a well - known material such as silica ( optical fibers ) can not be used , as its absorption becomes prohibitively large above 2.4 m. chalcogenide glasses , on the other hand , display very low absorption up to 16 m @xcite .
other appealing characteristics of chalcogenide glasses include a zero - dispersion wavelength typically above 4 - 5 m ( close to the pump wavelength , thus minimizing the phase mismatching of nonlinear interactions ) , an almost linear dispersion profile with a small dispersion slope , and a very high nonlinear index ( up to 1000 times that of silica ) .
as such , mid - ir sc generation in chalcogenide media has gathered considerable attention in recent years .
there have been demonstrations of sc generation in chalcogenide step - index fibers up to 8 m by pumping at 4 m with 320-fs pulses @xcite , and from 1.4 to 13.3 m by pumping at 6.3 m with 100-fs pulses @xcite , including some impressive demonstrations in bulk glasses from 2.5 to 7.5 m , by pumping at 5.3 m with 150-fs pulses of 20 mw peak power @xcite , and from 0.2 to 8 m by pumping at 1.6 m with 180-fs pulses with an astonishing 1.13 gw peak power @xcite . in this paper
we focus on the study of sc generation in the mid ir , and the observation of the accompanying rogue wave phenomenon , by means of low - power pumping of a chalcogenide waveguide in both short - pulse and cw regimes .
furthermore , we show that the low and flat dispersion characteristics of chalcogenide media in the spectral region of interest , conveniently modeled by an expansion of only up to the fourth - order in the propagation constant , allow to capture the relevant pulse dynamics and to attain most of the sc spectral width in both regimes .
we start by solving the generalized nonlinear schrdinger equation ( gnlse ) expressed as @xcite @xmath0 the number of terms @xmath1 on the left hand side is 4 in most of the simulations .
some simulations were performed using @xmath2 in order to show that the inclusion of only four terms is enough to capture the main aspects of the resulting behavior .
@xmath3 ps^2^/km , @xmath4 ps^3^/km , @xmath5 ps^4^/km , and @xmath6 ps^5^/km are the dispersion parameters , and were so chosen to be consistent with the measured dispersion profile of a chalcogenide glass in the anomalous dispersion region @xcite , i.e. , a small ( negative ) and flat dispersion in the spectral region of interest , yielding a zero - dispersion wavelength at @xmath74.7 m. @xmath8 ( w - km)@xmath9 and @xmath10 are the nonlinear terms around @xmath11 , the angular frequency corresponding to 5 m. it is important to point out that the chosen value for @xmath12 represents a lower limit for a chalcogenide glass , consistent with a nonlinear refractive index of @xmath7100 times that of silica taking as a reference the typical effective area of a step - index silica fiber @xcite .
the response function @xmath13 in the convolution integral of ( [ eq : gnlse ] ) is given by @xmath14 , where @xmath15 , the raman response of the medium , is approximated as @xmath16 @xmath17 ps , @xmath18 ps and @xmath19 were chosen to match the measured raman gain spectrum , with a gain peak shift of 10 thz , in a chalcogenide glass @xcite .
we consider a pump at 5 m , in two cases : short pulses with hyperbolic secant envelope @xmath20 , @xmath21 fs and @xmath22 kw ; and a 1-kw continuous wave . in both cases
, noise is added to the input electric field in order to achieve a realistic optical signal - to - noise ratio of 40 db ( pulsed ) and 60 db ( cw ) , where the noise power is measured in a 50-ghz ( pulsed ) and a 61-ghz ( cw ) bandwidth .
spontaneous raman emission is not taken into account in the analysis as its effect can be neglected compared to that of shot noise @xcite .
results are presented for an average of many noise realizations .
last , the waveguide is assumed to be an arbitrary 50-cm long . for a waveguide
this long , attenuation can be neglected , and , thus , is not included in equation [ eq : gnlse ] , as it can be lower than 0.6 db / m for a chalcogenide glass fiber at the pump wavelength @xcite .
also , the waveguide length was chosen to guarantee stability of the results , as we verified that longer lengths and/or higher peak powers lead to numerical instability when using standard double - precision , and reasonable small step sizes , in conjunction with the rk4-ip algorithm @xcite used to solve the gnlse .
other simulation parameters include ( resp . , pulsed and cw ) : integration step sizes of 100 and 2.5 micra ; 8192 and 16384 points for time / frequency representation ; 2.4 fs and 1 fs time resolutions ; and , 50 ghz and 61 ghz frequency resolutions .
output spectra are shown in figs . [
fig : sfig11 ] and [ fig : sfig12 ] for the pulsed ( average of 10000 noise realizations ) and cw ( average of a 1000 noise realizations ) cases , respectively .
we observe sc generation beyond 10 m ( measured at 20-db width ) .
several differences can be observed between the pulsed and cw cases .
while the maximum is above 3 m in the former case , it is close to 2 m in the latter .
moreover , in the pulsed case there is another local maximum close to 6 m which is not coincident with the pump wavelength .
finally , the spectrum decreases faster with decreasing frequency in the case of short pump pulses than in the continuous wave pump . in order to show that the generated sc is observed at a low pump power ,
it is interesting to compare our findings with results in the literature .
reference @xcite shows sc generation from 1.8 to 10 m ( measured at the 40-db point ) , by pumping a 11-cm chalcogenide fiber , @xmath23 ( w - km)@xmath9 , with 330 fs pulses of 3 kw peak power produced by an optical parametric amplifier ( opa ) . in terms of nonlinearity ,
the accumulated phase ( @xmath24 ) is 60% higher than in our case . in ref .
@xcite , a 10 m sc spectral width ( at the 30-db point ) is demonstrated by pumping a short chalcogenide waveguide at 4.184 m with 330 fs , 4.5 kw pulses , yielding an accumulated nonlinear phase similar to that in our case . in the aforementioned references
, the observed spectral width is narrower than the one reported here ; we also note that , while in both papers the pump wavelength is located between two zero - dispersion wavelengths ( zdws ) , we assume a simple dispersion profile with a single zdw .
moreover , the pump wavelength and power level in these references require the use of complex opas , while our proposal requires ready - to - use tabletop ( pulsed or cw ) co@xmath25/co lasers .
finally , simulation results in both papers use a dispersion profile modeled up to the 20th order . in this work , however , we show that it suffices to employ a simple 4th - order model capturing general dispersion characteristics of chalcogenide media . in another recent work @xcite , a comparable sc width
is reported , but with a pump power at the megawatt level .
as mentioned before , the dispersion profile is essentially flat in the region of interest , and this leads us to expand the propagation constant up to the fourth order . to justify this choice , in fig .
[ fig : sfig31 ] we show the output spectra ( pulsed case ) when considering dispersion effects up to the third , fourth , and fifth order , respectively .
we readily observe that the inclusion of the fifth order term leads to similar results as compared to the case where up to fourth - order dispersion is taken into account , in particular when looking in a one - octave span .
figure [ fig : sfig32 ] shows results for the cw case , where we observe no significant differences due to higher - order dispersion .
figure [ fig : fig2 ] shows the time evolution along the waveguide , displaying the complex pulse interaction expected from soliton dynamics ( fig .
[ fig : sfig21 ] , pulsed case ) and the signature of pulse trains and ensuing soliton dynamics characteristic of modulation instability - induced sc generation ( fig .
[ fig : sfig22 ] , cw case ) .
it is interesting to note that the emergence of new pulses occurs at approximately the same traveled distance into the fiber , @xmath26 cm , in both cases .
although one may be tempted to propose a soliton fission mechanism for the initial formation of sub - pulses in the pulsed - pump case , a few calculations show that this can not be the dominant factor .
the input pulse corresponds to a soliton of order @xmath27 @xcite .
the soliton fission length is usually estimated as @xmath28 cm @xcite , which is longer than the waveguide .
thus , modulation instability appears to be the cause of the initial optical pulse breakdown .
[ fig : sfig51 ] shows the evolution of the spectrum along the waveguide for the pulsed case .
the symmetrical broadening in the form of a ` v ' shape , observed in the first few centimeters into the waveguide , is a typical signature of self - phase modulation ( spm ) .
[ fig : sfig52 ] shows the average spectrum at 8 cm together with the spectrum in the case where the input pump is noiseless .
it also shows the gain spectrum due to modulation instability ( mi ) , calculated taking into account high - order dispersion @xcite , as a direct calculation using the standard formula ( @xmath29 , see @xcite ) yields a higher frequency offset .
taking into account that the sub - pulses that emerge near the input end of the waveguide are approximately 30 fs wide ( fig .
[ fig : sfig22 ] ) and noting that non - negligible spectral content at @xmath30 30 thz appears only in the noisy - input case ( fig .
[ fig : sfig52 ] ) , it can be concluded that the initial breakdown into pulses is due to noise - seeded modulation instability .
similarly , in the cw case , noise - seeded modulation instability leads to the formation of a pulse train with a period @xmath7 30 fs ( fig .
[ fig : sfig22 ] ) consistent with the location of the maxima of the mi - gain spectrum ( fig .
[ fig : sfig52 ] ) . from the previous results , and
the large amount of noise realizations , it is only natural to explore the possible occurrence of rogue events .
figures [ fig : sfig41 ] and [ fig : sfig42 ] show results obtained for pulse - intensity statistics for 5000 ( pulsed ) and 500 ( cw ) noise realizations , respectively .
it is worth noting that the cw case produces a much larger number of pulses for each realization as compared to the pulsed case ; as such , a significant number of events can be attained with a relatively small number of noise realizations .
in both cases , the classical signature of a long tail ( l shape ) distribution is observed , pointing at the enhanced probability of extreme events .
specifically , by choosing the usually adopted threshold for such events @xcite , namely the so - called significant wave height , corresponding to twice the average of the upper third of the intensity distribution , we find that approximately 0.3% and 0.15% correspond to rogue waves for the pulsed and cw cases , respectively .
insets in figures [ fig : sfig41 ] and [ fig : sfig42 ] show examples of some of the highest - power resulting pulses , @xmath7 7 kw in the pulsed case and @xmath7 14 kw in the continuous wave case .
it is interesting to emphasize that we observe the occurrence of these extreme events resorting to a very simple dispersion profile , as compared to results in the literature that include up to 10th - orden dispersion , and at similar levels of nonlinearity @xcite .
we showed , through numerical modeling , the feasibility of supercontinuum generation in the mid ir by pumping a chalcogenide waveguide with a pump wavelength at 5 m , both in the short pulse and cw regimes .
furthermore , in both cases initial breakdown into short pulses appears to be due to noise - seeded modulation instability , and supercontinuum generation proceeds through soliton dynamics and raman self - frequency shift .
we found substantial spectral content beyond 10 m , a promising result as it points to the feasibility of covering a larger part of the mid - ir window by means of sc generation through low - power pumping readily available from co~2~/co lasers .
further , we investigated the role of the dispersion profile of the chalcogenide medium , showing that inclusion of only up to the fourth - order term is sufficient to account for the observed rich pulse dynamics and sc spectral width within one - octave span .
extensive simulation of noise realizations in both pulsed and cw cases allowed us to study some of the statistical properties of the output pulses , and found evidence of rogue wave events . deeper analysis and control of such events may lead to more stable sc sources for real mid - ir applications such as molecular spectroscopy , metrology and tomography , among others .
the authors wish to thank e. cortizo , i. tarnapolsky , and j. bentez for enlightening discussions . | we present numerical results of supercontinuum ( sc ) generation in the mid - ir spectral region , specifically addressing the molecular fingerprint window ranging from 2.5 to 25 m. by solving the generalized nonlinear schrdinger equation ( gnlse ) in a chalcogenide waveguide , we demonstrate low - power sc generation beyond 10 m from a pump at 5 m. further , we investigate the short - pulse and cw regimes , and show that a simple linear dispersion profile , applicable to a broad range of chalcogenide media , is sufficient to account for the broad sc generation , and yield rich pulse dynamics leading to the frequent occurrence of rogue wave events .
results are encouraging as they point to the feasibility of producing bright and coherent light , by means of single low - power tabletop laser pumping schemes , in a spectral region that finds applications in such diverse areas as molecular spectroscopy , metrology and tomography , among others , and that is not easily addressable with other light sources .
_ keywords _ : mid - ir , supercontinuum , rogue wave , chalcogenide . |
discovery of grb afterglows has become almost commonplace . however , we are still in a regime where nearly every well - sampled afterglow contributes to our understanding of the phenomenon in new and meaningful ways .
observationally , grb 021211 distinguishes itself in two ways : ( 1 ) it is the second grb for which an optical afterglow was observed within minutes of the burst , thanks to rapid responses by the hete-2 satellite ( crew et al .
2002 , 2003 ) and three robotic telescopes raptor ( wozniak et al . 2002 ) ; kait ( li et al .
2002 , 2003 ) ; and super - lotis ( park , williams & barthelmy 2002 ; this paper ) ; and ( 2 ) it is the first grb for which filtered detections ( beginning 143 seconds after the burst ) and color information ( beginning 38 minutes after the burst ) were obtained at early times .
in addition to observations presented in grb coordinates network ( gcn ) circulars , many groups have presented their observations in peer - reviewed journals : li et al .
( 2003 ) present an unfiltered light curve beginning 105 seconds after the burst ; fox et al .
( 2003 ) present an unfiltered light curve beginning 21 minutes after the burst and filtered optical , nir , and radio observations beginning 2.0 hours after the burst ; pandey et al .
( 2003 ) present filtered optical observations beginning 6.8 hours after the burst ; holland et al .
( 2004 ) present filtered optical and nir observations of both the afterglow and host galaxy beginning 17 hours after the burst and measure the spectral flux distribution of the afterglow around 21 hours after the burst ; and smith et al .
( 2005 ) present submillimeter observations around 25 hours and 10 days after the burst .
finally , della valle et al .
( 2003 ) present photometric and spectral evidence for an associated supernova at late times . as in the case of grb 990123 ( akerlof et al .
1999 ) , the optical afterglow faded more rapidly at first , presumably due to a reverse shock ( wei 2003 ; fox et al . 2003 ; li et al . 2003 ; holland et al .
2004 ) . however , these afterglows differ in that grb 021211 was @xmath03 4 mag fainter , despite a lower redshift [ @xmath1 for grb 021211 ( vreeswijk et al . 2002 ; della valle et al . 2002 ) vs. @xmath2 for grb 990123 ( hjorth et al . 1999 ) ] ( fox et al . 2003 ; li et al . 2003 ; pandey et al .
2003 ; crew et al .
if it were not for the rapid response of the grb community , grb 021211 might have been called a `` dark burst '' : it faded from r @xmath0 14 mag at @xmath090 sec after the burst ( wozniak et al .
2002 ) to r @xmath3 21 mag about three hours later .
many bursts that would have been called `` dark '' in the bepposax era are being and will be called `` dim '' in the hete-2 , integral , and swift era due to faster responses .
some authors have modeled grb 021211 with an emphasis on its environment .
kumar & panaitescu ( 2003 ) argue that the grb and afterglow were produced by the same shock and within this framework constrain physical parameters for both constant - density and wind - swept media .
panaitescu & kumar ( 2004 ) consider the early - time afterglows of both grb 021211 and grb 990123 in the context of reverse - forward shock ( for both constant - density and wind - swept media ) and wind - bubble scenarios and find that the reverse - forward shock scenario is preferred .
chevalier , li & fransson ( 2004 ) argue for a wind - swept medium with the cooling break redward of the r band and within this framework find wind densities that are low compared to galactic wolf - rayet stars ( see also panaitescu & kumar 2004 ) .
finally , dado , dar & de rujula ( 2003 ) model grb 021211 within the framework of their cannonball model . in 2
, we present fun grb collaboration observations of grb 021211 , which include the earliest filtered detections and color information for an afterglow in the pre - swift era . in 3 , we fit standard afterglow and extinction curve models to these and other groups data and show that within the first few hours after the burst one and possibly two significant chromatic variations are superimposed on this `` standard model '' behavior . in 4 , we compare our results to previous modeling results and discuss possible reasons for these chromatic variations , including the possibility that they are due to a dust echo .
we also constrain physical parameters that describe the afterglow and surrounding medium for a variety of scenarios and discuss why grb 021211 s afterglow is so faint .
we draw conclusions in 5 .
long - duration , x - ray rich grb 021211 was detected by hete-2 s fregate , wxm , and sxc instruments on december 11 , 2002 at 11:18:34 utc ( crew et al .
the initial spacecraft localization was 14@xmath4 in radius and reported in near - real time , only 22 seconds after the burst .
ground analysis of the wxm and sxc data , reported 131 minutes after the burst , improved the localization to 2@xmath4 in radius and was consistent with the initial localization .
fox & price ( 2002 ) announced the discovery of an r @xmath5 18 mag and fading , stationary point source in the error circle 53 minutes after the burst .
while the pair labored , the robotic telescopes of three groups had already responded to the alert . for only the second time in the afterglow era ,
robotic telescopes extended the light curve of an afterglow back to within tens of seconds of the burst ( wozniak et al .
2002 ; li et al . 2002 , 2003
; park , williams & barthelmy 2002 ; this paper ) .
the dim and quickly fading afterglow soon grew too faint for small telescopes , and a possible host galaxy was detected ( lamb et al .
2002a , 2002b ; mcleod et al .
2002 ) but later confirmed under better seeing conditions to be cleanly separated from the afterglow by 1.5@xmath6 ( caldwell et al .
vlt spectroscopy of the true host galaxy resulted in a measured redshift of @xmath7 ( vreeswijk et al .
2002 ; della valle et al .
late - time observations indicate both a re - brightening at the time expected for a supernova at @xmath8 , and a spectrum that resembles that of type ic sn 1994i ( fruchter et al .
2002 ; della valle et al .
2003 ) .
we summarize fun grb collaboration observations of grb 021211 in table 1 .
we have calibrated all of our measurements using the field calibration of henden ( 2002 ) .
super - lotis imaged the entire grb 021211 field in r band beginning 143 and 309 seconds after the burst ( park , williams & barthelmy 2002 ) .
super - lotis is a fully automated f/3.5 0.6-meter diameter perkin - elmer telescope on a boller & chivens mount at kitt peak national observatory .
the camera is a 2048 @xmath9 2048 loral ccd , which yields a large , 51@xmath4 @xmath9 51@xmath4 field of view .
observations began automatically after receiving the hete-2 alert via a socket connection to the gcn .
both exposures were 60 seconds in duration .
the mean times that we list in table 1 are flux weighted using an iterated power - law index of @xmath10 , since the exposure time is comparable to the age of the burst , at least for the first exposure .
this results in small shifts of 2.4 and 1.2 seconds in the mean times of these observations .
tenagra observatories , ltd .
, 0.81-meter tenegra ii telescope imaged the grb 021211 field beginning 37 minutes after the burst .
we obtained four sets of four images , each set in a 2 @xmath9 2 arrangement to cover the initial 28@xmath4-diameter localization and each in a different filter ( i@xmath11r@xmath11vb ) .
we then re - pointed to the candidate afterglow of fox & price ( 2002 ) and cycled through i@xmath11r@xmath11 thrice more . of these , we combined the first two i@xmath11 and r@xmath11 images to optimize signal to noise , but the final two images were not usable due to the onset of morning .
this resulted in three detections ( i@xmath11r@xmath11v ) , a limit ( b ) , and two more detections ( i@xmath11r@xmath11 ) .
we reduced the images using iraf s ccdred package and performed psf photometry using iraf s daophot package .
we imaged the central 11@xmath4 @xmath9 11@xmath4 of the initial 28@xmath4-diameter localization in r@xmath11 band beginning 85 minutes after the burst from gunma astronomical observatory , located in agatsuma , gunma , japan ( kinugasa et al . 2002 ) .
we used the f/12 0.65-meter diameter cassegrain telescope , which is equipped with an apogee ap8 1024 @xmath9 1024 back - illuminated site ccd .
we obtained a total of 28 images , which we combined to optimize signal to noise .
we reduced the images using iraf s ccdred package and performed psf photometry on the combined image using iraf s daophot package .
we reacquired the field with the 1.34-meter diameter tautenburg schmidt telescope 11.7 hours after the burst and imaged in r and i bands for the next 1.1 hours using the 2048 @xmath9 2048 prime - focus ccd ( klose et al .
however , we did not detect the afterglow .
we began observations with the 3.5-meter diameter astrophysics research consortium ( arc ) telescope at apache point observatory 22.0 hours after the burst , and returned to the field on december 28 and march 23 , 17 and 102 days after the burst ( lamb et al .
2002a , 2002b ) .
all images were taken in i@xmath12 band using spicam , a 2048 @xmath9 2048 back - illuminated site ccd .
three 2000-second images were taken on the first night , and seven 1200-second images were taken on each of the following nights .
we reduced , combined , and calibrated these images using iraf s ccdred and daophot packages . finally , we re - observed the field on december 13 with the 1.0-meter diameter telescope at the u.s .
naval observatory s flagstaff station for purposes of calibration ( henden 2002 ) .
bvr@xmath11i@xmath11 images were taken with a 2048 @xmath9 2048 back - illuminated site / tektronix ccd under 2.2@xmath6 seeing conditions .
upon inspection of the images , the afterglow was still marginally visible in the 8-minute v - band image .
the afterglow was measured using a two - fwhm diameter aperture with iraf s daophot package .
supernova signatures had been found for many grbs prior to grb 021211 ( e.g. , galama et al . 1998 ; bloom et al . 1999 ; reichart 1999 ; galama et al .
2000 ; bloom et al . 2002 ; garnavich et al .
2003 ; price et al . 2003 ; stanek et al .
2003 ; see also zeh , klose & hartmann 2004 for a systematic analysis ) . for grb 021211 , fruchter et al .
( 2002 ) and della valle et al .
( 2003 ) found evidence for excess red light @xmath025 days after the burst , and della valle et al .
( 2003 ) obtained a vlt spectrum at 27 days .
this spectrum exhibits ca ii absorption with a relative velocity of @xmath014,440 km / s for @xmath13 and is similar to other type ic spectra .
our late - time arc observations neither confirm nor contradict the existence of this underlying supernova .
subtraction of our second and third i@xmath12 epochs using isis-2 ( alard 2000 ) does not reveal any residual flux .
however , this is likely due to the timing of our observations : the re - brightening reported by della valle et al .
( 2003 ) occurs mostly between our observations at 17 and 102 days after the burst . in figure 1
, we plot our i@xmath12 light curve and the fitted afterglow model of 3.1 .
to better investigate possible chromatic variations that occurred during the unfiltered kait and neat observations ( see 3.2 ) , we have recalibrated these measurements from the r@xmath11 band to broad bands given by the spectral responses of their respective ccds ( pravdo et al . 1999 ; li et al .
2003 ) : w. li ( private communication ) and p. price ( private communication ) kindly provided us with their calibration stars . using the bvr@xmath11i@xmath11 magnitudes of these stars from henden ( 2002 ) , we fitted extinguished blackbody functions to each of these stars and then integrated these fitted functions against the appropriate spectral response curve .
this resulted in small , 0.05 and 0.03 magnitude offsets in the calibration of the kait and neat measurements , respectively .
we now fit standard afterglow and extinction curve models to these and other groups data and show that within the first few hours after the burst one and possibly two significant chromatic variations are superimposed on this `` standard model '' behavior .
the data that we include in this analysis are plotted in figure 2 and consist of fun grb collaboration data ( 2.1 ) , data previously published in peer - reviewed journals ( pandey et al .
2003 ; fox et al .
2003 ; li et al . 2003 ; holland et al .
2004 ) , and data from the gcn archive ( mcleod et al .
these data span the first @xmath02.5 days after the burst , after which the host galaxy and supernova become contaminants .
all magnitudes have been converted to spectral fluxes as prescribed by bessell ( 1979 ) and bessell & brett ( 1998 ) .
we now model these data and constrain model parameters .
we model the afterglow with two components , corresponding to reverse and forward shocks .
each component has a power - law light curve and a power - law spectrum , but the spectrum is extinguished by dust in the source frame and in our galaxy and absorbed by hydrogen in the source frame and the ly@xmath14 forest : @xmath15,\ ] ] where @xmath16 is the galactic extinction curve model of cardelli , clayton & mathis ( 1989 ) , @xmath17 is the ly@xmath14 forest absorption model of reichart ( 2001a ) , @xmath18 is the source - frame extinction curve and lyman limit absorption model of reichart ( 2001a ) , @xmath19 and @xmath20 are the temporal indices of the two components , @xmath21 and @xmath22 are the spectral indices of the two components , @xmath23 is the effective frequency of the r band , @xmath24 is the time when these two components are of equal brightness at this frequency , and @xmath25 is this brightness . since the extinction and absorption models have features that are narrower than most photometric bands , we integrate equation 1 against the appropriate filter transmissivity curve ( or ccd spectral response curve for the unfiltered measurements ; 2.3 ) before fitting it to the data .
we fit this model to the data using bayesian inference ( e.g. , reichart 2001a ; lee et al .
2001 ; galama et al .
2003 ) : the posterior probability distribution is equal to the product of the prior probability distribution and the likelihood function .
the likelihood function is given by : @xmath26 ^ 2}{\sigma_i^2 + \sigma^2}\right\},\ ] ] where @xmath27 is the number of measurements , @xmath28 is the above described integration of equation 1 against the spectral curve of the @xmath29th measurement at the time of the @xmath29th measurement ; @xmath30 is the @xmath29th measurement in units of log spectral flux ; @xmath31 is the uncertainty in the @xmath29th measurement in the same units , and @xmath32 is a parameter , sometimes called the slop parameter , that models the small systematic errors that are unavoidably introduced when data are collected from many sources , and other small sources of error ( reichart 2001a ) . ignoring this parameter can lead to erroneous fits and significantly underestimated uncertainties in the fitted parameter values when the scatter of the measurements about the fitted model exceeds that which can be accounted for by the measurement uncertainties alone .
many of the parameters of the source - frame extinction curve model and all of the parameters of the ly@xmath14 forest absorption and galactic extinction curve models can be constrained a priori .
the source - frame extinction curve model of reichart ( 2001a ) is a function of eight parameters : the source - frame v - band extinction magnitude @xmath33 , @xmath34 , the intercept @xmath35 and slope @xmath36 of the linear component of the source - frame uv extinction curve , the strength @xmath37 , width @xmath38 and center @xmath39 of the uv bump component of the extinction curve , and the strength @xmath40 of the fuv excess component of the extinction curve .
the ly@xmath14 forest absorption model of reichart ( 2001a ) is a function of a single parameter , @xmath41 , the flux deficit .
reichart ( 2001a ) determines prior probability distributions for @xmath42 , @xmath35 , @xmath38 , @xmath39 , and @xmath41 , which means that the values of these parameters can be weighted by fairly narrow distributions , the description of which sometimes depends on other parameters ( @xmath36 and @xmath43 ) , a priori .
we adopt these priors here , which can be thought of as increasing the degrees of freedom by five . also , the galactic extinction curve model of cardelli , clayton & mathis ( 1989 ) is a function of @xmath44b@xmath45v@xmath46 mag for this line of sight ( schlegel , finkbeiner & davis 1998 ) and a single parameter , @xmath47 .
we adopt a prior for this parameter that is log normally distributed with mean @xmath48 and width 0.1 , which closely approximates the distribution of values of this parameter along random lines of sight through the galaxy ( e.g. , reichart 2001a ; lee et al .
2002 ; galama et al .
2003 ) .
we fit our model to the data for each of the four standard cases of sari , piran & narayan ( 1998 ) and chevalier & li ( 2000 ) , which relate @xmath20 to @xmath22 assuming ( 1 ) propagation through either a constant - density ( ism ) or wind - swept ( wind ) medium , and ( 2 ) a cooling break that is either redward ( red ) or blueward ( blue ) of the observed optical and nir frequencies : for the ism - red and wind - red cases , @xmath49 ; for the ism - blue case , @xmath50 ; and for the wind - blue case , @xmath51 , where @xmath52 is the power - law index of the electron - energy distribution .
since the temporal index is well constrained by the data , these additional constraints can be powerful tools for separating the intrinsic spectrum from extinction effects ( see 4.1 ) .
for purposes of comparison , we also fit our model to the data free of constraints on @xmath20 and @xmath22 .
best fits are found by maximizing the posterior .
compared to the wind - blue case , we can rule out the ism - red and wind - red cases at the 7.3@xmath32 credible level , and the ism - blue case is disfavored at the 3.1@xmath32 credible level .
furthermore , the wind - blue fit is consistent with the constraint - free fit , differing from it at only the 0.6@xmath32 credible level .
the primary difference between these cases is that the wind - blue case requests a shallow intrinsic spectrum , @xmath53 , and a small amount of extinction , @xmath54 mag , where the other cases request steeper intrinsic spectra and would fit better if @xmath55 mag were possible ( see 4 ) .
for the wind - blue case , we find that @xmath56 @xmath57jy , @xmath58 day , @xmath59 , @xmath60 , @xmath61 , @xmath62 , @xmath63 mag , @xmath64 ( 1@xmath32 ) , and @xmath65 mag
. the parameters @xmath37 and @xmath40 could not be constrained by the data .
we plot best - fit light curves for 13 spectral bands in figure 2 and best - fit spectral flux distributions for six epochs in figure 4 .
we plot the residuals of figure 2 in figure 3 .
one and possibly two significant chromatic variations can be seen from @xmath040 minutes after the burst until possibly @xmath06.0 hours after the burst .
the first of these is an increase relative to the best - fit model of the unfiltered neat and kait data , which is also clearly visible in figure 2b , concurrent with a decrease relative to the best - fit model of our r@xmath11 and possibly i@xmath11 data from tenagra and gunma .
since the neat and kait bandpasses are broad , encompassing the i@xmath11 and r@xmath11 bands on their red ends , this suggests that there was an excess of blue light at this time . to explore this further
, we plot the spectral flux distribution of the afterglow in six time slices in figure 4 : in figure 4a , we plot the best - fit spectral flux distribution at 67 minutes after the burst and have scaled all of the data between 39 and 94 minutes after the burst to this time using the best - fit light curve .
these data consist of i@xmath11r@xmath11vb data from tenagra and unfiltered neat data .
we plot the weighted average of the scaled neat data for clarity .
the combined neat point is only 0.14 mag above the best - fit model , but significantly so , at the 5.2@xmath32 confidence level . in figure 4b , we plot the best - fit spectral flux distribution at 2.2 hours after the burst .
we have scaled all of the data between 1.8 and 2.5 hours after the burst to this time using the best - fit light curve and plot weighted averages of the scaled data when there are multiple points per spectral band .
these data consist of a k@xmath66 point from fox et al .
( 2003 ) , i@xmath11r@xmath11 data from tenagra , an r@xmath11 point from gunma , and unfiltered kait data .
the combined kait point is even farther above the best - fit model , 0.35 mag , this time at the 3.3@xmath32 confidence level .
the i@xmath11 point is now below but still consistent with the best - fit model and the r@xmath11 point is below the best - fit model at the 2.3@xmath32 confidence level .
consequently , the kait point differs from the r@xmath11 point at the 4.0@xmath32 confidence level with respect to the best - fit model .
since the kait bandpass , like the neat bandpass , is broad , encompassing the i@xmath11 and r@xmath11 bands on its red end , this suggests that there was an excess of blue light at this time .
if we model this excess as an additional power - law component , just in this time slice , we find it to be bluer than @xmath67 at the 2@xmath32 credible level . in figure 4c
, we plot the best - fit spectral flux distribution at 4.4 hours after the burst .
we have scaled two points from fox et al .
( 2003 ) a b point at 3.1 hours after the burst and an r@xmath11 point at 5.7 hours after the burst to this time using the best - fit light curve .
one possibility is that the excess light has changed from blue to red : the r@xmath11 point is above the best - fit model at the 3.1@xmath32 confidence level and the b point is below the best - fit model at the 3.7@xmath32 confidence level .
consequently , these points differ at the 4.8@xmath32 confidence level with respect to the best - fit model .
however , given the sparsity of the data in this time slice a temporal variation can not be ruled out either .
the remaining panels , corresponding to 6.8 11 , 17 25 , and 46 48 hours after the burst , show no evidence for significant chromatic variations at later times .
although the third time slice is too sparsely sampled for a temporal variation to be ruled out , the first two time slices , which span the first proposed chromatic variation , are better sampled .
consider the following simple model : let @xmath68 be the beginning of this variation .
prior to @xmath68 , the afterglow is described by equation 1 . between @xmath68 and 2.5 hours after the burst , the neat and kait data are instead described by temporal index @xmath69 and the r@xmath11 and i@xmath11 data are instead described by temporal index @xmath70 .
if this were a temporal variation , @xmath69 would equal @xmath70 . instead
, we find that @xmath71 min and @xmath72 with @xmath73 at the 3.5@xmath32 credible level . here
, we have fixed all of the other parameters to their previous best - fit values so we can also plot this best fit in figure 3 .
allowing all of the parameters to vary , we find that @xmath73 at the 3.3@xmath32 credible level , which again suggests that this is a chromatic variation .
in the above fit , we also find that @xmath74 , which is somewhat steeper than the fitted value of @xmath20 .
this suggests that the light curve might be steepening during the first and second time slices , and this is consistent with the b point at the beginning of the third time slice also undercutting the model ( e.g. , figure 4c ) .
however , the data are consistent with the model in the fourth , fifth , and sixth time slices , which suggests a minor rebrightening during the third time slice .
such minor temporal variations are now commonplace
grbs 021004 and 030329 are extreme examples but further modeling of such variations is beyond the scope of this paper , and frankly beyond the quality of this data set .
however , this does lend some credibility to the possibility that the second variation is temporal instead of chromatic .
a final possibility is that the r point at the end of the third time slice is a statistical variation : given 80 points , the probability of encountering a 4.8@xmath32 variation is 1 in 7900 ( ruled out at the 3.8@xmath32 confidence level ) .
finally , we refit the four standard cases to the data , but this time we accommodate the first , chromatic variation with the above simple model and eliminate the second variation , whether chromatic or temporal , by not fitting to the two points of the third time slice . compared to the wind - blue case , we now rule out the ism - red and wind - red cases at the 6.3@xmath32 credible level and the ism - blue case at the 3.2@xmath32 credible level .
for the wind - blue case , we find that @xmath75 @xmath57jy , @xmath76 day , @xmath77 , @xmath78 , @xmath79 , @xmath80 , @xmath81 mag , @xmath82 ( 1@xmath32 ) , @xmath83 min , @xmath84 , @xmath85 , and @xmath86 mag .
the primary difference between this fit and the wind - blue fit of 3.1 in which these variations are not treated is we now find more source - frame extinction , @xmath81 mag with @xmath87 mag at the 2.8@xmath32 credible level .
also , @xmath70 is now consistent with @xmath20 , but @xmath88 is still greater than zero at the 3.0@xmath32 credible level .
our finding that the data are best described by the wind - blue case differs from the findings of others .
fox et al .
( 2003 ) discount this case in favor of the ism - blue case , arguing that if the early - time emission is due to a reverse shock , in a wind - swept medium it is expected to fade quickly and they measure a slower fading : @xmath89 .
however , chevalier , li & fransson ( 2004 ) point out that this measurement depends sensitively on how one subtracts out ( or models ) the forward - shock component , arguing that the value is closer to @xmath90 .
using final instead of gcn data , we find that @xmath59 ( variations untreated ; 3.1 ) or @xmath91 ( variations treated ; 3.2 ) .
however , in 4.2 we point out that emission from the reverse shock is not necessarily expected to fade quickly in a wind - swept medium if @xmath92 and other physical parameters are lower than expected , which appears to be the case for this grb .
holland et al . (
2004 ) also adopt the ism - blue case . the primary difference between their fit and ours
is that we permit source - frame extinction .
when we fit the ism - blue case , we find that @xmath93 ( variations untreated ) or @xmath94 ( variations treated ) with @xmath95 mag , which is very similar to their fit in a time slice around 0.88 days after the burst : @xmath96 with @xmath33 assumed to be zero . however , if source - frame extinction is permitted and one fits to all of the data , we find that the wind - blue case with a small to moderate amount of source - frame extinction , @xmath63 mag ( variations untreated ) or @xmath97 mag ( variations treated ) , is preferred at the 3.1@xmath32 ( variations untreated ) or 3.2@xmath32 ( variations treated ) credible level . figures 4e and 5e can be directly compared to figure 3 of holland et al .
( 2004 ) .
finally , chevalier , li & fransson ( 2004 ) adopt the wind - red case , guided by sparse color information that was available at the time , including the two points of figure 4c , which we have already identified as discrepant , possibly due to excess red light at this time ( 3.2 ) .
permitting source - frame extinction and fitting to all of the data , we rule this case out at the 7.3@xmath32 ( variations untreated ) or 6.3@xmath32 ( variations treated ) credible level .
the wind - blue case , however , requires a relatively hard electron energy distribution
so a break at higher energies is required .
bhattacharya ( 2001 ) determines the effect of @xmath99 on the standard equations : by introducing a cut - off frequency @xmath100 such that @xmath101 and assuming that @xmath100 evolves directly with the bulk lorentz factor of the shock , they find results similar to the standard prescriptions .
galama et al .
( 2003 ) found a similar hard electron energy index for grb 010222 , though other ideas , such as a continuous injection of energy ( bjornsson et al .
2002 ) or an early transition to non - relativistic motion ( int zand et al . 2001 ; masetti et al . 2001 ) , have been proposed . following the analysis of chevalier , li & fransson ( 2004 ) ,
but for the wind - blue case , and using the analytic expressions of granot & sari ( 2002 ) , we now constrain physical parameters that describe the afterglow and surrounding medium for a variety of scenarios .
the first constraint comes from the expression of granot & sari ( 2002 ) for the brightness of the afterglow in the frequency range of our observations , which for the wind - blue case is max\{@xmath102,@xmath103 } @xmath104 , where @xmath102 is the self - absorption frequency , @xmath103 is the typical synchrotron frequency , and @xmath105 is the electron cooling frequency .
this corresponds to segment g in their figure 1 .
for @xmath106 , a luminosity distance of @xmath107 cm ( assuming that @xmath108 , @xmath109 , and @xmath110 km s@xmath111 mpc@xmath111 ) , and an extinction - corrected @xmath112 = 19 @xmath57jy at 0.1 days after the burst , we find : @xmath113 where @xmath114 is the electron energy fraction when @xmath99 , @xmath115 is the magnetic field energy fraction , @xmath92 measures the density of the wind - swept medium , and @xmath116 erg is the total energy of the shock , if spherical .
the second constraint comes from the expression of granot & sari ( 2002 ) for @xmath117 and the fact that the data are well described by the wind - blue case even at early times ( see 4.3 ) .
taking @xmath118 prior to 3.9 minutes after the burst the time when the forward shock first outshines the reverse shock in the r band yields : @xmath119 where @xmath120 is the time that @xmath105 passes above the r band .
the third constraint is similar to the second in that we take @xmath121 prior to 3.9 minutes after the burst , else the light curve would have faded much more slowly at this time , as @xmath122 ( chevalier & li 2000 ; chevalier , li & fransson 2004 ) : @xmath123 where @xmath124 is the time that @xmath103 passes below the r band .
the final constraint comes from the expression of granot & sari ( 2002 ) for the brightness of the afterglow at 8.5 ghz , given that @xmath125 @xmath57jy at a mean time of 13 days after the burst ( fox et al .
2003 ) . here
we consider four scenarios : ( a ) @xmath126 ghz @xmath127 , ( b ) 8.5 ghz @xmath128 min\{@xmath102,@xmath103 } , ( c ) max\{@xmath102,@xmath103 } @xmath129 ghz , and ( d ) @xmath130 ghz @xmath131 . for scenario
a , using the expression of granot & sari ( 2002 ) that corresponds to their segment d , we find : @xmath132 combining equations 3 , 4 , 5 , and 6 yields : @xmath133 @xmath134 @xmath135 @xmath136 if we additionally require that @xmath137 ghz @xmath127 at 13 days after the burst , using the expressions of granot & sari ( 2002 ) for @xmath102 and @xmath103 , we find that @xmath138 @xmath57jy and @xmath139 . since the former of these constraints contradicts equation 10 for any value of @xmath140 min , we rule out this scenario . for scenario
b , using the expression of granot & sari ( 2002 ) that corresponds to their segment b , we find : @xmath141 combining equations 3 , 4 , 5 , and 11 yields : @xmath142 @xmath143 @xmath144 @xmath145 if we additionally require that 8.5 ghz @xmath128 min\{@xmath102,@xmath103 } at 13 days after the burst , we find that @xmath146 min , which is technically consistent with @xmath147 min , and @xmath148 @xmath57jy , which is technically consistent with @xmath149 @xmath57jy
. however , this constrains these parameters values to narrow ranges and by equation 15 implies a value for @xmath150 that is much too low , given that the isotropic - equivalent energy in gamma rays alone was @xmath151 erg ( holland et al .
2004 ) or @xmath152 erg ( lamb et al .
2004 ) . assuming that the efficiency at which energy is converted to gamma rays is @xmath520% ( e.g. , beloborodov 2000 ) , then @xmath153 many .
consequently , we rule out this scenario as well . for scenario c , using the expression of granot & sari ( 2002 ) that corresponds to their segment g , we find : @xmath154 combining equations 3 , 4 , 5 and 16 yields : @xmath155 @xmath134 @xmath135 @xmath156 if we additionally require that max\{@xmath102,@xmath103 } @xmath129 ghz at 13 days after the burst , we find that @xmath157 minutes and @xmath158 , neither of which are problematic . taking @xmath150 to be as low as 4 and @xmath120 and @xmath124 to be as low as the duration of the burst ( @xmath159 sec in the 30 85 kev band , in which @xmath160 peaks ; crew et al .
2003 ) , yields @xmath161 . in this case ,
@xmath162 and @xmath163 .
however , for @xmath115 to be this low requires considerable fine tuning : if @xmath150 is as high as 11 , @xmath120 is as high as 54 sec , @xmath124 is as high as 11 sec , or lesser combinations of these three , @xmath164 .
consequently , @xmath115 is likely considerably more than 0.04 , in which case @xmath165 can be no more than a factor of three greater and is likely less and @xmath92 can only be less . finally , for scenario d , using the expression of granot & sari ( 2002 ) that corresponds to their segment a
, we find : @xmath166 combining equations 3 , 4 , 5 and 21 yields : @xmath167 @xmath168 @xmath169 @xmath170 if we additionally require that @xmath130 ghz @xmath131 at 13 days after the burst , we find that @xmath146 min , which is technically consistent with @xmath140 min , and @xmath171 @xmath57jy , which is consistent with @xmath125 @xmath57jy . once again , this constrains these parameters values to relatively narrow ranges and by equation 25 implies a value for @xmath150 that is much too low .
consequently , we rule out this scenario as well .
consequently , we find that @xmath172 after @xmath1283.9 minutes after the burst and max\{@xmath102,@xmath173 } @xmath129 ghz around 13 days after the burst . in this scenario ,
@xmath114 and @xmath92 are considerably lower than canonical values .
since @xmath174 ( equation 3 ) , both of these contribute to the faintness of the afterglow ( 1 ) .
if we allow ourselves to be guided by the standard - energy result ( frail et al . 2001 ; bloom , frail & kulkarni 2003 ) , @xmath150 is also lower than what one might expect : for wind - swept media , the total energy released in gamma rays is typically measured to be many @xmath9 10@xmath175 erg ( bloom , frail & kulkarni 2003 ) .
given that the isotropic - equivalent energy in gamma rays for grb 021211 was @xmath176 erg ( holland et al .
2004 ) or @xmath152 erg ( lamb et al .
2004 ) , this implies a jet opening / viewing angle of @xmath520@xmath177 , which is about three times the canonical value .
hence , @xmath150 is probably about an order of magnitude less than what one might have expected .
consequently , we find that grb 021211 s afterglow is faint for a combination of 3 4 reasons : ( 1 ) a low fraction of energy in relativistic electrons , ( 2 ) a low density for the wind - swept medium , implying either a low mass - loss rate and/or a high wind velocity for the progenitor , ( 3 ) a wide opening / viewing angle for the jet , and possibly ( 4 ) moderate source - frame extinction ( 3.2 ) .
furthermore , with @xmath178 and @xmath115 likely much greater than 0.04 the jet appears to be significantly far from equipartition and magnetically dominated .
this is similar to sn 1993j , for which the magnetic field energy density is @xmath510@xmath179 times the relativistic particle energy density ( chandra et al .
2004 ) , but dissimilar to sn 1998bw ( kulkarni et al . 1998 ) and sn 2002ap ( bjornsson & fransson 2004 ) , which appear to be near equipartition .
these findings are supported by the existence of the bright reverse shock : fox et al .
( 2003 ) dismiss the possibility of a wind - swept medium because for canonical values of @xmath180 , @xmath181 , @xmath92 , and @xmath150 , @xmath182 is expected to be significantly less than @xmath183 , in which case the reverse shock is expected to fade away quickly and not be bright ( chevalier & li 2000 ) . however , this is not the case when @xmath180 , @xmath181 , and/or @xmath92 are sufficiently low .
taking @xmath184 and @xmath185 and substituting equations 17 , 18 , and 19 into equations 45 and 47 of chevalier & li ( 2000 ) , we find : @xmath186 for @xmath187 , @xmath188 , and @xmath189 .
for @xmath190 minute after the burst , corresponding to the beginning of the first detection of the reverse shock ( wozniak et al .
2002 ) , @xmath150 need only be greater than @xmath50.03 for @xmath191 .
globally , the data are well described by the model of 3.1 , but superimposed on this global behavior are small , but significant variations from @xmath040 minutes after the burst until possibly @xmath06.0 hours after the burst ( 3.2 ) .
this model does not explain these variations .
it merely attemps to accommodate them with a higher value of the slop parameter : @xmath192 mag .
the slop parameter is a global measure of the scatter of the data around the model , beyond what can be accounted for by the data s error bars ( 3.1 ) .
furthermore , at least the first of these variations appears to be chromatic , with a relative increase of blue light with respect to red light around two hours after the burst , and possibly , but less certainly , a reversal of this two hours later ( 3.2 ) .
when we modify the model and fit to better treat these variations , @xmath32 decreases to @xmath193 mag ( 3.2 ) .
one potential explanation for these variations is that we have undersampled a light curve that is varying in time in such a way as to mimic a chromatic effect .
indeed , the high signal - to - noise light curves of grbs 021004 and 030329 revealed temporal variations and a variety of explanations have been proposed , including variations in the density of the external medium ( e.g. , lazzati et al .
2002 ) , refreshed shocks ( e.g. , granot , nakar & piran 2003 ) , and patchy shells ( e.g. , kumar & piran 2000 ) . however , none of these occur on a sufficiently short timescale to explain the variation around two hours after the burst ( figure 3 ) .
however , a temporal variation can not be ruled out for the possible reversal of this effect two hours later ( 3.2 ) .
another potential explanation is that a spectral break , presumably the cooling break , is passing through our data around 2 4 hours after the burst .
although this is difficult to reconcile with the blue excess of figure 4b , since the spectrum is supposed to be half of a spectral index steeper blueward of the cooling break , it is not necessarily inconsistent with the possible red excess of figure 4c .
however , the spectra in figures 4a and 4e would then differ by half of a spectral index , which would be noticeable .
another potential explanation is that we observed a dust echo light scattered by dust into the line of sight and received with a time delay due to the greater path length .
waxman & draine ( 2000 ) and esin & blandford ( 2000 ) originally proposed dust echoes as an alternative explanation for the supernova - like components to the afterglows of grb 980326 ( bloom et al .
1999 ) and grb 970228 ( reichart 1999 ; galama et al
reichart ( 2001b ) modeled and computed dust echo light curves and spectral flux distributions and found that while dust echoes can mimic supernova light curves they can not mimic supernova spectral flux distributions , at least not near the spectral peak .
moran & reichart ( 2004 ) take the model of reichart ( 2001b ) and instead of applying it to dust shells of inner radius @xmath510@xmath194 cm , which is what is required to mimic supernova light curves , they apply it to dust shells of inner radius @xmath510@xmath195 cm , which is typical of late - type wc wolf - rayet stars , the likely progenitors of grbs .
moran & reichart ( 2004 ) find that for ( 1 ) wind - swept media , ( 2 ) bright optical flashes reverse shocks that outshine the forward shock at early times and ( 3 ) wide jet opening angles , dust echoes may be observable on a timescale of minutes to hours after the burst .
furthermore , the characteristic signature will be an excess of blue light ( since blue light scatters preferentially ) that quickly transitions to an excess of red light ( due to increasing path lengths through dust with increasing time delay ) . since all of these conditions appear to be met for grb 021211 ( 3.1 ,
4.1 , 4.2 ) , and excess blue light is observed on this timescale , as well as a possible transition to excess red light hours later , we now test this hypothesis in two ways : \1 .
equation 6 of moran & reichart ( 2004 ) gives the turn - on time of an idealized , on - axis dust echo as a function of the inner radius @xmath196 of the circum - progenitor dust shell and the initial opening angle @xmath197 of the jet .
dust echoes should not be visible at very early times because x rays from the burst and uv light from the optical flash should sublimate the dust within @xmath197 of the jet axis this close to the burst ( e.g. , waxman & draine 2000 ; fruchter , krolik & rhoads 2001 ; reichart 2001c ) .
taking the turn - on time to be @xmath50.4 of the peak time ( reichart 2001b ; moran & reichart 2004 ) and taking the peak time for grb 021211 to be @xmath52.2 hours , we solve for @xmath196 : @xmath198 for a wide jet opening angle , this yields @xmath199 cm , which is the expectation if late - type wc stars are indeed the progenitors of grbs .
however , given that the isotropic - equivalent energy in gamma rays alone was @xmath151 erg ( holland et al .
2004 ) or @xmath152 erg ( lamb et al . 2004 ) and that @xmath200 ( 4.2 ) , the deceleration radius @xmath201 was likely greater than @xmath202 cm ( e.g. , equation 7 of moran & reichart 2005 ) and consequently a value for @xmath196 likely can not be deduced .
equation 5 of moran & reichart ( 2004 ) , but with @xmath201 substitued for @xmath196 , gives the peak brightness of the optical flash off of the jet axis , at angles around @xmath197 , as a function of the peak brightness of the dust echo , the deceleration radius , the optical depth @xmath203 through the rest of the dust shell at frequency @xmath204 , also at angles around @xmath197 , and the duration of the optical flash . taking the peak brightness of the dust echo to be @xmath519 mag and the duration of the optical flash to be @xmath530 seconds ( e.g. , sari & piran 1999 ) , we find : @xmath205 using our best fit ( 3.1 ) , extrapolation of the reverse shock light curve back to @xmath030 seconds after the onset of the burst yields r @xmath206 mag .
however , given the known distance to grb 021211 @xmath207 would have to be significantly fainter lest the dust be sublimated in these directions as well , this close to the burst .
for longer optical flash durations , @xmath208 , and/or a greater forward scattering probability than what esin & blandford ( 2000 ) , reichart ( 2001b ) , and moran & reichart ( 2004 ) assume , @xmath207 would be fainter , but not sufficiently .
alternatively , prior fragmentation of the dust to pah levels by gamma rays from the burst might harden it against sublimation , since atomic bonds would then be more difficult to break
. finally , we point out that neither of these estimates hold in the case of a jet with a narrow opening angle but a large viewing angle .
grb 021211 is one of only a handful of grbs for which processes other than the forward shock have been identified at optical wavelengths , which has made it one of the most studied grbs . in this paper
, we present additional , multi - band observations of this event , ranging from minutes to months after the burst , which in combination with all previously published observations have allowed us to deeply probe the physics of this grb and properties of its circum - progenitor environment .
coupling the standard afterglow model with a general - purpose extinction curve model , we find that the afterglow is best described by propagation through a wind - swept medium , which implies a massive - star progenitor ( e.g. , price et al .
the jet itself appears to be significantly far from equipartition and magnetically dominated .
indeed , the low fraction of energy in relativistic electrons appears to be the primary reason that this afterglow is so faint .
this , combined with a low - density medium , a wide jet opening / viewing angle , and possibly moderate extinction might be important clues as to why many afterglows are dark / dim in the swift era , even at early times after the burst .
these findings are supported by the existence of the bright reverse shock in a wind - swept medium this should only be possible if @xmath92 is low and/or the jet is significantly far from equipartition , meaning that either @xmath180 or @xmath181 is low as well . finally , we observed one and possibly two significant chromatic variations hours after the burst .
we discuss possible reasons for these variations , including the possibility that they are due to a dust echo : the three primary requirements for an observable dust echo are a wind - swept medium , a bright optical flash , and a wide jet opening angle , and the characteristic signature should be an excess of blue light that quickly reddens , all of which appear to be satisfied for grb 021211 . however ,
in the case of grb 021211 this would imply an off - axis brightness and hence luminosity for the optical flash that would probably sublimate too much dust in these directions .
rapid , multi - band , and preferably simultaneous multi - band observations of future grbs might shed more light on this interesting possibility .
der very gratefully acknowledges support from nsf s mri , career , prest , and reu programs , nasa s apra , swift gi and ideas programs , unc s junior faculty development award , and especially leonard goodman and henry cox .
hsp acknowledges support from nasa s sr&t program and doe / uc llnl contract w-7405-eng-48 .
kk acknowledges support from ministry of education , culture , sports , science and technology of japan grant - in - aid 16740121 .
dql acknowledges support from nasa subcontract nagw-4690 ( hete-2 ) and nasa s ltsa program .
lcccc dec 11.4732 & 2.84 min & r & 15.24 @xmath209 0.07 & 0.60 m super - lotis + dec 11.4751 & 5.63 min & r & 16.26 @xmath209 0.12 & 0.60 m super - lotis + dec 11.4986 & 39.4 min & i@xmath11 & 18.60 @xmath209 0.11 & 0.81 m tenagra ii + dec 11.5114 & 57.9 min & r@xmath11 & 19.52 @xmath209 0.13 & 0.81 m tenagra ii + dec 11.5239 & 75.9 min & v & 20.06 @xmath209 0.41 & 0.81 m tenagra ii + dec 11.5366 & 94.1 min & b & @xmath319.8 & 0.81 m tenagra ii + dec 11.5479 & 1.84 hr & i@xmath11 & 19.99 @xmath209 0.24 & 0.81 m tenagra ii + dec 11.5525 & 1.95 hr & r@xmath11 & 20.74 @xmath209 0.42 & 0.81
m tenagra ii + dec 11.5566 & 2.05 hr & r@xmath11 & 20.70 @xmath209 0.16 & 0.65 m gunma + dec 11.9583 & 11.7 hr & r@xmath11 & @xmath322.0 & 1.34 m tautenburg + dec 11.9744 & 12.1 hr & i@xmath11 & @xmath320.7 & 1.34 m tautenburg + dec 12.3883 & 22.0
hr & i * & 23.02 @xmath209 0.12 & 3.5 m arc + dec 13.4680 & 47.9 hr & v & 23.0 @xmath209 0.5 & 1.0 m usno + dec 28.4283 & 17.0 day & i * & 24.41 @xmath209 0.22 & 3.5 m arc + mar 23.1335 & 102 day & i * & 24.51 @xmath209 0.29 & 3.5 m arc + | we present follow - up network for gamma - ray bursts ( fun grb ) collaboration observations of the optical afterglow of grb 021211 made between 143 seconds and 102 days after the burst .
our unique data set includes the earliest filtered detections and color information for an afterglow in the pre - swift era .
we find that the afterglow is best described by ( 1 ) propagation through a wind - swept medium , ( 2 ) a cooling break that is blueward of the observed optical frequencies , and ( 3 ) a hard electron energy distribution .
however , superimposed on this `` standard model '' behavior we find one and possibly two significant chromatic variations during the first few hours after the burst .
we consider possible reasons for these variations , including the possibility that they are due to a dust echo . finally , we constrain physical parameters that describe the afterglow and surrounding medium for a variety of scenarios and find that grb 021211 s afterglow is faint for a combination of 3 4 reasons : ( 1 ) a low fraction of energy in relativistic electrons , ( 2 ) a low density for the wind - swept medium , implying either a low mass - loss rate and/or a high wind velocity for the progenitor , ( 3 ) a wide opening / viewing angle for the jet , and possibly ( 4 ) moderate source - frame extinction .
the jet appears to be significantly far from equipartition and magnetically dominated .
more extreme versions of this might explain the darkness of many afterglows in the swift era . |
in the present work we deal with the problem of the behavior of wave packets in a two dimensional square lattice under the action of magnetic and electric fields .
the magnetic field is perpendicular to the lattice , while the electric field is in the plane .
we can mention the ( pioneering ) works done on the subject of wave propagation in low dimensional systems that have attracted the interest , since the early days of quantum mechanics .
@xcite the subject of carriers in a two dimensional structure under the action of external magnetic and electric fields , has aroused intense interest since it has become recently experimentally accessible .
@xcite we have found very interesting properties of the time evolution of initial wave packets that were assumed of a gaussian structure with a given velocity .
we analyzed gaussians with different dispersions which in turn determine the type of propagation a wave packet will present .
another parameter which has a direct influence on the wave packet behavior is the assumed initial velocity .
obviously , the magnitude as well as the direction of the electric field has also a direct influence on propagation . clearly , one notices the wavy nature of the solution of the time dependent schrdinger equation but , a novel effect is observed , namely , the successive splitting and reconstruction of the wave function in two or more components as time goes .
we observe the formation of vortices due to the joint effect of the crystal potential and the external fields . to our understanding
this novel effect has not been reported until now . in order to comprehend the characteristic of propagation we resort to the study of the trajectories in reciprocal space
since they are connected with the ones in direct ( coordinate ) space by a rotation of @xmath0 .
a pioneering work dealing with the motion of an electron in a 2d lattice potential superimposed to a magnetic field is due to peierls @xcite , that considered an effective single band hamiltonian arising from a tight - binding dispersion relation . as a result of this model the single bloch band
is split into magnetic subbands according to the number of flux quanta that pierces the unit cell of the 2d lattice . at the same time , the parameter @xmath1 being the ratio between the magnetic flux through the unit cell to the quantum of flux , controls the kind of propagation of wave packets in a lattice under a uniform magnetic field . for a rational value
we recover translational symmetry , with an enlarged unit cell which in turn makes possible for a packet to propagate in the sample . on the contrary for an irrational value of @xmath2 we face the problem of incommensurability of the potential which in turn produces a localization of a wave packet in a definite region of the lattice @xcite .
recently , the electronic spectrum of a two - dimensional quantum dots array under magnetic and electric fields was presented @xcite .
the action of a magnetic field is analyzed along the peierls model @xcite , which consists in taking a dispersion relation for a square lattice @xmath3 and replacing in it the quasimomentum @xmath4 by @xmath5 to obtain a model hamiltonian . in the present work we used the symmetric gauge for the vector potential : @xmath6 the classical paper of hofstadter @xcite showed that such an approach leads to a spectrum as function of the magnetic field that presents a fractal structure , the so called hofstadter butterfly . on the other hand , hall measurements on gaas / algaas superlattice provided evidence of the existence of the structure of hofstadter s butterfly@xcite .
since this model is of a single band , is limited to analyze systems of large gaps and/or magnetic intensities such that no interband transitions occur .
since we want to study the kind of propagation of a particle in such a system , we expand the wave function in the wannier representation : @xmath7 where the ket @xmath8 is the ket associated with the corresponding site @xcite .
next , we assume a discrete set of coordinates such that @xmath9 , and @xmath10 .
the time dependent schrdinger equation in the wannier representation becomes the set of equations : @xmath11 where @xmath12 is the ratio between the flux through the unit cell in the @xmath13 plane to the quantum of flux @xcite @xmath14 , @xmath15 are the on - site energies and @xmath16 is the hopping term . we have used the runge kutta method of fourth order to integrate the equations of motion . in order to solve the time dependent schrdinger equation we chose as initial condition a gaussian wave packet with a certain width and a given velocity : @xmath17\ ] ] we decided for this since it is more realistic to assume that the injected electron is not extremely localized .
one difficulty faced during the calculations was to decide the right size of the lattice in order to avoid boundary effects . to be specific ,
we have taken : a lattice of @xmath18 sites , @xmath19 which corresponds to a magnetic field of intensity @xmath20 , the dimensionless units of time @xmath21 , corresponds to @xmath22 , the dimensionless units of electric field are @xmath23 , such a unit is equivalent to @xmath24 , and all lengths are in units of the lattice parameter @xmath25 .
after solving the set of equations we constructed the following : \i ) the mean square displacement @xmath26 @xmath27 \ii ) we follow the centroid of the wave packet by evaluating the following quantity , which give us the amount of the displacement form the initial position of the particle : @xmath28 @xmath29 \iii ) the participation function @xcite @xmath30^{-1 } \label{pt}\ ] ] an interesting feature of this function is that it indicates the sites that participate in the wave packet . at the same time it presents an abrupt decline once the packet reaches the boundary of the lattice , in this way we can note the presence of size effects .
we followed the anderson @xcite criterium for analyzing diffusion , namely , we can conclude that diffusion has occurred if at @xmath31 the wannier amplitudes at the starting sites go to zero .
if , these amplitudes remain finite decreasing rapidly with distance , we say we have a localized state .
we also plot the wave packet as it evolves in time which tells us the kind of propagation for the different cases in study .
more than that , by looking at the displacements of the maxima of the packet we can infer the kind of _ trajectory _ a particle would describe . besides that
, we follow the time evolution of the centroid of the wave packet which gives complementary information .
we would like at this point to describe a surprising behavior observed during the time evolution of wave packets .
first , we note that their gaussian structure means that in reciprocal space we have a certain dispersion in @xmath4 , which implies that several wave vectors will participate in the evolution of the wave function .
this plays an important role as long as we are considering the wave vector associated with the velocity of the initial wave packet , lying along and around the lines of zero energy in the brillouin zone of the square lattice . in fig .
1 we show the lines of constant energy where the arrows signal the orbit in reciprocal space described by the wave vector . as it is well described in the books of solid state@xcite , the quasimomentum satisfies an equation of motion analogous to the classical one ; where one takes the group velocity @xmath32 in the expression of the lorentz force , which in turn determines that the wave vector moves along the lines of constant energy .
first , consider @xmath4 _ inside _ the region of energy zero , but close to its boundaries , it will describe a clockwise orbit in reciprocal space . as for @xmath4 _ outside _ but close to the line , we get another trajectory described in counterclockwise sense .
this will result in the appearance of vortices rotating in opposite directions when describing the evolution of the wave packet in direct space , where the `` orbits '' are obtained after rotation by @xmath0 .
what is very interesting is to consider @xmath4 on one of the lines of zero energy .
in such a case due to the dispersion in @xmath4 because of the gaussian structure of the initial packet , we discussed above , we have to take into account @xmath4 values _
inside _ and _ outside _ the line , symmetrically distributed .
consequently , part of the wave ( half of it ) will describe a trajectory in the clockwise sense and the other half of the wave in the opposite sense , the resulting movements is similar to a swimmer doing breaststroke .
as time goes , the wave packet starts moving in the direction of the velocity and then splits into two components that will join in a single component and so on .
consequently , two vortices with opposite angular velocities are formed .
the wave remains stationary in a certain region of the lattice .
see fig . 2 where this remarkable effect is shown .
assuming now @xmath4 with components ( @xmath33 ) , as shown in fig . 1 ,
four squares in reciprocal space participate in the movement of the packet reinforcing each other such that at a certain time the wave appears like is shown in fig .
3 . to get the evolution of the wave packet one has to follow the arrows in each of the four regions .
again , we observe the splitting and reconstruction of the wave , this time in several components . in this case ,
several vortices are present .
one encounters a similar effect when considering a 1-d crystal under the action of a dc electric field .
when considering as an initial state a well localized state at a site in the system , the evolution of the packet is such that it is split into two symmetric parts , which oscillate with the bloch frequency in opposite directions@xcite .
let us assume @xmath4 inside the region limited by the lines of zero energy but close to one of them .
the wave function will split into two components in such a way that one part of the wave , the major part , will rotate clockwise while the rest will do counterclockwise . as for @xmath4 outside the region but close to one of the lines , the reciprocal is true , the major part of the wave will perform a rotation counterclockwise . to illustrate this effect , consider now the wave vector of the initial wave packet , inside the region of the lines for @xmath34 but close to one of it , we took for example @xmath35 .
as said above , the wave is split in two components , where the bigger part rotates clockwise and the smaller part rotates in the opposite sense . in this case
this case it results in the appearance of two asymmetric vortices .
we consider now the inclusion of a dc electric field in the equation of motion for the wannier amplitudes . as a general trend we observe that the packet will propagate in a more complex way , but _ always _ in direction perpendicular to the electric field , as it was shown previously@xcite .
this behavior is also present in the classical treatment of the problem@xcite . from the view point of quantum mechanics
, we understand this behavior since the electric field breaks the degeneracy of the on - site energies along the direction of the applied field , inhibiting hopping between these sites .
first we take @xmath4 on one of the lines of zero energy , for example ( @xmath36 ) and the electric field along the diagonal , @xmath37 , while the initial packet has @xmath38 .
the wave is split while propagating but , due to the presence of the electric field , one part of the wave proceeds with a greater velocity , with a centroid trajectory similar to the one a _ snake _
performs when moving around .
the other part that moves in opposite direction , remains close to the starting point .
5 where we also show the centroid trajectory , and the msd and participation as functions of time . for the same configuration but for @xmath39 , we observe that the wave is _ not _ split and follows a trajectory quite similar to the classical one , i.e. , the more so , the more extended is the initial wave packet , i.e. , the greater is sigma .
this comes about since a greater dispersion in direct space is related to a smaller one in reciprocal space .
6 . for the case
@xmath40 we confirm this behavior of the propagation of the wave .
it is interesting to mention that the trajectories for @xmath41 and @xmath42 are exactly the same . by increasing the intensity of the electric field
we obtain a displacement of the centroid along a trajectory similar to the former one , the only difference is that in the case of stronger field the amplitude of the displacement as well as the period of the oscillations are reduced .
this field effect is shown at the bottom left of fig .
a very peculiar effect is obtained by considering the electric field with components @xmath43 and taken @xmath44 . in this case
we note that the `` swimmer '' by doing breaststroke can displace itself along the perpendicular direction to the applied electric field .
the vortices displacement is such that the centroid moves along a _
line , as shown in fig.7 .
this should be compared with the case without electric field in which the packet remains stationary as shown in fig .
we show in this work a novel effect of wave packet propagation in a two dimensional crystalline system under the action of combined magnetic and electric fields . using the runge kutta method of forth order
, we integrate the equations of motion in the wannier representation , assuming as initial condition a gaussian wave packet with given velocity .
we found the very interesting behavior of the wave function , namely , that by taking the initial velocity with the associated wave vector close to the lines of zero energy , the wave is split and reconstructed as time goes , with the appearance of a series of vortices .
this effect comes about since we used a gaussian as an initial condition .
this in turn , implies that one has to take into account a dispersion in the reciprocal space , so there are contributions of the lines of constant energy on both sides of the zero energy line , as explained above . without the presence of the electric field ,
the wave remains in a definite region of the lattice .
the inclusion of the electric field , produces a displacement of the vortices along the perpendicular direction of the applied field , showing a more complex behavior since now , besides the displacement , the wave is split and reconstructed as time goes . as for the centroid trajectory , for a configuration such that the applied electron field is parallel to the initial velocity
, it describes a trajectory similar to the way a snake performs when moving around . in the classical treatment ,
the trajectory is a trochoid . for other configurations ,
i.e. , for the wave packet with initial velocity perpendicular to the field , the centroid trajectory is a strait line .
one last comment deserves to be made , the effects we described are the results of taking into account that the particle , besides being under the action of the fields , is subjected to a 2d crystal potential as well . | in the present work we deal with the dynamics of wave packets in a two - dimensional crystal under the action of magnetic and electric fields .
the magnetic field is perpendicular to the plane and the electric field is on the plane . in the simulations we considered a symmetric gauge for the vector potential while the initial wave packet was assumed to have a gaussian structure with given velocities .
the parameters that control the kind of time evolution of the packets are : the width of the gaussian , its velocity , and , the intensity and direction of the electric field as well as the magnitude of the magnetic field . in order to characterize the kind of propagation we evaluated the mean - square displacement ( msd ) , the participation function and which is more important we were able to follow the wave at different times , which allowed us to see the time evolution of the centroid of the wave packets .
a novel effect was observed , namely , the dynamics is such that the wave function _ splits _ into two or more components and _ reconstructs _ successively as time goes , vortices are forming . to our understanding
this is the first time such an effect is reported . as for the inclusion of the electric field ,
we observe a complex behavior of the wave packet as well as note that the vortices propagate in direction perpendicular to the applied electric field . a similar behavior presented by the classical treatment , in our case
we give a quantum mechanics explanation for that . |
driven by recent technological interests in magnetic recording and data storage , it has become possible to grow well - characterized thin films , multilayers , and nanostructures .
simultaneously , advances in synchrotron radiation instrumentation have made it possible to determine the electronic and magnetic properties of these systems in an element - specific way .
x - ray magnetic circular dichroism ( xmcd ) in conjunction with the sum rules enables the separation of the spin and orbital contributions to the total magnetic moments @xcite while complementary soft x - ray resonant magnetic scattering measurements can provide details about the layer dependence of the magnetic moments , periodicity of the magnetic domain structures , and roughness of the magnetic layers @xcite . developments of theoretical models which treat these magnetic systems have also been quite successful .
it has been known already for some time that the spin and orbital magnetic moments at surfaces are often enhanced compared to their bulk values due to symmetry breaking and @xmath2 band narrowing at the surface @xcite .
electron hybridization at interfaces can give rise to charge transfer across the interface @xcite resulting in a change of the density of states ( dos ) near the fermi level .
this can enhance the spin and orbital magnetic moments in magnetic layers and can induce magnetization in adjacent `` non - magnetic '' layers . in giant magnetoresistance ( gmr ) materials , such as co / cu multilayers , the magnetic layers are ( anti)ferromagnetically coupled by the ruderman - kittel - kasuya - yosida ( rkky ) interaction , which is accompanied by an induced oscillatory magnetic moment in the `` non - magnetic '' spacer layer .
the magnetic properties of two metals near their common interface are essentially determined by the differences in electronegativity .
similarly to the model of friedel @xcite the excess nuclear charge displaces locally the mobile electrons , i.e. the electrons close to the fermi level , until the displaced charges totally screen out the nuclear charge .
this effect occurs at the interface over distances of the order of a few interatomic distances , i.e. the screening is strongly localized . in fe with only few spin - up electrons close to @xmath3 ,
the screening is almost uniquely due to spin - down @xmath2 electrons .
most of the attention so far has been focused on the magnetism in @xmath4 materials , although the possibility to observe magnetism in @xmath1 and @xmath0 elements in a surface geometry has been noticed early on @xcite .
impurity systems have been thoroughly studied by relativistic calculations and in particular the systematics of @xmath5 , @xmath6 , and @xmath7 impurities in a magnetic host have been established @xcite .
experimentally , sizeable spin and orbital magnetic moments in @xmath0 transition metals due to the hybridization with a magnetic @xmath5 element were found in e.g. fepd @xcite and copt @xcite alloys , ni / pt multilayers @xcite , and ir impurities @xcite . also fe mono- and bi - layers on w(110 ) received a considerable theoretical interest @xcite .
recently , the spin and orbital moments for w / fe , ir / fe multilayers were reported in ref .
@xcite . using atomistic arguments the authors concluded solely from the measurements of these two @xmath7 elements that the systematic behavior across the series was different than for @xmath7 impurities .
our systematic study across the entire @xmath7 series shows that in the vicinity of w and ir a sign change of the orbital moment occurs .
therefore , it should come as no surprise that the orbital moment of w can be positive or negative at the interface or for the impurity , respectively . by contrasting the results of @xmath7 impurities with @xmath7 interfaces in fe
, we will be able to establish the role of band filling versus the importance of geometry or structure of both the impurity and interface .
this is an important issue since the work in ref .
@xcite implied that the induced magnetic behavior of @xmath7 layers may be radically different than that of impurities and alloys .
if this were to be true then we would expect the magnetic properties to be determined more or less by structural considerations only and not to see any commonality between impurity and interface as a function of atomic number z. in the case of an w impurity in fe , the magnetic properties of the former are determined by @xmath0-@xmath4 hybridization , whilst at a w interface in fe this @xmath0-@xmath4 hybridization is reduced by the in - plane @xmath0-@xmath0 hybridization .
a question arises as to whether this @xmath7-@xmath5 hybridization is sufficiently reduced to alter , drastically enough , the magnetic properties of w interfaces in fe in comparison to w impurities in fe to give rise to the proposed radical new behavior as suggested in ref . @xcite . in the present study ,
the systematics of @xmath7 and @xmath6 interfaces with a fe substrate will be considered .
the relative alignment of the spin and orbital moments will be compared with the predictions based on hund s third rule .
this rule states that spin and orbital moment should be parallel ( antiparallel ) for more ( less ) than half - filled shell , c.f .
[ hund ] where the atomic values of these quantities are shown .
although strictly valid only for single atoms , this rule seems to be applicable also to solids with only a few exceptions , such as u metal @xcite and vanadium in vau@xmath8 @xcite and vpt@xmath9 @xcite .
the violation reported in fe / w multilayers
@xcite seems surprising , since w as an impurity and in alloys does not show this violation @xcite .
the opposite situation arises for ir that obeys the third hund s rule in fe / ir multilayers @xcite but violates it as a 5d impurity in fe @xcite .
therefore , in this study we will compare the electronic and magnetic properties of a @xmath7 transition metal with those of an impurity .
calculations were performed within the framework of relativistic local spin density ( lsd ) band theory , where analogous to the non - relativistic case , a set of coupled kohn - sham - dirac equations is derived which describes the ground state of a relativistic many - electron system @xcite with the corresponding hamiltonian of the form @xmath10 + { \boldsymbol{\beta}}\sigma_z b^{\text{eff } } [ n{({\bf r)}},m{({\bf r ) } } ] \;\;,\end{aligned}\ ] ] @xmath11 & = & v^{\text{ext } } { ( { \bf r)}}+ \frac{\delta e^{\text{xc } } [ n{({\bf r)}},m{({\bf r ) } } ] } { \delta n{({\bf r ) } } } \nonumber \\ & & + \ ; e^2 \int \frac{n{({\bf r^{\prime})}}}{| { { \bf r}}- { { \bf r}^{\prime}}| } d{{\bf r}^{\prime}}\;\ ; , \\ b^{\text{eff } } [ n{({\bf r)}},m{({\bf r ) } } ] & = & \frac{eh}{2mc } \frac{\delta e^{\text{xc } } [ n{({\bf r)}},m{({\bf r ) } } ] } { \delta m{({\bf r ) } } } \;\ ; , \\ n{({\bf r)}}&= & \sum \psi^+_i { ( { \bf r)}}\psi_i { ( { \bf r)}}\;\ ; , \\ m{({\bf r)}}&= & \sum \psi^+_i { ( { \bf r)}}{\boldsymbol{\beta}}\sigma_z \psi_i { ( { \bf r)}}\;\;,\end{aligned}\ ] ] where @xmath12 is a four - component one - electron dirac spinor .
the matrices @xmath13 and @xmath14 are the standard dirac matrices while @xmath15 is the vector of the 4@xmath164 pauli matrix .
the spin - polarized relativistic lmto method @xcite was used to calculate the spin and orbital magnetic moments .
the hamiltonian and overlap matrices corresponding to eq .
( [ eq : relh ] ) are expressed in terms of the basis set of muffin - tin orbitals ( mto ) which are constructed from the solutions , @xmath17 , to the single - site dirac equation , given by @xmath18 where @xmath19 and @xmath20 are the radial functions and the spin - angular function is @xmath21 here @xmath22 is a clebsch - gordan coefficient and @xmath23 a spherical harmonic . the values for the spin and orbital magnetic moments are obtained using @xmath24 where @xmath25 and e@xmath26 are the eigenvector and eigenvalue , respectively , of the @xmath27-th band at the point @xmath28 of the brillouin zone .
self - consistent relativistic calculations were performed for the electronic and magnetic properties of periodic multilayers consisting of @xmath29fe(001)/@xmath30(@xmath7 ) where @xmath7 can be ta , w , re , os , ir , pt , or au . the layers were constructed to be consistent with the bcc structure .
the layer thickness @xmath29 of fe took on the values @xmath29=3 and 5 while the layer thickness @xmath30 of @xmath7 has the values @xmath30=1 , 3 , 5 , and 7 .
basis sets with @xmath31 were used for both fe and @xmath7 .
convergence tests were performed to determine the influence of the inclusion of the higher angular momentum basis states , @xmath32 , and the values of the magnetic moments were found to be well converged with the inclusion of the @xmath32 states changing the values of the spin and orbital moments typically by less than 0.01 and 0.003 @xmath33 , respectively .
the planar lattice constant of bcc fe was used for both the fe and the @xmath7 layers .
perpendicular to the planes the lattice parameter of fe was used for the fe layers , while for @xmath7 the lattice parameter perpendicular to the planes was determined such that the volume around @xmath7 was equal to the experimental volume for the elemental @xmath7 system .
this procedure was used to model the relaxation of the lattice at the interface .
in fig . [ lsthickness ] the variations of the spin and orbital moments are shown as a function of w thickness for a fixed fe thickness of 5 layers .
the numerical values can be found in table [ qtab1 ] .
note that at the fe side of the fe / w interface the spin moment is decreased in comparison with the other fe layers , whilst the orbital moment shows a strong enhancement in the interface layer . on the w side of the interface , a spin moment on w
is induced by the fe which is aligned antiparallel .
this w moment shows an oscillatory behavior , becoming aligned to the fe spin moment in the two subsequent w layers and antiparallel again in the fourth layer .
the w orbital moments are initially aligned antiparallel to the fe orbital moment and become parallel from the third w layer onwards .
from these results it can be seen that in the w interface layer the spin and orbital magnetic moments are parallel aligned demonstrating that hund s third rule is violated as reported by the experimental work of ref . .
this can also be seen in table [ qtab1 ] where for 5fe/7w the w1 interface and w3 layers show a violation of hund s third rule whereas layers w2 and w4 have spin and orbital magnetic moment anti - parallel aligned .
table [ qtab1 ] also relates the trends in spin and orbital magnetic moments to charge transfer .
the charge transfer at the interface can be sizeable for all systems considered : the 5fe / w , 5fe/3w , 5fe/5w and 5fe/7w systems . from the distribution of the fe and w electrons into d and sp channels , we see that for fe the spin magnetic moment contribution from the sp electrons is antiparallel to the contribution from the d electrons .
this is what one would expect from fe .
however , for w , these contributions are parallel and are in accordance with a magnetic moment induced by hybridization with fe . actually , the contribution of the sp electrons is as large as @xmath34 of the total spin moment .
one can also see that the induced spin magnetic moment is largest on the w1 interface layer where the hybridization with fe is most sizeable .
the oscillations of the spin moments in the w layers are further analysed in fig .
[ centroid ] where the centroid positions of the spin - up ( majority ) and spin - down ( minority ) @xmath2 bands are presented for each of the layers of 5fe/3w .
the reduction of the fe spin moment at the interface can be seen to be caused by the centroid position of the minority spin band lowering its energy . for the w layers the centroid positions of the spin bands lie above the fermi energy ( @xmath3 ) and a small splitting of these
is induced .
this splitting between the spin - up and spin - down bands reverses sign from the w interface layer to the next w layer .
the sign of the spin moment in the w interface layer being aligned antiparallel to fe is a consequence of the fe - w hybridization across the interface . the minority fe @xmath2 bands lie energetically close to the minority w bands and a relatively pronouned hybridization between these bands develops .
this pulls the w minority @xmath2 bands down in energy .
in contrast the majority fe and w @xmath2 bands are energetically well separated and substantially less hybridization occurs .
consequently , the w minority d bands are pulled down in energy .
therefore these w minority d bands will be preferentially occupied and this allows a w spin magnetic moment , opposite in sign to that of fe , to develop .
figure [ dos ] further demonstrates the hybridization of fe and w at the interface .
it shows the layer decomposed dos for the 5fe/5w system .
whilst in fe and w layers away from the interface the dos resembles that of bulk fe and w , respectively , a substantial hybridization at the interface is noted . in particular
the minority fe @xmath2 band loses structure and its center of mass is shifted to lower energy .
the majority and minority fe @xmath2 bands in the energy region of @xmath3 and 0.2 ry below @xmath3 hybridise strongly with the w @xmath2 bands , creating in the spin - resolved w @xmath2 bands a similar depletion and increase as in the fe dos .
this hybridization rapidly disappears in subsequent w layers .
these results are in contrast with the findings of ref . where for gd / w multilayers no hybridization - induced spin polarization in the w layer is reported and spin polarization is only found to occur as a consequence of expanding the volume around the w sites sufficiently to induce a volume prompted spin polarization . from the present study
we see that hybridization of @xmath2 electrons at the interface between fe , which has a substantial spin moment of 2.2 @xmath33 , and w induces the w spin polarization . in the case of gd
the spin polarization of the @xmath2 states is small ( 0.5 to 0.6 @xmath33 ) and this seems to be insufficient to induce spin polarization via hybridization .
this argumentation was checked by performing calculations for ni / w multilayers where ni has roughly the same spin polarization as the gd @xmath2 band .
additional differences between the calculations are that fe / w , gd / w , and ni / w multilayers have bcc , hcp , and fcc superstructures , respectively . our calculations found that ni induced a much smaller spin moment of 0.005 @xmath33 in the w layers .
the different crystal structures are presumably responsible for inducing a spin moment for w on ni but not for w on gd . whilst we have established conclusively that the spin magnetic moment on w is induced through hybridization , the oscillatory behaviour of the w spin magnetic moment needs more careful consideration . in table
i we compare charge - transfer modulations with the oscillation in the spin magnetic moment layer by layer .
the charge - transfer modulation varies as @xmath35 whilst the spin magnetic moment varies as @xmath36 and the orbital magnetic moment as @xmath37 .
this means that hund s third rule is violated in layers 1 and 3 , whilst not in layers 2 and 4 .
this is a striking result emphasizing an oscillatory behaviour in the violation of hund s third rule .
actually , this demonstrates that it is not very meaningful to discuss the behaviour of the spin and orbital magnetic moments in terms of hund s third rule .
the striking feature of this work is the sign of the induced orbital magnetic moment in the w interface layer . to determine the robustness of this feature we plot in fig .
[ filling ] the spin and orbital moments as a function of energy for a 5fe(001)/3w multilayer . at the energy position of @xmath3 ( @xmath38 ) the proper value of the spin and orbital moments in the interface w layer
are @xmath390.113 and @xmath390.031 , respectively .
we see that for `` bulk '' fe ( fe@xmath40 ) the spin and orbital moments are aligned , the spin moment is positive for all energies and the orbital moment changes sign around @xmath41 ry .
this would imply that if we reduce the number of electrons in the system and in this manner model the lower z systems , i.e.moving towards and lower than half filling , the spin and orbital moment would become antiparallel aligned , consistent with hund s third rule .
all fe layers possess this behavior . in the w layers
the spin moments as a function of band filling show an oscillatory behavior , very different from what was noticed in fe .
moreover this oscillatory behavior is very different between the w interface layer and the next w layer .
the orbital moments also oscillates as a function of band filling and the simple cross - over from negative to positive values as seen in fe and shown to be consistent with hund s third rule , does not occur .
hence , the behavior can be seen to be a consequence of hybridization between the @xmath7 and fe and has no resemblance to hund s third rule behavior .
the spin and orbital moments as well as the charge transfer for 5fe(001)/3(@xmath7 ) , where @xmath7 is ta , w , or re , is displayed in fig .
the numerical values are given in table [ qtab2 ] .
judging from the reduction of the fe spin moment at the interface , the hybridization of ta with fe is strongest . apart from this ,
the behavior of the spin moments as a function of layers is similar for ta , w , and re .
the figure shows that the oscillations in the charge transfer do follow the oscillations in the spin moments .
however , concerning the orbital moments their behavior varies from ta to w to re . for the 5fe/3ta
there is no sign change in the multilayer at all . in the case of 5fe/3re
the oscillation of the orbital moment does follow the oscillations in the spin moments . for this system
hund s rule is not violated either : in both re layers the spin and orbital moments are aligned .
it is remarkable to see how complicated the behavior of the orbital moment is as a function of atomic number and layer index , whilst we note again the correlation between oscillations in the spin magnetic moment and the charge oscillations .
this complicated behavior of the orbital moments in ta , w , and re and also as a function of layer index should not come anymore as a surprise given the complex behavior seen in fig .
[ filling ] as a function of energy or band filling .
in particular by considering the energy scales in w1 and w2 of fig . [ filling ] marked by a dashed box , we can see that for a reduced energy at the lower end of the box the orbital magnetic moment would become positive in both w1 and w2 . in a rigid band fashion
this would correspond to ta . increasing the energy towards the upper end of the box
one can see that the orbital magnetic moment can remain negative in w1 but can become positive in w2 .
the system that we consider is a three - layer system incorporated in an fe host .
this three - layer system is repeated periodically with the structure 5fe(001)/3(@xmath7 ) .
whilst such a geometry is far removed from a @xmath7 impurity in an fe host , comparison with calculations for such systems could be instructive .
figure [ 5dseries ] shows the spin and orbital moments obtained for the 5fe/3(@xmath7 ) multilayer systems and the @xmath7 impurity calculations from ref . .
the spin moments of @xmath7 impurities in fe are antiparallel aligned for the first part of the @xmath0 transition metal series , i.e. for the @xmath7 with less than half filling . from os
onwards the spin moments of the @xmath7 impurity and fe become aligned .
this is for the same reasons as discussed in sec .
[ sec : a ] .
for @xmath7 impurities to obey hund s third rule the orbital moments would have to be positive throughout this @xmath7 series ( c.f .
[ hund ] for the atomic case ) .
this is not the case and the orbital moments are negative for re , os , ir ; so that in these calculations hund s rule is violated for os and ir .
the spin and orbital moment curves for the multilayers show the same behavior as for the impurity . for the spin moment
the multilayer and the impurity calculation switch between parallel and antiparallel alignment at the same location .
the orbital moments which are more sensitive to the structure and chemical environment switch between parallel and antiparallel at slightly different locations , making the orbital moments negative for w , re , and os . therefore , hund s rule is not only violated for w but also for os as was also the case for the single impurity .
figure [ filling ] reinforces this universal band filling scenario . when analyzing the orbital moment in the w interface layer ( w@xmath42 ) as a function of band filling in the vicinity of @xmath3 ( the area in fig .
[ filling ] marked by the box ) we see that moving away from @xmath3 to lower energy the orbital moment changes sign .
moving from @xmath3 to lower energy in fig . [ filling ] means , in a rigid - band manner , mimicking lower z such as ta and therefore modelling the z lower than w in fig . [ 5dseries ] . moving from @xmath3 up in energy in fig . [ filling ]
also models the behavior of the orbital moment in fig .
[ 5dseries ] , but now for z larger than w. therefore , we see that the behavior of the orbital moment as a function of z as shown in fig .
[ 5dseries ] is present in the behavior of the orbital magnetic moment of the w interface layer as a function of band filling ( boxed area in fig . [ filling ] ) .
moreover , a similar behavior can also be noted from fig . [ filling ] in the layer w@xmath43 , adjacent to the interface layer .
the difference is that the area where the orbital moment is negative has been reduced . as a consequence re , which is just after w in the periodic table and
can be modelled in a rigid band fashion by moving up @xmath3 , has a negative orbital moment in the interface layer but a positive orbital moment in the adjacent w@xmath43 layer ( fig .
[ slq ] ) .
this implies that the energy has moved up in such a manner that the orbital moment in the interface remains negative but at the same time has moved up sufficiently to reach energies where the sign change in the orbital moment of the w@xmath43 occurs ( @xmath440.05 ry ) .
moving @xmath3 down by roughly the same amount ( 0.05 ry ) to model ta leads to positive orbital moments in both the interface and the adjacent layer , in agreement with the observed behavior of fig .
[ slq ] .
if the band filling is the crucial ingredient in determining the behavior of the orbital moments as a function of z for the @xmath7 in fe , then we would expect a similar behavior for the @xmath6 in fe . in fig .
[ 4dseries ] we show the induced spin and orbital magnetic moments for the @xmath6 obtained for a 5fe/3(@xmath6 ) system .
the spin moment is aligned antiparallel to fe in the first part of the transition metal series and becomes parallel aligned from rh onwards .
again this is a consequence of hybridization as explained in sec .
[ sec : a ] and therefore the @xmath6 shows the same behavior as a function of band filling as the @xmath7 .
also the orbital magnetic moment shows the same trends in the @xmath6 as in the @xmath7 , namely from parallel to antiparallel and again to parallel with respect to the fe orbital moment with zero crossings at mo and ru .
these @xmath1 elements are isoelectronic with w and os , and the zero crossings in the @xmath7 occur between ta and w and between os and ir .
this shows that the zero crossings are only slightly different in the @xmath6 and @xmath7 .
in particular , the loci of zero orbital magnetic moment are shifted and the region of negative orbital magnetic moment is larger for the @xmath0 series than for the @xmath1 series .
the loci of zero orbital magnetic moment for the @xmath6 impurity occur at slightly higher values that for the @xmath6 interface .
this trend is consistent with the difference between impurity and interface of the @xmath7 series as seen in fig .
[ 5dseries ] .
this indicates that the geometrical differences between impurity and interface alter in a consistent way the loci of zero orbital magnetic moment to higher z in the impurity for both the @xmath6 and @xmath7 series . for the @xmath7 impurity the zero crossings are between w and re , and between ir and pt .
the similar behavior of these orbital moments as a function of z with only small changes between the @xmath7 impurity and interface , and between the @xmath6 impurity and interface demonstrates that band filling determines these properties .
we have presented a systematic study of the interfacial spin and orbital magnetic moments over the entire @xmath6 and @xmath7 series .
so far only an extrapolation from two experimental points , namely ir and w , existed @xcite . for w / fe(001 ) systems with different w layer thicknesses we found that in the fe interface layer the fe spin moment is reduced whilst its orbital moment is approximately doubled .
in the w layers a spin moment is induced which is antiparallel to that of fe in the first and fourth w layers but parallel to fe in the second and third w layers .
the w orbital moment does not follow the spin moment .
it is aligned antiparallel to fe in the first two w layers and changes sign in the third and fourth w layers .
the calculations show that small changes in the band filling can lead to a reversal of the orbital moment of the first w layer .
hence , the behavior can be seen as a consequence of the hybridization between w and fe which is mainly due to band filling , and to a lesser extent geometrical effects of either single impurity or interface . in comparison with impurity calculations we found that the trend in the spin and orbital moments over the @xmath1 and @xmath0 series is quite similar .
the number of zero crossings is the same for the interface and impurity calculations , again suggesting that the behavior is primarily due to band filling .
furthermore , for the @xmath6 interface we find the same behavior as for the @xmath7 , despite the fact that the spin - orbit interaction of the @xmath1 electrons is much smaller than for @xmath0 electrons .
.layer - resolved charge transfer ( @xmath45 ) , @xmath2 and @xmath46 occupancy of the majority and minority spin channel , and spin and orbital magnetic moments for 5fe(001)/@xmath30w multilayer systems with @xmath30=@xmath471,3,5,7@xmath48 . [ cols="^,^,^,^,^,^,^,^ " , ] v. popescu , h. ebert , b. nonas , and p. h. dederichs , phys .
b * 64 * , 184407 ( 2001 ) .
p. kamp , a. marty , b. gilles , r. hoffmann , s. marchesini , m. belakhovsky , c. boeglin , h. a. drr , s. s. dhesi , g. van der laan , and a. rogalev , phys .
b * 59 * , 1105 ( 1999 ) .
w. grange , i. galanakis , m. alouani , m. maret , j. p. kappler , and a. rogalev , phys .
b * 62 * , 1157 ( 2000 ) .
f. wilhelm _
et al . _ ,
lett . * 85 * , 413 ( 2000 ) .
v. v. krishnamurthy , m. suzuki , n. kawarmura , t. ishikawa , and y. kohori , physica b * 312 * , 647 ( 2002 ) .
x. qian and w. hbner , phys .
b * 60 * , 16192 ( 1999 ) .
x. qian and w. hbner , phys .
b * 64 * , 092402 ( 2001 ) .
i. galanakis , m. alouani , and h. dreyss , phys .
b * 62 * , 3923 ( 2000 ) .
i. galanakis , a. debernardi , m. alouani , and h. dreyss , surf .
* 482 * -*485 * , 1030 ( 2001 ) .
f. wilhelm , p. poulopoulos , h. wende , a. scherz , k. baberschke , m. angelakeris , n. k. flevaris , and a. rogalev , phys .
lett . * 87 * , 207202 ( 2001 ) . | results of _ ab initio _ calculations using the relativistic local spin density theory are presented for the magnetic moments of periodic @xmath0 and @xmath1 transition metal interfaces with bcc fe(001 ) . in this systematic study we calculated the layer - resolved spin and orbital magnetic moments over the entire series . for the fe / w(001 )
system , the fe spin moment is reduced whilst its orbital moment is strongly enhanced . in the w layers a spin moment is induced , which is antiparallel to that of fe in the first and fourth w layers but parallel to fe in the second and third w layers .
the w orbital moment does not follow the spin moment .
it is aligned antiparallel to fe in the first two w layers and changes sign in the third and fourth w layers .
therefore , hund s third rule is violated in the first and third w layers , but not in the second and fourth w layers .
the trend in the spin and orbital moments over the @xmath1 and @xmath0 series for multilayers is quite similar to previous impurity calculations .
these observations strongly suggest that these effects can be seen as a consequence of the hybridization between @xmath0 ( @xmath1 ) and fe which is mostly due to band filling , and to a lesser extent geometrical effects of either single impurity or interface . |
complex networks , evolved from the erds - rnyi random graph @xcite , are powerful models for describing many complex systems in biology , sociology , and technology @xcite . in the past decade , the explosion of the general interest in the structure and the evolution of most real - world networks is mainly reflected in two striking characteristics .
one is the small - world property @xcite , which suggests that a network has a highly degree of clustering like regular networks and a small average distance among any two nodes similar to random networks .
the small - world phenomenon has been successfully described by network models with some degree of randomness @xcite .
the other is the scale - free behavior @xcite , which means a power - law distribution of connectivity , @xmath0 , where @xmath1 is the probability that a node in the network has @xmath2 connections to other nodes and @xmath3 is a positive real number determined by the given network . the origin of the scale - free behavior has been traced back to two mechanisms that are observed in many systems , growing and preferential attachment @xcite .
recently , with the progress of research in networks , many other statistical characteristics of networks appeared on the stage .
of particular renown is the so - called community(or modularity ) .
that is to say , a network is composed of many clusters of nodes , where the nodes in the same cluster are highly connected , while there are few links among the nodes belonging to different clusters .
for instance , groups are formed in scientific collaboration networks @xcite .
also , it has been found that dynamical processes on networks are affected by community structures , such as tendencies spread well within communities @xcite and diffusion between different communities is slow @xcite . in the study of community networks ,
most research has been directed in two distinct directions . on the one hand ,
attention has been paid to designing algorithms for detecting community structures in real networks .
a pioneering method was made by girvan and newman @xcite , who introduced a quantitative measure for the quality of a partition of a network into communities .
later , a number of algorithms have been proposed in order to find a good optimization with the least computational cost .
the fastest available procedures use greedy techniques @xcite and extremal optimization @xcite , which are capable of detecting communities in large networks . on the other hand ,
research has focused on modeling of networks with community structures . in ref .
@xcite , a static social network was introduced where individuals belong to groups that in turn belong to groups of groups and so on . in ref .
@xcite , a networked seceder model was suggested to illustrate group formation in social networks . in ref .
@xcite , a growing bipartite network for social communities with group structures was proposed .
each of those models is constructed based on one aspect of reality . in this paper
, we introduce a network model with communities that gives a realistic description of local events @xcite .
the model incorporates three processes , the addition of new nodes intra - community and new links intra- or inter - community . using growing and preferential attachment mechanisms
, we generate the community network with a good right - skewed distribution of nodes degrees , which has been observed in many social systems .
the barabsi - albert network @xcite only describes a particular type of evolving networks , the addition of new nodes preferential connecting to the nodes already present in the network .
systems in the real world , however , are much richer . for example , in scientific collaboration networks , a multidisciplinary scientist is not only collaborate with scientists in his research fields but also has a stronger desire to collaborate with scientists in other fields . in friendship networks
, a person usually makes friends with people belonging to different communities besides the community he belongs to . to give a realistic description of the network construction like that ,
we introduce a growing model of community networks based on local events , the addition of new nodes intra - community and new links intra- or inter - community .
the proposed model is defined as follows .
we start with @xmath4 ( @xmath5 ) isolated communities and each community consists of a small number @xmath6 of isolated nodes . at each time step , we perform one of the following three operations
. \(i ) with probability @xmath7 we add a new node in a randomly chosen community . here
the randomly chosen means that the community is selected according to the uniform distribution .
the new node is only connected to one node that already present in the selected community .
we denote it as the @xmath8th commnuity .
the probability that node @xmath9 in community @xmath8 will be selected is proportional to its intra - community degree @xmath10 where the sum runs over nodes in community @xmath8 and @xmath11 is the intra - community degree of node @xmath9 in community @xmath8 .
\(ii ) with probability @xmath12 we add a new link in a randomly chosen community . for this
we randomly select a node in a randomly chosen community @xmath8 as the starting point of the new link .
the other end of the link is selected in the same community with the probability given by eq .
( [ nodeprob ] ) .
\(iii ) with probability @xmath13 @xmath14 we add a new link between two communities . for this
we randomly select a node in a randomly chosen community @xmath8 as the starting point of the new link .
the other end @xmath9 of the link selected in the other community @xmath15 is proportional to its inter - community degree @xmath16 where the sum runs over nodes in all communities except for community @xmath8 and @xmath17 is the inter - community degree of node @xmath9 in community @xmath15 .
after @xmath18 time steps , this scheme generates a network of @xmath19 nodes and @xmath18 links .
the parameters @xmath7 , @xmath12 , and @xmath13 control the network structure .
in the case of small @xmath13 , the generated network will have a strong community structure .
notice that whatever process is chosen in the network growth , only one link is added to the system at each time step ( duplicate and self - connected edges are forbidden ) , however , this is not essential .
we choose link probabilities @xmath20 and @xmath21 to be proportional to @xmath22 and @xmath23 , respectively , such that there is a nonzero probability of isolated nodes acquiring new links .
in our community network , the degree of a node consists of two parts , the intra - community degree and the inter - community degree .
increase in the node s connectivity can be divided into two processes , the increases of the intra - community degree and the inter - community degree . in each process , we assume that @xmath24 and @xmath25 change continuously , and the probabilities @xmath26 and @xmath27 can be interpreted as the rates at which @xmath24 and @xmath25 change , respectively . thus , the operations ( i)-(iii ) all contribute to @xmath28 , each being incorporated in the continuum theory as follows .
\(i ) addition of a new node in a randomly chosen community with probability @xmath7 : @xmath29 \(ii ) addition of a new link in a randomly chosen community with probability @xmath12 : @xmath30,\ ] ] where @xmath31 is the number of total nodes .
the first term on the right - hand side ( rhs ) corresponds to the random selection of one end of the new link , while the second term on the rhs reflects the preferential attachment ( eq .
( [ nodeprob ] ) ) used to select the other end of the link .
\(iii ) addition of a new links between two communities with probability @xmath13 : @xmath32.\ ] ] the first term on the rhs represents the random selection of one end of the new link , while the second term on the rhs considers the preferential attachment ( eq .
( [ linkprob ] ) ) used to select the other end of the link in the other community . combing the contribution of above processes
, we have @xmath33 with @xmath34 we can simplify eqs .
( [ evolve1 ] ) and ( [ evolve2 ] ) for large @xmath18 @xmath35 the boundary conditions of the intra - community degree and the inter - community degree at initial time @xmath36 can be estimated in the sense of mathematical expectations , @xmath37 and @xmath38 , respectively .
so we write the solutions of eqs .
( [ eqintra ] ) and ( [ eqinter ] ) @xmath39 in random networks , the degree distribution can be calculated by @xmath40 which gives @xmath41^{-(3+\frac{p}{p+q } ) } , \label{degreeintra}\\ p(k^{\text{inter } } ) & = & \frac{2p-2pq - p^2}{2-p-4q-2p^2 + 2q^2 + 2p^2q+pq^2+p^3 } \notag\\ & \times & \left[\frac{2 - 2q+pk^{\text{inter}}}{2+p-2q - pq - p^2}\right]^{-(3+\frac{p}{1-p - q})}. \label{degreeinter}\end{aligned}\ ] ] thus , the degree distribution of our network obeys a generalized power - law form @xmath42^{-\gamma(p , q)}. \label{distribution}\ ] ] ) and ( [ degreeinter ] ) , respectively .
the solid line in ( c ) is guide to the eye with power - law decay exponent @xmath43 .
the experiment network has a total number of nodes @xmath44 with parameters @xmath45 , @xmath46 , @xmath47 , and @xmath48 , respectively.,width=529 ] , @xmath49 , @xmath47 , and @xmath48 , respectively.,width=377 ] in fig .
[ fig1 ] we present numerical results of distributions of the intra - community degree , the inter - community degree , and the total degree of nodes in log - log scale .
the experimental network is generated by the proposed scheme with @xmath44 , @xmath45 , @xmath46 , @xmath47 , and @xmath48 , respectively .
the distributions of the intra - community degree and the inter - community degree , shown in figs . [ fig1](a ) and [ fig1](b ) , agree with analytical results of eqs .
( [ degreeintra ] ) and ( [ degreeinter ] ) , respectively .
the small deviations between computer simulations and analytical solutions at both ends of the distributions appears to be the mathematical approximation of the boundary conditions and the finite size effect due to the relatively small network sizes used in the simulations . according to the evolving rule of our network ,
nodes with larger intra- ( or inter- ) degree have higher probability to gain new links , then the usual degree preferential attachment is reasonably kept .
this means that the right - skewed character of the network , such as the node s total degree , will retain . as shown in fig .
[ fig1](c ) , the total degree distribution of nodes is well expected showing a good right - skewed character , which is reasonably in agreement with the condition of many realistic systems @xcite . to illustrate the predictive power
, we also compare the numerical result of our network with the statistics of an econophysics collaboration network . in the econophysics collaboration network , each node represents one scientist .
if two scientists have collaborated one or more papers , they would be connected by an edge .
_ took the largest connected component of this network , which includes @xmath50 nodes and @xmath51 edges , and provided the best division , i.e. , @xmath45 @xcite . in fig .
[ fig2 ] we plot the degree distribution of econophysicists of the econophysics collaboration network which is fitted by computer simulations of our network starting with @xmath52 communities . to gain @xmath7 and @xmath12 , we fit the connectivity distribution @xmath1 obtained from this collaboration network with eq .
( [ distribution ] ) , obtaining a good overlap for @xmath53 and @xmath54 ( fig .
[ fig2 ] ) .
networks with community structures underlie many natural and artificial systems .
it is becoming essential to model and study this kind topological feature .
we presented a simplified mechanism for networks organized in communities , which corresponds to local events during the system s growth . the generated network is highly clustered and has a good rightskewed distribution of connectivity , which have been found very common in most realistic systems .
the present paper only suggests a simple way for generating community networks .
the shape of the resulting network is deterministic in some extent .
it is more interesting to model the evolution of communities , especially the self organization ( or emergence ) of communities in the natural world @xcite , e.g. , expansion and shrinkage , which is left to future work .
the authors acknowledge financial support from nsfc/10805033 , socialnets/217141 , stcsm/08zr1408000 , ptdc / fis/71551/2006 , and fct / sfrh / bpd/30425/2006 . | the study of community networks has attracted considerable attention recently . in this paper , we propose an evolving community network model based on local processes , the addition of new nodes intra - community and new links intra- or inter - community . employing growth and preferential attachment mechanisms , we generate networks with a generalized power - law distribution of nodes degrees . ,
, complex networks ; community networks |
one of the phenomenological nucleon - nucleon interactions that have been extensively used in transport models to study heavy - ion collisions is the isospin- and momentum - dependent mdi interaction @xcite . by using this interaction in the isospin - dependent boltzmann - uhling - uhlenbeck ( ibuu ) transport model , an extensive amount of works
have been carried out to study various isospin sensitive observables in intermediate - energy heavy - ion collisions ( for recent reviews see refs .
@xcite ) . from comparisons of the isospin diffusion data from nscl - msu for @xmath0sn+@xmath1sn reactions at @xmath2 mev / nucleon @xcite with results from the ibuu model ,
a relatively stringent constraint on the density dependence of the nuclear symmetry energy at subsaturation densities has been obtained @xcite .
the resulting symmetry energy has further been used to impose constraints on both the parameters in the skyrme effective interactions and the neutron skin thickness of heavy nuclei @xcite . the mdi interaction with the constrained isospin dependence
has also been used to study the properties of hot asymmetric nuclear matter @xcite as well as those of neutron stars @xcite , including the transition density which separates their liquid core from their inner crust @xcite .
new constraints on the masses and radii of neutron stars were then obtained from comparing the resulting crustal fraction of the moment of inertia of neutron stars with that of the vela pulsar extracted from its glitches @xcite .
although the momentum - dependent nuclear mean - field potential or the energy density functional of uniform nuclear matter was used in the above mentioned applications of the mdi interaction , the explicit form of the mdi interaction has never been given in the literatures .
it was , however , implicitly mentioned @xcite that this momentum - dependent nucleon mean - field potential was constructed according to a gogny - like interaction @xcite consisting of a zero - range skyrme - like interaction @xcite and a finite - range yukawa interaction as in the mdyi interaction @xcite .
similar to the gogny interaction , the momentum dependence in the nucleon mean - field potential from the mdi interaction is then from the contribution of the finite - range yukawa interaction to the exchange energy density of uniform nuclear matter .
the effect of the finite - range part of the mdi interaction in nonuniform nuclear matter was in previous applications either neglected or inconsistently included by using the average density - gradient terms from the phenomenological skyrme interactions . from comparing the potential energy density functional obtained from above
assumed interaction with that used in previous applications of the mdi interaction , we are able to find unique relations among the parameters of this interaction and those used in the potential energy density functional for uniform nuclear matter . as a first step to facility the consistent application of the finite - range mdi interaction , it is of interest to derive an effective zero - range interaction with density - dependent parameters using the density matrix expansion first introduced in ref .
@xcite . as shown in ref .
@xcite for other finite - range interactions , corresponding zero - range interactions derived from the density matrix expansion reproduce reasonably well in the self - consistent hartree - fock approach the binding energies and radii of nuclei obtained with the original interactions , particularly if the density matrix expansion is only used for the exchange energy from the finite - range interaction @xcite . in the present paper
, we carry out such a study as an attempt to include consistently the effect due to the finite - range part of the mdi interaction . as an application
, we use the resulting interaction to study the transition density and pressure in neutron stars and compare the results with those from previous studies using the mdi interaction but with inconsistent density - gradient coefficients @xcite .
this paper is organized as follows . in sec .
[ model ] , we review the isospin- and momentum - dependent mdi interaction and determine its underlying nucleon - nucleon ( nn ) interaction by fitting the energy density functional used in previous applications of the mdi interaction . in sec .
[ dme ] , the density matrix expansion is then used to derive from the finite - range exchange interaction a zero - range effective interaction with density - dependent parameters .
the application of the resulting interaction to study the transition density and pressure at the inner edge of neutron star crusts is given in sec .
[ app ] , using the stability conditions that are derived from the linearized vlasov equation for the neutron star matter . finally , the summary is given in sec .
[ sum ] . for details on the determination of the underlying nucleon - nucleon interaction potential in the mdi interaction and the application of the density matrix expansion to the exchange energy of nuclear matter , they are given in appendices [ appa ] and [ appb ] , respectively .
in applications of the momentum - dependent mdi interaction for studying isospin effects in intermediate - energy heavy - ion collisions as well as the properties of hot asymmetric nuclear matter and neutron star matter , one usually uses following potential energy density @xmath3for infinite nuclear matter of density @xmath4 and isospin asymmetry @xmath5 , with @xmath6 and @xmath7 being , respectively , the neutron and proton densities . in the above , @xmath8 is the nucleon isospin ; @xmath9 is the nucleon phase - space distribution function at position @xmath10 ; and @xmath11 @xmath12 is the saturation density of normal nuclear matter .
values of the parameters @xmath13 , @xmath14 , @xmath15 , @xmath16 , @xmath17 , @xmath18 and @xmath19 can be found in refs .
@xcite . for symmetric nuclear matter
, this interaction gives a binding energy of @xmath20 mev per nucleon and an incompressibility @xmath21 of @xmath22 mev at saturation density .
the parameter @xmath23 in eq .
( [ mdiv ] ) is used to model the density dependence of the symmetry energy . based on analyses of the isospin diffusion in intermediate - energy heavy - ion collisions and the neutron skin thickness of heavy nuclei @xcite ,
the slope parameter @xmath24 of the symmetry energy has been constrained to @xmath25 mev , which corresponds to @xmath26 .
more recent analyses of experimental data on isospin diffusion , double @xmath27 ratio , and neutron skin thickness favor , however , a softer symmetry energy of @xmath28 mev @xcite
. the momentum dependence in the mdi interaction can come from a number of different origins .
besides the finite - range exchange term in the nucleon - nucleon interaction , the intrinsic momentum dependence in the nucleon - nucleon interaction and the effects of short - range nucleon - nucleon correlations @xcite can also contribute to its momentum dependence .
for simplicity , we assume in the present study that all of the momentum dependence in the mdi interaction comes from the finite - range exchange term . in this case , the explicit form for the mdi interaction can be obtained from the energy density given in eq .
( [ mdiv ] ) by assuming that the interaction potential between two nucleons located at @xmath29 and @xmath30 has a form similar to the gogny interaction @xcite but with its gaussian form in the finite - range term replaced by a yukawa form , that is @xmath31 in the above equation , the first term is the density - dependent zero - range interaction which can be considered as an effective three - body interaction , whereas the second term is the density - independent finite - range interaction . in terms of this nucleon - nucleon interaction , the total potential energy of a nuclear system can be calculated from @xmath32 where @xmath33 is the quantum state of nucleon @xmath34 with the spatial state @xmath35 , the spin state @xmath36 , and the isospin state @xmath37 , and @xmath38 , @xmath39 and @xmath40 are , respectively , the space , spin and isospin exchange operators .
as shown in appendix [ appa ] , where the detailed derivation of eq .
( [ mdiv ] ) from eq .
( [ mdiv ] ) is given , the first and second terms in eq .
( [ mdiv ] ) come from the direct contribution ( the first term of eq .
( [ vv ] ) ) of the finite - range term in the nn interaction .
the third term in eq .
( [ mdiv ] ) is simply from the contribution of the zero - range term .
although different density - dependent zero - range terms can lead to different symmetry energies / potentials @xcite , we use in the present study the one in the standard skyrme interaction @xcite . the momentum - dependent terms in eq .
( [ mdiv ] ) come from the exchange contribution ( the second term of eq .
( [ vv ] ) ) of the finite - range interaction . comparing the resulting energy density functional with eq .
( [ mdiv ] ) leads to following unique relations between the 8 parameters in the nn interaction and those in the energy density functional of eq .
( [ mdiv ] ) : @xmath41 it is seen that the symmetry energy parameter @xmath23 is related to the coefficient of the spin - exchange term in the density - dependent zero - range interaction , whereas the width @xmath17 in the momentum dependence is related to the range of the finite - range interaction . in the original mdi interaction
@xcite , the symmetry energy parameter @xmath23 can only have the value of @xmath42 or @xmath43 . to get a larger range of density dependence for the symmetry energy while fixing the value of @xmath44 mev
, the parameters @xmath45 and @xmath46 have been expressed as @xcite @xmath47 which results in the @xmath23 dependence of @xmath48 .
therefore , with this constraint the @xmath23 dependence also appears in the finite - range direct interaction .
although the contribution of a finite - range interaction to the direct energy density of nuclear matter can be treated exactly , it is numerically challenging to evaluate its contribution to the exchange energy density .
the latter can be , however , approximated by that from a skyrme - like zero - range interaction using the density - matrix expansion of ref .
as shown in appendix [ appb ] , the resulting exchange energy density from the finite - range interaction at position @xmath10 in a nuclear system can be expressed in terms of densities @xmath6 and @xmath7 as well as kinetic energy densities @xmath49 and @xmath50 as @xmath51 \notag\\ & + & b[\rho_n(\vec{r}),\rho_p(\vec{r } ) ] \tau_n(\vec{r } ) + b[\rho_p(\vec{r}),\rho_n(\vec{r } ) ] \tau_p(\vec{r } ) \notag\\ & + & c[\rho_n(\vec{r}),\rho_p(\vec{r } ) ] |\nabla \rho_n(\vec{r})|^2 \notag\\ & + & c[\rho_p(\vec{r}),\rho_n(\vec{r } ) ]
|\nabla \rho_p(\vec{r})|^2 \notag\\ & + & d[\rho_n(\vec{r}),\rho_p(\vec{r } ) ] \nabla \rho_n(\vec{r } ) \cdot \nabla \rho_p(\vec{r}),\end{aligned}\ ] ] where @xmath52,\\ b(\rho_n,\rho_p ) & = & -2[\rho_n v^l(\rho_n ) + \rho_p v^u(\rho_p,\rho_n)],\\ c(\rho_n,\rho_p ) & = & -\frac{\partial f(\rho_n,\rho_p)}{\partial \rho_n},\\ d(\rho_n,\rho_p ) & = & - \frac{\partial f(\rho_n,\rho_p)}{\partial \rho_p } - \frac{\partial f(\rho_p,\rho_n)}{\partial \rho_n}.\notag\\\end{aligned}\ ] ] in above equations , @xmath53 is the potential energy density of infinite nuclear matter from the finite - range exchange interaction ( i.e. the momentum - dependent terms in eq .
( [ mdiv ] ) ) and is given by @xmath54 with @xmath55 other terms are defined as @xmath56 where @xmath57 with @xmath58 and @xmath59 being , respectively , the first- and third - order spherical bessel functions and @xmath60 is the fermi momentum . for the function @xmath61
, it is defined as @xmath62 in fig .
[ abcd ] , we show the dependence of @xmath63 , @xmath64 , @xmath65 and @xmath66 on @xmath6 and @xmath7 at subsaturation densities . it is seen that @xmath63 and @xmath66 are symmetric in @xmath6 and @xmath7 , while @xmath64 and @xmath65 are not .
furthermore , the density dependence of @xmath64 , @xmath65 and @xmath66 are strong at low densities but weak near the saturation density . as these functions are from the exchange contribution of the finite - range interaction , they are independent of @xmath23 . , @xmath64 , @xmath65 and @xmath66 on @xmath6 and @xmath7 at subsaturation densities .
, title="fig : " ] , @xmath64 , @xmath65 and @xmath66 on @xmath6 and @xmath7 at subsaturation densities . , title="fig : " ] , @xmath64 , @xmath65 and @xmath66 on @xmath6 and @xmath7 at subsaturation densities . , title="fig : " ] , @xmath64 , @xmath65 and @xmath66 on @xmath6 and @xmath7 at subsaturation densities . ,
title="fig : " ]
the explicit form of the finite - range nn interaction and the skyrme - like zero - range interaction with density - dependent parameters derived in previous sections make it convenient to use the mdi interaction for a wider range of studies . in this section
, we discuss its application in studying the transition density and pressure at the inner edge of neutron star crusts based on the stability conditions that are derived from the linearized vlasov equation for the neutron star matter .
the single - particle hamiltonian for a nucleon can be obtained from minimizing the total energy of a nuclear system with respect to its wave function . for a neutron ,
it is given by @xmath67 with the neutron effective mass @xmath68 and the potential @xmath69 where @xmath70 is the contribution from the the zero - range interaction with @xmath71\ ] ] being the corresponding energy density , and @xmath72 @xmath73 are the direct and exchange contributions from the finite - range interaction , respectively . expressing the kinetic energy density in the exchange potential in terms of the density via the extended thomas - fermi ( etf ) approximation @xcite @xmath74 where @xmath75 , @xmath76 and @xmath77 , the neutron potential can then be written as @xmath78 where @xmath79 with @xmath80 similarly , the proton single - particle hamiltonian is @xmath81 where the first term @xmath82 is due to the nuclear interaction and is obtained from eq .
( [ hn ] ) by interchanging the neutron and proton symbols ; the second term @xmath83 is the direct coulomb potential given by @xmath84 and the last term @xmath85 is the exchange coulomb potential which in the first order of the density matrix expansion @xcite is given by @xmath86 the vlasov equation has been widely used in studying the collective density fluctuation and spinodal instability in nuclear matter ( see ref .
@xcite for a recent review ) . for
a @xmath87-stable and charge neutral @xmath88 matter , the vlasov equation can be written as @xmath89 in terms of the wigner function for particle type @xmath90 @xmath91 where @xmath92 is the wave function of i - th particle of type @xmath93 . in eq .
( [ vlasov ] ) , @xmath94 denotes the particle velocity and @xmath95 . to study the density fluctuation due to a collective mode with frequency @xmath96 and wavevector @xmath97 in the nuclear matter , we follow the standard procedure @xcite by writing @xmath98 with @xmath99 as the momentum dependence of the wigner function is mainly from @xmath100 , we have @xmath101 and can thus rewrite the vlasov equation ( eq . ( [ vlasov ] ) ) as @xmath102 where @xmath103 is the single - particle energy . using @xmath104 and writing @xmath105 with @xmath106 which has same time and spatial dependence as @xmath107 , eq .
( [ vlasov1 ] ) can be rewritten as @xmath108 by substituting the above equation into eq .
( [ vlasov2 ] ) , we obtain in the low - temperature limit @xmath109 with @xmath110 and @xmath111 being the fermi velocity . carrying out the angular integration leads to the usual lindhard function @xmath112 in the low - temperature limit , the momentum integration can be approximately evaluated as @xmath113\end{aligned}\ ] ] for @xmath114 , with @xmath115 and @xmath116,\ ] ] for electrons , where @xmath117 is the electron chemical potential . for the factor @xmath118 in eq .
( [ drho ] ) , there are the local contribution @xmath119 , the direct contribution @xmath120 from the finite - range interaction , and the gradient contribution @xmath121 from the finite - range exchange interaction using the density matrix expansion . in the case of neutrons , they are given by @xmath122 after neglecting higher - order terms in @xmath123 .
in the long - wavelength case , the finite - range direct contribution can be rewritten as @xmath124 with @xmath125 for protons , there are additional direct and exchange coulomb contributions given , respectively , by @xmath126 for electrons , there are only direct and exchange coulomb contributions to @xmath118 .
after linearizing the vlasov equation , the collective density fluctuation @xmath123 then satisfies the following equation @xmath127 with @xmath128 we note that the three terms in the above equation are due to , respectively , the bulk , the density - gradient , and the coulomb contribution .
non - trivial solutions of eq .
( [ eqdrho ] ) are obtained when @xmath129 , which also determines the dispersion relation @xmath130 of the collective density fluctuation .
the transition density in a neutron star is the density at which the collective density fluctuation would grow exponentially , resulting in the instability of the neutron star matter , and this happens when the frequency @xmath96 becomes imaginary . to determine the condition for this to occur , we let @xmath131 @xmath132 and rewrite the lindhard function as @xmath133 .
since the values of @xmath134 are @xmath135 , the critical values @xmath136 , corresponding to @xmath137 , then determine the spinodal boundary of the system when they are substituted into @xmath129 .
expanding the nucleon and electron densities at low temperatures according to @xmath138,\\ \rho_q & \approx & \frac { ( 2m_q^\star\epsilon_q^f)^{3/2}}{3\pi^2}\left [ 1 + \frac{\pi^2}{8 } \left(\frac{t}{\epsilon_q^f } \right)^2\right ] , ~ q = n , p\end{aligned}\ ] ] we obtain from eqs .
( [ xq ] ) and ( [ xe ] ) the following relations : @xmath139 the bulk contribution in eq . ( [ matrix ] ) in the low - temperature limit can thus be rewritten as @xmath140 denoting the effective density - gradient coefficients as @xmath141 we then obtain same expressions for determining the spinodal boundary of @xmath88 matter in the so - called dynamical approach @xcite , if the coulomb exchange terms are neglected .
in particular , in the long - wavelength limit , the spinodal boundary is determined by the vanishing point of @xmath142 where @xmath143 to find the upper limit of the spinodal boundary for all possible wavevectors , we first find the minimum value of @xmath144 given by @xmath145 which occurs when @xmath146 . the density that makes eq .
( [ vdynmin ] ) vanish then determines the spinodal boundary in the neutron star matter or the transition density at the inner edge of neutron star crusts . for symmetric nuclear matter .
the finite - range direct and exchange contributions are compared , and the value extracted by oyamatsu @xcite is also shown . ] in fig .
[ dd ] , we show the density dependence of effective density - gradient coefficients from different values of @xmath23 for symmetric nuclear matter .
it is seen that the contribution from the finite - range direct interaction is much larger than that from the finite - range exchange interaction .
also , @xmath147 decreases while @xmath148 increases with increasing value of @xmath23 . furthermore , @xmath149 and @xmath150 for symmetric nuclear matter , whereas all four @xmath151
s have different values for asymmetric nuclear matter .
the density - gradient coefficients extracted from the mdi interaction are , however , much larger than the value of @xmath152 mev @xmath153 shown in the figure that was used by oyamatsu @xcite to fit the nuclear radii and is also consistent with the average value from different skymre interactions .
we note that the large density - gradient coefficients in the presence study is due to our assumption that all of the momentum dependence in the mdi interaction comes from the finite - range exchange term . and pressure @xmath154 at the inner edge of neutron star crust as functions of the slope parameter @xmath155 for @xmath156 mev and @xmath157 mev .
previous results @xcite based on density - gradient terms from ref .
@xcite are also shown for comparison . ] in fig .
[ rhotpt ] , we compare the transition density and pressure in this study with those from previous calculations @xcite that use the value of @xmath158 mev @xmath153 from ref .
@xcite for the density - gradient coefficients and neglect the exchange coulomb interaction .
it is seen that the present results are smaller than previous ones for neutron star temperatures @xmath156 mev and @xmath157 mev .
we note that including the exchange coulomb term for protons slightly increases @xmath159 while the larger density - gradient coefficients make @xmath159 smaller . from the latest constraint @xmath160 mev on the density slope of the nuclear symmetry energy ,
the transition density and pressure are constrained within @xmath161 @xmath162 @xmath12 and @xmath163 mev/@xmath164 mev/@xmath165 for @xmath156 mev , and @xmath166 @xmath167 @xmath12 and @xmath168 mev/@xmath169 mev/@xmath165 for @xmath157 mev in the present study . although the transition density is smaller at fixed @xmath155 compared to our previous results @xcite , which leads to a smaller crustal fraction of the moment of inertia for neutron stars and an even stricter constraint on the masses and radii of neutron stars , the upper limit values of @xmath159 and @xmath154 are larger because of the smaller values of @xmath155 . the final constraint on the neutron star mass - radius ( m - r ) relation is expected to be similar to that in the previous work @xcite .
assuming that the momentum dependence in the mdi interaction is entirely due to the finite - range exchange term in the nucleon - nucleon interaction , we have identified the nn interaction potential that underlies the mdi interaction , which has been extensively used in studying isospin effects in intermediate - energy heavy - ion collisions as well as the properties of hot nuclear matter and neutron star matter through the resulting potential energy density in infinite nuclear matter . using the density matrix expansion ,
we have obtained an effective zero - range skyrme - like interaction with density - dependent parameters from the finite - range exchange interaction .
compared to the density - gradient coefficients extracted by oyamatsu @xcite and the average value from different skyrme interactions , the values from the mdi interaction are much larger . since the momentum dependence in the mdi interaction could also come from the intrinsic momentum dependence in the elementary nucleon - nucleon interaction or be induced by the effects of short - range nucleon - nucleon correlations , we have thus overestimated the effect of the finite - range exchange term . as an application , the resulting interaction is used to determine the transition density and pressure at the inner edge of neutron star crusts based on the linearized vlasov equation for the neutron star matter .
fortunately , the instability of neutron star matter is dominated by the bulk properties of nuclear matter , the large density - gradient coefficients due to the finite - range interaction do not reduce much the transition density and pressure in neutron stars .
the two particle state @xmath170 in eq .
( [ vv ] ) contains the spatial , spin and isospin parts as @xmath171 for the inner product of spatial state , we have @xmath172 with @xmath173 being the space wave function for particle @xmath34 and the density @xmath174 , and @xmath175 with the off - diagonal density @xmath176 , which reduces to the density when @xmath177 .
for the zero - range term @xmath178 in the mdi interaction , its direct contribution to the energy can be calculated using the fact that @xmath179 when considering the inner product of spin state in spin saturated matter , and the result is @xmath180 with @xmath181 its exchange contribution @xmath182 can be evaluated by using @xmath183 and replacing @xmath40 by @xmath184 when considering the inner product of isospin state , and the result is @xmath185 with @xmath186 and @xmath187 is the isospin asymmetry .
the total contribution of the zero - range interaction to the potential energy density is thus @xmath188 comparing it with eq .
( [ mdiv ] ) leads to eqs .
( [ t3])@xmath189([alpha ] ) . for the contribution from the finite - range interaction in eq .
( [ mdiv ] ) to the energy , we use in the exchange term the relation @xmath190 to obtain the following total contribution to the potential energy @xmath191\notag\\ & \times&\left(w+\frac{b}{2}-h-\frac{m}{2}\right)+2\rho_n(\vec{r}_1)\rho_p(\vec{r}_2 ) \left(w+\frac{b}{2}\right)\notag\\ & + & [ \rho_n(\vec{r}_1,\vec{r}_2)\rho_n(\vec{r}_2,\vec{r}_1)+\rho_p(\vec{r}_1,\vec{r}_2 ) \rho_p(\vec{r}_2,\vec{r}_1)]\notag\\ & \times&\left(m+\frac{h}{2}-b-\frac{w}{2}\right ) \notag\\ & + & 2\left .
\rho_n(\vec{r}_1,\vec{r}_2 ) \rho_p(\vec{r}_2,\vec{r}_1)\left(m+\frac{h}{2}\right)\right\}\frac{e^{-\mu first and second terms being the direct contribution , and the third and fourth terms being the exchange contribution . to obtain the energy density functional from the finite - range interaction , we introduce the coordinate transformation @xmath192 for the direct contribution , we use the approximation of infinite nuclear matter with constant density and make use of the integral @xmath193 to obtain @xmath194 with @xmath195 for the exchange contribution , we express the density matrix in terms of the fourier transform of the wigner function , that is @xmath196 substituting eq .
( [ wigner ] ) into eq .
( [ er ] ) and evaluating the fourier transform of the yukawa interaction according to @xmath197 the exchange contribution of the finite - range interaction to the potential energy is then @xmath198 with @xmath199.\end{aligned}\ ] ] comparing eqs .
( [ vrd ] ) and ( [ vre ] ) with eq .
( [ mdiv ] ) then allows one to obtain eqs .
( [ lambda])@xmath189([m ] ) for the parameters in the finite - range yukawa potential in the mdi interaction .
here we follow exactly the same procedure for the density matrix expansion introduced in ref . @xcite . to obtain the energy density functional in eq .
( [ hfh ] ) from the exchange contribution in eq .
( [ er ] ) , we first use the coordinate transformation given by eq .
( [ trans ] ) to express the density matrix as @xmath200 with @xmath201 acting on the first ( second ) term on the right , and @xmath202 is the spatial wave function of particle @xmath34 with isospin @xmath203 .
the angular average over the direction of @xmath204 is then @xmath205}{\frac{\vec{s}}{2}\cdot(\nabla_1-\nabla_2 ) } \sum_i \phi_{\tau i}^\star(\vec{r})\phi_{\tau i}(\vec{r}).\end{aligned}\ ] ] using the bessel - function expansion @xmath206 with @xmath207 the @xmath208th - order spherical bessel function and @xmath209 one obtains by identifying @xmath210 and @xmath211 with @xmath212 being the fermi momentum and keeping up to the third - order bessel function the following result : @xmath213\end{aligned}\ ] ] with @xmath214 , @xmath215 and @xmath216 being the kinetic energy density . by making the approximation @xmath217 \notag\\ & + & \rho_{\tau^\prime}(\vec{r})\rho_{sl}^{}(k_{\tau^\prime }
s)g(k_\tau s)s^2\notag\\ & \times&\left[\frac{1}{4}\nabla^2 \rho_\tau(\vec{r})-\tau_\tau(\vec{r})+\frac{3}{5 } k_\tau^2\rho_\tau(\vec{r})\right],\end{aligned}\ ] ] and using the relation @xmath218 we obtain @xmath219 where @xmath220 is the skymre - like zero - range energy density from the finite - range exchange interaction defined in eq .
( [ hfh ] ) . | by assuming that the isospin- and momentum - dependent mdi interaction has a form similar to the gogny - like effective two - body interaction with a yukawa finite - range term and the momentum dependence only originates from the finite - range exchange interaction , we determine its parameters by comparing the predicted potential energy density functional in uniform nuclear matter with what has been usually given and used extensively in transport models for studying isospin effects in intermediate - energy heavy - ion collisions as well as in investigating the properties of hot asymmetric nuclear matter and neutron star matter .
we then use the density matrix expansion to derive from the resulting finite - range exchange interaction an effective skyrme - like zero - range interaction with density - dependent parameters . as an application
, we study the transition density and pressure at the inner edge of neutron star crusts using the stability conditions derived from the linearized vlasov equation for the neutron star matter . |
we work with decision problems derived from optimization problems . the reader is assumed to have some background in finite model theory .
if not , the book by ebbinghaus and flum @xcite serves as a good introduction .
publications @xcite and @xcite are also relevant material for this line of research .
our main result is the following theorem : existential second order logic with a first order part that is universal horn ( or simply , eso universal horn ) is insufficient to capture the class * p , the class of problems decidable in polynomial time , even assuming that the input structures come with a successor relation . @xmath0
[ thm : esohierarchy ] * we provide two proofs of theorem [ thm : esohierarchy ] : ( a ) one based on reduced products , and ( b ) a second proof based on approximability theory ( this proof assumes that * p @xmath1 * np ) . * * we first introduce some notation and definitions .
* eso universal horn is defined as existential second order ( eso ) logic where the first order part is a universal horn formula . @xmath0 * * maximum matching .
given an undirected graph @xmath2 and a non - negative integer @xmath3 , is there a matching in @xmath4 such that the number of matched edges is at least @xmath3 ?
assume that @xmath4 has ( i ) no self - loops ; and ( ii ) at most one edge between any pair of vertices .
@xmath0 [ def : matching ] * ( * a note about k )
. readers may ask , how is @xmath3 represented in the input ?
in binary or unary ? " .
this may be relevant for problems with large data , such as maximum flow or minimum cost flow , where edge weights are part of the input .
however , the issue is irrelevant for problems such as maximum matching where the solution value lies in the @xmath5 $ ] range , where @xmath6 and @xmath7 is the number of vertices . if @xmath8 , then the solution is immediately disqualified there can not be a solution to the maximum matching problem with the number of matched edges higher than @xmath9 . if @xmath10 , the input size for @xmath3 is at most that of @xmath9 , whether in binary or unary .
@xmath0 * for simplicity , assume the following for problem [ def : matching ] : @xmath11 , where @xmath7 is the number of vertices in the input graph of the problem instance .
also , throughout this article , @xmath3 is a part of the input instance ; it can vary from instance to instance ; it is _ not a fixed constant over all instances . @xmath0
[ ass : kleqn ] _
we represent decision versions of optimization problems as a conjunction of a single objective function constraint ( ofc ) and a set of basic feasibility constraints ( bfc ) .
the motivation for this comes from a general mathematical programming framework where optimization problems are expressed in the form : @xmath12 & \mbox{subject to the following constraints : } \\[2 mm ] & \mbox{$\ds g_i(\mb{x } ) \le 0 $ , ~ $ 1 \le i \le m$ } , \\[1 mm ] & \mbox{$\ds h_j(\mb{x } ) = 0 $ , ~ $ 1 \le j \le p$ } , \\[2 mm ] & \mbox{and $ \mb{x } \in x$ ( for example , $ x = \bb{r}^n$)}. \end{array}\ ] ] additionally , for the decision problem , we are given a number @xmath3 as a part of the input .
above , the bfc comprises the @xmath13 constraints @xmath14 and @xmath15 , and the ofc comprises the single constraint @xmath16 for maximization problems ( @xmath17 for minimization problems ) .
the proof is a counterexample .
it relies on the fact that universal horn formulae are preserved under reduced products .
that is , if @xmath18 is a universal horn formula and structures @xmath19 and @xmath20 satisfy @xmath18 , then so does their reduced product @xmath21 . from the definition of problem [ def :
matching ] , note that any feasible solution to _ maximum matching consists of at least @xmath3 matched edges .
_ maximum matching can not be expressed in eso universal horn logic , even assuming that the input structures come with a successor relation . @xmath0
[ lem : matchingreducedproducts ] suppose we have a signature divided into two parts : a set of relation symbols @xmath22 , and a separate set @xmath23 of relation and constant symbols .
@xmath24 is true iff @xmath25 is a matched edge .
the relations in @xmath23 are first order ( input ) , hence the horn restriction does not apply to these . on the other hand ,
the relations in @xmath26 are second order ( quantified ) , and hence the horn condition applies to these relations . in particular , it applies to @xmath27 , which is unknown ; it is not a part of the input .
let us introduce this definition ( henceforth referred to as property [ def : maxmatchprop ] ) : ( maximum matching property ) .
given @xmath28 and an integer @xmath3 , is there a matching @xmath29 such that @xmath30 ? @xmath0 [ def : maxmatchprop ] let * a and * b be two input structures on the same universe @xmath31 ( @xmath32 ) with a successor relation ; furthermore , for all symbols in the signature @xmath23 , let * a and * b agree on the interpretation of the symbols ; however , they may differ on relations in @xmath26 . * * * * now define a product structure * a.b to be the structure on the domain @xmath31 with a linear ordering , in which every symbol in the signatures @xmath26 and @xmath23 is interpreted as the _ intersection of the two interpretations in * a and * b. then it is easy to show that any horn formula @xmath33 that is satisfied in both * a and * b is also satisfied in * a.b .
( this is exactly by the argument that shows that horn formulae are closed under direct products ; see chapter 9 , page 417 , of @xcite , or see page 493 of @xcite . ) * * * * * _ * then suppose we let * a and * b be structures with , say , @xmath34 vertices @xmath31 .
assume that in both structures , we have edges from the set @xmath35 .
* * in * a , we let @xmath27 consist of the @xmath36 edges @xmath37 ; and in * b , we let @xmath27 consists of the @xmath36 edges @xmath38 . thus @xmath27 indeed defines a matching in both structures .
* * also , we let @xmath39 in both structures .
so , property [ def : maxmatchprop ] is true in both * a and * b. that is , there exists a matching of size at least @xmath36 in * a and * b. * * * * however , property [ def : maxmatchprop ] is false in * a.b , since in this structure , @xmath27 is empty , even though the lower bound @xmath3 is still equal to @xmath36 .
it follows that property [ def : maxmatchprop ] can not be expressed as a horn formula , and hence we conclude that _ maximum matching , a problem in the class * p , can not be expressed in eso universal horn logic .
@xmath0 * _ * notes on the above proof : * readers may ask , _ is it not necessary to look for a solution @xmath27 ( a matching ) in the product structure ? the answer is no , it is _ not necessary .
if property [ def : maxmatchprop ] can indeed be expressed in eso universal horn logic , and assuming that the property is satisfied in * a and * b , then it must be satisfied in the product structure * a.b , according to the results in @xcite and @xcite . * * * _ _ * recall that the proof is a counterexample .
this allows us to choose any value for @xmath27 ( that it could possibly take ) in the structures * a and * b. * * * also note that this proof does _ not deal with the sub - property @xmath40 " alone it deals with the whole _ maximum matching property ( property [ def : maxmatchprop ] ) . _ _ * in place of _ maximum matching , we could have used any decision problem in * p derived from a maximization problem . * _
this proof assumes that @xmath41 .
we use approximability theory to show that some decision problems in the class * p , derived from optimization problems , can not be expressed in eso universal horn logic .
in particular , we rely on proven upper and lower bounds on the approximation ratios ( defined below ) for the vertex cover problem .
we disprove the following proposition , which appeared as a theorem in @xcite ( also as theorem 3.2.17 in @xcite ) : * assuming a successor relation as a part of the input structure , every problem in the class * p can be expressed in eso universal horn logic . @xmath0
[ thm : esohornyes ] * @xcite a * np - optimization problem @xmath42 is a tuple @xmath43 , where * 1 .
@xmath44 is a set of instances to @xmath45 , 2 .
@xmath46 is the set of feasible solutions to instance @xmath47 , 3 .
@xmath48 is the _ objective function _ value to a solution @xmath49 of an instance @xmath50 .
it is a function @xmath51 \rightarrow \bb{r}^+_0 $ ] ( non - negative reals ) , computable in time polynomial in the size @xmath52 of the domain of @xmath47 , 4 .
for an instance @xmath50 , @xmath53 is either the minimum or maximum possible value that can be obtained for the objective function , taken over all feasible solutions in @xmath54 .
+ @xmath55 ( for np - maximization problems ) , + @xmath56 ( for np - minimization problems ) , 5 .
the following decision problem is in * np : _ given an instance @xmath47 and a non - negative integer @xmath3 , is there a feasible solution @xmath57 , such that @xmath58 ( for an np - maximization problem ) , or @xmath59 ( for an np - minimization problem ) ? _ * the set of all such np - optimization problems is the @xmath60 class . @xmath0
[ def : npoptproblem ] note : _ feasibility as defined in point ( ii ) above has nothing to do with any upper or lower bounds on the objective function value @xmath61 .
it only concerns the bfc , see sec .
[ sec : bfcofc ] .
this also applies to def .
[ def : decisionproblem ] below . _ * decision versions .
[ def : npoptproblem ] . given a non - negative integer @xmath3 and an instance @xmath50 , the decision version of an np - optimization problem @xmath42 asks whether there is a feasible solution @xmath57 , such that @xmath58 ( if @xmath42 is a maximization problem ) , or @xmath59 ( if @xmath42 is a minimization problem ) . @xmath0 [ def : decisionproblem ] * ( see theorem 3.2.17 in page 147 of @xcite . ) if a decision problem @xmath62 can be represented by an eso universal horn sentence , then @xmath62 can solved in polynomial time . @xmath0
[ thm : gradel ] the above theorem follows from the fact that hornsat ( satisfiability of a propositional horn formula ) can be solved in ptime @xcite .
* approximation ratio .
given an instance @xmath47 of an optimization problem @xmath42 ( as in def .
[ def : npoptproblem ] ) and a feasible solution @xmath57 , the approximation ratio is defined as @xmath63 for minimization problems and @xmath64 for maximization problems .
we assume that @xmath65 , where @xmath66 is the optimal solution value to instance @xmath47 . @xmath0
[ def : approxratio ] * * approximation problems @xmath67 and @xmath68 . * : an instance @xmath47 of a problem @xmath69 ( cf .
[ def : npoptproblem ] ) .
( upper bound ) problem @xmath70 .
is there a feasible solution @xmath71 , such that the approximation ratio is at most @xmath72 ? ( lower bound ) problem @xmath73 .
is there a feasible solution @xmath57 , such that the approximation ratio have the phrase at most " in their definitions .
this is * not a typographical error . * ] is at most @xmath74 ?
( @xmath75 . )
[ def : approxproblem ] @xmath0 some remarks about def .
[ def : approxproblem ] ( also see the appendix ) : a _ feasible solution as mentioned in the definition of @xmath76 and @xmath77 has nothing to do with the objective function value @xmath61 .
hence it has nothing to do with the approximation ratio either .
this feasibility only concerns the constraints in the bfc .
[ sec : bfcofc ] .
[ def : approxproblem ] is irrelevant for problems in * p , hence we assume that the decision version of problem @xmath42 is not known to be polynomially solvable .
@xmath0 * * the apx class .
consider a problem @xmath42 from the class @xmath60 ( cf .
[ def : npoptproblem ] ) .
then @xmath42 is defined to be a member of apx iff the decision problem @xmath78 can be solved in ptime for some constant @xmath79 . @xmath0
[ def : apx ] * strictly speaking , the upper bound problem @xmath78 depends on the parameter @xmath72 . however , for ease of discussion , let us assume that @xmath72 is pre - defined for @xmath80 ; assume that @xmath72 is in - built " into @xmath80 . thus henceforth , we simply write @xmath80 rather than @xmath78 .
we make a similar assumption for the lower bound problem @xmath81 and the related parameter @xmath74 ; we will write @xmath81 rather than @xmath82 . for a proven upper bound @xmath72 , @xmath83 is ptime solvable . similarly , for a proven lower bound @xmath74 , @xmath84 is np - complete .
example : steiner tree is a problem in apx , with parameter @xmath85 . hence given an instance @xmath47 of the steiner tree problem , there is a ptime algorithm which can decide whether there is a feasible solution to @xmath47 with an approximation ratio of at most two .
however , @xmath72 is an _ upper bound on the approximation ratio ; we know that there is a ptime algorithm that guarantees a feasible solution within this upper bound .
furthermore , many problems ( such as vertex cover , below ) in the class apx also have a proven _ lower bound @xmath74 . for such problems ,
the approximation problem @xmath86 is known to be * np - complete .
* _ _ * vertex cover ( optimization version ) . * : a graph @xmath28 .
: mark vertices in @xmath87 in such a way that ( i ) for any @xmath88 where @xmath89 , if @xmath90 , then either @xmath91 of @xmath92 is marked , and ( ii ) the number of marked vertices is to be minimized . :
a set of marked vertices that obeys property ( i ) .
[ prob : vertexcover ] @xmath0 let @xmath93 and @xmath94 be the approximation problems @xmath95 and @xmath96 respectively ( from def .
[ def : approxproblem ] ) , applied to vertex cover .
@xmath93 and @xmath94 are members of the class * np ; in particular , @xmath93 is a member of the class * p , and @xmath94 is * np - complete . * * * the upper bound @xmath97 and the lower bound @xmath98 @xcite . in fact , for @xmath99 , one could use any number @xmath100 such that @xmath101 . for ease of representation in an eso formula ,
let us use @xmath102
. here is the [ page : corearg ] core argument : proposition [ thm : esohornyes ] is incorrect .
[ lem : esohornno ] @xmath0 since @xmath103 , theorem [ thm : esohornyes ] can be applied .
then it follows that @xmath93 can be expressed as an eso universal horn sentence @xmath104 .
this sentence can be written as a conjunction of two sub - formulae : 1 .
( the feasibility condition ) one stating that every edge is covered , which is easily written as @xmath105 and 2 .
( the approximation ratio condition ) another stating that the approximation ratio @xmath106 is at most two , that is , @xmath107 however , if we replace the sub - formula for the condition @xmath108 with the sub - formula for @xmath109 , we obtain a new eso universal horn sentence @xmath110 which captures @xmath94 .
though @xmath111 is known to be * np - complete , we have just obtained an eso universal horn sentence ( @xmath112 ) for this problem , implying that @xmath111 can be decided in ptime , when we apply theorem [ thm : gradel ] .
but this contradicts our assumption that * p @xmath1 * np . @xmath0 * * * hence we have shown the following : assuming that * p @xmath1 * np , and assuming a successor relation as a part of the input vocabulary , @xmath93 , a problem solvable in ptime , can not be expressed as an eso universal horn sentence . @xmath0 * * this concludes the second proof of theorem [ thm : esohierarchy ] .
we note a discrepancy between the results in theorems [ thm : esohierarchy ] and [ thm : esohornyes ] .
we observe that there is a fundamental difference between the types of expressions dealt with by the two theorems .
theorem [ thm : esohierarchy ] is about _ structure level expressions ( sle ) which deal directly " with properties of mathematical structures , whereas theorem [ thm : esohornyes ] is about _ machine level expressions ( mle ) which represent encodings of machine level computations of properties ; these merely encode the * steps ( * moves ) of the computation . * * _ _ we note that ( eso universal horn ) = * p at the _ machine level , whereas ( eso universal horn ) @xmath113 * p at the _ structure level . _ * _ * observe that in the two proofs , we * do not use a turing machine ; the expressions * are not the result of an output from a turing machine computation . * * the precise definitions for _ structure level and _ machine level are provided later in the paper but first , we describe intuitive notions . _ _
a few observations are in order : 1 .
descriptive complexity is a very useful tool at the structure level , _ not at the machine level . _
2 . a large subset of problems in the class *
np is derived from optimization problems , hence this is a important sub - class .
( problem [ def : matching ] has also been derived from an optimization problem . ) * 3 . in spite of the important sub - class described above
, we were unable to locate any _ structure level expression for such problems in the literature ( although one can construct _ machine level expressions using the methods in @xcite , chapter 7 of @xcite or @xcite ) .
this is a key motivation for our line of research .
_ _ from a mathematician s viewpoint , sle s are the natural choice to express a problem / property , _ not the mle s . to logically express a property of a given mathematical structure , what is the need to put it through computation first ? _
to write an mle for a given property @xmath114 , we first need to design a tm which will decide whether @xmath114 is true ( for any given input ) . but
this means that the space / time complexity to decide @xmath114 is already known , hence it is pointless writing an mle for @xmath114 . to elaborate on the item numbered one above : the usefulness of descriptive complexity lies in the fact that if one could write a structure level expression in a certain logic for a decision problem @xmath115 , then it can be deduced that @xmath115 requires a certain amount of resource ( time and/or space ) for its computation ; that is , we can recognize the complexity class that @xmath115 belongs to
however , our main result ( theorem [ thm : esohierarchy ] ) shows that even though a large class of problems ( those derived from optimization ) is ptime solvable , they can not be expressed in eso universal horn logic at the structure level .
hence for such problems , the usefulness of descriptive complexity is lost while using this particular logic . * machine level expressions ( mle ) .
these express the _ computation of a property ( in a turing machine , for instance ) . here , _
computation of a property is defined as _ checking whether a property is true using a computation device , for a given input . _ _ _ * such an expression encodes the computation steps of a turing machine ( tm ) , for example , as in the proof by fagin @xcite that eso logic captures the class * np and the proof by immerman @xcite that lfp ( under ordered structures ) captures the class * p. these expressions state that ( i ) if we go through the steps of a computation device such as a tm , and ( ii ) if the tm finally reaches an accepting state , then the property must be true for the given input structure . * * for example , to write a machine level expression for _ hamiltonian cycle ( hc ) , we should first design a tm that solves hc , then encode the steps of the tm in an expression . _ * structure level expressions ( sle ) . on the other hand , structure level logical expressions
do not express the _ computation of a property ; rather , they directly express the truth of a property of a mathematical structure .
_ no computation is involved in such a description . _
_ * _ machine level expressions for the class * np ( class * p ) were provided by fagin @xcite ( grdel @xcite ) respectively . * * _ since the structure level is indeed proven to be different from the machine level , we can then state the following : at the structure level , the validity of fagin s theorem is completely open it is unknown whether eso sentences can characterize the class * np at this level . @xmath0 * in the last 40 years , important contributions have been made at the machine level " , such as fagin s theorem and the immerman - vardi theorem .
however , these do not represent the full picture .
the formal mathematical definitions for concepts such as mle and sle explain accurately as to why theorems [ thm : esohierarchy ] and [ thm : esohornyes ] are correct in their respective domains " ( the machine computation domain and the mathematical structure domain ) .
we now provide the formal definitions of mle and sle . given a structure @xmath116 as input , defined on a first order vocabulary @xmath117 , we provide the following definitions : * higher order predicates are those that are not defined in the first order vocabulary of the input structure . [
def : hopred ] * \(a ) let @xmath118 be the set of expressions in logic @xmath119 such that a structure @xmath116 possesses property @xmath114 iff @xmath120 for every expression @xmath121 .
\(b ) we let @xmath122 be the set of * machine level expressions ( mle ) for @xmath114 . naturally , @xmath123 . we define @xmath122 as follows . *
let @xmath124 be turing machines that have at least one halting computation iff @xmath116 possesses property @xmath114 where @xmath125 .
hence when @xmath116 satisfies @xmath114 , let @xmath126 be the @xmath127 halting computation of @xmath124 where @xmath128 .
over the set of all turing machines \{@xmath129 , the set of halting computations is the set @xmath130 .
let @xmath131 be an encoding of @xmath126 in @xmath119 .
then @xmath122 = @xmath132 .
the predicates that encode machine computations are of higher order ( cf .
[ def : hopred ] ) .
[ def : mle ] @xmath0 above , we say at least one halting computation " to take into account the fact that @xmath133 could be a non - deterministic tm .
next , we define structure level expressions , as a subset of @xmath118 .
consider the set @xmath118 in def .
[ def : mle ] .
we let @xmath134 be the set of * structure level expressions ( sle ) for @xmath114 . naturally , @xmath135 . we define @xmath134 as follows :
for every expression @xmath136 , ( i ) the higher order predicates of @xmath137 range over tuples from the universe of @xmath116 , and ( ii ) the variables and first order predicate symbols in @xmath137 are from @xmath117 .
[ def : sle ] @xmath0 * although def .
[ def : mle ] explains how to generate an mle , it does nt tell us how to recognise an mle when we see one ; it does nt tell us how to distinguish an mle from an sle .
this can be achieved quite easily since the vocabulary of the two expression types are different . for an sle ,
the language consists of variables and relation symbols from the input vocabulary @xmath117 of structure @xmath116 ( as in def .
[ def : sle ] ) .
however for an mle , the language consists of * machine encodings for ( i ) the tape symbols of a tm ; ( ii ) the tape head movement and transitions of the tm ; and ( iii ) the variables and relation symbols in @xmath117 as well as the higher order predicate symbols used to encode the computation . *
for the class * p , it is a case of
easier done than said " .
one can let the computation of a decision problem run through a turing machine and obtain an expression in eso universal horn logic
. however , _ without computation , eso universal horn is insufficient to capture * p ( we need a stronger logic ) , as the two proofs of theorem [ thm : esohierarchy ] show . * _ *
nerio borges , anuj dawar , ian hodkinson , leonid libkin and anand pillay . 10 i. dinur and s. safra . , 162(1):439485 , july 2005 . .
springer , 1999 .
r. fagin . .
in r. karp , editor , _ complexity of computations _ , pages 4373 .
siam - ams proceedings ( no.7 ) , 1974 . . .
stacs 1991 : proceedings of the 8th annual symposium on theoretical aspects of computer science lecture notes in computer science 280 _ , pages 466477 .
springer - verlag , 1991 .
e. grdel , p.g .
kolaitis , l. libkin , m. marx , j. spencer , m.y .
vardi , y. venema , and s. weinstein . .
texts in theoretical computer science .
springer , 2007 .
w. hodges . .
cambridge university press , 1993 .
w. hodges . .
in dov m. gabbay , c.j .
hogger and j.a .
robinson ( eds . ) , _ handbook of logic in artificial intelligence and logic programming ( vol .
1 ) _ , pages 449503 . oxford university press ( ny ) , 1993 . n. immerman . .
springer - verlag , 1999 .
n. immerman . . ,
68(1 - 3):86104 , 1986 .
jones and w.t .
laaser . . ,
3(1):105117 , october 1976 .
a. panconesi and d. ranjan .
quantifiers and approximation . , 107:145163 , 1993 .
this paper was presented at the asian logic conference , wellington , nz , in december 2011 .
some readers who read a draft of this paper , after reading def .
[ def : approxproblem ] , remarked ,
these problems are very easy " .
well , i doubt if these problems can be categorised as very easy " .
however , points to note include the following : 1 .
a _ feasible solution means that the solution obeys the bfc ( cf .
[ sec : bfcofc ] ) feasibility here has nothing to do with the objective function nor the approximation ratio ; _ 2 . if @xmath138 , then @xmath139 and @xmath140 has no meaning .
+ conversely , if we can design an algorithm that solves @xmath141 in polynomial time for @xmath139 , then @xmath142 ; 3 .
if the answer is yes ( to either of the problems defined ) , then a feasible " solution @xmath143 that obeys the bound ( @xmath72 or @xmath74 ) must be returned ; 4 .
the approximation ratio is _ not given to us ; we are only given a _
bound on the approximation ratio ( @xmath72 or @xmath74 , as the case may be ) ; _ _ 5 . hence , for the upper bound problem @xmath83 , to answer the question with a yes or no within polynomial time , over all instances
, we should develop a polynomial time algorithm for @xmath42 ( which is not a trivial task for many a problem @xmath69 ) ; and 6 . for a proven upper bound @xmath72 , @xmath83 is ptime solvable .
similarly , for a proven lower bound @xmath74 , @xmath84 is np - complete .
note that @xmath75 .
it is the gap between @xmath74 and @xmath72 that we seek to exploit .
a reader raised a concern as to why should the expression for @xmath93 consist of a sub - formula for the condition @xmath144 .
the answer is , it is _ not a question of * should , but a question of * can . yes , we _ can express @xmath93 as a conjunction of the ( sub - formulae ) for the two conditions the feasibility condition , and the approximation ratio condition . | we show that eso universal horn logic ( existential second logic where the first order part is a universal horn formula ) is insufficient to capture * p , the class of problems decidable in polynomial time .
this statement is true in the presence of a successor relation in the input vocabulary .
we provide two proofs one based on reduced products of two structures , and another based on approximability theory ( the second proof is under the assumption that * p is not the same as * np ) .
we show that the difference between the results here and those in @xcite , is due to the fact that the expressions this paper deals with are at the structure level " , whereas the expressions in @xcite are at the machine level " a case of _ easier done than said . _ * * * _ keywords : descriptive complexity , optimization , existential second order logic , machine level and structure level . _
acm subject classification : f.1.3 , f.4.1 . |
present - day elliptical galaxies are known to host mostly old stellar populations ( @xcite , @xcite , @xcite ) . whether their stars have formed in situ or whether ellipticals assembled their present - day morphology only over time ( for example by mergers )
is less clear .
an important clue on the assembly redshift of a galaxy is provided by its dark matter density .
for example , in the simple spherical collapse model @xcite the average density of virialized halos is proportional to the mean density of the universe at the formation epoch : halos which form earlier become denser .
similarly , in cosmological @xmath5-body simulations , the concentration ( and , thus , the inner density ) is found to be higher in halos that have assembled earlier ( e.g. @xcite , @xcite ) .
in addition to this connection between formation epoch and halo density , the final halo mass distribution also depends on the interplay between dark matter and baryons during the actual galaxy formation process ( e.g. @xcite , @xcite )
. then , the properties of galaxy halos provide valuable information about when and how a galaxy has assembled its baryons . despite its cosmological relevance ,
the radial distribution of dark ( and luminous ) mass in early - type galaxies is not well known : because of the lack of cold gas as a dynamical tracer , masses are difficult to determine .
stellar dynamical models require the exploration of a galaxy s orbital structure and have only recently become available for axisymmetric or more general systems ( @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . scaling relations for the inner dark matter distribution in early - types have by
now only been reported for round and non - rotating galaxies ( @xcite , @xcite ) and spirals ( @xcite and @xcite ) .
the aim of the present paper is to provide empirical scaling relations for generic cluster early - types ( flattened , with different degrees of rotation ) .
in particular , this paper is focussed on the inner dark matter density and its implications on the assembly redshift of elliptical galaxy halos .
the paper is organized as follows . in sec .
[ sec : data ] , we review the galaxy sample and its modelling .
dark matter scaling relations are presented in sec .
[ sec : scaling ] .
[ sec : density ] is dedicated to the dark matter density .
the effect of baryons on the dark matter density is discussed in sec .
[ sec : contraction ] , while sec .
[ sec : assembly ] deals with the halo assembly redshift . in sec .
[ sec : semi - analytic ] our results are compared to semi - analytic galaxy formation models .
a summary is given in sec .
[ sec : summary ] . in the following ,
we assume that the coma cluster is at a distance of @xmath6 .
the dark halo parameters discussed in this paper are derived from the axisymmetric , orbit - based dynamical models of bright coma galaxies presented in @xcite .
the original sample comprises two cd galaxies , nine ordinary giant ellipticals and six lenticular / intermediate type galaxies with luminosities between @xmath7 and @xmath8 .
the spectroscopic and photometric observations are discussed in @xcite , @xcite , @xcite and @xcite .
our implementation of schwarzschild s ( 1979 ) orbit superposition technique for axisymmetric potentials is described in @xcite . for a detailed discussion of all the galaxy models
the reader is referred to @xcite .
three of the 17 galaxies from @xcite are excluded from the analysis below .
firstly , we do not consider the two central cd galaxies ( gmp2921 and gmp3329 ; gmp numbers from @xcite ) , because their dark matter profiles may be affected by the cluster halo .
secondly , we omit the e / s0 galaxy gmp1990 , whose mass - to - light ratio is constant out to @xmath9 .
the galaxy either has no dark matter within this radius , or its dark matter density follows closer the stellar light profile than in any other coma galaxy . in either case ,
the mass structure of this object is distinct from the rest of the sample galaxies .
in addition to the remaining 14 galaxies we consider two further coma galaxies for which we collected data recently ( gmp3414 , gmp4822 ) .
the models of these galaxies are summarized in app .
[ app:3414 ] .
similar dynamical models as used here have been applied to the inner regions of ellipticals , where it has been assumed that mass follows light ( e.g. @xcite , @xcite ) .
in contrast , our models explicitly include a dark matter component ( cf .
we probed for two parametric profiles .
firstly , logarithmic halos @xmath10 which posses a constant - density core of size @xmath11 and have an asymptotically constant circular velocity @xmath12 .
the central density of these halos reads @xmath13 secondly , nfw - profiles @xmath14 which are found in cosmological @xmath5-body simulations @xcite .
the majority of coma galaxies are better fit with logarithmic halos , but the significance over nfw halo profiles is marginal . even if the best fit is obtained with an nfw - halo , then the inner regions are still dominated by stellar mass ( cf .
@xcite ) . in this sense , our models
maximize the ( inner ) stellar mass . in sec .
[ sec : scaling ] we will only discuss results based on logarithmic halo fits ( i.e. we use the halo parameters from columns 5 and 6 in tab . 2 of @xcite and columns 3 and 4 in tab . [ tab:3414 ] of app .
[ app:3414 ] , respectively ) . while these are not necessarily the more realistic profiles , they minimize systematics in the comparison with published scaling relations for spirals that were performed using cored profiles similar to our logarithmic halos .
the nfw fits are used in sec .
[ sec : contraction ] .
the @xmath2-band luminosities of coma galaxies used in this paper are taken from hyperleda .
we adopt a standard uncertainty of @xmath15 to account for zero - point uncertainties , systematic errors in the sky subtraction , seeing convolution , profile extrapolation and others .
effective radii are taken from @xcite and @xcite . here
we estimate the errors to be @xmath16 .
this is slightly higher than the uncertainties given in @xcite , but accounts for possible systematic errors @xcite .
stellar masses were computed from our best - fit stellar mass - to - light ratios @xmath17 and @xmath18-band luminosities of @xcite . in case
the best fit is obtained with a logarithmic halo , @xmath17 is taken from column 4 of tab . 2 in @xcite .
in case of an nfw fit , @xmath17 comes from column 8 of the same table
. the best - fit stellar mass - to - light ratios of gmp3414 and gmp4822 are given in column 1 of tab .
[ tab:3414 ] ( cf
[ app:3414 ] ) .
lcccc 0144 & @xmath19 & @xmath20 & @xmath21 & @xmath22 + 0282 & @xmath23 & @xmath24 & @xmath25 & @xmath26 + 0756 & @xmath27 & @xmath28 & @xmath29 & @xmath22 + 1176 & @xmath30 & @xmath31 & @xmath32 & @xmath33 + 1750 & @xmath34 & @xmath35 & @xmath36 & @xmath37 + 2417 & @xmath38 & @xmath39 & @xmath40 & @xmath41 + 2440 & @xmath42 & @xmath43 & @xmath44 & @xmath45 + 3414 & @xmath46 & @xmath47 & @xmath48 & @xmath49 + 3510 & @xmath50 & @xmath51 & @xmath52 & @xmath53 + 3792 & @xmath54 & @xmath55 & @xmath56 & @xmath57 + 3958 & @xmath58 & @xmath59 & @xmath60 & @xmath61 + 4822 & @xmath62 & @xmath63 & @xmath64 & @xmath65 + 4928 & @xmath66 & @xmath67 & @xmath68 & @xmath69 + 5279 & @xmath70 & @xmath71 & @xmath72 & @xmath73 + 5568 & @xmath74 & @xmath75 & @xmath76 & @xmath77 + 5975 & @xmath78 & @xmath79 & @xmath80 & @xmath33
[ fig : rhvh ] shows the scalings of dark halo core - radii @xmath11 and halo asymptotic circular velocities @xmath12 with @xmath2-band luminosity @xmath81 and stellar mass @xmath82 ( the corresponding galaxy parameters with errors are listed in tab .
[ tab : gals1 ] ) .
both , halo core - sizes and halo circular velocities tend to increase with luminosity and mass .
the case for a correlation between @xmath12 and @xmath81 is weak , if the sample as a whole is considered ( cf .
column 8 of tab .
[ tab : fits ] ) . however , four galaxies ( gmp0144 , gmp0756 , gmp1176 and gmp5975 ) separate from the rest of the sample galaxies in having both noticeably smaller @xmath11 and @xmath12 .
these galaxies are shown in light color in fig .
[ fig : rhvh ] . as a general trend ,
halo parameters tend to scale more tightly with luminosity and mass when these galaxies are omitted .
the solid lines in fig .
[ fig : rhvh ] show corresponding log - linear fits . for comparison , in tab .
[ tab : fits ] we give both , fits to all coma galaxies as well as fits to the subsample without the four galaxies offset in fig .
[ fig : rhvh ] .
the difference between these four galaxies and the rest of the sample is further discussed below .
the logarithmic halos of equation ( [ logdef ] ) have two free parameters .
any and pair of @xmath11 , @xmath12 or @xmath83 characterizes a specific halo .
[ fig : rhoh_rh ] shows a plot of @xmath83 versus @xmath11 .
both halo parameters are clearly correlated .
a linear relation fits the points with a minimum @xmath84 ( per degree of freedom ; cf .
[ tab : fits ] ) .
this rather low value partly derives from a degeneracy between the halo parameters in the dynamical modeling ( e.g. @xcite , @xcite ) which correlates the errors in both quantities . in tab .
[ tab : fits ] , such a correlation between the errors is not taken into account and the @xmath85 might be thus underestimated .
a @xmath85 much larger than unity would indicate some intrinsic scatter in fig .
[ fig : rhoh_rh ] , whereas the low @xmath85 quoted in tab .
[ tab : fits ] formally rules out any intrinsic scatter .
note that dark matter halos in cosmological @xmath5-body simulations can be approximated by a two - parameter family of halo models , where the parameters are correlated qualitatively in a similar way as revealed by fig .
[ fig : rhoh_rh ] ( e.g. @xcite , @xcite ) , but with some intrinsic scatter .
the four galaxies offset in fig .
[ fig : rhvh ] are also slightly offset in fig .
[ fig : rhoh_rh ] .
however , given the large uncertainties , this is not significant and fig .
[ fig : rhoh_rh ] is consistent with the halos of the four galaxies belonging to the same one - parameter family as established by the remaining coma galaxies .
this implies that the four galaxies primarily differ in the amount of stellar light ( and stellar mass , respectively ) that is associated to a given halo .
noteworthy , the four galaxies have stellar ages @xmath86 @xcite , while all other coma galaxies are significantly older ( mostly @xmath87 ) . a mere stellar population effect , however , is unlikely to explain the offset of the four galaxies . in this case
differences to other coma galaxies should vanish when galaxies are compared at the same stellar mass , which is not consistent with fig .
[ fig : rhvh]b / d .
it should be noted , though , that the stellar masses used here are taken from the dynamical models .
a detailed comparison with mass - to - light ratios from stellar population synthesis models is planed for a future publication ( thomas et al . , in preparation ) . in fig .
[ fig : rh_reff]a we plot @xmath11 against @xmath88 .
larger core - radii are found in more extended galaxies .
three of the four galaxies with young cores are again offset .
these three galaxies have similar @xmath88 than other coma galaxies of the same luminosity ( cf .
[ fig : rh_reff]b ) .
this makes a higher baryon concentration unlikely to be the cause of their small halo core - radii .
an exceptional case is gmp0756 : it has a small core - radius , a ratio @xmath89 which is typical for coma galaxies with old stellar populations and a relatively small @xmath88 .
the scalings of @xmath88 and @xmath11 with luminosity in coma galaxies with old stellar populations imply a roughly constant ratio @xmath90 .
moreover , disk galaxies of the same luminosity show a similar ratio .
the fact that all four galaxies offset in figs .
[ fig : rhvh ] and [ fig : rh_reff ] appear at projected cluster - centric distances @xmath91 and have young stellar populations suggests that they may have entered the coma cluster only recently .
we will come back to this point in the next sec .
[ subsec : round ] .
@xcite and @xcite studied dark matter halos of 21 nearly round ( e0-e2 ) and non - rotating galaxies ( k2000 in the following ) .
k2000 galaxies have similar luminosities @xmath92 and half - light radii @xmath88 as ours , but the k2000 sample contains a mixture of field ellipticals and galaxies in the virgo and fornax clusters .
the dynamical models of @xcite differ in some respects from the ones described in sec .
[ sec : data ] and this will be further discussed below .
however , in their mass decomposition @xcite assumed the same halo profile as in equation ( [ logdef ] ) , such that we can directly compare their halo parameters to ours ( cf .
small black dots in figs .
[ fig : rhvh ] - [ fig : rh_reff ] ) .
we find halo parameters of both samples in the same range , but coma galaxies of the same @xmath81 have on average larger halo core - radii than k2000 galaxies .
however , the halos of the k2000 galaxies themselves are not different from the ones around coma early - types , as both belong to the same one - parameter family ( cf .
[ fig : rhoh_rh ] ) .
the main difference is that k2000 galaxies are brighter ( and have higher stellar mass ) than coma early - types with a similar halo .
can this be an artifact related to differences in the dynamical models ?
many of the k2000 models are based on @xmath2-band photometry , while we used @xmath93-band images for the coma galaxies .
elliptical galaxies become bluer towards the outer parts and @xmath2-band light profiles are slightly shallower than @xmath18-band profiles .
likewise , mass profiles of galaxies are generally shallower than their light profiles , such that there might be less need for dark matter in @xmath2-band models than in @xmath18-band models .
@xcite checked for this by modelling one galaxy ( ngc3379 ) in both bands and found comparable results .
the photometric data is therefore unlikely to cause the differences between the two samples .
k2000 galaxies were modelled assuming spherical symmetry .
not all apparently round galaxies need to be intrinsically spherical . neglecting the flattening along
the line - of - sight can result in an underestimation of a galaxy s mass ( e.g. @xcite ) .
based on the average intrinsic flattening of ellipticals in the luminosity interval of interest here , @xcite estimated that the assumption of spherical symmetry should affect mass - to - light ratios only at the 10 percent level .
we expect the effect on @xmath12 to be correspondingly small .
in addition , it is not obvious why spherical symmetry should enforce systematically small @xmath11 , such that the different symmetry assumptions are also unlikely to explain the more extended cores and higher circular velocities in coma galaxy halos .
since the shape of a galaxy is related to its evolutionary history , the round and non - rotating k2000 galaxies could be intrinsically different from the mostly flattened and rotating coma galaxies .
structural differences could also be related to the fact that k2000 galaxies are located in a variety of environments , with less galaxies in high density regions like coma .
for example , stellar population models indicate that field ellipticals are on average younger and have more extended star formation histories than cluster galaxies @xcite . but
a mere difference in the stellar populations can not explain the difference between k2000 and coma galaxies , as in this case the scalings with @xmath82 should be similar in both samples .
this is ruled out by fig .
[ fig : rhvh]b / d .
noteworthy , most of the k2000 galaxies in figs .
[ fig : rhvh ] - [ fig : rh_reff ] appear similar to the four coma galaxies with distinctly small @xmath11 and @xmath12 . as it has been discussed above
, these galaxies may have entered the coma cluster only recently and in this respect are more representative for a field galaxy population rather than being genuine old cluster galaxies . a consistent explanation for both , the offset between young and old coma galaxies on the one side and the difference between old coma galaxies and k2000 galaxies on the other would then be that field galaxies have lower @xmath11 and lower @xmath12 than cluster galaxies of the same stellar mass .
because there are more field galaxies in the k2000 sample than in coma , coma galaxy halos would be expected to have on average larger cores and to be more massive ( consistent with fig .
[ fig : rhvh ] ) . a larger comparison sample of field elliptical halos is required to conclude finally upon this point .
two independent derivations of dark matter scaling relations for spiral galaxies are included in figs .
[ fig : rhvh ] - [ fig : rh_reff ] through the dotted and dashed lines .
dotted lines show scaling relations from @xcite .
they are based on maximum - disk rotation - curve decompositions with the halo density from equation ( [ logdef ] ) .
@xcite give halo core - radii scaled by the optical disk radius . to reconstruct the underlying relationship between @xmath11 and @xmath81 ,
we follow @xcite and assume exponential disks with @xmath94 ( the empirical fit in @xcite has been transformed to our distance scale ) . dashed lines in figs .
[ fig : rhvh ] - [ fig : rh_reff ] fit the combined sample of 55 rotation curve decompositions of @xcite .
these authors discuss various halo profiles , but we here only consider non - singular isothermal dark matter halos . though these are most similar to equation ( [ logdef ] ) , isothermal cores and circular velocities are not exactly identical as in logarithmic halos . to account for the difference , we fitted a logarithmic halo to a non - singular isothermal density profile ( cf . tab .
4.1 of @xcite ) .
the fit was restricted to the region with kinematic data ( typically inside two core - radii ) .
we found that the logarithmic halo fit yields a 3 percent larger core - radius @xmath11 and a 10 percent smaller circular velocity @xmath12 .
the central halo density is reproduced to 0.001 dex ( not surprising given that the cores of the two profiles were matched ) .
thus , under the assumption that fits performed with the two profiles indeed match inside two core - radii , a correction of the derived halo parameters is not needed . scaling relations from @xcite
are shown in figs .
[ fig : rhvh ] - [ fig : rhoh_rh ] without any correction .
@xcite and @xcite discuss the scaling of disk galaxy halos with @xmath2-band luminosity . in order to compare early - types and spirals also at the same stellar masses , we used @xmath95 for the stellar mass - to - light ratios of disks .
equation ( [ eq : mlspir ] ) is derived from the tully - fisher and stellar - mass tully - fisher relations of @xcite . both in luminosity and in stellar mass , spirals and ellipticals
follow similar global trends . however , while old coma early - types have halo core - radii of similar size as spirals with the same @xmath2-band luminosity , the asymptotic halo circular velocities are @xmath96 times higher than in corresponding spirals .
in contrast , early - types with young central stellar populations have about 4 times smaller core - radii than spirals , but similar asymptotic halo velocities .
when galaxies are compared at the same stellar mass , then differences between ellipticals and spirals become larger ( 60 percent smaller @xmath11 and @xmath97 times higher @xmath12 in old coma early - types compared to spirals ; 90 percent smaller @xmath11 and 20 percent smaller @xmath12 in coma galaxies with young central stellar populations ) .
in addition , the halos of early - types and spirals do not belong to the same one - parameter family ( cf .
[ fig : rhoh_rh ] ) . at a given @xmath11
dark matter densities in ellipticals are about @xmath98 dex higher than in spirals .
[ fig : rho_l ] shows scaling laws for dark matter densities .
the central dark matter density @xmath83 ( cf . equation [ logrho0def ] ) of the logarithmic halo fits is plotted in panels ( a ) and ( b ) versus luminosity and stellar mass . for panels ( c ) and
( d ) , the best - fit dark matter halo of each galaxy ( being either logarithmic or nfw ) is averaged within @xmath0 : @xmath99 here , @xmath100 equals the cumulative dark mass inside a sphere with radius @xmath101 .
average dark matter densities are quoted in column ( 2 ) of tab .
[ tab : gals2 ] .
the general trend for both densities is to decrease with increasing galaxy luminosity .
thereby the central densities @xmath83 scatter more than the averaged @xmath102 . for two reasons
, @xmath102 quantifies the actual dark matter density more robustly than @xmath83 .
firstly , our estimate of the very central dark matter density depends strongly on the assumed halo profile .
instead , averaged over @xmath0 differences between logarithmic halo fits and fits with nfw - profiles are small compared to the statistical errors ( averaged over the coma sample nfw - fits yield @xmath103 dex higher @xmath102 than fits with logarithmic halos ) .
secondly , the most significant differences between @xmath83 and @xmath102 occur in the four coma galaxies with distinct halos discussed in sec .
[ sec : scaling ] .
these galaxies have young central stellar populations .
if the bulk of stars in these galaxies is old , however , then the related radial increase of the stellar mass - to - light ratio could contribute to their small @xmath11 and large @xmath83 .
this , because in our models it is assumed that the stellar mass - to - light ratio is radially constant . by construction then , any increase in the mass - to - light ratio with radius ( being either due to a stellar population gradient or due to dark matter ) is attributed to the halo component . in galaxies with a significant increase of the stellar @xmath104 with radius ,
the halo component of the model thereby has to account for both , additional stellar and possible dark mass @xcite . any contamination with stellar mass will be largest at small radii , where the increase in the stellar @xmath104 dominates the shape of the mass profile . the averaging radius in equation ( [ avrhodef ] )
is therefore chosen as large as possible .
the value of @xmath0 is a compromise for the whole sample , because the kinematic data extend to @xmath105 and the averaging should not go much beyond the last data point . compared to the coma galaxies , the majority of k2000 galaxies have larger @xmath83 . after averaging inside @xmath0 ,
the halo densities in both samples become comparable , however . in this respect ,
the k2000 galaxies again resemble the four coma galaxies with young stellar cores .
dotted and dashed lines in fig .
[ fig : rho_l ] show spiral galaxies .
their halo densities need not to be averaged before comparison , because core sizes of spirals ( in the considered luminosity interval ) are larger than @xmath0 . averaged over the whole sample ,
we find dark matter densities in coma early - types a factor of @xmath1 higher than in spirals of the same luminosity .
if early - types are compared to spirals of the same stellar mass , then the overdensity amounts to a factor of @xmath3 .
does this imply that spirals and ellipticals of the same luminosity have formed in different dark matter halos ?
even if ellipticals and spirals would have formed in similar halos , then the final dark matter densities after the actual galaxy formation process could be different , since the baryons in ellipticals and spirals are not distributed in the same way .
this effect can be approximated as follows : assume that ( a spherical ) baryonic mass distribution @xmath106 condenses slowly out of an original halo+baryon distribution @xmath107 .
the halo responds adiabatically and contracts into the mass distribution @xmath100 .
if the original particles move on circular orbits then @xmath108 = r_i m_i ( r_i)\ ] ] turns out to be an adiabatic invariant @xcite . in case of the coma galaxies , @xmath82 and @xmath109
are known from the dynamical modeling and equation ( [ adiaeq ] ) can be solved for @xmath110c , d ) it is not always possible to find @xmath110 for logarithmic halos .
therefore , here we only consider the best - fit nfw - halo of each galaxy . ] .
it characterizes the original halo mass distribution before the actual galaxy formation . had a disk with baryonic mass @xmath111 grown in this original halo instead of an early - type then the halo contraction would have been different such that in general @xmath112 .
the difference between @xmath113 and @xmath109 actually determines how much dark matter densities of ellipticals and spirals would differ , if both had formed in the same original halos . to quantify this further ,
let s consider the spherically averaged mass distribution of a thin exponential disk for @xmath111 ( e.g. * ? ? ?
it is fully determined by a scale - radius and a mass . for a given coma elliptical with luminosity @xmath81
, the scale - radius and the mass of a realistic disk with the same luminosity can be taken from equations ( [ spireff ] ) and ( [ eq : mlspir ] ) .
then , given @xmath111 and @xmath110 ( the recontracted coma galaxy halo ) , equation ( [ adiaeq ] ) can be solved for the baryon - contracted halo @xmath113 around the comparison disk ( cf .
once @xmath113 is known , the average @xmath114 follows directly .
if ellipticals and spirals ( of the same @xmath81 ) would have formed in the same halos , then @xmath115 should fully account for the observed ratio of elliptical to spiral dark matter densities .
however , averaged over the coma sample we find @xmath116 and , thus , that the higher baryon concentration in early - types is not sufficient to explain the factor of @xmath1 between the dark matter densities of ellipticals and spirals at constant luminosity . in general
, the observed dark matter density ratio @xmath117 between ellipticals and spirals will be a combination of a difference in the halo densities before baryon infall and a factor that comes from the baryons .
let @xmath118 denote the baryon - corrected dark matter density ratio , then the simplest assumption is @xmath119 with @xmath120 from equation ( [ dbar ] ) . after applying this approximate baryon correction , dark matter densities in coma ellipticals are still a factor of @xmath121 higher than in spirals of the same luminosity . if the comparison is made at the same stellar mass , then @xmath122 .
note that our baryonic contraction corrections are likely upper limits , because in equation ( [ eq : mlspir ] ) we only account for the stellar mass in the disk . in the presence of gas
, the baryonic disk mass will be larger and so will be the halo contraction .
the dark matter density contrast relative to the original elliptical will be therefore smaller .
concluding , the differences between the baryon distributions of ellipticals and spirals are not sufficient to explain the overdensity of dark matter in ellipticals relative to spirals of the same luminosity or stellar mass .
ellipticals and spirals have not formed in the same halos .
instead , the higher dark matter density in ellipticals points to an earlier assembly redshift .
in order to evaluate the difference between elliptical and spiral galaxy assembly redshifts quantitatively , let s assume that dark matter densities scale with the mean density of the universe at the assembly epoch , i.e. @xmath123 ( we will discuss this assumption in sec .
[ subsec : sams_zzz ] ) .
let @xmath124 and @xmath125 denote the formation redshifts of ellipticals and spirals , respectively , then @xmath126 @xcite , with @xmath127 some measure of the dark matter density , e.g. @xmath128 . equation ( [ zscale ] ) can be solved for @xmath129 \times \delta^{1/3 } - 1,\ ] ] where we have omitted the dependency of dark matter densities and formation redshifts on @xmath81 and defined @xmath130 .
equation ( [ zformdef ] ) allows to calculate formation redshifts of coma ellipticals from @xmath125 and the observed @xmath131 .
two estimates based on different assumptions about @xmath131 and @xmath125 are shown in fig .
[ fig : zl ] and are further discussed below . for each case , formation redshifts were calculated with both disk halo scaling laws shown in fig .
[ fig : rho_l ] and the two results were averaged .
raw formation redshifts without any baryon correction ( @xmath132 ) are shown in fig .
[ fig : zl]a ( and listed in column 3 of tab .
[ tab : gals2 ] ) .
we considered a wide range of spiral galaxy formation redshifts @xmath133 $ ] and the related uncertainty in @xmath124 is indicated by vertical bars .
our fiducial value is @xmath134 , because regular disks become rare beyond @xmath135 @xcite .
we did not allow for a luminosity dependence of spiral galaxy formation times .
for the coma ellipticals we then find @xmath124 ranging from @xmath136 to @xmath137 , with the majority of galaxies having formed around @xmath138 .
brighter galaxies have assembled later than fainter galaxies .
coma galaxy assembly redshifts shown in fig .
[ fig : zl]b ( see also column 5 of tab . [
tab : gals2 ] ) include the baryon correction of sec .
[ sec : contraction ] , because we used @xmath139 in equation ( [ zformdef ] ) .
moreover , because fainter spirals have denser halos than brighter ones , we allowed for a luminosity dependent @xmath140
. in analogy to equation ( [ zscale ] ) assume @xmath141 and @xmath142 for a reference luminosity @xmath143 ( the dashed line in fig .
[ fig : z_sam]c illustrates the resulting @xmath125 as a function of @xmath144 ) . because @xmath145 , the baryon corrected @xmath124 in fig .
[ fig : zl]b are lower than the uncorrected ones in fig .
[ fig : zl]a .
the typical assembly redshift reduces to @xmath146 , as compared to @xmath138 without the correction .
the baryon correction is mostly smaller than the uncertainty related to our ignorance about @xmath125 ( vertical bars in fig .
[ fig : zl]b correspond to @xmath147 $ ] ) .
the trend for lower @xmath124 in brighter galaxies is slightly diminished by the baryon correction such that the dependency of @xmath124 on @xmath144 in fig .
[ fig : zl]b mainly reflects the luminosity dependence of spiral galaxy assembly redshifts .
[ fig : zz ] compares halo assembly redshifts with central stellar population ages @xmath148 from @xcite .
( we use @xmath149 , @xmath150 and @xmath151 to transform ages into redshifts . )
largely independent from applying the baryon correction or not , the agreement between the two redshifts is fairly good for about half of our sample . among the remaining galaxies , some have halos which appear younger than their central stellar populations
this could indicate that the stellar ages are overestimated ( they are sometimes larger than the age of the universe in the adopted cosmology ; cf .
[ tab : gals2 ] ) .
it could also point at these galaxies having grown by dry merging . in a dry merger ,
the dark matter density can drop , but the stellar ages stay constant . in coma galaxies with young stellar cores ,
the halo assembly redshifts are instead larger than the central stellar ages .
this indicates some secondary star - formation after the main epoch of halo assembly .
in the following we will compare our results to semi - analytic galaxy formation models . to this end
we have constructed a comparison sample of synthetic ellipticals and spirals using the models of @xcite , which are based on the millennium simulation @xcite .
comparison ellipticals are selected to rest in dark matter cluster structures with virial masses larger than @xmath152 and to obey @xmath153 @xcite .
we ignore galaxies at the centers of simulated clusters since we have omitted the two central coma galaxies from the analysis in this paper . likewise ,
we exclude from the comparison galaxies that have been stripped - off their entire halo , because the only coma galaxy that possibly lacks dark matter inside @xmath154 has been excluded from the analysis in this paper as well ( cf . sec .
[ sec : data ] ) .
isolated field spirals are drawn from objects with @xmath155 in the semi - analytic models @xcite .
simulated galaxies were chosen randomly from the catalogue of @xcite in a way such that each of six luminosity intervals ( between @xmath156 and @xmath157 ; width @xmath158 ) contains roughly 50 galaxies .
we use dust - corrected luminosities @xmath92 of the semi - analytic models .
dark matter halos of simulated galaxies are reconstructed from tabulated virial velocities @xmath159 , virial radii @xmath160 , and maximum circular velocities @xmath161 as follows .
it is assumed that the halos can be approximated by an nfw - profile ( cf .
equation [ nfwdef ] ) , in which case the circular velocity profile reads @xmath162 here @xmath163 and the halo concentration is defined by @xmath164 .
the maximum circular velocity @xmath161 of an nfw halo occurs at @xmath165 @xcite , such that ( with equation [ nfwcirc ] ) @xmath166 using the tabulated @xmath159 and @xmath161 this equation can be numerically solved for the halo concentration @xmath167 , which in turn determines @xmath168 and , thus , the entire nfw profile of the halo .
before compared to the coma galaxy models , halo densities are averaged within @xmath0 ( cf .
equation [ avrhodef ] ) . in case of simulated spirals
we use effective radii from the empirical relation ( [ spireff ] ) . for ellipticals ,
we assume @xmath169 which is a fit to the coma data .
[ fig : dm_sam ] shows that the average dark matter densities @xmath102 of the coma early - types match fairly well with semi - analytical models .
this is remarkable , because the simulations do not take into account the halo response during baryon infall .
therefore , either the net effect of the baryons on the dark matter distribution is small in the analyzed population of galaxies or there is actually a mismatch between the halos of observed galaxies and the @xmath5-body models .
it may also be that real galaxies do not have maximum stellar masses .
this can be checked by the comparison of dynamically derived stellar mass - to - light ratios with independent stellar population synthesis models ( thomas et al .
, in preparation ) . similarly to what is found in real galaxies ,
the dark matter densities of spirals are lower than in ellipticals in semi - analytic models ( cf .
[ fig : dm_sam ] ) , but the density contrast in observed galaxies is larger .
again , a major uncertainty here is that the simulations do not take into account the gravitational effect of the baryons .
formation redshifts of simulated and observed galaxies are compared in fig .
[ fig : z_sam ] .
coma galaxy @xmath170 are from sec .
[ sec : assembly ] and both cases discussed there with and without baryon correction are shown separately in panels ( a ) and ( b ) , respectively .
formation redshifts of simulated galaxies are defined as the earliest redshift , when a halo has assembled 50 percent of its mass .
since we are mainly interested in cluster ellipticals , we need to take into account that interactions between the cluster halo and a galaxy s subhalo cause a mass - loss in the latter .
although cluster - galaxy interactions happen in both simulated and observed galaxies , the mass - loss in the simulations may be overestimated because of the finite numerical resolution and the neglect of the baryon potential .
in particular , for simulated subhalos with very low masses at @xmath171 the derived formation redshifts may be artificially high , when defined according to the assembly of half of the final mass . to avoid such artificially large assembly redshifts , we define @xmath170 of simulated galaxies as the earliest time when half of the maximum mass was assembled , that a single progenitor in the merger tree of given galaxy had at some redshift .
our assumption is that even if dynamical interactions between cluster and galaxy halos take place , they do not significantly affect the very inner regions @xmath172 of interest here . in case of field spirals ,
formation redshifts defined either from the final or from the maximum mass are very similar . without a baryon correction ,
our estimates of coma galaxy formation redshifts are on average higher than in the semi - analytic models ( fig .
[ fig : z_sam]a ) .
this , although ( 1 ) the dark matter densities of ellipticals match with the simulations and ( 2 ) our assumption about the formation redshifts of spirals ( @xmath173 ) is consistent with the simulations .
the origin for the offset between coma galaxies and semi - analytic models in fig .
[ fig : z_sam]a is that the density contrast between halos of ellipticals and spirals is larger in observed galaxies than in the simulations . after applying the baryon correction ,
the coma galaxy formation redshifts become consistent with the simulations ( fig .
[ fig : z_sam]b ) .
this result indicates that the discrepancy between the measured and the simulated density ratio @xmath174 is due to baryon effects .
our coma galaxy formation redshifts are based on the assumption that @xmath175 .
[ fig : rho_zf_sam ] shows @xmath102 versus @xmath176 explicitly .
independent of including a baryon correction or not , the slope of the relationship between @xmath102 and @xmath176 in the coma galaxies is roughly parallel to simulated @xmath5-body halos .
this confirms that our assumption for the scaling between @xmath102 and @xmath170 is approximately consistent with the cosmological simulations .
concerning the absolute values of the dark matter densities it has been already stated above that they are only consistent with the simulations if either the net effect of the baryons is zero in the case of ellipticals or if galaxies do not have maximum stellar masses . the former case would imply that halos of spiral galaxies experience a net expansion during the baryon infall ( several processes have been proposed for this , e.g. @xcite ) .
@xcite quote a stellar assembly redshift below @xmath177 for simulated ellipticals more massive than @xmath178 .
the halo assembly redshifts in fig .
[ fig : z_sam ] are mostly above @xmath179 . in part , this is due to the fact that we only consider semi - analytic galaxies in high - density environments similar to coma .
in addition , formation redshifts defined according to the stellar mass assembly and the halo assembly , respectively , are not always equal .
for example , in our comparison sample of simulated cluster ellipticals we find an average dark halo assembly redshift @xmath180 for galaxies more massive than @xmath178 .
evaluating for the same galaxies the redshift @xmath181 ( when half the stellar mass is assembled ) yields @xmath182 . that @xmath183 is plausible if some star formation is going on between @xmath184 in the progenitor and/or in the subunits that are to be accreted after @xmath170 .
it should also be noted that the simulations do not take into account stellar mass - loss due to tidal interactions .
we have presented dark matter scaling relations derived from axisymmetric , orbit - based dynamical models of flattened and rotating as well as non - rotating coma early - type galaxies .
dark matter halos in these galaxies follow similar trends with luminosity as in spirals . thereby , the majority of coma early - types those with old stellar populations have halo core - radii @xmath11 similar as in spirals with the same @xmath2-band luminosity , but their asymptotic halo velocities are about @xmath96 times higher .
in contrast , four coma early - types with young central stellar populations have halo velocities of the same order as in comparably bright spirals , but their core - radii are smaller by a factor of 4
. differences between spirals and ellipticals increase , when the comparison is made at the same stellar mass .
the average halo density inside @xmath0 exceeds that of comparably bright spirals by about a factor of @xmath1 .
if the higher baryon concentration in ellipticals is taken into account , the excess density reduces to about a factor of 3 , but if ellipticals and spirals are compared at the same stellar mass , then it is again of the order of @xmath185 .
our measured dark matter densities match with a comparison sample of simulated cluster ellipticals constructed from the semi - analytic galaxy formation models of @xcite .
these synthetic ellipticals have @xmath186 and higher dark matter densities than simulated field spirals , which are appear on average around @xmath173 .
assuming for local spirals @xmath187 as well , and assuming further that the inner dark matter density scales with the formation redshift like @xmath188 , our results imply that ellipticals have formed @xmath189 earlier than spirals . without baryon correction
, we find an average formation redshift around @xmath190 , which is slightly larger than in semi - analytic galaxy formation models .
accounting for the more concentrated baryons in ellipticals , the average formation redshift drops to @xmath191 . for about half of our sample ,
dark halo formation redshifts match with constraints derived from stellar populations @xcite : the assembly epoch of these ( old ) early - types coincides with the epoch of formation of their stellar components .
we thank ortwin gerhard and the anonymous referee for comments and suggestions that helped to improve the manuscript .
jt acknowledges financial support by the sonderforschungsbereich 375 astro - teilchenphysik of the deutsche forschungsgemeinschaft .
emc receives support from grant cpda068415/06 by padua university . the millennium simulation databases used in this paper and the web application providing online access to them were constructed as part of the activities of the german astrophysical virtual observatory .
the best - fit model parameters for the galaxies gmp3414 and gmp4822 ( which were not included in the original sample of @xcite ) are given in tab .
[ tab:3414 ] .
the table is similar to tab .
2 of @xcite and we refer the reader to this paper , in case more detailed information about the parameter definitions are required .
the best - fit models with and without dark matter halo are compared to the observations in figs .
[ fig:3414 ] and [ fig:4822 ] . in both galaxies
the best - fit inclination is @xmath192 , but the 68 percent confidence regions include models at @xmath193 ( gmp3414 ) and @xmath194 ( gmp4822 ) .
gmp3414 and gmp4822 were observed with the wide field planetary camera 2 ( wfpc2 ) on board the _ hst _ as part of the _ hst _ proposal 10844 ( pi : g. wegner ) . for each galaxy
two exposures with 300s each were taken with the filter f622w .
four other objects were previously observed as part of this proposal and a full description of the respective observational parameters and the data analysis is given in @xcite .
a table with the final photometric parameters will be published in the journal version of this paper .
99 bell , e. f. , & de jong , r. s. 2001 , , 550 , 212 binney , j. , & tremaine , s. 1987 , galactic dynamics ( princeton : princeton university press ) binney , j. , gerhard , o. , & silk , j. 2001 , , 321 , 471 blumenthal , g. r. , faber , s. m. , flores , r. , primack , j. r. 1986 , , 301 , 27 cappellari , m. , et al .
2006 , , 366 , 1126 chanam , j. , kleyna , j. , & van der marel , r. 2008 , , 682 , 841 conselice , c. j. , blackburne , j. a. , & papovich , c. 2005 , , 620 , 564 corsini , e. m. , wegner , g. , saglia , r. p. , thomas , j. , bender , r. , thomas , d. 2008 , , 175 , 462 cretton , n. , de zeeuw , p. t. , van der marel , r. p. , & rix , h .- w . 1999 , , 124 , 383 de lorenzi , f. , debattista , v. p. , gerhard , o. , & sambhus , n. 2007 , , 376 , 71 de lucia , g. , springel , v. , white , s. d. m. , croton , d. , kauffmann , g. 2006 , , 366 , 499 de lucia , g. , & blaizot , j. 2007 , , 375 , 2 gebhardt , k. , et al .
2000 , , 119 , 1157 gebhardt , k. , et al .
2003 , , 583 , 92 gerhard , o. , jeske , g. , saglia , r. p. , & bender , r. 1998 , , 295 , 197 gerhard , o. e. , kronawitter , a. , saglia , r. p. , bender , r. 2001 , , 121 , 1936 godwin , j. g. , metcalfe , n. , & peach , j. v. 1983 , , 202 , 113 gunn , j. e. , & gott , j. r. i. 1972 , , 176 , 1 jrgensen , i. , franx , m. , & kjrgard , p. 1995
, , 273 , 1097 jrgensen , i. , franx , m. , & kjrgard , p. 1996
, , 280 , 167 kormendy , j. , & freeman , k. c. 2004 , in iau symp .
220 , dark matter in galaxies , ed .
s. d. ryder et al .
( san francisco : asp ) , 377 kronawitter , a. , saglia , r. p. , gerhard , o.
e. , bender , r. 2000 , , 144 , 53 mehlert , d. , saglia , r. p. , bender , r. , wegner , g. 2000 , , 141 , 449 mehlert , d. , thomas , d. , saglia , r. p. , bender , r. , wegner , g. 2003 , , 407 , 423 navarro , j. f. , frenk , c. s. , & white , s. d. m. 1996 , , 462 , 563 persic , m. , salucci , p. , & stel , f. 1996a , , 281 , 27 persic , m. , salucci , p. , & stel , f. 1996b , , 283 , 1102 press , w. h. , teukoslky , s. a. , vetterling , w. t. , flannery , b. p. 1992
, numerical recipes in fortran 77 , 2nd edn .
cambridge univ . press , cambridge saglia , r. p. , burstein , d. , baggley , g. , davies , r. l. , bertschinger , e. , colless , m. m. , mcmahan , r. k. , jr . , & wegner , g. 1997 , , 292 , 499 schwarzschild , m. 1979 , , 232 , 236 simien , f. , & de vaucouleurs , g. 1986 , , 302 , 564 springel , v. , et al .
2005 , , 435 , 629 terlevich , a. i. , & forbes , d. a. 2002 , , 330 , 547 thomas , d. , maraston , c. , bender , r. , mendes de oliveira , c. 2005a , , 621 , 673 thomas , j. , saglia , r. p. , bender , r. , thomas , d. , gebhardt , k. , magorrian , j. , richstone , d. 2004 , , 353 , 391 thomas , j. , saglia , r. p. , bender , r. , thomas , d. , gebhardt , k. , magorrian , j. , corsini , e. m. , wegner , g. 2005b , , 360 , 1355 thomas , j. 2006 , ph.d .
thesis ( univ . of munich ) thomas , j. , saglia , r. p. , bender , r. , thomas , d. , gebhardt , k. , magorrian , j. , corsini , e. m. , wegner , g. 2007 , , 382 , 657 thomas , j. , jesseit , r. , naab , t. , saglia , r. p. , burkert , a. , & bender , r. 2007 , , 381 , 1672 trager , s. c. , faber , s. m. , worthey , g. , & gonzlez , j. j. 2000 , , 119 , 1645 wechsler , r. h. , bullock , j. s. , primack , j. r. , kravtsov , a. v. , & dekel , a. 2002 , , 568 , 52 valluri , m. , merritt , d. , & emsellem , e. 2004 , , 602 , 66 wegner , g. , corsini , e. m. , saglia , r. p. , bender , r. , merkl , d. , thomas , d. , thomas , j. , mehlert , d. 2002 , , 395 , 753 van den bosch , r. c. e. , van de ven , g. , verolme , e. k. , cappellari , m. , & de zeeuw , p. t. 2008 , , 385 , 647 lcccccccc + @xmath195 & @xmath196 & @xmath197 & @xmath198 & 1.24 & 0.35 & 0.39 & 0.010 & + @xmath195 & @xmath199 & @xmath200 & @xmath201 & 0.85 & 0.31 & 0.39 & 0.002 & + @xmath202 & @xmath196 & @xmath203 & @xmath204 & 1.62 & 0.19 & 0.21 & 0.109 & + @xmath202 & @xmath199 & @xmath205 & @xmath206 & 1.04 & 0.17 & 0.21 & 0.002 & + @xmath207 & @xmath195 & @xmath208 & @xmath209 & 0.41 & 0.41 & 0.66 & 0.001 & + @xmath195 & @xmath210 & @xmath211 & @xmath212 & 1.15 & 0.31 & 0.39 & 0.026 & + @xmath207 & @xmath196 & @xmath213 & @xmath214 & 0.75 & 0.47 & 0.66 & 0.019 & [ fig : rho_l]a + @xmath207 & @xmath199 & @xmath215 & @xmath216 & 0.94 & 0.52 & 0.66 & 0.058 & [ fig : rho_l]b + @xmath217 & @xmath196 & @xmath218 & @xmath219 & 1.31 & 0.38 & 0.32 & 0.004 & [ fig : rho_l]c + @xmath217 & @xmath199 & @xmath220 & @xmath221 & 1.79 & 0.44 & 0.32 & 0.025 & [ fig : rho_l]d + + + @xmath195 & @xmath196 & @xmath222 & @xmath223 & 0.16 & 0.16 & 0.44 & 0.002&[fig : rhvh]a + @xmath195 & @xmath199 & @xmath224 & @xmath225 & 0.18 & 0.17 & 0.44 & 0.007&[fig : rhvh]b + @xmath202 & @xmath196 & @xmath226 & @xmath227 & 0.14 & 0.07 & 0.24 & 0.016&[fig : rhvh]c + @xmath202 & @xmath199 & @xmath228 & @xmath229 & 0.12 & 0.07 & 0.24 & 0.008&[fig : rhvh]d + @xmath207 & @xmath195 & @xmath230 & @xmath231 & 0.04 & 0.18 & 0.75 & 0.006&[fig : rhoh_rh ] + @xmath195 & @xmath210 & @xmath232 & @xmath233 & 0.10 & 0.14 & 0.44 & 0.017&[fig : rh_reff ] + @xmath207 & @xmath196 & @xmath234 & @xmath235 & 0.31 & 0.34 & 0.75 & 0.028 & + @xmath207 & @xmath199 & @xmath236 & @xmath237 & 0.31 & 0.36 & 0.75 & 0.090 & + @xmath217 & @xmath196 & @xmath238 & @xmath239 & 0.77 & 0.31 & 0.41 & 0.012 & + @xmath217 & @xmath199 & @xmath240 & @xmath241 & 0.86 & 0.33 & 0.41 & 0.020 & + lcccccc 0144 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath247 + 0282 & @xmath248 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 + 0756 & @xmath254 & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 + 1176 & @xmath260 & @xmath261 & @xmath262 & @xmath263 & @xmath264 & @xmath265 + 1750 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & @xmath271 + 2417 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 + 2440 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 + 3414 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & @xmath289 + 3510 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 + 3792 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath300 & @xmath301 + 3958 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & & + 4822 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & @xmath310 & @xmath311 + 4928 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath317 + 5279 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 + 5568 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 + 5975 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 + lccccccc gmp3414 & sc & @xmath336 & & & & & @xmath337 + & log & @xmath338 & @xmath339 & @xmath340 & & & @xmath341 + & nfw & @xmath342 & & & @xmath343 & @xmath344 & @xmath345 + gmp4822 & sc & @xmath185 & & & & & @xmath346 + & log & @xmath347 & @xmath348 & @xmath349 & & & @xmath350 + & nfw & @xmath351 & & & @xmath352 & @xmath344 & @xmath353 | axisymmetric , orbit - based dynamical models are used to derive dark matter scaling relations for coma early - type galaxies . from faint to bright galaxies halo core - radii and asymptotic circular velocities increase .
compared to spirals of the same brightness , the majority of coma early - types those with old stellar populations have similar halo core - radii but more than 2 times larger asymptotic halo velocities .
the average dark matter density inside @xmath0 decreases with increasing luminosity and is @xmath1 times larger than in disk galaxies of the same @xmath2-band luminosity .
compared at the same stellar mass , dark matter densities in ellipticals are @xmath3 times higher than in spirals .
different baryon concentrations in ellipticals and spirals can not explain the higher dark matter density in ellipticals . instead , the assembly redshift ( 1+z ) of coma early - type halos is likely about two times larger than of comparably bright spirals .
assuming that local spirals typically assemble at a redshift of one , the majority of bright coma early - type galaxy halos must have formed around @xmath4 . for about half of our coma galaxies the assembly redshifts match with constraints derived from stellar populations .
we find dark matter densities and estimated assembly redshifts of our observed coma galaxies in reasonable agreement with recent semi - analytic galaxy formation models . |
most cataclysmic variables ( cvs ) consist of a white dwarf accreting material from a late - type main sequence star via roche lobe overflow .
however , in some cvs , the mass donor is hydrogen deficient - either a semi - degenerate star or a white dwarf .
these binaries are called am cvn systems or ultra - compact binaries since they have orbital periods less than @xmath070 mins .
currently there are 17 confirmed systems with another 2 candidate systems .
this paper presents : the main results of a survey of these systems using _ xmm - newton _ ; an initial study of the high resolution x - ray spectrum of rx j1914 + 24 ; the results of a search for radio emission from rx j0806 + 15 and rx j1914 + 24 , and we also present initial results of the rapid temporal survey ( rats ) project whose aim is to detect new am cvn systems .
until the launch of _ xmm - newton _ , am cvn systems had been little studied in x - rays - only 3 of the then known systems were detected in the _ rosat _ all - sky survey . with the greater sensitivity of the epic x - ray detectors on board _ xmm - newton , _ coupled with the optical / uv telescope ( the om ) it was possible to study these systems in way that was not previously possible .
one of our goals was to search for evidence for a coherent modulation in the x - ray intensity of these systems .
of the 7 systems which we have so far observed ( excluding rx j0806 + 15 and rx j1914 + 24 ) , none show any evidence for pulsations in their light curve .
this suggests that we have not detected the spin period of the accreting white dwarf , and implies that they do not have strong magnetic fields .
the sources did , however , show variability ( cf figure 1 ) .
a second goal was to determine the nature of their x - ray emission .
we found that they were best modelled using optically thin thermal plasma models with highly non - solar abundances . on top of their hydrogen deficiency , many systems showed significant nitrogen over abundances .
our third goal was to determine their energy balance for the first time .
we showed that a large fraction of their accretion energy was emitted at uv wavelengths .
in particular , the uv luminosity was greatest at shorter periods where the accretion disc dominates .
moreover , with parallaxes for these systems now available ( roelofs et al 2006 ) we showed that their observed luminosity agrees remarkably well with the accretion luminosity predicted by gravitational wave losses .
our results are presented in detail in ramsay et al.(2005b , 2006c ) .
( 6,7.5 ) ( -3,-0.7 )
the x - ray sources rx j0806 + 15 ( hm cnc ) and rx j1914 + 24 ( v407 vul ) have been the subject of much interest .
if their periods ( 321 sec and 569 sec respectively ) can be confirmed as their orbital period , then their orbital periods will be the shortest among known binary systems .
here we concentrate on two aspects - the nature of the x - ray spectrum of rx j1914 + 24 and a search for radio emission from rx j0806 + 15 and rx j1914 + 24 .
the x - ray spectrum of rx j1914 + 24 is complex
. it can not be well fit using a simple absorbed blackbody or absorbed optically thin thermal plasma model . rather , other model components ( such as an absorption edge or gaussian line ) have to be added to achieve good fits to the _ xmm - newton _ epic spectra ( ramsay et al . 2005a , 2006a ) .
it is not clear if these extra components are physically realistic . to investigate this in more detail
we have extracted a spectrum of rx j1914 + 24 using the reflection grating spectrometer detectors on - board _ xmm - newton_. we summed up the data from both rgs detectors and from data taken in _ xmm - newton _ orbits 0718 , 0721 and 0880 .
( rgs data from orbit 0882 has various issues which we are still investigating ) .
we show the spectrum in figure [ rx1914-spec ] .
we find that an absorbed thermal plasma model plus broad gaussian feature centered at 0.68 kev gives a significantly better fit than either a model which replaces the gaussian component with an edge or a model consisting of a blackbody ( plus gaussian or edge ) .
determining if rx j1914 + 24 has an optically thick or thin x - ray spectrum has important consequences for its luminosity .
ramsay et al.(2006a ) show that the optically thin emission model gives a luminosity 23 orders of magnitude less than the optically thick based models .
if the emission layer is hotter than the optically thick atmosphere beneath , then the spectrum would be expected to show emission lines .
there is some evidence for line features in figure [ rx1914-spec ] .
however , only a spectrum with a longer exposure would be able to verify the presence of such lines .
the physical origin of the feature , to which we have fitted a gaussian , is not clear .
could it be some feature produced by a high magnetic field ?
for instance broad absorption ( rather than emission ) features in x - ray spectra of isolated neutron stars have been interpreted as due to cyclotron features from which magnetic fields of 10@xmath1 g have been inferred ( eg haberl 2004 ) . the unipolar - inductor model ( wu et al .
2002 ) predicts that radio emission should be produced due to electron - cyclotron maser emission ( willes & wu 2004 ; willes , wu & kuncic 2004 ) .
if we observe the system at an appropriate viewing which should detect strongly circularly polarised radio emission .
a search was made using the vla in sept 2005 to detect emission from both rx j0806 + 15 and rx j1914 + 24 at 6 cm .
each observation lasted a total of 3 hrs .
we did not detect rx j1914 + 24 ( 3@xmath2 upper limit of 42@xmath3 jy / beam ) , but detected rx j0806 + 15 at a level of 96@xmath3 jy ( a 5.0@xmath2 detection ) . for comparison we did not detect es cet ( an accreting am cvn system with @xmath4=10.4 min ) using atca at 6 cm ( 3@xmath2 upper limit of 74@xmath3 jy / beam ) we have further observations of rx j0806 + 15 approved with the vla .
( 6,6.5 ) ( -2.6,-1.2 )
less than 20 am cvn systems are currently known .
however , models predict that many more systems should be easily detectable using current instrumentation . for instance , nelemans et al .
( 2004 ) predict that nearly 4000 systems with periods shorter than 25 mins should be brighter than @xmath5=22 .
these shorter period systems show modulation periods in the optical band up to 0.1 - 0.2 mag .
there are a number of projects whose aim is to discover new am cvn systems .
the strategy of the rapid temporal survey ( rats ) is to use wide field cameras on medium sized telescopes to detect stellar sources which vary in their optical intensity on periods shorter than @xmath080 min . a series of short ( @xmath030 sec )
exposures are obtained for a given field over the course of 23 hrs ( ramsay & hakala 2005 ) .
candidate systems are then subject to followup spectroscopy to determine their nature . our pilot study was obtained using the isaac newton telescope on la palma with the wide field camera .
we obtained data of 12 fields which gave a coverage of 3 square degrees .
a total of 46 objects were found to vary in their optical brightness .
details of our analysis and results are given in ramsay & hakala ( 2005 ) .
many of the objects were contact binaries ( also known as w uma stars ) . however , 4 sources showed variability with periods on timescales less than @xmath070 mins .
followup spectroscopy and photometry revealed their nature .
three of these systems are sx phe or dwarf cepheid stars - only @xmath020 of these pulsating stars were previously known outside stellar clusters .
the object showing the shortest period , 374 sec , is a rare pulsating sdb star ( also known as ec 14027 stars ) .
its high pulsation amplitude is consistent with its location at the cool end of the ec 14026 instability strip ( see ramsay et al .
2006b for details ) .
since our pilot survey we have obtained data covering 12 fields using the eso 2.2 m telescope in chile and 8 fields using the int in la palma , both in june 2005 .
our analysis of the data from eso is complete . excluding those fields which included globular clusters , we found over 150 new variable objects .
similar to our pilot survey , many of these are contact binaries ( based on their light curves ) and around 10% of these variables showed modulation periods of less than @xmath080 min ( we show the light curves of two such systems in figure [ eso - rats ] ) . a similar number of systems were found to be eclipsing systems .
we have spectra of the brighter systems using the saao 1.9 m telescope .
it is also of interest to determine the number of ultra - compacts located in globular clusters
. this would give insight into how their formation process was sensitive to the density of their environment .
we obtained data of m4 and m22 using the 2.2 m telescope in la silla , and m71 using the int .
we had to adapt our software to suit the more crowded fields . however , the cluster centers ( up to the core radius ) were too crowded for aperture photometry to perform well a future study will develop tools which utilises profile fitting photometry .
we found a total of 24 , 70 and 80 variable sources in clusters m4 , m22 and m71 respectively .
a significant proportion of the sources are within the cluster tidal radius and therefore may be cluster members . we have discovered more than 40 eclipsing systems and more than 10 systems which show variations on periods less than 60 min . we have obtained spectra of the variable sources in m4 and m22 using aaomega at the anglo australian telescope in the summer of 2006 . in the red arm we used the 385r grism and in the blue the 580v grism .
the seeing was moderate to poor
so we did not obtain useful spectra for the fainter targets .
there was no evidence for any ultra - compact binaries in either cluster .
however , we did detect emission lines in one system , rat j162324262622 which showed h@xmath6-h@xmath7 in emission along with emission lines of forbidden n ii on the blue and red side of h@xmath6 ( figure [ pn - spec ] ) .
such a spectrum is more similar to planatary nebulae .
rat j162325 - 262911 shows emission in the core of an absorption line at h@xmath6 , while rat j162342 - 264255 shows deep balmer absorption lines .
so far , we have not found any am cvn systems in the rats survey . at face value
this suggests that the models of nelemans et al .
( 2001 , 2004 ) overestimate the space density of am cvn s .
however , the size of our survey is relatively small ( 6 square degrees have been currently analysed in full ) and secondly , the models predict that they should be concentrated in the galactic plane .
future field selections will be strongly biased towards the galactic plane ( @xmath8 ) .
however , we have shown that our strategy is a good means for discovering rare pulsating systems , some of which may eventually become white dwarfs .
haberl , f. , 2004 , mem .
s.a . it , 75 , 454 nelemans , g. , portegies zwart , s. f. , verbunt , f. , & yungelson , l. , 2001 , a&a , 368 , 939 nelemans , g. , yungelson , l. , & portegies zwart , s. , 2004 , mnras , 349 , 181 ramsay , g , & hakala , p. , 2005 ,
mnras , 360 , 314 ramsay , g , hakala , p. , & wu , k. , et al , 2005a , mnras , 357 , 49 ramsay , g , hakala , p. , marsh , t. , & nelemans , g. , et al , 2005b , a&a , 440 , 675 ramsay , g. , cropper , m. , & hakala , p. , 2006a ,
mnras , 367 , l62 ramsay , g , napiwotzki , r. , hakala , p. , & lehto , h. , 2006b , mnras , 371 , 957 ramsay , g , groot , p. , & marsh , t. , et al , 2006c , a&a , 457 , 623 , a&a roelofs , g. , et al , submitted , apj wu , k. , cropper , m. , ramsay , g. , & sekiguchi , k. , 2002 , mnras , 331 , 221 willes , a. , & wu , k. , 2004 , mnras , 348 , 285 willes , a. , wu , k. , & kuncic , z. , 2004 , pasa,21 , 248 | we present the results of some recent research on am cvn systems .
we present : x - ray / uv observations made using _ xmm - newton _ ; the x - ray grating spectrum of rx j1914 + 24 ; preliminary results of a search for radio emission from am cvn binaries , and discuss the strategy and first results of the rats project , whose main aim is to discover am cvn systems . |
the quench of systems , like ferromagnets or binary mixtures , to below the critical point is characterized by the formation and subsequent growth of domains .
the typical feature of the phase ordering process is dynamical scaling , whereby physical properties become time - independent if lengths are measured in units of @xmath3 , i.e. the typical domain size at time @xmath4 @xcite .
this scaling scenario is expected to be valid also in presence of quenched disorder when it does not prevent the phase - ordering process @xcite . in the case of pure systems power law growth @xmath5
is well established .
conversely , when disorder is present the nature of asymptotic growht has been much debated @xcite . in general terms ,
disorder generated energy barriers bring about a slowing down of the coarsening process @xcite and , as a rule of thumb , more disorder means slower growth .
however , recent studies have shown that the growth low does not behave monotonically with increasing disorder in the two dimensional site or bond diluted ising model @xcite . in these systems , where a fraction @xmath0 of sites or bonds is randomly removed from a regular lattice , one observes that , for sufficiently large @xmath0 , more disorder produces faster growth .
more precisely , starting from the pure case with @xmath6 , where the usual temperature - independent power - law @xmath7 is obeyed , one enters a region with an asymptotic logarithmic behavior as soon as @xmath8 .
however a temperature - dependent power - law growth is observed right at @xmath9 , where @xmath10 is the critical percolation density , thus leading to a non - monotonic dependence of the speed of growth upon the disorder strength @xmath0 .
this counterintuitive behavior has been interpreted in terms of the topology of the network of occupied sites or bonds @xcite with the temperature - dependent algebraic behavior associated to the fractal properties of the percolating network . in this article , by studying numerically the site - diluted ising model ( sdim ) and the bond diluted one ( bdim ) in @xmath11 we find a non - monotonous dependence of @xmath3 on disorder qualitatively similar to the two - dimensional case .
this suggests that the relation between growth - law and topology might be a rather generic property , a fact that could provide new insights into the problem of phase - separation in disordered ferromagnets .
this paper is organized as follows . in sec .
[ models ] we describe the models , the sdim and the bdim , that will be considered and studied throughout this paper . in sec .
[ gl ] we present an overview of the results in @xmath12 whereas numerical simulations of the @xmath11 models are discussed in sec .
[ 3d ] . in sec .
[ concl ] we conclude the paper with a final discussion of the results and of some open issues .
in the sdim a network is prepared by generating configurations of occupied sites as in random percolation . on the sites of a square three - dimensional lattice
there are independent random variables @xmath13 , which take values @xmath14 ( occupied site ) with probability @xmath15 and @xmath16 ( empty site ) with probability @xmath17 .
the substrate is formed by the set of occupied sites .
any couple of nearest - neighbour occupied sites is connected by a _
bond_. in the bdim , instead , starting from a regular square three - dimensional lattice one removes bonds between nearest - neighbouring sites @xmath18 at random with probability @xmath0 .
specifically , an independent random variable @xmath19 is defined on each couple of nearest sites , which takes values @xmath20 with probability @xmath0 ( bond is absent ) and @xmath21 with probability @xmath22 ( bond is present ) . for both models ,
@xmath0 will be referred to as dilution .
we now describe the geometrical structure of the network as @xmath0 goes from high to low values .
recalling that the percolation threshold is @xmath23 and @xmath24 for site and bond percolation respectively , for @xmath25 , the network is a patchwork of finite clusters of occupied sites ( bonds ) with a characteristic size that diverges at the threshold and a fractal dimension @xmath26 @xcite .
the dilution range @xmath27 will not be considered in this paper since , lacking an infinite cluster , perpetual coarsening can not be sustained . at @xmath28 ,
the size of the percolating cluster of sites ( bonds ) diverges .
as @xmath0 is lowered below @xmath28 , the infinite cluster has a fractal geometry over distances up to a typical length @xmath29 that diverges at the threshold as ( d)~(d_c - d)^- , [ divxi ] with @xmath30 , whereas it becomes compact over larger distances .
moreover , clusters of finite size are also present .
the structure of the network described above holds in the range @xmath31 , where @xmath2 is a dilution value where ( d^ * ) = a , [ sub.3 ] and @xmath32 is a microscopic length , like the lattice spacing . for @xmath33 ,
the infinite cluster is compact on any lengthscale and finite clusters are absent .
the remaining far - apart vacancies inside this cluster are essentially single - site objects .
their average distance defines a new characteristic length ( d ) = a(d^*/d)^. [ sub.4 ] the diluted ising model is described by the following hamiltonian h=-_<ij > k_ij_i _ j. [ ham ] here @xmath34 are spins on the 3-d square lattice , and the couplings @xmath35 for the sdim is chosen as @xmath36 , where @xmath37 is a constant , and @xmath16 or @xmath14 with probability @xmath0 and @xmath22 , respectively . for the bdim , instead , @xmath14 on all sites and @xmath38 or @xmath39 with probability @xmath0 and @xmath22 , respectively .
notice that , since @xmath40 , the model is ferromagnetic .
the equilibrium phase - diagram of the model is pictorially represented in fig .
[ fig_tc ] . for @xmath41 , at low enough temperature the system exhibits ferromagnetic order . in the @xmath42 plane
there is a critical line @xmath43 , which separates the paramagnetic from the ferromagnetic phase .
the critical temperature @xmath43 , which in the following will be measured in units of @xmath44 , where @xmath45 is the boltzmann constant , decreases from the pure value ( @xmath46 ) as the dilution is increased and vanishes at @xmath28 ( @xmath47 ) .
non - conserved dynamics@xcite is implemented by evolving the spins with single - spin - flip transition rates of the glauber form w(_i -_i)= .
[ transr ] here , @xmath48 is the local weiss field obtained by the sum h_i^w = _ j l_i k_ij _ j over the set of nearest - neighbors @xmath49 of @xmath50 .
the system is initially prepared in the infinite - temperature disordered state and , at the time @xmath51 , it is suddenly quenched to a finite temperature @xmath52 .
this temperature is set in the simulations to a value @xmath53 for the sdim and @xmath54 for the bdim .
these values are chosen as a compromise : on the one hand we want to work in the limit of low - temperatures because the interplay between the two types of growth - law we are interested in is expected in the @xmath55 limit @xcite . on the other hand
very low temperatures are useless in simulations for at least a couple of reasons .
the first is that at very low temperatures pinning effects make the dynamics very slow . as a result
the growing length @xmath56 does not vary appreciably in the time range accessed by the simulations and it is difficult to discriminate between different growth - laws .
the second related reason is that at low temperatures the pinning centres produce a stop and go behavior of interfaces which , in turn , gives rise to an oscillatory behavior on top of the growth of @xmath56 .
these oscillations prevent a reliable determination of the growth - law .
a similar behavior is observed on fractal substrates @xcite in sec .
[ 3d ] we will present the results of simulations of systems with different values of @xmath0 in the range @xmath57 $ ] .
it must be noticed that , as discussed more detailedly in @xcite , for dilutions @xmath58 or close to it the working temperature @xmath59 is necessarily larger than @xmath43 , see fig .
[ fig_tc ] .
hence we are quenching above the critical temperature @xmath60 , the system will eventually relax to a disordered state with a finite spin coherence length @xmath61 ( not to be confused with the substrate property @xmath62 ) and the growth of @xmath56 will eventually stop at @xmath63 when @xmath64 . in a preasymptotic stage with @xmath65 coarsening
takes place and , since @xmath61 diverges very fast for @xmath55 , this can last for a very long time .
the situation is similar to the one found in other ferromagnetic systems with @xmath66 as , for instance , the one - dimensional ising model @xcite .
we have checked that @xmath67 is sufficiently low as to never observe the equilibration of the system in the range of simulated times , and that the system keeps on coarsening at any simulated time .
the observable quantity of interest in this paper is the typical domain size @xmath56 . for a model defined on a disconnected network , as it happens for @xmath68 ,
phase ordering occurs independently on the various parts of the system and , correspondingly , different definitions of the growing length can be given .
following @xcite , we determine this quantity from the inverse excess energy as l(t , d)=[e(t , d)-e_(d)]^-1 , [ lt ] where @xmath69 is the energy at time @xmath4 , the angular brackets denote a non - equilibrium ensemble average , taken over random initial conditions and over dynamical trajectories and disorder realizations , and @xmath70 is the energy of the equilibrium state at the final temperature @xmath52 .
( [ lt ] ) , which is often used to determine @xmath56 in non - diluted systems @xcite , has , in the present case , the further advantage that the disconnected finite parts of the substrate which are already ordered do not contribute to the computation of @xmath56 .
indeed each droplet is surrounded by empty sites in the sdim or by null bonds in the bdim and hence there is no excess energy associated with it when all spins are aligned , as discussed in @xcite . using eq.([lt ] ) , then , one computes @xmath56 only on the regions where growth is active , which is the kind of information we are interested in .
in @xcite it was shown that the kinetic properties of the @xmath71-sdim and bdim are due to the presence of three fixed points , in the sense of the renormalization group , existing on the @xmath0 axis .
the first fixed point is the trivial one of the pure system , located at @xmath1 , and is associated to the usual power - law growth @xmath72 .
the second one is the _ percolative _ fixed point located at @xmath28 and is characterized by a temperature - dependent power - law growth l(t , d_c)~t^1/(t ) , [ ldc ] with @xmath73 .
this new power - law growth is due to the fact that diluted sites act as pinning centres where interfaces get stucked unless an activation energy @xmath74 is supplied by the thermal bath .
according to the argument presented in @xcite this energy scales as @xmath75 .
this is expected to be true for sufficiently low temperatures .
assuming an arrhenius time @xmath76 $ ] to exceed such energetic barriers one obtains eq .
( [ ldc ] ) with an exponent 1/(t)=k_bt / a [ expt ] proportional to the absolute temperature .
the two fixed points discussed above are repulsive , in the sense that as soon as @xmath77 , or @xmath78 the asymptotic dynamics is governed by a different fixed point , located at @xmath79 ( roughly half way between @xmath1 and @xmath58 ) , to which a logarithmic increase of @xmath80 is associated . due to these fixed point structure ,
if the system is prepared with an intermediate dilution between @xmath1 and @xmath2 ( or equivalently between @xmath2 and @xmath28 ) a crossover pattern is observed at a certain time @xmath81 from an initial transient regime governed by the nearest unstable fixed point ( @xmath1 or @xmath28 respectively ) , with a power - law increase of the domains size , to a late regime controlled by the attractive point at @xmath2 , characterized by a logarithmic @xmath56 .
the crossover phenomenon occurs when @xmath56 reaches the typical size associated to dilution , which is either @xmath82 or @xmath62 depending on @xmath83 or @xmath84 .
given that both @xmath85 and @xmath86 decrease approaching @xmath2 ( eqs .
( [ divxi],[sub.4 ] ) ) , the slowest possible growth , namely the one where @xmath81 is smaller , is obtained at @xmath79 .
therefore by comparing @xmath56 for different values of @xmath0 one finds a non - monotonous behavior : the growth slows down in going from @xmath1 to @xmath79 and then speeds up again when @xmath0 is further increased from @xmath79 up to @xmath58 .
this can represent a practical tool to identify @xmath2 . in the next section
we will show that a similar structure is observed also in @xmath87 .
the details of the simulations are as follows .
we have considered a three - dimensional square lattice system of @xmath88 sites , with @xmath89 for the sdim and @xmath90 for the bdim .
we have checked that , with this choice , no finite - size effects can be detected in the time - regime accessed by the simulations . for every choice of the parameters , we have performed a certain number ( in the range 10 - 100 ) of independent runs with different initial conditions , disorder realizations and thermal histories in order to populate the non - equilibrium ensemble needed to extract average quantities . to speed up the simulations we have used a modified dynamics where spin flips in the bulk of domains , namely those aligned with all the nearest neighbors ,
are prevented .
this modified dynamics does not alter the behavior of the quantities we are interested in , as has been tested in the @xmath71 version of the models and in a large number of different cases @xcite .
we have checked that this is also true in the present study .
let us start by considering the sdim .
the time dependence of @xmath56 , for various dilution values , is plotted in fig .
[ fig1 ] for the @xmath87-sdim .
notice that we plot @xmath91 in order to normalize the curves at the microscopic time @xmath92 when the late - stage scaling regime is entered . in the pure case the asymptotic power - law @xmath72 is expected . in the relatively small range of simulated times
we measure a somewhat smaller exponent ( in the last decade a best fit procedure yields @xmath93 , as already reported in @xcite , because the convergence to the true asymptotic behavior is very slow , as explained in @xcite . indeed the observed _ effective
_ exponent @xmath94 appears to increase with time .
however here we are mainly interested in the non - monotonic behavior of the speed of growth for which an accurate estimate of @xmath95 is not necessary . in fact , it is clearly observed that , by increasing @xmath0 , the growth law becomes slower and slower up to a certain value of @xmath96 that , following @xcite , we identify with @xmath2 . for @xmath68 growth becomes faster upon increasing @xmath0 up to @xmath28 .
this behavior results in a non - monotonic @xmath0-dependence of @xmath56 , at fixed time @xmath4 . this can be clearly appreciated in the inset of fig .
[ fig1 ] where the behavior of the effective exponent obtained by fitting the data for various @xmath0 with a power - law ( regardless of the fact that the actual growth - law might not be power - law ) in the time - range @xmath97 . having established the non - monotonous dependence of the speed of growth on @xmath0 , we make some comments on the form of the growth - law . at the dilution @xmath79 we observe a downwards banding of @xmath56 which indicates that the growth is slower than algebraic , suggesting in this case a logarithmic law .
we recall that a form @xmath98 was predicted in a model with bond disorder in @xcite , which , however , due its slow character has never been accurately demonstrated , to the best of our knowledge , even in the less numerically demanding 2d cases .
clearly in this @xmath87 case the task is much more demanding , and the time range accessible in the simulations is not sufficient for a precise determination of the analytical form of the growth - law . for @xmath58 , on the other hand , the behavior of @xmath99 is well fitted , as expected , by a power - law @xmath100 , with @xmath101 . according to the discussion of sec .
[ over ] , this behavior is expected irrespective of the quenching temperature ( but with a different temperature - dependent exponent @xmath102 ) , provided it is sufficiently low . in order to check this
we have performed a series of simulations by fixing @xmath58 and changing @xmath67 .
the behavior of @xmath56 in these cases is shown in fig .
the data are consistent with power - laws for any choice of @xmath67 .
notice however that for very small @xmath67 the power - law is decorated by a sort of log - time periodic oscillation .
the exponent @xmath103 , obtained by fitting all the curves in the last decade ( for @xmath104 we use two decades in order to smear out the oscillation ) , is plotted in the inset . according to eq .
( [ expt ] ) , one finds a linear behavior at sufficiently low temperatures .
fitting this curve we obtain @xmath105 ( eq.([expt ] ) ) .
notice that for large values of the temperature this linear behavior breaks down , due to the constraint @xmath106 since it is impossible to grow faster than in the pure case . in order to check the generality of the behaviors observed in the sdim we have performed a series of simulations also for the 3d - bdim .
unfortunately , for this model it is not possible to find a region of sufficiently low temperatures to be considered representative of the @xmath55 limit , as discussed in sec .
[ timev ] , but high enough to eliminate the stop and go behavior of the interfaces , which is responsible for the oscillatory character of @xmath56 .
we have explored several values of @xmath107 , and in all cases we observe a non - monotonic behavior of @xmath56 as @xmath0 is varied . in fig .
( [ fig3 ] ) we plot results for @xmath54 . for quenches at this temperature oscillations are clearly evident ruling out the possibility to make any statement on the form of the growth - law ( whether logarithmic or power - law ) . despite this , the non monotonic behavior of @xmath56 as @xmath0 is varied at fixed @xmath4 is quite evident from fig .
[ fig3 ] . starting from the pure case @xmath1 , where a fast growth is observed (
although much slower than the asymptotic expected behavior @xmath72 , due to the long - lasting pre - asymptotic corrections discussed above ) the growth gets slower as @xmath0 is increased up to @xmath108 , and then speeds up again until @xmath28 is reached .
this shows that the non - monotonicity of the speed of growth as disorder is increased is a general property of diluted system , irrespective of dimension and of the kind of dilution .
this suggests that an interpretation based on the presence of three fixed points , the pure one and those at @xmath2 and @xmath28 , and the attractive / repulsive nature of such fixed points , can be appropriated quite generally .
in this paper we have studied the diluted ising model in three dimensions . the aim of the paper was to show the generality of the phase - ordering behavior previously observed in the 2d - sdim and 2d - bdim , where it was first shown @xcite the non - monotonous dependence of the speed of growth on the amount of disorder , a fact that was interpreted as due to the close relation between the coarsening properties and some topological features of the network in disordered magnets . in order to do that we have considered three - dimensional systems .
our results show that the pattern of behavior observed in these models is analogous to the one found in the 2d sdim and rbim , and henceforth an analogous interpretation can be given . in the pure system phase - ordering occurs without activation energy by a smooth curvature - driven displacement of interfaces , leading to the well known growth - law @xmath72 .
the situation changes radically as soon as disorder is introduced , because inhomogeneities pin the interfaces in energy minima and thermal activation is required to proceed . in @xcite an argument was developed leading to the conjecture that the @xmath56-dependence of the pinning barriers changes when passing from @xmath109 to @xmath58 , due to a radically different large - scale topology of the network which turns from a compact to a fractal object .
this implies also a different growth - law , logarithmic for @xmath109 vs power - law @xmath110 right at @xmath28 . in a renormalization - group language
this features can be accomodated in a framework with three fixed points , located at @xmath1 , @xmath79 and @xmath58 . accordingly , a crossover phenomenon is observed where , for any @xmath111 , @xmath56 changes from an early behavior associated to the nearest repulsive fixed point to the asymptotic logarithmic law .
this gives rise to the non - monotonic behavior of @xmath56 as @xmath0 is varied keeping @xmath4 fixed .
the presented numerical study suggests that the pattern of behaviors observed in @xcite is quite a general feature of ferromagnetic disordered systems .
indeed we find that the non - monotonic behavior of @xmath56 is observed also in @xmath87 systems .
our results also indicate that , at least for the sdim , @xmath99 exhibits a power law growth with a growth exponent @xmath112 inversely proportional to the temperature . the growth law becomes slower as @xmath0 is decreased to @xmath2 , even if our data do not allows the precise determination of a logarithmic growth for @xmath80 .
it would be an interesting line of research to investigate if a similar structure underlies also the kinetics of other disordered systems such as , for instance , ising models with random fields .
f. corberi , l. f. cugliandolo , and h. yoshino , in _ dynamical heterogeneities in glasses , colloids , and granular media _ , edited by l. berthier , g. biroli , j .-
bouchaud , l. cipelletti , and w. van saarloos ( oxford university press , oxford , 2011 ) .
r. paul , s. puri and h. rieger , europhys .
lett . * 68 * , 881 ( 2004 ) ; r. paul , s. puri and h. rieger , phys . rev .
e * 71 * , 061109 ( 2005 ) ; r. paul , g. schehr and h. rieger , phys .
e * 75 * , 030104(r ) ( 2007 ) .
s. puri and n. parekh , j. phys .
a * 26 * , 2777 ( 1993 ) ; e. oguz , a. chakrabarti , r. toral and j.d .
gunton , phys .
b * 42 * , 704 ( 1990 ) ; e. oguz , j. phys . a * 27 * , 2985 ( 1994 ) ; m. rao and a. chakrabarti , phys . rev . lett .
* 71 * , 3501 ( 1993 ) ; c. aron , c. chamon , l.f .
cugliandolo and m. picco , j. stat .
p05016 ( 2008 ) . c. castellano , f. corberi
, u. marini bettolo marconi , and a. petri , j. phys .
iv france * 08 * , pr6 - 93 ( 1998 ) .
r. burioni , f. corberi , and a. vezzani , j. stat . mech . ( 2010 ) p12024 ; j. stat . mech .
( 2009 ) p02040 ; r. burioni , d. cassi , f. corberi , and a. vezzani , phys . rev .
e * 75 * , 011113 ( 2007 ) ; phys .
rev . lett . * 96 * , 235701 ( 2006 ) .
f. corberi , c. castellano , e. lippiello , and m. zannetti , phys .
e * 65 * , 066114 ( 2002 ) ; n. andrenacci , f. corberi , and e. lippiello , phys .
e * 74 * , 031111 ( 2006 ) ; e. lippiello , f. corberi , and m. zannetti , phys .
e * 71 * , 036104 ( 2005 ) ; r. burioni , f.corberi , and a. vezzani , phys .
e * 79 * , 041119 ( 2009 ) ; s. j. cornell , k. kaski , and r. b. stinchcombe , phys . rev .
b * 44 * , 12263 ( 1991 ) .
f. corberi , e. lippiello , and m. zannetti , phys .
e * 63 * , 061506 ( 2001 ) ; eur .
j. b * 24 * ( 2001 ) , 359 ; phys.rev .
e * 68 * , 046131 ( 2003 ) ; phys.rev .
e * 78 * , 011109 ( 2008 ) ; e. lippiello , f. corberi , a. sarracino , and m. zannetti , phys . rev .
e * 78 * , 041120 ( 2008 ) ; f. corberi , and l.f .
cugliandolo , j. stat .
p05010 ( 2009 ) . | we study numerically the phase - ordering kinetics of the site - diluted and bond - diluted ising models after a quench from an infinite to a low temperature .
we show that the speed of growth of the ordered domain s size is non - monotonous with respect to the amount of dilution @xmath0 : starting from the pure case @xmath1 the system slows down when dilution is added , as it is usually expected when disorder is introduced , but only up to a certain value @xmath2 beyond which the speed of growth raises again .
we interpret this counterintuitive fact in a renormalization - group inspired framework , along the same lines proposed for the corresponding two - dimensional systems , where a similar pattern was observed . |
urface plasmon effects in gold nanoparticles is a physical phenomena that has been observed in colored glass objects since ancient times @xcite .
the most fascinating and useful features of the plasmonic resonances in metal nanoparticles is first of all the mere existence of these resonances that may occur at free - space wavelengths that are many order of magnitudes larger than the structure itself , and secondly ( and contrary to intuition ) that the corresponding resonance frequencies are virtually independent of the size of the particles ( if they are sufficiently small ) , but does depend on its shape and orientation , see _
e.g. _ , @xcite . today
, new theory and applications of plasmonics are constantly being explored in technology , biology and medicin .
the topic includes the study of surface plasmonic resonances in small structures of various shapes , possibly embedded in different media , see _
e.g. _ , @xcite .
the present study is restricted to passive surrounding materials , but future applications of plasmonics may even include amplifying ( active ) media as described in _
e.g. _ , @xcite .
the classical theories as well as most of the recent studies on plasmonic resonance effects are concerned with metal nanoparticles and photonics where the exterior domain is lossless , see _
e.g. _ , @xcite .
there are very few results developed for absorption and plasmonic resonance effects in particles or structures surrounded by lossy media . as _
e.g. _ , in @xcite is given geometry independent absorption bounds for the plasmonic resonances in metal nanoparticles in vacuum , and an indication is given about how their results can be extended to lossy surrounding media .
there exists a general mie theory for the electromagnetic power absorption in small spherical particles surrounded by lossy media , with explicit expressions and asymptotic formulas for the corresponding absorption cross section , see _
e.g. _ , @xcite .
even though these formulas are derived for spherical geometry , they are in general quite complicated and difficult to interpret .
however , as will be demonstrated in this paper , a new simplified formula for the absorption cross section can be derived which is valid for small ellipsoidal particles embedded in lossy media , and which facilitates a definition of the corresponding optimal plasmonic resonance .
a new potentially interesting application area for the plasmonic resonance phenomena is with the electrophoretic heating of gold nanoparticle suspensions as a radiotherapeutic hyperthermia based method to treat cancer @xcite . in particular ,
gold nanoparticles ( gnps ) can be coated with ligands ( nutrients ) that target specific cancer cells as well as providing a net electronic charge of the gnps @xcite .
the hypothesis is that when a localized , charged gnp suspension has been taken up by the cancer cells , it will facilitate an electrophoretic current and a heating that can destroy the cancer under radio or microwave radiation , and this without causing damage to the normal surrounding tissues @xcite .
hence , the potential medical application at radio or microwave frequencies provides a motivation for studying optimal plasmonic resonances in lossy media .
however , it is also important to consider the complexity of this clinical application with many possible physical and biophysical phenomena to take into account , including cellular properties and their influence on the dielectric spectrum @xcite , as well as thermodynamics and heat transfer , see _ e.g. _ , @xcite .
it is also interesting to note that several authors have questioned whether metal nanoparticles can be heated in radio frequency at all , see _
e.g. _ , @xcite . based on the above mentioned results @xcite as well as our own pre - studies in @xcite
, we are proposing that straightforward physical modeling can be used to show that the most basic electromagnetic heating mechanisms , such as standard joule heating and inductive heating , most likely can be disregarded for this medical application , whereas the potential application remains with the feasibility of achieving plasmonic ( electrophoretic ) resonances .
recently , an optimal plasmonic resonance for the sphere has been defined as the optimal conjugate match with respect to the surrounding medium , _
i.e. _ , the optimal permittivity of the spherical suspension that maximizes the absorption at any given frequency @xcite .
it has been demonstrated in @xcite that for a surrounding medium consisting of a weak electrolyte solution ( relevant for human tissue in the ghz range ) , a significant radio or microwave heating can be achieved inside a small spherical gnp suspension , provided that an electrophoretic particle acceleration ( drude ) mechanism is valid and can be tuned into resonance at the desired frequency . in this paper , we generalize the results in @xcite to include small structures of ellipsoidal shapes embedded in lossy media , and we provide explicit expressions for the corresponding absorption cross section and optimal conjugate match ( optimal plasmonic resonance ) .
we investigate the necessary and sufficient condition regarding the feasibility of tuning a drude model to optimal conjugate match at a single frequency , and we discuss the relation between the optimal conjugate match and the classical frlich resonance condition .
a relative absorption ratio is defined to facilitate a quantitative and unitless indicator for the achievable local heating , and some general expressions are finally given regarding the orientation of the ellipsoid in the polarizing field .
numerical examples are included to illustrate the theory based on simple spheroidal geometries , and which at the same time are relevant for the potential medical application with electrophoretic heating of gnp suspensions in the microwave regime .
the following notation and conventions will be used below .
classical electromagnetic theory is considered based on si - units @xcite and with time convention @xmath0 for time harmonic fields .
hence , a passive dielectric material with relative permittivity @xmath1 has positive imaginary part .
let @xmath2 , @xmath3 , @xmath4 and @xmath5 denote the permeability , the permittivity , the wave impedance and the speed of light in vacuum , respectively , and where @xmath6 and @xmath7 .
the wavenumber of vacuum is given by @xmath8 , where @xmath9 is the angular frequency and @xmath10 the frequency .
the cartesian unit vectors are denoted @xmath11 and the radius vector is @xmath12 where @xmath13 is the radial unit vector in spherical coordinates .
finally , the real and imaginary part and the complex conjugate of a complex number @xmath14 are denoted @xmath15 , @xmath16 and @xmath17 , respectively .
consider a small spherical region of radius @xmath18 ( @xmath19 ) consisting of a dielectric material with relative permittivity @xmath20 and which is suspended inside a lossy dielectric background medium having relative permittivity @xmath1 .
both media are assumed to be homogeneous and isotropic .
the absorption cross section @xmath21 of the small sphere is given by @xmath22 where the scattering cross section @xmath23 , the extinction cross section @xmath24 and the absorption cross section with respect to the ambient material @xmath25 , are given by @xmath26 , \label{eq : cext } \\
c_{\rm amb}=\frac{8\pi}{3}k_0r_1 ^ 3{\operatorname{im}}\{\sqrt{\epsilon}\},\label{eq : cinc}\end{aligned}\ ] ] see _ e.g. _ , @xcite . by algebraic manipulation of through , exploiting relations such as @xmath27 , @xmath28 and @xmath29 , it can be shown that the absorption cross section can also be expressed in the simplified form @xmath30 see also @xcite .
in particular , from it can be shown that the optimal conjugate match @xmath31 is the maximizer of @xmath21 for @xmath32 , and which defines the optimal plasmonic resonance for the sphere in a lossy surrounding medium @xcite .
the polarizability of the sphere is given by @xmath33 where @xmath34 is the volume of the spherical particle , see _
e.g. _ , @xcite . by inserting into through
, the following expression can be obtained @xmath35 alternatively , can be rewritten as @xmath36 and inserted into to yield @xmath37 at this point , it is emphasized that both expressions and have been derived based on the spherical assumption via .
when @xmath1 is real valued the expression reduces to the well known expression for the absorption cross section of small particles of arbitrary shape that are surrounded by lossless media , _ i.e. _ , @xmath38 , see @xcite . on the other hand , the expression is in a more simple form which is well suited for the derivation of the optimal plasmonic resonance in connection with .
it should be noted that the denominator @xmath39 in does not represent a pole of @xmath21 at @xmath40 , its significance is instead to cancel the corresponding zero that is present in the polarizability @xmath41 given by . to derive the polarizability of a small homogeneous structure or a particle
, it is assumed that the excitation is given by a constant static electric field @xmath42 , with the polarization defined by the direction of the @xmath43th cartesian axis .
the fundamental equations to be solved are given by @xmath44 where @xmath45 and @xmath46 are the electric field intensity and the electric flux density ( electric displacement ) , respectively , and where @xmath47 denotes the complex valued relative permittivity which is assigned the appropriate constant values inside and outside the structure .
the equations in are solved by introducing the scalar potential @xmath48 where @xmath49 , and where @xmath48 satisfies the laplace equation @xmath50 , together with the continuity of @xmath48 as well as the continuity of the normal derivative @xmath51 at the boundary of the structure .
finally , the scalar field must satisfy the asymptotic requirement @xmath52 .
the resulting dipole moment relative the background is then given by @xmath53 where @xmath54 denotes the electric field inside the structure and the letter @xmath55 is used to denote the domain of the structure as well as its volume .
consider now a small ellipsoidal region consisting of a dielectric material with relative permittivity @xmath20 and volume @xmath55 , and which is suspended inside a lossy dielectric background medium having relative permittivity @xmath1 , see figure [ fig : rfsetup4pdf ] .
both media are assumed to be homogeneous and isotropic .
let the largest spatial dimension of the ellipsoid be denoted @xmath56 and assume that @xmath57 .
a solution to the electrostatic problem for the ellipsoid is provided by @xcite , and it is shown that when the applied field is aligned along one of the axes of the ellipsoid the resulting electric field @xmath58 is constant inside the particle and parallel to the applied field @xmath59 . from the analytical solution of this problem ,
the polarizability @xmath60 of the ellipsoid is then finally obtained from the definition @xmath61 the resulting formula for the polarizability of the ellipsoid with semiaxes @xmath62 parallel to the cartesian axes @xmath63 , @xmath64 , and excitation @xmath42 is given by @xmath65 where @xmath66 for @xmath67 and where @xmath68 and @xmath69 , see @xcite . here ,
@xmath70 , @xmath71 and @xmath72 are geometrical factors satisfying @xmath73 .
( 50,150 ) ( 40,10)(150,120 ) and volume @xmath55 , surrounded by a lossy background material with permittivity @xmath1 .
the figure also illustrates some typical dimensions of coated gold nanoparticles constituting the ellipsoidal suspension with spatial dimension @xmath56 , see also @xcite.,title="fig:",width=226 ] ( 152,94 ) @xmath55 ( 140,75 ) @xmath20 ( 95,90 ) @xmath1 ( 132,25 ) @xmath74@xmath75 ( 189,25 ) @xmath76@xmath75 ( 175,70)@xmath77 note that @xmath78 is the additional dipole moment added to the background polarization .
this is obvious from the expression implying that @xmath79 when @xmath40 .
note also that @xmath39 is the additional permittivity inside the particle with respect to the background .
hence , the total polarization of the medium inside the particle can be written @xmath80 and the additional polarization @xmath81 yields the additional dipole moment relative the background @xmath82 where @xmath58 is a constant vector . by comparison of and , and exploiting that @xmath59 and @xmath58
are parallel , it is found that the interior field of the particle is given by @xmath83 the poynting s theorem gives the total power loss inside the particle as @xmath84 where has been used .
the power density of a plane wave in a lossy medium is given by @xmath85 where @xmath86 and @xmath87 .
hence , the absorption cross section is finally obtained as @xmath88 which is identical to the formula given in .
consider the real valued function @xmath89 where @xmath20 is a complex variable with @xmath90 and @xmath91 a constant with @xmath92 .
take the complex derivative of @xmath93 with respect to @xmath94 to yield @xmath95 showing that @xmath96 is a stationary point .
it has furthermore been shown in @xcite that @xmath93 is a strictly concave function with a local maximum at @xmath97 , and hence we refer to @xmath91 as the optimal conjugate match .
the absorption cross section for the ellipsoid with polarizability , is given by @xmath98 by comparison of and , it is immediately seen that the optimal conjugate match for the ellipsoid is given by @xmath99 and which hence defines the optimal plasmonic resonance for the ellipsoid in a lossy surrounding medium .
the sphere is a special case of the ellipsoid with @xmath100 yielding @xmath31 , and which reproduces the corresponding result given in @xcite .
the notion of the optimal resonance defined in as being `` plasmonic '' is motivated by the fact that a `` normal '' lossy background medium would have @xmath101 and hence @xmath102 , which is a typical feature of plasmonic resonances and which can be achieved _
e.g. _ , by tuning a drude model .
if we consider the optimal conjugate match @xmath91 in as a function of frequency , then it represents a metamaterial in the sense that it has a negative real part ( a dielectric medium with inductive properties ) , and which can not in general be implemented as a passive material over a fixed bandwidth , see also @xcite .
however , as will be shown below , in many cases a drude model can be tuned to optimal plasmonic resonance at any desired center frequency . a generalized drude model for the permittivity of the ellipsoidal particle
is given by @xmath103 where @xmath104 corresponds to the background material and where the static conductivity @xmath105 and the relaxation time @xmath106 are the parameters of the additional drude model .
it is assumed that the background material is a `` normal '' material with @xmath107 and @xmath108 over the bandwidth of interest .
the drude parameters may correspond to _
e.g. _ , an electrophoretic mechanism where @xmath109 and @xmath110 , where @xmath111 is the number of charged particles per unit volume , @xmath112 the particle charge , @xmath113 the friction constant of the host medium and @xmath114 the mass of the particle , see _
e.g. _ , @xcite .
the drude parameters can be tuned to the optimal conjugate match by solving the equation @xmath115 where @xmath116 is given by and @xmath117 is the desired resonance frequency .
this means that the following two equations corresponding to the real and imaginary parts of must be satisfied @xmath118 to find a solution to , it is necessary and sufficient that both equations have a right - hand side that is positive . for a `` normal '' surrounding material with @xmath101 , it is readily seen from that @xmath102 and hence that @xmath119 .
for the imaginary part , the requirement that @xmath120 together with leads directly to the condition @xmath121 when the condition is fulfilled , the system can be solved to yield the following tuned drude parameters @xmath122 see also @xcite .
consider the interpretation of the condition in the case with spheroidal shapes .
choose for example the @xmath123 axis as the direction of the applied electric field @xmath124 , and let @xmath125 where @xmath126 . the ellipsoid is then a prolate spheroid when @xmath127 , a sphere when @xmath128 and an oblate spheroid when @xmath129 .
the interpretation of is that the sphere and the prolate spheriod can always be tuned by a drude model to match the optimal value @xmath116 at any desired center frequency @xmath130 for which @xmath131 .
an oblate spheroid , however , can only be tuned into optimal plasmonic resonance using the drude model , when the shape is not too flat and @xmath132 .
this result agrees well with intuition , since polarizability ( and hence resonance ) is enhanced by prolongation of the particle shape in the direction of the polarizing field .
the result generalizes the classical frhlich condition @xcite in the sense that gives the condition for an optimal plasmonic resonance of a small homogeneous ellipsoid , which is not an approximation and which is valid for a surrounding lossy medium .
hence , the frhlich condition for the ellipsoid can be obtained from in a sequence of approximations as follows .
first , the criterion is approximated as @xmath133 assuming that the imaginary parts of both @xmath1 and @xmath20 are small . using the following form of the drude model @xmath134 where @xmath135 is the plasma frequency given by @xmath136
, the equation can be solved to yield the following frhlich resonance frequency @xmath137 where the last approximation is valid when @xmath138 . for a lossless surrounding medium with real valued @xmath1 , the frhlich resonance frequency for a sphere consisting of a drude metal
is given by @xmath139 , see @xcite .
the absorption cross section of a small volume with respect to the ambient material is given by @xmath140 , and which is valid for volumes of arbitrary shape , see also .
a unitless relative absorption ratio for the ellipsoid can now be defined as @xmath141 where has been used , as well as the relationship @xmath142 . by inserting the optimal conjugate match into , the following optimal relative absorption ratio is obtained for excitation along the @xmath143 axis of the ellipsoid @xmath144 the relative absorption ratio given by and can be useful as a quantitative unitless measure showing how much more heating that potentially can be obtained in a small resonant region in comparison to the ambient local heating .
it is important to note , however , that a complete system analysis would take into account not only the local heating capabilities , but also the significance of the frequency dependent penetration ( skin ) depth of the bulk material , see also @xcite .
finally , a general expression is given for the absorption cross section of a small homogeneous ellipsoidal particle with arbitrary orientation with respect to the applied field .
consider a small ellipsoidal region with its semiaxes @xmath62 aligned along the cartesian unit vectors @xmath63 , @xmath64 , and an applied electric field given by @xmath145 .
due to the linearity of the fundamental equations , it is straightforward to generalize the expressions on absorption cross section given in sections [ sect : abshomellipsoids ] and [ sect : optplasmresellipsoid ] above .
the polarizability can now be expressed in terms of the diagonal polarizability dyadic @xmath146 where @xmath147 , and the interior field @xmath58 is given by @xmath148 instead of .
the total power loss inside the particle is now given by @xmath149 and the corresponding absorption cross section @xmath150 where @xmath151 . by using ,
the absorption cross section and the relative absorption ratio finally becomes @xmath152 and @xmath153 it is immediately seen that the two expressions in and are strictly concave functions in terms of the complex variable @xmath20 for @xmath90 ( a positive combination of concave functions is a concave function , etc ) and the corresponding optimal plasmonic resonance is therefore well - defined and unique .
however , it is no longer possible to obtain a simple closed form expression for the optimal conjugate match @xmath154 as in .
to illustrate the theory , a numerical example is considered with parameter choices relevant for the application with microwave absorption in gold nanoparticle suspensions , see _
e.g. _ , @xcite .
hence , the resonant frequency is chosen here as @xmath155 to mimic a typical system operating in the microwave regime , see _
e.g. _ , @xcite . as a lossy ambient medium
is taken the typical characteristics of human tissue .
information about the dielectric properties of biological tissues can be found in _
e.g. _ , @xcite giving measurement results of most organs including brain ( grey matter ) , heart muscle , kidney , liver , inflated lung , spleen , muscle , etc . from these measurement results
we conclude that human tissue can be realistically modelled by using a conductivity in the order of @xmath156@xmath157 and a permittivity similar to water at a frequency of @xmath156@xmath158 .
hence , a simple conductivity - debye model for saline water is considered here where the surrounding medium is a weak electrolyte solution with relative permittivity @xmath159 where @xmath160 , @xmath161 and @xmath162 are the high frequency permittivity , the static permittivity and the dipole relaxation time in the corresponding debye model for water , respectively , and @xmath163 the conductivity of the saline solution . in the numerical examples below , these parameters are chosen as @xmath164 , @xmath165 , @xmath166@xmath167 and @xmath168@xmath157 . in figures [ fig : matfig503 ] through [ fig : matfig504b ] are shown the calculated relative absorption ratios for the ellipsoid with optimal , tuned drude and mismatched drude parameters , respectively .
the optimal parameter @xmath91 is given by , the tuned drude parameter @xmath169 is given by and , and the mismatched drude parameter @xmath170 is again the drude parameter given by and , but which is constantly mismatched to the sphere using @xmath31 .
a spheroidal shape is considered with the geometrical factors @xmath125 and @xmath126 , and where the applied electric field @xmath124 is aligned along the @xmath123 axis of the spheroid and hence @xmath171 is given by .
the relative absorption ratios in the three cases described above are denoted @xmath172 , @xmath173 and @xmath174 corresponding to the parameters @xmath91 , @xmath169 and @xmath170 respectively .
the parameter choices in figures [ fig : matfig503 ] through [ fig : matfig504b ] , are @xmath175 ( prolate spheroid ) , @xmath128 ( sphere ) and @xmath176 ( oblate spheroid ) which is close to the limiting case @xmath177 expressed in . from these examples , it is seen how the increased conductivity and losses ( figures [ fig : matfig503b ] and [ fig : matfig504b ] ) limits the usefulness of the local heating .
but even in the latter example , where @xmath178@xmath157 , the potential of local heating amounts to a relative absorption ratio of about 10:1 . in the case with the mismatched drude model , it is interesting to see how a prolongation of the spheroid lowers the resonance frequency , and a flattening of the spheroid yields a higher resonance frequency .
( 50,140 ) ( 100,0)(50,130 ) with @xmath172 corresponding to the optimal parameter @xmath154 and @xmath173 corresponding to the tuned drude parameter @xmath169 . here
, the surrounding medium is a saline solution with @xmath179@xmath157.,title="fig:",width=321 ] ( 50,140 ) ( 100,0)(50,130 ) with @xmath172 corresponding to the optimal parameter @xmath154 and @xmath174 corresponding to the mismatched drude parameter @xmath170 tuned to a sphere . here
, the surrounding medium is a saline solution with @xmath179@xmath157.,title="fig:",width=321 ] ( 50,140 ) ( 100,0)(50,130 ) with @xmath172 corresponding to the optimal parameter @xmath154 and @xmath173 corresponding to the tuned drude parameter @xmath169 . here
, the surrounding medium is a saline solution with @xmath178@xmath157.,title="fig:",width=321 ] ( 50,140 ) ( 100,0)(50,130 ) with @xmath172 corresponding to the optimal parameter @xmath154 and @xmath174 corresponding to the mismatched drude parameter @xmath170 tuned to a sphere . here
, the surrounding medium is a saline solution with @xmath178@xmath157.,title="fig:",width=321 ]
a new general formula has been derived for the absorption cross section of small ellipsoidal particles that are surrounded by lossy media .
the new formula is expressed explicitly in terms of the polarizability of the particle and can be used to define an optimal plasmonic resonance for a given surrounding medium .
the new formula can be derived from general mie scattering theory for a spherical particle in a lossy medium which generalizes to particles of ellipsoidal shape in the limiting case with small particles .
the formula can furthermore be derived directly from the knowledge about the static solution to the ellipsoidal polarizability problem .
a canonical example is presented based on the polarizability of a homogeneous spheroid .
the example shows how an optimal plasmonic resonance can be designed based on a tuned drude model and illustrates the typical shape dependent resonance frequency of the surface plasmon .
the numerical example is furthermore motivated by the medical application with radiotherapeutic hyperthermia based on electrophoretic heating of gold nanoparticle suspensions using microwave radiation .
this work was supported by the swedish foundation for strategic research ( ssf ) .
s. link and m. a. el - sayed , `` shape and size dependence of radiative , non - radiative and photothermal properties of gold nanocrystals , '' _ int . reviews in physical chemistry _ , vol . 19 , no . 3 , pp .
409453 , 2000 .
o. d. miller , a. g. polimeridis , m. t. h. reid , c. w. hsu , b. g. delacy , j. d. joannopoulos , m. soljacic , and s. g. johnson , `` fundamental limits to optical response in absorptive systems , '' _ optics express _ , vol .
24 , no . 4 , pp . 33293364 , 2016 .
t. lund , m. f. callaghan , p. williams , m. turmaine , c. bachmann , t. rademacher , i. m. roitt , and r. bayford , `` the influence of ligand organization on the rate of uptake of gold nanoparticles by colorectal cancer cells , '' _ biomaterials _ , vol .
32 , pp . 97769784 , 2011 .
g. w. hanson , r. c. monreal , and s. p. apell , `` electromagnetic absorption mechanisms in metal nanospheres : bulk and surface effects in radiofrequency - terahertz heating of nanoparticles , '' _ j. appl .
_ , vol . 109 , 2011 , 124306 .
s. j. corr , m. raoof , y. mackeyev , s. phounsavath , m. a. cheney , b. t. cisneros , m. shur , m. gozin , p. j. mcnally , l. j. wilson , and s. a. curley , `` citrate - capped gold nanoparticle electrophoretic heat production in response to a time - varying radio - frequency electric field , '' _ j. phys . chem .
116 , no .
45 , pp . 2438024389 , 2012 .
s. nordebo , m. dalarsson , y. ivanenko , d. sjberg , and r. bayford , `` on the physical limitations for radio frequency absorption in gold nanoparticle suspensions , '' _
_ , vol .
50 , no .
15 , pp . 112 , 2017 .
m. marquez , e. garcia , and m. camacho , `` hyperthermia devices and their uses with nanoparticles , '' jun .
11 2013 , us patent 8463397 , https://www.google.com/patents/us8463397 .
[ online ] .
available : https://www.google.com/patents/us8463397 mariana dalarsson received the m.s .
degree in microelectronics in 2010 , licentiate degree in electromagnetic theory in 2013 and ph.d .
degree in electromagnetic theory in 2016 from the royal institute of technology , stockholm , sweden .
since 2016 , she is a postdoctoral researcher at the department of physics and electrical engineering , linnus university .
her research interests are in mathematical physics , metamaterials , electromagnetic wave propagation and absorption , inverse problems and imaging .
sven nordebo received the m.s .
degree in electrical engineering from the royal institute of technology , stockholm , sweden , in 1989 , and the ph.d .
degree in signal processing from lule university of technology , lule , sweden , in 1995 .
he was appointed docent in signal processing at blekinge institute of technology , in 1999 . since 2002
he is a professor of signal processing at the department of physics and electrical engineering , linnus university .
his research interests are in statistical signal processing , optimization , electromagnetics , wave propagation , inverse problems and imaging .
daniel sjberg received the m.sc .
degree in engineering physics and ph.d .
degree in engineering , electromagnetic theory from lund university , lund , sweden , in 1996 and 2001 , respectively . in 2001
, he joined the electromagnetic theory group , lund university , where , in 2005 , he became a docent in electromagnetic theory .
he is currently a professor and the head of the department of electrical and information technology , lund university .
he serves as the chair of swedish ursi commission b fields and waves since 2015 .
his research interests are in electromagnetic properties of materials , composite materials , homogenization , periodic structures , numerical methods , radar cross section , wave propagation in complex and nonlinear media , and inverse scattering problems .
richard bayford received the m.sc .
degree in engineering from cranfield institute of technology , uk , in 1981 and ph.d .
degree in engineering , from middlesex university , uk , in 1994 .
he is currently professor of bio - modelling , head of biophysics at the middlesex university centre for investigative oncology , head of biophysics and engineering group , and visiting professor at ucl , department of electrical and electronic engineering , uk .
his expertise is in biomedical imaging , bio - modelling , nanotechnology , deep brain stimulation , tele - medical systems , instrumentation and biosensors .
he has had long collaborations with research groups both in the uk and overseas .
he has published over 270 scientific papers and served as editor - in - chief for physiological measurements . | a new simplified formula is derived for the absorption cross section of small dielectric ellipsoidal particles embedded in lossy media .
the new expression leads directly to a closed form solution for the optimal conjugate match with respect to the surrounding medium , _
i.e. _ , the optimal permittivity of the ellipsoidal particle that maximizes the absorption at any given frequency .
this defines the optimal plasmonic resonance for the ellipsoid .
the optimal conjugate match represents a metamaterial in the sense that the corresponding optimal permittivity function may have negative real part ( inductive properties ) , and can not in general be implemented as a passive material over a given bandwidth .
a necessary and sufficient condition is derived for the feasibility of tuning the drude model to the optimal conjugate match at a single frequency , and it is found that all the prolate spheroids and some of the ( not too flat ) oblate spheroids can be tuned into optimal plasmonic resonance at any desired center frequency .
numerical examples are given to illustrate the analysis .
except for the general understanding of plasmonic resonances in lossy media , it is also anticipated that the new results can be useful for feasibility studies with _
e.g. _ , the radiotherapeutic hyperthermia based methods to treat cancer based on electrophoretic heating in gold nanoparticle suspensions using microwave radiation . particle absorption , plasmonic resonances . |
many observations suggest that clusters of galaxies contain substantial amounts of mass in the form of a background intergalactic stellar component ( e.g. , vlchez - gmez et al .
1994 ; scheick , kuhn 1994 ; ferguson et al .
1998 and references therein ) .
it has usually been thought that such a component has its origin in stars stripped from galaxies by galaxy
galaxy encounters , or by the cluster tidal field . however , detailed numerical simulations of galaxy clusters are required to confirm how efficient such mass stripping is and how the structure of the background component evolves .
it was only recently that fully self - consistent @xmath0-body simulation of galaxy clusters , with each galaxy represented by many particles , was made possible , because such a full simulation requires huge computing resources .
funato et al .
( 1993 ) investigated the evolution of isolated clusters of galaxies using self - consistent @xmath0-body simulations .
similar simulations were also performed by bode et al .
( 1994 ) and garijo et al .
more recently , sensui et al . ( 1999 , sfm99 ; 2000 , sfm00 ) followed up the study of funato et al .
( 1993 ) by performing larger @xmath0 simulations for a wider variety of initial conditions . in their simulations , clusters
were initially in virial equilibrium , and all of the cluster mass was initially bound to individual galaxies ; a cluster was composed of about 100 identical galaxies , and each galaxy was modeled with 5004000 particles . sfm99 and sfm00 found that , in all cases , more than half of of the total mass is stripped from galaxies in a few crossing times of the cluster , and that the stripped matter forms a smooth cluster - wide common halo .
they also found that a density cusp , approximated as @xmath6 ( @xmath3 being form @xmath7 to @xmath8 ) , develops in the central region .
however , the physical mechanism of the cusp formation was not clear .
even if we are allowed to regard clusters of galaxies as `` pure '' self - gravitating many - body systems composed of point masses ( stars and dark matter particles ) , we have to admit that their evolution is complicated by many physical processes which take place simultaneously : gravitational encounters , physical collisions , and merging between galaxies ; galaxy background particle interactions , mass stripping from galaxies , internal evolution of galaxies , etc .
therefore , there is no doubt that fully self - consistent @xmath0-body simulations are most desirable to obtain qualitatively and quantitatively reliable results on cluster evolution ( see garca - gmez et al .
1996 ) .
nevertheless , more idealized or more simplified models that include ( part of ) relevant physical processes in approximate ways , such as fokker planck ( fp , hereafter ) models which we used in the present study , are still very useful regarding two points .
first , we can quickly obtain an overview of cluster evolution based on various initial conditions by using approximate models , which are inexpensive in computation time .
second , it is expected that we can understand more clearly the role and importance of each physical process , because we can easily separate the effects of various processes in approximate models . on the other hand , it is often difficult to `` understand '' the results of @xmath0-body simulations , where all processes are automatically included and connected through a complicated procedure
. thus , @xmath0-body and approximate models are complementary to each other , and comparison between these models will be particularly helpful to our further understanding of the physics of cluster evolution , regardless of whether agreement or disagreement is found .
merritt ( 1983 , 1984 , 1985 ) explored the evolution of clusters of galaxies using such approximate models based on an isotropic orbit - averaged fp equation ( cohn 1980 ) .
the fp equation describes the evolution of the phase - space distribution function driven by gravitational two - body encounters , and has been applied mainly to globular clusters ( see , e.g. , spitzer 1987 for a review ) . a typical cluster of galaxies is composed of @xmath9 1001000 galaxies , and thus the time scale of two - body relaxation between galaxies , @xmath10 , is of the same order as the crossing time of the cluster , @xmath11 ( @xmath12 ) .
therefore , two - body relaxation is expected to play an important role in the evolution of galaxy clusters . in his models ,
merritt also took account of mass stripping from galaxies due to tidal interactions between galaxies .
since the stripped particles ( stars or dark matter particles ) were considered to be much lighter than galaxies , these particles were treated as being massless .
dynamical friction between the galaxies and the background massless particles was also included .
thus , in merritt s models the evolution of the galaxies and the background was treated self - consistently .
fp models are continuum models , and therefore have an advantage that they are free from statistical fluctuations , which are unavoidable in @xmath0-body models and make it difficult to understand the underlying physics
. yepes et al .
( 1991 ) and yepes and domnguez - tenreiro ( 1992 ) studied the evolution of galaxy clusters by solving the moment equations of the fp equation .
they also included a background component of massless particles in their models , but did not take account of tidal stripping .
the purpose of the present work was to investigate the dynamical evolution of clusters of galaxies after virialization .
in particular , we aimed to clarify the physical mechanisms of the growth of the common halo , and of the development of the central density cusp , which were investigated by sfm99 and sfm00 with @xmath0-body models . for this purpose ,
we used fp models , which are based on merritt s ( 1983 ) formulation , but are improved in many aspects , and compare them with the self - consistent @xmath0-body models of sfm99 and sfm00 .
as shown below , the results of the fp and @xmath0-body models agree with each other surprisingly well .
this suggests that our fp models correctly include all important physical processes , and thus can be used as a useful tool for studying the dynamics of galaxy clusters .
this paper is organized as follows .
the basic equations of the fp models are given in section 2 .
section 3 describes the initial conditions of our simulations , and section 4 briefly describes the numerical schemes used for our fp and @xmath0-body simulations .
the simulation results are given in sections 5 and 6 ; in section 5 , the fp and @xmath0-body models are compared with respect to the growth of the common halo and the development of the central cusp ; in section 6 , the physical mechanism of the cusp formation is examined in detail based on fp models .
we summarize and discuss our results in section 7 .
in this section , we describe fp models of clusters of galaxies . in fp models , individual galaxies are treated as point particles , except that the effects of tidal stripping are considered .
detailed fp simulations of galaxy clusters were first performed by merritt ( 1983 ) .
we closely follow his formulation .
there are , however , three major differences between his and our models : ( 1 ) merritt used isotropic fp models , but we used anisotropic fp models .
( 2 ) we used a new cross - section formula of mass stripping from galaxies based on recent @xmath0-body experiments by funato and makino ( 1999 , fm ) . ( 3 ) dissipation of the orbital kinetic energies of galaxies during galaxy galaxy encounters was also considered , while it was neglected in merritt s simulations .
consider a self - gravitating many - body system composed of different mass components with particle mass @xmath13 ( @xmath14 ) .
we write the _ number density _ in @xmath15-space , i.e. the distribution function , of component @xmath13 as @xmath16 . by assuming that the system is spherically symmetric and in dynamical equilibrium
, @xmath17 can be written as a function of the energy per unit mass , @xmath18 , and the angular momentum per unit mass , @xmath19 .
then , the orbit - averaged fp equation , under the fixed gravitational potential , can be written as follows ( cohn 1979 ; takahashi 1997 ) : @xmath20 where @xmath21 denotes the scaled angular momentum , @xmath22 is a weight function , and @xmath23 , etc .
are flux coefficients .
see takahashi ( 1997 ) for a complete description of this equation . to follow the evolution of the system ,
this fp equation is solved in combination with poisson s equation ( cohn 1979 ) .
we introduce the _ mass density _ in phase space , @xmath24 here , we choose the @xmath25 component for representing the background , or a common halo .
we can assume that the mass of a background particle ( star ) is negligible compared to that of a galaxy .
however , the mass density of background particles is comparable to that of galaxies .
we now take the limit @xmath26 while keeping @xmath27 finite ( i.e. @xmath28 ) . in the fp equation ( [ eq : fp ] ) , component mass @xmath13 appears only in the flux coefficients . in the limit @xmath26
, the expressions for the first and second - order flux coefficients look like @xmath29 note that we write here only the dependence on @xmath13 and @xmath30 symbolically .
exact expressions can be easily obtained from the equations given in the appendix of takahashi ( 1997 ) .
the fp equation for the background component ( @xmath25 ) becomes @xmath31 which is obtained by multiplying equation ( [ eq : fp ] ) by @xmath32 and then taking the limit @xmath26 .
since the background particles are massless , they are completely collisionless , i.e. , they do not interact with each other .
however , galaxies suffer dynamical friction from the background particles that have a finite mass density .
thus , the orbital energy of the galaxies is transferred to the background .
consider an encounter between _ a test galaxy _ of mass @xmath33 and _ a field galaxy _ of mass @xmath34 , with impact parameter @xmath35 and relative velocity @xmath36 .
fm showed , both numerically and analytically , that the relative change of the test galaxy s mass in an encounter can be approximated by @xmath37 with @xmath38 23 and @xmath39 23 , for the cases with @xmath40 , @xmath41 , and @xmath42 . here , @xmath43 is a dimensionless constant , and @xmath44 and @xmath45 are the size ( represented by the virial radius in the following ) and the three - dimensional internal velocity dispersion of the test galaxy , respectively .
since fm did not carry out simulations of encounters between unequal - mass galaxies , the value of @xmath46 is not determined from their simulations . in the impulse approximation ,
the velocity change of a star , @xmath47 , is proportional to @xmath34 .
therefore , if the effect of the term @xmath48 is dominant in removing mass from the test galaxy , we may expect @xmath49 ; on the other hand , if the effect of the term @xmath50 is dominant , @xmath51 ( see merritt 1983 ; fm ) .
now suppose that a test galaxy experiences successive encounters with field galaxies of mass @xmath34 .
the rate of change in the test galaxy s mass is given by @xmath52 where @xmath53 is the number density of the field galaxies and @xmath54 is the ( normalized ) distribution function of the relative velocity between the test and field galaxies ; @xmath55 and @xmath56 are the minimum and maximum impact parameters , respectively .
if we assume a maxwellian distribution with a three - dimensional velocity dispersion , @xmath57 , for @xmath58 , this equation becomes @xmath59 here , we define @xmath60 and @xmath61 .
the first integral becomes @xmath62 for @xmath63 , and @xmath64 for @xmath65 .
therefore , if we take @xmath55 and @xmath56 as the size of the galaxy and that of the cluster , we may consider that the integral remains roughly constant during the cluster evolution for @xmath66 23 .
the second integral is also a constant of the order of unity for @xmath67 23 ; it equals 1 when @xmath68 , and @xmath69 when @xmath70 .
therefore , we may rewrite equation ( [ eq : ml2 ] ) as @xmath71 where @xmath72 is a dimensionless constant factor and @xmath73 is the mass density of the field galaxies .
note that the definition of @xmath72 depends on @xmath74 and @xmath75 .
when @xmath68 , for example , @xmath76 for @xmath63 , and @xmath77 for @xmath65 .
equation ( [ eq : ml3 ] ) includes two parameters of galaxy s internal structure , @xmath44 and @xmath45 .
because fp models do not follow the internal evolution of galaxies , we have to adopt some relations to express the changes in these parameters by @xmath33 only .
one such relation should be the virial theorem , @xmath78 .
as the other relation we assume @xmath79 with @xmath80 @xmath81@xmath82 ( see fm and sfm99 ) . using these relations
we may write the two parameters as @xmath83 finally , we obtain @xmath84 where a sum over all galaxy components @xmath85 is taken . we then orbit - average ( see , e.g. , takahashi 1997 ) equation ( [ eq : ml4 ] ) to obtain the mass - loss rate as a function of @xmath18 and @xmath19 .
stars escape from galaxies via galaxy galaxy encounters .
this is possible because part of the orbital kinetic energies of galaxies is converted into the energy of internal motions of the stars in the galaxies .
that is , these encounters are inelastic . if the sum of the galaxy binding energies is much smaller than the cluster binding energy , i.e. , if @xmath86 ( @xmath87 is the cluster velocity dispersion ) , the effect of energy dissipation due to inelastic encounters is not very important .
however , for our model clusters shown below , this condition is not well satisfied [ @xmath88 , and therefore energy dissipation is expected to play a non - negligible role .
the binding energy of a galaxy ( in virial equilibrium ) is @xmath89 .
the rate of its change is therefore given by @xmath90 here , in the second equality , we used equation ( [ eq : vg - m ] ) .
the mass - loss rate is given by equation ( [ eq : ml4 ] ) .
the decrease in the binding energy of a galaxy should be compensated by a decrease in the orbital kinetic energies of the galaxies involved in encounters with that galaxy . for simplicity , we assume that the change in the binding energy of a galaxy is just equal to the change in its orbital energy , and that the stripped stars have the same orbital energy per unit mass as that of the parent galaxy .
this is valid at least on average .
the dissipation rate of the orbital energy of galaxy @xmath33 is given by @xmath91 ( remember @xmath18 represents the energy per units mass ) .
the above assumption ensures that the total energy of a galaxy cluster , including the internal energies of the galaxies , is conserved .
the distribution function evolves owing to mass stripping from galaxies as well as owing to two - body relaxation .
the effect of mass stripping can be included in the fp equation by adding a source ( or sink ) term ( see , e.g. , quinlan , shapiro 1989 ) .
thus , in this case the fp equation becomes @xmath92 where @xmath93 denotes the fp collision term [ the right - hand side of equation ( [ eq : fp ] ) ] and @xmath94 represents the source term due to mass stripping .
we neglect merging events between galaxies for simplicity . for the initial conditions which we adopted in this work
, @xmath0-body simulations showed that mergers were rare ( sfm99 ; sfm00 ) .
therefore , neglect of merging events would not cause serious errors for our models .
the source term is calculated in practice as follows . for each galaxy component @xmath95 and for each grid point in @xmath96 space
, we can calculate the change in mass @xmath97 during a time step @xmath98 from the orbit - averaged stripping rate , @xmath99 .
thus , we know that galaxies of mass @xmath100 at time @xmath101 and position @xmath96 will have mass @xmath102 ( @xmath103 ) at time @xmath104 , and that each of these galaxies provides mass @xmath105 for the background .
the number of the galaxies does not change during this mass loss , unless they are completely disrupted .
since we use discrete mass components , a new mass @xmath106 is generally not equal to any components @xmath107 , and therefore we have to somehow distribute the galaxies of mass @xmath106 among some components . as lee ( 1987 ) and quinlan and shapiro ( 1989 ) did for stellar mergers in star clusters , we distribute the galaxies by linear interpolation : if @xmath108 galaxies of mass @xmath106 are to be formed and if @xmath109 , then @xmath110 galaxies are added to the @xmath107 component and @xmath111 galaxies are added to the @xmath112 component . if @xmath113 , where @xmath114 is a minimum galaxy mass assumed , the galaxies are distributed between @xmath115 ( background ) and @xmath114 .
further we assume that the galaxies after mass stripping and the ejected background particles have the same orbital energy and angular momentum per unit mass @xmath96 as the galaxies had before stripping .
the effect of energy dissipation is considered separately , as described below . by doing the above procedure over all of the galaxy components at each grid point in @xmath96 space
, we can calculate the change in the distribution function due to mass stripping , i.e. , the source term , @xmath94 . note that this procedure guarantees conservation of the total mass and energy of the cluster .
merritt ( 1983 ) formulated mass stripping as a first - order `` mass diffusion '' term in the fp equation [ see his equation ( 22 ) ] . in his formulation he assumed a distribution function , @xmath116 , which is continuous with respect to @xmath33 as well as @xmath18 . in integrating the fp equation numerically , he used , of course , discrete mass grids . on the other hand ,
we have assumed discrete mass components from the beginning .
although merritt s formulation and our formulation may look different apparently , they both consider only the first - order term @xmath117 and are essentially the same . the effect of energy dissipation ( cooling ) is included in the fp equation in the same way as the effect of binary heating in star clusters ( see , e.g. , takahashi 1997 ) .
that is , the cooling term , which may be calculated from equation ( [ eq : edis ] ) , is added to the first - order energy coefficient , @xmath118 . in practice , the amount of energy dissipation during @xmath98 is calculated so that it should be consistent with the above - described way of distributing stripped galaxies among mass components .
a commonly used definition of the two - body relaxation time is given by spitzer ( 1987 , equation [ 2 - 61 ] ) as @xmath119 . here
, a maxwellian velocity distribution of field particles is assumed and @xmath120 is the three - dimensional velocity dispersion ; @xmath121 is the mean square value of the change per unit time in the velocity component parallel to the initial velocity of a test particle , evaluated at @xmath122 . for the present cases , assuming that all galaxy components follow maxwellian velocity distributions with the same velocity dispersion , @xmath87
, we obtain @xmath123 where @xmath124 is the total density of the galaxies and @xmath125 is the _ mass - weighted _ mean galaxy mass . equation ( [ eq : tr ] ) represents the time scale of the velocity diffusion due to encounters between galaxies .
similarly , we define the dynamical friction time as @xmath126 , where @xmath127 is the mean change per unit time in the velocity component parallel to the initial velocity , evaluated at @xmath122 . on the same assumptions
as used in deriving equation ( [ eq : tr ] ) , we find that the dynamical friction time for a galaxy of mass @xmath33 is given by @xmath128 where @xmath129 is the total density including both the galaxy and halo components . unlike the relaxation time ( [ eq : tr ] ) , the dynamical friction time ( [ eq : tdf ] ) depends on the test - particle mass and on the halo density . from equations ( [ eq : tr ] ) and ( [ eq : tdf ] )
, it follows that @xmath130 in single - component clusters , @xmath131 . in multicomponent clusters ,
@xmath132 can be much smaller than @xmath10 for massive galaxies ; this holds also for light galaxies , if @xmath133 .
we define the tidal mass stripping time as @xmath134 .
using equation ( [ eq : ml4 ] ) with @xmath135 set equal to @xmath136 , we find @xmath137 using equations ( [ eq : tdf ] ) and ( [ eq : tts ] ) , with @xmath138 , we obtain @xmath139 for our standard set of parameters , @xmath140 , @xmath68 , and @xmath141 ( see section [ sec : r1 ] ) , it follows that @xmath142 this indicates that @xmath143 in our initial models ( see section [ sec : ic ] ) . as the fraction of the common halo increases , tidal stripping becomes less important ( merritt 1983 ) .
we use the same initial cluster model as that in sfm99 .
that is , the initial model is a plummer model composed of @xmath144 identical galaxies with no common halo .
exceptions are models fpa1 and fpa2 , which we describe later .
the initial galaxy model is also given by a plummer model as in sfm99 .
the ratio of the virial radius of each galaxy @xmath145 to the virial radius of the cluster @xmath146 is set equal to @xmath147 .
note that initial galaxy models affect fp models only through the mass - stripping rate , equation ( [ eq : ml4 ] ) .
this equation includes two parameters of the internal structure of galaxies , @xmath148 and @xmath149 , but these are related by the virial theorem ( we take @xmath150 ) for a given galaxy s mass .
hence , in fp models the structure of galaxies is represented by only one parameter , e.g. , @xmath145 . in reality
, even if the virial radii are the same , different galaxy models will give different stripping rates .
such differences in galaxy models could be reflected in fp models only by adjusting an undetermined numerical constant , @xmath72 , in equation ( [ eq : ml4 ] ) .
we use a system of units in which @xmath151 , @xmath152 , and @xmath153 , where @xmath154 is the gravitational constant ( heggie , mathieu 1986 ) ; hence @xmath155 . for the cluster
, we have @xmath156 , @xmath157 , and @xmath158 , where @xmath159 , @xmath146 , and @xmath160 are the mass , virial radius , and binding energy of the cluster , respectively .
the mean velocity dispersion of stars in a galaxy is @xmath161 , and that of galaxies in the cluster is @xmath162 .
the cluster crossing time is @xmath163 and the half - mass relaxation time ( spitzer 1987 , equation [ 2 - 63 ] ) is @xmath164 where the cluster half - mass radius @xmath165 and the coulomb logarithm @xmath166 . if we assume that @xmath167 and @xmath168 , the cluster model has @xmath169 , @xmath170 , @xmath171 , and @xmath172 .
only models fpa1 and fpa2 have a common halo initially ( see subsection [ sec : r1.1 ] ) ; the ratio of the common halo mass to the total mass , @xmath173 , is @xmath174 and @xmath175 , or if the mass is converted into the number of galaxies , @xmath176 and 32 , for fpa1 and fpa2 , respectively . galaxy and common halo components are distributed in space so that the ratio of their densities is independent of the radius .
the other conditions are the same as those of the other models .
details of the numerical integration scheme of the fp equation have been described by takahashi ( 1995 , 1997 ) .
we used 301 energy , 51 angular momentum , and 121 radial grid points .
we used @xmath177 mass components for galaxies .
these components are uniformly spaced in @xmath33 between 0 and 1 ; @xmath178 , @xmath179 , , @xmath180 .
in addition , one component of @xmath115 was assigned to a background ( common halo ) .
we confirmed that these numbers of grid points and mass components are large enough to obtain satisfactory convergence of the results .
the relative errors in the total mass were less than 0.1% and those in the total energy were less than 0.5% in all runs .
we calculated a number of fp models , switching on and off mass stripping and energy dissipation and changing the parameters of the stripping rate .
these models are summarized in table 1 .
we compared our fp models with the @xmath0-body models of sfm99 and sfm00 for the same initial conditions .
sfm99 performed a number of simulations for the same initial theoretical model using different random realizations of the model and varying the number of particles for representing each galaxy ; they confirmed that the results of those simulations were almost the same . in section
[ sec : r1 ] , the model pp of sfm00 , denoted by nb in this paper , is compared with the fp models .
the simulation was carried out using grape-4 ( makino et al .
1997 ) with the barnes hut tree algorithm ( barnes , hut 1986 ; makino 1991 ; athanassoula et al .
each initial galaxy model was represented by 2048 particles , and thus the total number of particles was 262144 . for more details of the numerical integration and for a galaxy identification scheme , see sfm99 and sfm00 .
first we discuss the evolution of model clusters in which mass stripping from galaxies is not allowed , but which initially have massive common halos , in order to reveal how such multicomponent systems evolve owing to two - body relaxation only .
although collisional evolution of multicomponent stellar systems has been investigated by many authors using fp models in the context of globular cluster evolution ( e.g. , inagaki , wiyanto 1984 ; cohn 1985 ; chernoff , weinberg 1990 ; takahashi 1997 and references therein ) , massless particle components were not considered in those studies . on the other hand , it is not very difficult to extrapolate the behavior of massless particles from the behavior of particles whose masses are much smaller than the mean particle mass . in that sense
, no essentially new results were found in our simulations , as discussed below .
however , it is useful to see first how `` pure '' collisional systems evolve before considering more complicated systems with mass stripping
. yepes et al .
( 1991 ) and yepes and domnguez - tenreiro ( 1992 ) also studied pure collisional evolution of galaxy clusters including massless particles , but using the moment equations of the fp equation .
figure 1a shows the evolution of density profiles for model fpa1 ; the solid lines and dotted lines represent the galaxy and common halo components , respectively .
figure 1b shows the corresponding profiles of the logarithmic gradient defined by @xmath181 figures 1c and 1d are for model fpa2 .
we have assumed that the number of galaxies , @xmath182 , is 64 and 32 for fpa1 and fpa2 , respectively .
however , it should be noted that @xmath182 is only concerned with time scaling ; if time is measured in units of the relaxation time , the evolution of these models does not depend on the number of galaxies , because only the two - body relaxation process is taken into account in them .
we set the coulomb logarithm equal to @xmath183 for time scaling for both models .
as is well known , energy redistribution among particles due to two - body relaxation eventually leads the cluster core to gravothermal core collapse ( lynden - bell , wood 1968 ; lynden - bell , eggleton 1980 ) ; after the onset of core collapse the central density continues to increase and the core radius continues to decrease , unless other physical mechanisms intervene .
late core collapse proceeds self - similarly with the development of the power - law density cusp , @xmath184 .
cohn ( 1980 ) found @xmath185 for single - component clusters . in multicomponent clusters
, the most massive component eventually dominates the central part , i.e. , strong mass segregation occurs , owing to a tendency toward equipartition of energy .
thereafter , the most massive component collapses as in single - component clusters and lighter components have shallower density cusps ( e.g. , inagaki , wiyanto 1984 ; chernoff , weinberg 1990 ) .
the development of the density cusps of the galaxy and common - halo components through gravothermal core collapse is clearly shown in figure 1 .
power - law cusps appear for @xmath186 .
the slopes of the cusps hardly depend on the difference in the initial conditions of fpa1 and fpa2 ; the slope of the galaxy cusp approaches @xmath185 and the slope of the common - halo cusp is much shallower , @xmath187 , as expected .
cohn ( 1985 ; see also chernoff , weinberg 1990 ) found that in deep collapse phases of multimass clusters , the power - law indices @xmath188 of the density profiles of different components are approximately related by @xmath189 where @xmath190 is the particle mass of the most massive component , which dominates the central potential well .
this equation is obtained on the assumption that the distribution function of component @xmath95 is given by @xmath191 with @xmath192 .
this relation for @xmath193 was originally found by bahcall and wolf ( 1977 ) for a steady state solution of the stellar distribution around a central massive black hole . in deriving equation ( [ eq : alphai ] ) the results of numerical simulation ( cohn 1980 ) , @xmath194 and @xmath195 ,
are also used .
equation ( [ eq : alphai ] ) gives @xmath196 as @xmath197 .
figures 1b and 1d show that the actual slope of the common halo cusp is slightly larger than this predicted value .
such deviation from equation ( [ eq : alphai ] ) at small masses was also reported by chernoff and weinberg ( 1990 ) . for models fpa1 and fpa2 , we followed the cluster evolution up to a rather deep collapse phase in order to see self - similar development of the density cusps .
however , we should keep in mind that it is very unlikely that actual galaxy clusters can reach such deep collapse , because galaxies have appreciable finite sizes compared to the cluster sizes . for our model cluster ,
the sizes of the galaxies are @xmath198 , and thus the core radius of the cluster can not be less than @xmath198 . in this section
we show the results of fp models in which mass stripping is included but energy dissipation is neglected , and compare these results with @xmath0-body results .
figure 2a shows the fraction of the common halo mass to the total mass as a function of time for models fpb1 and nb .
figure 2b shows the evolution of the central density of the galaxies and that of the common halo , @xmath199 and @xmath200 . for the @xmath0-body model the galaxy density
is not plotted , because it changes very noisily with time owing to the small number of galaxies ( see figure 6 ) .
the center of the cluster of the @xmath0-body model was chosen to be the density center ( casertano , hut 1985 ) of the common halo particles .
concerning the growth of the common halo mass , @xmath201 , the agreement between fpb1 and nb is pretty good , though the growth speed slows down somewhat earlier in fpb1 than in nb .
actually , the value of the numerical constant , @xmath202 , in equation ( [ eq : ml4 ] ) for fpb1 was chosen to achieve such a good agreement .
the initial growth rate of @xmath201 in the fp model is determined by only the stripping rate for given initial conditions .
this means that we can always adjust @xmath72 to obtain a desirable initial growth rate .
once its value is chosen , however , we have no freedom to adjust later growth of the common halo , which depends on evolving cluster properties .
therefore , the fine tuning of @xmath72 is not responsible for the overall good agreement observed in figure 2a . in principle
, the value of @xmath72 can be determined from the data of @xmath0-body experiments of galaxy galaxy encounters such as given by fm . however , its exact value depends on the detailed structure of the galaxies , which changes with time under the cluster environment . from equations ( [ eq : ml2 ] ) and ( [ eq : ml3 ] ) , we find , for @xmath68 , @xmath203 if @xmath63 , and @xmath204 if @xmath65 . hence @xmath202 corresponds to @xmath205 and 9.1 for @xmath63 and 3 , respectively .
equation ( [ eq : ml0 ] ) with these values of @xmath43 is consistent , in order of magnitude , with the data of fm ( see their figures 7 , 8 , 11 , and 12 ) .
models fpb1 and nb are in rather good agreement in the evolution of @xmath201 , but show a clear difference in @xmath200 evolution : the central density increases much more slowly in fpb1 than in nb .
we will see below that the inclusion of energy dissipation , which makes the treatment of tidal stripping in fp models fully self - consistent , removes this difference .
the initial quick increase in @xmath200 and the initial decrease in @xmath199 are simply due to the fact that mass stripping from the galaxies suddenly starts at @xmath206 . after a few cluster crossing times
, the central galaxy density begins to increase .
mass stripping works only to decrease the galaxy density by removing part of galaxy mass to the common halo .
therefore , the increase in the galaxy density should be due to two - body relaxation . since the processes of core collapse and mass stripping occur simultaneously , the central density of the galaxies and that of the common halo stay comparable for a long time , while the central density is soon dominated by galaxies in models fpa1 and fpa2 where there is no mass stripping .
figure 2b shows that the galaxy density finally exceeds the common halo density at the center even with mass stripping .
we discuss in more detail the process of core collapse with mass stripping in section [ sec : r2 ] . here , we mention the choice of the value of the coulomb logarithm .
as noted above , the central density increase in model fpb1 is significantly slower than that in model nb .
we found that rough agreement in the central density evolution was obtained when we artificially increased the two - body relaxation rate by about a factor of two without changing the mass stripping rate .
increasing the relaxation rate by a factor of two is equivalent to doubling the coulomb logarithm , @xmath207 , i.e. , replacing @xmath208 with @xmath209 in the present case .
there is some uncertainty for the choice of the exact value of @xmath210 , but it is very unlikely that such a large value as @xmath211 is justified .
we now take account of the effects of both mass stripping and energy dissipation in fp models ( sequence fpc in table 1 ) .
the evolution of model fpc1 ( our standard model ) is shown in figure 3 .
the growth of @xmath201 in fpc1 is similar to that in fpb1 , although the former shows a slightly larger growth rate at late epochs .
the difference between fpb1 and fpc1 is clear in the evolution of the central density , which increases more rapidly in fpc1 than in fpb1 .
we now see good agreement between models fpc1 and nb in the central density evolution as well as in the evolution of the common halo mass . it should be recalled that energy dissipation is included consistently with mass stripping , without introducing any additional free parameters . in the following ,
we investigate how sensitively the behavior of fp models depends on the parameters of the mass stripping rate ( @xmath72 , @xmath75 , @xmath46 , and @xmath212 ) .
the probable ranges of these parameters are limited by the results of @xmath0-body experiments and theoretical considerations ( see section [ sec : fpmodel2 ] ) . in figure 4
we compare models fpc1 , fpc2 , and fpc3 , which differ only in the value of @xmath72 ( @xmath213 25 , 15 , 35 ) .
larger @xmath72 gives a higher stripping rate , and consequently faster growth in the mass of the common halo . at early epochs the central density of the common halo also increases more quickly for larger @xmath72 , but at late epochs it increases more slowly because faster mass stripping prevents the density increase more severely .
we conclude that the value of @xmath202 gives the best agreement between the fp and @xmath0-body results .
figure 5 compares models fpc1 , fpc4 , fpc5 , and fpc6 , which differ in the value of @xmath75 , @xmath46 , or @xmath212 ( see table 1 ) . for model fpc4
it is necessary to adjust the value of @xmath72 so that the initial growth rate of the common halo becomes similar to those of the other models . these models are not very different from one another .
therefore , we conclude that the results of fp simulations do not very sensitively depend on the mass - stripping formula as long as the parameters are in reasonable ranges . one noticeable difference is seen in the evolution of the central density of model fpc5 ( @xmath214 ) ; at late epochs the central density increases more slowly compared with the other models .
we find that this is because massive galaxies are much more depleted in model fpc5 . in all models , the mean galaxy mass @xmath215 in the central regions first decreases quickly and then stays roughly constant as the galaxy density increases . at late epochs ,
@xmath216 at the center for model fpc5 and @xmath217 for the others .
this implies that the collapse rate of fpc5 is smaller by about a factor of two .
equation ( [ eq : tts ] ) tells us why such a difference occurs ; tidal stripping time @xmath218 is proportional to @xmath219 for @xmath140 , but @xmath218 does not depend on @xmath215 for @xmath214 .
therefore , while @xmath218 increases as @xmath215 decreases in the @xmath140 models , it stays almost constant ( actually somewhat decreases ) in model fpc5 .
consequently , @xmath215 decreases to a much lower value in fpc5 .
a comparison with the @xmath0-body model in figure 5b supports @xmath140 rather than @xmath214 .
figure 6 shows the mass - density profiles of models nb ( left ) and fpc1 ( right ) at selected epochs .
the density profiles for the galaxies ( middle ) and the common halo ( bottom ) are plotted separately as well as the total density ( top ) . for model nb , the density profiles of the galaxies are rather noisy because of the small number of galaxies
. we should note that a comparison of the profiles of the @xmath0-body and fp models on a length scale of @xmath220 is , at least for the galaxies , meaningless , since the sizes of the galaxies , @xmath44 , is @xmath198 .
thus , for @xmath221 in the @xmath0-body model , we see the internal structure of a galaxy if any galaxy is there , and a void if not . at late epochs , the central region of the cluster
is usually occupied by a few massive galaxies that have sunk there owing to dynamical friction . for @xmath222 , models nb and fpc1 show very similar evolution of the density profiles , apart from fluctuations in the @xmath0-body model . the similarity between the two models is more clearly shown in figure 7 , which directly compares their density profiles at @xmath223 .
in particular , almost perfect agreement is seen in the density profile of the common halo .
the agreement between the fp and @xmath0-body results shown above is enough to convince us that our fp models include all essential physical processes concerned and can describe the cluster evolution rather accurately . in figure 8
we show the logarithmic density gradient , @xmath3 [ see equation ( [ eq : alpha ] ) ] , as a function of the radius for model fpc1 at @xmath223 , together with the density profiles of fpc1 and nb again . for @xmath224 , the value of @xmath3 for the total density varies gently with the radius from @xmath225 to @xmath226 with a median value of @xmath198 .
this implies that the total density profile in this region may be reasonably approximated by a power law , @xmath227 , as sfm99 claimed .
the common halo component dominates the density for @xmath228 . for @xmath221
, the density of the galaxy component overwhelms that of the halo component in the @xmath0-body model , while they are comparable in model fpc1 .
as noted above , this difference arises because in the @xmath0-body model a massive galaxy stays around the cluster center at late epochs and its internal structure is observed on this length scale . in fact , the density of the particles in each galaxy is much larger than the density of the common halo particles for @xmath221 in nb .
this extra density increase in the @xmath0-body model strengthens our impression that the total density profile is given by a power law . it is not clear from figure 8 whether the central density distribution is really described by any single power law , since the central density has not yet increased sufficiently from the initial value . in principle , fp calculations can be ( formally ) continued further , if we neglect the fact that the galaxies have finite sizes .
however , we could actually not continue fp calculations to very advanced stages of core collapse for model fpc1 and other similar models , unlike the case for the models without stripping , owing to numerical instability .
this difficulty of fp calculations is linked to the divergence of the distribution function at the center caused by continuous decrease in the central velocity dispersion ; see section [ sec : r2 ] . in order to see the mass dependence of the galaxy distributions , we divided the galaxies into four mass groups , @xmath229 $ ] , @xmath230 $ ] , @xmath231 $ ] , and @xmath232 $ ] , and plotted the density profile for each group for model fpc1 at @xmath223 in figure 9 .
it is apparent that heavy galaxies are strongly depleted in the central regions , owing to tidal stripping , and have a very flat density profile .
if there were no stripping , the heavy galaxies would dominate the central regions as the result of two - body relaxation .
the development of `` inverse mass segregation '' is clearly shown in figure 10 , which plots the mass - weighted mean galaxy mass , @xmath233 , as a function of the radius for model fpc1 . the mean mass decreases toward the center .
this indicates that mass stripping is more efficient than dynamical friction for heavy galaxies . only at very late epochs ( @xmath234 ) , when the central density of the galaxies has increased sufficiently high , is a slight increase in the mean mass seen at the center .
similar inverse mass segregation was already found by funato et al .
( 1993 ) and sfm99 in @xmath0-body simulations and by merritt ( 1984 ) in fp simulations .
the density profiles of models fpc1c6 are similar , if they are compared at epochs with the same central density .
in particular , the halo density profiles are almost indistinguishable .
however , we show in the next section that the density cusp gradient does depend on the stripping rate .
it is obvious that the gradient , @xmath3 , should be close to 2.2 when mass stripping is almost negligible compared to two - body relaxation . on the other hand , if stripping is infinitely fast , all galaxies will dissolve instantaneously and no further evolution of the cluster can occur based on our present assumptions .
the stripping rates for models fpc1c6 are not very different ( see figures 4 and 5 ) , and hence the difference in their density profiles is insignificant .
in subsection [ sec : r1.3 ] we showed that a shallow density cusp , approximated by @xmath235 , develops in the central regions of a galaxy cluster .
now let us assume that a collapsing core leaves behind a cusp that exactly follows a power - law , @xmath184 .
this implies that the central density grows as @xmath236 , as the core radius @xmath237 decreases with time .
similarly , a dimensional analysis implies that the central velocity dispersion changes as @xmath238 ( but this is valid only for @xmath239 as we discuss later ) .
this relation predicts that the central velocity dispersion increases ( i.e. the core gets hotter ) as the core shrinks if @xmath5 , and that it decreases ( i.e. the core gets cooler ) if @xmath240 . for core collapse in star clusters , detailed numerical calculations (
cohn 1980 ; lynden - bell , eggleton 1980 ) show that @xmath241 , and that actually both the central density and velocity dispersion increase as the core shrinks `` hot collapse '' occurs . for core collapse in galaxy clusters ,
our calculations indicate @xmath240 .
therefore , we expect `` cool collapse '' in this case .
lynden - bell and eggleton ( 1980 ) proved that @xmath3 should be in the range @xmath242 for the existence of any self - similar collapse solution in equal - mass clusters .
the condition @xmath243 comes from the requirement that temperature must decrease outward so that the core can shrink by losing heat to the halo . in multimass clusters , however , heat is also transferred among different mass components .
therefore , it can happen that the core of heavy particles loses energy to light particles and shrinks even if @xmath240 .
let us see in more detail the case of a shallow density cusp , @xmath244 this produces a potential , @xmath245 , that is finite at the center and satisfies @xmath246 if we assume that the velocity distribution is isotropic ( this is not a bad assumption near the center ) , we find that the density distribution ( [ eq : rho.sc ] ) corresponds to the distribution function @xmath247^{-(6-\alpha)/2(2-\alpha)}. \label{eq : df.sc}\ ] ] thus , @xmath85 diverges as @xmath248 .
this divergence causes difficulty in numerical integration of the fp equation , and therefore we had to stop our calculations at relatively early ( low central density ) collapse phases for the models with mass stripping .
the depth of the central potential well becomes shallower as @xmath3 decreases ( tremaine et al .
1994 ) . as is argued by tremaine et al .
( 1994 ) , the dependence of the velocity dispersion , @xmath120 , on the radius for the density cusp represented by equation ( [ eq : rho.sc ] ) changes at @xmath249 .
we find @xmath250 which is the equation expected from a dimensional analysis .
for @xmath251 , the behavior of the velocity distribution is more subtle , since it is dominated by high - energy parts of the distribution function , which must deviate from equation ( [ eq : df.sc ] ) if the total mass is to be finite .
if we consider any finite - mass model , we find @xmath252 ( see also dehnen 1993 ) .
for any @xmath240 , @xmath253 as @xmath254 , and hence cool core collapse is expected . in figure 11
we plot the evolution of the central densities and the central velocity dispersions of the galaxy and common halo components for models fpa1 and fpc1 . here
the velocity dispersion of the galaxies is the mass - weighted average over all the galaxies .
the central density increases with time in both models , but the behavior of the velocity dispersion is very different between fpa1 and fpc1 .
as we expected , the central velocity dispersion of the galaxies does decrease as core collapse proceeds in model fpc1 , while it increases in model fpa1 . in model fpc1 ,
even the central velocity dispersion of the common halo finally decreases when core collapse accelerates .
it is considered that this is because cool halo particles are being supplied from cool central galaxies .
( remember we assumed that the particles stripped from a galaxy initially have the same energy and angular momentum per unit mass as those of its parent galaxy . )
we now consider why the velocity dispersion of the galaxies continue to decrease when mass stripping is included .
spitzer ( 1969 ) showed that equipartition among different mass components can not be achieved if self - gravity of heavy components is strong .
it is instructive to recall here his simple model .
consider a spherical system composed of particles of two different masses , @xmath114 and @xmath255 ( @xmath256 ) .
we assume that the total mass of the light component is much larger than that of the heavy component , @xmath257 , and that the heavy component is concentrated at the center .
then , the virial theorem for component 2 can be written as @xmath258 where @xmath259 is the mean velocity dispersion for component 2 , @xmath260 and @xmath261 are the median radii of components 1 and 2 , respectively , and @xmath262 and @xmath263 are numerical constants .
_ for fixed @xmath264 , @xmath265 , and @xmath260 , _ the right side has a minimum value of @xmath266 at @xmath267 .
the minimum of @xmath268 exists because the heavy particles form a self - gravitating system by themselves when @xmath269 , but behave as a non - self - gravitating system when @xmath270 .
therefore , after the concentration of the heavy component becomes sufficiently high , its velocity dispersion increases as the heavy component loses energy and contracts . however , if we now allow @xmath265 to change with @xmath271 fixed , we find that @xmath272 and @xmath273 as @xmath274 .
this implies that the velocity dispersion of the heavy component can continue to decrease to zero as the core shrinks , if mass stripping proceeds at an appropriate speed .
mass stripping prevents the heavy component from forming an independent self - gravitating system .
we may expect that core collapse changes its nature from cool collapse to hot collapse as tidal stripping becomes less and less important compared with two - body relaxation . to confirm this , we ran models fpc7 , c8 , and c9 , which are the same as model fpc1 but have @xmath213 2.5 , 0.25 , and 0.025 , respectively . in principle , we should change the initial models of galaxies and clusters rather than @xmath72 , which is considered to be roughly constant , in order to see the evolution of clusters with different stripping importance . from equation ( [ eq : ttstdf ] ) we find that stripping becomes less important as @xmath275 increases with
the other parameters fixed .
thus , very low stripping rates imply the condition @xmath276 .
( cf . in star clusters
the escape velocity from a star is much larger than the velocity dispersion of stars . )
however , we should not directly apply the equations for tidal stripping presented in subsections 2.2 and 2.3 to such cases , since those equations are considered to be valid only for @xmath277 . we therefore choose to change @xmath72 just to conveniently control the stripping efficiency in our theoretical experiments .
figure 12 shows the growth of the common halo mass for models fpc1 and fpc7c9 .
it is interesting that the difference in the growth rates at the late times is within an order of magnitude among these models ( @xmath278[dm_{\rm h}/dt ] \sim$]0.03 , 0.08 , 0.05 , and 0.01 for fpc1 , c7 , c8 , and c9 , respectively ) , while @xmath72 varies over three orders of magnitude .
this may be explained by a kind of self - regulation mechanism : if @xmath72 is large , the galaxies quickly become more compact through initial violent stripping , and then the stripping speed slows down ; if @xmath72 is small , the central density can increase to very high values , and this finally increases the stripping rate at the central regions .
figure 13 shows the evolution of the central density and velocity dispersion of the galaxy component for models fpc1 and fpc7c9 .
the transition from cool collapse to hot collapse occurs at @xmath279 . in model
fpc8 the central velocity dispersion increases until the central density increases by about a factor of @xmath280 , but decreases after that .
equation ( [ eq : ttstdf ] ) tells us that in the present models the stripping time eventually becomes shorter than the relaxation time as @xmath87 increases toward infinity .
thus , we expect that for any small value of @xmath72 , tidal stripping will become effective and the velocity dispersion will begin to decrease when the velocity dispersion has increased to a critical value , which is inversely proportional to @xmath72 .
the above results show that , unless the stripping rate is decreased to a very low value compared to the standard value , cool collapse occurs . therefore , under the conditions of real galaxy clusters ( @xmath281 ) , cool collapse is expected to occur .
figures 14 shows the profiles of the density and the logarithmic density gradient @xmath3 for model fpc7 ( @xmath282 ) , and figure 15 shows those for model fpc9 ( @xmath283 ) . in model
fpc7 a shallow density cusp of @xmath240 develops , but this cusp is steeper than the cusp in model fpc1 ( see figure 8) . in model
fpc9 a deep cusp of @xmath243 develops , and @xmath3 approaches 2.2 as the core collapse proceeds .
unlike in no - stripping models fpa , the central density of the halo component also significantly rises in model fpc9 due to continuous production of new halo particles .
figures 16a and b show the evolution of velocity dispersion profiles for models fpc7 and
fpc9 , respectively . in these figures qualitative difference between models fpc7 and fpc9 is clearly seen . in model fpc7
an inverse velocity dispersion gradient develops in the central regions .
in this study , we investigated the dynamical evolution of galaxy clusters after virialization using fp models and compared these models with the @xmath0-body models of sfm99 and sfm00 .
our fp models included the effects of gravitational two - body encounters between galaxies and between galaxies and common halo particles which are much lighter than the galaxies .
furthermore , tidal stripping of mass from the galaxies to the common halo and accompanying dissipation of the orbital kinetic energies of the galaxies were included using revised cross - section formulae .
we found that the results of the fp models agree very well with those of the @xmath0-body models regarding the growth of the common halo mass and the evolution of the cluster density profiles .
this indicates that all of the important physical processes that occur in the @xmath0-body models are properly included in the fp models .
although there is some ambiguity in the cross sections of mass stripping and energy dissipation , we found that the results of the fp simulations do not sensitively depend on the details of the cross - section formulae . in the evolution of galaxy clusters ,
tidal mass stripping from galaxies is very important . under the initial conditions that sfm99 and
sfm00 employed , where the total cluster mass is initially attached to galaxies , more than half of the total mass turns into the common halo during a first few crossing times .
heavy galaxies sink toward the cluster center owing to dynamical friction .
however , a high density of galaxies results in strong stripping , and thus it is usually seen in our models that the mean mass of galaxies decreases toward the center , which is inverse mass segregation .
dissipation of the orbital kinetic energies of galaxies due to tidal stripping also plays an important role in accelerating the sinking of the galaxies toward the cluster center , if the velocity dispersion of the stars in the galaxies is not very small compared to that of the galaxies in the cluster . as the central density increases , a cusp profile , approximated by a power law , @xmath184 , develops at the core region . in the model clusters of sfm99 and sfm00 , @xmath3 for the total density
is @xmath2841 .
this density distribution consists of the galaxy component with larger @xmath3 ( still @xmath285 ) and the common halo component with smaller @xmath3 .
the development of the density cusp is a consequence of the collisional evolution of galaxy clusters .
the slope of the density cusp , @xmath3 , depends on the ratio of the stripping rate to the relaxation rate .
when the effect of stripping is negligible compared to relaxation , `` hot core collapse '' , which is usual gravothermal collapse , occurs : i.e. , as the core shrinks both the central density and velocity dispersion increase , and a deep cusp with @xmath286 develops .
when stripping is important , `` cool core collapse '' occurs : while the central density increases , the central velocity dispersion decreases , and consequently a shallow cusp ( @xmath240 ) develops .
faster tidal stripping results in shallower cusps . under the conditions of real galaxy clusters , tidal
stripping is rather effective and cool collapse is expected to occur .
one might think that this conclusion is in contradiction with the results of sfm00 , who showed using @xmath0-body simulations that the profile with @xmath287 ( 11.5 ) develops regardless of the initial models of galaxies and clusters .
actually , their results are consistent with the results of this paper .
sfm00 varied the initial models of galaxies and clusters adopting plummer , king ( king 1966 ) , and hernquist models ( hernquist 1990 ) , but did not vary the number of galaxies and the ratio of the galaxy virial radius to the cluster virial radius . as a result
, the growth rates of the common halo mass , or the mean stripping rates are similar among all of their runs .
our fp simulations show that while @xmath3 depends on the stripping rate , it does so only weakly .
therefore , it is no wonder that no significant difference in @xmath3 has been detected among the @xmath0-body models of sfm00 .
hot core collapse terminates with the formation of hard binaries . because of subsequent energy release from these binaries ,
the core starts to expand ( see , e.g. , spitzer 1987 ) . what is the end state of cool collapse ?
through cool collapse , low - energy galaxies accumulate around the cluster center .
they finally merge and form a massive central galaxy .
the central galaxy will further grow by eating galaxies falling to the center in succession , and this process will continue until all galaxies , except for the central one , disappear .
thus , we will see at last one giant ( cd ? ) galaxy with an enormous halo . in the present study
we did not include the merging process in fp models , and hence we could not follow very late stages of the cluster evolution with fp models .
future @xmath0-body calculations as well as improved fp models will reveal the consequences of cool collapse in more detail .
athanassoula , e. , bosma , a. , lambert , j .- c . , & makino , j. 1998 , mnras , 293 , 369 bahcall , j. n. , & wolf , r. a. 1977 , apj , 216 , 883 barnes , j. e. , & hut , p. 1986 , nature , 324 , 446 bode , p. w. , berrington , r. c. , cohn , h. n. , & lugger , p. m. 1994 , apj , 433 , 479 casertano , s. , & hut , p. 1985 , apj , 298 , 80 chernoff , d. f. , & weinberg , m. d. 1990 , apj , 351 , 121 cohn , h. 1979 , apj , 234 , 1036 cohn , h. 1980 , apj , 242 , 765 cohn , h. 1985 , in dynamics of star clusters , iau symp no.113 , ed j. goodman , p. hut ( dordrecht : d. reidel publishing company ) p161 dehnen , w. 1993 , mnras , 265 , 250 ferguson , h. c. , tanvir , n. r. , & von hippel , t. 1998 , nature , 391 , 461 funato , y. , & makino j. 1999 , apj , 511 , 625 ( fm ) funato , y. , makino , j. , & ebisuzaki , t. 1993 , pasj , 45 , 289 garca - gmez , c. , athanassoula , e. , & garijo , a. 1996 , a&a , 313 , 363 garijo , a. , athanassoula , a. , & garca - gmez , c. 1997 , a&a , 327 , 930 heggie , d. c. , & mathieu , r. d. 1986 , in the use of supercomputers in stellar dynamics , ed p. hut , s. mcmillan ( berlin : springer ) p233 hernquist , l. 1990 , apj , 356 , 359 inagaki , s. , & wiyanto , p. 1984 , pasj , 36 , 391 king , i. 1966 , aj , 71 , 64 lee , h. m. 1987 , apj , 319 , 801 lynden - bell , d. , & eggleton , p. p. 1980 ,
mnras , 191 , 483 lynden - bell , d. , & wood r. 1968 , mnras , 138 , 495 makino , j. 1991 , pasj , 43 , 621 makino , j. , taiji , m. , ebisuzaki , t. , & sugimoto , d. 1997 , apj , 480 , 432 merritt , d. 1983 , apj , 264 , 24 merritt , d. 1984 , apj , 276 , 26 merritt , d. 1985 , apj , 289 , 18 quinlan , g. d. , & shapiro , s. l. 1989 , apj , 343 , 725 scheick , x. , & kuhn , j. r. 1994 , apj , 423 , 566 sensui , t. , funato , y. , & makino , j. 1999 , pasj , 51 , 943 ( sfm99 ) sensui , t. , funato , y. , & makino , j. 2000 , pasj , submitted ( astro - ph/0012092 ) ( sfm00 ) spitzer , l. , jr . 1969 ,
apj , 158 , l139 spitzer , l. , jr . 1987 , dynamical evolution of globular clusters ( princeton : princeton university press ) takahashi , k. 1995 , pasj , 47 , 561 takahashi , k. 1997 , pasj , 49 , 547 tremaine , s. , richstone , d. o. , byun , y .-
dressler , a. , faber , s. m. , grillmair , c. , kormendy , j. , & lauer , t. r. 1994 , aj , 107 , 634 vlchez - gmez , r. , pell , r. , & sanahuja , b. 1994 , a&a , 283 , 37 yepes , g. , & domnguez - tenreiro , r. 1992 , apj , 387 , 27 yepes , g. , domnguez - tenreiro , r. , & del pozo - sanz , r. 1991 , apj , 373 , 336 | we investigated the dynamical evolution of clusters of galaxies in virial equilibrium using fokker planck models and self - consistent @xmath0-body models .
in particular , we focused on the growth of a common halo , which is a cluster - wide halo formed by matter stripped from galaxies , and the development of a central density cusp .
the fokker
planck models include the effects of two - body gravitational encounters both between galaxies and between galaxies and common halo particles .
the effects of tidal mass stripping from the galaxies due to close galaxy galaxy encounters and accompanying dissipation of the orbital kinetic energies of the galaxies were also taken into account in the fokker planck models .
we find that the results of the fokker planck models are in excellent agreement with those of the @xmath0-body models regarding the growth of the common halo mass and the evolution of the cluster density profiles . in the central region of the cluster , a shallow density cusp , approximated by @xmath1 ( @xmath2 1 ) , develops .
this shallow cusp results from the combined effects of two - body relaxation and tidal stripping .
the cusp steepness , @xmath3 , weakly depends on the relative importance of the tidal stripping .
when the effect of stripping is important , the central velocity dispersion decreases as the central density increases and , consequently , a shallow ( @xmath4 ) cusp is formed . in the limit of no stripping ,
usual gravothermal core collapse occurs , i.e. the central velocity dispersion increases as the central density increases with a steep ( @xmath5 ) cusp left .
we conclude from our consideration of the origin of the cusp demonstrated here that shallow cusps should develop in real galaxy clusters . |
recent studies of the coupled schwinger - dyson ( sd ) equations for yang - mills theory have shown @xcite that the gluon propagator @xmath0 and ghost propagator @xmath1 in the landau gauge exhibit _ the power law behavior with a critical exponent @xmath2 _
@xmath3 @xmath4 in the infrared ( ir ) limit @xmath5 for euclidean momentum @xmath6 .
surprisingly , _ the gluon propagator is ir suppressed _ , while _ the ghost propagator is ir enhanced _ for @xmath7 .
however , 1 . the _ precise value of @xmath2 _ is unknown even if the ir power law is correct .
the transverse gluon propagator vanishes for @xmath8 , while it has a non - zero limit for @xmath9 .
2 . there is no _ analytical _ method to connect the ir asymptotic solution to the uv one , although numerical methods exist .
there is no guarantee for uniqueness of the solution obtained under the specific ansatz .
3 . there is no argument for the analytic continuation from euclidean region to minkowski region .
the purpose of this work @xcite is to search the solution which is consistent with the general principles of quantized gauge field theory : * non - perturbative multiplicative renormalizability * analyticity * spectral condition * poincar group structure without using the sd equations .
non - perturbative multiplicative renormalizability for the gluon form factor , @xmath10 @xmath11 yields the rg equation : @xmath12 \\
\times f_r \left({k^2 \over \mu^2},\alpha,\lambda \right ) = 0 . \end{gathered}\ ] ] where @xmath13 , @xmath14 @xmath15 and @xmath16 for a gauge parameter @xmath17 . measured from the negative real axis @xmath18 on the complex @xmath19 plane with singularities on the positive real axis.,height=132 ]
we consider the propagator along the ray in the cut @xmath19 plane ( see figure [ fig : ray ] ) . the asymptotic freedom ( i.e. , validity of perturbation theory for large @xmath19 ) and the general solution of the rg equation lead to the asymptotic form for @xmath20 along the ray : @xmath21 @xmath22 where @xmath23 with @xmath24 and @xmath25 .
_ both functions @xmath26 and @xmath27 vanish as @xmath20 along the ray _ , since @xmath28 .
on the complex @xmath19 plane with singularities on the positive real axis.,height=132 ] suppose i ) the complex function @xmath29 is _ analytic _ in the whole complex plane @xmath30 except for the positive real axis @xmath31 , ii ) @xmath29 _ vanishes asymptotically _ along any ray in the cut complex plane : @xmath32 as @xmath33 , and iii ) @xmath34 is _ real on the real axis _
@xmath35 ( at least for the space - like region @xmath18 ) . then , for a reference point @xmath19 , the _ dispersion relation _ follows @xmath36 we apply this relation to the form factor and the propagator .
from the assumptions : 1 ) poincare group structure ( representation ) , 2 ) spectral condition , and 3 ) completeness condition , the spectral representation for the renormalized gluon propagator follows : @xmath37 where @xmath38 the spectral weight function @xmath39 @xmath40 for @xmath41 has the asymptotic behavior : @xmath42 for @xmath43 . here
the anomalous dimension @xmath44 is gauge dependent . the sign of the asymptotic discontinuity @xmath45 is determined by the ratio @xmath46 , it is negative for @xmath47 .
for @xmath48 , the _ gluon propagator _ has an unsubtracted _ renormalized _ dispersion relation @xmath49 the renormalized @xmath50 is a function of @xmath51 where @xmath52 is the renormalization point and @xmath17 is the gauge parameter . in the similar way ,
the _ gluon dressed function or form factor _ @xmath53 has the _ renormalized _ dispersion relation
@xmath54 two relations are compatible if and only if @xmath55 this is the _ superconvergence relation @xcite in the generalized lorentz gauge _ with an _ initial _ gauge - fixing parameter @xmath17 and the fixed point @xmath56 .
we put the reference point @xmath19 on the euclidean region , i.e. , @xmath57 . _
the gluon form factor vanishes in the euclidean ir limit @xmath58 _ : @xmath59 and _ the gluon propagator converges to a constant _
@xmath60 the spectral function has the form @xmath61 where @xmath62 corresponds to a pole at @xmath63 and @xmath64 is the contribution of the continuous spectrum from more than two particle states , namely , @xmath65 .
the _ finiteness of the ir limit @xmath66 _ given by @xmath67 is interpreted as indicating the _ existence of a massive pole _
( @xmath68 ) , provided that @xmath69 .
( the last integral is finite for @xmath48 . )
thus , we have shown @xcite that , _ for gluon with massive spectrum , the power - series expansions for the gluon propagator and the form factor can be well - defined for small euclidean momenta @xmath70 _ for any gauge parameter . note that @xmath71 corresponds to @xmath9 .
( the gribov limit is @xmath72 . )
to which the cauchy theorem is applied on the complex @xmath19 plane with singularities on the positive real axis.,height=132 ] [ fig : momentcontour ] we recall the result of @xcite .
the _ renormalized _ ghost propagator @xmath73 has a dispersion relation in the arbitrary gauge .
however , the form factor @xmath74 has a dispersion relation only in the landau gauge @xmath75 .
hence , _ the superconvegence relation for the ghost holds , @xmath76 only in the landau gauge @xmath75 . _
_ for @xmath77 , however , the superconvergence relation does not hold _
, since the unsubtracted dispersion relation exists only for the propagator , not for the form factor , in sharp contrast with the gluon case .
if the spectral function @xmath50 has singularities accumulating toward the origin @xmath78 , however , we must replace an integration contour in figure [ fig : cutcomplexplane ] by another contour in figure [ fig : momentcontour ] to avoid the origin .
we conclude that _ the superconvergence for ghost does not hold even in the landau gauge _ @xcite .
from the viewpoint of general principles of qft , _ the transverse gluon can be massive and short range , while the fp ghost is singular and long - range_. the _ ir critical exponent of gluon is @xmath9 _ , since the gluon propagator behaves like @xmath79 _ the ghost propagator has a negative and non - integer exponent . _
it fulfills a sufficient condition for _ color confinement _ due to kugo and ojima .
@xmath80^{-1 } \equiv \lim_{q^2 \rightarrow 0 } [ g(q^2)]^{-1 } = 0 .
\label{ko}\ ] ] supposing the _ existence of ir fixed point _ for the gluon ghost
antighost coupling constant ( without dynamical quarks ) , as suggested from the sd equations @xcite , we have @xmath81 thus the ghost is expected to be a carrier of confinement in the covariant gauge . | we discuss the possible form of gluon and ghost propagators in the infrared region of yang - mills theory in the covariant gauge from the viewpoint of general principles of quantized gauge field theories . |
it is commonly believed that qcd at very high temperatures and/or densities allows for a new phase of strongly interacting matter , the quark - gluon plasma ( qgp ) .
compelling lattice qcd results point in that direction @xcite , and experiments in ultra - relativistic heavy - ion collisions @xcite under way at bnl - rhic @xcite and planned for cern - lhc attempt to glimpse at this elusive state of matter that was presumably present in the early universe @xcite . depending on the nature of the qcd phase transition
, the hadronization process of the expanding qgp generated in a high - energy heavy - ion collision may proceed in a number of different ways @xcite . in the analysis of phase conversion of strongly interacting matter from a chirally symmetric to a chirally broken state in the phase diagram of qcd , a mechanism of supercooling
is usually implied @xcite .
the kinetics of domain formation and growth can then be described by the the mechanisms of nucleation or spinodal decomposition , depending on the degree of supercooling @xcite .
an expanding plasma will probe , in principle , different temperatures and experiment both mechanisms . in the case of the early universe
, it is well known that the time scales for the expansion during the qcd transition are so slow that phase conversion is driven by the nucleation of bubbles of true vacuum inside the metastable phase , as the universe reaches temperatures below a critical value @xcite . on the other hand ,
the qgp presumably created in experiments in ultra - relativistic heavy - ion collisions expands at a pace several orders of magnitude faster than the primordial universe @xcite , and might simply bypass nucleation , entering the domain of spinodal decomposition @xcite . some results from cern - sps and bnl - rhic suggest what has been called _ sudden hadronization _
@xcite or _ explosive behavior _ @xcite . from the theoretical side ,
this phenomenon was recently associated with rapid changes in the effective potential of qcd near the critical temperature , such as predicted , for instance , by the polyakov loop model @xcite , followed by spinodal decomposition @xcite .
clearly , an understanding of the interplay between the typical space and time scales of the expanding plasma is welcome .
some attempts in this direction can be found in refs .
@xcite . in this paper
, we discuss the mechanism responsible for the onset of instabilities in an expanding plasma . as a phenomenological model to mimic the case of the qgp
, we use a relativistic plasma of quarks coupled to a chiral field .
although we derive a phenomenological _ chiral hydrodynamics _ from a variational principle , we do not focus on the numerical solution of the resulting hydrodynamic transport equations .
rather , we consider the role played by chiral waves and pressure waves in the generation of unstable modes .
we show that _ mechanical _ instabilities set in earlier than one would expect from the analysis of the thermodynamic potential decoupled from hydrodynamical modes .
therefore , the spinodal lines are shifted from the values one obtains studying only the chiral degrees of freedom , modifying the phase diagram .
it has been argued @xcite that the first - order transition line which starts at the @xmath0 axis of the baryon chemical potential ( @xmath1 ) vs. temperature ( @xmath2 ) phase diagram for qcd should come to an end at a critical point @xmath3 . beyond this endpoint , for @xmath4 ,
the qcd transition should become a smooth crossover , where spinodal instabilities do not occur .
results from lattice simulations at finite @xmath1 also point in this direction @xcite . here
, we consider two illustrative scenarios for the temperature ( @xmath2 ) vs. baryon chemical potential ( @xmath1 ) phase diagram in which a first order line starts at a point @xmath5 and ends in a critical point , @xmath6 , with @xmath7 , beyond which , for @xmath8 , the phase transition becomes a smooth crossover . to model the mechanism of chiral symmetry breaking present in qcd
, we adopt a simple low - energy effective chiral model : the linear @xmath9-model coupled to quarks @xcite , which in turn comprise the hydrodynamic degrees of freedom of the system . similar approaches , relying on low - energy effective models for qcd and making use of a number of techniques to treat the expanding plasma , can be found in the literature @xcite .
the gas of quarks provides a thermal bath in which the long - wavelength modes of the chiral field evolve .
the latter plays the role of an order parameter in a landau - ginzburg approach to the description of the chiral phase transition @xcite .
the paper is organized as follows . in section
ii , we present our effective field theory model . in section iii , we derive what we call phenomenological _ chiral hydrodynamics _ from a variational formulation of relativistic hydrodynamics . in section iv , we study the onset of instabilities . in section
v , we discuss our results .
let us consider a chiral field @xmath10 , where @xmath9 is a scalar field and @xmath11 are pseudoscalar fields playing the role of the pions , coupled to two flavors of quarks according to the lagrangian : @xmath12q + \frac{1}{2}\partial_{\mu}\phi \partial^{\mu}\phi - v(\phi)\ ; . \label{lagrangian}\ ] ] here @xmath13 is the constituent - quark field @xmath14 and @xmath15 is the quark chemical potential .
the interaction between the quarks and the chiral field is given by @xmath16 and @xmath17 is the self - interaction potential for @xmath18 .
the parameters above are chosen such that chiral @xmath19 symmetry is spontaneously broken in the vacuum .
the vacuum expectation values of the condensates are @xmath20 and @xmath21 , where @xmath22 mev is the pion decay constant .
the explicit symmetry breaking term is due to the finite current - quark masses and is determined by the pcac relation , giving @xmath23 , where @xmath24 mev is the pion mass .
this yields @xmath25 .
the value of @xmath26 leads to a @xmath9-mass , @xmath27 , equal to 600 mev . in mean field theory ,
the purely bosonic part of this lagrangian exhibits a second - order phase transition @xcite at @xmath28 if the explicit symmetry breaking term , @xmath29 , is dropped . for @xmath30 , the transition becomes a smooth crossover from the restored to broken symmetry phases . for @xmath31 , the finite - temperature one - loop effective potential also includes a contribution from the quark fermionic determinant @xcite . in what follows ,
we treat the gas of quarks as a heat bath for the chiral field , with temperature @xmath2 and baryon - chemical potential @xmath1 .
then , one can integrate over the fermionic degrees of freedom , obtaining an effective theory for the chiral field @xmath18 . to compute thermodynamic quantities
, one needs the partition function @xmath32[{\cal d}q][{\cal d}\overline{q } ] e^{-\int_0^{1/t } d\tau \int_{\cal v } d^3x { \cal l}_e}\quad , \label{partition}\ ] ] where @xmath33 is the euclidean lagrangian and @xmath34 is the ( infinite ) volume of the plasma . integrating over the fermions and using a classical approximation for the chiral field , we can write the thermodynamic potential as @xmath35 / t \ } \ ; , \label{tdpot}\ ] ] where @xmath36 is the fermionic euclidean propagator satisfying @xmath37 g_e(\tau,\vec{x};\tau',\vec{x}')= \delta^{(4)}(x - x')\ ; . \label{propagator}\ ] ] from the thermodynamic potential , ( [ tdpot ] ) , one can obtain all the thermodynamic quantities of interest .
the fermionic determinant that results from the functional integration over the quark fields can be calculated to one - loop order in the standard fashion @xcite .
the effective lagrangian for the chiral field , @xmath18 , in the presence of the quark thermal bath is then @xmath38
given the thermodynamic potential , one can derive the total pressure and energy density , and obtain the conserved energy - momentum tensor , @xmath39 , for an expanding quark _
perfect fluid _ by using standard methods of thermal field theory ( see , for instance , ref .
we prefer , instead , adopting an alternative approach to obtain the hydrodynamic equations for our system : the variational formulation @xcite .
this approach provides a natural and unified way of merging chiral and fluid dynamics once the action of the system is specified . for a different treatment of the hydrodynamics of nuclear matter in the chiral limit ,
see @xcite .
the variational formulation of relativistic hydrodynamics has been recently applied to several physical systems , especially in the realms of astrophysics and condensed matter . in the former
it has been used , for instance , to incorporate the effects of local turbulent motion in supernova explosion mechanisms , and can prove to be useful in the analysis of the relativistic motion of blast waves in models for gamma - ray bursts @xcite . in soft condensed matter
, it has been used to study the bubble dynamics in sonoluminescence experiments , in particular in deriving the relativistic generalization of the rayleigh - plesset equation @xcite . in this section ,
we briefly review how to obtain the usual equations of relativistic hydrodynamics within this framework , and derive in detail the case in which we have a chiral field coupled to an expanding quark fluid .
we describe the state of the fluid in terms of the four - velocity @xmath40 , the proper baryon density , @xmath41 , and the proper entropy density , @xmath42 . in what follows
we assume that @xmath43 and @xmath44 are conserved , _
i.e. _ : @xmath45 the action that yields the hydrodynamic equations is given by @xcite @xmath46 \ ; , \label{s_hydro}\ ] ] where @xmath47 is the proper energy density , from which temperature and chemical potential are obtained via the usual thermodynamic relations , @xmath48 and @xmath49 . the equation for the hydrodynamic motion of the fluid
is obtained by imposing the variational principle with respect to @xmath43 , @xmath44 and @xmath50 , under the constraints ( [ cons_ns ] ) and the normalization condition @xmath51 these constraints can be incorporated in the variational principle in terms of lagrangian multipliers , @xmath52 , @xmath53 and @xmath54 , so that one imposes : @xmath55 = 0 , \label{va-0}\end{aligned}\ ] ] equivalently , the fluid dynamics is given by the effective lagrangian , @xmath56 , @xmath57 now the variables @xmath43 , @xmath44 , @xmath50 , @xmath58 , @xmath59 , and @xmath60 are independent . variations with respect to @xmath43 , @xmath44 and @xmath50 yield @xmath61 while variations with respect to @xmath58 , @xmath59 and @xmath60 give simply the constraints ( [ cons_ns ] ) and ( [ co-3 ] ) . from ( [ co-3 ] ) and ( [ va-3 ] ) , one obtains @xmath62 where @xmath63 is the pressure , and identifies @xmath60 as the enthalpy density . using the gibbs - duhem relation , @xmath64 , and the property @xmath65 , it is straightforward to obtain @xmath66 which can be rewritten in the standard form @xmath67 where @xmath68 is the usual energy - momentum tensor of the fluid .
it is important to notice that the effective lagrangian , ( [ l_eff ] ) , evaluated in the proper comoving frame of the fluid , @xmath69 is nothing but _ minus _ the thermodynamic potential of the system .
this fact will be useful to couple the chiral degrees of freedom to the quark fluid motion .
the procedure outlined above for the derivation of relativistic hydrodynamics presents a number of nice features .
first , it can be easily generalized to include other degrees of freedom such as the chiral field . secondly , although we have implicitly used a minkowski metric , the generalization to the case of curved space - times is straightforward .
this will be discussed later , when we consider applications to cosmology .
furthermore , from a practical point of view , once the variational approach is established one can use this method to obtain the optimal parameters for a given family of trial solutions .
let us now describe how to couple a chiral field to the fluid within the framework of the variational principle .
the effective lagrangian for the chiral field , ( [ leff - fai ] ) , should be interpreted as the proper value of the lagrangian where the thermalized quark fluid is at rest .
therefore , eq . ( [ l_eff_prop ] ) suggests that the part corresponding to the quark pressure should be replaced by expression ( [ l_eff ] ) in order to reconstruct the fluid motion of thermalized quarks .
apart from the constraints , we have the following action for the coupled system of the chiral field , @xmath18 , and the thermalized quark fluid motion : @xmath70 here , the energy density , @xmath71 , is related to the thermodynamic potential , @xmath72 , through the legendre transformation @xmath73 . since @xmath74 ,
this is nothing but the general thermodynamic relation @xmath75 .
notice that now the energy density depends on the field variable @xmath18 through @xmath76 and , correspondingly , the second thermodynamic law should read @xmath77 the corresponding gibbs - duhem relation becomes @xmath78 where we have defined a new quantity @xmath79 ( with four components @xmath80 ) : @xmath81 the second line corresponds to maxwell s relation and comes from eq .
( [ gibbs - duhem-2 ] ) .
@xmath79 can be written as @xmath82 where @xmath83/t}+1}+(\mu_{q}\rightarrow-\mu_{q})\ ] ] is a scalar density for quarks . here
@xmath84 stands for the color - spin - isospin degeneracy of the quarks , @xmath85 , and @xmath86 plays the role of an effective mass for the quarks .
thus , our coupled system is described by an effective lagrangian @xmath87 the variation with respect to @xmath18 gives the equation of motion for the chiral field , @xmath88 the variation procedures with respect to the fluid variables are almost exactly the same as before , except for the use of the new gibbs - duhem relation , ( [ gibbs - duhem-2 ] ) .
thus , eq . ( [ div-2 ] ) is modified to @xmath89 or , in the standard form , @xmath90 where @xmath91 is given by eq .
( [ tmunu-1 ] ) and is not the total energy momentum tensor of the system , @xmath92 , which is of course conserved .
note that eq .
( [ divtmunu2 ] ) contains terms from the pure contribution of the field @xmath93 such as @xmath94 and @xmath95 , both in @xmath91 and in @xmath79 .
however , as we can see from the equivalent equation , ( [ eqm - fluid ] ) , some of these terms cancel out . to see more clearly the coupling between the fluid equation and the chiral field @xmath18
, it may be convenient to identify the fluid contribution of the pressure as @xmath96\;,\ ] ] and the energy - momentum tensor of the fluid , @xmath97 then the hydrodynamical motion is written as @xmath98
a numerical analysis of the system of differential equations consisting of ( [ conserv ] ) and the equations of motion for @xmath9 and @xmath99 was performed , for instance , in refs .
@xcite . here , we refrain from this approach and , instead , consider the behavior of small oscillations around equilibrium to study the onset of instabilities . in
what follows , we expand around equilibrium keeping terms up to first order in the perturbation , _
i.e. _ : @xmath100 since the normalization of the four - velocity implies @xmath101 , and @xmath102 , the conservation laws ( [ cons_ns ] ) yield @xmath103 and the chiral field equation of motion becomes @xmath104 where we have defined the ( diagonal ) chiral mass matrix @xmath105_{eq}\ ; .\ ] ] the time component of euler s equation , ( [ eqm - fluid ] ) , is automatically satisfied given the conservation laws above , and its spatial components yield @xmath106 for plane waves of the form @xmath107 , where @xmath108 and @xmath109 stands for @xmath110 , @xmath111 , @xmath112 or @xmath113 , we can rewrite the wave equation , ( [ wave ] ) , and the spatial component of euler s equation , ( [ euler2 ] ) , as ( dropping the tildes ) @xmath114where we have defined @xmath115_{eq}\ ; , \ ] ] @xmath116_{eq}\ ; , \ ] ] evaluated at constant @xmath117 and @xmath18 , and used the conservation laws ( [ conservation2 ] ) in the form ( dropping the tildes ) @xmath118 one can see that pions decouple from the hydrodynamic modes , to first order , and play no major role in the following analysis . in this approximation ,
the interplay between chiral modes and pressure modes happen entirely in the sigma sector of the chiral model .
the dispersion relation then reads @xmath119 for long wavelength fluctuations , we can approximate the roots of ( [ dispersion ] ) by @xmath120 @xmath121 these waves can be identified as pressure modes of frequency @xmath122 , moving with the speed of sound @xmath123 and chiral modes with frequency @xmath124 and mass @xmath125_{eq}\ ; .
\label{msigma}\ ] ] the onset of instabilities takes place when @xmath126 @xcite , _
i.e. _ , either for @xmath127 or for @xmath128 .
we see from eq .
( [ vsound ] ) that pressure modes become unstable before chiral modes do . to study the onset of instabilities within the phase diagram of the low - energy effective model for qcd described previously , we consider two illustrative cases .
the first corresponds to a coupling between quarks and the chiral field small enough for the transition at @xmath129 to be a smooth crossover .
we use @xmath130 , which yields a constituent quark mass of @xmath131mev , about @xmath132 of the nucleon mass . in this case
there is a line of first - order transitions starting at @xmath0 , @xmath133 mev , and ending at a critical point @xmath134 with @xmath135 mev and @xmath136 mev . in the second scenario
we take a larger coupling , @xmath137 , so that even at @xmath138 there is a first - order phase transition , in this case at @xmath139 mev @xcite .
let us now discuss the onset of instabilities in the phase diagram of our effective model . in fig .
[ mut-3.3 ] we plot the phase diagram in the @xmath140-plane for the couplings @xmath130 .
the coexistence line is represented by the solid curve .
the lines where pressure waves become unstable ( @xmath141 ) are shown as dashed curves .
they correspond to the supercooling and superheating spinodals .
the phase diagram in the @xmath143-plane is shown in fig . [ nt-3.3 ] .
the phase border of the coexistence region is represented by the solid line and dashed lines stand for the spinodals .
the sector between the lines on the right of the critical point @xmath134 corresponds to supercooled states in the chirally symmetric phase .
superheated states correspond to the area on the left .
the domain inside the dashed lines corresponds to mechanically unstable states which undergo spinodal decomposition .
results for the second choice of the coupling constant , @xmath137 , are shown in figs .
[ mut-5.5 ] and [ nt-5.5 ] .
as we have mentioned , the difference of this case with respect to the first is that the coexistence line now reaches @xmath129 . because of this
, the spinodals cross each other in the @xmath143 plane , and part of the phase coexistence region can be occupied by both supercooled and superheated states . below the crossing no metastable states exist .
we have shown that the process of domain formation and growth in the phase conversion of strongly interacting matter from a chirally symmetric to a chirally broken state in an expanding qgp is dominated by the exponential increase of hydrodynamical sound - like modes , while chiral - like modes remain stable .
we have mapped the boundaries of the unstable region for two illustrative cases , determining how much supercooling ( or superheating ) is necessary for the onset of spinodal instabilities . in the case of relativistic heavy ion reactions ,
the short time scale of expansion and effects due to the finite size of the system could change significantly our results . in order to address these points
one should perform extensive numerical simulations to study the development of instabilities in different scenarios .
results in this direction will be presented elsewhere .
a. dumitru and r.d .
pisarski , nucl .
a * 698 * , 444 ( 2002 ) .
o. scavenius , a. dumitru and a.d .
jackson , phys .
lett . * 87 * , 182302 ( 2001 ) .
pisarski , phys .
d * 62 * , 111501 ( 2000 ) ; a. dumitru and r. d. pisarski , phys .
b * 504 * , 282 ( 2001 ) .
csernai and i.n .
mishustin , phys .
rev.lett . * 74 * , 5005 ( 1995 ) ; a. abada and j. aichelin , phys . rev . lett .
* 74 * , 3130 ( 1995 ) ; a. abada and m. c. birse , phys .
d * 55 * , 6887 ( 1997 ) ; i.n .
mishustin and o. scavenius , phys .
lett . * 83 * , 3134 ( 1999 ) .
o. scavenius and a. dumitru , phys .
83 * , 4697 ( 1999 ) .
o. scavenius , a. dumitru , e.s .
fraga , j. t. lenaghan and a. d. jackson , phys .
d * 63 * , 116003 ( 2001 ) .
k. paech , h. stoecker and a. dumitru , phys .
c * 68 * , 044907 ( 2003 ) ; k. paech and a. dumitru , phys .
b * 623 * , 200 ( 2005 ) .
zabrodin , l.v .
bravina , l.p .
csernai , h. stocker and w. greiner , phys .
b * 423 * , 373 ( 1998 ) .
p. shukla and a. k. mohanty , phys . rev .
c * 64 * , 054910 ( 2001 ) . c. spieles , h. stocker and c. greiner , phys .
c * 57 * , 908 ( 1998 ) .
fraga and r. venugopalan , physica a * 345 * , 121 ( 2005 ) ; braz . j. phys .
* 34 * , 315 ( 2004 ) .
j. randrup , phys .
lett . * 92 * , 122301 ( 2004 ) ; j. randrup , nucl - th/0406031 ; v. koch , a. majumder and j. randrup , nucl - th/0509030 . e.s .
fraga and g. krein , phys .
b * 614 * , 181 ( 2005 ) .
stephanov , k. rajagopal and e.v .
shuryak , phys .
lett . * 81 * , 4816 ( 1998 ) ; phys .
d * 60 * , 114028 ( 1999 ) .
m. a. halasz , a. d. jackson , r. e. shrock , m. a. stephanov and j. j. verbaarschot , phys .
d * 58 * , 096007 ( 1998 ) ; j. berges and k. rajagopal , nucl .
b * 538 * , 215 ( 1999 ) ; t.m .
schwarz , s.p .
klevansky and g. papp , phys .
c * 60 * , 055205 ( 1999 ) .
o. scavenius , a. mocsy , i. n. mishustin and d.h .
rischke , phys .
c * 64 * , 045202 ( 2001 ) .
z. fodor and s.d .
katz , phys .
b * 534 * , 87 ( 2002 ) ; jhep * 0203 * , 014 ( 2002 ) .
allton _ et al .
_ , phys .
d * 66 * , 074507 ( 2002 ) .
m. gell - mann and m. levy , nuovo cim .
* 16 * , 705 ( 1960 ) ; r. d. pisarski , phys .
* 76 * , 3084 ( 1996 ) .
pisarski and f. wilczek , phys .
d * 29 * , 338 ( 1984 ) .
j. kapusta , _ finite temperature field theory _ ( cambridge university press , cambridge , 1989 ) . h.t .
elze , y. hama , t. kodama , m. makler and j. rafelski , j. phys .
g * 25 * , 1935 ( 1999 ) .
h. a. r. gonalves , s. b. duarte , t. kodama and v. davila , astron .
@xmath144 space science * 194 * , 313 ( 1992 ) ; t. kodama , r. donangelo and m. guidry , int .
c * 9 * , 745 ( 1998 ) . | we study the mechanism responsible for the onset of instabilities in a chiral phase transition at nonzero temperature and baryon chemical potential . as a low - energy effective model ,
we consider an expanding relativistic plasma of quarks coupled to a chiral field , and obtain a phenomenological _ chiral hydrodynamics _ from a variational principle . studying the dispersion relation for small fluctuations around equilibrium
, we identify the role played by chiral waves and pressure waves in the generation of instabilities .
we show that pressure modes become unstable earlier than chiral modes . |
wide band gap nitride semiconductors are now widely used for light emission in the blue and ultraviolet range.@xcite thanks to large band offsets , gan / aln heterostructures are also promising candidates for fast telecommunication devices based on intersubband transitions@xcite or for high - temperature single photon emitters.@xcite one of the specifics of nitride heterostructures is the existence of large internal electric fields due to spontaneous polarization and strains ( piezoelectricity).@xcite these built - in fields might transfer charges in the devices , leading for example to the formation of two dimensional ( 2d ) electron gases at the interfaces between gan and ( ga)aln layers.@xcite they might also separate the electrons from the holes in gan quantum wells and stranski - krastanov ( sk ) quantum dots , thereby reducing the band gap and oscillator strength ( quantum confined stark effect).@xcite it is therefore essential to understand and tailor the electric field in nitride heterostructures to suit a particular application .
whereas 2d layers usually feature a large density of dislocations , nitride nanowires offer the opportunity to make defect - free heterostructures thanks to the efficient strain relaxation.@xcite single gan quantum disks ( qds ) between two aln barriers have for example been grown on top of gan pillars ( 2050 nm diameter ) with plasma - assisted molecular beam epitaxy ( see fig .
[ figsystem]).@xcite the exciton and biexciton luminescence of 1 nm thick gan qds has been observed , showing the potential of such heterostructures for nitride optoelectronic devices.@xcite a strong red shift ( below the bulk gan band gap ) has been subsequently observed for larger disk thickness , a signature of the quantum confined stark effect.@xcite the apparent electric field is however smaller than expected from a comparison with gan / aln quantum wells.@xcite the effects of strain relaxation ( decrease of the piezoelectricity ) and finite lateral size on the electric field,@xcite as well as the screening mechanisms are still unclear .
nitride sk dots have been modeled before with @xmath2 or tight - binding approaches.@xcite the key role played by the charges transferred by band bending , which screen the electric field , has been emphasized in 2d gan / gaaln layers.@xcite in this paper , we model the electronic properties of @xmath0$]-oriented gan / aln nanowire heterostructures in an atomistic tight - binding framework.@xcite we compute strains with a valence force field method,@xcite then the pyro- and piezoelectric field .
we account for band bending with a semi - classical debye - hckel approach , making no assumption about the pinning of the fermi level , but including a distribution of surface states which act as a source or trap of carriers .
we show _ i _ ) that the piezoelectric component of the field can be significantly reduced by the efficient strain relaxation in the nanowire geometry ; _ ii _ ) that the spontaneous and piezoelectric polarizations create an electron gas at the lower gan / aln interface and are likely large enough to create a hole gas in the upper aln barrier . these electron and hole gases screen the electric field in the gan qd and reduce the stark shift ; _ iii _ ) that for carefully chosen dot and barrier thicknesses ( realized experimentally ) the gan qd is empty at equilibrium , consistent with the observation of exciton and biexciton transitions .
we discuss the magnitude of the electric field and the electronic stucture of the qds as a function of the dimensions of the heterostructure .
the paper is organized as follows : we review the methods in section [ sectionmethodology ] , then discuss the electric field in gan / aln nanowire heterostructures in section [ sectionefield ] .
we introduce a simple 1d model for the pyro- and piezoelectric field that reproduces the main trends .
we finally discuss the electronic structure of the gan qds and compare our calculations with the available experimental data in section [ sectionelectronic ] .
we analyze the dependence of the electronic and optical properties of the qds on the geometry of the heterostructures .
nm . ] in this section , we introduce the methods used to compute the structural and electronic properties of the gan / aln heterostructures .
each nanowire is modeled as a 30 nm diameter and 150 nm long@xcite cylindrical gan pillar oriented along @xmath0 $ ] , with the heterostructure on top ( we assume metal - face polarity the case of n - face polarity will be briefly discussed at the end of paragraph [ sectionelectronic ] ) .
the heterostructure consists of a lower aln barrier with thickness @xmath3 , a gan quantum disk with thickness @xmath4 and a upper aln barrier with thickness @xmath5 ( see fig . [ figsystem ] ) .
the dangling bonds at the surface of the nanowire are saturated with hydrogen atoms .
the strains in the nanowire are computed with keating s valence field model.@xcite this model provides an atomistic description of the elasticity of tetrahedrally bonded semiconductors .
it was originally designed for zinc - blende and `` ideal '' wurtzite materials with equal bond lengths and angles,@xcite and has been recently adapted to arbitrary wurtzite materials such as gan and aln.@xcite the elastic energy of the nanowire is minimized with respect to the atomic positions using a conjugate gradients algorithm .
the strains @xmath6 on each atom are then calculated from the atomic positions using a method similar@xcite to ref . .
lrr & gan & aln + @xmath7 ( c / m@xmath8 ) & @xmath9 & @xmath10 + @xmath11 ( c / m@xmath8 ) & @xmath12 & @xmath13 + @xmath14 ( c / m@xmath8 ) & @xmath15 & @xmath16 + @xmath17 ( c / m@xmath8 ) & @xmath18 & @xmath19 + @xmath20 ( @xmath21 ) & @xmath22 & @xmath23 + @xmath24 ( @xmath21 ) & @xmath25 & @xmath26 + @xmath27 ( ev ) & @xmath28 & @xmath29 + @xmath30 ( ev ) & @xmath31 & @xmath32 + @xmath33 ( mev ) & @xmath34 & @xmath35 + @xmath36 ( ev ) & @xmath37 & @xmath38 + @xmath39 ( ev ) & @xmath40 & @xmath41 + @xmath42 ( ev ) & @xmath43 & @xmath44 + @xmath45 ( ev ) & @xmath46 & @xmath47 + the pyro- and piezoelectric polarization density is next computed from the strains on each cation ( ga or al ) : @xmath48 where @xmath49 $ ] , @xmath7 is the spontaneous polarization , and @xmath11 , @xmath14 and @xmath17 are the piezoelectric constants of either gan or aln ( see table [ tabparams ] ) .
poisson s equation for the pyro- and piezoelectric potential @xmath50 : @xmath51 is then solved on a finite difference grid@xcite ( see appendix i of ref . for details ) .
@xmath52 is the dielectric constant ( @xmath53 inside the nanowire and @xmath54 outside ) .
the large pyro- and piezoelectric field @xmath55 in the heterostructure bends the conduction and valence bands and can therefore transfer charges from one part of the system to an other .
it is for example known that the spontaneous polarization in gaaln layers grown on gan pulls out electrons from the gaaln surface , which accumulate in a 2d electron gas at the gan / gaaln interface.@xcite these electrons leave positive charges at the gaaln surface , which can be ionized surface donors , emptied surface states , or even a hole gas .
this redistribution of charges creates , in turn , an electric field opposite to @xmath56 , which can screen the latter to a large extent .
the effects of band bending have been self - consistently computed in a semi - classical debye - hckel approximation . the local density of electrons , @xmath57 , and the local density of holes , @xmath58 are calculated as:@xcite [ eqnp ] @xmath59 \\
p({{\bf r}})&=&n_v f_{1/2}\left[+\beta\left(e_v - ev({{\bf r}})-\mu\right)\right ] \,,\end{aligned}\ ] ] where @xmath24 and @xmath20 are the effective conduction and valence band density of states of the material at point @xmath60 , @xmath30 and @xmath27 are its conduction and valence band edge energies ( see table [ tabparams ] ) , @xmath61 is the total electrostatic potential , and @xmath62 is the chemical potential or fermi energy .
@xmath63 is the fermi integral of order one - half and @xmath64 , where @xmath65 k is the temperature .
we have , additionally , assumed that the nanowires were non - intentionally @xmath66-doped , with a concentration of donor impurities ( silicon , oxygen or vacancies ) @xmath67 @xmath21 .
the density of ionized impurities is:@xcite @xmath68}}\ , , \label{eqnd}\ ] ] where @xmath33 is the binding energy of the donor , which typically ranges from a few tens to a few hundreds of mev .
as mentioned previously , surface states can play an important role in the electrostatics of nitride nanowires .
they might act as a source@xcite or as a trap@xcite of carriers , effectively pinning the chemical potential in the band gap .
little is however known about the electronic structure of nitride surfaces.@xcite on one hand , density functional theory ( dft ) calculations on reconstructed gan and aln surfaces@xcite suggest the existence of occupied ( donor - like ) surface states above the valence band edge and empty ( acceptor - like ) surface states below the conduction band edge ( as expected from simple considerations ) . on the other hand , the extensive literature about 2d electron gases in @xmath0 $ ] gan / gaaln
heterostructures@xcite suggests the existence of dense ( @xmath69 @xmath70ev@xmath71 ) surface donor states only @xmath72 ev below the conduction band of ga@xmath73al@xmath74n alloys ( @xmath75 ) .
although the nature of these surface donors is still debated , oxygen has often been put forward.@xcite it is not clear however that the same picture holds for non - polar ga(al)n surfaces and for surfaces of _
pure _ aln , where the oxide is not the same .
the situation is particularly tricky in nanowires , which expose different ( polar and non - polar ) surfaces .
for the sake of simplicity , we assume in this work the existence of a uniform density of occupied surface states in the @xmath76 $ ] energy range above the valence band edge , and of a uniform density of empty surface states in the @xmath77 $ ] energy range below the conduction band edge .
the density of ionized occupied surface states is therefore : @xmath78}}{1+\frac{1}{2}e^{\beta\left[e_1^+-ev({{\bf r}})-\mu\right]}}\ , , \label{eqnsp}\ ] ] while the density of electrons trapped in the empty surface states is : @xmath79}}{1 + 2e^{-\beta\left[e_2 ^ --ev({{\bf r}})-\mu\right]}}\ , .
\label{eqnsm}\ ] ] @xmath80 and @xmath81 are the density of occupied and empty surface states , respectively ( per unit surface and energy ) .
the values of @xmath36 , @xmath39 , @xmath42 and @xmath45 used in this work are also reported in table [ tabparams ] .
they are representative of _ ab initio _ calculations,@xcite and reproduce the pinning of the fermi level on non - polar gan surfaces.@xcite we have varied @xmath82 between @xmath83 @xmath70ev@xmath71 and @xmath84 @xmath70ev@xmath71 .
their effects will be discussed in paragraph [ sectionefield ] .
we will show , in particular , that the electric field in the qd is weakly dependent on the model for the surface states up to large @xmath80 and @xmath81 . in practice ,
the carrier densities @xmath57 and @xmath58 are computed on each ga , al and n atom , while the surface state densities @xmath85 and @xmath86 are computed on each hydrogen atom .
the charge on each atom is then transferred to the finite difference mesh , and poisson s equation for the total electrostatic potential @xmath61 is solved self - consistently with the newton - raphson method:@xcite @xmath87e\ , .
\label{eqv}\end{aligned}\ ] ] the chemical potential @xmath62 is adjusted to ensure overall charge neutrality of the nanowire .
finally , the electronic structure of the gan qd in the potential @xmath61 is computed with a @xmath1 tight - binding model.@xcite in order to access the relevant states directly , a slice containing the gan qd and 4 nm of each aln barrier is cut from the nanowire .
the bonds broken by this operation are saturated with hydrogen atoms , and a few conduction and valence band states are computed with a jacobi - davidson algorithm.@xcite the convergence of the electronic structure of the qd with respect to the thickness of the aln barriers has been checked .
the above methodology has been implemented in an in - house code called tb_sim .
and the valence band edge energy @xmath88 along the axis of a nanowire with @xmath89 nm and @xmath90 nm .
the reference of energy is the chemical potential @xmath91 .
the position of the aln barriers is outlined in gray , and the top surface is at @xmath92 . ] in this section , we discuss the electric field in gan / aln nanowire heterostructures .
we first analyze a particular case as an example .
we then show that the electric field can be reproduced by a simple 1d model in a wide range of dimensions .
we finally discuss the main trends as a function of the geometry of the heterostructures .
we focus as an illustration on a 30 nm diameter nanowire with a @xmath89 nm thick gan qd and @xmath90 nm thick aln barriers .
we set @xmath93 @xmath70ev@xmath71 .
the conduction band edge energy @xmath94 and the valence band edge energy @xmath88 are plotted along the axis of the nanowire in fig .
[ figbands_8_8 ] .
the reference of energy for this plot is the chemical potential @xmath91 .
the top of the nanowire is located at @xmath92 and the position of the aln barriers is outlined in gray .
the band discontinuities at the gan / aln interfaces are clearly visible .
the heterostructure undergoes a strong vertical electric field , which is almost homogeneous in the aln barriers and in the gan qd .
the latter is empty of carriers ( electrons and holes ) .
the chemical potential is however pinned at the valence band edge at the top surface of the nanowire , and crosses the conduction band at the interface with the gan pillar .
electrons therefore accumulate in the gan pillar , while holes accumulate in the upper aln barrier .
nm , @xmath90 nm ) .
the position of the aln barriers is outlined in gray , and the top surface is at @xmath92 . ]
this redistribution of charges follows from the pyro- and piezoelectric polarizations .
leaving aside piezoelectricity for the moment , the spontaneous polarization in gan is @xmath95 c / m@xmath8 , while the spontaneous polarization in aln is @xmath96 c / m@xmath8 .
this polarization is equivalent to a distribution of charges @xmath97 c / m@xmath8 at the top surface , and @xmath98 c / m@xmath8 at each gan / aln interface .
such a charge distribution , if unscreened , would create huge vertical electric fields and potentials of the order of 10 to 20 v in the nanowire ( see fig .
[ figvp_8_8 ] ) .
the pyro- and piezoelectric field however bends the conduction and valence bands and tends to draw positive charges at the top of the nanowire , which screen the polarization .
the potential actually rises the occupied surface states of the upper aln barrier above the fermi energy .
they therefore empty , leaving positive charges at the surface and releasing electrons in the gan pillar .
at moderate electric field , the surface states would be able to provide enough charge to reach equilibrium , and the chemical potential would lie in the band gap at the top aln surface .
here the electric field is however large enough to empty the @xmath99 @xmath70 occupied surface states .
the chemical potential then sinks into the valence band ; a gas of holes forms at the top surface and provides the missing charges . in the @xmath100 plane containing the axis of the nanowire ( @xmath89 nm , @xmath90 nm ) .
the gan and aln layers are delimited by dotted lines . ]
according to this picture , the charge in the system is mostly distributed at the top aln surface and at the gan / aln interfaces . as a consequence ,
the electric field is typical of a series of parallel plate capacitors , being almost homogeneous in the gan qd and aln barriers .
this is further emphasized in fig .
[ figv_8_8 ] , which represents the electrostatic potential @xmath61 in a @xmath100 plane containing the axis of the nanowire .
the equipotential lines are indeed parallel to the interfaces .
the electrostatic corrections due to the finite cross section of the nanowire are therefore limited in the gan qd and barriers in this range of dimensions .
the effective density of states in the conduction and valence bands of gan and aln are large enough to `` lock '' the potential at the interface with the gan pillar and at the top surface ( as a small variation of potential leads to exponential variations of the charge densities once the fermi energy is in the bands ) .
hence , [ eqmu ] @xmath101 at the top surface , and @xmath102 at the interface @xmath103 with the gan pillar .
the voltage drop @xmath104 across the heterostructure is therefore : @xmath105 the voltage drop across the heterostructure is thus primarily defined by the band gap @xmath106 of gan and the valence band offset between gan and aln once the fermi energy is pinned in the valence band of aln at the top surface .
the validity of this assumption will be discussed in the next paragraphs .
we can derive a simple 1d model for the electric field @xmath107 in the gan qd from the above observations .
for that purpose , we neglect finite size effects ( @xmath108 ) and doping .
we assume that the polarization is homogeneous in the lower aln barrier ( @xmath109 ) , gan quantum disk ( @xmath110 ) and upper aln barrier ( @xmath111 , see fig .
[ figmodel ] ) .
this polarization is equivalent to a charge density @xmath112 on the upper qd interface and @xmath113 on the lower qd interface .
we also assume that the difference of potential @xmath114 across the heterostructure is set by band structure effects [ eq . ( [ eqdv ] ) or equivalent if other pinning of the fermi level ] .
the electric field is then homogeneous in each layer and fulfill the continuity and integral equations : @xmath115 we therefore get : @xmath116 where : @xmath117 is an average polarization discontinuity at the interfaces of the qd .
additionally , the electric field in the barriers is : [ eqeinfsup ] @xmath118+\frac{\delta v}{t_{\rm het}}\\ e_{\rm sup}&=&\frac{1}{\kappa_0\kappa}\left[\delta p_{\rm sup}-\frac{t_{\rm inf}+t_{\rm sup}}{t_{\rm het}}\delta\bar{p}\right]+\frac{\delta v}{t_{\rm het}}\,.\end{aligned}\ ] ] the above equations hold as long as the qd is empty which is also often desired experimentally . neglecting quantum confinement in a first approximation ,
the qd is empty as long as the conduction band edge is above the fermi energy , and the valence band edge below the fermi energy throughout the dot .
assuming @xmath119 ( which is the case here ) , the qd is therefore free from holes if @xmath120 , and free from electrons if @xmath121 . using eqs .
( [ eqmu ] ) , ( [ eqdv ] ) and ( [ eqeinfsup ] ) , these conditions respectively translate into the following constraints on @xmath3 and @xmath5 : [ eqempty ] @xmath122 assuming fixed pyro- and piezoelectric polarizations , @xmath123 and @xmath124 are independent on @xmath3 and @xmath5 for given @xmath4 and @xmath125 . equations ( [ eqempty ] ) then show that the qd can be empty only in a finite range of positions within the heterostructure .
the qd is indeed filled with electrons if it is too far from the surface , and filled with holes if it is too far from the pillar .
note , however , that quantum confinement will practically hinder the charging of the qds by rising the electron and hole energies .
the above constraints thus provide safe bounds for the design of nanowire heterostructures
. equations ( [ eqeqd ] ) and ( [ eqeinfsup ] ) also give an estimate of the charge densities @xmath126 and @xmath127 accumulated at top surface and interface with the pillar , respectively .
the continuity equation for the electric field indeed reads at this interface : @xmath128 where @xmath129 is the electric field in the pillar .
since @xmath129 decreases rapidly away from the interface , @xmath130 assuming that the tip of the nanowire is charge neutral at equilibrium , we then get : @xmath131 we can further split @xmath126 and @xmath127 into polarization and induced charges : @xmath132 where @xmath133 , @xmath134 is the density of the electron gas at the interface with the pillar , and @xmath135 is the density of charges ( ionized surface states+holes ) at the top surface .
the latter thus finally read : [ eqnsnp ] @xmath136 note that @xmath125 must be large enough for the electron gas to form at the interface with the pillar ( @xmath137),@xcite but this is usually not limiting the design of the heterostructure .
we can get a rough estimate of @xmath107 , @xmath134 and @xmath135 by neglecting piezoelectricity [ @xmath138 , @xmath139 .
we then get from equations ( [ eqeqd ] ) and ( [ eqns ] ) : @xmath140+\frac{\delta v}{t_{\rm het } } \label{eqeqd2 } \\ n_p e&=&\frac{t_{\rm inf}+t_{\rm sup}}{t_{\rm het}}\left[p_0({\rm gan})-p_0({\rm aln})\right]-\kappa_0\kappa\frac{\delta v}{t_{\rm het } } \\
n_s e&=&-\frac{t_{\rm qd}p_0({\rm gan})+\left(t_{\rm inf}+t_{\rm sup}\right)p_0({\rm aln})}{t_{\rm het}}-\kappa_0\kappa\frac{\delta v}{t_{\rm het}}\,.\end{aligned}\ ] ] as a simple example , the pyroelectric field in a 4 nm thick qd embedded in an infinitely long nanowire ( @xmath141 ) would be @xmath142 mv / cm . in a finite heterostructure with @xmath143 nm , the induced charges screen this field down to @xmath144 mv / cm .
the density of the electron gas at the interface with the gan pillar is then @xmath145 @xmath70 , while the total density of charges ( surface states+holes ) at the top surface is @xmath146 @xmath70 .
therefore , the fermi level is actually pinned in the valence band of aln as long as the total ( donor ) surface states density is lower than @xmath147 @xmath70 .
this critical density , although large , is yet not unreasonable for bare nanowire surfaces .
we will however give further evidence in paragraph [ sectionelectronic ] that the fermi level is pinned at ( or at least close to ) the valence band edge of aln . and
the valence band edge energy @xmath88 along the axis of a nanowire with @xmath89 nm and @xmath148 nm . the reference of energy is the chemical potential @xmath91 .
the position of the aln barriers is outlined in gray .
the electric field is not constant in the barriers due to the inhomogeneous strains and piezoelectricity . ]
we have tested this simple 1d model against the numerical solution of equations ( [ eqnp])([eqv ] ) .
it gives excellent account of the electric field in the qd when @xmath149 .
the effects of the non - intentional doping are indeed negligible with respect to the amount of charges transferred by the pyro- and piezoelectric field .
this model however tends to overestimate @xmath134 ( as the electric field in the pillar actually decreases over tens of nanometers ) and thus overestimates @xmath135 ( by around 25% in the above example ) . also , the piezoelectric polarization and field become inhomogeneous in thick heterostructures , as the strains are maximum at the interfaces and relax in between@xcite ( see fig .
[ figbands_16_16 ] ) .
the 1d model above is nonetheless very helpful in understanding trends and guiding the design of nanowire heterostructures . in the gan qd as a function of @xmath3 and @xmath5 ( @xmath89 nm ) .
( b ) the average polarization discontinuity @xmath150 deduced from ( a ) and eq .
( [ eqeqd ] ) .
the two dots in parenthesis are charged with holes.,title="fig : " ] + in the gan qd as a function of @xmath3 and @xmath5 ( @xmath89 nm ) .
( b ) the average polarization discontinuity @xmath150 deduced from ( a ) and eq .
( [ eqeqd ] ) .
the two dots in parenthesis are charged with holes.,title="fig : " ] the amplitude of the electric field @xmath151 , computed with eqs .
( [ eqnp])([eqv ] ) as the difference of potential along the qd axis divided by @xmath4 , is plotted in fig .
[ fige]a as a function of @xmath3 and @xmath5 ( @xmath89 nm ) . as expected from eq .
( [ eqeqd2 ] ) , the electric field increases with the total thickness @xmath125 of the heterostructure , and ranges from @xmath152 mv / cm for @xmath153 nm to @xmath154 mv / cm for @xmath155 nm . the electric field is slightly higher than expected from the spontaneous polarization , and does not fulfill the symmetry relation @xmath156 due to piezoelectricity .
this is further emphasized in fig .
[ fige]b , which represents the average @xmath150 obtained by inverting eq .
( [ eqeqd ] ) with the data of fig . [ fige]a .
three horizontal lines are also plotted on this figure for reference .
@xmath157 c / m@xmath8 is the spontaneous polarization discontinuity at the gan / aln interface , which should provide a lower bound for @xmath150 .
@xmath158 c / m@xmath8 is the spontaneous and piezoelectric polarization discontinuity in a heterostructure biaxially strained onto gan , and @xmath159 c / m@xmath8 is the polarization discontinuity in a heterostructure biaxially strained onto aln , which are the expected limits for thin and thick barriers , respectively .
the actual @xmath150 lies between these bounds , as an evidence for piezoelectricity . the piezoelectric field
, though still significant , is lower than in a 2d aln / gan / aln quantum well , due to strain relaxation .
the variations of @xmath150 result from a complex interplay between strain relaxation and charging ( see discussion below ) .
it is nonetheless worthwhile to note that a very good approximation to the electric field can be obtained with a constant @xmath160 c / m@xmath8 ( for given @xmath4 and @xmath161 ) in a wide range of @xmath3 and @xmath5 .
the value of @xmath150 slightly increases with decreasing @xmath4 , up to @xmath162 c / m@xmath8 for @xmath163 nm . as discussed above
, the qds might not be empty if they are are too far from the surface or from the pillar .
( [ eqnp])([eqv ] ) do not , however , properly take quantum confinement into account .
we have therefore refined the assessment of the charge state of the qds with the tight - binding model : we have tentatively assumed that the qds were empty ( setting @xmath164 in the dots so that they are free of carriers ) , and checked the position of the tight - binding band edges with respect to the fermi energy .
we find that all the qds of fig .
[ fige ] are actually empty , except those with @xmath165 nm and @xmath166 nm , which are filled with holes . as expected from eq .
( [ eqns ] ) , the total charge density in the aln barriers increases with @xmath125 , from @xmath167 @xmath70 for @xmath168 nm , to @xmath169 @xmath70 for @xmath90 nm , and @xmath170 @xmath70 for @xmath148 nm . the fermi energy is therefore pinned in the valence band of aln at the top surface in all heterostructures considered here ( @xmath171 @xmath70 ) .
we would finally like to discuss the role of the lateral surface states .
the top surface states play a key role by releasing electrons in the gan pillar , thereby screening the pyro- and piezo - electric field .
the occupied lateral surface states of the upper and lower aln barrier also act as a ( secondary ) source of electrons .
most of these extra electrons ( as well as the donor electrons ) are , however , trapped by the empty lateral surface states of the gan pillar . as a consequence
, the gan pillar is effectively depleted far away from the heterostructure , and the fermi level is pinned @xmath172 ev below the conduction band edge ( see fig .
[ figbands_8_8]).@xcite this does not , however , have significant influence on the physics of the heterostructure .
we now discuss the electronic and optical properties of the gan / aln nanowire heterostructures , and compare our results with experimental data .
the dotted line is a guide to the eye .
the corresponding ( @xmath3 , @xmath5 ) are given ( in nm ) between parenthesis .
the red square is the experimental structure discussed in the text . ]
nm ; @xmath90 nm ) .
the gray dots are the al atoms in the aln barriers . ] the tight - binding band gap energy @xmath173 of empty 4 nm thick gan qds is plotted as a function of the electric field @xmath107 in fig .
[ figeg ] .
the corresponding values of @xmath3 and @xmath5 are reported between parenthesis .
the excitonic correction is not included in this calculation and should further decrease the optical band gap by at most @xmath174 mev .
the band gap energy is strongly red - shifted ( below the bulk value ) by the electric field ( stark effect ) .
it depends almost linearly on @xmath107 and spans around 1 ev in the investigated range of @xmath3 and @xmath5 .
the ground - state electron and hole wave functions of a particular qd ( @xmath90 nm ) are plotted in fig .
[ figwfn ] .
as expected , the electron is confined at the upper interface , while the hole is confined at the lower interface of the qd .
the electron and hole are , interestingly , both localized around the axis of the nanowire , where the strains are maximum , hence the piezoelectric field slightly larger than at the surface .
this helps preventing one of the carriers from being trapped by the charged lateral surface states.@xcite finally , we compare our theoretical predictions with the experimental results of ref . .
in this work , the room - temperature luminescence of 1 to 4 nm thick gan qds embedded in 30 nm diameter nanowires showed clear evidence of the quantum confined stark effect .
we have therefore computed the electronic structure of a 1 nm thick qd ( @xmath175 nm , @xmath176 nm ) , of a 2.5 nm thick qd ( @xmath177 nm , @xmath178 nm ) and of a 4 nm thick qd ( @xmath179 nm , @xmath180 nm ) . the geometry and thickness of the barriers were chosen after a detailed analysis of the experimental tem images.@xcite the calculated and experimental band gap energies are plotted in fig .
[ figexp ] .
all the dots are empty , which is consistent with the observation of the biexciton in the 1 nm thick qds.@xcite the electric field ranges from @xmath181 mv / cm for @xmath89 nm to @xmath182 mv / cm for @xmath163 nm .
it is , as expected , much smaller than in gan / aln quantum wells@xcite ( qws ) and stranski - krastanov@xcite dots ( sks ) with similar sizes due to to strain relaxation and screening by the electron gas and surface charges ( see the comparison between qws , sks and qds in fig . [ figexp ] ) .
the calculation reproduces the luminescence energies of the 1 and 2.5 nm thick qds within the error bars , and the downward trend of the electric field with increasing qd size evidenced in the experiment .
still , the calculation underestimates the luminescence energy of the 4 nm thick qds , where the stark effect is the largest , by about @xmath183 ev .
looking at fig .
[ figeg ] , this suggests that the electric field in this qd is overestimated by @xmath184 % .
the reason for this discrepancy is unclear at present . increasing the density of surface states to pin the fermi level in the bandgap of aln increases the electric field and ultimately charges the dots .
the experimental data have , moreover , been collected at low enough excitation power to prevent screening by photogenerated multiple electron - hole pairs.@xcite the calculated electric field might be affected by the electromechanical coupling@xcite ( influence of the electric field on strains ) , by the uncertainties in the pyro- and piezoelectric constants of gan and especially aln , and by their dependence on strains ( non - linear piezoelectricity).@xcite a simple 1d model however shows that the electromechanical coupling should reduce the electric field by at most @xmath185 ( see appendix [ appendix ] ) .
although the nanowires of ref .
are likely metal - polar , their polarity has not actually been assessed experimentally.@xcite interestingly , we would like to point out that a n - face polarity would give a slightly better agreement between theory and experiment . in that case , the pyro- and piezo - electric field are reversed , so that a hole gas forms at the interface between the gan pillar and heterostructure and electrons accumulate at the top aln surface .
the difference of potential across the heterostructure [ eq .
( [ eqdv ] ) ] then becomes @xmath186 ev , which is slightly higher in magnitude than for the metal - face polarity [ @xmath187 ev ] . as a consequence , the pyro- and piezoelectric field in the qd
are better screened , so that the luminescence in the 4 nm thick qds is raised by @xmath188 mev .
further experiments ( for different barrier and qd thicknesses ) , as well as detailed polarity measurements might therefore be needed to get a complete picture .
we have modeled the quantum confined stark effect in @xmath0$]-oriented aln / gan / aln nanowire heterostructures using a tight - binding approach .
we have taken strain relaxation and band bending into account in the calculation of the pyro- and piezoelectric field .
we have shown that strain relaxation reduces the piezoelectric polarization , and that the electric field pulls out electrons from the occupied states of the top surface .
these electrons accumulate in the gan pillar below the heterostructure , thereby screening the pyro- and piezoelectric field .
we suggest that the electric field is likely strong enough to pin the fermi level in or close to the valence band edge of aln at the top surface . as a consequence , the electric field
is significantly reduced with respect to gan / aln quantum wells or stranski - krastanov quantum dots .
this is in agreement with recent experimental data on gan / aln nanowire heterostructures.@xcite the calculation however overestimates the electric field in thick quantum dots , which calls for further experiments with different geometries and detailed polarity measurements .
we have , for this purpose , provided a simple 1d model for the electric field to help the design of such heterostructures .
we thank b. daudin , b. gayral , j. renard and c. bougerol for fruitful discussions .
in this appendix , we give an estimate of the electromechanical coupling and of its effect on the optical band gap of the nanowire heterostructures .
we assume that the heterostructures are biaxially strained onto gan . according to fig .
[ fige]b , this should provide an upper bound to the electric field , hence to the electromechanical coupling .
the in - plane strains @xmath189 in the gan qd and aln barrier are then : following ref .
, the vertical strain @xmath191 reads in each material : @xmath192 where @xmath193 and @xmath194 are the macroscopic elastic constants.@xcite the first term is the `` uncoupled '' elastic strain while the second one describes the feedback of the electric field @xmath195 on the structure ( electromechanical coupling ) . solving eq .
( [ emc ] ) with eqs .
( [ eqeqd])([eqeinfsup ] ) for the electric field yields : in a 4 nm thick qd with @xmath179 nm and @xmath180 nm ( experimental geometry ) .
the electromechanical coupling can therefore reduce the electric field by at most 5.75% . according to fig .
[ figeg ] , a @xmath197 mv / cm decrease of the electric field would account for a @xmath198 mev increase of the bang gap .
additionally , the strains in the gan layer are : using the interband deformation potential @xmath200 ev in gan,@xcite the strains in the coupled system would further increase the band gap by @xmath201 mev . as a whole ,
the electromechanical coupling can not , therefore , be expected to increase the band gap by more than @xmath202 mev .
the same order of magnitude is obtained assuming the heterostructure is biaxially strained onto aln .
99 s. nakamura , m. senoh , s. nagahama , n. iwasa , t. yamada , t. matsushita , y. sugimoto , and h. kiyoku , appl .
. lett . * 69 * , 4056 ( 1996 ) .
f. a. ponce and d. p. bour , nature * 386 * , 351 ( 1997 ) .
a. j. fischer , a. a. allerman , m. h. crawford , k. h. a. bogart , s. r. lee , r. j. kaplar , w. w. chow , s. r. kurtz , k. w. fullmer , and j. j. figiel , appl .
. lett . * 84 * , 3394 ( 2004 ) .
l. nevou , n. kheirodin , m. tchernycheva , l. meignien , p. crozat , a. lupu , e. warde , f. h. julien , g. pozzovivo , s. golka , g. strasser , f. guillot , e. monroy , t. remmele , and m. albrecht , appl .
phys . lett .
* 90 * , 223511 ( 2007 ) .
s. kako , c. santori , k. hoshino , s. gtzinger , y. yamamoto , and y. arakawa , nature mat . * 5 * , 887 ( 2006 ) .
f. bernardini , v. fiorentini , and d. vanderbilt , phys .
b * 56 * , r10024 ( 1997 ) ; a. zoroddu , f. bernardini , p. ruggerone , and v. fiorentini , phys . rev .
b * 64 * , 045208 ( 2001 ) . o. ambacher , j. smart , j. r. shealy , n. g. weimann , k. chu , m. murphy , w. j. schaff , l. f. eastman , r. dimitrov , l. wittmer , m. stutzmann , w. rieger , and j. hilsenbeck , j. appl . phys . * 85 * , 3222 ( 1999 ) .
j. p. ibbetson , p. t. fini , k. d. ness , s. p. denbaars , j. s. speck , and u. k. mishra , appl .
. lett . * 77 * , 250 ( 2000 ) .
h. w. jang , c. m. jeon , k. h. kim , j. k. kim , s .- b .
bae , j .- h . lee , j. w. choi , and j .- l .
lee , appl .
. lett . * 81 * , 1249 ( 2002 ) .
b. jogai , j. appl .
phys . * 93 * , 1631 ( 2003 ) .
g. koley and m. g. spencer , appl .
phys . lett . * 86 * , 042107 ( 2005 ) .
f. widmann , j. simon , b. daudin , g. feuillet , j. l. rouvire , n. t. pelekanos , and g. fishman , phys .
b * 58 * , r15989 ( 1998 )
. j. simon , n. t. pelekanos , c. adelmann , e. martinez - guerrero , r. andr , b. daudin , le si dang , and h. mariette , phys .
b * 68 * , 035312 ( 2003 ) . c. adelmann , e. sarigiannidou , d. jalabert , y. hori , j .- l
. rouvire , b. daudin , s. fanget , c. bru - chevallier , t. shibata , and m. tanaka , appl .
lett . * 82 * , 4154 ( 2003 ) .
wu , y .- y .
lin , h .- h . huang and j. singh , j. appl .
105 * , 013117 ( 2009 ) .
t. bretagnon , p. lefebvre , p. valvin , r. bardoux , t. guillet , t. taliercio , b. gil , n. grandjean , f. semond , b. damilano , a. dussaigne , and j. massies , phys .
b * 73 * , 113304 ( 2006 ) .
e. ertekin , p. a. greaney , d. c. chrzan , and t. d. sands , j. appl . phys .
* 97 * , 114325 ( 2005 ) .
f. glas , phys .
b * 74 * , 121302(r ) ( 2006 ) .
j. renard , r. songmuang , c. bougerol , b. daudin , and b. gayral , nano lett . * 8 * , 2092 ( 2008 ) .
j. renard , r. songmuang , g. tourbot , c. bougerol , b. daudin , and b. gayral , phys .
b * 80 * , 121305(r ) ( 2009 ) .
a. d. andreev and e. p. oreilly , phys .
b * 62 * , 15851 ( 2000 ) .
v. ranjan , g. allan , c. priester , and c. delerue , phys .
b * 68 * , 115305 ( 2003 ) .
j. risti , c. rivera , e. calleja , s. fernndez - garrido , m. povoloskyi , and a. di carlo , phys .
b * 72 * , 085330 ( 2005 ) . o. marquardt , d. mourad , s. schulz , t. hickel , g. czycholl , and j. neugebauer , phys . rev .
b * 78 * , 235302 ( 2008 ) .
s. schulz , a. berube , and e. p. oreilly , phys .
b * 79 * , 081401(r ) ( 2009 ) .
d. p. williams , s. schulz , a. d. andreev , and e. p. oreilly , ieee j. sel .
electron . *
15 * , 1092 ( 2009 ) .
y. m. niquet , phys .
b * 74 * , 155304 ( 2006 ) . y. m. niquet and d. camacho , phys . rev .
b * 77 * , 115316 ( 2008 ) .
d. camacho and y. m. niquet , physica e * 42 * , 1361 ( 2010 ) .
the nanowires , though shorter than in the experiments of ref .
, are long enough to converge strains and potentials .
p. n. keating , phys .
145 * , 637 ( 1966 ) .
r. m. martin , phys .
b * 6 * , 4546 ( 1972 ) .
the strains are defined from the deformations of tetrahedra built from atoms belonging to the same sublattice . c. pryor , j. kim , l. w. wang , a. j. williamson , and a. zunger , j. appl .
83 * , 2548 ( 1998 ) .
i. vurgaftman and j. r. meyer , j. appl . phys .
* 94 * , 3675 ( 2003 ) .
h. morko , _ handbook of nitride semiconductors and devices _ , vol . 1 ( wiley - vch , weinheim , germany , 2008 ) .
t. l. tansley and r. j. egan , phys . rev .
b * 45 * , 10942 ( 1992 ) .
the heterostructure is placed at the center of the finite difference box , with the nanowire protruding out on one side of the box .
von neumann ( zero normal electric field ) boundary conditions are applied on that side of the box , while dirichlet ( @xmath203 ) boundary conditions are applied on the others .
the convergence of the electric field in the gan qd with respect to the size of the box has been carefully checked .
s. m. sze and kwok k. ng , _ semiconductor devices : physics and technology _
( @xmath204 edition , wiley - interscience , new - york , 2006 ) . v. schmidt , s. senz , and u. gsele , appl .
a * 86 * , 187 ( 2007 ) .
b. s. simpkins , m. a. mastro , c. r. eddy jr . , and p. e. pehrsson , j. appl .
phys . * 103 * , 104313 ( 2008 ) .
r. m. feenstra , y. dong , and c. d. lee , j. vac .
b * 23 * , 1174 ( 2005 ) .
j. e. northrup , r. di felice , and j. neugebauer , phys .
b * 55 * , 13878 ( 1997 ) . j. fritsch , o. f. sankey , k. e. schmidt , and j. b. page , phys .
b * 57 * , 15360 ( 1998 ) .
d. segev and c. g. van de walle , europhys
. lett . * 76 * , 305 ( 2006 ) . c. g. van
de walle and d. segev , j. appl .
* 101 * , 081704 ( 2007 ) .
m. s. miao , a. janotti , and c. g. van de walle , phys .
b * 80 * , 155319 ( 2009 ) .
y. dong , r. m. feenstra , and j. e. northrup , j. vac .
b * 24 * , 2080 ( 2006 ) ; appl .
. lett . * 89 * , 171920 ( 2006 ) .
w. h. press , s. a. teukolsky , w. t. vetterling , and b. p. flannery , _ numerical recipes in fortran _
( cambridge university press , 1992 ) .
j. c. slater and g. f. koster , phys . rev .
* 94 * , 1498 ( 1954 ) .
a. di carlo , semicond .
* 18 * , r1 ( 2003 ) . c. delerue and m. lannoo ,
_ nanostructures : theory and modelling _
( springer , new york , 2004 ) .
jancu , f. bassani , f. della sala , and r. scholz , appl .
* 81 * , 4838 ( 2002 ) .
g. l. g. sleijpen and h. a. van der vorst , siam journal on matrix analysis and applications * 17 * , 401 ( 1996 ) ; siam review * 42 * , 267 ( 2000 ) .
y. m. niquet , a. lherbier , n. h. quang , m. v. fernndez - serra , x. blase , and c. delerue , phys .
b * 73 * , 165319 ( 2006 ) .
j. renard , r. songmuang , g. tourbot , c. bougerol , b. daudin , and b. gayral , private communication . b. jogai , j. d. albrecht , and e. pan , j. appl . phys . * 94 * 3984 ( 2003 ) .
b. lassen , d. barettin , m. willatzen , and l. c. lew yan voon , microelectronics journal * 39 * , 1226 ( 2008 ) .
g. bester , x. wu , d. vanderbilt , and a. zunger , phys .
96 * , 187602 ( 2006 ) . y. m. niquet , nano lett .
* 7 * , 1105 ( 2007 ) . | we model the quantum confined stark effect in aln / gan / aln heterostructures grown on top of @xmath0$]-oriented gan nanowires .
the pyro- and piezoelectric field are computed in a self - consistent approach , making no assumption about the pinning of the fermi level , but including an explicit distribution of surface states which can act as a source or trap of carriers .
we show that the pyro- and piezoelectric field bends the conduction and valence bands of gan and aln and transfers charges from the top surface of the nanowire to an electron gas below the heterostructure . as a consequence ,
the fermi level is likely pinned near the valence band of aln at the top surface .
the electron gas and surface charges screen the electric field , thereby reducing the stark effect .
the efficient strain relaxation further weakens the piezoelectric polarization .
we compute the electronic properties of the heterostructures with a @xmath1 tight - binding model , and compare the theoretical predictions with the available experimental data . |
planetary nebulae represent only an ephemeral stage in the late evolution of low- to intermediate - mass stars , the predominant stellar population in late type galaxies , but their rich emission line spectra allow us to quantify the amounts of helium , and heavier elements ( mainly carbon , nitrogen , and @xmath6-process elements ) that these stars produce and return into the interstellar medium ( ism ) .
recent studies have indicated that pne may also contain non - negligible amounts of endogenous oxygen and neon synthesized by intrinsic nucleosynthesis in the stellar progenitors and expelled into the nebulae following the third dredge - up , especially more so in lower metallicity environments such as those of the galactic halo ( pquignot & tsamis 2005 ) and of satellites of the galaxy ( pquignot et al .
2000 ; zijlstra et al . 2006 ; leisy & dennefeld 2006 ) . their importance as vital sources of neutron - capture elements , such as for instance selenium and krypton among others ,
is now observationally confirmed ( pquignot & baluteau 1994 ; sterling et al .
2007 ) . with the advent of 8-m class telescopes ,
pne are increasingly being used as probes of the chemical evolution of local group galaxies ( e.g. , gonalves et al .
2007 ; saviane et al . 2008 ) , while by virtue of their intrinsically bright monochromatic [ o iii ] 500.7 nm emission they have been detected in the intracluster stellar population of the coma cluster at a distance of 100mpc ( gerhard et al . 2005 , 2007 ) .
a focal problem , however , remains in the astrophysics of these objects : namely that large discrepancies are measured between abundances and temperatures obtained from recombination lines and continua _ versus _ those obtained from the bright forbidden lines ( e.g. , tsamis et al .
2003b , 2004 ; liu et al .
2000 , 2004 ; wesson et al . 2005 ) . for the majority of more than 100 pne surveyed thus far by long - slit ( or chelle ) spectroscopy , abundances of the elements carbon , oxygen , nitrogen and neon derived from their optical recombination lines ( orls )
are higher than those derived from the classic ( ` forbidden ' ) collisionally excited lines ( cels ) by factors 23 ( see review by liu 2006 and references therein ) . for about 510 per cent of the surveyed objects however the discrepancies are in the 480 range , with the most pathological case to date being hf2 - 2 ( liu et al .
2006 ) , a southern planetary nebula . in the cases of carbon ,
nitrogen ( tsamis 2002 ; tsamis et al .
2004 ) and oxygen ( liu et al 2001 ; tsamis et al .
2004 ) , large metal overabundances are correlated with lower plasma temperatures derived from orls and continua than from forbidden lines .
the resolution of this problem remains of high priority as it continues to undermine our trust on secure elemental abundance determinations for pne , and for other types of emission line objects ( such as h ii regions ; e.g. tsamis et al 2003a ) . in the overwhelming majority of cases the ratio of orl / cel abundances for a given heavy element ion ( the _ abundance discrepancy factor _ , adf )
is not correlated with the excitation energies , @xmath7 , of the cels involved : meaning that , per individual pn , infrared ( ir ) or ultraviolet ( uv)/optical cels of low and high @xmath7 respectively , yield very comparable ionic abundances , which are uniformly lower than the corresponding abundances of the same ions derived from orls .
the most likely explanation for these spectroscopic results was judged to be the presence of cold plasma regions embedded in the nebular gas in the form of relatively dense , hydrogen - poor condensations clumps or filaments ( liu et al .
2000 ; tsamis 2002 , tsamis et al .
2004 ) . due to their elevated content in heavy elements ( several times solar )
, these plasma regions would have reached photoionization equilibrium at lower temperatures than the ambient ` normal ' composition gas by emitting far - ir lines , the primary nebular thermostat in relatively low temperatures . since the emissivity of heavy element recombination lines
is enhanced at lower temperatures , while at the same time that of the classic forbidden lines is diminished , the hydrogen - poor , heavy element - rich clumps would emit heavy element recombination lines profusely , yielding a truer estimate of the heavy element content of these regions .
whereas the long - slit surveys yielded concrete evidence for elevated recombination - line abundances , the temperature of the suspected hydrogen - deficient clumps was tougher to determine since the diagnostic lines involved are faint and often suffer from blends in lower resolution spectra .
nevertheless , there has been evidence from those studies for temperatures lower by several thousand k than the typical cel temperatures of photoionized nebulae ( tsamis et al 2004 ; liu et al .
this is one aspect of the proposed ` dual abundance model ' solution that renders it self - consistent , something rather lacking from alternative propositions , such as small - scale temperature fluctuations in a chemically homogeneous medium ( e.g. peimbert et al .
2004 ) , which do not invoke a physical process to explain the discrepancies .
a major unresolved issue remains the as yet unknown origin of the high abundance clumps , and their mass relative to the total ionized nebular gas .
heavy element recombination lines in nebulae can be a thousand times or more fainter than hydrogen recombination lines , in sharp contrast to the luminous forbidden lines emitted by the same heavy ions .
we therefore brought to bear on this thorny issue the combination of integral field spectroscopy ( ifs ) and the light collecting power of the very large telescope ( vlt ) .
ifs techniques allow spectra to be obtained from every spatial element in a two - dimensional field of view , combined with imaging capability of the field at any wavelength across a given @xmath8 range .
this is usually achieved by means of an ifu array composed of microlenses coupled to optical fibres or slicers ( e.g. allington - smith 2006 ) .
these are fed into a spectrograph and are subsequently individually imaged on the focal plane of the telescope .
the contiguous array of microlenses ( or spatial pixels ` spaxels ' ) can also be thought of as a series of long - slits stacked together and , in this sense , an ifu of @xmath9 @xmath1 @xmath10 spaxels used on a single telescope is equivalent to using an array of @xmath9 telescopes equipped with long slits of @xmath10 pixels each .
it could be argued that in the former case a given observing programme is executed @xmath9 times faster than when using a classic long - slit spectrograph . in reality
this is not strictly true as invariably there are some light losses associated with ifu fibres and their optical connection to the spectrograph ( particularly at the interfaces with the microlenses and the feed into the slit ) .
overall , ifu observations afford considerable advantages especially in a science programme like ours where faint diagnostic lines are targeted , and where 2d mapping of nebular regions is sought in the hope of isolating suspect metal - rich regions which may _ a priori _ have a random spatial distribution in the nebula . in addition , ifs data can provide the much - needed input to fully realize the capabilities of new plasma codes such as those developed in recent years for the 3d modelling of photoionized nebulae ( e.g. ercolano et al .
2003 ; morisset et al .
for example , such codes can produce simulated 3d or 2d spectral images of model nebulae at arbitrary wavelengths and inclinations that can be readily evaluated using images obtained by ifs .
moreover , high spectral resolution ifs data such as those we obtained for ngc7009 ( walsh et al . in preparation ) can yield detailed information on the velocity field of an object that could allow a 3d kinematical model to be built .
these are some of the wider aspects of this work . here
we pursue a physicochemical analysis of three representative pne belonging to the galactic - disk population with published oxygen adfs ranging from @xmath4 2 10 .
long - slit spectroscopy of the targets has been published : ngc5882 ( tsamis et al . 2003b , 2004 ) ; ngc6153 ( liu et al .
2000 ) ; ngc7009 ( liu et al . 1995 ; hereafter lsbc ) . in those studies , the first two targets were observed with a long slit scanned across the whole nebular surface .
this technique , which was applied throughout the long - slit surveys , yielded _
mean _ relative emission line fluxes for the whole pne , but naturally any variations of the physical properties in the direction perpendicular to the slit could not be sampled .
an analysis along a ngc6153 fixed slit sampling a ` slice ' of the nebula revealed substantial variations of the physical properties over the slit length . typically , however , the formerly employed @xmath4 2 arcsec - wide slits achieved a rather coarse spatial sampling of @xmath41.6 arcsec per pixel , without sampling the seeing which was of the order of 1 arcsec .
we obtained contiguous 2d spatial coverage of these pne at sub - arcsec resolution with the fibre large array multi element spectrograph ( flames)/giraffe ifu argus .
our main aim was to map out the nebular properties and their variations across large nebular areas and to investigate in detail the existence of any heavy element - rich nebular regions , so as to place constraints on models that have been put forward to explain this astrophysical quandary .
the sample pne were chosen because ngc6153 and ngc7009 exhibit the abundance anomaly problem to a high degree ( with known oxygen adfs of @xmath4 10 and 5 respectively ) and are excellent cases for a detailed study , whereas ngc5882 ( with an adf of @xmath4 2 ) was meant to be our ` control ' target as it represents the majority of galactic nebulae that show a moderate degree of abundance discrepancies .
the paper layout is as follows : the specifications of the observational data set are discussed in section 2 , the data reduction methods are presented in section 3 , and the results in sections 4 and 5 .
the discussion and our conclusions appear in section 6 .
an introduction to this study was presented by tsamis et al .
integral field spectra of ngc5882 , 6153 and 7009 were obtained in service mode between march
june 2005 in sub - arcsec seeing with the flames / giraffe argus ifu on the 8.2-m vlt ut2/kueyen ( p.i .
tsamis ; eso period 75 ) .
the observing log is presented in table 1 .
all targets were observed in the 396.4 507.8 nm range with the large @xmath11 @xmath1 @xmath12 ifu .
the argus unit is a rectangular array of 22 @xmath1 14 microlenses fed by optical fibres : the @xmath11 @xmath1 @xmath12 array thus projects 308 spaxels on the sky ; the individual spaxel size is @xmath13
@xmath1 @xmath13 .
eight spaxels are reserved for sky subtraction purposes ( from a total of 14 dedicated sky - fibres ) and three currently correspond to broken fibres ( appearing in the second row of a reconstructed argus image ; see fig .
6 ) ; the remainder provide 297 individual spectra per target .
the giraffe spectrograph is equipped with a 2048 @xmath1 4096 eev ccd which has 15@xmath14 m pixels .
two separate wavelength ranges were observed using the lr02 ( 39644567 ) and lr03 ( 45015078 ) grating set - ups .
the respective spectral resolving powers , @xmath15 ( @xmath16 @xmath8/@xmath17 ) , were 10200 and 12000 .
higher spectral resolution hr04 and hr06 spectra ( @xmath15 @xmath16 32500 ) of ngc7009 were also taken with the small @xmath18 @xmath1 @xmath19 ifu and will be published in a separate paper ( walsh et al . in preparation ) .
total exposure times were 6930sec ( ngc5882 ) , 24000sec ( ngc6153 ) , and 4260sec ( ngc7009 ) ; more time was devoted to ngc6153 than to the other two targets since this nebula has the lowest surface brightness , and exhibits the abundance anomaly to a high degree .
the average seeing during the ngc5882 exposures was 0.6 arcsec full width at half maximum ( fwhm ) ; during six out of ten observing blocks of ngc6153 it was better than 0.8 arcsec , while during the remaining four it was @xmath4 0.9 arcsec . during the ngc7009 exposures the seeing ranged between 0.50.8 arcsec .
the sky transparency conditions were generally in the ` clear ' category .
the positions observed on each pn were given due consideration . in figs .
13 we show the exact positions observed by overlaying argus spectral images taken in the light of h@xmath20 4861 on _ hubble space telescope _ ( _ hst _ ) wfpc2 images of the nebulae .
the argus maps have been drawn to scale in all cases . regarding ngc5882 ( fig . 1 ) , which is the nebula with the smallest apparent size of the three , the ifu array was positioned so that the largest part of the bright shell of the pn as well as part of the fainter halo to the south could be covered .
the centre of the ifu was positioned at @xmath21 @xmath16 15@xmath2216@[email protected] , @xmath25 @xmath16 @xmath2645@xmath2739@[email protected] ( j2000 ) . for ngc6153 ( fig .
2 ) we chose to align the large axis of argus roughly along the minor axis of the pn so as to cover both the bright patch visible on the _ hst _ image in the southeastern outskirts of the nebula and the inner fainter regions closer to the central star
this position also shares the orientation of the 2-arcsec wide fixed long slit employed by liu et al .
the centre of the ifu was positioned at @xmath21 @xmath16 16@xmath2231@[email protected] , @xmath25 @xmath16 @xmath2640@xmath3115@[email protected] ( j2000 ) .
ngc7009 ( fig .
3 ) has an elliptical shape with low ionization emission regions at both ends of its major axis of symmetry which give it a bipolar appearance .
it shows a clearly defined bright inner shell surrounded by a fainter outer envelope .
we thus chose to place the ifu array so that both the faint regions in the centre of the pn and the sharp rim of the bright inner shell could be covered .
the centre of the ifu was positioned at @xmath21 @xmath16 21@xmath2204@[email protected] , @xmath25 @xmath16 @xmath2611@xmath2721@[email protected] ( j2000 ) . in all cases
the central stars of the pne ( visible in the insets of the _ hst_/argus overlays ) fell within the ifu field of view and were a useful point of reference .
this further allowed us to examine the nebular physical conditions in the immediate vicinity of the stellar nuclei .
lccccc target & date & @xmath8-range & grating & exp .
time + & ( ut ) & ( ) & & ( sec ) + ngc5882 & 23/03/05 & 39644567 & lr02 & 3@xmath1440 + & 23/03/05 & 45015078 & lr03 & 3@xmath1330 + & 29/04/05 & 39644567 & lr02 & 6@xmath1440 + & 29/04/05 & 45015078 & lr03 & 6@xmath1330 + ngc6153 & 29/04/05 & 39644567 & lr02 & 6@xmath1800 + & 29/04/05 & 45015078 & lr03 & 6@xmath1400 + & 30/04/05 & 39644567 & lr02 & 12@xmath1800 + & 30/04/05 & 45015078 & lr03 & 12@xmath1400 + & 01/05/05 & 39644567 & lr02 & 4@xmath1800 + & 01/05/05 & 45015078 & lr03 & 4@xmath1400 + ngc7009 & 02/06/05 & 39644567 & lr02 & 8@xmath1270 + & 02/06/05 & 45015078 & lr03 & 12@xmath1175 +
the spectra were reduced at eso garching with the girbldrs pipeline in a process which included bias removal , localization of fibres on the flats , extraction of individual fibres , wavelength calibration and rebinning of the thar lamp exposures , and the full processing of the science frames which resulted in cosmic - ray cleaned , flat - fielded , wavelength - rebinned spectra . in practice
the pipeline actions were initiated by the creation of a master bias frame from the raw biases ( biasmast pipe ) which was subtracted from the science exposures at a later reduction step .
the localization of fibres on the flat field frames followed ( locmast pipe ) .
the full master wavelength solution was produced via the extraction , wavelength calibration and rebinning of the raw th - ar exposures ( wcalmast pipe ) .
this was followed by the full processing of the science frames via the extract recipe which , by using the products of the previous steps , subtracts the master bias from the science exposures , takes out an average rejecting any cosmic ray hits , performs the flat - fielding , and finally rebins the extracted spectra in wavelength space . in this way , four to six science frames per observing block were processed together and the resulting frame was corrected for atmospheric extinction within iraf .
the airmass during the exposures was @xmath4 1.1 . when this varied by more than @xmath4 0.1 during the course of a single observing block ( 1 hr ) the frames were extracted individually and were corrected for atmospheric extinction one at a time .
they were subsequently averaged using the imcombine task which performed the cosmic ray rejection as well .
the flux calibration was achieved using exposures of the white dwarf eg 274 using the tasks calibrate and standard .
the flux standard exposures were first individually extracted with girbldrs and the spaxels containing the star were summed up to form a 1d spectrum . the sensitivity function was determined using iraf s sensfunc and this was subsequently applied to the combined science exposures .
the sky subtraction was performed by averaging the spectra recorded by the sky fibres and subtracting this spectrum from that of each spaxel in the ifu .
each of the 14 sky - fibre spectra were examined and the average count in each compared .
four sky fibres had higher counts caused by contamination from the simultaneous calibration lamp ( see below ) , and were not included in the computation of the mean sky spectrum .
custom - made scripts allowed us to convert the row by row stacked , processed ccd spectra to data cubes .
the nebular emission line fluxes were generally measured automatically by fitting gaussian profiles via means of the dedicated program spectre which is based on minuit @xmath32 minimization routines ( james 1998 ) .
this enabled us to retrieve from the spectra associated with each spaxel : the line fluxes , the line fwhm , equivalent widths ( ew ) , line centroids , and the respective errors of these quantities which were propagated in all physical properties derived afterwards .
the same program was used to interactively fit the nebular lines in the vicinity of the pn nuclei , and the absorption lines underneath the hi lines over the same spaxels . in figs .
4 and 5 giraffe spectra of the pne from single spaxel extractions in the vicinity of 465.0 nm are shown , covering the full o ii v1 recombination multiplet , a few n ii and c iii orls , n iiibowen fluorescence lines , and he ii @xmath84686 .
the lower panels display the observed spectrum ( black line ) , along with the sum of the fitted continuum ( cubic spline fit ) and the individual gaussians ( red line ) .
the upper panels show the errors bars at each point in the spectrum ( red ) and the residuals from the line and continuum fitting ( black ) . in general
the fit residuals are comparable to the error bars .
the fits in figures 45 are for single gaussian line fits to each emission species ; for ngc6153 , 130 spaxels were fitted with double gaussians , since the velocity splitting ( due to nebular expansion ) of 0.6 could be resolved by two lines of fwhm 0.5 . in the case of ngc5882 ,
30 spaxels were fitted by double profiles .
the gaussian fits to the brightest lines in figures 45 show that the detailed profile deviates from a gaussian and in the case of ngc7009 , for example , the very strong he ii @xmath84686 line shows residuals in the line wings that are greater than the errors .
since only gaussians were used for fitting , without a detailed line profile describing the grating scattering line wings , these line wing features were not fitted in detail .
neglect of these faint features constitutes a loss of less than one per cent to the line flux , and it was considered that a consistent line profile shape for all lines ( i.e. gaussian ) was of highest priority for comparison of relative line fluxes ( on which plasma conditions and abundance determinations that are central to this analysis depend ) . in general ,
the spectral resolution achieved was sufficient to resolve any known blends that affect heavy element orls in lower resolution spectra .
extra care was taken to interactively fit any remaining blended lines and to also fit double gaussian profiles in cases where the spectral resolution was high enough to resolve the expansion of the nebula ( e.g. in the case of ngc6153 ) .
an unforeseen issue arose due to the use of the ` simultaneous calibration lamp ' during the science exposures .
use of this device is not really necessary for spectrophotometric work and it was left on unintentionally during the preparation of phase ii .
the lamp introduced an additive contamination pattern ( francesca primas , private communication ) on certain spaxels of the array ( see fig .
6 ) , which showed up as a mix of broad and narrow features .
all these features are light scattered from fibre - to - fibre and the broad features are presumably the scattering wings of strong lines in the comparison lamp spectrum . since the pn line emission covered the whole of the ifu area for all three nebulae , there was no opportunity with the current data to subtract pn emission - free spaxel spectra in order to remove the contamination pattern .
instead the contaminated spectra were interactively fitted by following the continuum in detail and any spurious narrower ` lines ' were fitted by additional gaussians .
the results were satisfactory , although of course the fit errors on the nebular line fluxes were increased at those spaxels .
the spectra of ngc6153 , which is the nebula with the lowest surface brightness , were most affected and 40 spaxels were interactively fitted and decontaminated in the manner described above . likewise , a scattered light feature appeared towards the last few ccd columns ( long wavelengths ) at the top - right corner of the stacked array .
this was due to a residual ccd signal , which is a known cosmetic defect of the chip .
the signal is additive ( in excess of the bias ) and , as it affected only certain spaxels in the reconstructed argus image , it was fitted with a low - order polynomial and subtracted out .
finally , apart from the three fibres which were broken during eso period 75 , the fibre corresponding to spaxel ( 4 , 4 ) of the reconstructed array is problematic in that it exhibits an anomalous spike in its signal , which often affected spaxels ( 3 , 4 ) and ( 5 , 4 ) too the respective spectra were therefore handled accordingly and generally suffer from large uncertainties .
after the line fluxes had been measured , the reconstructed argus data cubes were corrected for interstellar extinction using the @xmath33(h@xmath20 ) extinction maps derived from a comparison of the observed to the predicted h@xmath34/h@xmath20 line ratio using its case b value from storey & hummer ( 1995 ) .
the galactic extinction law of howarth ( 1983 ) with a ratio of total to selective extinction ( @xmath35 ) of 3.1 was used . since the lr02 and lr03 spectra , containing the h@xmath34 and h@xmath20 lines respectively , were taken in different exposures , and often on different nights , then it was necessary to tie the fluxes together in order to determine the extinction of the maps from the h@xmath34 to h@xmath20 ratio .
this comparison was performed by taking the total flux from the h@xmath34 and h@xmath20 maps and comparing the ratio with that observed from long - slit studies .
for ngc5882 the absolute value of the ratio was taken as 0.410 from the scanned - slit ( mean over the whole pn ) data of tsamis et al .
( 2003b ) ; for ngc6153 , the ratio for the whole nebula ( 0.333 ) was taken from liu et al .
( 2000 ) ; for ngc7009 , the ratio of 0.443 for a value of @xmath33(h@xmath20 ) @xmath16 0.20 was adopted ( cf .
it should be noted that the ifu area does not exactly correspond to the comparison regions from the long - slit data , so the absolute extinction , and hence dereddened line ratios can be systematically offset .
to gain an idea of the photometric quality of the observations , the lr02 and lr03 maps had to be corrected by 47 , 5 and 5 per cent for ngc5882 , 6153 and 7009 respectively .
potential shifts between the lr02 and lr03 maps , caused either by a slightly different pointing between the observations ( although note that the lr02 and lr03 setting were always taken in pairs ; table 1 ) and/or the changing differential atmospheric extinction between the settings , were determined by cross - correlation of features in the h@xmath34 and h@xmath20 maps .
for ngc5882 a shift of 0.25 spaxels in the ifu x - direction could be clearly detected ( this pn has the sharpest spatial features as it was observed under @xmath4 0.6 arcsec seeing ) .
it was nonetheless decided not to correct the maps for this small shift since data at the ifu edges would be compromised . the extinction map for ngc5882 does not show a very pronounced cusp at the position of the shell , as might be expected if the h@xmath34 and h@xmath20 maps were poorly aligned . in the paragraphs that follow we present spectral maps for representative emission lines , the plasma physical conditions which were based on spectral map ratios , and the resulting chemical abundances .
lccc & ngc5882 & ngc6153 & ngc7009 + @xmath33(h@xmath20 ) & 0.45 @xmath36 0.24@xmath37 & 1.15 @xmath36 0.12 & 0.18 @xmath36 0.11 + @xmath38(h@xmath20 ) ( erg@xmath39s@xmath40 ) & 2.66 @xmath41 10@xmath42 & 3.12 @xmath41 10@xmath43 & 3.77 @xmath41 10@xmath42 + @xmath44(he i @xmath84471 ) & 5.47 @xmath36 0.23 & 6.17 @xmath36 0.34 & 4.90 @xmath36 0.48 + @xmath44(he ii @xmath84686 ) & 4.67 @xmath36 3.27 & 17.4 @xmath36 8.68 & 25.60 @xmath36 8.27 + @xmath44(c ii @xmath84267 ) & 0.491 @xmath36 0.071 & 2.95 @xmath36 0.58 & 1.04 @xmath36 0.41 + @xmath44(o ii @xmath84089 ) & 0.203 @xmath36 0.108 & 0.679 @xmath36 0.173 & 0.343 @xmath36 0.098 + @xmath44(o ii
@xmath84649 ) & 0.433 @xmath36 0.066 & 1.88 @xmath36 0.41 & 0.882 @xmath36 0.152 + @xmath44([ne iii ] @xmath83967 ) & 30.09 @xmath36 1.77 & 32.70 @xmath36 1.68 & 44.15 @xmath36 6.71 + @xmath44([o iii ] @xmath84959 ) & 342.3 @xmath36 11.2 & 291.6 @xmath36 14.2 & 372.3 @xmath36 25.4 + @xmath44([ar iv ] @xmath84711 ) & 1.97 @xmath36 0.33 & 2.66 @xmath36 0.75 & 5.04 @xmath36 0.88 + [ ar iv ] @xmath84711/@xmath84740 & 0.867 @xmath36 0.085 & 0.972 @xmath36 0.086 & 0.896 @xmath36 0.034 + [ o iii ] @xmath84959/@xmath84363 & 250 @xmath36 5 & 274 @xmath36 47 & 187 @xmath36 18 + @xmath45 ( @xmath46 ) & 5160 @xmath36 1200 & 3650 @xmath36 1400 & 4780 @xmath36 490 + @xmath47 ( k ) & 9520 @xmath36 190 & 9270 @xmath36 410 & 10340 @xmath36 380 + @xmath48 ( k ) & 9560 @xmath36 2260 & 9180 @xmath36 11290 & 10270 @xmath36 1970 + @xmath5 & 0.00034 @xmath36 0.00021 & 0.00180 @xmath36 0.00234 & 0.00075 @xmath36 0.00016 + @xmath37 : : these are the uncertainties associated with the variation of the quantity across the ifu region ( the rms deviation from the mean ) .
they can not be directly compared to the formal error on an integrated flux measurement for the same region : the uncertainties on the total integrated fluxes and ratios are @xmath49 10 times smaller . in fig .
7 a set of ngc5882 maps is presented showing the observed h@xmath20 , the logarithmic extinction at h@xmath20 , and a sample of representative dereddened spectral maps in the light of he i @xmath84471 , he ii @xmath84686 , [ o iii ] @xmath84959 , [ ne iii ] @xmath83967 , and the recombination lines c ii @xmath84267 and o ii @xmath84649 .
the top two rows of the array appear very bright in the ` unmasked ' h@xmath20 and was judged to be a combination of the effects of differential atmospheric extinction at the edge of the field of view and the fact that the line flux there shows quite a steep gradient this is also evident from a comparison of our line map with the _ hst _ f555w image of the pn which shows a confluence of clumped , relatively high surface brightness emission in that area ( corradi et al .
such a gradient at the edge of the array could introduce errors in ratios of lines originating from the different lr02 and lr03 gratings and so these two rows were masked out and are not considered in the physical analysis of ngc5882 that follows .
we also decided to exclude the faint pn halo from a detailed analysis as the line s / n ratio there , especially for the o ii orls , was much lower than the ifu mean . a mask based on the h@xmath20 signal - to - noise ratio
was therefore applied to exclude the halo region which appears with zero value spaxel intensities ( coloured deep blue ) in the @xmath38(h@xmath20 ) map of fig .
mean values of the physical conditions in the halo are given in sections 4.2.1 and 4.3.1 .
the emission lines peak on the main shell of ngc5882 and especially on the north - west corner which coincides with the bright patch visible on the _ hst _ image ( see inset of fig . 1 ) , our standard measure of comparison .
the pn appears elliptical as in the hubble image , and the southern portion of its halo is clearly registered in the ` unmasked ' h@xmath20 map .
the south segment of its main shell appears faint in all lines and coincides with a low surface brightness blister - like feature on the _ hst _ image which extends towards the halo .
the central regions are also much fainter relative to the main shell .
the [ o iii]and [ ne iii ] maps appear very similar , as do the c ii and o ii orl maps .
he ii @xmath84686 is more concentrated on the west part of the pn than on the east side , in contrast to the he i @xmath84471 emission which traces almost the full nebular ring .
the dust extinction is variable across the nebula ; @xmath33(h@xmath20 ) has a mean value of 0.449 @xmath36 0.034 with an rms deviation of 0.241 . assuming a uniform interstellar extinction screen over the face of the pn
, the central - eastern regions appear to be less dusty than the nebular shell . in fig .
8 a set of ngc6153 maps is presented showing the observed h@xmath20 , the logarithmic extinction at h@xmath20 , and a sample of representative dereddened spectral maps of metallic cels , the helium recombination lines and the c iiand o ii orls .
the argus ifu targeted the brightest portion of the south - east quadrant of the nebula .
the most striking feature in the maps is the high surface brightness patch peaking at around spaxel ( 8 , 17 ) and which is clearly identified with the bright area in the south - eastern quadrant of the _ hst _ image ( cf .
fig . 2 ) .
this area seems to have lower than average dust extinction associated with it , at @xmath33(h@xmath20 ) @xmath50 0.9 the extinction is relatively higher towards the inner part of the nebula , and in the vicinity of the central star , with values closer to 1.3 .
the variations may not be very significant statistically ; @xmath33(h@xmath20 ) has a mean value of 1.146 @xmath36 0.180 with an rms deviation of 0.120 .
a lower surface brightness discontinuity is present in the bright area at coordinates @xmath4 ( 5 , 14 ) in all emission lines except [ ar iv ] @xmath84711 ( which appears bright over a large region ) .
this discontinuity can be identified with the ` thinning ' of the bright patch towards the north - east , something also seen in the inset of fig . 2 .
he i@xmath84471 covers the same area of the ifu as h@xmath20 .
the he ii @xmath84686 peak does not coincide with the h@xmath20-bright region but lies interior to it , by 2 spaxels , towards the inner nebula where the plasma ionization state is higher there is also a local he ii enhancement near the pn nucleus .
the [ ne iii]and [ o iii ] maps are very similar and the lines peak at the same spots .
the c ii and o ii orl maps are also very similar and show peaks near the h@xmath20maximum , but also local peaks in the vicinity of the central star .
spaxels ( 35 , 22 ) for the brightest emission lines of this pn appear with higher values than the mean , and may suffer to some unspecified degree from the same problem as row y @xmath16 22 of ngc5882 . at the same time , the inset of fig .
2 shows that there is a local flux enhancement within that ifu area . in fig .
9 we present a similar set of maps for ngc7009 as for ngc6153 .
the argus array targeted the northern part of the inner shell of the nebula and the long ifu axis lies in approximately the north - east to south - west direction ( cf .
fig . 3 ) .
the most conspicuous feature in all maps is the bright arc that runs from top to bottom and shows high surface brightness spots at either end this is mostly evident in the light of h@xmath20 but is also present in other wavelengths too .
it can be safely identified with the rim of the inner bipolar nebular shell visible in the inset of fig .
the rim appears fainter in the middle in all lines , except he ii @xmath84686 and [ ar iv ] @xmath84711 the dust extinction also drops there .
the extinction constant displays some variability with a mean value of 0.183 @xmath36 0.084 and an rms deviation of 0.111 .
the [ ar iv ] line peaks at the lower segment of the rim , whereas he iishows high surface brightness spots in three positions along it .
characteristically , these spots are all shifted by one spaxel towards the direction of the pn nucleus , relative to local enhancements of the h@xmath20emission , and should correspond to higher ionization zones perhaps associated with the star - facing side of an expanding shell , that is , zones that lie deeper in the nebula from the edge of the main ionization front .
the c ii@xmath84267 and o ii @xmath84649 orl maps have similar appearances and the lines also peak at the lower part of this nebular boundary .
the dereddened emission line maps discussed above were used to compute line ratio maps and hence the electron temperature ( @xmath47 ) and density ( @xmath45 ) , and heavy ion abundances in the pne relative to hydrogen .
the nebular to auroral [ o iii ] @xmath84959/@xmath84363 ratio was used to derive @xmath47 , and the [ ar iv ] @xmath84711/@xmath84740 ratio was used to derive @xmath45 .
atomic data for [ o iii ] and [ ar iv ] were taken from mendoza ( 1983 ) and zeippen et al .
( 1987 ) respectively .
mean values of dereddened ratios of lines used in this study , and the corresponding plasma conditions for the pne are reported in table 2 .
these values are in very good agreement with results from the long - slit surveys .
table 2 also tabulates the fractional mean - square temperature fluctuation parameters , @xmath5 , and average electron temperatures , @xmath48 , across the ifu area for each pn ( that is , in the plane of the sky ) .
we have used equations 11 and 12 of mesa - delgado et al .
( 2007 ) for the derivation of these quantities ( see also rubin et al .
2002 ) , based on the observed ( @xmath47 , @xmath45 ) spatial distributions .
very small values of the @xmath5 parameter in the plane of the sky are found in all three nebulae .
the large uncertainties attached to ( @xmath5 , @xmath48 ) for ngc6153 are dominated by the relatively large uncertainties in the per spaxel determination of @xmath45 for this nebula . in the case of ngc7009
our @xmath5 estimate of 0.00075 is about five times smaller than that derived by rubin et al .
( 2002 ) from _ hst_/wfpc2 imagery of the whole nebula in the [ o iii ] lines .
these results do not preclude the existence of large scale temperature variations along a given line of sight .
for instance , barlow et al . (
2006 ) have presented preliminary evidence for large line - of - sight @xmath47([o iii ] ) variations in ngc6153 ; however , as discussed in previous works , such large scale temperature gradients can not be the main cause of the abundance discrepancy problem ( cf . section 1 ) .
furthermore , barlow et al .
( 2006 and in preparation ) discussed new evidence that o ii orls in ngc6153 and 7009 arise from cool plasma , based on their narrower line widths compared to that of the [ o iii ] @xmath84363 line . in fig .
10 maps for ngc5882 of the dereddened [ ar iv ] and [ o iii]diagnostic ratios are presented , along with their respective uncertainties determined from the line fitting ( which incorporate the uncertainties derived from the pipeline processing ) , and the corresponding nebular @xmath45 and @xmath47maps and their errors .
the uncertainties on the [ ar iv ] ratio are of the order of 10 per cent , whereas those on the [ o iii ] ratio are typically smaller .
the uncertainties on the @xmath45 map are of the order of 10 per cent or less along the bright pn shell , but increase in lower surface brightness regions ( towards the centre , and towards the south - east ) , where the [ ar iv ] lines are weaker .
the mean density is 5160 @xmath36 465 @xmath46 with an rms deviation of 1200 @xmath46 .
the @xmath47 uncertainties are of the order of five per cent or less .
the mean temperature is 9520 @xmath36 50k with an rms deviation of 190k .
the eastern part of the pn has the lowest ( @xmath47 , @xmath45 ) .
the observed range in temperature is @xmath4 1000k ; a couple of local maxima are seen , the most conspicuous of which coincides with the h@xmath20-bright patch on the northwest part of the nebular shell .
the mean electron temperature and density in the masked halo region , contained in the lower part of the ifu aperture [ 86 spaxels excluding ( 34 , 4 ) ] , are somewhat lower than in the main shell , at 9380 @xmath36 150k and 2530 @xmath36 600 @xmath46 respectively .
the @xmath45 and @xmath47 maps of ngc6153 ( with the accompanying @xmath47uncertainties only ) are presented in fig .
11 ( top ) .
this pn shows a temperature range across the ifu of @xmath4 1800k , with a mean of 9270 @xmath36 150k and an rms deviation of 410k .
the demarcation line running diagonally across the error map is caused by the transition from single gaussian profile fits to the [ o iii ] @xmath514363 , 4959 lines ( lower left ) to double peaked profiles ( upper right ) .
the errors are typically larger across the transition zone because of the increased uncertainty of fitting marginally resolved lines with double profiles .
where double gaussian profiles were fitted the combined flux of both was used in the analysis .
the spectral resolution was high enough to allow a separate @xmath47 determination for both nebular velocity components but , as the line s / n ratio was not sufficiently high , this was not attempted .
there is a distinct negative temperature gradient from the nucleus towards the outer nebula ( with increasing y spaxel coordinates out to an offset of 10 arcsec from the nucleus ; see also fig.15b ) .
the highest values are observed near the nucleus , i.e. around spaxel ( 9 , 3 ) , but also in the neighbourhood of ( 2 , 7 ) .
a @xmath47 gradient of very similar magnitude along the minor axis of the pn was also observed by liu et al .
( 2000 ) , whose fixed long - slit position is contained within the ifu orientation ( the ifu long - axis covers roughly half the extent over which those authors performed their spatial analysis ) .
our observations further resolve distinct local minima such as those at @xmath4 ( 13 , 8) .
liu et al . concluded that the major part of the @xmath47 gradient can not be due to contamination of the @xmath84363 line from recombination excitation . in sections 5.4 and 6
the temperature gradient is discussed in the light of the inhomogeneous dust distribution and the higher dust extinction in the inner nebula . from our analysis the @xmath45 surface variation in ngc6153 does not show a clear correlation with distance from the nucleus .
the range in [ ar iv ] density is @xmath4 5000 @xmath46 with a mean of 3650 @xmath36 1520 @xmath46 and an rms deviation of 1400 @xmath46 .
there is some evidence for slightly higher densities over the h@xmath20 bright patch , and lower densities at spaxels beyond its outer edge [ at @xmath4 ( 3 , 19 ) ; bottom - left ifu region ) .
a similar rise in density was also seen by liu et al .
roughly correlating with h@xmath20 surface brightness .
no firm correlation between h@xmath20 and @xmath45 can be however established from the present analysis as the density uncertainties per spaxel are fairly high .
this nebula shows a mean @xmath47 of 10340 @xmath36 90k with an rms deviation of 380k and a range across the ifu of 1700k .
in contrast to ngc6153 , the temperature in ngc7009 ( fig . 11 , bottom ) is not the highest close to the nucleus at map coordinates ( 3 , 12 ) , but at a position @xmath4 5 arcsec north from it at ( 12 , 13 ) .
that position coincides with the he ii - bright high excitation zone discussed in section 4.1.3 ( where the he@xmath3/he ionization fraction shows a local maximum ; see also fig .
14 , bottom row ) , and marks the edge of the ionization front of the inner nebular bubble .
the @xmath45 surface variation across the ifu shows enhancements in the top and centre - bottom regions , approximately corresponding to local h@xmath20-bright spots ; it is lower over spaxels ( 14 , 1320 ) where the h@xmath20 surface brightness is low .
the mean electron density across the ifu is 4780 @xmath36 320 @xmath46 with an rms deviation of 490 @xmath46 and a range of @xmath4 3000 @xmath46 . in this
section maps of the chemical abundances for the pne are discussed .
the o@xmath3/h@xmath52 abundance ratio for all nebulae was derived from the [ o iii]@xmath84959 forbidden line , and also from the o ii @xmath84089 3d4f and @xmath84649 3s3p orls .
the c@xmath3/h@xmath52ratio was derived from the c ii@xmath84267 orl .
the he@xmath52/h@xmath52and he@xmath3/h@xmath52abundance ratios were derived from the he i @xmath84471 and he ii @xmath84686 recombination lines respectively ; the sum of the two yielded total he / h abundance ratios for the three nebulae .
the @xmath47([o iii ] ) and @xmath45([ar iv ] ) maps discussed above were adopted in all cases .
the reader is referred to tsamis et al .
( 2003b , 2004 ) for details on the method of calculating ionic abundance ratios based either on heavy element cels or helium and heavier element orls .
effective recombination coefficients are taken from : storey ( 1994 ) for o ii @xmath84649 ; liu et al .
1995 for o ii @xmath84089 ; davey , storey & kisielius ( 2000 ) for c ii@xmath84267 ( including both radiative and dielectronic processes ) ; storey & hummer ( 1995 ) for he i and he ii . table 3 summarizes the mean values from the maps presented in figs .
1214 , including the abundance discrepancy factors for o@xmath3 . for all targets ,
the mean _ ionic _
c@xmath3/h@xmath52 and o@xmath3/h@xmath52 orl abundance ratios are higher than the corresponding solar abundances of carbon and oxygen ( e.g. lodders 2003 ) , by factors of @xmath4 10 ( ngc6153 ) , @xmath4 3 ( ngc7009 ) , and @xmath4 2 ( ngc5882 ) .
we posit that these overabundance factors point towards the presence of heavy element - rich ( hydrogen - poor ) regions within the nebulae .
the total carbon and oxygen recombination - line abundances , relative to hydrogen , for these regions should be even higher .
the he / h abundances are also higher than solar for all pne ( by up to 40 per cent in the case of ngc6153 ) . in the presented maps the posited hydrogen - deficient zones are identified as undulations in the orl - based diagnostics .
llccc & & ngc5882 & ngc6153 & ngc7009 + o@xmath3/h@xmath52&[o iii ] @xmath84959 & ( 4.28 @xmath36 0.02 ) @xmath41 10@xmath53 & ( 4.09 @xmath36 0.26 ) @xmath41 10@xmath53 & ( 3.54 @xmath36 0.03 ) @xmath41 10@xmath53 + o@xmath3/h@xmath52&o ii @xmath84089 & ( 1.83 @xmath36 0.21 ) @xmath41 10@xmath54 & ( 6.12 @xmath36 0.63 ) @xmath41 10@xmath54 & ( 3.12 @xmath36 0.25 ) @xmath41 10@xmath54 + adf(o@xmath3 ) & @xmath84089/@xmath84959 & 4.33 @xmath36 0.49 & 16.1 @xmath36 2.0 & 9.10 @xmath36 0.74 + o@xmath3/h@xmath52&o ii @xmath84649 & ( 9.10 @xmath361.60 ) @xmath41 10@xmath53 & ( 4.17 @xmath36 0.61 ) @xmath41 10@xmath54 & ( 1.78
@xmath36 0.13 ) @xmath41 10@xmath54 + adf(o@xmath3 ) & @xmath84649/@xmath84959 & 2.14 @xmath36 0.35 & 10.9 @xmath36 2.3 & 5.14 @xmath36 0.35 + c@xmath3/h@xmath52&c ii @xmath84267 & ( 4.68 @xmath36 0.37 ) @xmath41 10@xmath53 & ( 2.81 @xmath36 0.17 ) @xmath41 10@xmath54 & ( 1.01 @xmath36 0.05 ) @xmath41 10@xmath54 + he@xmath52/h & he i @xmath84471 & 0.1042 @xmath36 0.0023 & 0.1182 @xmath36 0.0062 & 0.0950 @xmath36 0.0040 + he@xmath3/h & he
ii @xmath84686 & 0.0039 @xmath36 0.0001 & 0.0142 @xmath36 0.0010 & 0.0213
@xmath36 0.0003 + he / h & ( he@xmath52 @xmath55 he@xmath3)/h & 0.1081 @xmath36 0.0023 & 0.1324 @xmath36 0.0063 & 0.1163 @xmath36 0.0040 + abundance maps for ngc5882 are presented in fig .
12 ; the corresponding error maps are shown for the oxygen diagnostics . the o@xmath3/h@xmath52 ratio based on the [ o iii ] @xmath84959 cel has a mean value of ( 4.28 @xmath36 0.02 ) @xmath41 10@xmath53 with an rms deviation of 2.97 @xmath41 10@xmath56 .
the abundance of this ion shows a local maximum at the position of the pn nucleus at map coordinates ( 6 , 15 ) , surrounded by a @xmath4 19 arcsec@xmath57 area where the abundance is depressed , and two further maxima at positions symmetrically offset by @xmath4 2 arcsec from the nucleus on the north - south direction .
the cel abundance of this ion is further depressed over the bright nebular shell but increases over the mean value on the eastern side of the pn , where the h@xmath20surface brightness and @xmath47([o iii ] ) are low . in the masked halo region
the o@xmath3/h@xmath52 cel ratio is ( 4.52 @xmath36 0.28 ) @xmath41 10@xmath53 .
the o@xmath3/h@xmath52abundance ratio based on the o ii @xmath84649 3s3p orl shows a mean of ( 9.10 @xmath36 1.60 ) @xmath41 10@xmath53 with an rms variation of 1.22 @xmath41 10@xmath53 .
the lowest values of the ratio are found in the south - east corner of the ifu .
the ionic abundance based on this line is perhaps slightly depressed over a four spaxel area in the vicinity of the nucleus with a value of ( 6.00 @xmath36 3.13 ) @xmath41 10@xmath53 . on the other hand ,
the o@xmath3/h@xmath52ratio based on the o ii @xmath84089 3d4f line shows a very different surface distribution relative to the @xmath84649 transition . whereas the mean value is ( 1.83 @xmath36 0.21 )
@xmath41 10@xmath54 with an rms deviation of 9.76 @xmath41 10@xmath53 , the region in the vicinity of the pn nucleus over a nine spaxel area shows a three times higher abundance with a value of ( 5.52 @xmath36 0.61 ) @xmath41 10@xmath54 . from the long slit surveys it was established that , on average , the @xmath84089 transition typically returns higher o@xmath3/h@xmath52 ratios than the @xmath84649 transition when both line emissivities are calculated under the assumption of @xmath47([o iii ] ) for their emitting regions
this effect , when attributed to the different @xmath47 dependence of the emissivities of the two classes of o ii transitions ( 3s3p _ versus _
3d4f ) , has been used to provide evidence for the existence of cold plasma regions in several pne consistent with the existence of hydrogen - deficient regions ( tsamis et al .
2004 ; liu et al .
2004 ) . in tsamis
( in preparation ) we will revisit this issue and present a detailed analysis of the o ii @xmath84089/@xmath84649 ratios from the current data - set based on the latest effective recombination coefficients . in the masked halo region
the o@xmath3/h@xmath52 orl ratio from the @xmath84649 and @xmath84089 lines is ( 6.96 @xmath36 1.21 ) @xmath41 10@xmath53 and ( 1.37 @xmath36 0.49 ) @xmath41 10@xmath54 respectively , rather consistent with the values across the main shell of the nebula .
we also present the respective o@xmath3 adf maps , that is the o@xmath3@xmath84649/o@xmath3@xmath84959 and o@xmath3@xmath84089/o@xmath3@xmath84959 quantities .
these appear different ( caused by the different behaviour of the o@xmath3 orl maps discussed above ) with the former showing local maxima along the nebular shell and the latter showing a sharp maximum near the pn nucleus .
the c@xmath3/h@xmath52 ratio derived from the c ii @xmath84267 orl shows a mean of ( 4.68 @xmath36 0.37 ) @xmath41 10@xmath53 with an rms deviation of 6.82 @xmath41 10@xmath56 .
higher values are observed at the northwest and southeast segments of the h@xmath20-bright shell . the value of the ratio over four spaxels in the vicinity of the central star is ( 4.26 @xmath36 1.11 ) @xmath41 10@xmath53 and is therefore consistent with the mean across the ifu ( the strong stellar continuum in the neighbourhood of the pn nucleus can sometimes introduce higher errors in the line fitting higher spectral resolution data would help in such cases ) . in the masked halo region
this abundance ratio is ( 3.88 @xmath41 0.42 ) @xmath41 10@xmath53 .
finally , fig . 12 ( bottom ) shows the helium abundance maps of ngc5882 .
the mean he / h ratio is 0.1081 @xmath36 0.0023 with an rms deviation of 0.0029 .
the he@xmath52/h@xmath52 abundance ratio is depressed in the central and western nebular regions , but is enhanced in the south and eastern area which is bordering on the pn halo .
in contrast , the he@xmath3/h@xmath52 abundance ratio shows two prominent maxima offset symmetrically on either side of the pn nucleus along the east - west axis .
the he@xmath52/h@xmath52 ratio is inversely correlated with the he@xmath3/h@xmath52ratio across the ifu region ( see table a1 ) . in the masked halo region
the he@xmath52/h@xmath52 and he@xmath3/h@xmath52 ratios are 0.1142 @xmath36 0.0056 and ( 9.14 @xmath36 6.10 ) @xmath41 10@xmath53 respectively . abundance maps for ngc6153 are presented in fig .
the o@xmath3/h@xmath52 ratio based on the [ o iii ] @xmath84959 cel has a mean value of ( 4.09 @xmath36 0.26 ) @xmath41 10@xmath53 with an rms deviation of 8.51 @xmath41 10@xmath56 .
the forbidden line abundance of this ion shows an inverse correlation with @xmath47([o iii ] ) , being high where the plasma temperature is low ( characteristically in the outer nebula mainly ; see section 5 ) and vice - versa .
in contrast , the o@xmath3/h@xmath52ratio based on the o ii @xmath514089 , 4649 orls generally increases towards the inner nebula , being proportional to the @xmath44(o ii)/@xmath44(h@xmath20 ) brightness ratio surface distribution .
both of the o iirecombination line abundance diagnostics show local maxima in the vicinity of the pn nucleus , but as in the case of ngc5882 ( and for the same reasons ) , their respective maps have a different appearance . for this reason ,
the @xmath84089/@xmath84959 adf map shows values as high as @xmath4 40 and the @xmath84649/@xmath84959 adf map values as high as @xmath4 20 at spaxels near the pn nucleus . in the same vein ,
the c@xmath3/h@xmath52 ratio derived from the c ii @xmath84267 orl , with a mean of ( 2.81 @xmath36 0.17 ) @xmath41 10@xmath54 and an rms deviation of 5.72 @xmath41 10@xmath53 , strongly peaks near the pn nucleus as well .
characteristically , both the ionic carbon abundance and the ionic oxygen abundance ( from the o ii @xmath84649 line ) peak in the neighbourhood of spaxel ( 10 , 2 ) about 2 arcsec from the nucleus ; the o@xmath3/h@xmath52ratio from @xmath84649 has a value of ( 6.14 @xmath36 0.81 ) @xmath41 10@xmath54 over a three spaxel area when the mean over the ifu is ( 4.17 @xmath36 0.61 ) @xmath41 10@xmath54 ; this represents a local enhancement of 47 per cent over the mean . across the same region
the c@xmath3/h@xmath52 orl abundance is ( 3.92 @xmath36 0.27 ) @xmath41 10@xmath54 , that is 40 per cent higher than the mean .
it therefore seems that in the case of ngc6153 the pn nucleus , either in itself or via its hard radiation field , is strongly implicated in the causes of the abundance discrepancy , and in the high abundances derived from oxygen and carbon orls .
the peaking of the c iiand o ii abundance diagnostics in the central pn regions had also been inferred from the long - slit analysis ( liu et al .
2000 ) . in fig .
13 ( bottom ) we show the helium abundance maps of ngc6153 .
the mean he / h ratio is 0.1324 @xmath36 0.0063 with an rms deviation of 0.0039 . the he@xmath52/h@xmath52 abundance ratio is lower in the inner nebular regions ( mid - top ifu area ) , but is enhanced over its mean value in the outer nebula .
in contrast , the he@xmath3/h@xmath52 ratio is higher in the inner pn ( see also section 5 ) . as with ngc5882
, there is an inverse linear correlation between the he@xmath52/h@xmath52and he@xmath3/h@xmath52ratios across all spaxels ( see table a1 ) .
the total he / h abundance possibly shows a small positive gradient towards the inner nebula ( top - left ifu area ) . in fig .
14 abundance maps for ngc7009 are presented . the o@xmath3/h@xmath52 ratio based on the [ o iii ] @xmath84959 cel has a mean value of ( 3.54 @xmath36 0.03 ) @xmath41 10@xmath53 with an rms deviation of 4.42 @xmath41 10@xmath56 . in this pn
also , the forbidden - line abundance of the o@xmath3 ion shows an inverse correlation with @xmath47([o iii ] ) , being high where the plasma temperature is low ( see section 5 ) and vice - versa .
the abundance is generally depressed over a triangular region in the middle of the ifu .
in contrast , over the same area the o ii and c ii recombination line abundance diagnostics are enhanced .
there are prominent maxima in the o@xmath3/h@xmath52 orl ratio in the vicinity of the pn nucleus near map coordinates ( 3 , 12 ) where the ratio based on the @xmath84649 line has a value of ( 3.36 @xmath36 0.62 ) @xmath41 10@xmath54 over a two spaxel area ; this constitutes a local enhancement of 89 per cent over the corresponding mean across the map of ( 1.78 @xmath36 0.13 ) @xmath41 10@xmath54 .
the respective adf maps show peaks near the nucleus as well .
the c@xmath3/h@xmath52 ratio peaks close to the nucleus with a value of ( 1.14 @xmath36 0.16 ) @xmath41 10@xmath54 over a four spaxel area , but also near the edge of the h@xmath20-bright rim , some 5 arcsec away from the central star ; it has a value of ( 1.49 @xmath36 0.04 ) @xmath41 10@xmath54 at coordinates ( 11 , 18 ) .
its mean value across the ifu is ( 1.01 @xmath36 0.05 ) @xmath41 10@xmath54 with an rms deviation of 1.57 @xmath41 10@xmath53 . in fig .
14 ( bottom ) the helium abundance maps of ngc7009 are shown .
the mean he / h ratio is 0.1324 @xmath36 0.0063 with an rms deviation of 0.0039 .
the he@xmath52/h@xmath52 abundance ratio is generally low over the triangular area where the o@xmath3 and c@xmath3 orl abundances are enhanced , whereas the he@xmath3/h@xmath52 ratio is correspondingly high there , peaking near the rim of the inner nebular bubble .
there is a strong linear inverse correlation between the he@xmath52/h@xmath52and he@xmath3/h@xmath52ratios in this nebula too ( see table a1 ) .
the total he / h abundance shows slightly lower values in the top of the ifu map , but does not vary across the edge of the inner nebular bubble where the he@xmath3/he ionization fraction peaks .
in this section we investigate several correlations between physical properties of the pne that can be established following this analysis .
evidence for a few of these existed from the long - slit surveys , whereas others are new . a unique feature of this study , however , is that the large number of data points ( @xmath4 300 ) per pn allowed us to put all these on a firm footing for the first time .
moreover , the 2d aspect of this analysis is particularly revealing , and in the following paragraphs it shall be brought to the fore .
we present results for each nebula individually so as to first highlight the differences and then address the similarities ; ngc6153 is examined first . the o@xmath3 adf and the [ o iii ] electron temperature for this pn show a very distinct correlation with distance from the pn nucleus .
we have taken the argus maps of these physical quantities ( figures 11 and 13 ) and binned all non - zero value spaxels at equal radii outwards from the position of the central star ( with a radial increment of one spaxel ) ; in fig .
15a b the result is shown .
as the [ o iii ] temperature rises from the outer to the inner regions of the pn from @xmath4 8400 to 10000k , characteristically the oxygen adf rises as well . the quantity plotted in fig .
15a is the @xmath84649/@xmath84959 adf ( and is also used in figs .
1618 ) , but a linear correlation is obtained for the @xmath84089/@xmath84959 adf as well .
15c shows the relationship between the oxygen adf and @xmath47([o iii ] ) for the 296 spaxels with a measurement of these quantities .
the solid line is an error - weighted least - squares fit to the data with a correlation coefficient ( @xmath58 ) of 0.92 . in fig .
15d the inverse relationship between the o@xmath3/h@xmath52abundance ratio based on the [ o iii ] @xmath84959 line and @xmath47([o iii ] ) is shown ; the correlation is very tight in this case too with a linear correlation coefficient of @xmath260.98 . in figs .
15 and 16 , black solid lines are least - squares fits weighted by the errors , red circles denote data from spaxels with coordinates ( from the corresponding argus maps ) of ( x , y ) @xmath16 ( 114 , 111 ) , whereas blue diamonds correspond to coordinates ( 114 , 1222 ) .
this colour scheme allows us to examine the behaviour of nebular properties in nebular regions that are near and far out from the central star ( red and blue symbols respectively ; cf .
also the orientation of the argus ifu long axis viz .
the position of the pn nucleus in fig . 2 ) .
from the juxtaposition of fig .
15a b and 15c d a consistent picture emerges .
regions closer to the nucleus of ngc6153 have higher temperature , lower forbidden line o@xmath3/h@xmath52abundance and exhibit higher o@xmath3 adfs than regions farther out in the nebula .
some unavoidable intermixing of the colour - coded data is seen due to projection ( line of sight ) effects and/or due to the fact that the steady progression from lower to higher temperatures has only been highlighted with a two - colour scheme . in fig .
15e f the correlations between the orl abundance ratios c@xmath3/h@xmath52and o@xmath3h@xmath52 _ versus _
@xmath47([o iii ] ) are plotted ; the correlation coefficients of the error - weighted fits are in this case 0.88 and 0.70 respectively .
these plots show that the recombination line abundances of doubly ionized carbon and oxygen , which are temperature - insensitive to a large degree , ( i ) rise towards the central nebular regions , and ( ii ) are typically the highest where the forbidden line [ o iii ] temperature is highest .
the o@xmath3/h@xmath52orl ratio used in the correlations discussed and plotted in this section is based on the o ii @xmath84649 line .
we also find a correlation of the c@xmath3/h@xmath52 orl abundance ratio and the he@xmath3/he ionization fraction _ versus _ the extinction constant , @xmath33(h@xmath20 ) ; we performed least - squares fits for the inner nebula ( @xmath4 148 data points ) and show the results in fig .
15g h . fig.15i shows that an inverse trend is established between the oxygen adf and the dereddened h@xmath20 flux from each argus spaxel .
the trend is clearly non - linear .
regions of low surface brightness ( in h@xmath20 ) that lie in the inner nebula exhibit the highest oxygen adfs , whereas the low adf ` tail ' corresponds to brighter zones farther out in the nebula ( most points in this tail correspond to spaxels of the last few ifu rows large y coordinates on the far side from the pn nucleus ) .
the correlations between the orl abundance ratios c@xmath3/h@xmath52 and o@xmath3/h@xmath52 _ versus _ the forbidden - line o@xmath3/h@xmath52 abundance ratio are shown in fig .
16a b , where an inverse trend is seen in both cases with @xmath58 s of @xmath260.86 and @xmath260.69 respectively .
these are mainly driven by the very strong linear correlation between the temperature - dependent forbidden - line o@xmath3/h@xmath52ratios and @xmath47([o iii ] ) that was discussed earlier .
in stark contrast , a positive correlation between the temperature - insensitive orl c@xmath3/h@xmath52 and orl o@xmath3/h@xmath52 abundance ratios is established with @xmath58 @xmath16 0.87 ( fig .
the fact that the trends between the orl cel and orl orl abundance diagnostics lie in completely opposite directions shows that ( i ) _ vastly different assumptions govern the physical origins and the metallicity of the gas traced by orl diagnostics on the one hand and cel diagnostics on the other _ , and that ( ii ) _ consistent nebular metallicity estimates can in principle only be derived using pure orl / orl or cel / cel emission line ratios_. this further tells us that the c@xmath3/o@xmath3 ( pure orl ) abundance ratio is approximately constant throughout the portion of ngc6153 observed in this programme with a mean value of 0.68 @xmath36 0.11 .
16d e shows that the c@xmath3/h@xmath52 abundance ratio and the oxygen adf are positively correlated with the he@xmath3/he ionization fraction with @xmath58 s of 0.85 and 0.86 respectively .
the he@xmath3/he fraction rises close to the pn nucleus where the ionization degree of the nebular plasma is higher due to the harder radiation field , so these plots are again consistent with an origin for the orl - emitting heavy element - rich plasma in the inner regions of the nebula . on the other hand
an inverse linear trend is established between the o@xmath3/h@xmath52cel abundance ratio and he@xmath3/he ( @xmath58 @xmath16 @xmath260.87 ; fig .
16f ) . in fig .
16g h the correlations between the c@xmath3/h@xmath52 orl abundance , the oxygen adf , and the o@xmath3/h@xmath52 orl abundance _ versus _ the he / h abundance ratio are shown .
positive , but weaker , trends are seen in all cases with @xmath58 s of 0.68 , 0.59 , and 0.55 respectively .
these plots also reveal that typically regions in the inner nebula show a somewhat higher helium abundance ( but note that this large he / h is challenged in section 6 ) .
finally , we do not find any obvious trends of the c@xmath3/h@xmath52 or o@xmath3/h@xmath52 orl abundances ( or of the o@xmath3adf ) with electron density .
the detailed parameters of our regression analysis for this nebula are listed in table a1 .
parameters of fits obtained when the o@xmath3/h@xmath52 orl abundance is based on the @xmath84089 line are also listed in table a1 ( the corresponding plots are not shown ) : the fits are all in the same direction as previously , and of similar strength , but with a slightly larger scatter [ as the per spaxel uncertainties on @xmath44(@xmath84089 ) are somewhat higher ] . in this section
we examine whether the various correlations between diagnostics that were established for ngc6153 hold for this nebula too . in order to check for any markedly different behaviour between nebular regions the data from the argus maps
have been separated according to the following scheme : in the following plots , data from spaxels with coordinates
see fig .
9 ( x , y ) @xmath16 ( 19 , 519 ) are represented by red circles , whereas data points outside this rectangular area are marked by blue diamonds .
the red symbols thus correspond to a nebular area of lower mean surface brightness ( in dereddened h@xmath20 ) and of somewhat closer proximity to the pn nucleus than the blue symbols ( cf .
fig . 3 ) .
the blue symbols comprise the zones of higher surface brightness and specifically the bright h@xmath20 rim of the inner pn shell which runs from top to bottom in the h@xmath20 image of the nebula ( cf .
fig . 9 and
the inset of fig . 3 where this sharp boundary is clearly seen in the _ hst _ image ) .
this nebula is however larger than ngc6153 and the argus ifu targeted the central regions of the inner bipolar shell ; hence the adopted grouping of spaxels in the plots that follow may not be as clear - cut regarding their distance from the nucleus as it was for ngc6153 where the ifu was both more favourably placed and captured a larger portion of the pn . in fig .
17a the inverse relationship between the o@xmath3/h@xmath52 abundance ratio based on the [ o iii ] @xmath84959 line and @xmath47([o iii ] ) is plotted for the 296 spaxels with a measurement of these quantities .
the solid line is an error - weighted least - squares fit to the data with a correlation coefficient of @xmath260.94 .
the positive trend between the oxygen adf and @xmath47([o iii ] ) is shown in fig .
17b and has a correlation coefficient of 0.78 . in fig .
17c d the positive correlations between the orl abundance ratios c@xmath3/h@xmath52 and o@xmath3/h@xmath52 _ versus _
@xmath47([o iii ] ) are shown ; the @xmath58 s of the error - weighted fits through the 135 red data points of the inner nebular field are in this case 0.66 and 0.70 respectively ; they are 0.50 and 0.35 , respectively , when fitting all the data .
these plots show that the recombination line abundances of doubly ionized carbon and oxygen for ngc7009 follow the same trend with forbidden line temperature as was the case of ngc6153 .
the o@xmath3/h@xmath52 orl ratio used in the correlations plotted in this section is based on the o ii @xmath84649 line . in fig .
17f g we plot the correlations between the orl abundance ratios o@xmath3/h@xmath52 and c@xmath3h@xmath52 _ versus _ the forbidden - line o@xmath3h@xmath52 abundance ratio , where an inverse trend is seen in both cases with @xmath58 s of @xmath260.68 and @xmath260.63 respectively for fits to the red data points .
when fits to all data are performed the coefficients are less than 0.50 due to the larger scatter of some of the blue data points which correspond to nebular zones of higher mean surface brightness .
the general behaviour of the trends is the same as for ngc6153 . a positive correlation between the temperature - insensitive orl c@xmath3/h@xmath52 and orl o@xmath3/h@xmath52 abundance ratios is established ( @xmath58 @xmath16 0.86 ; fig . 17e ) .
the c@xmath3/o@xmath3 ( pure orl ) abundance ratio is approximately constant throughout the field but there are a few ( red ) points a factor of @xmath41.27 higher ( but within 2@xmath59 ) than the mean ratio of 0.57 @xmath36 0.05 . fig .
17i j shows that the c@xmath3/h@xmath52 orl abundance ratio and the o@xmath3 adf are positively correlated with the he@xmath3/he ionization fraction with @xmath58 s of 0.35 and 0.55 respectively ( for fits through the red data points only ) .
the o@xmath3/h@xmath52cel ratios are inversely correlated with the helium ionization state ( the weighted fit to all data has an @xmath58 of @xmath260.84 ) .
there is more scatter in these plots than was the case for ngc6153 : the cluster of low ionization points within the dashed squares in fig .
17i j ( at adf @xmath4 5 ) come from the lower - right corner of the corresponding he@xmath3/he map and also from ifu row y @xmath16 14 ; they are also of relatively high o@xmath3/h@xmath52(forbidden - line ) abundance .
the main trends identified in these plots are in the same vein as for ngc6153 , with high ionization gas exhibiting higher oxygen adfs and higher c@xmath3/h@xmath52 orl abundance ratios than lower ionization , high o@xmath3/h@xmath52cel regions . fig .
17k m shows the orl o@xmath3/h@xmath52 and c@xmath3h@xmath52 abundance ratios and the o@xmath3 adf _ versus _ the he / h abundance ratio .
weak positive trends are seen in all cases for the o@xmath3 adf vs. he / h linear fit , @xmath58 @xmath16 0.58 .
the trends are in the same direction as in the case of ngc6153 .
we further plot the o@xmath3 adf _ versus _ the dereddened h@xmath20 intensity per spaxel in fig .
no clear trend is seen in contrast to the ngc6153 case .
we do not find any obvious trends of the c@xmath3/h@xmath52 or o@xmath3/h@xmath52 orl abundances ( or of the o@xmath3adf ) with electron density .
no firm trend of the heavy element orl abundances with the extinction constant is found in contrast to the case of ngc6153 .
the detailed parameters of our regression analysis for this nebula are listed in table a1 .
parameters of fits obtained when the o@xmath3/h@xmath52 orl abundance is based on the @xmath84089 line are also listed in table a1 ( the corresponding plots are not shown ) : the fits are in the same direction as previously .
this pn is the smallest of the three and its symmetrical inner shell was almost fully captured by the argus ifu .
it also displays lower mean oxygen adfs and was meant to be our ` control ' target . in this section
we explore correlations between the same set of quantities for ngc5882 as for the other two nebulae .
the argus spaxels were grouped based on their distance from the nucleus in the following manner ( cf .
fig . 7 ) : points with spaxel coordinates ( x , y ) @xmath16 ( 410,1118 ) are coded as red circles whereas those exterior to this boxed region are coded as blue diamonds .
18a shows the o@xmath3/h@xmath52 abundance ratio based on the [ o iii]@xmath84959 line _ versus _
@xmath47([o iii ] ) for the 203 spaxels with a measurement of these quantities ( that is , excluding the masked halo region of the nebula ) .
the solid line is an error - weighted least - squares fit to the data with a correlation coefficient of @xmath260.89 .
the positive trend between the oxygen adf and @xmath47([o iii ] ) ( fig .
18b ; @xmath58 @xmath16 0.52 ) is not as strong as for the other two pne . in fig .
18c d we plot the c@xmath3/h@xmath52 and o@xmath3h@xmath52 orl abundance ratios _ versus _
@xmath47([o iii ] ) ; both show loose positive trends ( with equal @xmath58 @xmath16 0.27 ) .
the o@xmath3/h@xmath52 orl ratio used in the correlations plotted in this section is based on the o ii @xmath84649 line .
the o@xmath3/h@xmath52 and c@xmath3h@xmath52 orl abundances are plotted against the forbidden - line o@xmath3/h@xmath52 abundance ratio in fig .
18f g ; loose negative trends are established in both cases , but in the same vein as for ngc6153 and ngc7009 where the trends were stronger . the c@xmath3/h@xmath52 orl abundance is positively correlated with the o@xmath3/h@xmath52orl abundance ratio ( @xmath58 @xmath16 0.61 ; fig .
18e ) following the same trend as for the other targets .
the c@xmath3/o@xmath3(pure orl ) abundance ratio of ngc5882 is approximately constant across the ifu region with a mean value of 0.52 @xmath36 0.11 .
a loose negative trend is seen between the o@xmath3/h@xmath52 cel abundance and the he@xmath3 ionization fraction ( mostly for low ionization gas represented by the blue data points with @xmath58 @xmath16 @xmath260.68 ; fig .
the c@xmath3/h@xmath52orl abundance shows a flat trend with increasing he@xmath3/he ( fig .
18i ) , in contrast to the cases of ngc6153 and 7009 , while the o@xmath3 adf shows a week positive correlation with he@xmath3/he ( fig .
18j ; @xmath58 @xmath16 0.42 ) .
the excitation state of the gas in this pn is however the lowest of the three , with less than 10 per cent of helium being doubly ionized ( compared with values of up to 20 and 30 per cent for ngc6153 and ngc7009 respectively ) .
no clear trends are seen between the doubly ionized carbon and oxygen orl abundances , and the o@xmath3 adf _ versus _ the total he / h abundance ( fig .
18k m ) when positive trends were established for the other two nebulae .
finally , no trend is established between the o@xmath3 adf and the h@xmath20intensity per spaxel either ( fig .
the trend between the c@xmath3/h@xmath52 orl abundance ratio and the extinction constant is flat and is not shown ( parameters of the fit are however listed in table a1 ) .
detailed parameters of our regression analysis of ngc5882 are listed in table a1 , including those obtained when the o@xmath3/h@xmath52 orl abundance is based on the @xmath84089 line ; those fits are in the same vein as previously .
a strong inverse correlation between the cel - based o@xmath3/h@xmath52 abundances and @xmath47([o iii ] ) is established for all three sources .
the o@xmath3/h@xmath52 cel abundance decreases by factors of approximately 2.0 , 1.5 and 1.3 over @xmath60@xmath47 s of @xmath4 @xmath551600 , @xmath551100 and @xmath55700k in ngc6153 , 7009 , and 5882 respectively ( figs . 15d , 17a , 18a ) .
the exponential dependence of cel - based abundances on the adopted electron temperature ( the boltzmann factor ) mathematically suppresses most of the scatter in o@xmath3/h@xmath52 induced by errors in the @xmath47([o iii ] ) determination , thus strengthening the correlation .
however , the correlation mostly reflects a physical process since the variation of temperature with position is not random ( see ngc6153 , but also ngc7009 where regions closer to the nucleus have a higher mean temperature ) , and @xmath47([o iii ] ) shows a high degree of correlation with other essentially independent quantities such as the c@xmath3/h@xmath52 and o@xmath3/h@xmath52 orl abundances .
a physical correlation is expected as well , since [ o iii ] is a dominant coolant of the gas : where the concentration of o@xmath3 is lower , @xmath47([o iii ] ) will tend to be higher .
this may be due to a real variation of the o / h abundance or to a variation of the o@xmath3/o ionization fraction . assuming a constant o / h , the higher forbidden - line temperature in the inner nebular regions can be qualitatively understood if part of the o@xmath3 ionic component shifts into o@xmath61 ; this is a likely possibility since o@xmath61 must coexist with he@xmath3(based on ionization stratification considerations ) , and the o@xmath3/h@xmath52 cel abundance inversely correlates with he@xmath3/he ( figs . 16f , 17h , 18h ) .
along the same lines , contrasting ngc6153 with ngc5882 , the observed range of @xmath47([o iii ] ) across the respective ifu regions is larger where the he@xmath3zone is more prominent ( i.e. , in the former rather than in the latter nebula ) .
the large opacity of he@xmath52 and the relatively large average energy of the corresponding photoelectrons both contribute to further heat the high - ionization region . that the inner nebula electron temperature can be explained in this way should await quantitative analysis by means of photoionization models tailored to these observations .
extra heating can be provided by the photoelectric effect on small dust grains that may populate the vicinity of the central stars ( see section 6 ) .
the variety of potential heat sources and the associated uncertainties can make potential spatial variations of o / h in the gas phase that emits [ o iii ] @xmath84959 difficult to ascertain .
one intriguing aspect is the fact that @xmath47([o iii ] ) continues to decrease outwards in ngc6153 , even in regions where he@xmath3 and o@xmath61 should in principle be minor species .
it may be that he@xmath3 can survive relatively farther out from the nucleus than expected as the gas ` porosity ' induced by density fluctuations can blur out the theoretical ion stratification .
the second fundamental piece of information from this analysis is the clear positive correlation of the c@xmath3/h@xmath52 and o@xmath3/h@xmath52 orl abundances _ versus _
@xmath47([o iii ] ) for all three targets even for the low adf ` control ' ngc5882 .
throughout the ifu region of ngc6153 , for instance , these abundances increase by a factor of @xmath42 ( fig .
15e f ) from the outer to the inner nebula , so that the corresponding increase in the o@xmath3 adf is a factor of @xmath44 ( fig .
this is paradoxical in that orls should _ a priori _ be enhanced in low @xmath47 conditions ( due to their emissivities obeying an inverse power law with electron temperature ) .
if a sizeable fraction of the ( optical ) [ o iii ] flux were to arise from the same parcel of gas emitting a sizeable fraction of orl flux , then part of the [ o iii ] emission would necessarily be produced at relatively low electron temperatures and the ( observed ) average @xmath47([o iii ] ) would be lower ( _ not higher _ ) , where orl emission is strongest . _
thus , this paradox demonstrates beyond reasonable doubt that heavy element cels and orls must essentially arise from unrelated gas components . _ the ` dual abundance ' picture of two distinct components of highly ionized gas at very different temperatures can solve the paradox , but the only known way to maintain a strongly photoionized gas at low @xmath47 is thermostasis by fine - structure ( fs ) infrared lines .
this mechanism can only be effective for sufficiently large heavy - element abundances ( and sufficiently low electron densities so that the low critical density fs lines are not quenched ) .
this interpretation of the data is qualitatively coherent since cool hydrogen - deficient ionized gas propitiously emits orls .
therefore , at another quantitative level ( separate from orls ) , an essential check can be provided by the 2d spatial distribution of infrared fs line emission ; this should be tractable by means of _ spitzer space telescope _
observations and future _ herschel _ surveys .
we have presented the first detailed analysis of integral field unit spectrophotometry of three galactic pne ( ngc 5882 , 6153 , and 7009 ) taken with the 8.2-m vlt and demonstrated the superb capability of flames argus data for the accurate mapping of the physical properties of nebulae .
spatially resolved maps of 11.5@xmath29 @xmath1 7.2@xmath29 areas of the pne with 0.52@xmath57 arcsec@xmath57 spaxels were made in the light of [ o iii ] @xmath84959 and [ ne iii]@xmath83967 lines , and the h@xmath20 @xmath84861 , he i @xmath84471 , he ii@xmath84686 , c ii @xmath84267 , o ii @xmath84089 and @xmath84649 recombination lines .
the dust extinction and the plasma temperature and density were mapped out , along with the ionic abundances of helium , c@xmath3 , and o@xmath3 , relative to hydrogen .
the o@xmath3/h@xmath52 ratio was derived from the o ii optical recombination lines and the [ o iii ] @xmath84959 collisionally excited line .
maps of the resulting o@xmath3 abundance discrepancy factor were made .
the o@xmath3adf varies between the targets and across the argus field of view in each case . in ngc6153 and ngc7009 the o@xmath3 adf and the c@xmath3
orl abundance show distinct spikes at or near the nucleus . in ngc7009
these quantities further peak at positions coinciding with the high excitation boundary between the inner nebular shell and the outer envelope ; this feature is well registered on both the argus maps and the _ hst _ wfpc2 image . in ngc5882 , where the s / n ratio of o ii orls is lower , the o@xmath3adf derived from the @xmath84089/@xmath84959 diagnostics also peaks at the position of the central star . in all cases these local maxima
are observed over about two spaxels ( @xmath41 arcsec ) and are therefore at the limit of our spatial resolution .
this means that the hypothetical structures associated with the hydrogen - poor plasma should have physical dimensions of @xmath62 1000 astronomical units ( a.u . ) , assuming a typical pn distance of the order of 1kpc .
correlations between the o@xmath3 adf and the ionization state of the gas as gauged by the he@xmath3/he ionization fraction were established for ngc6153 and 7009 ; likewise strong correlations of the o@xmath3 adf and the c@xmath3/h@xmath52 and o@xmath3/h@xmath52 orl abundances _ versus _ the forbidden line electron temperature were established for these nebulae ( for ngc5882 these correlations were weaker but in the same direction ) .
this constitutes new evidence that the ` metallic ' orls in these pne are somehow associated with highly ionized plasma , and that heavy element orls and cels arise from largely unrelated parcels of gas , in accordance with dual - abundance photoionization models ( pquignot et al .
we argue that this hydrogen - deficient , highly ionized plasma may well ( i ) have been recently ejected from the nucleus in the form of clumped gas , or ( ii ) be part of circumstellar material very close to the star in the form of a disk - torus structure ( based on the spatial correlation of the diagnostics with the position of the central star ) , or equally that ( iii ) it originates from high excitation gas currently being photoevaporated from milli - pc scale neutral condensations embedded in the nebulae at the limit of our spatial resolution . in case ( i ) , _ the physical origin of the enhanced heavy - element orl emission could be an ensemble of hydrogen - poor clumps analogous to the h - poor knots observed in the class of nebulae , such as abell 30 or abell 58 _
( e.g. borkowski et al .
1993 ; guerrero & manchado 1996 ) , which may have undergone a late helium - flash ( iben et al .
1983 ) or could be associated with neon novae ( wesson et al .
2007 ) . in this context
, the range of adfs found amongst our targets could reflect an evolutionary effect with h - poor zones becoming progressively more readily observed as the density and temperature contrast between them and the ambient expanding nebula increases over time .
one should recall that ngc5882 is positioned in the lower end of the ` adf sequence ' established from the long - slit surveys according to which dense , compact and lower ionization pne exhibit smaller integrated abundance discrepancies than more diffuse , larger and higher ionization nebulae . ngc6153 and 7009 , on the other hand , appear to be in a more advanced stage of the ` adf evolution ' ( tsamis et al .
2004 ; liu et al .
2004 ) . in case ( ii ) , _ the hydrogen - deficient gas could arise from circumstellar structures _ such as those recently discovered by vlt interferometric observations in association with j - type c - stars ( iras 18006 - 3213 deroo et al
. 2007a ) , post - agb stars ( v390 vel deroo et al .
2007b ) and pne ( mz3 chesneau et al .
these toroidal structures have estimated radii of @xmath62 1000 a.u .
and there is evidence that they contain large amounts of silicate dust .
the inference about their chemistry is important as the c / o orl abundance ratio derived for our pn sample is lower than unity ; this means that an association of the hydrogen - poor , o - rich gas with silicate - based ( i.e. oxygen - rich ) dust is not unlikely ( see below ) . in case ( iii )
_ the h - poor plasma could originate in evaporating dusty ` cometary knots ' analogous to those in the helix nebula _
( ngc7293 ; e.g. meaburn et al .
1992 ) immersed in high ionization nebular zones both near ( ngc5882 , 6153 , 7009 ) and farther out ( ngc7009 ) from the stars .
the helix globules however are rich in molecular hydrogen ( meixner et al . 2005 ) and this hypothesis would require a mechanism for the efficient removal of hydrogen and/or the chemical differentiation of the evaporated gas for the spectroscopic signature of h - poor plasma to be observed .
also , in the helix nebula at least , there is no evidence of globules inside the high ionization he ii zone or of he ii emission from the observed globules , and these features are mostly associated with emission from lower ionization species ( odell et al .
2002 ; odell , henney and ferland 2007 ) ; in contrast , as we have shown , the h - poor plasma in our pne is associated with he ii emission . on the other hand
, this _ does not _ preclude that ` molecular ' globules can be effectively destroyed in the inner regions of pne whilst leaving pockets of h - poor plasma behind .
this hypothesis can clearly be put to the test with observations of the type presented in this work .
we have presented evidence that , in the central regions of ngc6153 , the c@xmath3orl abundance , the he@xmath3/he ionization fraction and the o@xmath3 adf are weakly correlated with the extinction constant .
there is marginal evidence that this may hold for the inner parts of ngc7009 as well .
this correlation , if real , could corroborate either the ` dusty clump evaporation ' scenario or the ` dusty disk ' scenario for the origin of the hydrogen - deficient gas .
dust - rich regions can contribute to the heating of their plasma environs via photoelectric gas heating ( e.g. dopita & sutherland 2000 ; stasiska & szczerba 2001 ; weingartner , draine and barr 2006 ) and to the metal enrichment of the gas via grain destruction , thus bringing about this correlation . according to dopita &
sutherland dust grain photoelectric heating is more effective in regions with a population of small grains , it increases with increasing grain metallicity , and is more important in highly ionized regions .
this effect could also play a role in the observed rise of @xmath47([o iii ] ) in the inner ngc6153 regions ( in parallel to usual photoionization which can also enhance the electron temperature in the he@xmath3 zone ) , although coupled dust / plasma models would be needed to ascertain and quantify this .
alternatively , since the theoretical h@xmath34/h@xmath20 line ratio decreases with decreasing electron temperature ( storey & hummer 1995 ) , and since the extinction constant ( based on h@xmath34/h@xmath20 ) was computed here adopting @xmath47([o iii ] ) , the positive correlation of @xmath33(h@xmath20 ) with the adf may be indicative of the fact that hilines in pne partly arise from cool gas which emits most of the heavy element orls , in accordance with dual - abundance models .
relatively weak correlations between the o@xmath3 adf and the o@xmath3/h@xmath52 , c@xmath3/h@xmath52orl abundances _ versus _ the he / h ratio in ngc6153 and ( in the inner regions ) of ngc7009 have been established .
this result , which should be backed up by further studies , constitutes new observational evidence that corroborates the predictions of the dual - abundance pn models of pquignot et al .
( 2002 , 2003 ) . in those models ,
the heavy element - rich ( h - poor ) gas component , which is responsible for the bulk of the heavy element orl emission , is also helium - rich and as such it contributes non - negligibly to the emissivity of the helium orls .
when this effect is not accounted for , as in empirical methods like the one we have used in this work , there is an insidious bias towards overestimating the helium to hydrogen abundance ratio .
based on our results it is therefore clear that total he / h abundances of pne as traditionally determined from empirical methods , and which assume chemical uniformity of the plasma , will be systematically overestimated .
again detailed modelling in the context of a chemically inhomogeneous nebula is needed to remove this bias . with respect to helium
it should be noted that , in contrast to he@xmath3 , he@xmath52 is not perturbed by h - poor material , since nebular regions with high he@xmath3/h@xmath52 give the same ( he@xmath52 @xmath55 he@xmath3)/h ratio as regions with low he@xmath52/h@xmath52 . in conclusion ,
our detailed ifs analysis of a few suitable objects has allowed us to identify important new elements of the still incomplete puzzle of planetary nebula astrophysics , and we have put forward testable proposals for the likely nature of some of the missing pieces . furthermore , this study has underlined a number of clear requirements that a detailed solution to the ` orl vs. cel ' problem should meet , if it were to be successful .
we thank a. blecha and g. simond of the geneva observatory for answering our queries on girbldrs and for providing extra calibrations and patches .
marina rejkuba is thanked for early advice on the reduction of argus data .
francesca primas is thanked for an informed explanation of the ` scattered light ' issue mentioned in section 3 .
we appreciate comments from bob odell and thank the referee for a constructive report .
this study has made use of the nasa ads database .
ygt gratefully acknowledges the hospitality of eso garching where part of this work was completed during a science visit .
barlow , m. j. , hales , a. s. , storey , p. j. , liu , x .- w . ,
tsamis , y. g. , & aderin , m. e. 2006 , planetary nebulae in our galaxy and beyond , proceedings of the international astronomical union , symposium no . 234 . edited by michael j. barlow and roberto h. mndez .
cambridge : cambridge university press , 2006 , p. 367
davey , a. r. , storey , p. j. , & kisielius , r. 2000 , a&asupplement , 142 , 85 deroo , p. , acke , b. , verhoelst , t. , dominik , c. , tatulli , e. , & van winckel , h. 2007b , a&a , 474 , l45 deroo , p. , van winckel , h. , verhoelst , t. , min , m. , reyniers , m. , & waters , l. b. f. m. 2007a , a&a , 467 , 1093 gerhard , o. , arnaboldi , m. , freeman , k. c. , kashikawa , n. , okamura , s. , & yasuda , n. 2005 , apjl , 621 , l93 gerhard , o. , arnaboldi , m. , freeman , k. c. , okamura , s. , kashikawa , n. , & yasuda , n. 2007 , a&a , 468 , 815 gonalves , d. r. , magrini , l. , leisy , p. , & corradi , r. l. m. 2007 , mnras , 375 , 715 guerrero , m. a. , & manchado , a. 1996 , apj , 472 , 711 liu x .- w . ,
luo s .- g . , barlow m. j. , danziger i. j. , storey p. j. , 2001 , mnras , 327 , 141 liu x .- w . , storey p. j. , barlow m. j. , clegg r. e. s. , 1995 , mnras , 272 , 369 ( lsbc ) liu x .- w . , storey p. j. , barlow m. j. , danziger i. j. , cohen m. , bryce m. , 2000 , mnras , 312 , 585 liu , y. , liu , x .- w .
, barlow , m. j. , & luo , s .-
2004 , mnras , 353 , 1251 liu , x .- w . , barlow , m. j. , zhang , y. , bastin , r. j. , & storey , p. j. 2006 , mnras , 368 , 1959 lodders , k. 2003 , apj , 591 , 1220 meaburn , j. , walsh , j. r. , clegg , r. e. s. , walton , n. a. , taylor , d. , & berry , d. s. 1992 , mnras , 255 , 177 meixner , m. , mccullough , p. , hartman , j. , son , m. , & speck , a. 2005 , aj , 130 , 1784 mendoza , c. 1983 , in planetary nebulae ; proceedings of the symposium , london , england , august 9 - 13 , dordrecht , d. reidel publishing co. , 1983 , p. 143
mesa - delgado , a. , esteban , c. , & garca - rojas , j. 2007 , arxiv e - prints , 710 , arxiv:0710.1285 morisset , c. , stasiska , g. , & pea , m. 2005 , mnras , 360 , 499 odell , c. r. , balick , b. , hajian , a. r. , henney , w. j. , & burkert , a. 2002 , aj , 123 , 3329 odell , c. r. , henney , w. j. , & ferland , g. j. 2007 , aj , 133 , 2343 pquignot , d. , amara , m. , liu , x .- w . ,
barlow , m. j. , storey , p. j. , morisset , c. , torres - peimbert , s. , & peimbert , m. 2002 , proceedings ionized gaseous nebulae , mexico city , 2000 ( eds .
w. j. henney , j. franco , m. martos , & m. pea ) revista mexicana de astronoma y astrofsica ( serie de conferencias ) vol .
142 - 143 pquignot , d. , liu , x .- w . ,
barlow , m. j. , storey , p. j. , & morisset , c. 2003 , planetary nebulae : their evolution and role in the universe , proceedings of the 209th symposium of the international astronomical union , canberra , 2001 .
edited by sun kwok , michael dopita , and ralph sutherland
. published by the astronomical society of the pacific , p. 347
saviane , i. , exter , k. m. , tsamis , y. g. , gallart c. , pquignot , d. 2008 , a&a , submitted stasiska , g. , & szczerba , r. 2001 , a&a , 379 , 1024 sterling , n. c. , dinerstein , h. l. , & kallman , t. r. 2007 , apjs , 169 , 37 storey , p. j. 1994 , a&a , 282 , 999 storey , p. j. , & hummer , d. g. 1995 , mnras , 272 , 41 tsamis y. g. , 2002 , ph.d .
thesis , university of london tsamis , y. g. , barlow , m. j. , liu , x .- w . , danziger , i. j. , & storey , p. j. 2003a , mnras , 338 , 687 tsamis y. g. , barlow m. j. , liu x .- w . , danziger i. j. , storey p. j. , 2003b , mnras , 345 , 186 tsamis , y. g. , barlow , m. j. , liu , x .- w . , storey , p. j. , & danziger , i. j. 2004 , mnras , 353 , 953 tsamis , y. g. , walsh , j. r. , pequignot , d. , barlow , m. j. , liu , x .- w . , & danziger , i. j. 2007 , the messenger , 127 , 53 weingartner , j. c. , draine , b. t. , & barr , d. k. 2006 , apj , 645 , 1188 wesson , r. , liu , x .- w . , &
barlow , m. j. 2005 , mnras , 362 , 424 wesson , r. , barlow , m. j. , liu , x. , storey , p. j. , ercolano , b. , & de marco , o. 2007 , arxiv e - prints , 711 , arxiv:0711.1139 zeippen , c. j. , le bourlot , j. , & butler , k. 1987 , a&a , 188 , 251 zijlstra , a. a. , gesicki , k. , walsh , j. r. , pquignot , d. , van hoof , p. a. m. , & minniti , d. 2006 , mnras , 369 , 875 & + o@xmath3/h@xmath52 cel vs. @xmath47([o iii ] ) & 2.25(@xmath263 ) @xmath36 2.00(@xmath265 ) & @xmath261.99(@xmath267 ) @xmath36 2.20(@xmath269 ) & @xmath260.98 & 0.57 & 296 + adf(o@xmath3@xmath84649 ) vs. @xmath47([o iii ] ) & @xmath2658.2 @xmath36 1.70 & 7.39(@xmath263 ) @xmath36 1.90(@xmath264 ) & 0.92 & 0.82 & 296 + adf(o@xmath3@xmath84089 ) vs. @xmath47([o iii ] ) & @xmath2696.7 @xmath36 3.70 & 0.012 @xmath36 4.11(@xmath264 ) & 0.86 & 2.2 & 295@xmath37 + c@xmath3/h@xmath52 orl vs. @xmath47([o iii ] ) & @xmath269.56(@xmath263 ) @xmath36 3.79(@xmath264 ) & 1.33(@xmath266 ) @xmath36 4.19(@xmath268 ) & 0.88 & 2.2 & 295 + o@xmath3/h@xmath52 @xmath84649 orl vs. @xmath47([o iii ] ) & @xmath269.13(@xmath263 ) @xmath36 7.66(@xmath264 ) & 1.42(@xmath266 ) @xmath36 8.40(@xmath268 ) & 0.70 & 1.2 & 296 + o@xmath3/h@xmath52 @xmath84089 orl vs. @xmath47([o iii ] ) & @xmath262.34(@xmath262 ) @xmath36 1.56(@xmath263 ) & 3.15(@xmath266 ) @xmath36 1.72(@xmath267 ) & 0.73 & 2.8 & 291@xmath37 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 @xmath84649 orl & 6.82(@xmath265 ) @xmath36 8.13(@xmath265 ) & 0.645 @xmath36 0.021 & 0.87 & 2.3 & 294 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 @xmath84089 orl & 6.47(@xmath264 ) @xmath36 6.86(@xmath265 ) & 0.347 @xmath36 0.013 & 0.85 & 2.4 & 290@xmath37 + o@xmath3/h@xmath52 @xmath84649 orl vs. o@xmath3/h@xmath52 cel & 6.72(@xmath263 ) @xmath36 1.80(@xmath264 ) & @xmath266.46 @xmath36 0.39 & @xmath260.69 & 1.3 & 296 + o@xmath3/h@xmath52 @xmath84089 orl vs. o@xmath3/h@xmath52 cel & 1.15(@xmath262 ) @xmath36 3.81(@xmath264 ) & @xmath2614.1 @xmath36 0.82 & @xmath260.71 & 2.8 & 291@xmath37 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 cel & 5.16(@xmath263 ) @xmath36 9.72(@xmath265 ) & @xmath265.87 @xmath36 0.20 & @xmath260.86 & 2.4 & 295 + o@xmath3/h@xmath52 cel vs. he@xmath3/he & 5.45(@xmath264 ) @xmath36 5.77(@xmath266 ) & @xmath261.36(@xmath263 ) @xmath36 4.48(@xmath265 ) & @xmath260.87 & 1.5 & 296 + c@xmath3/h@xmath52 orl vs. he@xmath3/he & 1.72(@xmath263 ) @xmath36 3.35(@xmath265 ) & 9.83(@xmath263 ) @xmath36 3.58(@xmath264 ) & 0.85 & 2.5 & 295 + adf(o@xmath3@xmath84649 ) vs. he@xmath3/he & 4.62 @xmath36 0.15 & 53.1 @xmath36 1.90 & 0.86 & 1.1 & 295 + adf(o@xmath3@xmath84089 ) vs. he@xmath3/he & 5.17 @xmath36 0.24 & 87.4 @xmath36 3.20 & 0.85 & 2.3 & 293@xmath37 + o@xmath3/h@xmath52
@xmath84649 orl vs. he / h & @xmath260.0127 @xmath36 0.0015 & 0.125 @xmath36 0.011 & 0.55 & 1.5 & 296 + o@xmath3/h@xmath52 @xmath84089 orl vs. he / h & @xmath260.0321 @xmath36 0.0031 & 0.282 @xmath36 0.024 & 0.57 & 3.3 & 289@xmath37 + c@xmath3/h@xmath52 orl vs. he / h & @xmath260.0118 @xmath36 0.0010 & 0.109 @xmath36 0.007 & 0.68 & 3.4 & 295 + adf(o@xmath3@xmath84649 ) vs. he / h & @xmath2654.2 @xmath36 4.90 & 476 @xmath36 37.9 & 0.59 & 1.6 & 296 + adf(o@xmath3@xmath84089 ) vs. he / h & @xmath26105 @xmath36 9.90 & 882 @xmath36 75.8 & 0.57 & 3.6 & 289@xmath37 + c@xmath3/h@xmath52 orl vs. @xmath33(h@xmath20 ) & @xmath265.50(@xmath264 ) @xmath36 4.57(@xmath264 ) & 3.03(@xmath263 ) @xmath36 3.94(@xmath264 ) & 0.54 & 1.9 & 148@xmath37 + adf(o@xmath3@xmath84649 ) vs. @xmath33(h@xmath20 ) & @xmath268.86 @xmath36 3.28 & 18.0 @xmath36 2.84 & 0.47 & 1.0 & 147@xmath63 + adf(o@xmath3@xmath84089 ) vs. @xmath33(h@xmath20 ) & @xmath2619.4 @xmath36 4.55 & 31.8 @xmath36 3.90 & 0.55 & 1.9 & 147@xmath63 + @xmath47([o iii ] ) vs. @xmath33(h@xmath20 ) & 7530 @xmath36 248 & 1693 @xmath36 211 & 0.56 & 1.2 & 146@xmath63 + he@xmath3/he vs.
@xmath33(h@xmath20 ) & @xmath260.167 @xmath36 0.026 & 0.253 @xmath36 0.023 & 0.68 & 2.0 & 147@xmath64 + he@xmath52/h@xmath52 vs. he@xmath3/h@xmath52 & 0.128 @xmath36 3.41(@xmath264 ) & @xmath260.673 @xmath36 0.022 & @xmath260.87 & 0.45 & 292@xmath63 + & + o@xmath3/h@xmath52 cel vs. @xmath47([o iii ] ) & 1.63(@xmath263 ) @xmath36 1.52(@xmath265 ) & @xmath261.23(@xmath267 ) @xmath36 1.47(@xmath269 ) & @xmath260.98 & 3.1 & 294 + adf(o@xmath3@xmath84649 ) vs. @xmath47([o iii ] ) & @xmath2626.2 @xmath36 1.40 & 3.05(@xmath263 ) @xmath36 1.41(@xmath264 ) & 0.78 & 2.5 & 296 + adf(o@xmath3@xmath84089 ) vs. @xmath47([o iii ] ) & @xmath2650.8 @xmath36 2.40 & 5.78(@xmath263 ) @xmath36 2.34(@xmath264 ) & 0.82 & 2.0 & 293@xmath37 + c@xmath3/h@xmath52 orl vs. @xmath47([o iii ] ) & @xmath263.23(@xmath263 ) @xmath36 4.23(@xmath264 ) & 4.08(@xmath267 ) @xmath36 4.06(@xmath268 ) & 0.66 & 1.6 & 135@xmath64 + o@xmath3/h@xmath52 @xmath84649 orl vs. @xmath47([o iii ] ) & @xmath269.48(@xmath263 ) @xmath36 9.88(@xmath264 ) & 1.08(@xmath266 ) @xmath36 9.48(@xmath268 ) & 0.70 & 1.6 & 135@xmath64 + o@xmath3/h@xmath52 @xmath84089 orl vs. @xmath47([o iii ] ) & @xmath268.56(@xmath263 ) @xmath36 8.72(@xmath264 ) & 1.13(@xmath266 ) @xmath36 8.49(@xmath268 ) & 0.62 & 2.0 & 293@xmath37 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 @xmath84649 orl & 7.50(@xmath265 ) @xmath36 3.23(@xmath265 ) & 0.519 @xmath36 0.018 & 0.86 & 1.8 & 296 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 @xmath84089 orl & 3.11(@xmath264 ) @xmath36 3.31(@xmath265 ) & 0.230 @xmath36 0.011 & 0.78 & 2.2 & 293@xmath37 + o@xmath3/h@xmath52
@xmath84649 orl vs. o@xmath3/h@xmath52 cel & 4.78(@xmath263 ) @xmath36 2.81(@xmath264 ) & @xmath268.70 @xmath36 0.811 & @xmath260.68 & 1.6 & 135@xmath64 + o@xmath3/h@xmath52 @xmath84089 orl vs. o@xmath3/h@xmath52 cel & 6.10(@xmath263 ) @xmath36 2.53(@xmath264 ) & @xmath268.41 @xmath36 0.694 & @xmath260.58 & 2.0 & 293@xmath37 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 cel & 2.15(@xmath263 ) @xmath36 1.22(@xmath264 ) & @xmath263.27 @xmath36 0.350 & @xmath260.63 & 1.6 & 135@xmath64 + o@xmath3/h@xmath52 cel vs. he@xmath3/he &
4.39(@xmath264 ) @xmath36 2.90(@xmath266 ) & @xmath264.22(@xmath264 ) @xmath36 1.64(@xmath265 ) & @xmath260.84 & 7.9 & 289 + c@xmath3/h@xmath52 orl vs. he@xmath3/he & 7.39(@xmath264 ) @xmath36 6.37(@xmath265 ) & 1.36(@xmath263 ) @xmath36 3.10(@xmath264 ) & 0.35 & 1.9 & 135@xmath64 + adf(o@xmath3@xmath84649 ) vs. he@xmath3/he & 1.19 @xmath36 0.53 & 19.3 @xmath36 2.57 & 0.55 & 2.6 & 135@xmath64 + adf(o@xmath3@xmath84089 ) vs. he@xmath3/he & 5.78 @xmath36 0.28 & 15.9 @xmath36 1.62 & 0.50 & 3.0 & 293@xmath37 + o@xmath3/h@xmath52 @xmath84649 orl vs. he / h & @xmath264.79(@xmath264 ) @xmath36 4.64(@xmath264 ) & 0.0201 @xmath36 0.0040 & 0.28 & 2.5 & 294 + o@xmath3/h@xmath52 @xmath84089 orl vs. he / h & @xmath266.38(@xmath263 ) @xmath36 1.11(@xmath263 ) & 0.0819 @xmath36 0.0096 & 0.45 & 2.3 & 290@xmath37 + c@xmath3/h@xmath52 orl vs. he / h & @xmath263.13(@xmath264 ) @xmath36 2.88(@xmath264 ) & 0.0115 @xmath36 0.0025 & 0.26 & 3.4 & 295 + adf(o@xmath3@xmath84649 ) vs. he / h & @xmath2619.5 @xmath36 2.01 & 214 @xmath36 17.5 & 0.58 & 3.3 & 295 + adf(o@xmath3@xmath84089 ) vs. he / h & @xmath2638.5 @xmath36 3.90 & 409 @xmath36 34.0 & 0.58 & 2.9 & 293@xmath37 + he@xmath52/h@xmath52 vs. he@xmath3/h@xmath52 & 0.110 @xmath36 3.64(@xmath264 ) & @xmath260.789 @xmath36 0.018 & @xmath260.93 & 0.86 & 293@xmath37 + & + o@xmath3/h@xmath52 cel vs. @xmath47([o iii ] ) & 1.71(@xmath263 ) @xmath36 4.52(@xmath265 ) & @xmath261.34(@xmath267 ) @xmath36 4.72(@xmath269 ) & @xmath260.89 & 8.2 & 204 + adf(o@xmath3@xmath84649 ) vs. @xmath47([o iii ] ) & @xmath265.39 @xmath36 0.87 & 7.92(@xmath264 ) @xmath36 9.09(@xmath265 ) & 0.52 & 1.0 & 202 + adf(o@xmath3@xmath84089 ) vs. @xmath47([o iii ] ) & @xmath2621.8 @xmath36 2.62 & 2.70(@xmath263 ) @xmath36 2.74(@xmath264 ) & 0.58 & 1.8 & 197@xmath37 + c@xmath3/h@xmath52 orl vs. @xmath47([o iii ] ) & @xmath263.55(@xmath264 ) @xmath36 2.06(@xmath264 ) & 8.75(@xmath268 ) @xmath36 2.16(@xmath268 ) & 0.27 & 1.8 & 203 + o@xmath3/h@xmath52 @xmath84649 orl vs. @xmath47([o iii ] ) & @xmath266.16(@xmath264 ) @xmath36 3.85(@xmath264 ) & 1.60(@xmath267 ) @xmath36 4.05(@xmath268 ) & 0.27 & 1(@xmath264 ) & 203 + o@xmath3/h@xmath52 @xmath84089 orl vs. @xmath47([o iii ] ) & @xmath262.21(@xmath263 ) @xmath36 1.13(@xmath263 ) & 4.07(@xmath267 ) @xmath36 1.19(@xmath268 ) & 0.24 & 1.8 & 190@xmath37 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 @xmath84649 orl & 1.56(@xmath264 ) @xmath36 2.98(@xmath265 ) & 0.347 @xmath36 0.031 & 0.61 & 1.5 & 203 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 @xmath84089 orl & 5.13(@xmath264 ) @xmath36 1.90(@xmath265 ) & @xmath260.0164 @xmath36 0.0109 & @xmath260.11 & 1.9 & 196@xmath37 + o@xmath3/h@xmath52 @xmath84649 orl vs. o@xmath3/h@xmath52 cel & 1.37(@xmath263 ) @xmath36 1.12(@xmath264 ) & @xmath261.07 @xmath36 0.260 & @xmath260.28 & 1(@xmath264 ) & 203 + o@xmath3/h@xmath52 @xmath84089 orl vs. o@xmath3/h@xmath52 cel & 2.95(@xmath263 ) @xmath36 3.15(@xmath264 ) & @xmath263.01 @xmath36 0.739 & @xmath260.28 & 1.82 & 193@xmath37 + c@xmath3/h@xmath52 orl vs. o@xmath3/h@xmath52 cel & 7.13(@xmath264 ) @xmath36 5.95(@xmath265 ) & @xmath260.547 @xmath36 0.140 & @xmath260.26 & 1.8 & 203 + o@xmath3/h@xmath52 cel vs. he@xmath3/he & 4.47(@xmath264 ) @xmath36 2.97(@xmath266 ) & @xmath268.48(@xmath264 ) @xmath36 7.62(@xmath265 ) & @xmath260.68 & 14 & 148@xmath65 + c@xmath3/h@xmath52 orl vs. he@xmath3/he & 4.88(@xmath264 ) @xmath36 7.70(@xmath266 ) & @xmath261.44(@xmath264 ) @xmath36 1.73(@xmath264 ) & @xmath260.06 & 1.9 & 199 + adf(o@xmath3@xmath84649 ) vs. he@xmath3/he & 1.95 @xmath36 0.04 & 5.62 @xmath36 0.87 & 0.42 & 0.31 & 200 + adf(o@xmath3@xmath84089 ) vs. he@xmath3/he & 3.52 @xmath36 0.13 & 12.6 @xmath36 3.19 & 0.27 & 2.7 & 200@xmath37 + o@xmath3/h@xmath52
@xmath84649 orl vs. he / h & 2.23(@xmath263 ) @xmath36 2.94(@xmath264 ) & @xmath260.012 @xmath36 0.003 & @xmath260.30 & 1(@xmath264 ) & 203 + o@xmath3/h@xmath52 @xmath84089 orl vs. he / h & 4.87(@xmath263 ) @xmath36 1.22(@xmath263 ) & @xmath260.029 @xmath36 0.011 & @xmath260.18 & 2.6 & 203@xmath37 + c@xmath3/h@xmath52 orl vs. he / h & 9.29(@xmath264 ) @xmath36 1.72(@xmath264 ) & @xmath264.12(@xmath263 ) @xmath36 1.59(@xmath263 ) & @xmath260.18 & 1.9 & 203 + c@xmath3/h@xmath52 vs. @xmath33(h@xmath20 ) & 4.76(@xmath264 ) @xmath36 8.1(@xmath266 ) & 1.38(@xmath265 ) @xmath36 1.61(@xmath265 ) & 0.06 & 1.9 & 203@xmath37 + adf(o@xmath3@xmath84649 ) vs. he / h & 7.07 @xmath36 0.84 & @xmath2645.4 @xmath36 7.80 & @xmath260.38 & 0.31 & 203 + adf(o@xmath3@xmath84089 ) vs. he / h & 12.9 @xmath36 2.46 & @xmath2683.1 @xmath36 22.7 & @xmath260.25 & 2.2 & 197@xmath37 + he@xmath52/h@xmath52 vs. he@xmath3/h@xmath52 & 0.108 @xmath36 3.27(@xmath264 ) & @xmath261.13 @xmath36 0.071 & @xmath260.75 & 1.4 & 201@xmath37 + @xmath37 : : corresponding plot not shown .
@xmath64 : : fit only for the inner nebula ( coloured red ) data points ; see text for details .
@xmath65 : : fit only for the outer nebula ( coloured blue ) data points ; see text for details . | results from the first dedicated study of galactic planetary nebulae ( pne ) by means of optical integral field spectroscopy with the vlt flames argus integral field unit ( ifu ) are presented .
three typical galactic - disk pne have been mapped with the 11@xmath0 @xmath1 7@xmath2 argus array : two dimensional spectral maps of the main shell of ngc5882 and of large areas of ngc6153 and 7009 with 297 spatial pixels per target were obtained at sub - arcsec resolutions .
a corresponding number of 297 spectra per target were obtained in the 396.4 507.8 nm range .
spatially resolved maps of emission lines and of nebular physical properties such as electron temperatures , densities and ionic abundances were produced .
the abundances of helium and of doubly ionized carbon and oxygen , relative to hydrogen , were derived from optical recombination lines ( orls ) , while those of o@xmath3 were also derived from the classic collisionally excited lines ( cels ) .
the occurrence of the abundance discrepancy problem , pertaining to oxygen , was investigated by mapping the ratio of orl / cel abundances for o@xmath3 ( the _ abundance discrepancy factor _ ; adf ) across the face of the pne .
the adf varies between targets and also with position within the targets , attaining values of @xmath4 40 in the case of ngc6153 and @xmath4 30 in the case of ngc7009 .
correlations of the adf with geometric distance from the central star and plasma surface brightness ( for ngc6153 ) , as well as with [ o iii ] electron temperature , plasma ionization state and other physical properties of the targets are established .
very small values of the temperature fluctuation parameter in the plane of the sky , @xmath5(o@xmath3 ) , are found in all cases .
it is argued that these results provide further evidence for the existence in run - of - the - mill pne of a distinct nebular component consisting of hydrogen - deficient , super - metal - rich plasma .
the zones containing this posited component appear as undulations in the c ii and o ii orl abundance diagnostics of about 2 spatial pixels across , and so any associated structures should have physical sizes of less than @xmath4 1000 astronomical units . regarding the origin of the inferred zones , we propose that circumstellar disks , abell 30-type knots , or helix - type cometary globules may be involved .
implications for emission line studies of nebulae are discussed .
* key words : * ism : planetary nebulae : general planetary nebulae : individual ngc 5882 , ngc 6153 , ngc 7009 ism : abundances |
in previous publications @xcite , an objective description of a quantum system in the time interval between two complete measurements has been proposed in terms of _ two _ state vectors , together with a new type of physical quantity , the weak value " of a quantum mechanical observable .
specifically , for a system drawn from an ensemble preselected in the state @xmath0 and postselected in the state @xmath1 , the weak value for the observable @xmath2 is defined as @xmath3 where the real part is the quantity of primary physical interest ( and to which the term weak value " shall henceforth apply unless otherwise noted ) .
the suggestion was motivated operationally by the fact that both real and imaginary parts of weak values can be linked to conditional measurement statistics predicted by standard quantum mechanics for the general class of weak measurements " , defined so as to minimize the disturbance to the system as a result of a diminished interaction with the measuring instrument . under these conditions ,
joint weak measurements of two non - commuting observables can be made with negligible mutual interference , thus ensuring that the simultaneous assignment of weak values to all elements of the observable algebra is operationally consistent .
the usefulness of this description has been demonstrated , both theoretically and experimentally , in a number of applications in which novel aspects of quantum processes have been uncovered when analyzed in terms of weak values .
these include photon polarization interference @xcite , barrier tunnelling times @xcite , photon arrival times @xcite , anomalous pulse propagation @xcite , correlations in cavity qed experiments @xcite , complementarity in which - way " experiments @xcite , non - classical aspects of light @xcite , communication protocols @xcite and retrodiction paradoxes " of quantum entanglement @xcite . a certain amount of skepticism @xcite has nevertheless prevailed regarding the physical status of weak values , particularly in the light of the unconventional range of values that is possible according to ( [ weakvdef ] ) .
indeed , the real part of @xmath4 , describing the pointer variable " response in a weak measurement , may lie outside the bounds of the spectrum of @xmath2 .
manifestly eccentric " weak values , as are negative kinetic energies @xcite or negative particle numbers @xcite , are not easily reconciled with the physical interpretation that is traditionally attached to the respective observables .
less intuitive yet is when @xmath2 stands for a projection operator , in which case the weak value suggests weak probabilities " taking generally non - positive values @xcite .
such bizarre interpretations call for a sharper clarification of what physical meaning should be attached to the weak value of an observable .
another item of skepticism surrounding the physical significance of weak values has to do with their general domain of applicability .
it seems reasonable to demand of any new physical concept that it be applicable to a wide variety of situations outside the restricted context in which it is defined operationally .
although progress has been made in this direction @xcite , convincing evidence of the general validity of the concept of the weak value is still lacking . with these questions in mind , the aim of this paper is two - fold : first , we address the physical meaning of weak values by showing that there exists an unambiguous interpretation of the real part of the weak value as a _
definite _ mechanical effect of the system on a measuring probe that is specifically designed to minimize the dispersion in the back - reaction on the system .
second , based on this interpretation , we present a new framework for the analysis of general von neumann measurements , in which the measurement statistics are interpreted as _ quantum averages of weak values _ ( qawv ) .
we believe this framework is physically intuitive and provides compelling evidence for the ubiquity of weak values in more general measurement contexts .
in particular , we show that for arbitrary system ensembles , the expectation value of the reading of any von neumann - type measurement is an average of weak values over a suitable posterior probability distribution .
we furthermore show how qawv framework has a natural correspondence in the classical limit with the posterior analysis of measurement data according to the classical inference framework .
thus , we can establish a correspondence between weak values and what in the macroscopic domain are regarded as objective classical dynamical variables .
the paper is structured as follows : in sec .
[ eigvvswv ] , we motivate the idea of averaging weak values by discussing the connection between pre - selected and pre- and postselected statistics in arbitrary measurements von - neumann type measurements . in sec .
[ mechint ] we present the operational definition of the weak value as a definite mechanical effect associated with infinitesimally uncertain unitary transformations .
the qawv framework is then introduced in sec . [ qaves ] for arbitrary strength measurements .
we provide an illustration in section [ likecc ] , where we discuss a number of measurement situations in which the framework gives a simple characterization of the outcome statistics .
finally , we establish in sec .
[ clas ] the classical correspondence of the qawv framework .
some conclusions are given in sec .
[ concl ] .
the conventional interpretation of a quantum mechanical expectation value , such as @xmath5 for an observable @xmath2 , is as an average of the eigenvalues of @xmath2 over a probability distribution that is realized in the context of a complete strong measurement of @xmath2 .
our main suggestion in this paper is that for a wide class of generalized conditions on the von neumann measurement of @xmath2 , the statistics of measurement outcomes can alternatively be related to an underlying statistics of a different quantity , the weak value of @xmath2 , which is to be regarded as a definite physical property of an unperturbed quantum system in the time interval between two complete measurements .
we shall therefore begin by discussing in this preliminary section the connection between pre- and pre - and post - selected measurement statistics of arbitrary strength von neumann measurements , and from this discussion show an instance in which averages of weak values more aptly describe the posterior break - up of the measurement outcome distribution . in the von neumann measurement scheme @xcite ,
the device is some external system , described by canonical variables @xmath6 and @xmath7 , with @xmath8=i$ ] ( @xmath9 ) .
the system - device interaction is designed so that the measurement result is read - off from the effect on some designated device pointer variable " , which we take to be @xmath7 . for a measurement of the system observable @xmath2 at the time @xmath10
, this interaction is modelled by the impulsive hamiltonian @xmath11 ( note that a possible coupling constant can always be absorbed by canonically redefinig @xmath12 and @xmath13 . )
the effect of the measurement is then described by the unitary operator @xmath14 .
since we will only be concerned with the effect of this interaction from times immediately before to immediately after @xmath15 , we shall henceforth assume the all additional free evolution is already contained in the states .
we first consider the pointer variable statistics from an ensemble defined by pure initial conditions on the system and the apparatus , described by states @xmath0 and @xmath16 , respectively .
for later convenience , we shall term this ensemble the _ preselected measurement ensemble _ ( pme ) @xmath17 .
further , we introduce the notation @xmath18 or @xmath19 to denote times immediately before or immediately after the measurement time @xmath15 .
now , for the pme @xmath17 , the effect of the measurement interaction is easily described by the heisenberg picture transformation @xmath20 induced by the evolution operator @xmath21 . since the initial system plus apparatus state is separable , the final statistics of the pointer variable are easily obtained from the spectral decomposition of @xmath2 , and are given by the probability distribution @xmath22 where @xmath23 , and @xmath24 is the projector onto the eigenspace of the system hilbert space with eigenvalue @xmath25 . in this description
, a strong " or projective measurement corresponds to the limit @xmath26 , ( i.e. , @xmath27 ) , in which case the pointer distribution mimics the spectral distribution of the born interpretation , @xmath28 .
note however that even if the spectrum can not be resolved , the resulting expression ( [ prestats ] ) for the pointer statistics can still be interpreted as if , on every single trial , the pointer variable is displaced in proportion to one of the eigenvalues of @xmath2 , with the eigenvalues distributed randomly throughout the sample according to @xmath29 .
thus , _ regardless _ of the form of @xmath30 the mean and variance of the distribution ( [ prestats ] ) will always satisfy @xmath31 where @xmath32 is the variance in @xmath13 of the state @xmath16 and we have assumed @xmath33 for simplicity .
= 2.65truein now suppose that after time @xmath15 , a postselection is performed on the system , and we wish to concentrate on the subset of measurement outcomes arising only from those systems that ended up in some specific state , @xmath34 .
this final condition defines for us a subensemble @xmath35 of the pme @xmath17 , that we call a _ pre- and post- selected measurement ensmble _ ( ppme ) , the measurement statistics of which can be obtained from the conditional final state of the apparatus@xcite @xmath36 where the normalization @xmath37 is shorthand for the transition probability @xmath38 ( i.e. , the average relative size of the ensemble @xmath35 ) . from this state ,
the corresponding pointer variable distribution is given by @xmath39 .
let us briefly discuss some relations between the pme and ppme statistics .
suppose the postselection involves a complete measurement of some non - degenerate observable @xmath40 , with eigenstates @xmath41 .
a pooling of the data from all the subensembles @xmath42 of @xmath17 must then yield the preselected distribution ( eq .
( [ prestats ] ) ) , in other words @xmath43 where @xmath44 is the relative size of each ppme .
two important consequences follow from this decomposition : first , the pme expectation value of the pointer @xmath45 breaks up in a similar fashion as @xmath46 ; assuming that the prior expectation value of @xmath13 vanishes , this entails the sum rule @xmath47 i.e , the weighted average of the ppme pointer expectation values has to yield the standard expectation value of @xmath2 .
a second consequence of ( [ pooldist ] ) is a covering " condition satisfied by the individual ppme distributions , @xmath48 for all values of @xmath13 and all final outcomes@xmath49 .
this imposes a constraint on how rare a ppme @xmath50 should be were the corresponding @xmath51 to be peaked somewhere in the tail region of @xmath52 .
the relevance of ( [ avesumrule ] ) and ( [ covercond ] ) is that indeed the weight of a ppme measurement outcome distribution need not lie within the normal " region of expectation defined by the bounds of the spectrum of @xmath2 ( fig.[fig : poolings ] ) , contrary to what one would have naively expected given the generality of the spectral expansion of the pme distribution ( [ prestats ] ) .
this may not be obvious under strong measurement conditions , such that when the appartus wave function in @xmath13 is expanded as a superposition of shifted wave functions @xmath53 the overlap between two shifted functions @xmath54 and @xmath55 for all @xmath56 is negligible .
indeed , in such a case , the resulting p.d.f .
for the pointer variable takes the form of a mixture of strong " measurement distributions , i.e. , @xmath57 , each centered at one of the eigenvalues of @xmath2 with weights given by the aharonov , bergmann and liebowitz rule for projective measurement sequences @xcite ; the weight of this distribution is , of course , within the bounds of the spectrum of @xmath2 .
however , away from strong measurement conditions @xmath58 will involve interference terms between the shifted wave functions @xmath59 with coefficients @xmath60 that are not generally real nor positive - definite , preventing the resolution of the individual shifted peaks and allowing for destructive interference effects that may place the weight of @xmath61 beyond the spectrum of @xmath2 . for a wide class of wave functions of the apparatus ,
weak values emerge from the limiting behavior of these interference effects in a complementary limit to that of strong measurement conditions @xcite , namely when @xmath12 , the conjugate to the pointer variable , satisfies @xmath62 ( @xmath63 ) .
in particular , if @xmath64 , one obtains the weak value as the limiting conditional expectation value @xmath65 this limiting behavior is furthermore accompanied by the limit @xmath66 , as if indeed no measurement had taken place , justifying the term weak limit " .
hence , in this limit the posterior break - up ( [ pooldist ] ) of the pme pointer distribution is essentially that of a mixture of distributions , each of which is centered at the weak value defined by its corresponding final state in the post - selection and weighted by the corresponding ( unperturbed ) transition probability .
thus we have an instance in which the expectation value of @xmath2 is more appropriately interpreted operationally as an average of weak values than as an average of eigenvalues .
indeed , it is easily verified that the weighted average of the weak values defined by a complete post - selection is the standard expectation value of @xmath2 @xmath67 as expected from the sum rule ( [ avesumrule ] ) .
note that this is a _ classical _ averaging process , as it arises from the mixing of the distributions conditioned on the distinguishable outcomes of the post - selection .
the sum rule ( [ avewvs ] ) embodies a general rule of thumb , namely that _ eccentric weak values are unlikely " _ , according to which weak values lying outside the spectrum of @xmath2 must be weighted by correspondingly small relative probabilities , ensuring that the average over all pre- and postselected subensembles yields a quantity within the spectral bounds of @xmath2 .
this generic property of weak values is at the heart of the qawv framework presented in section ( [ qaves ] ) , where we show that pre - and post - selected statistics away from weak measurement conditions can also be interpreted from a _
quantum _ averaging process involving weak values .
implicit in the suggestion that standard expectation values can be interpreted ( at least under certain conditions ) as averages of weak values , is the idea that weak values are in some sense sharp " physical properties .
we therefore expand on this notion of sharpness " by giving an operational sense in which the weak value can indeed be regarded as a definite mechanical property of a system that is known to belong to an enesemble defined by complete pre- and post - selections .
the functional dependence on @xmath12 of the transition amplitude @xmath68 in ( [ statetransf ] ) furnishes the necessary elements to build a description of the ppme statistics based on a picture of action and reaction " , in which , if the variable @xmath12 is sharply defined , then a ) the measured system is subject to a sharply - defined unitary transformation generated by @xmath2 , and b ) the measuring apparatus suffers a sharply - defined response given by the weak value of @xmath2 .
this elementary picture serves the basis for the more general qawv framework discussed in the following section .
let us look at the polar decomposition of @xmath69 , which we choose to express as @xmath70 where @xmath71 ( see also refs .
@xcite ) gives the transition probability from @xmath72 to @xmath34 , but mediated by an intermediate unitary transformation @xmath73 .
thus , the variable @xmath12 can be regarded as the parameter of a back - reaction on the system , generated by the operator @xmath2 , inducing the transformation of the initial state @xmath74 ( alternatively , the reaction can be viewed as the inverse transformation @xmath75 on the final state @xmath34 ) . on the other hand , the phase factor in ( [ polardecomp ] ) can be viewed as the generator of a certain reaction of the system on the apparatus corresponding to a specific rotation parameterized by @xmath12 : viewed as a unitary operator on the apparatus degrees of freedom , @xmath76 induces in the heisenberg picture the generally nonlinear canonical transformation of the pointer operator @xmath77 where @xmath78 a straightforward derivation then shows that @xmath79 is indeed a weak value @xmath80 namely the weak value of @xmath2 for the rotated state @xmath81 and the final state @xmath34 .
equation ( [ canshift ] ) therefore shows that for a definite value of @xmath12 , there is an associated definite reaction on the measuring device pointer variable by the weak value for the corresponding pair of states @xmath82 .
more precisely , note that for the general pointer variable statistics for the ppme @xmath35 , eq .
( [ specexp ] ) , we can equivalently express the final apparatus wave function @xmath61 as the fourier integral @xmath83 } \ , .\ ] ] let us now suppose that @xmath12 is constrained to lie exclusively within a finite range around some value @xmath84 , by taking @xmath85 to be the window " function of width @xmath86 centered at @xmath84 .
@xmath87 in this case the wave function in the @xmath13-representation is a modulated sinc " function @xmath88 of characteristic width @xmath89 .
now let @xmath86 be small enough that variations of @xmath90 and @xmath79 are negligible within the interval @xmath91 .
thus , we can approximate @xmath92 , and perform the fourier integral in the group velocity approximation " , i.e. , by expanding the phase about @xmath93 to first order and replacing @xmath90 by @xmath94 ; this yields @xmath95 where we define @xmath96 hence , in the limit @xmath97 , where the apparatus wave function approaches an eigenstate of @xmath6 with eigenvalue @xmath93 , the final wave function for the pointer becomes ( up to a phase ) the initial wave function _ rigidly shifted by a definite weak value _ , the weak value @xmath98 for the rotated state @xmath99 and the final state @xmath34 .
from the point of view of the system , the limit @xmath100 can be regarded as an idealization of a situation often encountered in more general contexts , where the evolution of a quantum system is treated as effectively unitary despite the fact that certain parameters of the evolution are actually physical variables of some external ( and typically macroscopic ) system ; for example a spin rotation , where a macroscopic external magnetic field sets the rotation angle . that such parameters can be treated as classical numbers
is a consequence of a negligible uncertainty of the quantum variable of the external system acting as the parameter for the transformation .
the interaction with such an external system may thus be idealized as an _ infinitesimally uncertain unitary transformation _ at a given parameter value .
this idealization provides the desired mechanical definition of weak values : _ the weak value @xmath79 corresponds to a definite conditional reaction of the system on the variable conjugate to the external physical
parameter variable " @xmath6 of an infinitesimally uncertain unitary transformation generated by @xmath2 at parameter value @xmath12_. the essence of a weak measurement is thus to approach , as close as possible , the ideal conditions of an infinitesimally uncertain transformation .
the above definition presents no ambiguities in the physical interpretation of eccentric " weak values or in the sometimes unexpected relationships that may arise between the weak values of say , @xmath2 and @xmath101 ( e.g. , negative weak variances " , etc . ) . to the extent that we associate weak values to infinitesimal unitary transformations , no _ a - priori _ connection between the weak values of two commuting observables
should be expected ; typically , commuting operators such as @xmath2 and @xmath101 generate entirely different types of un unitary transformations .
rather , relations between weak values follow from the _ linear _ , vector space structure of the lie algebra of hermitian operators generating infinitesimal transformations .
the vector space structure is reflected , for instance , in the fact that for any two initial and final states that are eigenstates of the the observables @xmath2 and @xmath40 , with eigenvalues @xmath25 and @xmath49 respectively , the reaction to an infinitesimal unitary transformation generated by the linear combination @xmath102 at @xmath103 is the linear combination @xmath104 .
finally , let us emphasize the significance of the present mechanical interpretation of weak values in connection with certain quantum mechanical operators , such as kinetic energy or particle number @xcite , for which any association with negative values would appear to be forbidden .
the fact that the reactions associated with weak values will almost always lie within the range of the observable s spectrum is what gives us a reference from which to identify , in those unlikely circumstances where the reaction is eccentric " , what are unique quantum - mechanical effects associated with the role of the observable as a generator of infinitesimal transformations .
one would hardly suspect that such effects could indeed be possible given the physical interpretations that we have traditionally attached to the eigenvalues of a quantum mechanical observable .
the framework of quantum averages of weak values ( qawv ) is the extension of the previous analysis to general von neumann measurements , with arbitrary pure initial states of the apparatus not necessarily satisfying a weakness condition " .
given a _ pure _
ppme @xmath35 , we shall show how the conditional average of measurement outcomes can nevertheless be interpreted as a quantum average of weak values over a suitable distribution . for more general initial and final conditions on the system ( as well as more general initial conditions on the apparatus ) , the corresponding averages can then be obtained by a classical averaging process , similar to that of ( [ avewvs ] ) , given that any such ensemble can always be broken - up into complete pre - and postselected measurement subensembles with appropriate relative weights .
the heuristics of the framework are straightforward : a general apparatus pure state @xmath16 entails indefiniteness in the parameter value @xmath12 driving the back - reaction on the system according to ( [ backreact ] ) , so that a generally finite range of system configurations are sampled in the orbit of transformed initial states @xmath105 .
correspondingly , the pointer measurement statistics should reflect the sampling of a certain range of weak values @xmath106 associated to this orbit .
however , once @xmath12 is allowed to take arbitrary values , a new element in the description comes into play .
this has to do with the relative weights associated with the sampled values of @xmath12 , which reflect a probability - reassessment in the light of the additional conditions entailed by the post - selection .
the central idea of the framework is then that an arbitrary strength von neumann measurement on a pre- and postselected system may be viewed as a superposition of weak measurements at different sampling points @xmath12 , with a re - assessment of the weights of each sample in accordance with bayes theorem .
let us for simpliciity consider an initial apparatus function that is real and smooth .
this function may then be represented as the limit of a superposition of infinitesimally - wide window functions @xmath107 centered at the sampling points " @xmath108 with @xmath109 and @xmath110 . from the results of the previous section , and by linearity , the corresponding final apparatus state wave function may be represented as @xmath111 the final apparatus state can therefore be viewed as a superposition of weak measuerements at the sampling points @xmath112 but with the initial weights @xmath113 replaced by new weights @xmath114 . as is easily seen ,
this re - assessment of weights is in correspondance with @xmath115 , the p.d.f . for strong measurements of @xmath12 ( performed either before or after the measurement interaction ) on the ppme @xmath35 . consistently with bayes theorem , @xmath115 is the posterior distribution for @xmath12 after a re - assessment of the prior p.d.f .
@xmath116 by the likelihood @xmath117 of the post - selection given the @xmath12-dependent rotation of the initial state : @xmath118 note that in accordance with the eccentric weak values are unlikely " rule of thumb , the likelihood factor @xmath119 will tend to suppress the contributions in the superposition for which the weak value falls outside the spectrum of @xmath2 . as we shall illustrate in the coming section ,
it is this mechanism that ensures , together with quantum mechanical interference , that the strong measurement distributions peaked at the eigenvalues of @xmath2 can nevertheless be understood as a quantum superpositions of weak measurements .
it becomes convenient to capture in compact form the two conceptually different processes involved in the updating of the apparatus state @xmath120 as a result of the measurement .
the first step , the generally irreversible process of probability re - assessment , can be expressed conveniently by defining a fiducial state @xmath121 , which we term the _ re - assessed initial state _ of the apparatus . defining the state by its wave function in @xmath12
, it corresponds to the ( prior ) initial wave function @xmath85 multiplied by the square root of the likelihood factor in ( [ postdef ] ) : @xmath122 the other process is the reversible mechanical action of the system on the measurement apparatus generated by the unitary operator @xmath123 defined by the polar decomposition ( [ polardecomp ] ) .
the final conditional state of the measuring device can then be expressed as a unitary transformation applied to the re - assessed state @xmath121 , @xmath124 .
equivalently , one can compute pointer statistics from the heisenberg picture transformation ( [ canshift ] ) using the apparatus state @xmath121 . in particular , the conditional p.d.f . of pointer readings
can be expressed as a quantum - mechanical analogue of a marginal distribution of shifted pointer values , @xmath125 in other words , as a _ quantum average of weak values _
, where the average is taken with respect to the re - assessed initial state @xmath121 . as we shall see in section [ clas ] , eq .
( [ quantpdf ] ) has a natural correspondence in the classical limit ; it can be shown to correspond with the marginal posterior p.d.f .
for the measurement outcomes of the classical function corresponding to @xmath2 on a classical canonical system specified by initial and final boundary conditions in time . using the heisenberg picture ,
we finally obtain the pointer reading mean and variance for the ppme @xmath35 [ finalmoms ] @xmath126 where the subscripts @xmath127 and @xmath128 stand for expectation values in the state @xmath121 and @xmath129 , and where , the quantum weak value average @xmath130 and variance @xmath131 are directly evaluated using the posterior p.d.f .
( [ postdef ] ) .
these expressions can be further simplified if the initial apparatus state has a real @xmath85 @xcite and vanishing expectation value of @xmath13 , in which case the posterior expectation @xmath132 and the correlation @xmath133 vanish . under such conditions , the first two central moments of ( [ quantpdf ] ) are indistinguishable from those obtained from classically averaging weak values with a variability defined through the posterior distribution ( [ postdef ] ) .
note therefore that a condition for a weak measurement that is more general than the one discussed in the previous section is that we have a sharp _ posterior _ p.d.f in @xmath12 around some value @xmath134 , in which case the pointer average reflects a measurement of a sharply - defined weak value @xmath135 with a small uncertainty @xmath131 .
examples of how such effective weak measurements are attained will be given in the next section .
equation ( [ quantpdf ] ) provides a statistical characterization of the pointer variable response as a quantum average of weak values , given the most restrictive conditions possible for a pre- and postselected measurement ensemble . statistics from less restrictive measurement ensembles can then be obtained using standard probability assessments on the ppme @xmath35 consistent with the specified conditions . in particular , for the preselected measurement ensemble @xmath17 and some specific post - selection measurement , the pointer variable distribution ( [ quantpdf ] ) obeys equation eq .
( [ pooldist ] ) . this correspondence yields a generalization of the sum rule ( [ avewvs ] ) to arbitrary measurement strengths , involving both classical and quantum averages @xmath136 and which is easily verified using eqs .
( [ postdef ] ) and ( [ avewvs ] ) .
even more generally , since the statistics for any set of less restrictive conditions on the system and/or the apparatus will involve a classical averaging over the states @xmath16 , @xmath72 , and @xmath34 , the final expectation value of any von - neumann type measurement can always be connected to a suitable average of weak values .
= 2.4truein = 2.0truein let us then illustrate how the qavw framework provides new insight into the measurement statistics of arbitrary - strength von neumann measurements in pre - and post - selected ensembles .
a particularly graphic example of the interrelationship between the orbit of weak values @xmath137 and the corresponding likelihood function @xmath138 is that of spin - component measurements given initial and final spin-@xmath139 coherent states @xcite .
let @xmath140 be a unit vector with direction parameterized by the polar angles @xmath141 and @xmath142 , and @xmath143 the maximal weight @xmath144 eigenstate ( @xmath145 ) ; a spin-@xmath139 coherent state is then defined as @xmath146 and is hence an eigenstate of @xmath147 .
calculations are simplified by the fact that this state can be realized as a product state of @xmath148 copies of the spin-@xmath149 coherent state @xmath150 .
in particular , the transition probability between two coherent states is @xmath151 while the weak value of all spin components are easily captured by a weak spin vector @xmath152 for which the projection onto both @xmath153 and @xmath154 is @xmath139 ( fig .
[ fig : spin ] ) . note the relation between @xmath155 and the length @xmath156 of the weak spin vector , @xmath157 in consistency with the eccentric weak values are unlikely " rule .
= 3.5truein in a measurement of the spin component @xmath158 on a ppme defined by initial and final coherent states @xmath159 and @xmath160 , the back - reaction corresponds to a spin rotation of the initial state about the @xmath161-axis by the angle @xmath162 : @xmath163 this reaction in turn entails an orbit for the weak spin vector @xmath164 ( see fig [ fig : sl ] ) , from which the the weak value function @xmath165 for @xmath158 can be obtained by projecting onto the @xmath161-axis .
note that since the @xmath161 component of @xmath154 is unaffected by the rotation , @xmath166 ; thus , the likelihood factor satisfies @xmath167 figure [ fig : spincurves ] illustrates the correlated behavior of @xmath165 and @xmath168 for the case @xmath169 and with initial and final spin coherent states with @xmath170 and @xmath171 . for these conditions , the weak value is given by @xmath172 oscillating between @xmath173 at @xmath174 ( full rotations of the initial state ) , and @xmath175 at @xmath176 when @xmath177 coincides with @xmath178 ) .
as the figure shows , for @xmath179 , the likelihood @xmath180 shows essentially an exponential behavior similar to a modular gaussian distribution ; in particular , near values @xmath181 or @xmath182 ( both periodic ) , for which the magnitude of @xmath165 is respectively either maximal or minimal on the orbit , we have the approximations for large @xmath139 @xmath183 the exponential suppression of ( [ likeapp ] ) near @xmath181 , where the weak value is maximal in magnitude , is generic of the phenomenon of fourier superoscillations @xcite , exhibited by the amplitude @xmath184 near @xmath185 .
this suppression imposes a robustness " condition on the prior distribution in @xmath12 if one is to measure eccentric weak values near @xmath181 : not only must the prior distribution be sharp " around @xmath186 , but additionally it must show a sufficiently fast fall - off to overcome the exponential rise in likelihood .
= 3.50truein on the basis of the generic correlated behaviors of the likelihood function and the weak value , the ppme pointer statistics for a relatively wide range of von neumann measurement conditions ranging from weak to strong measurements can easily be described in the qawv framework using simple sampling profiles . as discussed in the previous section , a sharp _ posterior _ distribution @xmath187 about some well - defined
sampling point " @xmath188 satisfies the conditions for a weak measurement .
more generally , however , the reassessment by the likelihood factor of the initial apparatus state @xmath16 may yield a state @xmath189 for which the wave function in @xmath12 shows several well - separated narrow peaks , each satisfying weak measurement conditions . in this case
, the ppme pointer distribution will be the result of a coherent superposition of weak measurement pointer wave functions , and will therefore exhibit interference fringes . in simple examples , the existence of just two peaks may be all that is needed to produce the statistical distributions associated with strong measurement conditions ( i.e. , with maxima at the eigenvalues of the measured observable ) .
figures [ fig : likeffex]a to [ fig : likeffex]g show how such single or multiple weak measurement conditions are attained from prior distributions in @xmath12 , of various shapes and locations , for the spin-@xmath139 ppme setting of fig .
[ fig : spincurves ] ( weak value ranging between @xmath190 to @xmath191 ) , with real wave functions for the initial state of the apparatus .
starting with figs .
[ fig : likeffex]a and [ fig : likeffex]b , we illustrate the likelihood effects on an initial robust state of the apparatus given by a narrow window function in @xmath12 of the form given by eq .
( [ weaktrans ] ) , with two different locations @xmath93 . such profiles guarantee that @xmath12 , and hence the average weak value , will always lie within a specific interval ; thus , the effect of the likelihood factor will primarily be a distortion in the shape of the pointer distribution , with minimum effect on the expectation value of @xmath13 . in figs .
[ fig : likeffex]c through [ fig : likeffex]e , we show the likelihood effects on robust gaussian priors of variance @xmath192 at different locations . here
, the prior sampling region may be significantly altered while still preserving a gaussian profile with relatively narrow width . for general initial and final state
, these effects can be described by performing a gaussian approximation of the posterior distribution around its maximum @xmath188 , determined by the equation @xmath193 showing that the imaginary part of the complex weak value can be interpreted as a bias function " for the posterior sampling point .
the resulting pointer distribution will be approximately a gaussian centered at @xmath194 with a corrected width determined by the gaussian approximation .
two interesting effects are then worth noting from these examples : first , as illustrated in fig .
[ fig : likeffex]d , if the bias function is large at the prior sampling point @xmath93 , the posterior sampling point @xmath188 may lie in the tail region of the prior distribution . thus ,
even if the prior distribution is quite narrow , the sampled weak value @xmath195 may differ significantly from the weak value @xmath98 at the prior sampling point .
the second effect has to do with appreciable alterations of the widths as illustrated in figs .
[ fig : likeffex]c and [ fig : likeffex]e : if the prior sampling point @xmath93 is set at a minimum ( maximum ) of the likelihood function ( cases for which @xmath196 in the gaussian approximation ) , the respective posterior distributions in @xmath12 will be widened ( narrowed ) with respect to the prior ; correspondingly , the pointer distributions may be narrowed ( widened ) with respect to the prior distribution @xmath197 . in particular
, it follows that gaussian measurement conditions probing the most eccentric " weak value on the orbit will generically show a _ squeeze _ of the prior pointer distribution a surprising effect if the statistics are viewed as the result of sampling eigenvalues .
= 3.45truein turning finally to figs .
[ fig : likeffex]f and [ fig : likeffex]g , we show the effects on two quite dissimilar non - robust priors centered at @xmath103 : a wide window function of width @xmath198 , and a narrow lorentzian of half - width @xmath199 ( comparable to the prior widths in figs . [
fig : likeffex]a and [ fig : likeffex]c ) , both encompassing the maximum likelihood regions around @xmath200 with either no suppression or insufficiently slow tail suppression of the likelihood factor .
the resulting posterior distributions in @xmath12 are then both qualitatively very similar and similar in turn to the likelihood factor @xmath201 within the region @xmath202 $ ] , which from ( [ likeapp ] ) is spproximately the sum of two equally - shaped narrow gaussians at @xmath203 .
thus , conditions are achieved for the superposition of two weak measurements at @xmath204 , both sampling in this case the least eccentric weak value on the orbit , @xmath205 .
the two peaks in these cases are in fact quite similar to the single peak from the gaussian profile of fig .
[ fig : likeffex]e at @xmath206 ; thus , even while the prior pointer distributions in [ fig : likeffex]e through [ fig : likeffex]f differ substantially in their shapes , the resulting ppme pointer distributions for all three cases share essentially the same envelope , with the last two cases showing interference fringes from the superposition of the two weak measurement sampling points .
this interference pattern can then be connected to the spectral distribution expected from a strong measurement : given weak value and likelihood curves symmetric about @xmath103 , and a posterior distribution in @xmath12 with two similarly - shaped narrow peaks at locations @xmath207 , the resulting ppme pointer distribution will be the ppme pointer distibution for the single peak weak measurement at @xmath188 , but modulated by the term @xmath208 describing the interference pattern .
for the situation depicted in fig .
[ fig : likeffex ] , the phase shift is easily obtained from eq .
( [ wvalj ] ) and is given by @xmath209 for @xmath210 , we have @xmath211 ; hence , interference patterns similar to those of figs [ fig : likeffex]f and [ fig : likeffex]g will show maxima at integer values of @xmath13 ( corresponding to integer values of @xmath139 ) , or at half - integer values of @xmath13 for half - integer @xmath139 , consistently with spectrum of @xmath158 .
the foregoing suggests a fairly general picture underlying the transition from weak to strong measurement conditions for fixed initial and final conditions , as the width of the prior distribution in @xmath12 is varied . illustrating this passage with a gaussian prior of variable width @xmath212 centered at @xmath103 ( fig .
[ fig : trans ] ) for the same @xmath158 measurement , we find the onset of a transitional behavior at a critical value of @xmath212 ( fig .
[ fig : trans]b ) where the gaussian approximation fails . beyond this critical value ,
the exponential rise of the likelihood factor dominates the prior on both sides , thus producing two symmetrically opposed peaks , the locations of which gradually move towards @xmath203 as @xmath212 is increased . this transitional behavior is reflected in the resulting pointer distribution by the emergence of an interference pattern with increasingly closer fringes , modulated by an envelope that gradually shifts with the sampled weak value @xmath195 from the eccentric to the normal region of expectation .
the pattern eventually settles at the characteristic shape of the strong measurement distribution when the location of the two peaks reaches @xmath203 , only becoming sharper with increasing @xmath212 when the tails of the gaussian prior activate " the next likelihood peaks at @xmath213 , etc .
the connection between macroscopic classical " properties and weak values has already been suggested in the literature @xcite . in this section
we give further evidence of this connection by showing the correspondence of the qawv framework in the classical limit .
in particular , we show that in the semi - classical limit , the necessary conditions for a precise measurement of a classical dynamical quantity @xmath214 according to classical mechanics are at the same time the conditions that in the quantum description guarantee a weak measurement of the corresponding observable @xmath2 yielding the same numerical outcome .
let @xmath215 be the configuration variable of a classical system , with free dynamics described by the lagrangian @xmath216 .
for simplicity , we concentrate on a measurement of a function @xmath217 of the configuration variable @xmath215 alone , with a measurement lagrangian of the form @xmath218 coupling the system and an external classical apparatus with pointer variable @xmath13 and canonical conjugate @xmath12 . to connect with the results of section [ qaves ] , we interpret the pre- and post- selection as the fixing of initial and final boundary conditions on the system trajectory : @xmath219 and @xmath220 with @xmath221 .
let us assume for simplicity throughout that only one solution is possible for the euler - lagrange equations . for non - zero @xmath12 ,
the trajectory of the system will differ from its free trajectory due to a modification of the equations of motion by an additional @xmath12-dependent impulsive force @xmath222 arising from the back reaction of the apparatus on the system . then , since the actual trajectory will be some function @xmath223 of the boundary conditions and @xmath12 , the quantity @xmath224 will generally depend on @xmath12 as well . in analogy with our previous notation ,
define the function @xmath225 as one can show from the equations of motion , the classical action for the total lagrangian @xmath226 evaluated on the trajectory @xmath227 , @xmath228 \ , , \ ] ] serves as a generating function for @xmath229 , i.e. , @xmath230 .
thus , from the equations of motion for the apparatus , we find that the pointer variable suffers the impulse @xmath13 at the time @xmath15 @xmath231 in direct correspondence with eq .
( [ canshift ] ) .
we now turn to the probabilistic aspects of the measurement . allowing for uncertainties in the initial state ( i.e. , the point in phase space ) of the apparatus
, we describe our knowledge with a prior p.d.f .
@xmath232 for the state of the apparatus before the measurement , where @xmath233 denotes all available prior information .
we also assume that initial conditions on the system are irrelevant for this prior assessment of probabilities so that @xmath234 .
since the variable @xmath12 enters the equations of motion of the system , knowledge of the final condition @xmath235 becomes relevant for inferences about @xmath12 at the time of the measuring interaction , and will therefore determine a re - assessment of prior probabilities .
we must therefore compute the posterior p.d.f .
@xmath236 for the apparatus , conditioned on the endpoints of the system trajectory , at the time _ before _ the interaction .
the dynamics of the measurement can then be described by the liouville evolution generated by @xmath237 , i.e. , @xmath238 using bayes theorem , we find that @xmath239 where we have used the fact that @xmath12 is the only relevant apparatus variable entering the dynamics of the system , thus yielding a likelihood factor @xmath240 analogous to @xmath241 in the quantum case . finally , evolving to the time after the measurement through eq .
( [ classtatetarns ] ) and marginalizing , we obtain for the pointer variable distribution after the measurement : @xmath242 where the dummy variables @xmath243 and @xmath244 are averaged over the reassessed initial phase space p.d.f . for the apparatus @xmath245 .
this distribution is in complete analogy with eq .
( [ quantpdf ] ) if averages over @xmath246 are identified with averages over the reassessed state @xmath121 and if @xmath229 is identified with the @xmath12-dependent weak value @xmath79 . with this identification , eq .
( [ finalmoms ] ) for the associated moments can be used for both the classical or quantum descriptions .
furthermore , the terms @xmath247 and @xmath248 in ( [ finalmoms ] ) can also be eliminated in the classical case by requiring that the prior phase space distribution factors as @xmath249 with the expectation value of @xmath13 vanishing over @xmath250 . we can now show that under appropriate semi - classical conditions on a corresponding quantum system , the above analogy is not only formal but rather constitutes a true numerical correspondence between classical and quantum averages .
for this , we need to calculate the so - far unspecified likelihood factor @xmath251 in eq .
( [ clasbayes ] ) , which plays the role of @xmath252 in the state reassessment of eq .
( [ statereassess ] ) . in the classical description ,
the probability of being at @xmath235 at the time @xmath253 is proportional to the integral @xmath254 over all possible initial momenta @xmath255 of the system , yielding @xmath256 where @xmath257 is the value of the initial momentum as determined from the boundary conditions .
this initial momentum can be obtained from a variation of the classical action , @xmath258 @xcite , so that @xmath259 ( known as van vleck determinant @xcite from its extension to higher dimensions ) .
correspondence with the quantum description can now be established by calculating the quantum mechanical propagator @xmath260 for the corresponding quantum system , with @xmath261 being the time evolution operator associated with the classical lagrangian @xmath262 . as is easily verified , this is the relevant amplitude for the von neumann measurement of @xmath263 at the time @xmath15 with the given boundary conditions .
under appropriate semiclassical conditions @xcite ( e.g. , small times , large masses , slowly varying potentials , etc . )
, the propagator reduces to the semiclassical or wkb form @xmath264 where @xmath265 is classical action of eq .
( [ clasact ] ) . consequently
, under semiclassical conditions , the weak value @xmath79 of @xmath263 at the time @xmath15 coincides with the classical @xmath229 ; similarly , the likelihood factor in the re - assessment of the initial state of the apparatus ( eq .
[ statereassess ] ) is the square root of the the likelihood factor @xmath266 involved in the re - assessment probabilities in the classical description .
thus , assuming the conditions ensuring @xmath247=0 , the final posterior mean value of @xmath13 will be given both in the classical and quantum descriptions by the average value @xmath267 over the respective posterior distributions in @xmath12 , which can be made to coincide .
this allows us to claim a stronger correspondence between the classical and quantum descriptions when the system satisfies semiclassical conditions : for the same prior distributions in @xmath12 , the classical and quantum expectation values and variances of @xmath214 _ are numerically equal _ and hence , in particular , the final pointer expectation values are equal .
it follows that the minimum dispersion conditions on the variable @xmath12 that in a classical description are required for a precise measurement of @xmath214 ( i.e. , @xmath268 ) , are at the same time the conditions that in the quantum description will guarantee a _
weak _ measurement of @xmath2 yielding the same numerical value .
this correspondence strongly suggests that indeed , what we call macroscopic classical " properties , are in fact weak values .
let us elaborate on this assertion : the use of classical mechanics to describe macroscopic systems or other quantum systems exhibiting classical behavior relies on the fact that individual measurements may be devised so that : a ) the effect on the measurement device accurately reflects the numerical value of the classical observable being measured , b ) no appreciable disturbance is produced on the system as a result of the measurement interaction ; and c ) the effect on the measurement device is statistically distinguishable ( i.e. , the signal to noise ratio is large ) .
the three conditions can be stated as follows : a ) @xmath269 , b ) @xmath270 , @xmath271 and c ) @xmath272 . in the quantum description , conditions a ) and
b ) are weak measurement conditions and can be attained asymptotically by making the posterior uncertainty @xmath273 tend to zero , with the posterior average fixed at @xmath103 ; however , condition c ) can not be upheld in the limit @xmath274 since @xmath275 due to the uncertainty principle .
equivalently , conditions a ) and b ) can not be fulfilled if condition c ) is to be satisfied by demanding @xmath276 as in the case of an ideally strong measurement .
while it is therefore impossible to satisfy the three conditions either in the absolute strong or weak limits , relatively weak measurement conditions can nevertheless be found as a compromise in the uncertainty relations so that conditions a ) , b ) and c ) are simultaneously satisfied for all practical purposes " when classical - like physical quantities are involved .
indeed , for such quantities one expects @xmath277 to be in a sense large " relative to atomic scales , or more precisely , to scale extensively with some scale parameter @xmath278 growing with the size or classicality " of the system ( such the mass , or the number of atoms ) .
one can then choose a scaling relation for @xmath279 , i.e. , @xmath280 so that @xmath281 in which case conditions a ) b ) and c ) can be satisfied in the limit @xmath282 .
assuming that @xmath283 scales as @xmath277 , then with the aid of the uncertainty relation @xmath284 and @xmath285 , we find that this is possible in the quantum description if @xmath279 can be made to scale as @xmath286 in which case @xmath287 as was recently shown @xcite , this is precisely the scaling relation of the optimal compromise for measurements of classical " collective properties ( such as center of mass position or total momentum ) of a large number @xmath288 of independent atomic constituents .
in this paper we have advanced the claim that weak values of quantum mechanical observables constitute legitimate physical concepts providing an objective description of the properties of a quantum system known to belong to a completely pre- and postselected ensemble .
this we have done by addressing two aspects , namely the physical interpretation of weak values , and their applicability as a physical concept outside the weak measurement context . regarding the physical meaning of weak values
, we have shown that the weak value corresponds to a definite mechanical response of an ideal measuring probe the effect of which , from the point of the system , can be described as an infinitesimally uncertain unitary transformation .
we have stressed how from this operational definition the weak value of an observable @xmath2 is tied to the role of @xmath2 as a generator of infinitesimal unitary transformations .
we believe that this sharper operational formulation of weak values in terms of well - defined mechanical effects clarifies the sense in which weak values describe new and surprising features of the quantum domain . regarding the applicability of the concept of weak values in more general contexts
, we have shown that arbitrary - strength von neumann measurements can be analyzed in the framework of quantum averages of weak values , in which dispersion in the apparatus variable driving the back - reaction on the system entails a quantum sampling of weak values .
the framework has been shown to merge naturally into the classical inferential framework in the semi - classical limit .
it is our hope that the framework introduced in the present paper may serve as a motivation for a refreshed analysis of the measurement process in quantum mechanics .
y. a. acknowledges support from the basic research foundation of the israeli academy of sciences and humanities and the national science foundation .
a.b . acknowledges support from colciencias ( contract no .
245 - 2003 ) .
this paper is based in part on the latter s doctoral dissertation @xcite , the completion of which owes much to prof .
yuval neeman and financial support from colciencias - bid ii and a one - year scholarship from icsc - world laboratory . | we re - examine the status of the weak value of a quantum mechanical observable as an objective physical concept , addressing its physical interpretation and general domain of applicability .
we show that the weak value can be regarded as a _
definite _ mechanical effect on a measuring probe specifically designed to minimize the back - reaction on the measured system .
we then present a new framework for general measurement conditions ( where the back - reaction on the system may not be negligible ) in which the measurement outcomes can still be interpreted as _ quantum averages of weak values_. we show that in the classical limit , there is a direct correspondence between quantum averages of weak values and posterior expectation values of classical dynamical properties according to the classical inference framework . |
this work was supported in part by the national science council , taiwan , roc under the grants nsc nsc 97 - 2112-m-032 - 007-my3 ( gcl ) , 98 - 2112-m-001 - 009-my3 ( kwn ) , and the national center for theoretical sciences , taiwan , roc .
g. r. farrar and p. j. e. peebles , astrophys . j. * 604 * , 1 ( 2004 ) [ arxiv : astro - ph/0307316 ] .
s. lee , k. a. olive , and m. pospelov , phys .
d * 70 * , 083503 ( 2004 ) [ arxiv : astro - ph/0406039 ] . s. lee , g .- c .
liu , and k .- w .
ng , phys .
d * 73 * , 083516 ( 2006 ) [ arxiv : astro - ph/0601333 ] .
l. amendola , phys .
d * 69 * , 103524 ( 2004 ) [ arxiv : astro - ph/0311175 ] .
r. bean , e. e. flanagan , i. laszlo , and m. trodden , phys .
d * 78 * , 123514 ( 2008 ) [ arxiv:0808.1105 ] . ; g. l. vacca , j. r. kristiansen , l. p. l. colombo , r. mainini , and s. a. bonometto , jcap 0904:007 ( 2009 ) [ arxiv:0902.2711 ] .
w. lee and k .- w .
ng , phys . rev .
d * 67 * , 107302 ( 2003 ) [ arxiv : astro - ph/0209093 ] . ; r. r. caldwell , m. doran , c. m. mueller , g. schaefer , and c. wetterich , astrophys . j. * 591 * , l75 ( 2003 ) [ arxiv : astro - ph/0302505 ] .
t. koivisto , phys .
d * 72 * , 043516 ( 2005 ) [ arxiv : astro - ph/0504571 ] .
l. amendola and d. tocchini - valentini , phys .
d * 66 * , 043528 ( 2002 ) [ arxiv : astro - ph/0111535 ] .
m. tegmark _
_ , phys . rev .
d * 74 * , 123507 ( 2006 ) [ arxiv : astro - ph/0608632 ] .
l. wang and p. j. steinhardt , astrophys .
j. * 508 * , 483 ( 1998 ) [ arxiv : astro - ph/9804015 ] .
m. doran , j .-
schwindt , and c. wetterich , phys .
d * 64 * , 123520 ( 2001 ) [ arxiv : astro - ph/0107525 ] .
m. bartelmann , f. perrotta , and c. baccigalupi , astron .
astrophys . * 396 * , 21 ( 2002 ) [ arxiv : astro - ph/0206507 ] .
m. kunz , p .- s .
corasaniti , d. parkinson , and e. j. copeland , phys .
d * 70 * , 041301 ( 2004 ) [ arxiv : astro - ph/0307346 ] .
d. j. eisenstein _ et al .
_ , astrophys .
j. * 633 * , 560 ( 2005 ) [ arxiv : astro - ph/0501171 ] .
m. crocce and r. scoccimarro , phys .
d * 77 * , 023533 ( 2008 ) [ arxiv:0704.2783 ] . | we investigate the effects of the nonminimal coupling between the scalar field dark energy ( quintessence ) and the dark matter on the two - point correlation function .
it is well known that this coupling shifts the turnover scale as well as suppresses the amplitude of the matter power spectrum .
however , these effects are too small to be observed when we limit the coupling strength to be consistent with observations . since the coupling of quintessence to baryons is strongly constrained , species dependent coupling may arise .
this results in a baryon bias that is different from unity .
thus , we look over the correlation function in this coupled model .
we are able to observe the enhancement of the baryon acoustic oscillation ( bao ) peak due to the increasing bias factor of baryon from this species dependent coupling . in order to avoid the damping effect of the bao signature in the matter power spectrum due to nonlinear clustering , we consider the coupling effect on the bao bump in the linear regime .
this provides an alternative method to constrain the coupling of dark energy to dark matter .
@xmath0_institute of physics , academia sinica , taipei , taiwan 11529 , r.o.c . _ @xmath1_leung center for cosmology and particle astrophysics , national taiwan university , taipei , taiwan 10617 , r.o.c .
_ @xmath2_department of physics , tamkang university , tamsui , taipei county , taiwan 251 , r.o.c .
_ @xmath3_institute of astronomy and astrophysics , academia sinica , taipei , taiwan 11529 , r.o.c .
_ due to the strong constraint on the coupling of the scalar field dark energy ( quintessence ) to baryons from the local gravity , we investigate the effect of the species dependent coupling @xcite by considering a model in which the quintessence @xmath4 is only coupled to the cold dark matter ( cdm ) .
we assume a yukawa - type coupling , @xmath5 , where @xmath6 is the bare mass of the cdm @xcite .
this specific choice of coupling requires that the present value of the scalar field vanishes in order to satisfy @xmath7 at present .
then we are able to write the general action including this interaction as s = - d^4x \
{ + v(q ) - l_c - l_r - l_b } , [ s ] where @xmath8 is the reduced planck mass , @xmath9 is the potential of @xmath4 , and @xmath10 and @xmath11 ( @xmath12 ) denote the lagrangian of radiation , cdm , and baryons respectively .
we adopt @xmath13 with @xmath14 in the following @xcite . however , the main conclusions are independent of the form of the scalar field potential ( see below ) . due to the coupling ,
the scalings of the cdm and the quintessence energy densities are changed respectively to @xcite _ c ( a ) & = & _ c^0 a^-3 + ( a ) = n_c , [ rhoc ] + _ q & = & -3 h ( 1 + _
q^eff ) _ q _
q^eff = _ q + q , [ drhoq ] where @xmath15 , @xmath16 is the equation of state ( eos ) of the quintessential dark energy ( de ) , @xmath17 denotes the present value of the cdm energy density , the present value of scale factor @xmath18 , and primes mean the differentiation with respect to the conformal time @xmath19 . generally , the sign of @xmath20 depends on both the model and the form of the coupling .
the linear perturbation equations for the cdm and the scalar field @xmath4 in the synchronous gauge are @xcite _ c & = & - _ c - h + n_c q , [ deltac ] + _ c & = & -h _ c + n_c ( k^2 q - q _ c ) , [ thetac ] + q &
+ & 2 h q + k^2 q + q = - h q - _ c n_c _ c , [ deltaq ] where @xmath21 is the wave - number , @xmath22 is the metric perturbation , @xmath23 , and @xmath24 is the gradient of the cdm velocity flow . also from the perturbed einstein equations , we obtain ( h + h h ) = - , [ heq ] where @xmath25 is the sound speed of baryons .
note that we will adopt the adiabatic initial conditions and thus @xmath26 term is absent in eq .
( [ thetac ] ) .
the coupling strength @xmath27 is commonly constrained through the comparison with the observed cosmic microwave background anisotropy and matter power spectrum . in ref .
@xcite , we found @xmath28 . even though the actual value of the upper limit depends on the form of the quintessence potential and that of the coupling , the obtained limits for other potentials and couplings are of the same order as shown in ref .
@xcite .
furthermore , eqs .
( [ deltac ] ) and ( [ thetac ] ) , which describe the evolutions of the cdm density and velocity field respectively , imply that the influence of the quintessence field are dwarfed by the background evolution , @xmath29 .
thus , as long as we have the late - time dominated quintessence model , the evolution behaviors of eqs .
( [ deltac ] ) and ( [ thetac ] ) are quite similar and almost independent of the form of potential .
however , we have also checked the early dark energy model , for example , with the potential given by @xmath30 , in which the dark energy component is not negligible at early times @xcite . in this model
, @xmath29 is quite different from the quintessence model with the exponential potential and produce a quite different behavior of @xmath31 .
here we will concentrate on the late quintessence model . even though we use the full set of the above equations for our calibration
, we investigate the effects of coupling on the matter power spectrum with some approximation in what follows .
it is well known that averaging out the small and oscillatory @xmath32 and @xmath33 is a good approximation for all scales @xcite . from this fact
, we obtain the approximate expression of @xmath34 from eq .
( [ deltaq ] ) q - , [ deltaq2 ] where @xmath35 .
we obtain the evolution equation of @xmath31 from eqs .
( [ deltac ] ) and ( [ thetac ] ) by using eq .
( [ deltaq2 ] ) , _ c +
h _ c - h^2 _ c ( 1 + 2n_c^2 ) _ c 0 , [ dddeltac ] where @xmath36 is the quintessence energy density relative to the critical density and we use @xmath37 @xcite .
the evolution of the linear perturbation of the baryon @xmath38 which is not coupled to the scalar field is the same as the above equation ( [ dddeltac ] ) except that now the coupling terms are absent : _ b + h _ b - h^2 _ c _ c 0 .
[ dddeltab ] it is convenient to rewrite the above equations ( [ dddeltac ] ) and ( [ dddeltab ] ) in terms of @xmath39 as & & + + & & - ( 1 + 2n_c^2 ) _ c _ c 0 , [ dddeltacx ] + & & + ( - _ q _ q ) - _ c _ b 0 , [ dddeltabx ] + & & where _ q 1 - _ c ^-1 . [ omegaq2 ] in eq . ( [ dddeltabx ] ) , we define a baryon bias factor @xmath40 by @xmath41 with a small time variation ( _ i.e. _ we ignore @xmath42 and @xmath43 terms ) .
if we further use the assumption that the linear growth of the cdm is given by @xmath44 with @xmath45 , then we will obtain the analytic form of @xmath40 b = ^-1 .
[ b ] here we use the ansatz for the coupling between cdm and de as @xmath46 . in our model , the scalar field evolves from @xmath47 to @xmath48 during @xmath49 .
we put a limit on the magnitude of the coupling constant @xmath50 , where the choice of the sign of @xmath27 is still arbitrary .
the effective mass of the cdm varies at most around @xmath51 % .
@xmath52 ( _ equally _ , @xmath53 ) increases ( decreases ) as it evolves to the present for the positive ( negative ) @xmath27 .
thus , the evolutions of the background quantities are slightly changed dependent on the sign of @xmath27 .
the effect of the coupling on the @xmath40 also depends on the sign of @xmath27 as given in eq .
( [ b ] ) , which is used to estimate @xmath40 as shown in fig .
[ fig1 ] . in the left panel of fig .
[ fig1 ] , we show the evolution of @xmath40 when @xmath54 . for the positive @xmath27 , the last term in eq . ( [ b ] ) is negative and its magnitude is bigger than @xmath55 for the given model .
thus , @xmath56 .
the evolution of @xmath40 for @xmath57 is depicted in the right panel of fig .
[ fig1 ] . in this case ,
the last term in eq .
( [ b ] ) is positive and @xmath40 is always smaller than @xmath58 .
thus , we are able to constrain not only the magnitude but also the form of the coupling between cdm and de from accurate observations of the baryon power spectrum .
a similar but slightly different conclusion was drawn in ref .
@xcite ; however , their conclusion is only true for the tracking region solutions .
first , we study the cdm density fluctuation @xmath31 given in eq .
( [ dddeltacx ] ) for different cases .
we denote respectively the cosmological model including cdm component with the cosmological constant as @xmath59cdm , with the non - coupled q field ( @xmath60 ) as qcdm , and with the coupled q field as cqcdm .
we also denote @xmath61 for each model as @xmath62 , @xmath63 , and @xmath64 .
@xmath63 has the same evolution equations as those of the @xmath59cdm model except the difference in @xmath65 given in eq .
( [ omegaq2 ] ) .
@xmath16 in the qcdm model changes from @xmath66 in the radiation - dominated epoch ( the so - called `` early tracking region '' ) to around @xmath67 at present .
thus , @xmath68 during the entire epoch .
this causes the suppression of @xmath61 in qcdm model compared to @xmath59cdm model when we use the same present values of the cosmological parameters .
we illustrate this in the left panel of fig .
[ fig2 ] .
the diamond and the rectangular points correspond to @xmath62 and @xmath63 , respectively .
we also compare @xmath63 and @xmath64 .
when the cdm is coupled to q field , the scaling of @xmath69 is changed as given in eq .
( [ rhoc ] ) .
also @xmath16 is increased during the matter - domination epoch in the cqcdm model as shown in the right panel of fig .
[ fig2 ] . the dot - dashed and the dashed lines depict @xmath16 when @xmath70 and @xmath48 , respectively .
thus , this causes slightly further suppression of @xmath61 in the cqcdm model .
however , if we want to compare @xmath64 with @xmath63 at the relevant sub - horizon scale , then we should constrain their evolutions at late times @xmath71 ( _ equally _ , @xmath72 ) . but then @xmath16s of the two models are almost identical as shown in the right panel of fig .
[ fig2 ] and the discrepancy between @xmath64 and @xmath63 becomes negligible .
if we use the definition of the baryon bias factor @xmath73 , then we would obtain that @xmath74 for the same present values of the cosmological parameters .
now we consider the previously mentioned differences in the two - point correlation functions in different models . instead of using the above approximations , we will run the numerical evolution of the full system of equations .
the coupling of the quintessence to the dark matter modifies both the turnover scale and the amplitude of the matter power spectrum .
however , the shift in the turnover scale and the suppression in the amplitude of the matter power spectrum ( defined by @xmath75 ) due to this coupling are too small when we limit the coupling strength to be consistent with observations .
thus , this gives us the motivation to probe the coupling effects on the baryon acoustic peak in the correlation function , @xmath76 , where @xmath77 is the spherical bessel function . in order to avoid the damping effect of the bao signature in the matter power spectrum due to nonlinear clustering
, we put the limitation of @xmath78 in our correlation function calculation .
we show the correlation function times the comoving separation square ( @xmath79 ) in fig .
[ fig3 ] .
the solid , dashed , and dotted lines correspond to the @xmath59cdm , qcdm , and cqcdm models , respectively .
the cosmological parameters that we use in this figure are @xmath80 , @xmath81 , @xmath82 , @xmath83 , and the galaxy bias factor @xmath84 .
we also normalize the matter power spectrum to @xmath85 to be consistent with the wmap and the luminous red galaxies ( lrg ) observations @xcite .
note that the influence of de on the present value of @xmath86 , where @xmath87 is the fourier transform of a top - hat window function and @xmath88 , is well studied @xcite . in fig .
[ fig3 ] , we are able to clearly see the differences of the correlation functions for the different models .
the first peak in the correlation function corresponds to the turnover scale which is related to the scale factor @xmath89 when the radiation and the matter densities are equal .
the @xmath59cdm and qcdm models have the same @xmath90 for the same set of cosmological parameters while their comoving distances of the matter and radiation equality , @xmath91 , are @xmath92 @xmath93 and @xmath94 @xmath93 , respectively .
the discrepancy comes from the fact that the comoving distance to the @xmath90 is @xmath95 and the two models have slightly different @xmath96 .
also , the location of @xmath91 in the cqcdm model is different from that in the qcdm model due to the change in the scaling of the cdm density . in our model , @xmath97 and it causes the delay of the radiation and matter equality epoch ( @xmath98 ) .
thus , the comoving distance @xmath91 is shifted to about @xmath99 @xmath93 in the coupled case .
we have already shown for the relevant @xmath21-modes , @xmath100 for the same present values of the cosmological parameters .
thus , if we normalize the power spectra of the latter two models on cluster scales @xmath101 , they acquire a larger amplitude of primordial fluctuations compared to the @xmath59cdm model .
the bao bump in the correlation function of both qcdm and cqcdm models are larger than that of the @xmath59cdm model .
however , the reason for the enhancements of the bao bumps in both models are different .
the enhancement in the qcdm model compared to the @xmath59cdm one is due to the choice of the larger amplitude of primordial fluctuations .
the enhancement in the cqcdm model is due to the coupling between q and cdm .
we observe this effect in fig .
[ fig3 ] .
if we compare the bao peak of cqcdm with that of @xmath59cdm or qcdm , then we observe that it is enhanced in the cqcdm model .
this is consistent with our early explanation that the baryon bias factor in the cqcdm model is enhanced as given in eq .
( [ b ] ) .
the overall amplitudes of all three models in fig .
[ fig3 ] are smaller than those given in ref .
@xcite .
this is due to the different choices of the galaxy bias factor .
we use the lrg galaxy bias factor @xmath102 @xcite whereas the authors in ref .
@xcite claim that they use the scale dependence bias factor which is however not given therein .
however , both the shapes of the correlation functions and the amplitudes of the bao bumps of the qcdm and cqcdm models fit to the data points better than the @xmath59cdm model . as we mentioned above , if we choose a negative @xmath27 , then the amplitude of the bao bump in the cqcdm model is decreased due to anti - biasing of baryons with @xmath103 .
this effect mimics the nonlinear effect @xcite .
it will suppress the amplitude and shift the location of the bao bump , and thus making harder for us to fit the data .
however , we limit the calculation in the linear regime and this effect is irrelevant for our consideration . in the figure 3 of ref .
@xcite , it is claimed that @xmath104 shows the better fit to the data compared to the case using @xmath105 .
however , the amplitude of the bao bump with @xmath104 shows the bigger discrepancy with the data than that with @xmath105 .
we are able to give a better direction for fitting both the shape and the amplitude of the correlation function in the qcdm and cqcdm models .
we have done a simple @xmath106 analysis and found that the @xmath106 values are @xmath107 , @xmath108 , and @xmath109 for the @xmath59cdm , qcdm , and cqcdm models , respectively .
although the improvement has low statistical significance , the species - dependent coupling effects to the correlation function in the coupled quintessence model may be potentially important and are being further studied .
note that there is a slight shift in the location of the bao peak in the qcdm model .
this is due to the change in @xmath96 in the qcdm model compared to the @xmath59cdm model .
this affects the sound horizon that is given by @xmath110 , where the sound speed of the baryon - photon fluid @xmath111 is the same for every model .
hence , for each model we have a different @xmath96 and @xmath112 will vary
. however , the difference is quite small . |
let @xmath4 be a graph of order @xmath5 ( without loops or multiple edges ) with @xmath6 edges .
alspach @xcite defined the following notions . the _ matching number _ of a linear ordering @xmath7,[2],\ldots,[m]$ ] of the edges of @xmath8 ( we use @xmath9 $ ] to denote the @xmath10th edge in an ordering of the edges of @xmath8 )
is the largest number @xmath0 such that every @xmath0 consecutive edges in the ordering form a matching of @xmath8 .
the _ matching sequencibility _ of @xmath8 , denoted @xmath11 , is the maximum matching number of a linear ordering of the edges of @xmath8 . clearly , the matching sequencibility of @xmath8 is bounded above by the largest number of edges in a matching of @xmath8 and this in turn is bounded above by @xmath12 if @xmath5 is odd .
if @xmath5 is even and @xmath8 itself is a perfect matching , then @xmath13 .
if @xmath5 is even and @xmath8 is not a matching , then @xmath11 can not equal @xmath14 because the edge @xmath7 $ ] and the edge @xmath15 $ ] would have to be identical .
thus , if @xmath5 is even , provided @xmath8 is not a matching , @xmath16 .
the matching sequencibility of @xmath17 is clearly 1 .
using the walecki decomposition of the complete graph @xmath18 with odd @xmath19 into hamilton cycles and the decomposition of @xmath18 with even @xmath20 into hamilton paths , alspach @xcite showed how to order the edges of @xmath18 ( @xmath19 ) to get a matching number equal to @xmath12 .
thus @xmath21 let @xmath22 be a bipartite graph with @xmath6 edges .
, and let @xmath23 be a @xmath24 biadjacency matrix of @xmath8 .
then a ordering of the @xmath6 edges of @xmath8 corresponds to a bijective replacement of the @xmath6 1s of @xmath23 with the integers @xmath25 resulting in a matrix @xmath26 .
the matching sequencibility of this ordering is the largest integer @xmath27 such that every set of @xmath27 consecutive integers lie in different rows and columns . for a complete bipartite graph @xmath28 with @xmath29 ,
it is not difficult to show that @xmath30 the matching sequencibility is certainly bounded above by these numbers . using the biadjacency matrix @xmath23
, it is straightforward to order the edges ( the 1s of @xmath23 ) to show that these upper bounds can be attained . in this representation ,
a matching of size @xmath27 corresponds to a set of @xmath27 1s of @xmath23 no two from the same row and column .
we illustrate this for @xmath31 and @xmath32 whose biadjacency matrices are the @xmath33 and @xmath34 matrices @xmath35 and @xmath36 of all 1s , respectively .
@xmath37\mbox { and } ( k_{4,6 } ) : \widehat{a}=\left[\begin{array}{cccccc } 1&13&9&21&5&17\\ 18&2&14&10&22&6\\ 7&19&3&15&11&23\\ 24&8&20&4&16&12\end{array}\right].\ ] ] henceforth , to avoid trivialities , we assume that @xmath19 .
we define a variation of the matching sequencibility of @xmath8 by considering cyclic orderings of the edges of @xmath8 .
the _ cyclic matching number _ of a cyclic ordering @xmath7,[2],\ldots,[m],[1]$ ] of the edges of @xmath8 is the largest number @xmath0 such that every @xmath0 consecutive edges in the cyclic ordering form a matching of @xmath8 .
the _ cyclic matching sequencibility _ of @xmath8 , denoted @xmath38 , is the maximum cyclic matching number of a cyclic ordering of the edges of @xmath8 .
we have @xmath39 clearly , @xmath40 .
the main result of this note is the following theorem .
[ th : main ] for @xmath20 , @xmath41 combining theorem [ th : main ] and equation ( [ eq : alspach ] ) , we have : [ cor : main ] for @xmath20 , @xmath42
the proof is in three parts : constructions for the even case ( i ) , constructions for the odd case ( ii ) , and a nonexistence argument for the odd case ( iii ) .
* figure 1 * \(i ) we first show that if @xmath43 , then there is a cyclic labeling of the edges of @xmath1 whose cyclic matching number is @xmath3 . in view of ( [ eq : ineq ] ) , this will establish the first equality in theorem [ th : main ] .
let the vertices of @xmath1 be labeled @xmath44 , and consider the vertices @xmath45 as equally spaced points on the unit circle centered at a point labeled @xmath46 . consider the perfect matching @xmath47 of @xmath1 .
we label the edges of @xmath48 as listed in the order @xmath7,[2],\ldots,[m]$ ] .
let @xmath49 be the rotation about the origin by @xmath50 degrees .
then @xmath51 is a 1-factorization of @xmath1 , that is , @xmath52 perfect matchings of @xmath1 partitioning the edges of @xmath1 .
we note that @xmath53 . for @xmath54 ,
the edges @xmath55 receive labels @xmath56,[2+mk],[2+mk]\ldots,[m+mk],\ ] ] respectively . in this way we obtain an ordering @xmath7,[2],\ldots,[2m^2-m]$ ] of the edges of @xmath1 which we regard as a cyclic ordering with [ 1 ] following @xmath57 $ ] . to verify that this cyclic labeling has cyclic matching number @xmath3 , it suffices to show that each set of @xmath3 consecutive edges of @xmath58 is a matching for each @xmath54 .
but this is straightforward to check using the definitions of @xmath48 and @xmath49 .
we illustrate this construction for @xmath59 , that is , @xmath60 , in figure 1 .
\(ii ) we next show that if @xmath43 , then there is a cyclic labeling of the edges of @xmath2 whose cyclic matching number is @xmath3 .
a _ near - perfect matching _ of @xmath2 is a matching of @xmath6 edges .
such a matching meets all but one vertex of @xmath2 , and we call the missing vertex the _ isolated vertex of the matching_. let the vertices of @xmath2 be labeled @xmath61 , and consider the vertices @xmath61 as equally spaced points on the unit circle centered at the origin
. the near - perfect matching @xmath62 has vertex @xmath46 as its isolated vertex .
we label the edges of @xmath48 as listed in the order @xmath7,[2],\ldots,[m]$ ] ( according to the smallest vertex the edge contains ) .
let @xmath49 be the rotation about the origin by @xmath63 degrees
. then @xmath64 are @xmath65 near - perfect matchings of @xmath2 and they partition the edges of @xmath2 .
each vertex of @xmath2 is an isolated vertex of one of these near - perfect matchings .
for @xmath66 , the edges @xmath67 receive labels @xmath56,[2+mk],\ldots,[m+mk],\ ] ] respectively . in this way we obtain an ordering @xmath7,[2],\ldots,[2m^2+m]$ ] of the edges of @xmath2 which we regard as a cyclic ordering with [ 1 ] following @xmath68 $ ] .
( we remark that the walecki decomposition of @xmath2 into hamilton cycles proceeds with @xmath69 vertices equally spaced on the unit circle and one at the origin and with one hamilton cycle @xmath70 .
then with @xmath71 equal to the rotation about the origin through @xmath72 degrees @xmath73 is a hamilton decomposition of @xmath2 . ) to verify that this cyclic labeling has cyclic matching number ( at least ) @xmath3 , we first note that each @xmath74 is a near - perfect matching and thus each of the two sets of @xmath3 consecutive edges of @xmath74 ( that is , its first @xmath3 edges and its last @xmath3 edges ) is a matching . also since @xmath75 including @xmath76 , it is enough to show that any @xmath3 consecutive edges from @xmath77 form a matching .
but again this is straightforward to check using the definitions of @xmath48 and @xmath49 .
we illustrate this construction for @xmath59 , that is , @xmath78 , in figure 2 .
* figure 2 * \(iii ) finally we show that if @xmath43 , then there does not exist a cyclic ordering of the edges of @xmath2 with cyclic matching number @xmath6 . in view of ( ii )
we conclude that @xmath79 .
assume to the contrary that there is a cyclic ordering @xmath7,[2],\ldots,[2m^2+m],[1]$ ] of the edges of @xmath2 with cyclic matching number equal to @xmath6 .
each set of @xmath6 consecutive edges in the cyclic ordering is a near - perfect matching of @xmath2 with a unique isolated vertex .
let @xmath80 } = \{[i],[i+1],[i+2],\ldots,[i+m-1]\ } \quad ( i=1,2,\ldots,2m^2+m)\ ] ] where the edges in the matching @xmath81}$ ] are taken in the cyclic ordering . since @xmath82 } = \{[i+1],[i+2],\ldots,[i+m-1 ] , [ i+m]\},\ ] ] the edge @xmath83 $ ] joins the isolated vertex of @xmath81}$ ] to one of the vertices of edge @xmath9 $ ] .
for each @xmath84 , we get a partition @xmath85}= \{m_{[i ] } , m_{[i+m]},m_{[i+2m]},\dots , m_{[i+(m-1)m]}\}\ ] ] of the edges of @xmath2 into @xmath86 near - perfect matchings ; here the subscripts are taken modulo @xmath87 .
since the degree of each vertex of @xmath2 equals @xmath69 , each vertex of @xmath2 is the isolated vertex of exactly one near - perfect matching of @xmath88}$ ] . for each @xmath84 ,
let @xmath89}$ ] be the spanning subgraph of @xmath2 consisting of the @xmath86 edges @xmath90,[i+m],[i+2m],\ldots,[i+2m^2]\}.\ ] ] note that @xmath91=[i]$ ] in our cyclic ordering , since @xmath2 has exactly @xmath87 edges .
each graph @xmath89}$ ] has exactly @xmath86 vertices and exactly @xmath86 edges .
* claim 1 : * each @xmath89}$ ] is a connected unicyclic graph consisting of a cycle with pendent edges @xmath92possibly none@xmath93 at its vertices .
in fact , since the number of edges of @xmath89}$ ] equals the number of its vertices , if @xmath89}$ ] is connected and the edges span all its vertices , @xmath89}$ ] must be unicyclic . in constructing @xmath89}$ ] in the order of the edges listed ,
each edge of @xmath89}$ ] has a vertex in common with the edge of @xmath89}$ ] preceding it in the order ( interpreted cyclically ) given in ( [ eq : g_i ] ) ; thus , starting with one of the vertices of @xmath89}$ ] , we include one new vertex at a time , giving an ordering of the vertices for each @xmath89}$ ] , and attach an edge from the new vertex to an old vertex .
it follows that @xmath89}$ ] is connected , and since each vertex of @xmath2 is an isolated vertex of one of the near - perfect matchings of @xmath94 , the edges of @xmath89}$ ] span its vertices .
thus @xmath89}$ ] is a connected unicyclic graph , and so @xmath89}$ ] consists of a cycle with a tree ( possibly an empty tree ) rooted at each of its vertices .
moreover , since each edge of @xmath89}$ ] has a vertex in common with the previous edge , these trees can not have paths of length greater than 1 .
this establishes claim 1 .
consider the connected unicyclic graph @xmath95}$ ] .
let @xmath27 be the length of its cycle @xmath96 where we label the vertices so that @xmath97 .
we start the cyclic ordering of the edges of @xmath95}$ ] with the edge @xmath98 , and we can assume that @xmath98 is edge [ 1 ] in the cyclic ordering of the edges of @xmath2 .
let there be @xmath99 pendent edges of @xmath95}$ ] at vertex @xmath10 .
since each edge of @xmath95}$ ] has a vertex in common with the previous edge , it follows that the cyclic ordering of the edges of @xmath95}$ ] is of the form @xmath100 @xmath101 where @xmath102 are the pendent vertices of @xmath89}$ ] joined to vertex @xmath10 ( @xmath103 ) .
assuming , without loss of generality , that vertex 1 ( as opposed to vertex @xmath104 ) is the isolated vertex of the near - perfect matching @xmath105}$ ] , then it follows that the isolated vertices of the near - perfect matchings in the partition @xmath106}$ ] are in the order @xmath107 thus , in constructing @xmath108}$ ] in the manner described above , the vertices are added in this order with an edge from a vertex to some preceding vertex .
* claim 2 : * the edges @xmath109 @xmath110 and the edges @xmath111 @xmath112 are not edges of @xmath108}$ ] .
suppose @xmath113 is an edge of @xmath108}$ ] .
then @xmath114 must be joined to either @xmath115 or @xmath116 in @xmath108}$ ] , and it follows that @xmath117 is joined by an edge to one of @xmath118 .
now vertex @xmath119 is joined by an edge in @xmath108}$ ] to one of @xmath120 , a contradiction since all such edges belong to @xmath95}$ ] .
similar reasoning shows that @xmath111 can not be an edge of @xmath108}$ ] .
this establishes claim 2 .
the vertices of @xmath108}$ ] are added in the order given in ( [ eq : order ] ) with its edge @xmath121 $ ] equal to @xmath122 for some vertex @xmath123 .
the edge @xmath124 $ ] is the edge @xmath125 , since by claim 2 , it can not be the edge @xmath126 .
similarly , by claim 2 , the next edge is @xmath127 since it can not be the edge @xmath128 .
continued application of claim 2 shows that all subsequent vertices must be joined by an edge to @xmath123 in @xmath108}$ ] .
this implies that @xmath108}$ ] is a graph @xmath129 and thus is not unicyclic , a contradiction .
this completes the proof of theorem [ th : main ] .
* figure 3 *
let @xmath43 . by theorem [ th : main ] , @xmath79 . as already remarked , in @xcite , the walecki decomposition of @xmath2 into hamilton cycles
is used to show that @xmath130 .
if one uses that construction twice , one obtains that @xmath131 where @xmath132 is the multigraph obtained from @xmath2 by doubling each edge .
this is illustrated in figure 3 for @xmath59 . in this figure ,
the ordering of the edges of @xmath78 given by the integers @xmath133 is the ordering from the walecki decomposition used by alspach @xcite to show that the matching sequencibility of @xmath78 equals 3 ; note that this ordering does not give that the cyclic matching sequencibility of @xmath78 equals 3 . if @xmath134 is a cycle of @xmath5 vertices with @xmath5 odd , then it is easy to show that @xmath135 , the maximum possible .
this follows by assigning the integers @xmath136 to the edges , starting with any edge and going around the cycle twice taking alternate edges .
this is illustrated in figure 4 for @xmath137 .
now the difference of two positive integers in the same row or column of @xmath26 is @xmath145 modulo @xmath146 .
this implies that every ( cyclical ) set of @xmath147 consecutive positive integers lie in different rows and columns .
since @xmath148 and @xmath149 are at most @xmath147 , it follows that @xmath150 the matrix @xmath26 when @xmath151 is @xmath152.\ ] ] now suppose that @xmath143 is odd , and let the cyclic ordering of the edges of @xmath138 be given by the matrix @xmath26 equal to @xmath153.\ ] ] the difference of two positive integers in the same row and column of @xmath26 is @xmath145 or @xmath143 modulo @xmath146 implying that every ( cyclical ) set of @xmath147 consecutive positive integers lies in different rows and columns .
thus @xmath154 the matrix @xmath26 when @xmath155 is @xmath156.\ ] ] now consider a path @xmath158 of @xmath5 vertices .
first assume that @xmath159 .
then a @xmath160 biadjacency matrix @xmath23 of @xmath161 is obtained from the biadjacency matrix of the cycle @xmath138 by replacing the 1 in its @xmath162-position with a zero .
it is easy to obtain an ordering of the edges of @xmath161 ( the @xmath163s in @xmath23 ) that shows that @xmath164 .
for example , if @xmath165 , the following ordering works and generalizes in the obvious way : @xmath166.\ ] ] now assume that @xmath167 is odd
. then a @xmath168 biadjacency matrix @xmath23 of @xmath169 is obtained from the biadjacency matrix of the cycle @xmath170 by deleting its last row .
it is easy to show that @xmath171 .
for example , if @xmath165 , the following ordering works and generalizes in the obvious way : @xmath172.\ ] ] note that this construction does not give @xmath173 .
in fact , @xmath174 , as we now argue by contradiction .
suppose there were a cyclic ordering of the @xmath146 edges of @xmath169 such that every cyclic set of @xmath143 edges form a matching of @xmath169 .
since we have a cyclic ordering , we may assume that the edge @xmath7 $ ] is one of the pendent edges of @xmath169 .
but then , no matter which matching is determined by the edges @xmath7,[2],\ldots,[q]$ ] , all of the other edges meet a vertex of one of the edges @xmath175,\ldots,[q]$ ] , a contradiction .
the ( cyclic ) matching sequencibility has been computed for complete graphs , complete bipartite graphs , cycles , and paths .
it may be of interest to compute it for other important classes of graphs , but such computations may be very difficult and they will in general depend on the particular graph in the class and thus be of less interest . for a tree , the ( cyclic ) matching sequencibility may differ considerably from the _ matching number _ ( the maximum number of edges in a matching ) .
for example , in figure 5 there is a tree @xmath177 of order @xmath178 with a perfect matching ( matching number equal to 9 ) and an ordering of the edges showing that @xmath179 .
it is not hard to see that @xmath180 and @xmath181 .
if one takes any graph @xmath8 of order @xmath5 and adjoins @xmath182 ( respectively , @xmath183 ) pendent edges at one of its vertices to get a graph @xmath184 of order @xmath185 ( respectively , @xmath186 ) , then @xmath187 ( respectively , @xmath188 ) , since in any ordering ( respectively , cyclic ordering ) of the edges of @xmath184 , some two of these pendent edges would have to be consecutive .
another class of graphs of potential interest are the @xmath27-regular bipartite graphs @xmath189 .
for example , consider the 3-regular bipartite graph @xmath189 whose biadjacency matrix is @xmath190 ( equivalently , @xmath191 ) where @xmath192 is as before the full cycle permutation matrix of order @xmath5 . then @xmath193 .
we illustrate this in terms of the biadjacency matrix @xmath23 and the labeling given by @xmath26 for the @xmath5 odd and @xmath5 even cases with @xmath137 and @xmath194 : @xmath195,\quad \left[\begin{array}{c|c|c|c|c|c|c|c } 1&18&&&&&&8\\ \hline 17&2&10&&&&&\\ \hline & 11&3&20&&&&\\ \hline & & 19&4&12&&&\\ \hline & & & 13&5&22&&\\ \hline & & & & 21&6&14&\\ \hline & & & & & 15&7&24\\ \hline 9&&&&&&23&16\end{array}\right].\ ] ] | we define the cyclic matching sequencibility of a graph to be the largest integer @xmath0 such that there exists a cyclic ordering of its edges so that every @xmath0 consecutive edges in the cyclic ordering form a matching .
we show that the cyclic matching sequencibility of @xmath1 and @xmath2 equal @xmath3 . * key words and phrases :
graph , matching number , cyclic matching sequencibility * * mathematics subject classifications : 05c70 , 05c38 , 05c50 * |
one of the main challenges in theoretical nuclear physics is a microscopic description of the properties of finite nuclei , using realistic nuclear forces . for a given interaction model ,
appropriate observables showing impact on different features of the nuclear potential need to be tested in a many - body system through direct comparison with experimental data . in few - particle systems , where the quantum many - body problem of nucleons can be solved exactly both for bound and scattering states
, one has an optimal setting to probe off - shell properties of different potentials . for photodisintegration of light nuclei ,
exact calculations predict a slightly different behavior when using different two - body realistic interactions , while a more evident effect of three - body forces is found . however , the experimental situation is still unsatisfactory , as the data are not sufficiently precise to discriminate clearly among different interactions models . among light nuclei ,
the photoexcitation of the alpha particle has recently attracted much attention both in theory , where a calculation with realistic two- and three - body forces was carried out @xcite , and in experiments @xcite , where the aim was to further clarify whether @xmath1he exhibits a pronounced giant dipole resonance or not , a question that was raised in the first microscopic calculation of the reaction @xcite , where semirealistic two - body forces where used .
unfortunately , experimental data do not yet lead to a unique picture . nevertheless , a comparison of the impact on observables of different potentials on a theoretical level is already instructive .
these exact calculations of the photodisintegration cross section can also offer valuable theoretical guidance in a class of astrophysical studies .
nuclear absorption of high - energy @xmath4-rays through excitations of giant resonances may potentially become an important diagnostic tool for the dense environments of compact astrophysical sources .
for example , black holes emit an intense high - energy @xmath4-ray continuum , against which nuclear absorption features would uniquely probe the baryonic - matter density , and possibly some compositional information , near the source .
since the dependency on the potential model used in our calculations turns out to be small in comparison to the discrepancy in experiments , our precise microscopic reaction calculations can help to investigate interstellar gas , mostly isotopes of hydrogen and helium , surrounding the astrophysical @xmath4-ray source @xcite .
the purpose of this paper is to investigate the total photodisintegration cross section of light nuclei using the potential constructed within the unitary correlation method ( ucom ) , denoted with @xmath2 .
this is aimed at testing for the first time this interaction in a continuum reaction , where many disintegration channels are open , and simultaneously studying the capability of describing electromagnetic reactions via non - local two - body interactions _ versus _ local two - and three - body models .
exact calculations are performed making use of the lorentz integral transform ( lit ) approach @xcite , which enables to fully take into account the final state interaction , while circumventing the difficulties arising from the continuum many - body scattering states .
the problem is in fact reduced to the solution of a bound - state - like equation , which we solve with an hyperspherical harmonics ( hh ) basis .
the paper is organized as follows . in sec .
[ sec : overview ] the theoretical background is set : an overview of the ucom is given , then the lit method is briefly summarized . in sec .
[ sec : results ] results are shown and conclusions are finally drawn in sec .
[ sec : conclusions ] .
different models for the nucleon - nucleon ( nn ) interaction are found in the literature , as _
e.g. _ the argonne v18 potential ( av18 ) @xcite , the cd bonn @xcite and the nijmegen @xcite potentials .
these modern two - body interactions reproduce the experimental nn data with high precision , but they underbind nuclei with @xmath5 .
this drawback has traditionally been overcome introducing phenomenological three - body forces fitted to reproduce experimental binding energies and first excited states of light nuclei .
furthermore , chiral perturbation theory can provide a systematic way to build two- , three- and more - body forces @xcite .
the above mentioned potential models , including two- and three - body forces have been used to calculate bound state properties of light nuclei ( up to @xmath6 and more ) , mainly within the green s function monte carlo ( gfmc ) method @xcite and the no - core shell model ( ncsm ) approach @xcite .
however , the addition of a three nucleon potential requires a very intensive computational effort when used in many - body systems . in ref . @xcite it was shown that different , but phase equivalent two - body interactions are related by a unitary non - local transformation .
it is expected that the addition of non - locality to the nn interaction could reduce or even cancel the need of three - body force , leading to a big simplification of many - body calculations .
based on this idea , a new category of nn potentials has then emerged , among them one can recall : ( i ) the inside non - local outside yukawa ( inoy ) interaction by doleshall _ et al . _
@xcite , ( ii ) the j - matrix inverse scattering potential ( jisp ) by shirokov _
_ @xcite and ( iii ) the unitary correlation operator method ( ucom ) by feldmeier _
the @xmath2 potential derived from the av18 interaction has already been used as universal input in quite a variety of many - body techniques , ranging form hartree - fock calculations @xcite , to random phase approximation @xcite and to nuclear structure calculations in the framework of fermionic molecular dynamics , see _
e.g. _ @xcite . in the ucom framework
the dominant short range - correlations , induced by the repulsive core and by the tensor part of the nn interaction , are explicitly described by a state independent unitary transformation . when the correlation operator is applied to the hamiltonian , a phase - shift equivalent correlated interaction is obtained .
a correlated operator is defined via a similarity transformation @xmath7 where @xmath8 and @xmath9 are the unitary radial and tensor correlation operators , respectively .
the correlator @xmath10 introduces a radial distance - dependent shift which keeps nucleons away from each other when their uncorrelated distance is smaller than the range of the strongly repulsive core of the nn interaction .
@xmath9 produces tensor correlations by further spatial shifts perpendicular to the radial direction , depending on the orientation of the spins of the two nucleons with respect to their distance . in a many - body system the correlated operator of eq .
( [ correlated_op ] ) becomes an a - body operator with irreducible @xmath11-body contributions , @xmath12 , @xmath13}+\tilde{o}^{[2]}+ \tilde{o}^{[3 ] } + \dots \tilde{o}^{[a]}\ , , \label{cluster}\ ] ] where @xmath14}$ ] indicates the irreducible @xmath11-body part . in the ucom one
usually employs a two - body cluster approximation , neglecting @xmath15}$ ] and higher cluster orders .
the effect of this approximation was tested in the framework of ncsm , where one can perform exact few - body calculations @xcite .
in particular , in this reference it was shown that the omitted three- and more - body terms of the cluster expansion can be tuned changing the range of the tensor correlators in order to compensate the genuine three - body force to a large extent , still preserving the phase shift equivalence . the value of the tensor `` correlation volume '' @xmath16 fm@xmath0 ( for details see @xcite ) was found to give the best description of binding energies of @xmath0h and @xmath1he on the tjon line . in this paper
we will use this very same value of @xmath18 in order to investigate whether the non - local two - body @xmath2 interaction , that minimizes the effect of three - body forces on the binding energies , can also describe the continuum photoabsorption cross section . as already pointed out in @xcite ,
phenomenological non - local interaction terms introduce modifications on the electromagnetic current operator , conserved by gauge invariance .
in fact , if the nuclear potential @xmath19 does not commute with the charge operator , then two - body currents , usually called meson exchange currents ( mec ) , have to be introduced , for consistency . in case of phenomenological non - local potentials the explicit construction of a consistent mec
could be rather involved .
the study of electromagnetic reactions at low energies , where one can use the siegert theorem ( see _ e.g. _ @xcite ) , allows to investigate the role of implicit electromagnetic exchange mechanisms without requiring an explicit knowledge of the two - body current operator . for this reason ,
the test of a given potential model on the prediction for photoabsorption cross section is very important .
the total photoabsorption cross section is given by @xmath20 where @xmath21 is the electromagnetic coupling constant and , at low photon energy @xmath22 , @xmath23 is the inclusive unpolarized dipole response function , generally defined as : @xmath24 here , @xmath25 and @xmath26 indicate the total angular momentum of the nucleus in its initial ground state and its projection , while @xmath27 and @xmath28 denote wave function and energies of the ground and final states , respectively .
the dipole operator is @xmath29 where @xmath30 and @xmath31 are the third components of the position in the center of mass reference frame and isospin of the @xmath32-th particle , respectively .
the dipole approximation has been proven to be very good at low photon energy for the deuteron @xcite and for the triton case @xcite . with the dipole operator the major part of the meson exchange currents ( mec )
is implicitly taken into account , via the siegert theorem . in the lit method @xcite
one obtains @xmath23 after the inversion of an integral transform with a lorentzian kernel @xmath33 where @xmath34 is the unique solution of the inhomogeneous `` schrdinger - like '' equation @xmath35 because of the presence of an imaginary part @xmath36 in eq .
( [ 3 ] ) and the fact that the right - hand side of this equation is a localized state , one has an asymptotic boundary condition similar to a bound state .
thus , one can apply bound - state techniques for its solution .
the response function @xmath23 is recovered via an inversion of the integral transform of eq.([2 ] ) ( for inversion methods see refs . @xcite and @xcite ) . up to now
this method has allowed calculations of electromagnetic reactions for @xmath0h and @xmath0he @xcite and recently even for @xmath1he @xcite with realistic two- and three - body forces .
nuclei with six- @xcite and even seven nucleons @xcite have been investigated with semirealistic interactions .
a similar formalism has been applied to describe exclusive electromagnetic processes in the four - body system with semirealistic potentials @xcite .
recently , the method has been also applied to the harmonic oscillator basis in the framework of ncsm @xcite . in the present paper
we calculate the lit using an hh expansion of the internal wave function . for the antisymmetrization of the wave function we make use of the powerful algorithm developed by barnea _
_ @xcite . the hh approach , which is natural in coordinate space for a local interaction , has been recently extended to the use of non - local potentials given in terms of two - body matrix elements between harmonic oscillator ( ho ) states @xcite , like the jisp potential . in this paper
we will make use of the same method using as input the correlated ho matrix elements of the av18 potential , constructed within the ucom .
[ sec : results ] we begin the discussion with the results for the ground state energies of @xmath1he and @xmath0h calculated in the hh expansion with @xmath2 .
even if the starting point is the local av18 potential , the correlated @xmath2 contains derivative operators . in principle
, one could extend the hh formalism to the use of such terms and consider the rather involved operator form of the @xmath2 , the result being a more intensive numerical computation for each matrix element .
the alternative approach we use , consists in representing the two - body potential operator on an ho basis , similarly to the case of the jisp potential .
the input of our hh calculation are the correlated relative two - body ho matrix elements , @xmath37 where @xmath11,@xmath38 and @xmath39 are radial quantum number , orbital angular momentum , spin and total angular momentum of the two - body sub - system , respectively ( isospin is omitted for the sake of simplicity ) . in the operator form
the ucom is independent on the ho parameters , therefore the ho representation should be viewed as a parameterization rather than a formulation of the potential .
in fact , we expand the two - body potential as @xmath40 the sum over two - body quantum numbers @xmath41 and @xmath42 has to be performed up to maximal value of @xmath43 and @xmath44 for the radial quantum number and angular momentum , respectively .
the expansion has to be pushed forward , till independence of observables , like binding energies , on the ho parameters is reached .
we find out that good convergence of the angular part is reached with @xmath45 , while at least @xmath46 oscillator quanta are needed for the radial part .
since we work in an hyperspherical harmonics many - body hilbert space , convergence as a function of the hyperspherical grand angular momentum @xmath47 needs to be further investigated .
.ground state energies in mev with the @xmath2 potential .
ncsm results from @xcite .
hh calculations performed with @xmath43=15 . [ cols="^,^,^",options="header " , ] in the following an investigation of two photonuclear sum rules , the already mentioned @xmath48 ( eq .
( 10 ) ) and the thomas - reiche - kuhn ( trk ) sum rule @xcite @xmath49|\psi_0 ; m_0\right>\,,\end{aligned}\ ] ] is briefly presented ( see also ref .
@xcite ) . with respect to the av18+uix case ,
the @xmath2 potential predicts a very similar @xmath48 for @xmath0h and a lower @xmath48 for @xmath1he .
this fact is also reflected in the size of the radii of eq.([oddo ] ) . in table
[ table2 ] we summarize the situation : for @xmath0h we present only the proton radius , since the there is no proton - proton radius , while for @xmath1he we show the squared mass radius ( @xmath50 if isospin mixing is neglected ) and @xmath51 .
one can see that @xmath2 leads almost to the same proton radius as av18+uix within a @xmath52 deviation , but smaller radii for the alpha particle are obtained , if compared to av18+uix . a consequence of that is the lower @xmath1he peak cross section found with @xmath2 with respect to av18+uix .
the trk sum rule is of interest since it contains information on the exchange mechanisms induced by the non - commuting part of the potential @xmath53 .
for example , for @xmath1he , we find that @xmath54 mb mev for @xmath2 , which is about @xmath55 lower in comparison to the value of @xmath56 mb mev found for av18+uix @xcite .
this is more strongly connected to the fact that with @xmath2 the cross section goes faster to zero beyond pion threshold and indicates that the meson exchanges underlying the two potential models are different .
he and @xmath1he with three different potential models , ( thin line ) @xmath2 , ( thick line ) av18+uix and ( dashed line ) jisp potential from @xcite , in comparison with the available experimental data : empty squares from @xcite for @xmath0he , empty triangles from @xcite , full triangles from @xcite , full squares from @xcite , shaded area sum of data from @xcite and @xcite and box from @xcite for @xmath1he . ]
we now compare the theoretical predictions with the available experimental data . among the theoretical curves we add the result with the jisp potential @xcite , recently published in @xcite . in fig .
[ figure6 ] , we show the situation for the @xmath0he nucleus and for @xmath1he . in case of the three - body nucleus the error bars of the data from fetisov _ et al .
_ @xcite are unfortunately too big to allow us to discriminate among the different potential models . in case of the alpha particle the experimental situation is more involved .
close to threshold several data were taken in different experiments , which unfortunately show fairly large discrepancies . only the data from arkatov _ et al . _
@xcite cover a larger energy range .
they present a rather high peak cross section , in fair agreement with recent data from @xcite , which favors the av18+uix potential model . in the energy region
@xmath57 mev the results with @xmath2 agree better with this set of data than those with av18+uix .
finally , at higher energies the measurements are precise enough to conclude that av18+uix and @xmath2 can explain , within the error bars , the photoabsorption cross section , in contradistinction to the jisp potential . in ref .
@xcite it was argued that a possible reason of this discrepancy of the jisp model could be due to the probably incorrect long range part of the potential , whose construction is not constrained by the meson exchange theory . regarding this issue
we can only state that @xmath2 , which explicitly contains the long - range pion exchange term of av18 , unchanged by the short range action of the correlations , exhibits a correct tail behavior in the photoabsorption cross section of light nuclei . here
we would also like to point out that the consideration of large hilbert spaces ( large @xmath47 and @xmath43 ) is essential to achieve a stable and ho - independent result on the whole photon energy range .
the fact that the jisp potential is constructed on a small ho basis , for defined @xmath43 and @xmath58 values , could limit the correct description of the tail of this sensitive observable .
we have presented the results of an _ ab initio _ calculation of the @xmath0h , @xmath0he and @xmath1he photodisintegration cross section with the @xmath2 potential .
the difficulty of the scattering continuum problem is circumvented via the lorentz integral transform method , where the final state interaction is fully taken into account .
an expansion of the wave function on hyperspherical harmonics , extended for non - local interaction , is used for the solution of the `` schrdinger - like '' equation .
the first investigation of the unitary correlation operator method on a continuum observable is presented in this paper .
the sensitivity of the photoabsorption cross section on non - local terms of the interaction is investigated . with respect to the traditional local av18 potential , a reduction of the peak and an enhancement of the tail of the cross section
is produced by the non - locality of @xmath2 .
the comparison between the av18+uix potential model and @xmath2 allows to investigate , whether the omission of a genuine three - body force can be replaced by a non - local interaction on a continuum observable .
though binding energies of very light nuclei are well described by @xmath2 with a @xmath59 difference with respect to the av18+uix model , the situation is different in the photodisintegration cross section below pion threshold .
our analysis shows that the @xmath2 potential leads to a very similar result as the av18+uix for high photon energy , while in the region close to threshold larger differences are found , particularly in case of @xmath1he .
the larger deviation in @xmath1he with respect to the @xmath0h case could be related to the fact that smaller mean square radii are obtained with the @xmath2 .
the higher density of the alpha particle makes it certainly more sensitive to characteristic short range properties of the three - body force and non - local parts of the nn interaction , magnifying the differences .
the similarity of the high - energy cross section shows that also in the continuum genuine three - body forces can be partly simulated with the non - local @xmath2 potential .
but the differences found at low - photon energies point out that the two potential models are not completely equivalent .
unfortunately , the lack of precise experimental data limits the possibility to use such calculations as a discriminant test of the potentials .
the author would like to thank nir barnea for providing the hh code extended for the use of non - local interaction , and robert roth for supplying the correlated harmonic oscillator matrix element of @xmath2 .
further gratefulness is owed to hans feldmeier , winfried leidemann and giuseppina orlandini for many useful discussions and for a critical reading of the manuscript .
numerical calculations are partly performed at cineca ( bologna ) . | an exact calculation of the photodisintegration cross section of @xmath0h , @xmath0he and @xmath1he is performed using as interaction the correlated argonne v18 potential , constructed within the unitary correlation operator method ( @xmath2 ) .
calculations are carried out using the lorentz integral transform method in conjunction with an hyperspherical harmonics basis expansion .
a comparison with other realistic potentials and with available experimental data is discussed .
the @xmath2 potential leads to a very similar description of the cross section as the argonne v18 interaction with the inclusion of the urbana ix three - body force for photon energies @xmath3 mev , while larger differences are found close to threshold . |
ammonia is the only significant nitrogen - bearing compound that can be readily observed in brown dwarfs .
its presence in very cool brown dwarf atmospheres has been expected for some time @xcite .
@xcite reported a detection of weak nh@xmath5 features in the near infrared spectrum of the t7p dwarf gl 229b but a clear detection had to await the _ spitzer _
infrared spectrograph ( irs ) observation of the t1/t6 binary @xmath6 indi bab @xcite .
irs spectra also reveal the presence of nh@xmath5 in all observed dwarfs with spectral types of t2 and later @xcite .
in addition to its importance as a strong absorber in very cool brown dwarfs , nh@xmath5 provides an important window into the phenomenon of non - equilibrium chemistry .
because some chemical reactions proceed very slowly at the relatively low temperatures encountered in the atmospheres of t dwarfs , vertical transport can lead to chemical abundances that depart from local thermodynamic equilibrium .
this phenomenon has long been established in jovian planets @xcite . in brown dwarfs ,
the strongest expected signatures of departures from equilibrium chemistry are an overabundance of co and depletion of nh@xmath5 .
the prediction of non - equilibrium chemistry in brown dwarfs @xcite was quickly confirmed by the detection of the 4.7@xmath7 m band of co in gl 229b @xcite , and subsequently by the detection of nh@xmath5 in a high - resolution near - ir spectrum @xcite .
nonequilibrium chemistry may be common in cool brown dwarfs since @xmath8 photometry shows a systematic flux depletion in t6
t9 dwarfs that can be attributed to enhanced co 4.7@xmath7 m absorption @xcite relative to models in chemical equilibrium that with a spectral type of t7.5 @xcite , gl 570d is one of the latest t dwarfs known and its irs spectrum shows the strongest 10 11@xmath7 m nh@xmath5 absorption observed so far @xcite .
it is a particularly interesting target since it is a wide companion to the well - studied k4 v star gl 570a and to a pair of m dwarfs , gl 570bc .
the distance , metallicity , and age of the system very effectively reduce the parameter space allowed to model gl 570d .
@xcite used these constraints to determine the effective temperature @xmath9 and gravity @xmath10 of gl 570d and to analyze its optical and near infrared spectrum . here , we take advantage of the new spectroscopic information provided by the irs spectrum to redetermine @xmath11 .
we also perform the first quantitative analysis of the nh@xmath5 10 11@xmath7 m band for any brown dwarf .
our analysis of gl 570d is based on spectroscopic data covering most of its spectral energy distribution .
we combine a keck lris 0.63
1.01@xmath7 m optical spectrum @xcite with a ukirt cgs4 near ir spectrum covering 0.79 - 1.35@xmath7 m and 1.43 2.52@xmath7 m @xcite , and a irs short low module spectrum which covers 5.43 14.68@xmath7 m @xcite . absolute flux calibration for the ukirt and irs spectra was obtained by using mko photometry @xcite and irac band 4 photometry ( b. patten , priv .
comm . ) , respectively .
the far red optical spectrum was then scaled to match the flux density level of the flux - calibrated ukirt spectrum .
we analyze the spectrum of gl 570d with a combination of model atmospheres , synthetic spectra and evolution sequences .
its t7.5 spectral type strongly suggests that its atmosphere is free of condensates and we use cloudless model atmospheres and synthetic spectra which were most recently described in @xcite and references therein . averaging recent metallicity determinations for gl 570a
@xcite gives [ fe / h]=0.09@xmath12 .
this slight enrichment is negligible in view of the other uncertainties in the data and models and our entire analysis assumes solar metallicity . we have developed an evolution code for brown dwarfs that assumes an adiabatic internal structure .
it uses the h / he equation of state of @xcite . nuclear energy generation through the first branch of the pp - chain @xcite is included .
we use the nacre nuclear reaction rates @xcite and ion - ion and electron - ion screening factors ( chabrier , priv . comm . ) the surface boundary condition is obtained from our atmosphere models .
all brown dwarf atmosphere models become convective at depth , where the specific entropy of the gas becomes essentially constant . the value of the entropy at the bottom of the atmospheres @xmath13 gives the entropy for the adiabatic interior .
the boundary condition is described as a tabulated function @xmath14)$ ] .
thus , our cooling sequences are fully consistent with the atmosphere models @xcite .
details of the evolution calculation will be provided in a future publication .
we follow the method developed by @xcite and initially applied to gl 570d by @xcite .
briefly , we used the optical , near - ir and mid - ir spectra to determine the integrated flux observed at earth ( @xmath15erg / s/@xmath16).5.7% is based on conservatively assuming that the individual calibration uncertainties all add in the same direction .
the 1% uncertainty on the distance @xcite is negligible by comparison . ]
this represents about 70% of the bolometric flux of gl 570d .
we use our grid of model spectra to compute the bolometric correction and the bolometric luminosity , @xmath17 , which depends on the yet - to - be - determined @xmath9 and @xmath10 .
an independent relation @xmath18 is given by our evolution sequences .
a consistent solution between the spectrum and the evolution is obtained by imposing @xmath19 .
this gives a family of solutions @xmath20 ( fig .
a combination of age indicators shows that the gl 570 system is between 2 and 5@xmath1gyr old @xcite , which constrains the range of the gravity to that shown by the isochrones in fig . 1 and the models in table 1 .
models a and c bracket the age limits and model b corresponds to the midpoint .
this new determination of the physical parameters of gl 570d differs from that of @xcite in several ways .
the observed sed has been greatly extended by the _
irs spectrum and the near - ir spectrum has been recalibrated using the latest mko photometry .
the atmosphere models have steadily improved , most notably with the inclusion of a more extensive line list for ch@xmath21 @xcite .
finally , the evolution is computed with an independent code , using our cloudless atmosphere grid as the surface boundary condition , rather than relying on the sequences of @xcite .
despite all these changes , the parameters obtained for gl 570d are remarkably similar to those of @xcite ( crosses on fig .
1 ) . most of the difference can be attributed to a shift of the isochrones in the @xmath11 plane between our evolution sequences and those of @xcite caused mainly by the different surface boundary conditions .
this shows that the method is quite robust and that further improvements in the models will not affect the values in table 1 significantly .
figure 1 shows that the main uncertainty in the @xmath11 of gl 570d arises from the uncertainly in the age of the system . by comparison ,
the uncertainty on @xmath22 from the flux calibration of the data is small .
given @xmath9 , @xmath10 , the radius @xmath23 from the evolution , and the parallax of gl 570a , we can compute model fluxes at earth that can be compared directly to the data . the model b spectrum
is shown in fig .
this is not a fit of the spectrum _
per se _ , but it is obtained solely from the procedure outlined above . the agreement with the optical and near - ir data is generally excellent . the deviations in the 1.6 1.7@xmath7 m region are due to the incomplete ch@xmath21 line list .
the reason for the underestimated @xmath24 band flux is unclear .
models computed with the slightly higher metallicity of gl 570a would give a higher @xmath24 band flux .
a similar effect could also be obtained by flux redistribution from a stronger ch@xmath21 1.6@xmath7 m band computed from a more complete line list .
we find that a cloudy model with a thin cloud deck parameterized by a sedimentation parameter @xmath25 @xcite , can also improve substantially the agreement with the observed @xmath24 band peak . in the mid - ir
, the model systematically underestimates the flux beyond 9@xmath7 m , which coincides with strong absorption by nh@xmath5 .
the outlier models ( a & c in table 1 ) give very similar synthetic spectra .
the differences remain below @xmath26mjy in the near - ir and below @xmath27mjy in the mid - ir , with higher gravity models giving lower fluxes .
vertical transport in the atmosphere of a t dwarf will result in a reduced nh@xmath5 abundance , which will increase the model flux for @xmath28 m and improve the agreement with the irs spectrum .
we consider atmosphere models that depart from chemical equilibrium to optimize the fit to the data .
the kinetics of the chemistry of carbon and nitrogen in giant planets and brown dwarfs has been described in details in @xcite and the effect on the spectra of brown dwarfs in @xcite .
we only provide a brief summary of nonequilibrium chemistry here . the chemistry of carbon and nitrogen in brown dwarfs atmospheres is essentially described by the net reactions @xmath29 @xmath30 because of the large binding energies of co and n@xmath31 , both reactions proceed much more slowly to the right than to the left .
vertical mixing dredges up hot gas that is relatively rich in co and n@xmath31 to the cooler part of the atmosphere , where co and n@xmath31 are converted very slowly into ch@xmath21 and nh@xmath5 , respectively .
the net result is a relative overabundance of co and n@xmath31 in the upper atmosphere and decreased abundances of ch@xmath21 and nh@xmath5 in the upper atmosphere .
the abundance of h@xmath31o is also reduced by conservation of the total number of oxygen atoms .
the correct chemical pathway and reaction time scale for the conversion of co into ch@xmath21 remains uncertain @xcite .
we adopt the `` fast '' chemical time scale of @xcite . for the n@xmath31 conversion
, we adopt the time scale given in @xcite .
slow mixing can occur in the radiative zones of atmospheres through a variety of processes such as the turbulent decay of waves propagated upward from the convective / radiative boundary and instabilities arising from rapid rotation . in the absence of any detailed modeling of these processes ,
we adopt a simple model of vertical mixing by eddy turbulence @xcite .
this process is analogous to diffusion and occurs over a characteristic time scale @xmath32 , where @xmath33 is the pressure scale height and @xmath24 is the coefficient of eddy diffusion .
@xmath24 ranges from @xmath34 to @xmath35@xmath16/s in planetary stratospheres and it is the only free parameter in our non - equilibrium models .
the time scale for mixing in the atmospheric convection zone is @xmath36 where @xmath37 and @xmath38 are the convective mixing length and velocity , respectively , and are evaluated with the mixing length theory . for convenience , we choose @xmath39 .
note that @xmath40 .
the chemical abundance profiles @xcite in model atmosphere b are shown in fig .
3 . depth in the atmosphere is indicated by the local temperature .
the equilibrium abundances are shown with dashed lines that overlap solid lines for h@xmath31o and ch@xmath21 .
the transition in the carbon chemistry from co deep in the atmosphere to ch@xmath21 near the surface occurs at @xmath41 .
the n@xmath31 to nh@xmath5 transition occurs around @xmath42 .
remarkably , and in contrast to the carbon chemistry , the n@xmath31/nh@xmath5 equilibrium ratio remains nearly constant deep in the atmosphere due to a near cancellation of the increased temperature ( favoring n@xmath31 ) and the increased pressure ( favoring nh@xmath5 ) along this portion of the atmosphere profile .
the chemical time scales for the conversion of co and n@xmath31 are shown by the heavy solid and dotted lines , respectively .
these vary over more than 20 orders of magnitude throughout the atmosphere and become extremely long in the upper atmosphere .
mixing occurs rapidly in the convection zone but it is about 7 orders of magnitude slower in the radiative zone for our choice of @xmath43@xmath16/s . in regions where @xmath44 , chemical equilibrium prevails . where @xmath45 , however , abundances of co , ch@xmath21 , n@xmath31 , nh@xmath5 and h@xmath31o will depart from equilibrium ( solid lines ) .
the nonequilibrium mole fractions are approximately determined by their values where @xmath46 . in our calculation
, we adopt the scheme of @xcite to determine this `` quenching level . ''
as can be seen in fig .
3 , there can be more than one level where @xmath46 , resulting in alternating zones in and out of chemical equilibrium . for simplicity
, we consider only the uppermost ( lowest @xmath47 ) crossing .
it turns out that in all cases , this is an excellent approximation to the full solution for the chemical profile since the deeper quenching occurs well below the level of formation of the corresponding molecular bands .
figure 3 shows that above the co quenching level ( @xmath48 ) the nonequilibrium abundance of co is orders of magnitude higher than at equilibrium .
the abundances of ch@xmath21 and h@xmath31o are barely affected for this choice of @xmath24 ( slow mixing ) .
the nh@xmath5 abundance is reduced by about one order of magnitude in the upper atmosphere . in this model
, n@xmath31 is quenched below the bottom of the atmosphere model , in a region where n@xmath31/nh@xmath5 is still nearly constant .
the fact that the equilibrium mole fraction of nh@xmath5 is nearly constant for @xmath49 implies that the non - equilibrium nh@xmath5 abundance profile is completely insensitive to the choice of the eddy mixing coefficient @xmath24 over a very wide range ( @xmath50 ) .
this means that _ all _ plausible choices of @xmath24 will give the same nh@xmath5 abundance and therefore the same mid - ir spectrum . in view of the above remarks , we arbitrarily choose @xmath51@xmath16/s to compute nonequilibrium spectra for the three models in table 1 .
besides the choice of equilibrium or nonequilibrium chemistry , the only fitting parameter left is the gravity , constrained to be between the values for models a and c ( table 1 ) .
we determine the goodness of fit of a synthetic spectrum by computing the @xmath52 between the model and the data for @xmath53 m , which is where nh@xmath5 features are seen in the spectrum ( fig .
2 ) . for this purpose ,
model spectra are smoothed to the irs short low point source spectral resolution of @xmath54 with a gaussian filter and then binned to the instrument s wavelength sampling . the uncertainty ( @xmath55 ) on the fitted @xmath52
is obtained by fitting 5000 simulated data sets generated by adding a random gaussian noise distribution to each pixel with a width given by the individual pixel noise ( @xmath56mjy ) .
we also fit renormalized spectra to account for the @xmath57% calibration uncertainty , which turns out to be the dominant source of uncertainty .
nonequilibrium model c always gives the best fit .
a reasonable fit ( @xmath58 worse than model c ) is obtained with nonequilibrium model b if the observed flux is increased by the full 6.5% uncertainty .
the best equilibrium model fit is @xmath59 worse than the best nonequilibrium model fit .
the nonequilibrium model c spectrum is shown by the dashed ( blue ) curve on fig .
the agreement with the data is remarkable considering that all we have done is include non - equilibrium chemistry to the calculation and increase the gravity slightly , with no other free parameter .
we conclude that the nh@xmath5 abundance is reduced by nonequilibrium chemistry and that the parameters of gl 570d fall between models b and c , and much more likely closer to the latter .
we briefly consider five possible explanations that could account for the enhanced mid - ir flux in gl 570d : a low nitrogen abundance , an error in the temperature profile of the atmosphere , uncertain nh@xmath5 opacity , time variability , and photodissociation of nh@xmath5 .
the metallicity of gl 570d is very nearly solar , based on abundance determinations of na , si , ca , ti , v , cr , fe , co , and ni in gl 570a that range from [ x / h]@xmath60 to 0.24 , with a typical spread of @xmath61 between independent determinations @xcite . to the best of our knowledge
however , the n abundance has not been measured . a low [ n / h ] value would result in a decreased nh@xmath5 abundance that could potentially explain the relatively high mid - ir flux .
the dependance of the mole fraction of ammonia on the nitrogen abundance is @xmath62[n / h ] @xcite . in our models ,
the strongest nh@xmath5 absorption occurs at a depth of @xmath63 in the atmosphere where the non - equilibrium chemistry reduces @xmath64 by a factor of 10 ( fig .
3 ) . to obtain a similar reduction from a low nitrogen abundance would require [ n / h]@xmath65 , which would be an extreme departure from the abundances of other metals in gl 570a . in principle ,
n - bearing condensates may also cause a depletion of n - bearing gases in the atmosphere .
however , there are no n - bearing condensates expected along the @xmath66 profile of gl 570d that can significantly remove nitrogen , and temperatures are not low enough to allow nh@xmath21sh or nh@xmath5 condensation as in the much cooler upper atmospheres of the jovian planets .
our determination of @xmath22 and our solution for @xmath11 are in excellent agreement with the values obtained by @xcite without the benefit of the irs spectrum .
this indicates that while the near - ir flux has remained constant , the unknown mid - ir flux of gl 570d in march 2000 could not be significantly different from that measured in feb 2005 . the fact that our equilibrium model is in very good agreement with the full sed of gl 570d ( fig .
2 ) supports the notion that all pieces of the sed are consistent with each other .
finally , @xmath9 would have to increase to @xmath67k to obtain an equilibrium nh@xmath5 mole fraction similar to the non - equilibrium value needed to fit the irs spectrum .
such a high @xmath9 is completely incompatible with _ all _ of the gl 570d observations .
if we consider the temperature profile of the atmosphere , an increase of over 200@xmath1k ( @xmath68% ) at @xmath69bar would be required ; an implausible error in our models , an inconsitent with the near - ir spectrum of gl 570d .
the opacity of nh@xmath5 is computed from a line list obtained by combining the geisa @xcite and hitran @xcite data bases , complemented with recent laboratory measurements and theoretical calculations ( see @xcite for more details ) . to our knowledge , this compilation of nh@xmath5 opacity is the most complete presently available . nonetheless , the line list is only complete up to @xmath70k and transitions arising from levels that are excited only at higher temperatures are underrepresented or missing .
qualitatively , the nh@xmath5 opacity we calculate for gl 570d is a lower limit to the actual opacity , implying that the derived underabundance of nh@xmath5 is an _
upper _ limit .
more quantitatively , the 914@xmath7 m band of nh@xmath5 arises from the ground state ( @xmath71 ) , with a small overlaping contribution from the @xmath72 excited state ( @xmath73 ) .
the energy differences between other excited vibrational states of nh@xmath5 is too large for the corresponding hot bands to overlap with the @xmath71 band .
we are confident that the line list for the 914@xmath7 m band on which our analysis of the abundance of nh@xmath5 is based is quite complete .
the spectroscopic data analyzed here was gathered over a period of several years : the optical spectrum is from observations taken in march 2000 and feb 2001 @xcite , the near - ir spectrum in march 2000 @xcite and the _ spitzer _ irs spectrum in february 2005 @xcite .
it is conceivable that the luminosity of gl 570d may have varied during this 5 year period , affecting our determination of @xmath9 and @xmath10 and fortuitously result in our excellent fit of the combined spectrum with a model departing from chemical equilibrium .
@xcite do not report any variation in the optical flux of gl 570d between their two observations taken one year apart , and the reported equivalent width of cs i and rb i lines show no variation within 1.5@xmath55 .
the near ir spectrum is calibrated to @xmath74 mko photometry taken in feb 2000 @xcite .
repeat observations in july 2005 are fully consistent ( @xmath75 ) with the original photometry .
the 2mass photometry of gl 570d was obtained in may 1998 . applying the color transformations of @xcite shows agreement to @xmath76 ( @xmath77 ) , @xmath78 ( @xmath33 ) and @xmath75 ( @xmath24 ) with the mko magnitudes .
based primarily on the stability of the near - ir mko photometry , we conclude that any time variability in gl 570d over a period of 7 years is at most at the 2 - 3% level .
this is less than our estimate of a 5.7% uncertainty on @xmath22 .
nevertheless , we consider the possibility that the absolute mid - ir flux may not be representative of the average or typical state of gl 570d .
if we relax the absolute flux calibration of the irs spectrum , do we still see a nh@xmath5 depletion ? for this purpose , we have repeated the fitting procedure described in 5.4 , with the flux calibration adjusted by a factor that minimizes the value of @xmath52 . allowing for the estimated noise in each pixel of the observed spectrum , we find that the non - equilibrium models always fit better than the equilibrium models at the @xmath79 level , with insignificant variation between models a , b and c. nh@xmath5 is a molecule that is easily dissociated by uv photons @xmath80 . since gl 570d is a companion to a main sequence star , it is possible that the uv flux from the primary results in a depletion of nh@xmath5 in gl 570d
. this process was discussed in details by @xcite in the case of another late t dwarf with a companion , gl 229b , where it was found to be completely negligible at the level of the photosphere . by comparison ,
the uv flux of the dk4 star gl 570a is about 10 times larger than that of the dm1 star gl 229a ( based on the nextgen models of @xcite ) but the flux incident on gl 570d is smaller because of the much greater projected separation ( 1525@xmath1au @xcite rather than 44@xmath1au @xcite for gl 229b ) .
since the atmospheres of both brown dwarfs are quite similar , we can expect that the photodissociation of nh@xmath5 in gl 570d will be @xmath81 times that in gl 229b . the gl 570 bc pair of m dwarfs lies about 10 times closer to gl 570a than gl 570d and their contribution to the uv flux incident on gl 570d is a small correction to the above estimate .
our method to determine @xmath9 and the gravity of brown dwarfs is most effective when applied to objects with known parallax and which are members of systems where the primary provides the metallicity and age indicators .
the method is based on integrated fluxes and evolution sequences rather than a direct fit of the spectrum , an approach that largely eliminates biases due to remaining systematic problems in the model spectra . with recalibrated optical and near - ir spectra and , most importantly , a _
irs spectrum , we have redetermined @xmath11 for gl 570d to find nearly the same values as @xcite .
an excellent fit of the irs spectrum can only be obtained if the nh@xmath5 abundance reduced by nearly one order of magnitude in the upper atmosphere .
this is most readily explained by vertical mixing that drives the nitrogen chemistry away from chemical equilibrium @xcite .
we have considered several other possible explanations but none of them are plausible .
we find that in late t dwarfs , the nonequilibrium n@xmath31/nh@xmath5 ratio generally does not depend on the efficiency of mixing in the atmosphere .
therefore , while the 10 11@xmath7 m band of nh@xmath5 is a sensitive indicator of nonequilibrium chemistry in brown dwarf atmospheres , it provides little information about the efficiency of mixing . on the other hand
, the co abundance is very sensitive to the choice of the mixing parameter @xmath24 , and photometric or spectroscopic observations of the 4.7@xmath7 m co band is the most effective way to determine @xmath24 .
our study provides the most accurate parameters of any known brown dwarf ( fig .
1 ) : @xmath0 820@xmath1k , @xmath82 5.23 , @xmath83 0.0853 , @xmath84 to @xmath4 , @xmath85 47 and an age of 3 to 5gyr . the upper limit of this range ( fig .
1 ) is at least @xmath58 more likely than the lower limit .
gl 570d joins gl 229b as the second t dwarf where the effect of nonequilibrium chemistry due to vertical transport has been demonstrated explicitly .
this work is based ( in part ) on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under nasa contract 1407 .
t. roellig and m. marley acknowledge the support of the nasa office of space sciences .
k. lodders is supported by nsf grant ast04 - 06963 .
part of this work was supported by the united states department of energy under contract w-7405-eng-36 , and nasa through the spitzer space telescope fellowship program , through a contract issued by the jet propulsion laboratory , california institute of technology under a contract with nasa .
this research has benefited from the m , l , and t dwarf compendium housed at dwarfarchives.org and maintained by c. r. gelino , j. d. kirkpatrick , and a. j. burgasser .
ackerman , a. s. & marley , m. s. 2001 , , 556 , 872 allard , f. hauschildt , p. h. 1999 , , 525 , 871 allard , f. hauschildt , p. h. , alexander , d. r. , tamanai , a. , & schweitzer , a. 2001 , , 556 , 357 angulo , c. et al .
1999 , nucl .
a , 656 , 3 barshay , s. s. & lewis , j. s. 1978 , icarus , 33 , 593 bzard , b. , lellouch , e. , strobel , d. , maillard , j .- p . & drossart , p. 2002 , icarus , 159 , 95 burgasser , a. j. , kirkpatrick , j. d. , cutri , r. m. , mccallon , h. , kopan , g. , gizis , j. e. , liebert , j. , reid , i. n. , et al .
2000 , , 531 , l57 burgasser , a. j. , kirkpatrick , j. d. , liebert , j. & burrows , a. 2003 , , 594 , 510 burgasser , a. j. , geballe , t. r. , leggett , s. k. , kirkpatrick , j. d. & golimowski , d. a. 2005 , , 637 , 1067 burrows , a. & liebert , j. 1993 , rev .
phys . , 65 , 301 burrows , a. , et al .
1997 , , 491 , 856 burrows , a. , sudarsky , d. , & lunine , j. i. 2003 , , 596 , 587 chabrier , g. & baraffe , i. 1997 , , 327 1039 champion , j. p. , loete , m. & pierre , g. 1992 , in spectroscopy of the earth s atmosphere and the interstellar medium , k. n. rao and a. weber , eds . , 339 cushing , m. c. , roellig , t. r. , marley , m. s. , saumon , d. , leggett , s. k. , kirkpatrick , j. d. , wilson , j. c. , sloan , g. c. , et al .
2006 , submitted to fegley , b. jr . &
lodders , k. 1994 , icarus , 110 , 117 fegley , b. jr . & lodders , k. 1996 , , 472 , l37 fegley , b. jr . & prinn , r. g. 1985 , , 299 , 1067 fortney , j. j. , saumon , d. , marley , m. s. , lodders , k. & freedman , r. s. 2006 , , in press .
freedman , r. s. & lodders , k. 2006 , in prep .
geballe , t. r. , saumon , d. , leggett , s. k. , knapp , g. r. , marley , m. s. & lodders , k. 2001 , , 556 , 373 golimowski , d. a. , burrows , c. j. , kulkarni , s. r. , oppenheimer , b. r. & brukardt , r. a. 1998 , , 115 , 2579 golimowski , d. a. , leggett , s. k. , marley , m. s. , fan , x. , geballe , t. r. , knapp , g. r. , vrba , f. j. , henden , a. a. et al . , 2004 , , 127 , 3516 griffith , c. a. & yelle , r. v. 1999 , , 519 , l85 husson , n. , bonnet , b. , scott , n. a. , & gredin , a. 1992 , , 48 , 509 leggett , s. k. , golimowski , d. a. , fan , x. , geballe , t. r. , knapp , g. r. , brinkmann , j. , csabai , i. , gunn , j. et al . 2002 , , 564 , 452 lewis , j. s. & prinn , r. g. 1980 , , 238 , 357 lodders , k. 2003 , , 591 , 1220 lodders , k. & fegley , b. jr .
1994 , icarus , 112 , 368 lodders , k. & fegley , b. jr .
2002 , icarus , 155 , 393 lodders , k. & fegley , b. jr .
2006 , astrophysics update 2 ( new york : praxis springer ) , 1 marley , m. s. , saumon , d. , guillot , t. , freedman , r. s. , hubbard , w. b. , burrows , a. , & lunine , j. i. 1996 , science , 272 , 1919 noll , k. s. , geballe , t. r. & marley , m. s. 1997 , , 489 , l87 perryman , m. a. c. , et al .
1997 , , 323 , l49 roellig , t. l. , van cleve , j. e. , sloan , g. c. , wilson , j. c. , saumon , d. , leggett , s. k. , marley , m. s.,cushing , m. c. , et al , 2004 , , 154 , 418 rothman , l. s. et al .
2005 , , 96 , 139 santos , n. c. , israelian , g. , mayor , m. bento , j. p. , almeida , p. c. , sousa , s. g. & ecuvillon , a. 2005 , , 437 , 1127 saumon , d. , chabrier , g. & van horn , h. m. 1995 , , 99 , 713 saumon , d. , geballe , t. r. , leggett , s. k. , marley , m.s . , freedman , r. s. , lodders , k. , fegley , b. , jr .
& sengupta , s. k. 2000 , , 541 , 374 saumon , d. , marley , m. s. , lodders , k. & freedman , r. s. 2003 , in brown dwarfs , iau symp .
martn , ed . , ( asp : san francisco ) , 335 smith , m. d. 1998 , icarus , 132 , 176 stephens , d. c. & leggett , s. k. 2004 , , 116 , 9 thorn , p. & feltzing , s. 2000 , , 363 , 692 tsuji , 1964 , annals of the tokyo astronomical observatory , 9,1 valenti , j. a. & fisher , d. a. 2005 , , 159 , 141 visscher , c. , & fegley , b. jr .
2005 , , 623 , 1221 yung , y. l. , drew , w. a. , pinto , j. p. & friel , r. r. 1988 , icarus , 73 , 516 | we present the first analysis of an optical to mid - infrared spectrum of the t7.5 dwarf gliese 570d with model atmospheres , synthetic spectra , and brown dwarf evolution sequences .
we obtain precise values for the basic parameters of gl 570d : @xmath0 820@xmath1k , @xmath2 5.23 , and @xmath3 to @xmath4 .
the _ spitzer _
irs spectrum shows prominent features of ammonia ( nh@xmath5 ) that can only be fitted by reducing the abundance of nh@xmath5 by about one order of magnitude from the value obtained with chemical equilibrium models .
we model departures from chemical equilibrium in the atmosphere of gl 570d by considering the kinetics of nitrogen and carbon chemistry in the presence of vertical mixing .
the resulting model spectrum reproduces the data very well . |
being one of the main sources of information on strangeness changing baryon interactions , the non mesonic weak decay of @xmath1hypernuclei has attracted a great deal of attention for several years . in this decay , the final , observable product is the emission of two or more nucleons from the hypernucleus .
the non mesonic decay width , @xmath8 , is built up from the one and two nucleon induced decays , @xmath9 and @xmath10 , respectively : @xmath11 .
all possible isospin components are : @xmath12 , @xmath13 , @xmath14 , @xmath15 and @xmath16 .
one should note that the only observables in hypernuclear weak decay are the lifetime @xmath17 , the mesonic rates @xmath18 and @xmath19 and the spectra of the emitted particles ( nucleons , pions and photons ) . none of the non mesonic partial decay rates ( @xmath20 , @xmath21 , @xmath22 , etc ) is an observable from a quantum mechanical point of view .
each one of the possible elementary non mesonic decays occurs in the nuclear environment , thus subsequent final state interactions ( fsi ) modify the quantum numbers of the weak decay nucleons and new , secondary nucleons are emitted as well : this prevents the measurement of any of the non mesonic partial decay rates . among the weak decay observables one has the spectra of emitted neutrons ( @xmath23 ) and protons ( @xmath24 ) as well as the @xmath25 and @xmath26 nucleon coincidence spectra , @xmath27 and @xmath28 . in the present contribution
we discuss a microscopic model to take care of the @xmath0 ratio and of the single
nucleon spectra for the emission of protons and neutrons .
we report results for @xmath2c .
the paper is organized as follows : in section [ form ] we outline the formalism , results for @xmath2c are discussed in section [ results ] and finally , in section [ conclu ] , some conclusions are given .
using the formalism developed in @xcite , the number of particles , @xmath29 , and of pair of particles , @xmath30 , where @xmath31 or @xmath32 , emitted in non mesonic decay can be written as follows : @xmath33 all terms which contain the functions @xmath34 represent the action of fsi .
we refer the reader to @xcite for a complete explanation of the meaning of these functions .
the factors @xmath35 are the numbers of nucleons of the type @xmath36 contained in the multinucleon state @xmath37 . in the same way ,
the @xmath38 s are the numbers of @xmath39 nucleon pairs in the state @xmath37 .
the summation over @xmath40 and @xmath37 runs over all possible initial ( @xmath41 and @xmath42 ) and final ( @xmath37 ) states .
we have used a normalization for which @xmath43 .
the five quantities @xmath29 and @xmath30 of eqs .
( [ nn1f])([npp1f ] ) are observables .
five decay widths , @xmath44 and @xmath45 , also enter into these equations . the fact that fsi can not be neglected makes it impossible to invert the above relations in order to obtain expressions for the decay widths in terms of the observables @xmath29 and @xmath30 . by starting from the experimental values of @xmath29 and @xmath30 it is possible to extract the
so called experimental values for the decay widths , once a model for the fsi is implemented .
this is done in the hope that several models for fsi will lead to the same values for the extracted decay widths . in this contribution
we adopt a different working method . instead of extracting experimental values for the decay widths , we work out a microscopic model to report results for @xmath29 and @xmath30 given by eqs .
( [ nn1f])([npp1f ] ) .
fsi are modeled by the set of feynman diagrams of fig .
it should be noted that by considering all the possible time
orderings of these diagrams , some of the corresponding goldstone diagrams are pure gsc contributions ( to @xmath46 ) , others are pure fsi terms and others are interferences between gsc and fsi contributions .
pauli exchange terms have not been evaluated . and
@xmath47 , respectively . ]
here we report results on the coincidence number ratio , @xmath0 , and on the single proton and single
neutron emission spectra , @xmath3 and @xmath4 , as a function of the nucleon kinetic energy .
the calculation is performed in nuclear matter ; by the local density approximation we obtain results for @xmath2c .
the weak transition potential , @xmath48 , is represented by the exchange of @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 and @xmath54 mesons , with the coupling constants and cut off parameters deduced from the nijmegen soft core interaction nsc97f @xcite . for the residual strong interaction , @xmath55
, we use a @xmath56 potential with the addition of a @xmath57 landau migdal parameter , with @xmath58 . in table
[ table1 ] we present our results for the @xmath0 ratio .
the inclusion of fsi brings to a good agreement with data ( without fsi , one would have @xmath59 ) .
conversely , the effect of the quantum interference terms ( which correspond to the diagrams @xmath60 , @xmath61 , @xmath62 and @xmath63 in fig . [ fig1 ] ) is not important in the determination of @xmath0 .
however , from table [ table2 ] we see that the individual values of the interference terms are sizable in comparison with the dominant ones ( @xmath64 ) in both @xmath27 and @xmath28 . from table
[ table2 ] we also see that the reduction due to the energy and angular cuts suffered by the fsi terms is much bigger than the reduction for the fsi free , 1n ind term .
this is expected , as fsi give a dominant contribution in the low energy region .
the single
nucleon spectra obtained with the microscopic model are shown in fig .
it is clear that an improvement has been obtained with respect to the fsi free predictions .
we observe that it is the action of the residual strong interaction ( through fsi and gsc ) the mechanism which could lead , once additional many
body terms are considered , to an agreement with the experimental spectra . ) and single neutron ( @xmath23 ) kinetic energy spectra for @xmath65 . with continuous ( dashed ) lines
we give our results with ( without ) fsi .
data are from kek e508 @xcite . ]
from the experimental spectra of nucleons produced in hypernuclear non mesonic weak decay it is not straightforward to determine the decay widths @xmath66 , @xmath67 , @xmath22 , etc : a theoretical model for fsi is required . in this contribution , a quantum mechanical microscopic model for nucleon emission has been discussed .
it has been shown that with this approach it is possible to obtain a rather good agreement with experiment for the @xmath0 ratio . however , we note that one should also reproduce the individual spectra @xmath27 , @xmath28 , @xmath23 , @xmath24 , etc . from a comparison with data ,
the predictions of our microscopic model for fsi shows a clear improvement over the fsi free results .
our results suggest that the interference terms are not important .
the present approach requires further improvements : @xmath68 a more realistic residual strong interaction , @xmath69 the inclusion of pauli exchange terms for the fsi diagrams and @xmath70 the study of the effect of the @xmath71 .
finally , we note that the microscopic model allows us to set constraints not only on @xmath48 but also on @xmath47 .
one of us ( eb ) would like to thank the hyp - x organizing committee for their support to attend the conference . | the ratio @xmath0 between the number of neutron neutron and neutron proton pairs emitted in the non mesonic weak decay of @xmath1hypernuclei is calculated within a nuclear matter formalism extended to @xmath2c via the local density approximation . the single
nucleon emission spectra , @xmath3 and @xmath4 , are also evaluated .
our formalism takes care of both ground state correlations ( gsc ) and final state interactions ( fsi ) .
the evaluation of @xmath0 which , unlike @xmath5 , is an actual observable quantity in non mesonic decay is performed within a fully microscopic model where a proper treatment of fsi , gsc and ground state normalization is considered .
all the isospin channels contributing to one and two nucleon induced decays are included .
our final result for the coincidence number ratio , @xmath6 , is in agreement with the kek
e508 datum , @xmath7 .
@xmath1hypernuclei , non mesonic weak decay , two
nucleon induced decay , fsi 21.80.+a , 25.80.pw |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.